mechanisms and clinical implications of the neuroendocrine … · mechanism by which the observed...
TRANSCRIPT
Mechanisms and clinical implications of the
neuroendocrine response to a novel carbon dioxide
stressor in man
Dr Joey Michael Kaye MBBS FRACP
This thesis is presented for the degree of Doctor of
Philosophy
The University of Western Australia
School of Medicine and Pharmacology
2005
ii
DEDICATION
To my wife - CMK
iii
ABSTRACT
Maintenance of normal health requires an intact stress system capable of
mounting the metabolic, autonomic, behavioural and motor responses required
for coping with or avoiding physiological and pathological challenges. The
neuroendocrine component of this response principally involves the
hypothalamic-pituitary-adrenal (HPA) and sympatho-adrenomedullary (SAM)
axes. Impaired regulation of these axes has been implicated in the pathogenesis
and expression of numerous disease states, however, it has proved very difficult
to reproducibly activate the HPA and SAM axes and no single test exists that can
reliably and safely be used to study these systems in man.
Carbon dioxide (CO2) is the principal regulator of respiration, acid-base balance
and behavioural-state arousal in humans. Paradigms of CO2 inhalation have been
used in psychiatric research to investigate panic and anxiety disorders, but
evaluation of other components of the stress response to CO2 has not previously
been performed. I hypothesised that a single breath of 35% CO2 would be a
simple and reliable tool for the evaluation of the stress response in humans. A
single breath of four doses of CO2 (5%, 25%, 35% and 50%) was administered to
9 healthy volunteers in a randomised, single blind fashion. Subjective symptoms
of anxiety increased in a dose-dependent manner. Inhalation of a single breath of
35% CO2 stimulated significant ACTH (p = 0.006), noradrenaline (p < 0.0001),
cortisol (p = 0.02) and prolactin (p = 0.002) release. It also provoked an acute
pressor response and an associated bradycardia (p < 0.0001 for both). No
significant habituation of psychological, HPA or cardiovascular responses was
iv
seen when this dose was repeated after one week (n = 10) or 6 months (n = 5). It
was apparent that a single breath of 35% CO2 reliably and safely produced SAM
and HPA axis activation and further studies were then undertaken to assess the
mechanism by which the observed responses occurred and its potential clinical
implications.
Administration of naltrexone (an opiate antagonist) to 10 normal volunteers
disinhibited the HPA axis (p < 0.0004), whilst administration of metyrapone (a
cortisol synthesis inhibitor) significantly reduced baseline cortisol (p < 0.03)
levels. However, this alteration in HPA axis activity had no effect on either
cardiovascular or psychological responses. Further, in a study of 8 breastfeeding
mothers (a state associated with physiological suppression of the HPA axis)
suckling significantly reduced plasma cortisol levels compared with control (p =
0.002) and bottle-feeders (p = 0.003). Despite this cortisol, systolic blood
pressure (SBP), heart rate and psychological responses to 35% CO2 were not
affected.
Continuous cardiovascular monitoring of subjects immediately following a single
breath of 35% CO2 identified bradycardia as the first response followed by an
acute pressor response and an increase in noradrenaline. In a study of 20 male
diabetic subjects, 11 with autonomic neuropathy (with early parasympathetic
dysfunction), the autonomic neuropathy subjects failed to demonstrate any CO2-
induced bradycardia (p < 0.0001) suggesting an early effect of autonomic
neuropathy on cardiac vagal innervation. Further studies of patients with
abnormalities of central CO2 chemoreceptor activation and integration identified
v
dissociation between the sympathetic (pressor response) and parasympathetic
(bradycardic response) components of the response.
The role of central autonomic centres in determining the overall response to CO2
was addressed in a study of 9 patients with multiple systems atrophy (MSA) and
9 with pure autonomic failure (PAF) (central vs peripheral with autonomic
failure). Resting noradrenaline was significantly lower in PAF (p < 0.0001).
Following the CO2 challenge SBP increased in both PAF (p < 0.001) and MSA
(p = 0.002), although this increase was significantly blunted and delayed
compared with controls. Cortisol responses were lower in MSA patients who also
experienced fewer somatic symptoms of fear. Finally, control subjects showed a
reduction in skin blood flow that was blunted in MSA subjects, whilst PAF
subjects showed a striking increase in skin blood flow (p = 0.006). This increase
was due to non-neurally mediated vasodilation produced by CO2 that is normally
masked by the centrally-mediated vasoconstriction. In addition to providing
insight into the mechanism of CO2 mediated physiological changes this
difference may prove clinically important in differentiating the autonomic failure
syndromes.
In summary, it appears that the SAM response to the 35% CO2 challenge is
driven by activation of brainstem autonomic centres with secondary activation of
the HPA axis. The vagal response occurs independently whilst the psychological
response occurs probably as a result of both direct activation of limbic centres by
CO2 as well as indirect activation from brainstem sympathetic systems.
vi
The single breath 35% CO2 inhalation test has shown itself to be a safe, reliable
and reproducible means of generating a stress response in humans and should
prove a useful addition to currently available laboratory tests of the human stress
response.
vii
TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION……………………………………………...1
1.1 The neuroendocrine stress response in man…………………………..….2
1.1.1 Clinical importance………………………………………………...2
1.1.2 Background………………………………………………………...7
1.1.2.1 The HPA axis……………………………………………….12
1.1.2.2 AVP………………………………………………………...16
1.1.2.3 The sympatho-neural and sympatho-adrenomedullary
axis………………………………………………………….17
1.1.2.4 Prolactin…………………………………………………….19
1.1.2.5 Opiates…………………………………...…………………20
1.1.2.6 Serotonin (5-HT)……………………………………...……21
1.1.2.7 The growth, thyroid and reproductive axes…………...……22
1.1.2.8 Other neurohormonal systems…………….………………..23
1.2. Investigating the stress response………………………………………...25
1.2.1. Currently available tools and their limitations………………...….25
1.2.1.1. Psychological challenges…………………………………...26
1.2.1.2. Physiological challenges……………………………………28
1.2.1.3. Pharmacological challenges………………………………...30
1.2.2. CO2 as a potential neuroendocrine stressor……………………….33
1.3. Summary, aims and hypotheses………….………………………...……35
viii
CHAPTER 2: METHODOLOGY…………………………………..………..37
2.1 The single breath 35% CO2 model………………………………...……38
2.1.1 Design…………………………………………………………….42
2.1.2 Procedure…………………………………………………………45
2.1.3 Inclusion/Exclusion criteria………………………………………47
2.1.4 Monitoring…………………………….………………………….49
2.1.4.1 Ventilation………………………………………………….49
2.1.4.2 Psychology…………………………………………………50
2.1.4.3 Cardiovascular physiology…………………………………53
2.1.4.4 Biochemistry……………………………………………….55
2.2 Plasma catecholamine HPLC…………………………………………..63
2.3 Salivary amylase……………………………………………………….73
CHAPTER 3: THE 35% CO2 MODEL: INITIAL DESCRIPTION………78
3.1 Rationale for the use of CO2 as a neuroendocrine stressor……………..79
3.1.1 CO2 physiology…………………………………………………..80
3.1.2 CO2 as a psychological stressor…………………………………..81
3.1.3 CO2 as a cardiovascular stressor………………………………….83
3.1.4 CO2 as a neurohormonal stressor…………………………………84
3.2 Experimental plan……………………………………………………….86
3.3 Dose response study…………………………………………………….87
3.3.1 Introduction and methods…………………………………………87
3.3.2 Results…………………………………………………………….88
3.3.3 Discussion…………………………………………...……………97
ix
3.4 Reproducibility studies………………………………………………...103
3.4.1 Introduction and methods………………………………………..103
3.4.2 Results…………………………………………………………...104
3.4.3 Discussion…………………………………………...…………..112
3.5 Summary………………………………………………………………113
3.6 Conclusion…………………………………………………………….119
CHAPTER 4: THE 35% CO2 MODEL: MECHANISMS UNDERLYING
THE NEUROENDOCRINE RESPOSNE………………………………….121
4.1 Introduction……………………………………………………...…….122
4.2 Experimental plan……………………………………………………...126
4.3 Central neurotransmitter study………………………...………………127
4.3.1 Introduction and methods…………….………………………..127
4.3.2 Results…………………………………………………………129
4.3.3 Discussion…………………………………………...…………137
4.4 Mineralocorticoid/glucocorticoid study……………………………….143
4.4.1 Introduction and methods…………….………………………..143
4.4.2 Results…………………………………………………………146
4.4.3 Discussion…………………………………………...…………152
4.5 Lactation study……………………………………...…………………154
4.5.1 Introduction and methods…………….………………………..154
4.5.2 Results…………………………………………………………158
4.5.3 Discussion…………………………………………...…………165
4.6 Peripheral versus central autonomic nervous system effects………….167
4.6.1 Introduction and methods…………….………………………..167
x
4.6.2 Results…………………………………………………………168
4.6.3 Discussion…………………………………………...…………172
4.7 Conclusion……………………………………………………………..176
CHAPTER 5: THE 35% CO2 MODEL: RESPONSES IN SPECIFIC
SUBPOPULATIONS – FURTHER MECHANISMS AND POTENTIAL
CLINICAL RELEVANCE…………………………………………………180
5.1 The role of the HPA axis …………………………………...…………181
5.2 Addison’s disease……………………………………………...………185
5.2.1 Introduction and methods…………….………………………..185
5.2.2 Results…………………………………………………………188
5.2.3 Discussion…………………………………………...…………193
5.3 The role of the autonomic nervous system…………………………….196
5.4 Experimental plan……………………………………………………...199
5.5 Diabetic autonomic neuropathy………………………………………..200
5.5.1 Introduction and methods…………….………………………..200
5.5.2 Results…………………………………………………………206
5.5.3 Discussion…………………………………………...…………212
5.6 Chronic autonomic failure syndromes…………………………………216
5.6.1 Introduction and methods…………….………………………..216
5.6.2 Results…………………………………………………………223
5.6.3 Discussion…………………………………………...…………235
5.7 Conclusion……………………………………………………………..241
xi
CHAPTER 6: SUMMARY, CONCLUSION AND FUTURE
DIRECTIONS………………………………………………………………...245
6.1 Summary …………………………………………………...………….246
6.2 Conclusions and future directions……………………………………..252
CHAPTER 7: REFERENCES………………………………………………254
APPENDICES……………………………………………………………..….273
Appendix 1 Visual Analogue Scale
Appendix 2 Somatic Symptom Questionnaire
Appendix 3 Edinburgh Post-Natal Depression Score Questionnaire
Appendix 4 Protocol for the extraction of catecholamines #1
Appendix 5 Protocol for the extraction of catecholamines #2
Appendix 6 Protocol for the extraction of catecholamines #3
Appendix 7 Protocol for the extraction of catecholamines #4
LIST OF TABLES
Table 3.01 Dose response study - baseline characteristics……………….…91
Table 3.02 Diurnal variability study - baseline characteristics…………….109
Table 4.01 Neurotransmitter study - baseline characteristics…………...…131
Table 4.02 Lactation study - baseline characteristics………………….…..160
Table 5.01 Addison’s disease study - baseline characteristics………….…190
Table 5.02 Diabetic autonomic neuropathy study –
baseline characteristics……………………………………...…208
Table 5.03 Autonomic failure - baseline characteristics………………..…227
xii
Table 5.04 Autonomic failure – hormonal and cardiovascular
responses………………………………………………….……229
Table 5.05 Autonomic failure – Skin blood flow response…………….….232
LIST OF FIGURES
Figure 2.01 Schematic of the breathing circuit used to deliver
the CO2 breath………………………………………………..…44
Figure 2.02 Arrangement for patient monitoring……………………….……58
Figure 2.03 Arrangement for patient monitoring…………………………….59
Figure 2.04 Arrangement for patient monitoring…………………………….60
Figure 2.05 Arrangement for CO2 delivery………………………………….61
Figure 2.06 Arrangement for CO2 delivery………………………………….62
Figure 2.07 Chromatogram – unextracted standard………………...….…….67
Figure 2.08 Chromatogram – successful extraction………………………….68
Figure 2.09 Chromatogram – successful extraction………………………….69
Figure 2.10 Chromatogram – poor recovery……...………………………….70
Figure 2.11 Chromatogram – unsuccessful extraction……………………….71
Figure 2.12 Chromatogram – unsuccessful extraction……………………….72
Figure 3.01 Dose response study – psychological responses………………...92
Figure 3.02 Dose response study – cardiovascular responses…………...…..93
Figure 3.03 Dose response study – hormonal responses……………...……..94
Figure 3.04 Dose response study – hormonal responses……………...……..95
Figure 3.05 Dose response study – correlations……………………………..96
Figure 3.06 Repeatability study – responses at 1 week and at 6 months...…106
xiii
Figure 3.07 Repeatability study – cardiovascular responses at 1 week
and at 6 months……………………………...…………………107
Figure 3.08 Salivary amylase response and correlations…………...………108
Figure 3.09 Diurnal variability study – morning
versus evening responses………………………………..……..110
Figure 3.10 Diurnal variability study – cardiovascular responses……...…..111
Figure 4.01 Overview of the neuroanatomical stress pathways……………125
Figure 4.02 Neurotransmitter study – baseline cortisol and prolactin……..132
Figure 4.03 Neurotransmitter study – individual hormonal responses….….133
Figure 4.04 Neurotransmitter study – peak hormonal responses………..….134
Figure 4.05 Neurotransmitter study – cardiovascular responses………..….135
Figure 4.06 Neurotransmitter study – psychological responses………...….136
Figure 4.07 Glucocorticoid/mineralocorticoid study – baseline cortisol.….148
Figure 4.08 Glucocorticoid/mineralocorticoid study – cortisol response….149
Figure 4.09 Glucocorticoid/mineralocorticoid study –
psychological responses…………………………………….…150
Figure 4.10 Glucocorticoid/mineralocorticoid study - cardiovascular
responses…………………………………………………….…151
Figure 4.11 Lactation study – baseline and peak cortisol response………...161
Figure 4.12 Lactation study – baseline and peak prolactin response……….162
Figure 4.13 Lactation study – cardiovascular responses……………………163
Figure 4.14 Lactation study – psychological responses……………….……164
Figure 4.15 Peripheral vs central neuropathy –
cortisol and prolactin response………………………..……….169
xiv
Figure 4.16 Peripheral vs central neuropathy –
cardiovascular responses…………………………………..…..170
Figure 4.17 Peripheral vs central neuropathy - psychological responses…..171
Figure 4.18 Principal components of the CO2 mediated stress response…...179
Figure 5.01 Addison’s disease study - cardiovascular responses…………..191
Figure 5.02 Addison’s disease study - psychological responses………..….192
Figure 5.03 Diabetic autonomic neuropathy –
cortisol and prolactin response………………….……………..209
Figure 5.04 Diabetic autonomic neuropathy - cardiovascular responses…..210
Figure 5.05 Diabetic autonomic neuropathy - psychological responses…...211
Figure 5.06 Calculation for change in skin blood flow…………………….222
Figure 5.07 Autonomic failure - cardiovascular responses…………..…….228
Figure 5.08 Autonomic failure – hormone responses………………..…….230
Figure 5.09 Autonomic failure - psychological responses…………...…….231
Figure 5.10 Autonomic failure – skin blood flow response from single
subjects…………..…………………………………………….233
Figure 5.11 Autonomic failure – mean skin blood flow response………….234
Figure 5.12 Principal components of the CO2 mediated stress response -
revised………………………………………………………….244
xv
ACKNOWLEDGMENTS
The advice, support, wisdom and encouragement of many people has contributed
to the success of this project. Principally, Professor Stafford Lightman, for
giving me the opportunity to work and study alongside the extraordinary
members of the Henry Wellcome Laboratories for Integrative Neuroscience and
Endocrinology as well as the wonderful staff and patients of the Bristol Royal
Infirmary. His patience, encouragement and never-failing enthusiasm for this
project has ensured its success and has provided me with an appreciation of
research performed at the highest level. I am deeply grateful for all his support
and direction, but particularly for showing me that things are never quite so
difficult or frustrating if approached with just the right amount of humour, wit
and intelligence.
To Professor Peter Leedman for encouraging me to do this in the first place. To
Dr Chris Lowry, Phil Johnson, Jacob Hollis, David Knight and all the members
of the Integrative Neuroscience and Endocrinology laboratories for making me
feel at home and for their guidance, patience and direction particularly when not
everything went to plan. To Moira Hunt who gave so much of her time and
experience to assisting me with the clinical experiments.
I would also like to acknowledge the generous support of many members of the
University of Bristol and the Bristol Royal Infirmary whose time and expertise
has contributed greatly to this project.
xvi
Professor David Nutt and Jayne Bailey for sharing their expertise and experience
with their CO2 model. Dr Adrian Kendrick and Fiona Buchanan for all their time
and assistance in modifying and optimising the CO2 delivery system. Professor
Peter Soothill and Dr Roger Corrall for their willingness to share their knowledge
and expertise. The nursing staff of the Antenatal Day Assessment Unit at St
Michael’s Hospital and the nursing staff of the Diabetes Outpatient Clinic at the
BRI for their advice, patience and good humour. Dr Rob Andrews who helped in
the design, organise and run the studies in patients with Addison’s disease.
To the staff and patients of the Autonomic Unit at the National Hospital for
Neurology and Neurosurgery, Queen Square, London, including Professor Chris
Mathias, Katherine Bleasedale-Barr and Laura Watson for your willingness and
support in expanding our work into your midst. I’m particularly grateful to Dr
Tim Young for his thoughtfulness and unselfish support of this project. For
dedicating himself to its success and for giving up so much of his time and
energy.
This project was generously supported by the University of Western Australia
through a FA Hadley Travelling Fellowship, the Neuroendocrinology Charitable
Trust, UK and the Charitable Trustees of the United Bristol Healthcare Trust.
To all the people of Bristol, who welcomed us, supported us and nurtured us,
thank you for giving us some of the most enjoyable years of our lives.
xvii
Finally and most importantly to my wife, Carolyn, whose unerring love and
dedication through some of our very best and very worst times, has given my life
a sense of purpose and meaning and who, with our children, has made it all
worthwhile.
Joey Kaye
xviii
PERSONAL CONTRIBUTION OF THE CANDIDATE The candidate was involved in the discussions and decision-making with the
other principal investigators regarding the development of the concept of using
inhaled CO2 as a challenge for the investigation of the human stress response.
The candidate undertook the design, development and modification of the
methodology for the administration of the CO2 as well as the protocols for
monitoring and recording of responses. With regard the test procedure, the
candidate was responsible for:
• Administration of all of the CO2 challenges (apart from the study performed
at the National Hospital for Neurology and Neurosurgery where Dr Tim
Young performed some challenges after instruction by the candidate).
• Collection, alliquoting and storage of all biochemical samples.
• Collection and recording of all physiological responses.
• Administration and recording of all psychological response questionnaires.
• Quality control, identifying and documenting any adverse responses for
safety and tolerability.
With regards the individual studies, the candidate was responsible for:
• The design of each study.
• All patient/subject recruitment (assisted by Dr Tim Young for the autonomic
neuropathy study and Dr Rob Andrews for the Addison’s study).
• Administration of all patient/subject information sheets and consent forms.
• Storage and dispensing all administered medication/placebo capsules.
xix
• All data collection, storage and statistical analysis with advice from Dr
Valerie Burke, statistician with the School of Medicine and Pharmacology,
UWA.
• The administration and interpretation of all the autonomic function studies in
the diabetic autonomic neuropathy study.
With regard biochemical analyses, the candidate was responsible for:
• Establishing and performing a high performance liquid chromatography
(HPLC) assay for plasma catecholamines.
• Performing all ACTH radioimmunoassays for the dose response study.
• Storing and arranging the analysis of other blood and saliva samples.
xx
ABBREVIATIONS
5-HT Serotonin ACEI Angiotensin Converting Enzyme Inhibitor ACTH Adrenocorticotropic Hormone AN Autonomic Neuropathy Ang II Antagonist Angiotensin II Receptor Antagonist ANOVA Analysis of Variance AVP Arginine Vasopressin BF Breast Feeding BO Bottle Feeding BP Blood Pressure C Control CCHS Congenital Central Hypoventilation Syndrome CCK-4 Cholecystokinin 4 CnA Central Nucleus of the Amygdala CNS Central Nervous System CO Cardiac Output CO2 Carbon Dioxide CRH Corticotropin Releasing Hormone CRH-R1 Corticotropin Releasing Hormone Receptor 1 CRH-R2 Corticotropin Releasing Hormone Receptor 2 CV Co-efficient of Variation DAN Diabetic Autonomic Neuropathy DBP Diastolic Blood Pressure DHBA 3,4 dihydroxybezylamine hydrogen bromide DMNX Dorsal Motor Nucleus of the Vagus E/I ratio Expiration / Inspiration ratio ECG Electrocardiogram EDTA Ethylenediaminetetraacetic Acid EGTA Ethylene Glycol-bis N,N,N,N-Tetraacetic Acid ELISA Enzyme Linked Immunoabsorbent Assay EPDS Edinburgh Post-natal Depression Score FSH Follicle Stimulating Hormone GABA Gamma Aminobutyric Acid GH Growth Hormone GR Glucocorticoid Receptor HbA1c Glycosolated Haemoglobin HPA Hypothalamic-Pituitary-Adrenal HPLC High Performance Liquid Chromatography HR Heart Rate IGF-1 Insulin-like Growth Factor 1 IL-1 Interleukin 1 IL-6 Interleukin 6 ITT Insulin Tolerance Test KH2PO4 Potassium dihydrogen orthophosphate dihydrite LC Locus Coeruleus LH Leutinising Hormone LiHep Lithium Heparin
xxi
MHPG 3-Methoxy-4-Hydrophenylglycol MR Mineralocorticoid Receptor MRI Magnetic Resonance Image mRNA Messenger Ribonucleic Acid MSA Multiple Systems Atrophy NaCl Sodium Chloride NPY Neuropeptide Y NREM Non-Rapid Eye Movement O2 Oxygen PaCO2 Arterial Partial Pressure of carbon Dioxide PAF Pure Autonomic Failure PD Parkinson's Disease PNMT Phenoxyethanolamine-N-methyltransferase PNS Parasympathetic Nervous System PRA Plasma Renin Activity PU Perfusion units PVN Paraventricular Nucleus RIA Radioimmunoassay SAM Sympatho-adrenomedullay SBP Systolic Blood Pressure SEM Standard Error of the Mean SSQ Somatic Symptom Questionnaire SSRI Selective Serotonin Reuptake Inhibitor SV Stroke Volume T3 Tri-iodothyronine T4 Thyroxine TNF-α Tumour Necrosis Factor - alpha TPR Total Peripheral Resistance TRH Thyrotropin Releasing Hormone TSH Thyroid Stimulating Hormone TSST Trier Social Stress Test VAS Visual Analogue Scale VC Vital Capacity VLM Venterolateral Medulla VO2max Maximum Oxygen Consumption
xxii
PUBLICATIONS ASSOCIATED WITH THIS THESIS
1. KAYE JM, Buchanan F, Kendrick A, Johnson P, Lowry C, Bailey J, Nutt D
and Lightman S. (2004) Acute carbon dioxide exposure in healthy adults:
Evaluation of a novel means of investigating the stress response. J
Neuroendocrinol 16 : 256-264
2. KAYE JM, Hunt M, Soothill, P and Lightman S. (2004) Responses to the
35% CO2 challenge in post-partum women. Clinical Endocrinology 61 :
582-588
3. KAYE JM, Corrall R and Lightman S. (2005) A new test for autonomic and
neuroendocrine responses in diabetes mellitus: evidence for early vagal
dysfunction. Diabetologia 48 (1) : 180-186
4. KAYE JM and Lightman S. Endocrine response to psychological stress.
(2003) In: An Introduction to Human Psychoneuroimmunology. (Vedhara K,
Ramakalawan T, Tallon D and Johnston M. Eds) Oxford University Press
2005
5. KAYE JM, Young TM, Mathias CJ, Watson L and Lightman SL. (2005)
Neuroendocrine and Behavioural Responses to CO2 Inhalation in Central
versus Peripheral Autonomic Failure. Neurology (in press)
xxiii
6. Young TM, KAYE JM, Lightman SL, Mathias CJ. Acute effects of CO2
inhalation on systemic and regional blood haemodynamics in two forms of
sympathetic denervation: multiple system atrophy and pure autonomic
failure. J Physiology (submitted July 2005)
CHAPTER 1
INTRODUCTION
1
1.1 The neuroendocrine stress response in man
1.1.1. Clinical Importance
Throughout the history of medicine, reference has been made to the influence of stress,
particularly in the form of negative emotions and psychological distress, on physical
health [Sternberg 1997]. However, it has only been in the last few decades that clear
scientific evidence supporting such a notion has been forthcoming. In numerous recent
epidemiological surveys, stress and impaired psychological functioning have been
shown to be associated with an increased prevalence of serious medical conditions.
These have included psychiatric conditions such as depression and post-traumatic stress
disorder [Checkley 1996, Chrousos and Gold 1998, Gold et al 1998, Ehlert et al 2001],
vascular disease such as coronary heart disease [Rozanski et al 1999, Krantz and
McCeney 2002], immune-mediated conditions including asthma [Sternberg et al 1992a,
Wright et al 1998] and other conditions such as osteoporosis, diabetes, dementia and
premature death [Chrousos 2000, Mathe 2000]. Why some individuals manifest stress
as psychiatric illness, whilst others are more prone to physical disease and yet others
seem resistant to the effects of stress exposure is not well understood. The
neurobiological pathways that underlie these processes and the factors that contribute to
the expression of disease within an individual are still being explored. One of the most
important factors limiting the investigation of stress and its consequences is the lack of
a universally accepted scientific definition of stress [McEwen 1998]. Whilst stress as a
concept is understandable to most, an accepted definition remains elusive [Harbuz and
Lightman 1992]. Noble [2002], using Stedman’s Medical Dictionary [Stedman 1990],
defines stress as the ‘reactions of the body to forces of a deleterious nature, infection,
and various abnormal states that tend to disrupt normal physiologic equilibrium
2
(homeostasis)’. In practical terms, he then describes three methods that can be used to
measure the body’s response to stress. These include: questionnaires such as the Life
Stress Inventory [Boone and Christensen 1998] and the Perceived Stress Scale [Cohen
et al 1983]; biochemical measures particularly cortisol and catecholamines; and
cardiovascular responses including blood pressure, heart rate and heart rate variability.
Chrousos [1998] also defines stress as ‘a state of threatened homeostasis’ with those
factors, both intrinsic and extrinsic, that are challenging this state termed stressors and
the complex adaptive physiologic, hormonal and behavioural responses that occur to
restore homeostasis is the stress response. By implication, stress is ultimately
damaging with negative consequences for the individual in whom it is occurring.
McEwen [2000] on the other hand, highlights the protective role of the stress response
in the short term, and the importance of associated learning and adaptation (a process
that requires plasticity of brain responses) following stress exposure to the longer term
health and survival of the individual. It is only when these responses occur in excess of
the body’s requirements, or continue for longer than is necessary then do damaging
effects result [McEwen 1994, McEwen 1998, McEwen 2000]. Moreover, stress
includes all of the many day-to-day events of normal life that result in activation of one
or more of the components of the stress response system. This process of continual
activity of the stress response system in order to maintain homeostatic equilibrium has
been referred to as allostasis [McEwen 2000]. The cost of this process, or allostatic
load, reflects a state of wear and tear for the individual. The extent of this load is
determined by genetic make-up, co-morbid illness, individual habit such as diet,
substance abuse and physical fitness and developmental experiences that set life-long
patterns of behavioural and physiological reactivity [McEwen 1998, McEwen 2000
Goldstein and McEwen 2002, McEwen 2002].
3
Repeated exposure to a particular stressor typically results in subsequent stress
responses that are smaller in amplitude and shorter in duration – a process of learning
and adaptation that minimises the impact of repeated exposure to the same stressor
[McEwen 2000]. Increasing allostatic load with its damaging consequences can result
from repeated stress; failure to undergo adaptation; prolonged or excessive responses; or
an inadequate initial response with or without a compensatory increase in the activity of
other systems [McEwen 2000]. For example, as described by McEwen [2000],
repeated stress that causes frequent surges in blood pressure and catecholamine release
is associated with accelerated atherosclerosis and an increased risk of myocardial
infarction [Rozanski et al 1999, Krantz and McCeney 2002].
Impaired adaptation or failure to appropriately terminate the stress response results in
chronic hyperactivity of the stress response systems. The most widely described
example of this is melancholic depression that is associated with overactivation of one
of the major stress response systems – the HPA axis [Chrousos 1998, Gold et al 1998,
Gold and Chrousos 1999]. This state of excessive cortisol exposure is associated with
marked hippocampal atrophy an area important in the formation of emotional memories
[McEwen 1998, McEwen 2002]. It is also increasingly apparent that HPA axis
dysregulation appears well before clinical symptomatology and is a predictor of
treatment resistance. Similarly, failure to normalise HPA axis responses with treatment
is a strong predictor of relapse [Holsboer 2000]. The hypercortisolaemic state also
appears to promote metabolic features seen more overtly in patients with Cushing’s
disease. Glucocorticoids regulate adipocyte differentiation and stress-induced excess
cortisol is associated with increased abdominal fat accumulation [Rosmond et al 1998,
Miller and O'Callaghan 2002]. Visceral adiposity is associated with the metabolic
syndrome and includes insulin resistance, hypertension, dyslipidaemia and premature
4
atherosclerosis [Chrousos 2000, Miller and O'Callaghan 2002]. Further, this syndrome
is epidemiologically linked to increased rates of cardiovascular and cerebrovascular
disease [Chrousos 1998, Chrousos 2000, Miller and O'Callaghan 2002]. Others have
also described increased bone loss with an increased predisposition to osteoporosis as
well as immune dysfunction and an increased risk of infectious and neoplastic disease
[Chrousos 2000, McEwen 2002, Miller and O'Callaghan 2002]. Indeed patients with
depression that is associated with chronic hyperactivity of the HPA axis have been
shown to have a reduced life expectancy predominantly as a result of an excess of
cardiovascular deaths [Sternberg et al 1992a, Chrousos 2000, Miller and O'Callaghan
2002] and an increased morbidity due to conditions associated with glucocorticoid
excess. These include osteoporosis, obesity, hypertension and immunosuppression
[Chrousos 1998, Webster et al 2002]. In addition to depression, hypercortisolism is
associated with other mood and affective disorders including anorexia nervosa, chronic
anxiety, obsessive-compulsive disorder, chronic alcoholism, and other situations such as
childhood sexual abuse [Chrousos 1998].
HPA axis activation with associated CRH hypersecretion has also been shown to
influence the activity of other regulatory systems in the body and may have a role in
producing some of the other clinical manifestations of stress. CRH hyperactivity is
associated with gastro-intestinal symptoms such as pain, increased gut motility and
diarrhoea – typical features of the irritable bowel syndrome that is commonly associated
with stress [Chrousos 1998, Fukudo et al 1998]. Similarly, glucocorticoids inhibit the
growth axis and it has been postulated that the severe growth retardation associated with
psychosocial abuse or deprivation during childhood is, in part, related to chronic HPA
axis activation [Chrousos 1998].
5
Chronic hypoactivation of the HPA axis in contrast is also associated with specific
disease states. Post-traumatic stress disorder, chronic fatigue syndrome and atypical
depression [Yehuda et al 1994, Chrousos 1998, Gold and Chrousos 2002, Miller and
O'Callaghan 2002] are associated with CRH hypoactivity and reduced cortisol
production. Similarly, immune dysregulation is an important consequence of altered
HPA axis activity. Relative CRH deficiency, as exemplified in the Lewis rat (in
comparison with the histocompatible Fischer rat), is associated with an enhanced
immune response and an increased resistance to infections and tumours, but an
increased susceptibility to some autoimmune conditions [Sternberg et al 1992b,
O’Connor et al 2000]. In humans studies, rheumatoid arthritis appears to be associated
with HPA axis hypoactivation [O’Connor et al 2000] with blunted cortisol diurnal
rhythms and reduced ACTH and cortisol levels [Templ et al 1996].
Dysregulation of another of the major stress response systems, the sympatho-
adrenomedullary axis, is also associated with significant clinical consequences.
Hyperactivity of the LC and other central noradrenergic centres have been shown to
influence anxiety and behavioural arousal, with dysregulation of this system postulated
as contributing to the pathogenesis of mood disorders [Gold and Chrousos 2002]
particularly depression [Leonard 1997, Aston-Jones et al 1999, Johnston et al 1999].
Further, the bidirectional relationship between this system and the HPA axis suggests
these systems reinforce and enhance each other thereby promoting and enhancing their
negative consequences [Kirschbaum et al 1993]. It is also apparent that hyperactivity of
the SAM system may also contribute to the somatic consequences of chronic stress
exposure [McEwen 2002] including the increased frequency of cardiovascular
complications associated with chronic stress [McEwen 1998, Yergani et al 2001].
6
Understanding the biological mechanisms that underlie these various disease states and
their long term sequelae is essential for the development of new and effective
therapeutic strategies and for the development of effective preventative practices.
Investigating these pathways, in the form of molecular and pharmacological studies,
animal models and human studies are all important in order to achieve these aims. For
example, understanding the role of CRH and its receptors in the pathogenesis of
depression has led to human studies with a CRH-receptor antagonist that is showing
early promise in the treatment of this disorder [Habib et al 2000, Zobel et al 2000,
Habib et al 2001]. The stress response is, however, complex and the tools available for
its study in humans are limited. The principal objective of this thesis is to describe a
model for use in the evaluation of the stress response in humans and its possible clinical
relevance.
1.1.2. Background
In the face of any threat or challenge, either real or perceived, an organism must mount
a series of coordinated and specific hormonal, autonomic, immune and behavioural
responses that allow it to escape or adapt to this threat [Chrousos 1998, McEwen 1998,
Habib et al 2001, Carrasco and Van Der Kar 2003]. In order to cope successfully with a
particular stressor, the characteristics and intensity of the stress response must match
that posed by the threat itself. Further, the duration of the response should be no longer
than is otherwise necessary for a successful outcome. [Kiecolt-Glaser et al 2002]. The
consequences of a response that is inadequate or excessive in terms of its specificity,
intensity or duration may be one or more of a multitude of psychological or physical
7
pathologies [Checkley 1996, Chrousos and Gold 1998, Gold et al 1998, McEwen 1998,
Rozanski et al 1999, Chrousos 2000, Ehlert et al 2001, Habib et al 2001, Krantz and
McCeney 2002, Vanltallie 2002].
In the early 1900’s, Walter Cannon introduced the concept of homeostasis - an ideal
steady state for all physiological processes, disruption of which initiated an emergency
‘fight or flight’ response typical of sympatho-adrenomedullary activation where
coordinated body processes would work together to restore this ideal balance [McCarty
1994]. Hans Selye [Selye 1956, Levine 2000] also emphasised the importance of
multiple, integrated systems that respond in a coordinated fashion to a particular stressor
[Carrasco and Van Der Kar 2003]. He highlighted the importance of the HPA axis in
the stress response when he described the General Alarm Reaction - an early response to
a noxious stimulus characterised by non-specific activation of the principal endocrine
response systems, the HPA and SAM axes. Continued stress exposure to the same
noxious agent had lasting effects on endocrine, immune and other systems characterised
by adrenal gland enlargement, gastric ulceration and lymphatic atrophy [McEwen 1998,
Koob 1999]. Seyle termed this the General Adaptation Syndrome and he noted that
recovery from this state was possible provided the stress was terminated. Continued or
repeated exposure, however, usually resulted in exhaustion and ultimately death [Koob
1999, Levine 2000].
In addition to noxious stimuli and other physical stressors such as exertion, physical
extremes, trauma and injury, early Freudian theory also suggested psychological stress
could produce a similar response. In the 1950’s Freud held that maternal influences and
early childhood trauma had profound and long lasting effects on the psychological and
physical health of the individual [Sternberg 1997, Levine 2000]. Animal studies by
Levine [2000], and many others [Brunson et al 2001] subsequently have shown long
8
term endocrine changes as a consequence of both psychological and physical stress
experienced in early development. Further work by Mason [1968] has demonstrated
that psychological stressors are some of the most potent stimuli of the endocrine stress
response particularly when they involve elements of novelty, uncertainty and
unpredictability [Ursin 1998]. This was highlighted by the observation that anticipating
an event can be as potent an activator of the stress response as the event itself [Levine
2000].
The notion that all stress is damaging and this damage occurs as a consequence of
excessive hormone production was initially implied by the observation that various
endocrine diseases were often associated with overt psychopathology and that this
corrected itself once the endocrine condition resolved [Brambilla 2000]. This led
several authors to postulate that single hormone abnormalities (either excesses or
deficiencies) were responsible for specific behavioural consequences such as depression
or schizophrenia [Brambilla 2000]. Whilst this notion is no longer held, it did lead on
to the idea that abnormalities of neurotransmitters and neurotransmitter systems are
more likely responsible for certain psychopathologies, and that pharmaceutical agents
that target these systems could produce successful treatments [Brambilla 2000]. It has
become apparent more recently, that the stress response is beneficial in protecting an
individual from a harmful situation and that the brain can learn, adapt and adjust its
future response to be more efficient and effective [McEwen 1998]. Problems arise
when the stress is sustained or becomes repetitive. The principal stress responsive
systems are energy expensive and failure to adequately deal with or recover from these
challenges may result in ill health. The ability to continually adapt to and recover from
stress (or allostasis [McEwen 2000]) is further diminished when we choose a lifestyle
consisting of a poor diet, little exercise, disturbed sleep and the consumption of
9
cigarettes and alcohol and so on. The consequences of not adequately recovering from
each stressful episode (allostatic load) is the promotion of ill-health through high blood
pressure, hypercortisolism, suppressed immune function and poor growth and
[Chrousos 1998, McEwen 1998, McEwen 2000, Kiecolt-Glaser et al 2002]. Ultimately,
the consequences for long-term physical and mental health are significant.
The stress response system has evolved as both an early warning system capable of
recognising potential or existing threats, and as a response system that can initiate and
drive the necessary processes required to escape or confront the threat. By its very
nature, the response must be dynamic, beginning rapidly with brain and behavioural
activation followed quickly by physiological activation [Richer et al 1996]. These
processes are characterised by positive-feedback and feed forward loops that enhance
and reinforce themselves as well as recruiting other arms of the stress response. Slower
acting hormone systems are recruited into the cascade providing checks and balances to
the already active, but energy expensive systems, putting a brake on the whole response
to ensure it is kept appropriate to the type of stress faced, to its intensity and duration,
and to ensure the response is switched off when the threat has been adequately dealt
with [Eriksen et al 1999, Sapolsky et al 2000].
Changes in the internal or external environment that represent either real or potential
threats are recognised with the parts of the brain responsible for receiving, integrating,
interpreting and then relaying this information on to those areas responsible for co-
ordinating the necessary response. This brain activation can be detected within
milliseconds and proceeds over seconds to minutes as the response continues to unfold.
Stereotypical orienting behaviour, initiated within seconds, gradually gives way to more
10
goal-directed behaviour that is specific to the stressor being faced and the environment
in which it is occurring [Eriksen et al 1999].
Activation of the autonomic nervous system occurs within seconds, mediated by the
release of catecholamines from sympathetic nerves and the adrenal medulla and
enhanced by the withdrawal of parasympathetic activity [McCarty 1994, Young and
Landsberg 1998]. These systems promote the immediate physiological, motor and
behavioural responses needed in the face of acute physical or psychological stress.
Within minutes of the onset of this cascade of events occurring, hypothalamic releasing
hormones stimulate the release of pituitary hormones with the appearance of ACTH
signalling the recruitment of the HPA axis into the process [Harbuz and Lightman 1992,
Chrousos 1998, Thorner et al 1998]. The cortisol response is much slower, with peak
levels not seen for 15-20 minutes after the onset of the stress. Early actions of the HPA
system provide additional energy resources for the stress response, whilst slower gene-
related effects over the next few minutes to hours serve to restrain ongoing actions of
the stress response which, if left unchecked, may prove to be unsustainable for the
individual [Chrousos 1998, Habib et al 2001].
The HPA and SAM axes are the principal endocrine effector arms of the stress
response. However, a number of other hormone axes and neurotransmitter systems are
either directly stress responsive themselves, or modulate these other hormone systems.
11
1.1.2.1. The HPA Axis
Corticotropin releasing hormone (CRH) was identified by Vale and others [Vale et al
1981] in 1981 as the 41-amino-acid peptide responsible for promoting the synthesis and
release of anterior pituitary ACTH. It is also widely distributed throughout the CNS,
including within the cortex where it has important effects on behaviour and cognitive
processing. Within the brainstem interactions with sympathetic and parasympathetic
centres influence autonomic functioning whilst within limbic and para-limbic regions
such as the amygdala, CRH influences the expression of mood and anxiety-type
behaviours [Owens and Nemeroff 1991, Harbuz and Lightman 1992, Thorner et al
1998, Carrasco and Van Der Kar 2003].
ACTH release from the anterior pituitary under the influence of CRH acts directly on
the adrenal cortex to promote the release of large amounts of adrenal glucocorticoids
into the circulation [Harbuz and Lightman 1992, Chrousos 1998, Habib et al 2001].
Glucocorticoids, in general, have two fundamental roles in the stress response. Firstly,
during stress free periods, basal levels have a role in preparing the organism for future
stress exposure. This involves energy storage and conservation by promoting glucose
and fat uptake and opposing energy utilisation. They also prime the immune system for
future activation and promote memory formation of previous stressors so that future
exposure to the same or similar stressor may facilitate a more rapid and efficient
response [Sapolsky et al 2000].
The second role is that of a modulating effect at the time of stress exposure itself.
Initially glucocorticoids enhance the cardiovascular effects of catecholamines and
vasopressin (AVP), promote energy provision and utilisation, influence and enhance
12
appropriate stress-related behaviours and stimulate certain aspects of the immune
response [O’Connor et al 2000, Sapolsky et al 2000, Tsigos and Chrousos 2002].
However, most of these responses would occur even in the absence of any circulating
glucocorticoids. It is perhaps even more important that once the stress response has
been initiated, some of the principal actions of glucocorticoids are to suppress and
restrain the activity of these systems, in particular the SAM and immune systems. In
doing so, glucocorticoids provide an essential regulatory balance to ensure the stress
response is appropriate in terms of both its intensity and duration and that all these
responses are ‘switched off’ when the stress has been successfully dealt with [Sapolsky
1994, McEwen 1998, Goldstein and McEwen 2002]
The functional role of CRH and its receptors within the brain has been extensively
researched in animals using anatomical lesion studies, CRH-receptor knockout mice,
and antibodies or CRH-like compounds with either inhibitory or stimulatory effects on
CRH receptors [Ma and Lightman 1998, Neumann et al 1998, Reichlin 1998, Bakshi
and Kalin 2000, Jessop et al 2001]. Overall, the addition of CRH seems to promote
anxiety-like behaviour and increase the physiological effects of stress, whilst inhibiting
CRH suppresses both stress-related behaviour and physiological changes. In several
studies the administration of CRH directly into the brains of rodents decreased
exploratory behaviour, increased fear and conflict behaviour, and suppressed feeding
and sexual activity [Jones et al 1998, Bakshi and Kalin 2000]. CRH also stimulated
stress-related physiological changes including increases in heart rate and blood pressure
[Jezova et al 1999]. At least two CRH receptors have been identified within the brain.
CRH-R1 is widely distributed in the pituitary, brainstem, cerebellum and amygdala and
appears to mediate stress induced HPA axis activation as well as stress related (and
possibly spontaneous) anxiety behaviour. Some evidence for this includes the
13
demonstration that CRH-R1 knockout mice display increased behavioural activity
consistent with reduced anxiety in situations where normal mice display increased
anxiety. Additionally, following an acute stress, these animals show a blunted ACTH
and glucocorticoid response [Owens and Nemeroff 1991]. Similarly, studies of
compounds that block CRH-R1 suggest treated animals display reduced anxiety to
some, but not all, anxiety provoking situations as well as displaying a reduced HPA
hormone response to high intensity stressors [Bakshi and Kalin 2000].
CRH-R2 has a more restricted distribution and a much higher affinity for CRH-like
peptides, particularly urocortin II and III, than for CRH. Studies of CRH-R2 knockout
mice as well as CRH-R2 inhibitors have yielded less consistent results with animals
displaying a mix of both increased and decreased anxiety behaviour and mixed
hormonal responses to stress that were dependent on such factors as the gender of the
animal and the type of provoking stimulus [Bakshi and Kalin 2000].
In addition to driving both HPA and behavioural responses to stress, CRH also
modulates the autonomic, immune and other endocrine responses to stress. CRH
activity in the brain activates central sympathetic systems with subsequent increases in
adrenaline and noradrenaline release, decreased parasympathetic outflow and an overall
increase in sympathetic tone with increased heart rate, blood pressure and cardiac output
[Jezova et al 1999, Habib et al 2001, Gammatopoulos and Chrousos 2002].
Bidirectional influences between CRH and the immune system result in cytokine (IL-1,
IL-6 and TNF-α) mediated stimulation of the HPA axis with increased cortisol release.
Cortisol itself has widespread anti-inflammatory actions, although there is evidence that
locally produced CRH and autonomic-immune interactions stimulated by CRH can
have pro-inflammatory effects [O’Connor et al 2000, Habib et al 2001].
14
Whilst much of the understanding of the biology of CRH is derived from studies in
rodents, human studies have indicated that dysregulation of CRH function contributes to
the pathophysiology of several disorders including: depression, post-traumatic stress
disorder, disorders of sleep, stress-related menstrual irregularity and infertility,
manifestations of the metabolic syndrome and functional gastro-intestinal disorders
[Chrousos 1998, Koob 1999, Gammatopoulos and Chrousos 2002]. Further, targeting
CRH peptides and their receptors are increasingly seen as an effective means of
managing these various conditions and a number of CRH agonists and antagonists are
currently under evaluation in clinical and pre-clinical studies. Oral CRH-R1 antagonists
have been shown to decrease anxiety-related endocrine, autonomic and behavioural
responses in rat and primate models [Habib et al 2000, O’Connor et al 2000]. They
have also been safely trialed in a human phase II clinical study of patients with major
depression where patients showed an improvement in depression and anxiety symptoms
on treatment with a significant worsening following treatment discontinuation [Zobel et
al 2000].
Cortisol secretion is precisely controlled by the complex feedback system that regulates
the HPA axis. Whilst significant variability exists between individuals, within an
individual both the circadian and ultradian rhythms of cortisol are tightly controlled and
highly stable. Repeated or chronic stress with consequent CRH hyperactivity and
excessive or unrestrained cortisol exposure has been shown to result in hippocampal
neuronal damage and impaired hipppocampal function [McEwen 1998]. The
hippocampus is involved with contextual memory formation and is also an essential part
of the cortisol negative feedback system. Impaired memory formation particularly for
emotional events may exacerbate future stress responses under similar circumstances,
15
whilst impaired negative feedback further exacerbates the hypercortisolism. Rosmond
et al [Rosmond et al 1998] was able to demonstrate that the HPA axis in non-stressed
individuals is characterised by a wide circadian rhythm (with distant morning zeniths
and evening nadirs), a discrete but small response to an acute stress and appropriate
suppression of cortisol levels following the administration of an exogenous
glucocorticoid indicating a normal pattern of secretion with an intact feedback
mechanism. In contrast, chronically stressed individuals displayed a decreased
circadian variability (due to decreased morning zeniths and increased evening nadirs), a
large acute stress response and inadequate cortisol suppression indicating an altered
pattern of secretion and impaired negative feedback with resultant cortisol
hypersecretion. The effect of this hypercortisolism in the brain includes depression and
anxiety [Chrousos and Gold 1998], whilst in the periphery it is associated with bone
loss, obesity, hypertension, insulin resistance and other features of the metabolic
syndrome [Chrousos and Gold 1998]. Further, some patients with cancer have been
shown to have alterations to their neuroendocrine and immune axes that may have
implications for their disease progression and outcome [Speigel et al 1998, Sephton et
al 2000].
1.1.2.2. Arginine Vasopressin
AVP released from magnocellular cells of the PVN and the supraoptic nucleus in
response to osmotic and haemodynamic stimuli promotes water and electrolyte
retention in the kidneys [Carrasco and Van Der Kar 2003] whilst AVP derived from
parvocellular cells of the PVN acts synergistically with CRH to stimulate the release of
ACTH [Ma et al 1997, Sephton et al 2000]. In animal studies, AVP appears to be an
16
important mediator of ACTH release during chronic stress. Chronic stress increases
AVP expression within the PVN [Ma et al 1997, Aguilera and Rabadan-Diehl. 2000],
whilst inhibitors of AVP impair the HPA response to various stressors including
insulin-induced hypoglycaemia and restraint stress [Linton et al 1985]. As mentioned,
during chronic or repeated stress, glucocorticoid levels are initially high with negative
feedback downregulating CRH receptors and subsequently suppressing CRH and
ACTH responses. As a result, CRH responses often decline with time in the face of
high cortisol levels (habituation). However, exposure to a novel stress during this time
is associated with a normal HPA response and this restored response is mediated by
AVP whose receptors that don’t show the same downregulation (and in some
circumstances may be upregulated) [Ma et al 1999, Aguilera and Rabadan-Diehl. 2000].
Recently, Scott and Dinan [1998] have suggested the impaired HPA axis regulation that
commonly accompanies major depression in humans reflects AVP activity rather than
CRH. AVP may therefore be an important target in the design of therapeutic agents for
this condition. AVP, it appears, sensitises the pituitary to the effects of a superimposed
novel stressor suggesting AVP and CRH are regulated independently.
1.1.2.3. The Sympatho-Neural and Sympatho-Adrenomedullary Axis
The hallmark sympathetic ‘fight or flight’ response, as described by Cannon [McCarty
1994], is characterised by global activation of the SAM system and features typical
physiological and behavioural activation including heart rate quickening, increased
blood pressure and rapid breathing. Fear, vigilance, sensory arousal and motor
activation often with trembling, goose bumps and piloerection (hair standing on end)
also occur. Release of glucose stores, immune activation and increased blood flow to
17
essential organs such as the brain occur whilst non-essential activity like digestion is
inhibited. This produces a state ‘of emergency’ which can rapidly attend to a sudden
change in physiological balance [Young and Landsberg 1998, Habib et al 2001]. This
response is characterised by its speed of onset, its ability to begin in anticipation of an
event being stressful, and by its interaction with other stress-responsive systems [Young
and Landsberg 1998]. This interaction can occur either through neural connections or
through increased blood flow that transports other messengers (such as hormones and
cytokines) more rapidly to their respective sites of action [Young and Landsberg 1998].
Whilst the ‘fight or flight’ reaction is a useful way of describing the global SAM
response to various stressors, it is clear from several studies that the adrenomedullary
and sympathetic nerve responses to stress are regulated independently and the
components of each vary in their response to different stressors [Bornstein and
Chrousos 1999]. In humans, for example, the physical stress of cold or exercise is
associated with both adrenaline and noradrenaline responses, whereas the response to
insulin-induced hypoglycaemia is mediated predominantly by adrenaline alone [Young
and Landsberg 1998]. Psychological stress is also a potent stimulus of the SAM axis,
with activation occurring at generally lower levels of stress than that required to
generate an HPA response [Singh et al 1999, Habib et al 2000, Zobel et al 2000].
Unlike cortisol and the HPA axis, SAM (catecholamine, heart rate and BP) responses to
either exercise or psychological stress do not show the same continuum of responses,
although differences between individuals are apparent when anticipating a stress, or in
response to cold stress [Cacioppo et al 1995, Kirschbaum et al 1995, Negrao et al
2000]. Concordance between high anticipatory or cold-induced heart rate responses and
high stress-induced cortisol responses has allowed the classification of individuals as
18
high or low responders, with the hypothesis that either hypo- or hyper-reactivity of the
stress response will influence an individual’s susceptibility to developing various
psychological, metabolic or immune-related disorders [Biondi and Picardi 1996].
Concordance between SAM responses and the HPA axis has its foundation in a large
number of neuroanatomical and behavioural studies indicating the importance of strong,
bi-directional neuronal influences of the brainstem catecholamine centres and CRH
mediated pathways including the PVN [Kvetnansky et al 1995, Koob 1999, Zeigler et al
1999, Gerra et al 2001]. This translates into a powerful feed-forward system where
stress-induced activation of catecholaminergic systems activates, and in turn is further
activated by, stress-responsive CRH neurones [Neumann et al 1998, Carrasco and Van
Der Kar 2003]. Such an interaction would serve to appropriately reinforce the principal
early stress response systems following an acute threat, but altered sensitivity of one or
other may significantly contribute to the disordered activity of the endocrine stress
response that underlies many psychological conditions. Hyperactivity of the brainstem
catecholamine system with subsequent CRH hyperactivity and reduced feedback
sensitivity, for example, has been postulated as a likely mechanism for the endocrine
changes seen in depression and anxiety states [Kvetnansky et al 1995, Gold et al 1998,
Koob 1999, Ressler and Nemeroff 1999].
1.1.2.4. Prolactin
Prolactin, released from the anterior pituitary under tonic inhibitory control from
dopamine neurons is required predominantly for milk production during lactation
[Reichlin 1998.]. It has, however, been shown in both animal and human experiments
19
to be released in response to acute stress, although its role is yet to be fully determined
[Chrousos 1998, Biondi and Picardi 1999, Habib et al 2001]. High prolactin levels in
rodents, such as those occurring during lactation, are associated with reduced expression
of anxiety behaviour and with HPA axis suppression [Torner et al 2001]. Further,
intracerebroventricular administration of prolactin in rats reduced both behavioural
anxiety and corticosterone responses to a superimposed stress, whilst a prolactin
receptor antagonist given by the same route enhanced anxiety behaviour [Torner et al
2002]. A few studies of physical and psychological stress have measured prolactin
responses in humans, but little consistency between these responses has been shown
[Richer et al 1996, Biondi and Picardi 1999]. In some studies prolactin levels have
been shown to increase, whilst in others, prolactin has either not changed or decreased
[Richer et al 1996, Biondi and Picardi 1999].
1.1.2.5. Opiates
Endogenous opiates and their receptors are ubiquitously distributed throughout the
central, peripheral and autonomic nervous systems. Opiate pathways have been shown
to influence a broad range of functions and behaviours related to stress including the
regulation of pain, reinforcement and reward, and the modulation of autonomic and
neuroendocrine axes [Olson et al 1996, Drolet et al 2001]. At least three families of
endogenous opiates have been identified (enkephalin, endorphin and dynorphin) which
act through three major opiate receptors – mu (µ), delta (δ) and kappa (κ) [Drolet et al
2001]. Each receptor subtype maintains a distinct anatomical distribution within the
brain, with µ and κ receptors in particular being associated with structures involved in
the neurobiology of stress. Sites include the hippocampus, the hypothalamic
20
paraventricular and arcuate nuclei, the central nucleus of the amygdala, locus coeruleus,
ventral lateral medulla and dorsal motor nucleus of the vagus [Mansour et al 1995,
Drolet et al 2001, Habib et al 2001]. Reciprocal innervation between CRH producing
neurones of the PVN, brainstem noradrenergic centres, the amygdala and limbic system
and the opiate system suggests a significant role for opiate pathways in the modulation
and regulation of the principal stress response centres [Drolet et al 2001, Habib et al
2001]. Acute stress exposure, for example, activates the HPA and SAM axes and
reciprocal innervation increases opiate expression and action within the PVN and other
stress responsive centres [Young and Lightman 1992, Mansi et al 2000, Drolet et al
2001]. Opiate pathway stimulation inhibits the activity of the central components of the
stress system thereby dampening their output. Further, projections to the hindbrain and
spinal cord produce analgesia [Habib et al 2001]. In other words, the opiate system
seems to act to diminish the impact of the stress response by dampening the physiologic
and behavioural responses that, if left unchecked, would be ultimately harmful to the
organism [McCubbin 1993, Drolet et al 2001].
1.1.2.6. Serotonin (5-HT)
Serotonergic systems are important regulators of the behavioural, autonomic and
endocrine responses to stress [Dinan 1996, Lowry 2002]. Originating primarily from
the brainstem raphe complex, serotonergic neurones project to multiple forebrain
centres including the central nucleus of the amygdala, hippocampus and medial septum,
areas involved in the emotional fear response [Lowry 2002]. 5–HT projections have
been shown to influence HPA axis responses through their regulation of CRH output at
the level of the PVN, but recent evidence also indicates an additional influence of 5-HT
21
on anterior pituitary ACTH release and on adrenal cortical glucocorticoid release
[Dinan 1996]. Projections to the locus coeruleus and other autonomic brainstem centres
influence autonomic ‘fight-and-flight’ responses whilst behavioural motor responses are
influenced by projections to areas such as the substantia nigra [Lowry 2002].
The influence of these serotonergic systems on the stress response is, however, complex
as they can either facilitate or inhibit responses. The complexity of these systems arises
in part by the vast array of 5-HT receptors within the brain [Chaouloff 2000] and in part
from evidence that the serotonergic system itself is made up of separate systems that are
also anatomically and functionally discrete [Lowry 2002]. Activation of serotonergic
systems on the one hand are associated with enhanced autonomic arousal, anxiety and
fear, whilst on the other hand stress-induced activation via separate serotonergic
pathways results in inhibition of HPA and SAM responses [Lowry 2002].
1.1.2.7. The Growth, Thyroid and Reproductive Axes
Anterior pituitary growth hormone (GH) release is stimulated following some acute
stressors [Chrousos 1998]. GH itself stimulates the release of IGF-I from the liver that
acts on many different tissues, but whose main role in acute stress is the release of
energy stores from the liver [Thorner et al 1998]. Physical stressors including exercise
are generally associated with an increase in GH levels [Richer et al 1996, Thorner et al
1998], although most psychological stressors are not [Negrao et al 2000]. Chronic
emotional disorders such as anxiety or depression are associated with suppressed GH
levels through the combined influence of CRH, which stimulates GH inhibitory
22
peptides such as somatostatin, and through a direct action of glucocorticoids on the GH
gene [Habib et al 2001].
Similarly, the thyroid axis is also suppressed in response to chronic stress-induced HPA
axis activation and thyroid hormone (T4) levels seem to follow a similar pattern to GH
levels in response to acute stress [Biondi and Picardi 1999, Habib et al 2001].
Hypothalamic TRH is inhibited by CRH and somatostatin, whilst glucocorticoids inhibit
both TSH release from the pituitary and T4 release from the thyroid gland [Chrousos
1998, Thorner et al 1998]. Further, they also reduce the conversion of T4 to its more
active form, triiodothyronine (T3), in peripheral tissues [Thorner et al 1998].
The reproductive axis is also very sensitive to the inhibitory influences of CRH,
glucocorticoids and components of other stress responsive pathways such as
inflammatory cytokines and endorphins [Habib et al 2001]. Chronic HPA axis
activation from physical or psychological stress, as in highly trained athletes or people
with anorexia nervosa, is commonly associated with suppression of reproductive
hormone release, and particularly with menstrual cycle inhibition in women [Knol 1991,
Chrousos 1998, Biondi and Picardi 1999].
1.1.2.8. Other Neurohormonal Systems
Oxytocin, a posterior pituitary hormone required for the induction of labour, is also
released during stress [Van de Kar and Blair 1999]. Currently its precise function
during stress is unclear, but some recent evidence suggests it acts to oppose or modulate
the action of AVP [Van de Kar and Blair 1999].
23
The renin-angiotensin system is an important hormonal system that regulates
circulating blood volume and blood pressure. It is activated in response to
haemodynamic stressors such as blood loss, but renin is also released during acute
psychological stress as part of the activation of the sympathetic nervous system and may
have a role in the anticipatory phase of the stress response [Van de Kar and Blair 1999].
Substance P is activated during pain and chronic inflammatory stress and appears to
add to the inhibitory influence some inflammatory mediators have on CRH and the
HPA axis. In addition, substance P increases sympathetic activation in response to pain
[Habib et al 2001].
Neuropeptide Y (NPY), a 36-amino acid peptide, is an additional important component
of the sympathetic nervous system. It is found within sympathetic brain centres where it
has important regulatory effects on both the HPA and SAM axes, influences appetite
and feeding behaviour and may also have anxiolytic properties [Habib et al 2001]. In
the periphery, it is found within sympathetic nerve fibres associated with blood vessels
and immune cells. Released with noradrenaline, neuropeptide Y contributes to the
control of blood pressure, blood flow and lymphocyte traffic [Elenkov et al 2000].
24
1.2 Investigating the stress response
1.2.1 Currently available tools and their limitations
Survival of an individual depends on the ability of the stress response systems to mount
a biologically appropriate response when threatened. Failure to do so is likely to be
deleterious and result in significant harm. Normal baseline or resting cortisol and other
stress hormone levels are not necessarily indicative of a system that is capable of
responding dynamically to perceived or actual threats. Partial or even complete ACTH
deficiency, for example, is often associated with normal resting cortisol levels and is
typically asymptomatic during non-stressed conditions. However, when stressed these
individuals are often unable to mount an appropriate cortisol response, a state that can
be life threatening. It is well established that baseline or even diurnal cortisol profiles
do not reflect the dynamic integrity of the HPA axis and are therefore not particularly
useful for its assessment [Erturk et al 1998]. Rather, it is necessary to use challenges
that assess the dynamic ‘responsiveness’ of the various components of the stress system.
Numerous laboratory based challenges have been developed and used for the study of
the stress response in humans. In addition, real life events and scenarios have also been
examined in order to try and predict the impact of stressful life events on responses and
disease outcomes. Laboratory paradigms that have been described have used a variety
of psychological, physiological or pharmacological challenges.
25
1.2.1.1 Psychological challenges
Psychological challenges will activate stress response systems if the challenge is
threatening or demanding without adequate perceived resources being available for
coping with the particular challenge [Ehlert et al 2001]. In other words, these
challenges are usually characterised as being novel, unpredictable and uncontrollable
[Ehlert et al 2001]. Typical challenges have involved performance tasks such as mental
arithmetic and the Stroop colour-word conflict test, speech or interview tasks, video
games and films or videos with a strong emotional content [Biondi and Picardi 1999].
Combined tasks are also commonly used, the Trier Social Stress Test (TSST) being the
most widely used example [Kirschbaum et al 1993, Ehlert et al 2001]. This task
involves a 10 minute stress anticipation period, followed by a speaking task and a serial
subtraction task in front of at least three unfamiliar observers that is videotaped and
recorded [Kirschbaum et al 1993]. Whilst cortisol levels in response to this task are
reproducibly elevated, albeit with a wide inter-individual variability, repeated
challenges do not stimulate the HPA axis in most individuals as the task, when repeated,
loses its novelty and unpredictability [Kirschbaum et al 1995].
Mental arithmetic alone and most other public speaking or performance tasks are not
reliably associated with a significant cortisol response although most do produce
significant increases in catecholamine release with expected increases in blood pressure
and heart rate [Jorgensen et al 1990, Sgoutas-Emch et al 1994, Biondi and Picardi
1999]. Combined tasks, including the TSST, consistently increase SAM responses that
are usually reproducible when the task is repeated [Al’Absi et al 1997, Biondi and
Picardi 1999]. Cortisol responses on the first occasion show wide inter-individual
variation with a gradual decline in responses with repeated exposure in most individuals
[Kirschbaum et al 1995]. As will be discussed later, a group of individuals will continue
26
to demonstrate HPA axis activation with repeated challenges. These individuals,
termed ‘high responders’ may be more prone to the adverse effects of chronic activation
of the stress response systems [Cacioppo et al 1995].
A few studies of both laboratory and real life psychological stress have measured
prolactin responses, but little consistency between responses has been shown [Richer et
al 1996, Biondi and Picardi 1999]. In some studies of academic exam stress and
parachute jumping, prolactin levels have been shown to increase, whilst other forms of
psychological stress have shown either no change or even decreases in prolactin levels
[Richer et al 1996, Biondi and Picardi 1999]. Similarly, laboratory psychological
stressors are generally not associated with a change in GH or thyroid hormone levels
(TSH or T4) [Biondi and Picardi 1999], although some studies of parachute jumping or
bereavement recall have shown a response [Richer et al 1996, Biondi and Picardi 1999,
Ehlert et al 2001].
Films, videotapes and some interview protocols have been inconsistent in the
production of both HPA and SAM responses and it is apparent that a response is
dependent on the subject matter being viewed or discussed [Biondi and Picardi 1999]
and the duration of exposure. As might be expected, challenges that involve exposure
to subject matter that the individual does not perceive as being emotionally relevant to
themselves is not likely to provoke a response [Biondi and Picardi 1999].
Whilst psychological stressors are widely used for the assessment of stress responses in
the laboratory setting, they have a number of specific limitations. Firstly, not all
challenges will reliably provoke an HPA axis response (for example mental arithmetic
or other performance tasks alone). Secondly, HPA responses rapidly decline with
repeated exposure making them unreliable for experiments evaluating different stress
27
conditions or interventions in the same individual. Thirdly, the challenges are often
labour intensive (requiring several investigators per participant) and time consuming
(with many typically lasting several hours). Finally, standardising the challenge
between individuals can be difficult as the success of the challenge depends on its
relevance to the subject and individuals vary considerably according to such factors as
their level of experience with public speaking, education and so on.
Parachute jumping, for example, is considered to be a psychological as opposed to a
physical challenge and has been studied previously as a paradigm that produces a
significant HPA, SAM, GH and TSH response [Richer et al 1996]. This paradigm is,
however, illustrative of some of the difficulties faced. The HPA response only occurred
in first-time jumpers. The challenge is inherently dangerous and is expensive with a
very restricted pool of potential volunteers. Sampling was also very difficult, requiring
a complex arrangement required for samples to be taken automatically from a brachial
arterial line (which itself has significant ethical, safety, cost and practicality issues)
[Richer et al 1996].
1.2.1.2 Physical challenges
Physical exertion (for example from treadmill running or cycling, knee bends or
handgrip) and pain (cold, venepuncture) have been the principal paradigms used as
physical challenges. Intense exercise, usually as running or cycling, has been
extensively studied as a means of generating a stress response. These challenges require
the subject to exercise to exhaustion, based on measuring maximum oxygen
consumption (VO2max) and reliably produce activation of the HPA, SAM, GH and in
28
some cases prolactin systems [Petrides et al 1994, Singh et al 1999, de Vries et al
2000]. HPA responses do, however, show the same continuum as are seen in response
to psychological challenges and there is a good correlation between high responders to
exercise stress and high responders to psychological stress [Singh et al 1999]. It is
postulated that there is a dissociation in the neuroendocrine (HPA and SAM) response
to exercise stress with SAM responses reflecting work-load and HPA responses
reflecting the mental effort or distress associated with exercising to exhaustion [Singh et
al 1999].
Several potential limitations to the use of this challenge exist. Principally, the
equipment is expensive and the challenge labour intensive usually requiring several
operators including someone specifically skilled in exercise physiology and the use of
the equipment. The equipment itself is not portable, hence limiting where it can be
performed and there may be significant safety issues associated with this level of
exercise. The test would also be restricted to those physically capable of performing
this type of exercise to this intensity. Finally, there are a number of confounding
variables influencing neurohormonal responses that make standardising exercise
protocols difficult. These include static vs dynamic exercise; position (supine vs erect);
the mass of muscle engaged in the exercise; the intensity and duration of the exercise;
and importantly an individuals level of training [de Vries et al 2000, O'Sullivan et al
2001].
The cold pressor test is a relatively well tolerated stressor that has been shown to
increase both SAM and HPA axis activity through combined effects on temperature and
pain sensing afferents [Bullinger et al 1984, Durel et al 1993, Pascualy et al 2000]. The
test requires the subject to immerse their hand in an ice bath (4°C) for a brief period (1-
29
3 minutes or longer depending on the protocol). Its principal use has been in examining
the pressor response as a predictor of risk of hypertension [Velasco et al 1997] and in
pain research [Washington et al 2000]. Little literature exists on the HPA response and,
in particular, stress responses to repeated exposure in the same individual. Whilst this
challenge is simple to perform and appears to produce reliable SAM and HPA
responses, doubts remain over the response to repeated exposure, and as with other
types of painful stimuli such as venepuncture, ethical difficulties exist related to
repeated exposure.
1.2.1.3 Pharmacological challenges
The administration of numerous pharmacological agents have been used in the
evaluation of various components of the stress response with the insulin tolerance test
considered to be the ‘gold standard’ for the evaluation of the integrity of the HPA axis.
In psychiatric research, pharmacological agents have been widely used in the study of
anxiety, fear and panic and there is extensive crossover between the components of the
fear/anxiety response and the general stress response.
Insulin induced hypoglycaemia or the insulin tolerance test (ITT) is considered the
‘gold standard’ test for the dynamic assessment of HPA axis integrity [Erturk et al
1998, Lange et al 2002]. The test involves the intravenous administration of a bolus
dose of insulin with regular blood glucose monitoring. Subjects must develop
symptoms consistent with hypoglycaemia (sweating, hunger, tachycardia and
confusion) and must have a recorded blood glucose level <2.5 mmol/l [Erturk et al
1998]. A test is considered normal if the peak cortisol response is >550 nmol/l at 60
30
minutes, and abnormal if lower than this. ACTH responses correlate poorly with
cortisol responses and are not useful in evaluating the axis [Borm et al 2003]. The test
has a high positive predictive value for determining a normal HPA response, but the
result can only be interpreted as either normal or abnormal. Unlike other dynamic tests
of the HPA axis that produce a continuum of cortisol values with the degree of cortisol
elevation predictive of their susceptibility to responses to other stressful conditions and
to disease outcomes [Greiz et al 1990a, Cacioppo et al 1995, Kirschbaum et al 1995,
Singh et al 1999], the ITT is unable to provide this sort of information. Further, the ITT
represents a well established test of HPA and GH axis integrity, but is not a well
established test of other components of the stress response particularly the SAM axis
and emotional arousal. Finally, the complexity and safety of the ITT has often been
questioned in its use as a routine clinical test [Erturk et al 1998]. A study by Nye et al
[Nye et al 2001] suggested that for reliability, the glucose nadir should be <1.6 nmol/l,
significantly increasing the potential for hypoglycaemic-associated complications such
as seizures.
Other pharmacological tests of HPA axis activity that have been employed include the
dexamethasone suppression test, the naloxone test, metyrapone administration, and
CRH with or without dexamethasone administration. These agents all act on different
levels of the HPA axis either directly stimulating anterior pituitary ACTH release
(CRH), interfering with cortisol synthesis (metyrapone) or altering negative feedback
(dexamethasone and naloxone) and specific response criteria have been set [Inder et al
1995, Checkley 1996, Gold et al 1998, Thorner et al 1998, Ehlert et al 2001]. All these
tests suffer from low sensitivity and specificity with considerable inter-individual
variation and significant numbers of non-responders. In addition they provide little if
any information on non-HPA components of the stress response.
31
The physiological and psychological features of fear, anxiety and at their extreme,
panic, in many ways resembles that of the generalised stress response, including
emotional arousal, cardiorespiratory and autonomic activation and HPA axis stimulation
[Gorman et al 2000]. Numerous pharmacological agents have been used to generate
fear responses. Yohimbine (an α2-adrenergic antagonist), cholecystokinin-β receptor
antagonists (CCK), fenfluramine (a 5-HT releasing agent), meta-chlorophenylpiperizine
(a mixed 5-HT receptor agonist), caffeine and β-carboline esters have all been used to
generate panic attacks that are associated with significant HPA axis activation as well as
cardiovascular stimulation [Bourin et al 1998, Sinha 1999, Argyropoulos et al 2002].
Not all are respiratory stimulants and catecholamine responses have been inconsistent
and variable [Bourin et al 1998]. Sodium lactate, bicarbonate, isoproterenol and
doxapram on the other hand will all induce panic attacks but without any significant
HPA axis activation [Bourin et al 1998, Sinha 1999, Argyropoulos et al 2002]. Carbon
dioxide exposure, in a number of different protocols, will produce emotional arousal,
autonomic activation and in some, but not all cases, HPA axis activation [Hardgrove et
al 1938, Sechzer et al 1960, Cross and Silver 1962, Gorman et al 1988, Woods et al
1988, Greiz et al 1990a, Greiz et al 1990b, Perna et al 1994, Papp et al 1997, Bourin et
al 1998, Verberg et al 1998, Sinha 1999, Argyropoulos et al 2002, Bailey et al 2002].
The rationale for choosing to evaluate this in more detail as a model of the stress
response is outlined below. Of the other agents, CCK-4 has been considered safest and
most reliable in terms of generating panic, autonomic and HPA axis activation
[Bradwejn et al 1991a, Bradwejn et al 1991b, Bourin et al 1998,]. Panic rates are at
least equivalent to a single breath of 35% CO2 [Koszycki et al 1991] and its major
drawbacks are its need to be given by intravenous infusion and cost [Bourin et al 1998].
32
In summary, the number of different challenges that have been used previously are
indicative of the lack of a single safe, simple and reliable test of the neuroendocrine
stress response. Many paradigms only target specific components of the stress response
system such as the behavioural response or the HPA axis, whilst others are either
difficult to perform and reproduce, costly or potentially dangerous. There was a
perceived need, therefore, to develop a paradigm that could safely and easily produce
reliable stress responses that could then be used to investigate the importance of these
systems in health and disease.
1.2.2. CO2 as a potential neuroendocrine stressor
Carbon dioxide is known to be the principal regulator of respiration, acid-base balance
and behavioural-state arousal in humans [West 1974]. Acting through central
chemoreceptors located predominantly within the ventrolateral medulla (VLM) of the
brainstem, small incremental changes in CO2 produce large changes in minute
ventilation (respiratory rate x tidal volume) to provide fine control over acid-base
homeostasis [West 1974]. In addition to regulating ventilation, changes in CO2
concentration are also associated with behavioural state arousal [Woods et al 1988]. In
particular, a rise in CO2 concentration is associated with increased arousal and a
perception of breathlessness or air hunger as well as sympathetic activation (see below).
This arousal is associated with activation of a number of different brain centres as
evidenced by functional MRI imaging during CO2 exposure. These brain centres are
also associated with other primitive feelings and behaviours including hunger, thirst and
pain [Brannan et al 2001, Liotti et al 2001]. It is likely this arousal, sympathetic
activation and fear produced by CO2 exposure has evolved as an instinctive protective
33
mechanism to warn and escape from an impending threat to respiratory homeostasis,
such as may occur with suffocation [Brannan et al 2001, Liotti et al 2001].
As a challenge paradigm, CO2 has been studied extensively in relation to its respiratory
physiology, its autonomic response (particularly the cardiovascular response to CO2
when it was used as an anaesthetic induction agent) and more recently its emotional and
behavioural response in anxiety research. There is also extensive literature on its safety
and methods of administration. A detailed review of the literature regarding CO2 and
the rationale for its choice in the development of a novel neuroendocrine stress
paradigm is provided in Chapters 2 and 3. Most importantly there was local expertise
within the University of Bristol on its use in psychiatric research involving human
subjects. Whilst formal neuroendocrine studies were limited and inconsistencies existed
in the described autonomic and cortisol responses to CO2 exposure, the current literature
did support the potential for CO2 as useful challenge in the investigation of the stress
response.
34
1.3 Summary, aims and hypotheses
At present, there is no good single test that can safely and reliably assess the activity of
the HPA, autonomic and prolactin stress response systems in man in either a clinical or
research setting. The aim of this thesis therefore, was to develop a novel stress test
based on the detailed study of the response to CO2 inhalation. A further aim was to use
this test to evaluate the stress response in normal individuals and selected patient
groups. In particular, the thesis aimed to:
i. evaluate the safety, reliability and reproducibility of a single breath of
35% CO2 as a model of the neuroendocrine stress response,
ii. evaluate the mechanisms by which the physiological and
neurohormonal responses occur and are regulated,
iii. evaluate various clinical populations using this model to determine the
influence of disease pathophysiology on the performance of the stress
response, and
iv. determine whether the characteristics of the stress response in specific
disease states are predictive of clinical outcome.
The principal hypothesis that this thesis sought to evaluate, therefore, was that a single
breath of 35% CO2 would reliably produce activation of the HPA and SAM axes.
Activation of other neurohormonal and autonomic components of the stress response
system may also occur and all of these responses could be readily measured in a clinical
setting. Further, it was hyopthesised that activation of these axes was dependent on
initial activation of brainstem noradrenergic centres, but that manipulation of the HPA
35
axis would either inhibit or augment other responses. Finally, it was also hypothesised
that the pattern of response to the challenge would vary predictably in circumstances
where underlying disease processes impacted on the function of the various components
of the stress response and that this information would have clinical relevance.
36
CHAPTER 2
METHODOLOGY
37
2.1. The single breath CO2 model
Existing tools available for the laboratory or clinical study of the neuroendocrine stress
response lack the necessary simplicity, safety, practicality and reproducibility that
would make any one test ideal. One of the principal objectives of this project was the
development of a test that would more closely meet these requirements and would
therefore be a more useful and applicable model for the study of the neuroendocrine
response to stress in man. The preferred model, a single breath of 35% CO2, was
chosen based on recent experience with various CO2 challenges as research tools for the
investigation of panic and anxiety. This included some experience with the
psychological responses to these challenges in experiments performed within the
Neuroendocrinology Research Centre and the Department of Psychopharmacology at
the University of Bristol.
The ability of CO2 to induce anxiety and the increased sensitivity of anxiety sufferers to
CO2 exposure was recognised in the early twentieth century [Drury 1919] and
subsequently during evaluation of its physiological and clinical properties, particularly
as an anaesthetic induction agent [Cohen and White 1951, La Verne 1953, Sechzer et al
1960, Davy 1972, Woods et al 1988, Rassovsky and Kushner 2003]. More recently,
researchers in psychiatry evaluated carbon dioxide inhalation as a potential treatment of
people suffering panic and anxiety disorders [Gorman et al 1984, Van den Hout MA
and Griez 1984, Verberg et al 2001]. Their demonstration that panic sufferers are, in
fact, more sensitive to the effects of CO2 administration led to the more detailed
evaluation of CO2 inhalation and the neurobiological pathways mediating this increased
susceptibility. Since then, several methods have been developed for the administration
of CO2 and the assessment of its response. There has, however, been little in the way of
38
standardisation of models in terms of equipment used, procedure and order of
administration, instruction given and response monitoring [Cohen and White 1951,
Lejuez et al 1998, Verberg et al 2001]. Most response monitoring has been in relation
to the induction of panic attacks the definition of which and its means of assessment has
also varied widely [Lejuez et al 1998].
Amongst anxiety researchers, the CO2 inhalation challenge has become the principal
tool for investigating panic as it has allowed the study of the neurobiological
mechanisms that underlie anxiety, and has also been useful in evaluating panic in the
clinical setting including the evaluation of pharmacological and cognitive therapies. In
contrast to other panic-inducing challenges such as the infusion of sodium lactate,
noradrenaline, CCK-4 and isoproterenol, or the administration of oral yohimbine or
caffeine [Cohen and White 1951], CO2 has proven to be a more efficient and more
sensitive challenge for provoking anxiety [Greiz et al 1990b, Sapolsky 1994, Perna et al
1995a, Lejuez et al 1998, Verberg et al 1998]. CO2 inhalation has also been described
as being a safer challenge with intense, but short-lived symptoms of anxiety [Cohen and
White 1951], with no significant adverse effects [Verberg et al 2001] that was relatively
easy to administer and non-invasive [Lejuez et al 1998].
In general, there exist three separate approaches to the administration of CO2 in humans
for the investigation of panic. Each method has been used widely by several groups and
a degree of variation in the approach to each exists. The three methods, as outlined
below, vary considerably in the equipment used for CO2 delivery, the relative
concentration of CO2 used and the duration of the exposure. The three methods, and
their respective advantages and disadvantages will be described briefly as these
significantly influenced the choice of method used for our studies.
39
i. The steady-state method.
This technique involves the delivery of a fixed concentration of CO2, typically 5 - 7.5%
CO2, for approximately 10-20 minutes or until a panic attack occurs. The subject either
wears a plastic hood placed over the head into which the CO2 is vented, or they breath
through a mask fitted over the nose and mouth into which a mixture of air and the
required CO2 concentration is delivered. The principal advantage of this technique is
that the timing of the CO2 exposure can be carefully controlled thereby reducing the
confounding effects of anticipatory anxiety [Lejuez et al 1998]. Disadvantages are the
significant anxiogenic effects of wearing the respiratory canopy which may also
stimulate other components of the stress response system [Sanderson and Wetzler
1990], the length of time each test takes, the variability in dose received [Gorman et al
1988] and the prolonged length of time subjects are experiencing significant anxiety
symptoms [Lejuez et al 1998].
ii. The Read re-breathing method.
Introduced in 1967 by Read [Read 1967], this technique requires the subject to breath
within a closed system attached to a mouth-piece whilst the nose is occluded with a
nose clip. The CO2, typically 5-7%, is combined with O2, however, as re-breathing of
expired air proceeds, the inspired CO2 concentration increases progressively.
Significant anxiety usually occurs after about 15 minutes. Problems with this method
include the length of time the test takes [Lejuez et al 1998], the variability in CO2
concentration [Gorman et al 1988] and the confounding anxiogenic effect of wearing a
mask and nose-clip for a considerable length of time [Askanazi et al 1980]. Its main
advantage is that it has been the most widely used technique historically.
40
iii. Single or double breath inhalation.
This method involves one or two full vital capacity inhalations of 35% CO2 combined
with either 65% O2 or air [Lejuez et al 1998]. Most commonly the gas mixture is
breathed in through a mask that covers the mouth and nose. The mask is connected to a
gas reservoir via a 3-way tap that is manually opened and closed as the subject receives
specific instructions on when to breathe. Its main advantage is its speed of onset and
very short duration of action. Thus it produces the highest level of systemic CO2 among
the three methods, but only lasts a very short amount of time. Anxiety symptoms are at
least as intense as the other methods (at their peak), but are extremely short lived. Its
main disadvantage is the potential confounding effect of anticipatory anxiety since the
onset of the breath (and therefore the CO2 exposure) is obvious to the patient. In
addition, some subjects find taking a second full vital capacity breath difficult.
Very little work has been done to directly compare the efficacy, safety and reliability of
the three methods. In a small study by Woods et al [1988] the panic rates produced by a
5% CO2 steady-state challenge compared to the Read re-breathing technique were
equivalent (67% vs 75%, respectively). Gorman et al [1994] and Papp et al [1997] both
administered 5% and 7% CO2 using the respiratory canopy method and demonstrated
increased panic rates with the higher dose compared to the lower dose. Further, it is
also not clear whether the three methods represent biologically similar or distinct
challenges based on their mechanism of action. Continuous exposure to low dose CO2
produces hypercapnia and respiratory acidosis [Sanderson and Wetzler 1990, Lejuez et
al 1998], whilst it has been suggested that a single breath of high dose CO2 produces an
initial brief hypercapnia that stimulates ventilation with the resultant hyperventilation
causing hypocapnia (as CO2 is blown) off with subsequent respiratory alkalosis
[Zandbergen et al 1989]. In the former situation panic is caused by the acidosis,
41
whereas in the latter situation it is caused by the alkalosis akin to the infusion of sodium
lactate [Peskind et al 1998].
The simplicity of the 35% single breath method and the rapidity of its apparent response
suggested this method was more suited to our stated test requirements. Additional
evidence that this method would be advantageous came from two studies of the HPA
response to CO2 exposure. As will be discussed in more detail, few studies have
examined stress, particularly HPA, responses to CO2. Results from studies of low dose
continuous CO2 exposure were contradictory with the most recent study and general
consensus favouring no significant HPA response to low dose CO2, despite the
induction of panic in panic disorder patients [Sinha 1999]. On the other hand, a pilot
study of a single breath of 35% CO2 in healthy volunteers performed at the University
of Bristol [Argyropoulos et al 2002] did demonstrate a cortisol response. No detailed
study of the neuroendocrine stress response to 35% CO2 has been conducted. Based on
the evidence above, there appeared a pressing need to evaluate the potential for a single
breath CO2 test to perform reliably in this setting. Consequently, I performed a
systematic and detailed review of the response to a single breath of 35% CO2 was
undertaken with particular regard to developing this as a potentially useful test in
evaluating the role of neuroendocrine stress response systems in health and disease in
man.
2.1.1. Design
Several models of CO2 administration have been described for both experimental and
clinical use. Most designs have relied on some form of manual administration of CO2
42
using a mask or mouthpiece, a 3-way valve and a balloon reservoir. Some groups have
developed computer-controlled automated delivery models [Lejuez et al 1998] that have
greater precision in gas delivery but are significantly more expensive and are useful
particularly for the delivery of low dose CO2 over many minutes. Manually controlled
delivery systems are more suited to and have been used more extensively in single
breath experiments. This was felt to be more appropriate for our model. Pre-mixed 10
L cylinders of medical grade CO2 and O2 were obtained from a commercial source
(BOC Gases, Guildford, Surrey, UK). For most experiments a concentration of 35%
CO2 / 65% O2 was obtained. A 10 L silicone reservoir (Douglas) bag (Hans Rudolf
Inc., Kansas City, Missouri, USA) was connected to the cylinder with silicone tubing
and could be filled directly from the source cylinder. Outflow from the Douglas bag
was through an analogue flow meter (Ohmeda Medical, Columbia, Maryland, USA)
and was controlled manually by a 3-way tap (Hans Rudolf). One port of the 3-way tap
received gas from the Douglas bag via the flow meter, with the opposite port connected
to either a silicone mouthpiece or face mask. The third port remained open to air. With
the subject breathing through the mouthpiece or facemask, the 3-way tap allowed the
investigator to control whether the subject was breathing normal air or the gas from the
Douglas bag. A simplified schematic diagram of the breathing circuit is given in Figure
2.01.
The flow meter allowed inspired vital capacity to be recorded. Based on previous
literature of the 35% CO2 breath test in panic, recording vital capacity is necessary as it
had been suggested that the challenge is not likely to succeed if an inadequate intake of
the gas is taken [Verberg et al 2001]. As such, vital capacity was recorded and a test
breath was considered adequate if it was at least 80% of a baseline air vital capacity
breath.
43
Subject Mouth- 3-way Flow Douglas CO2 piece tap meter bag cylinder
ii. CO2
i. Normal air
Figure 2.01. Schematic diagram of the breathing circuit used to deliver a
single breath of 35% CO2. Subjects breathe through a mouthpiece or facemask. With the tap vertical they would inhale and exhale room air only (i - top panel). With the tap turned horizontally, inhalation would be from the pre-filled Douglas bag containing the CO2/O2 mixture. Flow through the flow meter records inspired vital capacity (ii -lower panel). The tap would be manually returned to the vertical position before expiration so that exhaled gases would be to room air.
44
2.1.2. Procedure
Most tests were performed in a dedicated area of the University of Bristol’s Research
Centre for Neuroendocinology’s Clinical Investigation Unit. As will be discussed, for
the collaborative study with the Autonomic Unit at the National Hospital for Neurology
and Neurosurgery, Queen’s Square, London and St Mary’s Hospital, London, testing
was performed in the Clinical Investigation Unit of the National Hospital for Neurology
and Neurosurgery. For the study involving patients with Addison’s disease, testing was
performed in a dedicated room within the University of Bristol Exercise Physiology
department.
Each unit contained a comfortable chair and a bed and subjects could either remain
seated or resting on the bed during cannula insertion and for the time prior to the test
commencing. All tests were performed with the subject in the seated position.
Written informed consent was provided in all cases. Studies were approved by the local
ethics committee of the North Bristol United Healthcare Trust and where appropriate by
the ethics committees of the National Hospital for Neurology and Neurosurgery,
Queen’s Square, London and St Mary’s Hospital, Praed St, London.
For the initial study, which was essentially a descriptive pilot study, a sample size
power calculation was not performed. For subsequent studies where group comparisons
were being made, a power calculation was performed based on the size needed to detect
cortisol responders from non-responders. This was determined from the difference
between the cortisol response to 35% CO2 (responders) compared to 5% CO2 (non-
responders). Based on an approximate 30% increase in cortisol over baseline with a
standard deviation of the difference between responders and non-responders of 29.4
45
mmol/l, the calculated sample size was 10-12. Based on this, future studies aimed for a
sample size of 10 in each group.
Upon arrival subjects underwent a full clinical interview and physical examination
including baseline blood pressure and had an ECG performed. A 20 gauge intravenous
cannula (Venflon, Viggo Spectramed, Helingsborg, Sweden) was inserted into the
subjects antecubital vein and kept patent with 0.9% saline. Subjects then rested quietly
for 30 minutes and were allowed access to neutral reading material during this time. At
the end of the rest period, patients where shown the device used for the delivery of the
CO2 and were instructed in its use. All subjects received the same instructions.
Subjects where told that they would have their nose occluded with a nose-clip whilst
they breathed through the mouthpiece. The 3-way tap was demonstrated and it was
explained to the subjects that the tap would be held in such a position that they would
only be breathing room air. When instructed they would be asked to take a deep breath
in and exhale fully. The operator would then turn the tap to the open the Douglas bag
port, and the subject would be asked to take one full deep inspiration and hold it for
slow count of 4. The operator would then turn the tap back to the original position and
the subjects allowed to exhale and then asked to breathe normally. After 3 or 4 breaths
of room air through the mouthpiece, the device and nose-clip would be removed from
the subject.
In addition to instruction regarding the procedure, subjects were also given a brief
description of what they might experience from the challenge. Descriptions were
always the same with subjects being told that “the CO2 is harmless but they may feel
some very transient and short-lived feelings that might include feeling breathless and an
some cases anxious. The intensity of these feelings are variable with some people
46
experiencing very little whilst others may experience them more intensely”. Panic
attacks were not referred to at any stage.
At the end of the rest period, with the Douglas bag removed and using the mouth-piece,
3 way tap and flow meter assembly only, subjects performed a number of practice
breaths in order to get used to using the device, understand the instructions and to
record baseline vital capacity. The Douglas bag was then reconnected and filled with
the CO2 / O2 mixture. Physiological monitoring commenced and baseline blood samples
and questionnaires where obtained. Baseline cardiovascular monitoring continued for 5
minutes and at the end of this period the test breath was taken. As soon as the device
was removed, subjects were asked to complete a ‘peak response’ questionnaire and at 2
minutes post-exposure the first blood sample was taken. Further sampling and
questionnaires were collected at 10, 20, 30 and in some cases 40 and 60 minutes
following exposure. Cardiovascular monitoring continued for 5 to 6 minutes following
exposure.
2.1.3. Inclusion/exclusion criteria CO2 challenge tests have been used in psychiatric research for many years and standard
exclusion criteria have been published [Verberg et al 2001]. These criteria are not
based on actual occurrences of harm, but have been surmised based on the known
physiological effects of CO2. Challenge tests to date have been performed without
evidence of any significant adverse events [Lejuez et al 1998, Verberg et al 2001]. The
most likely side effect and therefore potential risk, is the generation of acute anxiety and
panic attacks. Unlike psychiatry research where the frequency of panic attacks is often
the principal outcome measure, panic provocation was not an intended objective of the
47
experiments conducted in this thesis. Indeed, whilst anxiety reactions and emotional
arousal were recorded, all attempts were made to exclude subjects who were felt to be at
significant risk of CO2-induced panic. As described earlier, anxiety sensitivity to 35%
CO2 varies considerably with panic disorder patients being most susceptible even
compared to sufferers of other forms of anxiety disorders such as obsessive-compulsive
disorder, social phobias and generalised anxiety disorder [Greiz et al 1990a, Greiz et al
1990b, Perna et al 1994, Perna et al 1995a, Perna et al 1995b, Verburg et al 1995,
Caldirola et al 1997]. However, some patients with anxiety disorders did respond in a
similar fashion to those with panic disorder [Caldirola et al 1997]. Healthy volunteers
have very low rates of panic in response to the 35% CO2 challenge. Depending on how
panic is defined, panic rates in healthy volunteers have been reported to range from 0%
to 2% [Griez et al 1987, Greiz et al 1990a, Perna et al 1995c]. First degree relatives of
patients with panic disorder, who themselves have never previously experienced panic
attacks, are however, more susceptible to panic in response to 35% CO2 compared to
normal individuals without a family history implying a genetic predisposition and
familial clustering of panic sensitivity [Perna et al 1995b, Perna et al 1995c, Bellodi et
al 1998, Verberg et al 2001]. Since the generation of panic attacks was not a desired or
necessary outcome of these studies, it was felt that subjects with a personal or family
history of panic disorder or other severe anxiety disorder who are more likely to be at
risk of panic attacks in response to the CO2 challenge should be excluded from
participation. Further, it should be noted that even in susceptible individuals, the
experience of experimentally induced panic with CO2 did not increase the likelihood of
spontaneous panic or the risk of panic in response to other panicogens in the weeks and
months following the original exposure [Harrington et al 1996, Perna et al 1999,
Verberg et al 2001].
48
In view of the cardiovascular changes that have been reported to occur in response to
CO2 exposure, particularly an increase in blood pressure, subjects with uncontrolled
hypertension, or those with a history of previous stroke, transient cerebral ischaemia,
cerebrovascular disease, angina, ischaemic heart disease or previous cardiac
arrhythmias were excluded. Severe chronic airways disease and asthma were also
exclusions on the grounds that high dose CO2 may be irritant to the airways and induce
bronchospasm. Other medical exclusion criteria included cerebral aneurysm, epilepsy
and pregnancy.
Normal volunteers were required to be in good physical health without any present or
past psychiatric history and to be medication free for the 2 weeks prior to participation.
Incidental use of simple analgesics was permitted up to the day before the test, but
inhaled or oral glucocorticoids or β-agonists were specifically excluded. As will be
explained in detail, clinical studies involving patients with diabetes, Addison’s disease,
autonomic neuropathy or post-lung transplantation were permitted to remain on their
usual treatments as long as this did not include α or β - agonists or antagonists, anti-
depressants or psychotropic medications or opiates. All subjects were asked to refrain
from alcohol, caffeine or nicotine consumption for 12 hours prior to undergoing testing.
Other inclusion/exclusion criteria that relate to the specific experiments of the clinical
population being tested will be detailed in the description and discussion of these
studies in subsequent chapters.
49
2.1.4. Monitoring
2.1.4.1. Ventilation
CO2 is the principal regulator of ventilation and small increases in PaCO2 significantly
increase ventilatory responses including respiratory rate, tidal volume and minute
ventilation. Several studies have examined changes in these respiratory variables in
panic disorder patients in response to the low dose CO2 challenge [Gorman et al 1988,
Papp et al 1997, Lejuez et al 1998]. Whilst results vary, it has been suggested that
disordered ventilatory responses (baseline hyperventilation, more rapid rises in
respiratory rate and minute ventilation and irregular breathing patterns) were more
common in panic disorder patients and were more likely to predict a panic attack in any
one individual [Gorman et al 1994, Papp et al 1997, Lejuez et al 1998]. Recording
ventilatory responses, however, requires additional equipment including a
pneumotachograph arrangement that can measure both inspiratory and expiratory flow
rates as well as calculate minute ventilation. This needs to be built into the breathing
circuit and requires the subject to be breathing through the circuit at all times.
Given minute ventilation (respiratory rate x tidal volume) is the principal respiratory
variable that changes in response to CO2 exposure, the speed of onset and recovery of
ventilatory responses to a single breath 35% CO2 makes this a less useful physiological
marker of this challenge. This is in contrast to challenges involving prolonged exposure
to low dose CO2. In addition, the additional equipment expense and the lack of
portability of the circuit further limit the utility of measuring respiratory variables if it
includes a pneumotachograph and the necessary amplifiers. Finally, many subjects find
it difficult breathing through a mouth-piece for more that a few breaths increasing the
50
potential for confounding from anxiety provocation associated with the use of a
pneumotachograph.
In view of this, it was felt that recording respiratory variables in this particular challenge
would add unnecessary expense, limit the portability of the test and add little in the way
additional useful information.
2.1.4.2. Psychology
Traditionally, the CO2 challenge has been used as a stimulus of panic, and panic
response rates have been the principal outcome measure of most challenge test to date.
Despite this, there is significant variability in the tools used by different groups to
measure panic or anxiety responses and little consistency between them in the definition
of panic attacks. Studies vary depending on whether they used a behavioural definition
of panic, observation of the subject by an experimenter or self-report measure [Lejuez et
al 1998]. A behavioural definition of panic, that is a request by the subject to terminate
the procedure, and observation of the subject are methods specifically designed to
record panic ‘frequency’. Self-report methods, on the other hand, are designed more
specifically to measure the ‘reactivity’ of an individual to the challenge and may also
allow the intensity of the reaction to be quantified [Lejuez et al 1998, Verberg et al
2001]. In the development of the 35% CO2 challenge, we were more interested in
assessing the degree of emotional arousal generated rather that in inducing panic
attacks, hence self-report methods were preferred.
Self-report tools are essentially pen-and-paper questionnaires that contain specific
symptom checklists. Each symptom has its own analogue scale whereby the subject can
51
rate their experience of that symptom on a continuum from ‘0 – I don’t feel this
symptom at all at the moment’ through to ‘100 – This is the worst I’ve ever experienced
this particular symptom’. These self-report questionnaires, also termed visual analogue
scales, are usually divided into two. The first set of questions relates to subjective global
feelings of anxiety. Questions, for example, include ‘To what degree do you feel
anxious’; ‘to what degree do you feel happy’; ‘to what degree do you feel fearful’ and
‘to what degree do you feel relaxed?’ The second set of questions includes a list of
somatic symptoms that are typically associated with acute anxiety or a panic attack.
Each symptom is rated on a similar scale. Several validated somatic symptom
questionnaires of panic exist including the one used in our studies developed by Nutt et
al [1990]. Questionnaires are administered to assess baseline (immediately before
exposure) and peak (immediately after exposure) experiences, the calculated difference
being the effect of the challenge itself. Examples of the VAS used in the subsequent
studies are attached as appendix 1 and 2.
One problem with this technique is the ‘ceiling effect’. That is responses greater than
100 are not possible therefore for an individual with baseline score of 90, an increase of
only 10 is possible. In other words whether a change from 90 to 100 is the same as a
change from 0 to 10 can be argued.
Some groups have used similar scales to arbitrarily define a panic attack in response to
CO2. In one study [Rapee et al 1992] for example, panic was defined as a score of 1 out
of a maximum of 8 (a score equivalent to 13 out of 100 on the VAS we used) for feeling
fear or panic. Using this criterion, they found panic rates of 65%. The same authors
then altered the criteria to a score of at least 5 (equivalent to at least 63 out of 100) and
found panic rates of 41%. Other groups required a similar a score (50-60 out of 100) in
4 of 13 panic-like symptoms to define a panic attack [Verberg et al 2001]. Neither of
52
these criteria, however, took into account the duration of symptoms, that is whether they
were felt just fleetingly or for several minutes. In the studies described herein, self-
reporting scales were used to rate individual reactivity to the challenge and were not
used to define panic. Panic attacks were defined according to a combination of
observation, behavioural responses and self-reporting whereby subjects who described
severe anxiety with panic-like symptoms that lasted for at least several minutes were
considered to have had a panic attack. As will be detailed for each experiment, all
responses to the 35% CO2 challenge were transient and well tolerated. No subject
described or was observed to have panicked in response to the challenge.
2.1.4.3. Cardiovascular physiology
Cardiovascular responses to CO2 exposure have been studied for many years. Rodents,
mammals and non-human primates have been extensively studied. In addition, many
human studies have been performed, although most have involved chronic or sub-acute
exposure. Studies of acute exposure (single breath) are more limited particularly
recently. As has been described in more detail, results have varied depending on the
method and concentration of CO2 administration, but in general responses have
included an increase in systolic blood pressure and tachycardia [Hardgrove et al 1938,
La Verne 1953, Tenney 1956, Sechzer et al 1960, Tenney 1960, Cross and Silver 1962,
Cullen and Edgar 1974, Woods et al 1988, Argyropoulos et al 2002, Bailey et al 2002].
Increased diastolic blood pressure, cardiac output and stroke volume have been reported
in some, but not all studies in which they were measured [Tenney 1956, Sechzer et al
1960, Tenney 1960, Woods et al 1988]. Similarly, catecholamine release, particularly
noradrenaline and on some occasions adrenaline, has been reported in some but not all
53
studies [Hardgrove et al 1938, Tenney 1956, Sechzer et al 1960, Tenney 1960, Cross
and Silver 1962, Woods et al 1988, Krystal et al 1989]. Reports of changes in total
peripheral resistance have been variable [Hardgrove et al 1938, Cullen and Edgar 1974,
Woods et al 1988].
Techniques for non-invasive measurement of cardiovascular parameters including
systolic and diastolic blood pressure, heart rate, cardiac output, stroke volume, total
peripheral resistance and the pre-ejection period have increased substantially in recent
years. The first experiment described herein was performed using a Finapress®
(Ohmeda) device that measures beat-to-beat blood pressure and heart rate responses
(see Chapter 3). Subjects wear a self-inflating finger cuff that contains a photosensitive
cell connected via a servo-controlled pump that inflates the cuff to maintain a constant
pressure. The cuff is worn on the opposite hand to the arm in which the cannula is
inserted so as to avoid problems with blood sampling. The hand must be held still,
approximately at the level of the heart. This recording device was a very useful tool for
recording continuous changes in cardiovascular parameters. The main problem that
arose with the use of the device was interference related to hand movement. Any
movement of the hand on which the cuff was worn caused significant variation in
readings causing falsely elevated or reduced readings. This was a particular problem at
the time when the CO2 was being inhaled as subjects found it difficult to keep from
moving their hands. The Finapress® proved difficult to use with several results
uninterpretable. Instead the simpler Dynamap (Critikon, Tampa, Florida, USA)
automated blood pressure recording device was used for the majority of studies. With
this device, a self-inflating upper arm blood pressure cuff that measures heart rate and
mean arterial pressure is worn. The device then derives systolic and diastolic blood
54
pressures. It can be set to record values every minute. The device is not sensitive to
the effects of movement in the hand.
A collaborative study performed at the University College of London’s Autonomic Unit
(located within the National Hospital for Neurology and Neurosurgery) took advantage
of the availability of more sophisticated cardiovascular monitoring equipment (see
Chapter 5). In this study, a more advanced form of the Finapress®, the Finometer®
(Ohmeda) was used. This monitor also used the self-inflating finger cuff principal but
was able to also self-correct for hand position relative to the heart and was more
insensitive to interference from hand movement. The Finometer® was also used to
obtain other cardiovascular parameters including cardiac output, stroke volume, total
peripheral resistance and the pre-ejection period.
Also available in this collaboration was a laser flow doppler (Perimed) that was used to
measure skin blood flow responses in the hand (see Chapter 5). The laser doppler uses
two laser probes to measure changes in blood flow in very small blood vessels of the
hand compared to the forearm. Flow changes in these blood vessels of the hand
represent sympathetically mediated vasoconstriction or vasodilatation whilst the
forearm measures correct for the effect of movement.
Samples for plasma catecholamines were taken from most subjects and a High
Performance Liquid Chromatography (HPLC) assay was developed within the
University Research Centre for Neuroendocrinology (see below) for their measurement.
For technical reasons only samples from the first experiment were successfully analysed
with this assay. The Autonomic Unit, National Hospital for Neurology and
Neurosurgery, Queen Square, London, using an established HPLC method, performed
55
the assays on samples from the collaborative study. This group has published
exensively using this assay [see Kimber et al 2000; 2001 as examples]. Intra- and inter-
assay coefficients of variation were 4.7 and 4.3% respectively for noradrenaline, 4.6 and
5.1% respectively for adrenaline.
2.1.4.4. Biochemistry
Blood samples for hormone levels were taken at various time points before and after the
CO2 exposure through the intravenous cannula. For the first experiment, on each
sampling occasion, 15 ml of venous blood was taken in 3 x 5 ml aliquots. One aliquot
was collected into a tube containing SST and assayed within 24 hours for cortisol,
prolactin, GH, TSH, FSH and LH. The remaining 2 aliquots were collected into pre-
chilled tubes containing lithium heparin (LiHep) and EDTA respectively. These
samples were then immediately cold centrifuged at 4°C, the plasma separated and stored
at –50°C until assayed for plasma renin activity, AVP (LiHep) and for ACTH (EDTA).
In the studies where plasma catecholamines were measured, samples were collected into
pre-chilled LiHep tubes to which 1ml of EGTA (anti-oxidant) was added. Samples
were immediately cold centrifuged at 4°C, the plasma separated and stored at –80°C
until assayed. Saliva samples collected using salivettes (Sarstedt) were stored at –20°C
until assayed for cortisol and salivary amylase.
Standard immunometric assays (Immulite 2000, DPC, Los Angeles, California, USA)
were used for plasma cortisol, prolactin, GH, TSH, FSH, LH and salivary amylase.
These assays were performed at cost by the Department of Chemical Pathology, Bristol
Royal Infirmary. Intra- and inter-assay coefficients of variation (CV’s) were 5.3% and
56
7.2%; 2.5% and 6.9%; 3.4% and 5.5%; 2.1% and 4.3%; 3.6% and 6.7%; 3.7% and
5.3%; 2.0 and 3.4%, respectively.
Salivary cortisol assays were performed at cost by Cultech Ltd (Swansea, UK) using the
Neogen cortisol ELISA kits (Neogen, Lamsing, Missouri, USA). Renin was assayed at
cost by the Blood Pressure Unit, Western Infirmary (Glasgow, UK) according to the
method of Miller et al [Miller et al 1980].
ACTH assays were performed by myself within the URCN using a commercially
available radioimmunoassay (DSL, Webster, Texas, USA). Samples were analysed in
duplicate and the CV’s for the assay were 5.9% and 7.3% respectively.
AVP assays were kindly performed by Dr Mary Forsling, King’s College, London using
a double antibody radioimmunoassay following separation from plasma proteins by
methanol extraction and chromatography.
Figures 2.02, 2.03 and 2.04 demonstrate how physiological monitoring equipment was
applied and Figures 2.05 and 2.06 demonstrate the administration of the CO2.
57
a
bc
Figure 2.02. Arrangement used for patient monitoring. a: Laser doppler probes b: Intravenous cannula c: Finometer® finger probe and transducer
58
c
b
a
Figure 2.03. Arrangement used for patient monitoring. a: Laser doppler probes b: Intravenous cannula c: Finometer® finger probe d: Laser doppler
d
59
c
b
a
Figure 2.04. Arrangement used for patient monitoring. a: Finometer® monitor b: Laser doppler c: Laser doppler monitor
60
a
b
c d
e
Figure 2.05. Arrangement for CO2 delivery.
a: Nose-clip b: Mouth-piecec: Manual control valve (3-way tap) d: Flow meter e: Douglas bag
61
a c d
be
f
Figure 2.06. Arrangement for CO2 delivery.
a: Nose-clip b: Mouth-piecec: Manual control valve (3-way tap) d: Flow meter e: Douglas bag f: CO2 cylinder
62
2.2. Plasma catecholamine HPLC
A number of methods have been developed for the measurement of catecholamines
(adrenaline, noradrenaline and dopamine) in human plasma. HPLC with
electrochemical detection being the most commonly used technique as it provides the
greatest sensitivity and specificity. Since its initial description by Kissinger et al
[Refshauge et al 1974] in 1973, several modifications and variations have been
described, however, it remains a difficult assay to perform [Forster and Macdonald
1999, Raggi et al 1999].
The principal of electrochemical detection is based on the detection of a change in an
electrical current maintained between two electrodes that occurs as the result of
oxidation or reduction of the substance of interest [Raggi et al 1999]. The electrical
potential needed to oxidise or reduce a particular substance (polarising potential) is
maintained between two electrodes. An electrochemical substance passing between the
electrodes will be oxidised or reduced leading to a gain or loss of electrodes. The
resulting current is detected by a measuring instrument, amplified and displayed as a
chromatographic signal. Since only a limited number of substances have the same
polarising potential, electrochemical detection typically has a high sensitivity and
specificity for identification of the substance of interest. Specificity, however, is
diminished if several compounds are oxidised or reduced at a similar polarising
potential. Interference from catecholamine metabolites and other substances within the
plasma or derived from the extraction procedure (see below) or mobile phase is one of
the major problems with plasma catecholamine HPLC assays [Forster and Macdonald
1999]. The second major problem is the low yield of plasma catecholamines in the
plasma (especially under resting conditions) [Raggi et al 1999]. Specificity and
63
sensitivity of the assay can be improved by extracting catecholamines from the plasma
first (purification step) in order to remove interfering substances and to increase the
yield of catecholamines within the sample as much as possible. This purification step
involves binding the catecholamines to alumina [Benedict 1987, Raggi et al 1999], a
boric acid gel [Imai et al 1988] or organic solvents [Smedes et al 1982, Forster and
Macdonald 1999] under alkaline conditions, washing or removing the remaining plasma
and as many impurities as possible and then, under acidic conditions, unbinding and
collecting the catecholamines prior to undertaking chromatographic analysis [Raggi et
al 1999].
Sample collection is equally important, as catecholamines are rapidly oxidised at room
temperature. Sample collection requires pre-chilled heparinised tubes, with the
immediate addition of an anti-oxidant such as EGTA, cold centrifugation at 4°C and
freezing at –20 to –80°C prior to analysis.
Commercially available kits are available for the extraction phase of the assay.
However, we initially chose to develop an assay based on published protocols for the
extraction of catecholamines from human plasma [Forster and Macdonald 1999, Raggi
et al 1999] and advice from Dr Ian MacDonald, School of Biomedical Sciences,
Nottingham University Medical School, Queen’s Medical Centre, Nottingham, UK.
These protocols used the addition of alumina to a plasma sample for catecholamine
binding, washing of the alumina to remove impurities and the unbinding the
catecholamines with perchloric acid. Protocols used in the development of the assay are
given in appendices 3, 4 and 5.
Once extracted, catecholamines were run through a 15 x 3.9 mm reverse phase
Bondapak column (Waters, Milford, Massachusets, USA) with a Waters 510 pump and
64
a programmable sample processor (Waters 710-B WISP). Quantification was on a
Waters M-460 dual electrochemical detector with a working potential of +0.74 V.
Mobile phase was HPLC grade water with 130 ml methanol, 35 mg EDTA, 9.53 g
potassium dihydrogen orthophosphate dihydrate (KH2PO4) and 200 g octanesulphonic
acid to a pH of 3.4.
Standard solutions of adrenaline and noradrenaline with 3,4-dihydroxybenzylamine
hydrogen bromide (DHBA) as internal standard were made and standard curves created.
Sample concentrations were estimated from comparisons of their peak heights with
known concentrations of pure standards. Calculated catecholamine concentrations were
expressed as pg/ml plasma. Examples of obtained chromatograms are shown in Figures
2.07, 2.08 and 2.09. To determine the sensitivity of the assay, a lower limit of
dectection study was performed. This involved serial dilutions of known concentrations
of stantards (noradrenaline and adrenaline). Concentrations tested were 1000, 500, 250,
25 and 2.5 pg/ml. Serial dilutions of plasma samples spiked with known concentrations
of standards were also performed to assess recovery rates. Technical difficulties (as
described below) hampered the interpretation of these procedures and further attempts
to optimise the essay were not pursued.
Sample recovery using this technique varied from 50 – 65%, but the technique was used
to measure catecholamines from the first CO2 dose response study performed. Shortly
after this, however, a laboratory fire damaged several pieces of equipment including the
HPLC system. It was decided by the laboratory to replace the system with the more
up-to-date Coulochem II® system (ESA Inc, Chelmsford, Massachusets, USA). This
system combined the Coulochem II® Electrochemical detector with a 582 solvent
delivery module (ESA) and an ESA 542 Autosampler. The system also used the EZ
Chrom Elite™ (ESA) software system for chromatogram display and calculations.
65
With this system, it was also decided to obtain catecholamine extraction kits from the
same company. These kits used syringes pre-loaded with an alumina matrix (PCAT
analysis kits, ESA) and a vacuum manifold (ESA) in order to run plasma samples
through the alumina. Once bound, a supplied acidic eluting solution was run through
the same matrix to extract the catecholamines. Samples were then run through an MD-
150, 3 µm, 3.2 x 15 mm reverse phase column (ESA) and quantified with a dual
electrode detector (ESA Coulochem II®). Internal standard was again DHBA and
calculated catecholamine concentrations were expressed as pg/ml plasma. Mobile
phase was the Cat-A-Phase® II Mobile Phase (ESA) provided as part of the kit. Pure
standards provided with the kit were used to create a standard curve. Recovery using
this method varied from 30 – 75% and was often inconsistent. Examples of
chromatograms are shown in Figures 2.10, 2.11 and 2.12. Several attempts using
spiked plasma and experimental samples were made using this new system and
methodology were made but results were inconsistent and unreliable and have not been
reported in any of the studies where this methodology was used.
66
Figure 2.07. HPLC chromatogram of a standard mixture of
catecholamines (noradrenaline, adrenaline, DHBA (internal standard) and dopamine.
67
Figure 2.08. HPLC chromatogram following successful extraction of catecholamines from a plasma sample.
68
Figure 2.09. HPLC chromatogram following successful extraction of catecholamines from a plasma sample.
69
Figure 2.10. HPLC chromatogram demonstrating poor recovery of
catecholamines (including internal standard) from a plasma sample.
70
Figure 2.11. HPLC chromatogram following unsuccessful extraction of catecholamines from a plasma sample.
71
Figure 2.12. HPLC chromatogram following unsuccessful extraction of catecholamines from a plasma sample.
72
2.3. Salivary amylase
In 1996 Chatterton and others [Chatterton et al 1996] described the use of salivary α-
amylase as a measure of the plasma catecholamine response to various stressors in
humans. The physiological role of salivary amylase is principally in the initiation of
starch digestion, but it also has a role in the maintenance of tooth integrity, inhibition of
bacterial adherence and colonisation and reducing available sugar substrate for bacterial
growth [Pedersen et al 2002]. The overall composition of saliva, including the
concentration of α-amylase, is strongly dependent on salivary flow rate [Pedersen et al
2002]. This was thought to be the likely mechanism by which α-amylase levels reflect
plasma catecholamine levels. With sympathetic stimulation, salivary flow and therefore
α-amylase levels, increase proportionally. However, salivary flow rate is dependent on
many different factors, including the type and size of the principal salivary gland from
which the saliva is being secreted as these differ in their response to stimulation and in
their secretory composition [Busch et al 2002, Pedersen et al 2002]. Other factors
include the state of hydration; nutritional state; the time of day as salivary flow follows
a circadian rhythm; the nature and duration of the stimulus; emotional state and gender
[Pedersen et al 2002].
In the study by Chatterton et al [1996], salivary amylase and plasma noradrenaline and
adrenaline levels were measure following exposure of normal individuals to a variety of
stressors including exercise, written examination and thermal stress. They found a
significant correlation between α-amylase and noradrenaline following both exercise
and examination stress. Neither adrenaline nor salivary cortisol correlated with α-
amylase. The α-amylase level also increased with both heat and cold, although
corresponding catecholamine levels were not measured. Based on these results, it was
73
concluded that salivary α-amylase levels reflected plasma catecholamine levels
particularly noradrenaline [Chatterton et al 1996].
Several animal studies exist that have shown sympathetic stimulation increases both
salivary flow and α-amylase levels, although this is dependent on both the stimulus and
the gland being studied. Busch et al [2002] studied α-amylase release from rat parotid
and sub-mandibular glands following administration of isoproterenol (a β-adrenergic
agonist). In this study, β-adrenergic stimulation increased α-amylase release from the
parotid but not the sub-mandibular glands. Similar studies of the effect of sympathetic
stimulation and inhibition on salivary flow and α-amylase release have not been
performed.
Following the initial study by Chatterton et al [1996], he and several other groups have
published studies that have reported salivary α-amylase as a surrogate marker of plasma
catecholamine/noradrenaline release. In a study of the hormonal response to skydiving,
Chatterton et al [1997] identified significantly elevated α-amylase levels on the day of
the jump compared to a rest day, with a significant rise in α-amylase levels from
baseline following the jump. This pattern was similar to noradrenaline levels measured
in other parachute jump experiments [Richer et al 1996] and the authors conclude this is
further evidence validating the use of salivary amylase as a marker of catecholamine
activation [Chatterton et al 1997]. Interestingly, however, in the first parachute jumping
study [Chatterton et al 1996], there is no mention of other features of sympathetic
stimulation including heart rate and blood pressure changes. In the second study,
[Richer et al 1996], heart rate responses correlated with adrenaline release but not with
noradrenaline.
74
Nater et al [2002] measured salivary α-amylase and plasma catecholamine responses to
a psychological stressor (the TSST). Whilst the test increased salivary α-amylase,
noradrenaline and adrenaline significantly, there was no correlation between the α-
amylase response and the noradrenaline or the adrenaline response.
Several other groups have reported salivary α-amylase responses to a variety of
stressors as a surrogate of plasma catecholamines. Xiao et al [2000] examined the
salivary α-amylase response to three simulated trauma cases in 10 anaesthetists. The
first case simulated routine management (and therefore low stress), whilst the 2nd and
3rd cases were designed to contain stressful management events. Noradrenaline, blood
pressure and heart rate responses were not measured. Salivary amylase levels were high
before starting and after the first 2 cases (routine and stressful), but not after the third
(stressful) case indicating a lack of correlation between the degree of perceived stress
and α-amylase responses.
Hojo et al [2003] measured α-amylase as well as blood pressure, heart rate and cortisol
responses to a distressing video challenge in healthy volunteers. Their results showed a
significant increase in both α-amylase and cortisol in response to the challenge, but no
change in either blood pressure or heart rate. Similarly, Morrison et al [2003] studied
intensive care nurses during real life working stress. They found a significant
correlation between noise stress, subjective annoyance and heart rate changes. Salivary
amylase levels showed large inter-individual variation and did not correlate with heart
rate or subjective measures of stress.
In summary, following a single study showing a correlation between salivary α-amylase
and noradrenaline during an acute stress paradigm [Chatterton et al 1996], several
75
different authors have accepted this as a surrogate marker of catecholamine (particularly
noradrenaline) production. However, one study failed to show a significant correlation
between salivary α-amylase and either noradrenaline or adrenaline, whilst others have
failed to measure or show a relationship between α-amylase and other markers of
sympathetic activation including heart rate and tachycardia. Given the difficulty in
collecting and measuring plasma catecholamines, a simple surrogate marker would be
extremely useful in stress research. The simplicity of measuring α-amylase and the
apparent initial correlation makes this test extremely attractive, however, the evidence
as summarised above appears conflicting.
As part of the studies described in this thesis, salivary amylase levels were measured in
both healthy subjects undergoing the 35% CO2 challenge, as well as several of the
clinical populations. The further evaluation and validation of salivary α-amylase as a
non-invasive surrogate marker of sympathetic activation is due to be further
investigated by other researchers at the University of Bristol. Planned studies include
the measurement of α-amylase in response to the insulin hypoglycaemia challenge;
further studies of the 35% CO2 challenge correlating α-amylase levels with
catecholamine responses; and studies of α-amylase in response to α and β-adrenergic
stimulation and inhibition in humans.
In this thesis, for each individual study where α-amylase was measured, the effect of the
CO2 challenge will be described in detail. However, in summary, there appeared to be
marked inter-individual variation in the α-amylase response to the challenge. This
variation included both the intensity and the timing of the response. Further, in healthy
individuals, there was no correlation between amylase responses and noradrenaline,
SBP, DBP, HR, cortisol, prolactin or any of the psychological parameters measured.
76
Similarly, despite some controls and patients groups differing significantly in baseline
noradrenaline, cortisol, SBP or heart rate, there was no difference in their resting
amylase levels.
It appears therefore that salivary α-amylase may not be a good surrogate marker of
sympathetic stimulation, at least in this challenge paradigm. The further studies
mentioned above should help to clarify its role as a non-invasive measure of activation
of the sympathetic stress system.
77
CHAPTER 3
THE 35% CO2 MODEL: INITIAL DESCRIPTION
78
3.1. Rationale for the use of CO2 as a neuroendocrine stressor
There were several reasons for thinking that CO2 exposure would be an appropriate
challenge for the investigation of the stress response. Ideally, the most useful challenge
paradigm would be one that:
i. Is safe for use in humans,
ii. Reliably produces emotional, hormonal and autonomic responses akin to those
described for the general stress response,
iii. Produces responses that are consistent, reproducible and occur in a dose dependent
manner [Bourin et al 1998],
iv. Is simple to administer and produces responses that are easily recorded,
v. Is inexpensive and portable.
As mentioned, several laboratory based paradigms for the evaluation of the stress
response exist and the specific pros and cons of these existing tests have already been
detailed. Existing paradigms have used a variety of psychological, physical and
pharmacological stressors. However, all of these tests have their own specific
limitations, risks and benefits and as yet none fulfils all the criteria listed above. The
lack of an existing suitable neuroendocrine stressor has led us to evaluate CO2 as a
potential model of the stress response in order to develop a test that might have broad
use in both a research and clinical setting.
79
3.1.1. CO2 physiology
Spontaneous respiration is an automatic function that originates in the brainstem and
relies on the rhythmic firing of ‘inspiratory’ and ‘expiratory’ neurons [Nattie 1999].
Control of breathing relies on several chemical (O2, CO2 and pH) and mechanical
factors (lung volume changes, respiratory and other muscle and joint inputs) [Thews
1983, Nattie 1999]. Central chemoreceptors, located predominantly on the ventral
medullary surface of the brainstem are sensitive particularly to changes in arterial and
alveolar CO2 concentrations and provide the principal means of regulating metabolic
acid-base balance [West 1974, Nattie 1999]. Small changes in CO2 levels produce
rapid and dramatic changes in minute ventilation (the product of respiratory rate and
tidal volume). Ventilation, for example, doubles in response to a 2 mmHg rise in
PaCO2 [Nattie 1999]. This is in stark contrast to changes in O2 levels which, sensed
exclusively through peripheral chemoreceptors, have little effect on minute ventilation
until PaO2 levels have fallen from normal values (100 mmHg) to around 60-70 mmHg
[Thews 1983]. More recently, evidence has accumulated indicating that central CO2
chemoreceptors are more widely distributed than previously thought. In addition to the
ventral medullary surface, CO2 chemoreceptors are present throughout the brainstem,
including such areas as the locus coeruleus, raphe complex and pons and are also
present within the cerebellum [Nattie 1999]. In addition to their anatomical
distribution, these chemoreceptors also appear to vary in their sensitivity and thresholds
to activation by CO2 and it has been postulated that this reflects the importance of CO2
in the regulation of other homeostatic functions beyond ventilation [Nattie 1999]. Some
of these functions include sleep and sleep arousal, the regulation of metabolic rates for
temperature control and change in ventilation in order to sustain the demands of
exercise [Nattie 1999].
80
From an evolutionary perspective, it has been suggested that hypercapnia (such as may
occur in impending suffocation) is perceived by the organism as an immediate threat to
life, and the response demands not only a change in ventilation, but also behavioural
arousal, autonomic cardiovascular activation and a motor response in order to escape
from this threat [Nattie 1999, Brannan et al 2001, Liotti et al 2001]. It is not surprising
therefore, that exposure to CO2 will consistently produce not only hyperventilation and
an increase in minute ventilation, but will also produce cardiovascular activation
(including tachycardia and hypertension) as well as marked behavioural and emotional
arousal [Hardgrove et al 1938, Davy 1972, Sechzer et al 1960, Cross and Silver 1962,
Gorman et al 1988, Woods et al 1988, Greiz et al 1990a, Greiz et al 1990b, Perna et al
1994, Papp et al 1997, Bourin et al 1998, Verberg et al 1998, Sinha 1999, Argyropoulos
et al 2002, Bailey et al 2002].
3.1.2. CO2 as a psychological stressor
The feelings generated when inhaling high doses of CO2 were noted as far back as 1800
[Davy 1972], and it was first postulated as a trigger for anxiety in 1951 by Cohen and
White [Cohen and White 1951]. Detailed experiments, however, were only first
conducted in the early 1980’s. Gorman and others [Gorman et al 1984, Verberg et al
2001], based on the association of panic with hyperventilation (with resultant
hypocapnia) and the observation that attacks could be prevented by breathing into a
paper bag (to raise PaCO2), hypothesised that breathing low doses of CO2 (5%) should
reduce anxiety sensitivity. Paradoxically, however, it was found that more subjects
panicked in response to CO2 than did in response to hyperventilation [Gorman et al
81
1984]. Subsequent to this, several groups have studied the anxiogenic effects of CO2
exposure in more detail. As has been discussed, several different paradigms have been
used to administer CO2 to both healthy volunteers and patients with a variety of anxiety
related disorders [Read 1967, Sanderson and Wetzler 1990, Lejuez et al 1998]. Low
dose CO2 exposure (for example, breathing 5-7.5% CO2 over 10 – 15 minutes) will
provoke the gradual onset of hyperventilation and feelings of breathlessness or air
hunger [Gorman et al 1988, Woods et al 1988, Papp et al 1997, Brannan et al 2001,
Liotti et al 2001]. Additionally, it is associated with many of the somatic symptoms of
acute anxiety including dizziness, blurred vision, feeling hot or flushed and feeling that
the heart is racing. In addition, in some individuals low dose CO2 will provoke panic
attacks [Gorman et al 1984, Van den Hout MA and Griez 1984, Rapee et al 1992,
Gorman et al 1994, Rassovsky and Kushner 2003].
A single breath of 35% CO2, based on a model used by Wolpe [1958, Verberg et al
2001] in 1954 for the treatment of anxiety, was investigated further by several groups
including Van den Hout and Griez [1984]. Studies by these groups found that this
challenge provoked strong somatic symptoms associated with anxiety as well as
autonomic activation in most subjects, but induced specific feelings of anxiety and
panic only in subjects with an underlying diagnosis of panic disorder [Van den Hout
MA and Griez 1984, Griez et al 1987, Greiz et al 1990a, Verberg et al 2001]. Since
then, numerous groups performing similar studies have reinforced the reliability and
safety of CO2 as a biological challenge [Greiz et al 1990b, Perna et al 1994, Perna et al
1995a, Verburg et al 1995, Verberg et al 1998]. These groups have also identified
panic disorder patients, as having a specific sensitivity to CO2 exposure, as opposed to
patients with other forms of anxiety disorders [Perna et al 1995a, Verburg et al 1995,
Caldirola et al 1997]. Further, they have also indicated an inherited tendency towards
increased CO2 sensitivity in first degree relatives of panic disorder patients [Perna et al
82
1995b, Perna et al 1995c, Harrington et al 1996, Perna et al 1999] and the importance
of specific pharmacological agents in treatment of panic attacks [Van den Hout et al
1987, Pols et al 1991, Pols et al 1993, Pols et al 1996a, Pols 1996b, Bertani et al 1997,
Perna et al 1997, Nardi et al 2000, Battaglia et al 2001, Meiri et al 2001, Perna et al
2002, Bertani et al 2003].
3.1.3. CO2 as a cardiovascular stressor
Much of the existing literature on autonomic cardiovascular responses to CO2 exposure
stems from physiological studies of low dose CO2 related to the use of CO2 in
anaesthesia and mild to moderate hypercapnia as it occurs particularly in respiratory
disease and artificial ventilation. The commonly used models of both low and high
dose CO2 exposure in psychiatric research have only very infrequently recorded
autonomic responses.
Studies of chronic low dose (5-10%) CO2 exposure in animals and non-human primates
have described sympathetic nervous system stimulation with tachycardia, increased
systolic and diastolic blood pressures and an increase in catecholamine metabolites
[Tenney 1956, Woods et al 1988, Krystal et al 1989]. In one animal study [Witzleb
1983], CO2 exposure during artificial ventilation (with respiratory rate held constant)
resulted in a significant pressor response but with bradycardia.
In humans, similar CO2 concentrations have consistently reported increases in systolic
blood pressure and most have reported tachycardia [Hardgrove et al 1938, Sechzer et al
1960, Tenney 1960, Rammana Reddy et al 1986, Gorman et al 1988, Woods et al
1988]. Studies vary, however, in the observed changes in other cardiovascular
83
parameters, with some showing increases in cardiac output, stroke volume and
decreases in total peripheral resistance [Sechzer et al 1960, Gorman et al 1988]. Other
studies in contrast have shown no change in cardiac output or stroke volume and
increases in total peripheral resistance [Kety and Schmidt 1948, Tenney 1960, Witzleb
1983]. Similarly, despite the consistency of blood pressure responses, catecholamine
production has been inconsistent with some studies showing an increase in both
adrenaline and noradrenaline [Sechzer et al 1960, Tenney 1960] with others showing
either an increase in noradrenaline only [Rammana Reddy et al 1986] or no
catecholamine response at all [Woods et al 1988]. The variance in responses between
the studies has been attributed to the significant variation in CO2 dose and duration of
exposure as well as differences in monitoring techniques [Tenney 1960].
A single study of brief exposure (25 seconds) to 20% CO2 [Zvolensky et al 2001]
reported a significant increase in heart rate, but did not examine blood pressure or other
physiological responses. Very few of the 35% CO2 challenge studies have evaluated
cardiovascular responses. One study by Mieri et al [2001] showed an increase in heart
rate with a trend towards increased noradrenaline levels, whilst the recent study by
Argyropoulos et al [2002] showed a significant increase in systolic blood pressure that
was associated with a significant fall in heart rate.
3.1.4. CO2 as a neurohormonal stressor
Studies of the neurohormonal response to CO2 exposure are limited and the results
conflicting. In a dose ranging study of prolonged CO2 exposure in rhesus monkeys
(5%, 7.5% and 10% CO2 for up to 180 minutes), Krystal et al [1989] showed a dose
84
dependent increase in cortisol, GH and prolactin. Similarly, Woods et al [1988] showed
a dose dependent effect of low dose CO2 exposure in humans. In this study, 15 minutes
of 5% CO2 had no effect on cortisol levels, whilst there was a small increase in cortisol
following the same duration of 7.5% CO2. Neither dose produced a significant change
in GH or prolactin levels. Other similar low dose exposure studies have also shown a
significant increase in cortisol levels [Sechzer et al 1960, Tenney 1960]. In contrast,
Sinha et al [1999] did not demonstrate cortisol release in panic disorder patients with a
5% CO2 paradigm despite generating significant panic attacks in this subgroup.
In response to a single breath of 35% CO2 Argyropoulos et al [2002] showed a
significant cortisol rise, although in contrast, no cortisol increases where seen with this
challenge in studies by Meiri et al [2001] or Van Duinen et al [2004, Griez and
Schreurs 2003]. Meiri et al [2001] was also unable to demonstrate a prolactin response
despite significant anxiety responses. Differences in responses have again been
attributed to the small number of studies done with varying methodologies.
In summary, the existing literature regarding CO2 is extensive with a large amount of
data available on its use and safety in humans. Models of CO2 exposure used for the
investigation of anxiety disorders are well described and have been shown to be safe
and reliable. Data regarding autonomic and neurohormonal responses are more limited
but the existing literature supports the need for further evaluation of this model as a
potential means of investigating the stress response.
85
3.2. Experimental plan
The specific aim of this first series of experiments was to determine the ease, safety and
reliability of a single breath of carbon dioxide in producing a stereotypical
neuroendocrine stress response. In order to do this, a dose response study incorporating
four increasing concentrations of CO2 was undertaken so as to determine the optimum
concentration that would generate a stress response without significant adverse effects.
The four concentrations tested were 5%, 25%, 35% and 50%. All were combined with
pure oxygen to create mixtures that were all hyperoxic relative to air.
Once an optimal dose had been identified, the reliability and reproducibility of this dose
was assessed in three separate experiments. Firstly, reproducibility was assessed by
comparing responses to one exposure to 35% CO2 that was then repeated in the same
individual after an interval of either one week or 6 months. In a subsequent experiment,
responses at different times of the cortisol circadian cycle were evaluated.
86
3.3. Dose response study
3.3.1. Introduction and methods
Healthy adult volunteers were recruited by advertisement from within the University of
Bristol and were paid for participation in the studies. All potential participants
underwent a full clinical interview, examination and electrocardiogram. Exclusion
criteria, monitoring and procedural details were as described in the methodology section
(Chapter 2). The ethics committee of the United Bristol Healthcare Trust approved the
study and all subjects provided written informed consent prior to participation.
Nine subjects (5 male) with a mean age of 27.2 years (range 24 – 36) were enrolled and
all completed the study. Identical cylinders of pre-mixed CO2 and O2 (BOC gases)
containing each of the four concentrations to be tested were obtained and colour coded
to maintain blinding. Subjects attended once a week for four weeks and received each
of the four concentrations once, in random order. Visual analogue scale questionnaires
and blood samples were taken simultaneously every 15 minutes, commencing 30
minutes before exposure and continuing for 60 minutes following. An additional
sample was taken at 2 minutes following exposure. Blood and saliva sampling for
cortisol, prolactin, plasma renin activity (PRA), GH, TSH, FSH, LH, ACTH and plasma
catecholamines was as according to the protocol already described.
Continuous pulse and blood pressure recordings were made with the Finapress recorder
(Ohmeda). Cardiovascular responses to each concentration of CO2 were determined by
calculating the mean of every 5 beats derived from the continuous beat-to-beat monitor.
The maximum change from a baseline mean 30 seconds prior to exposure was then
87
calculated for both systolic blood pressure and heart rate. Similarly, for each specific
hormone, the absolute maximum change from baseline was also determined. These
were then subject to a mixed model analysis in order to account for the presence of both
fixed (dose) and random (subject) effects at each time point. Correlations were
determined using Pearson’s correlation coefficient. All data are presented as mean +/-
SEM. A p-value of less than 0.05 was considered significant.
3.3.2. Results
Anxiety symptoms were transient and well tolerated. There were no other significant
adverse effects from the CO2 exposure and no subjects described or were observed to
have signs or symptoms suggestive of panic.
Five of the 9 subjects were unable to take a full vital capacity breath of 50% CO2
(inspired VC < 80% of baseline). It has been suggested this may be related to
oropharyngeal spasm [146] induced by the high CO2 concentration. Since more than
half the subjects were unable to achieve an adequate breath of this dose, their results
have been not been included in the statistical analysis. The test breath from all other
exposures was considered adequate (VC > 80% of baseline).
Psychological response
Prior to exposure, subjective feelings of anxiety, fear and breathlessness were minimal
(Table 3.01) and did not change significantly in anticipation of the test. Peak increases
in these three symptoms occurred immediately following the exposure and had returned
to baseline within minutes. The peak increase in subjective feelings of anxiety,
breathlessness and in the 5 most commonly experienced somatic symptoms was
88
significantly greater following 25% than after 5% CO2 (p < 0.05), and similarly, these
increased further following 35% compared to 25% CO2 (p < 0.05). Subjective feelings
of fear increased significantly after 25% compared to 5% (p < 0.05), but did not
significantly increase further after 35% CO2. The symptoms experienced consistently
by most subjects and which accounted for most of the difference in experience between
the various CO2 doses included: feeling short of breath; feeling hot; dizziness; feeling
light-headed or faint, blurred or narrowed vision and having difficulty concentrating
(Figure 3.01).
Physiological response
Cardiovascular responses following CO2 exposure are shown in Figure 3.02. Baseline
cardiovascular parameters were not significantly different (Table 3.01). From the
continuous beat-to-beat recording, the mean of every 5 beats was calculated to give a
single reading from which peak changes were calculated. SBP and HR increased
slightly in all subjects immediately prior to all doses, however, no significant change in
blood pressure or heart rate occurred following exposure to either 5% or 25% CO2. In
contrast, following exposure to 35% CO2, a significant bradycardic effect occurred that
persisted for up to 90 seconds (mean peak change in heart rate –22.1 +/- 5.4 b/m
compared to 7.6 +/- 4.2 b/m following 5% CO2 and 0.5 +/- 1.9 b/m following 25% CO2;
p < 0.001). The bradycardia associated with 35% CO2 was followed by a significant
rise in SBP (mean peak change in SBP 28.5 +/- 6.7 mmHg following 35% CO2
compared to 0.2 +/- 3.2 mmHg following 5% CO2 and –7.1 +/- 4.6 mmHg following
25% CO2; p < 0.001).
89
Biochemical response
Baseline concentrations of measured hormones were equivalent prior to each CO2
exposure (Table 3.01). Neither 5% nor 25% CO2 had a significant effect on any of the
measured hormones. 35% CO2 on the other hand produced a significant increase in
ACTH, plasma and salivary cortisol, prolactin and noradrenaline but had no effect on
adrenaline, AVP, PRA, TSH, GH, LH or FSH. The peak change in ACTH and
noradrenaline occurred at 2 minutes post 35% CO2 exposure, whilst the peak effect on
plasma and salivary cortisol and prolactin occurred at 15 minutes. Figures 3.03 and 3.04
show peak individual and mean responses for these hormones.
Correlations
Following exposure to 35% CO2, there was a significant positive correlation between
the peak change in plasma cortisol and the peak change in SBP (p = 0.02, R2 = 0.607)
(Figure 3.05) but not with the peak HR change. A significant positive relationship
between plasma cortisol change and the peak change in anxiety scores was also present
(p = 0.04, R2 = 0.466). No other significant correlations were present.
90
Table 3.01 Baseline physiological, psychological and hormonal characteristics prior to each CO2 exposure.
5% 25% 35%
Physiology Systolic BP (mmHg) 129.3 +/- 11.4 135.8 +/- 9.0 135.9 +/- 8.9 Diastolic BP (mmHg) 79.6 +/- 8.7 75.1 +/- 5.5 77.1 +/- 4.2 HR (beats/min) 75.1 +/- 4.2 79.1 +/- 5.1 72.1 +/- 1.7 Temperature (°C)
36.5+/-0.2 36.5+/-0.1 36.3+/-0.2
Psychology Anxiety (mm) 8.3+/-2.2 10.0+/-3.1 12.2+/-3.1 Fear (mm) 5.0+/-2.7 4.4+/-1.7 5.0+/-1.8 Breathlessness (mm) 1.1+/-0.7 1.1+/-0.8 1.6+/-1.1 Relaxed (mm)
75.0+/-5.5 75.0+/-6.0 73.8+/-8.4
Endocrinology Plasma cortisol (nmol/l) 285.8+/-43.4 300.2+/-35.9 318.7+/-43.8 Salivary cortisol (nmol/l) 90.4+/-31.1 102.3+/-23.4 126.8+/-33.6 ACTH (pg/ml) 46+/-0.3 49+/-0.6 45+/-0.2 PRL (nmol/l) 128.1+/-11.6 156.7+/-15.3 155.4+/-16.1 AVP (pg/ml) 1.4+/-0.2 1.1+/-0.3 1.0+/-0.6 FSH (IU/l) 3.2+/-0.8 2.8+/-0.6 2.4+/-0.4 LH (IU/l) 2.8+/-0.7 2.1+/-0.3 2.0+/-0.5 GH (mU/l) 4.6+/-2.5 5.5+/-2.1 8.2+/-4.2 TSH (mU/l) 1.6+/-0.3 1.4+/-0.4 1.5+/-0.3 Renin (fmol/l) 16.7+/-3.1 17.0+/-3.0 12.1+/-2.8 Noradrenaline (pg/ml) 359.1+/-53.2 N/A 431.0+/-68.8 Adrenaline (pg/ml) 92.4+/-32.1 N/A 95.5+/-44.2 n=9 (all subjects received all three doses in random order) All differences are non-significant. BP = blood pressure; HR = heart rate.
91
Anxiety, fear andbreathlessness
anx0
25
50
75
1005%25%35%
mm
b
a
Con
cent
ratin
g0
25
50
75
mm
b
a
Diff
icul
ty
Figure 3.01. Subjspecific somatic sya:p<0.05 5% vs 25b:p<0.05 25% vs 3
a
iety fe
b
Somatic sy
Feel
ing
hot
b b
a
ective anxiety,mptoms (botto% CO25% CO2
a
ar breathle
mptoms
Diz
zine
ss
Ligh
thea
ded
b
a a
fear and breathm) following C
ssness
Blu
rred
vis
ion
b
a
lessness (top) and O2 exposure.
92
Systolic Blood Pressure
70 90 110 130 150
100
120
140
160
1805%25%35%
beat
mm
HgSystolic Blood Pressure
5% 25% 35%
-50
0
50
100
∆m
ax S
BP(
mm
Hg)
Heart Rate
70 90 110 130 150
50
60
70
80
90
1005%25%35%
beat
bea
ts/m
in
Heart Rate
5% 25% 35%
-50
-25
0
25
50
∆m
ax H
eart
Rat
e(b
eats
/min
)
a
b
Figure 3.02. Systolic blood pressure and heart rate responses to a single breath of 5%, 25% and 35% CO2 represented as mean response over time (left hand panels with exposure represented by hashed bar) and as maximum response from baseline for each individual (right hand panels). a and b: p<0.001
93
ACTH
5% 25% 35%-25
0
25
50
75∆
max
AC
TH (p
g/m
l)Plasma Cortisol
5% 25% 35%-200
-100
0
100
200
∆m
ax p
lasm
a co
rtis
ol(n
mol
/l)
Salivary Cortisol
5% 25% 35%-100
0
100
200
300
400
∆m
ax S
aliv
ary
Cor
tisol
(nm
ol/l)
Prolactin
5% 25% 35%-100
0
100
200
∆m
ax P
rola
ctin
(nm
ol/l)
Noradrenaline
5% 35%-250
0
250
500
750
∆m
ax N
orad
rena
line
(pg/
ml)
a b
c d
e
Figure 3.03. Individual and mean peak ACTH, plasma cortisol, salivary cortisol, prolactin and noradrenaline responses to a single breath of 5%, 25% and 35% CO2. a: p=0.006; b: p=0.02; c: p=0.02; d: p<0.001; e: p=ns
94
FSH
5% 25% 35%
-0.5
0.0
0.5
1.0
∆m
ax F
SH (I
U/l)
LH
5% 25% 35%
-1
0
1
2
∆m
ax L
H (IU
/l)
GH
5% 25% 35%
-30
-20
-10
0
10
20
∆m
ax G
H (m
U/l)
TSH
5% 25% 35%
-0.5
0.0
0.5
1.0
∆m
ax T
SH (m
U/l)
RENIN
5% 25% 35%
-10
0
10
20
∆m
ax R
enin
(fm
ol/l)
AVP
5% 25% 35%
-1
0
1
∆m
ax A
VP(p
g/m
l)
Figure 3.04. Individual and mean peak FSH, LH, GH, TSH, Renin and AVP responses to a single breath of 5%, 25% and 35% CO2. Differences between responses for each dose were not significant.
95
Plasma Cortisol vs Anxiety
0 50 100 150 200
0
50
100
p=0.04R2=0.4667
∆max Plasma cortisol (nmol/l)
∆m
ax A
nxie
ty V
AS
(mm
)
Plasma Cortisol vs SBP
0 50 100 150 200
0
25
50
75
p=0.02R2=0.607
∆max Plasma Cortisol (nmol/l)
∆m
ax S
BP
(mm
Hg)
Figure 3.05. Correlations between peak change plasma cortisol following exposure to a single breath of 35% CO2 and peak change in systolic blood pressure (upper panel) and peak change in anxiety score (lower panel).
96
3.3.3. Discussion
The principal aim of this first study was to detail the response to CO2 inhalation and
identify an optimum dose that would serve as the basis for future studies.
50% CO2 seemed to induce an involuntary inhibition of inspiration that prevented a
subject from taking a full vital capacity breath of this mixture. Whether this is due to
CO2 induced laryngeal spasm secondary to the action of CO2 chemoreceptors in the
nasopharynx as Davey [1972] suggests is uncertain. However, from volunteers
descriptions is was clearly evident that the bitter taste associated with CO2 was much
more obvious when breathing through a nasal-oral mask (when nasal chemosensitive
taste receptors [Coates 2001] are stimulated) as compared to a mouth-piece and nose-
clip. As a result, the frequency if inadequate breaths (< 80% VC), was far greater with
50% CO2 or when the nasal-oral mask was used.
Psychological response
The principal psychological changes seen in this study were a dose dependent increase
in subjective feelings of anxiety, breathlessness and a few specific somatic symptoms of
fear namely difficulty concentrating, dizziness or lightheadedness, blurred or narrowed
vision, a feeling of the heart racing or pounding and feeling hot or flushed. Other
symptoms were experienced rarely. For example few, if any subjects described nausea,
chest pain, butterflies or feeling tremulous. This is consistent with the data of Verberg
et al [1988] who has described the CO2 challenge as producing a narrow spectrum of
very specific somatic symptoms in both normal individuals and panic disorder patients.
In this study, no subject panicked in response to any of the exposures. This is again
consistent with published literature suggesting the CO2-induced panic rate, from either
low-dose chronic exposure or acute high dose (single breath of 35% CO2), in healthy
97
individuals is low [Rassovsky and Kushner 2003]. Whilst considerable variation exists
in the methods of administering CO2 and in the definition and recording of panic
attacks, a number of studies have examined the anxiety response to a single breath of
35% CO2 in both susceptible panic disorder patients, their relatives and healthy controls
[Battaglia and Perna 1995, Perna et al 1995a, Perna et al 1995b, Perna et al 1995c,
Perna et al 1995d, Perna 1995e, Verburg et al 1995]. Overall in susceptible individuals
panic attack rates occur in about 50-75%, whilst normal controls panicked in only 0-5%
of cases. First degree relatives panicked in about 22% of cases [Perna et al 1995b].
Further, even with the higher doses of CO2, anxiety symptoms were only moderate in
intensity and transient, resolving completely within 2-3 minutes. Subjects participating
in this study had, however, been carefully selected to exclude a personal or familial
predisposition to panic or anxiety.
Sympatho-neural and sympatho-adrenomedullary response
A single breath of 35% CO2 produced a marked systolic pressor response that was
preceded by a significant and persistent bradycardia. Five and 25% CO2 did not have
any significant effect on cardiovascular parameters or catecholamine release. The
pressor response to 35% CO2 was paralleled by a significant release of noradrenaline at
2 minutes following exposure. No significant adrenaline response was seen and this
most likely reflects the differential regulation of sympatho-neural and sympatho-
adrenomedullary systems with CO2 preferentially stimulating sympatho-neural
pathways but not affecting adrenaline release from the adrenal medulla [Young and
Landsberg 1998]. Cardiac output, stroke volume, peripheral resistance and other indices
of cardiovascular function were not assessed in these initial studies, however, as will be
described in more detail in Chapter 5, detailed cardiovascular responses have been
evaluated using non-invasive monitoring techniques. Results of these indicate the
98
pressor response is associated with a marked increase in total peripheral resistance
without a significant change in either stroke volume or cardiac output. This would be
consistent with noradrenaline mediated peripheral vasoconstriction without a significant
adrenaline mediated inotropic effect. Several groups, including Cullen and Edgar
[1974] have shown CO2-mediated peripheral vasoconstriction and increased total
peripheral resistance with an associated increase in blood pressure that was, in part,
mediated through peripheral chemoreceptors.
Parasympathetic response
It was initially thought that reflex baroreceptor activity in response to the increase in
blood pressure produced the bradycardia. Examination of the continuous beat-to-beat
monitoring of heart rate and blood pressure clearly shows the onset of the bradycardia
occurring before the pressor response, making reflex bradycardia unlikely. Three
possible mechanisms producing bradycardia can be postulated. Firstly, the bradycardia
may simply reflect the respiratory gasp reflex that occurs with a single vital capacity
breath through the mouth. The bradycardia associated with this reflex is, however, brief
and less marked than was observed. Indeed, continuous physiological monitoring as
performed in studies described in Chapter 5 clearly show a transient small gasp reflex
bradycardia followed by a more intense sustained bradycardia associated with CO2
inhalation. Secondly, this response might represent a negative chronotropic effect of
CO2 on the myocardium. Tenney [1956, 1960] has shown in isolated heart muscle
preparations a direct effect of CO2 producing a slowing of spontaneous heart muscle
contractions. Several experiments described fully in Chapters 4 and 5 have shown no
effect of CO2 on heart rate when the heart has been either surgically (as in heart-lung
transplantation) or pathologically (as in various forms of autonomic neuropathy)
denervated. This suggests CO2 is acting through neural mechanisms rather than directly
99
on the myocardium. This leaves the most likely mechanism – direct vagal action either
through CO2 stimulation of brainstem vagal centres or through activation of reflex arcs
between peripheral chemoreceptors (within the oropharynx or carotid bodies) and the
vagus nerve as have been described by Henry et al [1998].
Hypothalamic-pituitary-adrenocortical response
A single breath of 35% CO2 produced a significant release of ACTH at 2 minutes and
plasma cortisol at 15 minutes. A single breath of 5% or 25% CO2 had no effect on the
HPA axis. Plasma cortisol release was reproduced following a second exposure to 35%
CO2 after one week and after 6 months. Most psychological stress paradigms show a
significant inter-individual variability in the magnitude of the cortisol response
[Sapolsky 1994] and this was similarly seen in these studies. Repeated exposure to both
physical and psychological stressors typically results in a decline in cortisol responses
(habituation) [Singh et al 1999] making many of these laboratory stress paradigms
unsuitable for repeated evaluation of HPA axis activity in the same individual. Whilst
no significant habituation of cortisol responses was seen on repeated exposure, there
was a trend toward smaller cortisol responses in the group that returned after 6 months.
This was associated with overall smaller vital capacity test breaths and may reflect
reduced CO2 exposure rather than habituation. Sub-optimal vital capacity (< 80%
baseline) resulted in little or no cortisol response and this reinforces the importance of
achieving an adequate test breath. However, from this series of experiments we can’t be
sure that more frequent exposure would not result in a more significant reduction in
cortisol responses.
100
Other endocrine responses
Release of both prolactin and GH may occur as part of the endocrine stress response and
an increase of both has previously been demonstrated following hypercapnia [Woods et
al 1988]. Prolactin has been shown to have anxiolytic properties and it may play a role
in modifying anxiety-related behaviour [Torner et al 2001]. Whilst a significant
prolactin response did occur with 35% CO2, the anticipated rise in GH did not. Neither
was there a significant release of other anterior pituitary hormones (TSH, LH, or FSH).
AVP is a stress responsive neuropeptide capable of stimulating pituitary ACTH release
and plays an important role in the HPA response to both acute [Scott and Dinan 1998]
and chronic [Ma and Lightman 1998] stress. Hypercapnia induces marked increases in
c-fos expression in AVP containing neurones of the supraoptic and paraventricular
nuclei [Kc et al 2002], although, this was probably secondary to activation of pathways
projecting from the brainstem. In this study we did not observe any significant change
in AVP levels, although it is important to note that peripheral blood levels of AVP may
not necessarily reflect AVP levels within the hypophyseal portal blood.
The renin-aldosterone system has a role in maintaining cardiovascular integrity
particularly in response to haemodynamic challenges and is regulated by sympathetic β-
adrenergic innervations [DiBona and Kopp 1997]. Despite this there was no change in
plasma rennin activity in keeping with the known specificity of activation of different
parts of the sympathetic nervous system.
Correlations
As mentioned, wide inter-individual variation in cortisol responses to 35% CO2 was
seen in all these studies and this is consistent with most laboratory stress paradigms,
101
particularly those involving psychological stressors. Similarly SAM responses
including noradrenaline, blood pressure and heart rate changes showed wide inter-
individual variation. This variation occurs through the interplay of numerous influences
including genetic, developmental, social and individual personality characteristics
[Kirschbaum et al 1995, Gerra et al 2001, Levine 2000]. Repeated exposure,
particularly to psychological stress, is associated with a significant decline in ACTH
and cortisol responses (habituation), but not in catecholamine or blood pressure
responses. Individuals who show reduced habituation (termed high responders) tend to
perceive more stress and express higher rates of symptoms of anxiety and depression
[Kirschbaum et al 1995, Van Eck et al 1996, Gerra et al 2001]. These individuals also
appear to display reduced habituation to physical stressors as well as psychological
stressors [Singh et al 1999]. Cacioppo et al [1995] have also suggested that high
responders are characterised by higher levels of neuroticism, higher anticipatory anxiety
and poor psychological defence and these individuals typically show larger and longer
lasting cortisol and SAM stress responses to psychological stress. The authors imply
these individuals are more susceptible to the negative health consequences of an
impaired neuroendocrine stress response system.
102
3.4. Reproducibility studies
3.4.1. Introduction and methods
From the previous study, a single breath of 35% CO2 had been identified as the
optimum dose to use for ongoing studies. This was based on its ability to generate an
HPA, prolactin, SAM and parasympathetic response, whilst being well tolerated with
little adverse effect. Further studies were then undertaken to evaluate the
reproducibility of 35% CO2 exposure. In the first of these, ten male subjects with a
mean age of 24.4 (+/- 1.2) years (range 22-35) received a single breath of 35% CO2 on
two occasions separated by one week. Results from three were excluded because one of
their two test breaths was < 80% of their expected VC. Blood samples for cortisol and
prolactin were collected in the same manner as described for the dose response study at
the same intervals. Saliva, for α-amylase, was also obtained from these 10 subjects
using Salivettes (Sarstedt) and collected at the same times as blood samples were taken.
VAS questionnaires were administered immediately before and immediately following
the exposure only. Cardiovascular monitoring in these studies where performed with
the automated blood pressure cuff (Dynamap) that recorded responses every minute.
Since only one concentration was administered to the same individual on two occasions,
a paired t-test analysis was used to determine the difference in the maximum change for
each parameter following each exposure.
A separate group of 5 male subjects (21.1 +/- 1.4 years) received a single breath of 35%
CO2 according to the same protocol. These subjects then returned after 6 months when
the procedure was repeated. The CO2 administration occurred at the same time of the
day (between 1100 – 1300hours) for both visits in both of the above studies.
103
In order to assess the effect of the natural diurnal variation in plasma cortisol levels on
reproducibility, a fourth group of 10 healthy volunteers (4 male) aged 24.3 +/- 0.7 years
(range 22-30) was studied. Each subject attended twice, one week apart. One of the
visits was scheduled at 7 o’clock in the morning (peak baseline cortisol levels) and the
other at 8 o’clock in the evening (trough baseline cortisol levels). The time of the first
visit was assigned randomly. Each subject received a single breath 35% CO2 at each
visit according to the same protocol already described, however, instead of using a
mouth-piece and nose clip, this experiment was conducted using a nasal-oral face mask
that allowed the subject to breath through both the mouth and nose. Cardiovascular
monitoring using the Dynamap recorder, blood sampling for plasma cortisol and
prolactin and administration VAS questionnaires were as for the repeatability studies
described above. Data from 3 individuals was excluded because at least one of their test
breaths was < 80% of baseline VC.
3.4.2. Results
As shown in Figures 3.06 and 3.07, the increase in subjective feelings of anxiety,
plasma cortisol and SBP with the corresponding fall in HR that occurred following the
initial breath of 35% CO2 was not significantly different when the exposure was
repeated after 1 week. Similarly, following a single breath of 35% CO2 repeated after 6
months, similar increases in plasma cortisol, SBP, anxiety and breathlessness where
seen, with the same corresponding fall in heart rate (Figures 3.06 and 3.07). Whilst
differences in cortisol responses where not significant, there was a trend toward a lower
cortisol response at the 6 month visit, although measured VC for this visit also tended to
104
be lower compared to baseline (5.6 +/- 0.5 L at 0 months compared to 5.2 +/- 0.2 L at 6
months).
As shown in Figure 3.08, there was wide inter-individual variation in salivary amylase
levels both at rest and in following CO2 exposure. There was no significant increase in
amylase following the test. Further, as is also shown in Figure 3.08, there was no
significant correlation in the change in amylase levels with the SBP, DBP or heart rate
response in these subjects.
Table 3.02 shows baseline values for cardiovascular, hormonal and psychological
parameters at different times of the day. Baseline cortisol levels were significantly
higher in the morning compared to the evening (561.1 +/- 41.3 compared to 199.6 +/-
28.9 nmol/l respectively, p < 0.0001). There was no significant difference in other
parameters prior to exposure. A significant pressor response with associated
bradycardia was seen at both times of the day (Figures 3.09 and 3.10). Similarly, the
pattern of anxiety, breathlessness and other somatic symptom responses was the same as
in previous experiments, and there was no difference between morning and evening.
There was no statistically significant difference between the maximum change in
cortisol levels following CO2 exposure, although there was a trend toward lower
responses in the morning (higher baseline cortisol levels). No significant increase in
prolactin release was seen following either CO2 exposure (Figure 3.09). Whilst test
breaths were considered adequate, (> 80% of baseline) the average test breath using the
nasal-oral mask appeared lower that in the dose response study (2.8 +/- 0.3 compared to
4.0 +/- 0.3 L respectively) which had a similar sex distribution.
105
Response to 35% CO2 atbaseline and repeated after 1
week
Anx
iety
(mm
)
Bre
athl
essn
ess
(mm
)
Cor
tisol
(nm
ol/l)
SB
P (m
mH
g)
HR
(b/m
)-25
0
25
50
75Baseline1 week
mm
Response to 35% CO2 atbaseline and repeated after 6
months
Anx
iety
(mm
)
Bre
athl
essn
ess
(mm
)
Cor
tisol
(nm
ol/l)
SB
P (m
mH
g)
HR
(bpm
)-25
25
75
125Baseline6 months
mm
Figure 3.06. Psychological, cardiovascular and cortisol responses to a single breath of 35% CO2 administered at baseline and then repeated after 1 week (n=7, top panel) or, in a separate group of subjects, repeated after 6 months (n=5, lower panel). Differences between responses for each parameter were not significant.
106
Systolic blood pressureresponse to repeated 35% CO2
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
100
110
120
130
140
Time (minutes)
mm
Hg
Pulse rate response torepeated 35% CO2
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
50
60
70
80
Time (minutes)
bpm
O baseline • 1 week
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
60
70
80
90
Time (minutes)
bpm
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
110
120
130
140
150
160
Time (minutes)
mm
Hg
O baseline • 6 months
Figure 3.07. Cardiovascular responses for the 5 minutes before and 5 minutes after a single breath of 35% CO2 given at baseline and then repeated after 1 week (n=7, top panels) and in a separate group, repeated after 6 months (n=5, lower panels). Left hand panels represent systolic blood pressure. Right hand panels represent heart rate. Exposure is represented by hashed line.
107
Salivary amylase vs SBP
0.0 2.5 5.0 7.5 10.0 12.5
-25
0
25
50
R2=0.2371p=ns
∆max Amylase (U/l)
∆m
ax S
BP
(mm
Hg)
Salivary amylase vs DBP
0.0 2.5 5.0 7.5 10.0 12.5
-20
-10
0
10
R2=0.2695p=ns
∆max Amylase (U/l)
∆m
ax D
BP
(mm
Hg)
Salivary amylase vs HR
0.0 2.5 5.0 7.5 10.0 12.5
-30
-20
-10
0
10
R2=0.0627p=ns
∆max Amylase (U/l)
∆m
ax H
R(b
/m)
Salivary amylase response to35% CO2
Baseline Peak
0
100000
200000
300000
U/l
Figure 3.08. Salivary amylase response to 35% CO2 in 10 healthy volunteers including peak response and correlations between change in salivary amylase and change in SBP, DBP or HR.
108
Table 3.02. Baseline physiology, psychology and hormonal characteristics for the morning compared to evening visits.
am visit pm visit p
Physiology
SBP (mmHg) 107.9+/-4.9 118.7+/-4.0 ns
Heart Rate (bpm) 66.1+/-2.9 66.2+/-2.9 ns
Test Vital Capacity (L) 3.0+/-0.4 2.8+/-0.3 ns
Psychology
Anxiety (mm) 27.0+/-6.1 35.5+/-8.0 ns
Breathlessness (mm) 57.5+/-8.2 66.0+/-6.2 ns
Endocrinology
Plasma cortisol (nmol/l) 561.1+/-41.3 199.6+/-28.9 <0.0001
Prolactin (mU/l) 230.0+/-32.7 186.3+/-28.9 ns
109
Response to 35% CO2according to time of
administration
Anxi
ety
(mm
)
Brea
thle
ssne
ss (m
m)
Cor
tisol
(nm
ol/l)
Pro
lact
in (m
iu/l)
SBP
(mm
Hg)
HR
(bpm
)-25
0
25
50
75ampm
mm
Figure 3.09. Change in plasma cortisol, prolactin, cardiovascular and subjective anxiety responses following a morning (am) compared to evening (pm) breath of 35% CO2 (top panel) and in the lower panels is the cardiovascular response for the 5 minutes before and after exposure (represented by the hashed line). n=7. Differences between responses for each parameter were not significant.
110
Systolic blood pressureresponse according to time of
administration of CO2
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
90
100
110
120
130
140
Time (minutes)
mm
Hg
Pulse rate response accordingto time of administration of CO2
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
50
60
70
80
90ampm
Time (minutes)
bpm
Figure 3.10. Cardiovascular responses following a single breath of 35% CO2 given at 7 am compared to 8 pm (n=7). Upper panel represents systolic blood pressure. Lower panel represents heart rate. Exposure is represented by hashed line.
111
3.4.3. Discussion
Significant psychological habituation following repeated 35% CO2 exposure either at
one week or at 6 months did not occur. A similar persistence of responses has been
seen in other studies of 35% CO2. Anxiety responses were equivalent when exposure
was repeated after 1 [Van den Hout et al 1987], 2 or 6 days [Perna et al 1994, Perna et
al 1997]. This is in marked distinction to pure psychological stressors such as the Trier
Social Stress Test (TSST), which is usually only effective the first time it is performed
[Gerra et al 2001].
Repeat exposure, whether at 1 week or at 6 months, showed the cardiovascular response
to be robust and easily reproducible. No evidence of habituation or a decline in
responses was apparent event at 6 months when the mean vital capacity breath was
somewhat smaller compared to the initial visit. The study examining the effects of
diurnal variation showed the typical circadian variation in blood pressure was evident in
these normal subjects with baseline blood pressure tending to be lower in the morning
compared to the evening [Witzleb 1983]. Despite this, and the significantly different
baseline cortisol levels at these times, both systolic blood pressure and heart rate
responses were no different following 35% CO2 exposure. Similarly, baseline cortisol
levels did not have any significant effect on cortisol, prolactin or psychological
responsiveness following 35% CO2 exposure.
112
3.5. Summary
Carbon dioxide-induced anxiety has been extensively researched for several decades,
but the mechanism of CO2-anxiogenesis is still unclear. At present, the two commonly
held hypotheses regarding CO2-mediated panic induction refer to hypersensitivity of
either the brainstem locus coeruleus or the central nucleus of the amygdala as the
triggers of panic [Gorman 2003]. The former theory holds that panic is a form of
severe anticipatory anxiety whereby autonomic sensations (such as breathlessness or
suffocation-like symptoms produced by CO2) are misinterpreted by the individual as
representing a more serious and imminent threat. This then triggers the LC/sympathetic
response system with subsequent stimulation of limbic and HPA axes [Gorman 2003,
Griez and Schreurs 2003]. Evidence supporting this includes yohimbine (a stimulus of
brainstem noradrenergic centres) mediated panic attacks are inhibited by blocking
noradrenergic brainstem activity with clonidine [Gorman 2003]. Further, in animal
studies, lesions of the LC inhibit anxiety behaviour [Gorman 2003].
Opposing this view, the latter hypothesis describes a more complex view of anxiety and
CO2-induced panic as arising from a hypersensitivity of the alarm system that has
evolved to regulate CO2 and lactate levels [Klein 1993]. This system involves the
central nucleus of the amygdala, centres involved in emotional memory as well as
centres involved in cognitive processing of environmental cues [Gorman 2003].
Neuroimaging has clearly defined activation of these areas in response to a number of
fear inducing stimuli [Brannan et al 2001, Liotti et al 2001]. It is also argued that the
complexity of the panic response is such that it is more likely to result from stimulation
of these limbic centres with subsequent activation of the HPA and noradrenergic centres
[Gorman 2003].
113
Many different substances (for example lactate, doxopram, yohimbine, cholecystokinin-
B receptor agonists, caffeine and CO2), each with vastly different mechanisms of action,
can produce panic attacks [Sinha 1999] in humans. Similarly, both cognitive
behavioural therapy [Gorman et al 2000] and a diverse range of neurotransmitter
agonists [Bertani et al 1997, Nardi et al 2000], and antagonists [Battaglia et al 2001]
can inhibit such attacks. It is likely, therefore, that both pathways have important roles
in mediating some or all of the features of CO2-induced anxiogenesis and the
predominant mechanism are yet to be determined. It is also likely that the pathways
involved will vary between panic disorder patients and normal volunteers and between
the type of stimulus applied (low dose versus high dose CO2 for example) as well as
other factors.
Most studies of hypercapnia, however, demonstrate activation of central chemoreceptor
sites particularly noradrenergic centres within the brainstem with subsequent
sympathetic nervous system stimulation. In animal models, CO2-induced activation of
the LC increases respiratory drive, induces hypertension and tachycardia, increases
cerebral blood flow and is partly responsible for regulating arousal responses [Krystal et
al 1989, Haxhiu et al 2001]. Bilateral LC lesions in cats impair this response
[Rammana Reddy et al 1986]. In humans, cardiovascular stimulation is a feature of
CO2–induced anxiety in both healthy volunteers and panic disorder patients [Gorman et
al 1988]. Systolic hypertension, tachycardia and decreased total peripheral resistance
have been the most common features described, although catecholamine release has not
been a consistent finding. Sechzer et al [1960] and Tenney [1960] describe both
adrenaline and noradrenaline release, while Woods et al [1988] failed to show an
increase in the noradrenaline metabolite MHPG in response to prolonged exposure to
low dose CO2. All these studies involved CO2 concentrations of between 5 – 15%
114
administered over minutes to hours, and this, plus differences in sampling times may
account for some of the observed variation in catecholamine release.
HPA axis activation following CO2 exposure in other paradigms has been inconsistent
and variable, with some studies demonstrating cortisol release [Sechzer et al [1960,
Woods et al 1988, Krystal et al 1989, Argyropoulos et al 2002], whilst others have
failed to demonstrate a response despite some of them generating acute anxiety and
even panic [Van den Hout et al 1987, Sinha 1999, Coplan et al 2002, Van Duinen et al
2004]. The reason for this difference is not immediately clear, although, it appears the
threshold to produce behavioural arousal is significantly less than that required to
generate an HPA response. This threshold difference has been recognised in a number
of laboratory stress paradigms as well as in real life stressors [Biondi and Picardi 1999].
Similarly, numerous studies, including those by Pacak et al [1998, Pacak and Miklos
2001], have shown that varying intensities of a particular stressor will produce different
degrees or patterns of HPA and SAM stimulation. It is possible that differences in
intensity of the CO2-specific stress response are responsible for the behavioural arousal
(ie anxiety) seen with low doses, whilst HPA axis activation only occurs with much
higher doses.
An alternative explanation, however, is that acute and chronic hypercapnia might
represent qualitatively distinct stressors. Herman et al [1997] suggested stressors that
represent an immediate threat to life (such as hypoxia or hypotension) directly activate
brainstem and hypothalamic nuclei generating immediate physiological responses
required to counter these threats. In contrast, during situations such as novelty or
conditioned fear, there is a need for some form of higher cortical processing of the
challenge relative to past experience before the situation is perceived as stressful. Acute
115
hypercapnia from a single breath of 35% CO2 may be an example of the former
situation, being an immediate threat to respiratory homeostasis whereas the somatic
symptoms associated with low dose CO2 may first require interpretation as distressing
before a neuroendocrine response is generated [Klein 1993, Gorman et al 2000].
It is clear from studies of the biological basis of fear and panic that activation of a
number of different brain centres can result in anxiety and panic attacks that appear
outwardly similar. Differences do exist, however, particularly in whether or not HPA
axis activation is a feature of the attack. Exposure to some pharmacological agents
including yohimbine, cholecystokinin-B receptor agonists and caffeine will produce
acute panic without HPA axis activation, whereas lactate, isoproterenol and doxopram
all produce both panic and HPA axis activation [Sinha 1999]. Nattie [1999] has shown
that central CO2 chemoreceptors are distributed widely throughout the brainstem,
cerebellum and midbrain and these chemoreceptors retain discrete neuroanatomical and
functional properties. Further, CO2 sensitive receptors in different brain regions respond
selectively to a narrow range of CO2 concentration and it is possible that the different
responses seen to acute and chronic CO2 exposure could reflect activity of different
neuroanatomical pathways or different thresholds of activity of the same pathway.
There are several possible pathways that may mediate HPA axis activation in response
to CO2 exposure. Immediate early gene (c-fos) expression studies in rodents [Haxhiu et
al 2001] have demonstrated activation of the hypothalamic paraventricular nucleus
(PVN) following CO2, although the presence of CO2/pH sensitive chemoreceptors
within the PVN itself have not been clearly demonstrated. It seems much more likely
that the PVN is activated either in response to activation of brainstem noradrenergic
centres, or from limbic centres mediating the psychological response, or both. In
116
addition, other neurotransmitter pathways may be important in modifying the HPA axis
response to this challenge. Such possibilities include serotonergic pathways that are
involved in respiration, arousal [Haxhiu et al 2001] and CO2–induced anxiety [Klaasen
et al 1998, Kc et al 2002]; AVP, which is important in HPA responses to stress and has
been shown in animal studies to be activated by hypercapnia [Bornstein and Chrousos
1999], although did not appear to be affected by CO2 in this study (see below); and
finally non-ACTH mediated release of cortisol from the adrenal cortex secondary to
SAM activation [Bornstein and Chrousos 1999] is also possible.
Consistent with the literature, a significant correlation between change in cortisol and
change in anxiety was observed, as was a significant correlation between change in
cortisol and change in SBP. This suggests low responders to the CO2 stress are
characterised by low cortisol, anxiety and blood pressure responses as opposed to high
responders. Gerra et al [2001] attempted to identify whether significant SAM
habituation could occur from repeated exposure to psychological stress. Their studies
confirmed cortisol habituation in some, but not all subjects, but no habituation of SAM
responses. This dissociation of responses was unexpected in view of the links between
brainstem sympathetic centres and the HPA axis [Bugajski et al 1995, Gerra et al 2001],
nevertheless, similar dissociations have been reported elsewhere [Malarkey et al 1995].
In the above studies of repeated CO2 exposure, there was a non-significant trend
towards reduced cortisol responses when the dose was repeated either at 1 week or at 6
months, although at least in the 6 month study this may have been related to smaller
inspired vital capacity volumes. ACTH and noradrenaline responses were not
measured after repeated exposure, but blood pressure and heart rate responses were
robust and easily reproducible. Individual cortisol responses varied markedly with
some responses declining after the second exposure, some increasing and others
117
remaining relatively constant. The small sample sizes used precludes in conclusions
being drawn about the likelihood of habituation occurring or the proportion of
individuals who are likely to be low or high responders. Similarly, in these studies an
attempt to identify subjects as high or low responders prior to exposure was not made.
Thus whether cortisol responses to this model are predictive of personality type,
baseline anxiety or future risk is not known.
118
3.6. Conclusion
The aim of this first series of experiments was to explore the neuroendocrine,
psychological and cardiovascular effects of acute hypercapnia in normal individuals
with the intention of detailing the response and assessing the validity of using 35% CO2
as a simple means of evoking sympathetic and HPA axis responses. The results confirm
that the anxiogenic response to hypercapnia in normal individuals is dose dependent,
but anxiety is transient and 35% CO2 is well tolerated. Cardiovascular and cortisol
responses, on the other hand, were only activated by 35% CO2 but were reproducible
when repeated after short (1 week) and long (6 month) intervals. Haemodynamic
responses in particular were robust, easily measured, and unaffected by intervals
between doses or by time of exposure.
The mechanisms and neuroanatomical pathways behind the observed responses are still
to be determined and some of the experiments detailed in the subsequent chapters
address some of the possibilities. In summary, however, the description above suggests
that the initial response to a single breath of 35% CO2 is vagally mediated bradycardia
followed by noradrenaline mediated peripheral vasoconstriction producing an acute
pressor response as a result of stimulation of brainstem noradrenergic centres. HPA
axis activation is a feature of acute (35% CO2) exposure but is likely to occur indirectly
as a result of either brainstem noradrenergic projections to the PVN and/or from
projections from the limbic fear centre (including the CnA). Stimulation of the CnA
with subsequent involvement of the cortex and limbic areas associated with this fear
circuit is likely to be responsible for the observed emotional arousal. However, whether
this is a direct CO2 effect, as suggested by Gorman et al [2003], or an indirect effect of
119
brainstem activation, as suggested by Bailey et al [2003], or both, is yet to be
determined.
120
CHAPTER 4
THE 35% CO2 MODEL:
MECHANISMS UNDERLYING THE
NEUROENDOCRINE RESPOSNE
121
4.1. Introduction
The principal observations from the initial studies performed with 35% CO2 were
emotional arousal, initial bradycardia and a significant pressor response that was
associated with noradrenaline release. Hormonal responses included prolactin release
and HPA activation with ACTH and cortisol release. Further, cortisol responses were
well correlated with both the pressor response and with aspects of emotional arousal,
but not with heart rate changes.
As mentioned, studies of CO2 challenges and panic induction in the literature have
yielded two potential hypotheses regarding the likely mechanisms that mediate CO2
responses. The ‘false suffocation’ theory of Klein [1993] suggests that in panic
disorder, panic attacks originate from an increased sensitivity of the CnA and its
associated fear circuit as described by Gorman et al [2000]. Components of this fear
circuit are responsible for the various features of panic including anticipation of the
event, the acute attack and phobic avoidance. The circuit can be activated at different
regions explaining how panicogens with vastly different mechanisms of action can all
produce similar panic attacks and how different treatments can all be effective [Van den
Hout et al 1987, Greiz et al 1990b, Nutt et al 1990, Bertani et al 1997, Perna et al 1997,
Sinha 1999, Gorman et al 2000, Nardi et al 2000, Battaglia et al 2001].
Communication between the CnA and other components of this circuit with brainstem
noradrenergic centres would be responsible for the sympathoadrenal activation seen in
fear and panic. Projections between the CnA and brainstem with the hypothalamic
PVN, on the other hand, would be responsible for the HPA axis activation associated
with some, but not all panicogens [Sinha 1999, Gorman et al 2000].
122
More recently, an alternative hypothesis has been proposed based on studies that have
shown significant positive correlations between the emotional arousal and the cortisol
response and with sympathetic activation [Argyropoulos et al 2002, Bailey et al 2002].
Based on their work, these authors have proposed that stimulation of brainstem
noradrenergic centres (particularly the LC) is the initiating event following CO2
exposure and that a feed forward loop involving the CnA and the PVN results in
subsequent psychological responses and HPA axis activation respectively. In addition,
activation of CRH pathways that project back to the LC and brainstem as well as to the
CnA enhance these same responses. Sensitivity of these and other modulating
neurotransmitter pathways are then responsible for individual variability and the clinical
susceptibility of some individuals to anxiety and panic disorders [Bailey et al 2002].
Other neurotransmitter pathways such as serotonin [Dinan 1996, Klaasen et al 1998,
Ben Zion et al 1999, Miller et al 2000, Schruers et al 2000, Lowry 2002, Schruers et al
2002], endogenous opiates [Gritz et al 1976, Vythilingam et al 2000, Drolet et al 2001],
GABA, acetylcholine, substance P and neuropeptide Y [Harbuz and Lightman 1992,
Chrousos 1998, Habib et al 2001] have been implicated as regulators or modifiers of the
observed response to CO2 inhalation.
Neuroanatomical and immunohistochemistry studies in animals and neuroimaging
studies in humans lend support to both hypotheses. It would seem likely that rather than
being mutually exclusive, both hypotheses are complementary with one or other being
dominant depending on the nature of the hypercapnoeic stressor and the circumstances
in which exposure is occurring. Specifically, there is evidence indicating the presence
of CO2/pH sensitive chemoreceptors in the region of the VLM, LC and other
noradrenergic brainstem sites [Nattie 1999, Haxhiu et al 2001]; in association with 5-
HT neurons of the raphe complex [Wang et al 2001, Wang et al 2002]; and within
123
limbic centres [Nattie 1999]; but not within the PVN or pituitary gland [Haxhiu et al
2001, Kc et al 2002]. Functional neuroimaging in humans breathing low dose CO2 has
confirmed activation of multiple brain centres including the VLM, LC, CnA, as well as
other limbic and para-limbic centres, frontal lobes and the cerebellum [Brannan et al
2001, Liotti et al 2001].
A simplified overview of the principal neuroanatomical centres and their
interconnections likely to be involved in mediating the various components of the CO2
stress response is given in Figure 4.01. It is not yet clear, however, whether the
initiating event is activation of brainstem noradrenaline with subsequent psychological
or HPA responses, or whether the ‘central fear circuit’ (CnA) is activated first with
subsequent brainstem and HPA activation.
124
VLMLC
CnA
Pit
Hipp
V
PVN
Pit
CRH
ACTH
Cortisol
Vagus
SNS PNS
VLM
LC
CnA
Hipp
5-HT
opioid
Figure 4.01. Simplified overview of the principal centres likely to be involved ingenerating the psychological, HPA and SAM responses to acute hypercapnia. CnA: Central nucleus of amygdala; Pit: Pituitary; PVN: Hypothalamicparaventricular nucleus; Hipp: Hippocampus; VLM: Ventrolateral medulla; LC:Locus coeruleus; SNS: Sympathetic nervous system; PNS: Parasympathetic nervoussystem; 5-HT: Serotonin network
125
4.2. Experimental plan
In order to evaluate some of the mechanisms and pathways mediating the various
responses to CO2, a series of experiments were conducted to examine the specific roles
of the HPA axis as well as the central and peripheral components of the autonomic
nervous system. The contribution of three central neurotransmitter pathways
(serotonergic, opiate and noradrenergic) was examined in the first study, whilst the
impact of a suppressed HPA axis on the psychological, cardiovascular and prolactin
response to CO2 was assessed in two further studies. One study involved the
administration of a glucocorticoid synthesis inhibitor (metyrapone) or a
mineralocorticoid receptor antagonist (spironolactone) or a combination of both in
normal volunteers. The other experiment involved the assessment of breastfeeding
mothers who, during the lactation phase, are thought to have a physiologically
suppressed HPA axis.
Finally, examining two individuals with functional alterations in their baseline
autonomic nervous system assessed the role of the autonomic nervous system. The
first individual was an adult male with the idiopathic congenital central hypoventilation
syndrome (CCHS), a rare inherited condition of central CO2 insensitivity associated
with hypoventilation and other features of impaired cental autonomic integration. The
second individual studied, an adult male following dual lung-heart transplant,
represented a clinical model of peripheral cardio-pulmonary denervation.
126
4.3. Central neurotransmitter study
4.3.1. Introduction and methods
Central neurotransmitter concentrations were altered by the administration of
paroxetine, naltrexone or nortriptyline. Paroxetine, a potent selective serotonin re-
uptake inhibitor (SSRI), is used in clinical practice predominantly as an anti-depressant.
It acts acutely to inhibit the pre-synaptic re-uptake and storage of serotonin thereby
increasing intra-synaptic serotonin concentrations [Stahl 1998]. However, the initial
administration of this drug is often associated with a worsening of clinical symptoms
(especially anxiety), with a delay of several days to weeks before clinical anti-
depressant or anti-anxiety effects are seen. The delayed response is thought to be
related to modification of 5-HT receptor number and function [Stahl 1996, Stahl 1998].
The mu and kappa opiate receptors are the principal receptors responsible for opiate
regulation of HPA axis activity [Drolet et al 2001]. Naltrexone, an orally available
opiate antagonist is selective for mu receptors at low doses, but at higher doses will also
antagonise kappa receptors [Gritz et al 1976]. Nortriptyline is a tricyclic anti-
depressant and noradrenaline re-uptake inhibitor. In this study it was used in preference
to Reboxetine a more potent and more selective noradrenaline re-uptake inhibitor
because Reboxetine has been associated with causing significant hypertension and this
was felt to be potentially dangerous to use given the significant pressor response
associated with the CO2 challenge [Stahl 1996, Penttila et al 2001].
Twenty seven healthy volunteers who had not participated in any previous CO2 related
studies were recruited from within the University of Bristol. Volunteers were assigned
to one of three treatments in a randomised single blinded fashion. Treatment groups
127
were either placebo and paroxetine 20 mg; or placebo and nortriptyline 25 mg; or
placebo and naltrexone 50 mg. Identical capsules containing either a single dose of
placebo or the active drug were administered between 7 and 8 am the morning of the
test. All tests were conducted between 1 and 2 pm that same day. Each subject
attended for two visits one week apart and the order (placebo vs active drug) was
randomised and double blind. At each visit a single breath of 35% CO2 was
administered according to the same protocol as described above.
All subjects attended all scheduled visits. The administered medications were well
tolerated without significant side effect apart from naltrexone that caused mild dizziness
and occasional nausea in most subjects assigned to this treatment. One subject in this
group vomited the morning of the test. The test protocol was as described previously
and for all visits inspired VC was measured and test breaths were all considered
adequate, being at least 80% of baseline. The CO2 itself was well tolerated without
significant adverse effect.
Each group had 9 subjects. In the paroxetine group mean age was 24 years (range 22-
27) with 4 males. The mean age in the naltrexone group was 23 years (range 22-25)
with 4 males, whilst in the nortriptyline group, mean age was 23 years (range 22-25)
with 4 males.
Repeated measures analysis of variance was used to determine the effects of CO2
exposure and the differences between the groups with paired t-test analysis used to
determine differences between single time points including baseline measures for the
placebo and active treatment visits for each subject.
128
4.3.2. Results
Baseline hormone, cardiovascular and psychological measures are given in Table 4.01
with baseline hormone measures also represented graphically in Figure 4.02. As
illustrated, there were no differences in any baseline measures in the groups taking
either paroxetine or nortriptyline, whilst the group taking naltrexone had, on their
naltrexone visit, significantly higher basal cortisol and prolactin levels (p = 0.0004 and
p = 0.03 respectively). In addition, for the naltrexone visit, subjective VAS scores for
dizziness and nausea were significantly elevated at baseline compared to their placebo
visit (p = 0.02 for both dizziness and nausea). Similarly, cardiovascular measures in the
naltrexone group indicated a significantly higher baseline SBP (p = 0.003) compared to
the placebo visit, with no difference in baseline heart rates.
A single breath of 35% CO2 produced a significant increase from baseline in cortisol,
prolactin, SBP, anxiety fear, breathlessness and the somatic symptoms of ‘awareness of
heartbeat’, ‘feeling hot’, ‘blurred vision’ and ‘dizziness’, with a significant fall in HR (p
< 0.05 for all). Within each group, following exposure to a single breath of 35% CO2,
individuals treated with paroxetine showed no difference in their maximum change in
both cortisol and prolactin compared to their placebo visit. Naltrexone and nortriptyline
treated individuals showed no difference in their cortisol response but did show a non-
significant trend towards a smaller prolactin response. (Figure 4.03. and Figure 4.04.).
Between each group, there was no difference in cortisol or prolactin responsiveness.
The cardiovascular response (both SBP and HR) to 35% CO2, both within each group
(treatment visit compared to placebo visit) and between the three treatments, was no
different. Maximum change from baseline for SBP and HR as well as a temporal
129
profile of cardiovascular response to CO2 are shown in Figure 4.05. Psychological
responses, both within and between groups, were also equivalent, although there was a
non-significant trend towards a greater fear response after taking paroxetine, and a
smaller change in somatic symptoms (apart from dizziness) in those treated with
nortriptyline (Figure 4.06.).
130
Table 4.01. Baseline hormonal, cardiovascular and psychological parameters for each
visit: placebo vs paroxetine; placebo vs naltrexone; placebo vs nortriptyline.
Placebo Paroxetine Placebo Naltrexone Placebo Nortriptyline
Cortisol
(nmol/l)
358.8+/-41.2 386.3+/-54.9 288.1+/-27.1 675.4+/72.4* 371.6+/-34.9 365.7+/-57.3
Prolactin
(nmol/l)
156.1+/-25.6 168.2+/-24.4 124.3+/-16.5 249.8+/-45.7† 138.1+/-20.8 202.0+/-37.1
SBP
(mmHg)
116.0+/-2.3 112.4+/-1.6 111.9+/-4.0 117.6+/-3.1† 111.9+/-4.0 113.6+/-2.2
HR
(b/m)
72.0+/-2.5 68.3+/-2.8 71.6+/-3.3
69.7+/-2.4 74.4+/-7.2 76.1+/-2.5
Dizzines
s (mm)
0.0+/-0.0 0.5+/-1.6 1.1+/-3.3 15.0+/-17.6‡ 0.0+/-0.0 1.6+/-3.5
Nausea
(mm)
0.0+/-0.0 1.1+/-0.7 1.1+/-1.1 23.8+/-6.9‡ 0.0+/-0.0 2.2+/-1.2
For each group n=9. *p=0.0004; †p=0.03; ‡ p=0.02
131
Baseline cortisol
plac
ebo
paro
xetin
e
plac
ebo
naltr
exon
e
plac
ebo
nortr
ipty
line0
250
500
750
nmol
/l
Baseline prolactin
plac
ebo
paro
xetin
e
plac
ebo
naltr
exon
e
plac
ebo
nortr
ipty
line0
100
200
300
nmol
/l
†
132
Figure 4.02. Baseline cortisol (upper panel) and prolactin (lower panel) levels for the placebo/paroxetine; placebo/naltexone and placebo/nortriptyline visits. *p=0.0004, †p=0.03
Maximum change in cortisol -placebo vs paroxetine
Placebo Paroxetine
-200
-100
0
100
200
300nm
ol/l
Maximum change in prolactin -placebo vs paroxetine
Placebo Paroxetine
-50
0
50
100
150
nmol
/l
Maximum change in cortisol -placebo vs naltrexone
Placebo Naltrexone
-100
0
100
200
300
nmol
/l
Maximum change in prolactin -placebo vs naltrexone
Placebo Naltrexone
-50
0
50
100
150nm
ol.l
Maximum change in cortisol -placebo vs nortriptyline
Placebo Nortryptiline
-100
0
100
200
nmol
/l
Maximum change in prolactin -placebo vs nortriptyline
Placebo Nortriptyline
-25
0
25
50
75
nmol
/l
133
Figure 4.03. Individual and mean cortisol and prolactin responses to a single breath of 35% CO2 in each treatment group. Responses are maximum change from baseline and comparisons are made between an individuals placebo visit and their treatment visit. Individuals taking paroxetine are shown in the upper panels, whilst those taking naltrexone are shown in the middle panels, with those taking nortriptyline shown in the lower panels. Graphs on the left represent cortisol responses, with those on the right prolactin responses. All differences are non-significant by paired t-test.
Cortisol
Pla
cebo
Par
oxet
ine
Pla
cebo
Nal
trexo
ne
Pla
cebo
Nor
trypt
iline0
30
60
90
120
nmol
/l
Prolactin
Pla
cebo
Par
oxet
ine
Pla
cebo
Nal
trexo
ne
Pla
cebo
Nor
trypt
iline0
10
20
30
40
50
nmol
/l
Figue 4.04. Mean peak cortisol and prolactin responses to a single breath of 35% CO2 for the placebo/paroxetine; placebo/naltrexone and placebo/nortriptyline visits. All differences are non-significant.
134
SBP
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
100
110
120
130
140
150placeboparoxetinenaltrexonenortriptyline
Time (mins)
mm
Hg
SBP
plac
ebo
paro
xetin
e
naltr
exon
e
nortr
ipty
line
-10
0
10
20
30
40
50
mm
Hg
HR
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
50
60
70
80
90
100placeboparoxetinenaltrexonenortiptyline
Time (mins)
bpm
HRpl
aceb
o
paro
xetin
e
naltr
exon
e
nortr
ipty
line
-75
-50
-25
0
25
bpm
Figure 4.05. Cardiovascular responses to a single breath of 35%CO2 for the placebo (combined), paroxetine, naltrexone and nortriptyline groups. Differences are all non-significant.
135
Anxiety
plac
ebo
paro
xetin
e
plac
ebo
naltr
exon
e
plac
ebo
nortr
ipty
line0
25
50
75
100
mm
Fear
plac
ebo
paro
xetin
e
plac
ebo
naltr
exon
e
plac
ebo
nortr
ipty
line0
25
50
75
100
mm
Breathlessness
plac
ebo
paro
xetin
e
plac
ebo
naltr
exon
e
plac
ebo
nortr
ipty
line0
25
50
75
100
mm
Heart beat
plac
ebo
paro
xetin
e
plac
ebo
naltr
exon
e
plac
ebo
nortr
ipty
line0
25
50
75
100
mm
Feeling hot
plac
ebo
paro
xetin
e
plac
ebo
naltr
exon
e
plac
ebo
nortr
ipty
line0
25
50
75
100
mm
Blurred vision
plac
ebo
paro
xetin
e
plac
ebo
naltr
exon
e
plac
ebo
nortr
ipty
line0
25
50
75
mm
Dizziness
plac
ebo
paro
xetin
e
plac
ebo
naltr
exon
e
plac
ebo
nortr
ipty
line0
25
50
75
100
mm
Figure 4.06. Mean psychological (anxiety, fear, breathlessness, awareness of heart beat, feeling hot, dizziness and blurred vision) responses to a single breath of 35% CO2 for the placebo/paroxetine; placebo/naltexone and placebo/nortriptyline visits. All differences are non-significant.
136
4.3.3. Discussion
The administration of a single dose of paroxetine (20mg) or nortriptyline (25mg)
approximately 6 hours prior to testing had no significant effect on baseline cortisol,
prolactin or cardiovascular measurements. Similarly, neither of these agents produced
any significant side effect that volunteers were aware of. A single dose of naltrexone
(50 mg), administered at a similar time produced significant dizziness and some nausea
in most volunteers and was associated with elevated baseline cortisol, prolactin and SBP
measurements. Naltrexone is used clinically in the management of drug and alcohol
dependence [Hersh et al 1998, Anton et al 1999], however, several studies have
reported variable success depending on treatment compliance. One of the most
important factors determining compliance with naltrexone has been its side effect
profile. In a review of several studies examining naltrexone side effects and
compliance, Oncken et al [2001] classed the principal side effects as either
gastrointestinal (particularly nausea and vomiting) and neuropsychiatric (including
headache, dizziness, nervousness and fatigue). The specific side effects and the
frequency with which they occurred in the naltrexone treated group in this study is
consistent with the above published literature.
As already mentioned, endogenous opioids have an important role in modulating the
HPA axis response to stress. At rest, endogenous opioids act predominantly at the level
of the hypothalamic PVN to provide a tonic inhibition of ACTH and cortisol release.
They are probably responsible for dampening and restraining the stress response as a
means of protecting the organism from the detrimental effects of excessive activation of
the stress response [Drolet et al 2001]. Opioid antagonists, including a single dose of
naltrexone in non-human primates and humans, disinhibit the HPA axis resulting in a
137
significant increase in ACTH and cortisol release [King et al 2002, Williams et al
2003]. This increase occurs from about 2-4 hours after ingestion of the dose, and the
significantly elevated baseline cortisol seen in this study would be consistent with this
effect.
The cortisol response to CO2 following naltrexone pre-treatment was not significantly
altered. Whilst it was hypothesised that opioid antagonism would result in an enhanced
HPA response this has not been commonly observed with other stress paradigms.
Yohimbine, an α-2 adrenoreceptor antagonist, for example activates the HPA axis
through enhancement of LC and brainstem noradrenergic projections to the PVN
[Vythilingam et al 2000]. Increased cortisol responses to a yohimbine/opioid
antagonist combination were seen in one study when yohimbine was given to
individuals pre-treated with an extremely high dose of intravenous naloxone (1mg/kg)
[Charney and Heninger 1986]. In contrast, there was no synergistic effect on cortisol
release when yohimbine was given to individuals pre-treated with oral naltrexone for 8
days [Rosen et al 1999].
Opioid antagonists, however, do appear to alter emotional and behavioural responses to
stress. Principally, their clinical use stems from the decreased cravings and alcohol-
induced ‘high’ experienced by those dependent on alcohol [McCaul et al 2001].
Administered to patients with anxiety disorders, however, naltrexone and naloxone have
been shown to exacerbate feelings of anxiety and to increase physiological responses
when faced with an acute anxiety provoking situation [Rosen et al 1999]. Similarly,
naltrexone pre-treatment antagonised the anti-anxiety effects of benzodiazepines in rats
[Billingsley and Kubena 1978] and significantly increased anxiety responses to
138
yohimbine in humans [Rosen et al 1999]. Unlike these studies, however, naltrexone did
not appear to affect psychological responses to CO2 exposure.
Opiate peptides also have a role in cardiovascular homeostasis mediated both centrally
and peripherally through both receptor-dependent and receptor-independent pathways
[Gritz et al 1976]. Responses are complex, however, in general opiate agonists produce
a fall in systolic blood pressure with a more variable effect on heart rate. It might be
expected therefore that opiate antagonism would result in blood pressure increases, but
this has not been universally demonstrated particularly with acute dosing. Gritz et al
[1976] studied single dose effects of naltrexone and did not demonstrate blood pressure
or heart rate differences. McCubbin et al [1988] on the other hand has suggested opiate
antagonism is associated with exaggerated SAM and circulatory reactivity characteristic
of early stages of hypertension. In his study, naloxone pre-treatment, however, did not
affect resting blood pressure, but did exacerbate SBP responses to an acute
psychological stressor in individuals with normal resting blood pressures [McCubbin et
al 1988]. Naltrexone pre-treatment in this study increased baseline SBP, however,
blood pressure and heart rate changes were no different compared with placebo
following CO2 exposure. Alternatively, the higher resting blood pressures may simply
reflect the increased distress associated with naltrexone side effects that were not
present with either paroxetine, nortriptyline or any of the placebo visits.
Opiate peptides are also involved in the regulation of prolactin secretion predominantly
through their action on tuberoinfundibular dopamine levels. Dopamine pathways,
which provide tonic inhibition of prolactin secretion, are inhibited by opiate agonists
thereby increasing prolactin levels [Ellingboe et al 1980, Gilbeau et al 1985, McCubbin
et al 1988]. This regulation is mediated by µ and κ receptors [Kreek et al 1999,
Butelman and Kreek 2001, Andrews and Grattan 2003], antagonism of which would be
139
expected to produce a reduction in prolactin secretion. Several studies have confirmed
that opiate antagonists including naltrexone either decrease or have no effect on plasma
prolactin levels [Gold et al 1979, Rubin et al 1979, Volavka et al 1979, Mello et al
1989]. Paradoxically, baseline prolactin levels in this study were elevated in the
naltrexone group, although there was a non-significant trend towards a smaller prolactin
response to CO2 in this group. The reason for the difference in baseline prolactin is not
immediately clear, although opiate antagonism does vary in its effect on prolactin levels
according to diurnal rhythms (with smaller effects in the morning compared to the
evening) [Frecska et al 1988].
A single dose of the SSRI paroxetine had no significant effect on baseline or CO2
stimulated hormonal, cardiovascular or psychological parameters. Serotonin (5-HT)
pathways regulate a number of central functions including HPA, autonomic and
behavioural responses to stress and 5-HT can either facilitate or inhibit the activity of
these response systems [Lowry 2002]. Clinically, SSRI’s are used in the treatment of
anxiety, mood and affective disorders, although their precise mechanism of action is not
yet fully understood. Initially, SSRI’s increase 5-HT levels within the neuronal synapse
by reducing 5-HT re-uptake and storage [Barker and Blakely 1995, Stahl 1998]. This
early administration is often associated with a worsening of anxiety symptoms. Over
the next few days, anxiety symptoms improve and this is thought to be related to effects
on receptor number and function including receptor desensitisation and to effects on
other systems such as alterations in neurosteroid activity [Stahl 1996, Stahl 1998,
Nechmad et al 2003]. Studies of tryptophan depletion in both healthy volunteers and
panic disorder patients have shown a significant increase in anxiety symptoms with an
acute reduction in brain 5-HT following both high (35%) [Klaasen et al 1998, Schruers
et al 2000a, Schruers et al 2002] and low (5%) [Miller et al 2000] dose CO2 challenges.
140
Similarly, treatment with the 5-HT antagonist metergoline exacerbated CO2 induced
anxiety [Ben Zion et al 1999].
Conversely the tryptophan replete state appeared to protect against anxiety provocation
from CO2 [Klaasen et al 1998, Schruers et al 2000a]. Pre-treatment with d-
fenfluramine, a 5-HT releasing agent, in patients with panic disorder caused increased
baseline anxiety levels but a blunted anxiety response to 7% CO2 [Mortimore and
Anderson 2000]. Cortisol and cardiovascular measures before and after this low dose
CO2 challenge were no different with d-fenfluramine pre-treatment, although some
more severely affected patients had higher baseline and stimulated prolactin levels.
Chronic treatment of panic disorder patients with SSRI’s (1 week to 1 month)
significantly reduced their anxiety response to both high (35%) and low (5%) dose CO2
challenges [Bertani et al 1997, Perna et al 2002, Bertani et al 2003]. Acute
administration of paroxetine, on the other hand, could be postulated to be analogous to
d-fenfluramine use, although in our normal volunteer study no effects from paroxetine
were seen. The reasons for this are unclear, but may relate to the higher anxiety
responsiveness and sensitivity of panic disorder patients compared to healthy
volunteers. Alternatively, it is possible that the dose was insufficient to produce a
significant change in central 5-HT levels, although it is worth noting that a single dose
of 20 mg paroxetine was sufficient to enhance motor output in patients following acute
stroke [Loubinoux et al 2002].
Nortriptyline, a tricyclic antidepressant and noradrenergic re-uptake inhibitor, increases
central noradrenergic neurotransmission. A related agent, desipramine, has been shown
to increase baseline ACTH and cortisol levels in humans [Pomara et al 2001]
presumably through reciprocal pathways between brainstem noradrenergic centres and
141
the HPA axis. Acute nortriptyline administration (single dose of 50 mg) has also been
shown to increase resting systolic blood pressure [Torpy et al 1995]. Treatment with
imipramine, another related tricyclic antidepressant, for 7 days was noted to decrease
anxiety responses to 35% CO2 in subjects with panic disorder [Bertani et al 1997]. In
this study, however, a single dose of 25 mg nortriptyline did not affect baseline or
stimulated hormonal, cardiovascular or psychological measures.
Whilst is was anticipated that pre-treatment with naltrexone and perhaps paroxetine
might increase anxiety symptomatology and possibly hormonal and cardiovascular
responses to CO2 inhalation, it is interesting to note that other studies of pre-treatment
with anxiogenic agents either had no effect or in fact reduced anxiety responses to 35%
CO2. Pols et al [1994], for example, pre-treated healthy volunteers with yohimbine
(20mg) or placebo before giving 35% CO2 in a randomised cross-over study.
Yohimbine pre-treatment did not produce the anticipated increase in anxiety responses
to CO2 compared to placebo. Schruers et al [2000b], pre-treated healthy volunteers
with 10 µg of the anxiogenic agent CCK-4 in a randomised, double-blind, placebo
controlled manner. Following 35% CO2, instead of the anticipated increase in anxiety
responses, anxiety was in fact reduced compared to placebo. The authors postulate that
CCK-4 and CO2 acted through different neurobiological systems to inhibit rather than
enhance each other.
142
4.4. Mineralocorticoid/glucocorticoid study
4.4.1. Introduction and methods
The relationship between the two major effector limbs of the stress response (the HPA
and SAM axes) is a complex one that is maintained through direct and indirect
reciprocal connections at both central and peripheral levels. However, the importance
of these systems in the maintenance of normal health requires their regulation to be
finely controlled with multiple, multi-level inputs that have extensive redundancy and
plasticity. Stress responsive noradrenergic neurones from the LC, the caudal nucleus of
the solitary tract (A2 cell group) and from the medullary A1 group all project directly to
the hypothalamic PVN where they influence both CRH and ACTH activity [Zeigler et
al 1999, Habib et al 2001, Carrasco and Van Der Kar 2003]. Brainstem noradrenergic
neurones also project directly to the forebrain and amygdala which indirectly influence
HPA activity [Carrasco and Van Der Kar 2003]. CRH releasing neurones, in addition to
regulating ACTH release from the anterior pituitary, also project to brainstem
noradrenergic centres where they influence activity particularly of the LC [Pacak et al
1995, Koob 1999, Carrasco and Van Der Kar 2003].
Glucocorticoids, the final effectors of the HPA axis, regulate the stress response through
their action on two receptor subtypes. Within the brain, high affinity, type I or
mineralocorticoid (MR) receptors have a limited distribution, being found particularly
in the hippocampus, where they are involved in feedback processes important for the
normal day-to-day variation in cortisol levels. In contrast, low affinity type II
glucocorticoid (GR) receptors have a more widespread distribution and are important
regulators of the glucocorticoid response to stress [de Kloet 1991, Harbuz and Lightman
1992, Chrousos 1998, Young et al 1998, Sapolsky et al 2000]. Activation of
143
mineralocorticoid receptors in the hippocampus, for example, promotes GABA-ergic
activity that in turn inhibits the activity of the HPA axis [Carrasco and Van Der Kar
2003]. Peripherally, the adrenal cortex (responsible for glucocorticoid production) and
the adrenal medulla (responsible for catecholamine, particularly adrenaline, production)
is regulated by multiple extra- and intra-adrenal inputs. ACTH is the principal regulator
of cortisol biosynthesis and release from the adrenal cortex, however, steroidogenesis is
also influenced by the SAM axis, with stimulation of neural (noradrenergic) inputs to
the adrenal cortex increasing glucocorticoid synthesis and release [Bornstein and
Chrousos 1999]. Similarly, a wide variety of other peptides released from sympathetic
and other nerves, including adrenaline, serotonin, neuropeptide Y and substance P, have
also been shown to modulate adrenocortical activity [Bornstein and Chrousos 1999].
Glucocorticoids, on the other hand, are also important in the regulation of adrenaline
release from the adrenal medulla. The medullary enzyme phenylethanolamine-N-
methyltransferase (PNMT) controls the synthesis of adrenaline from noradrenaline
within the adrenal medulla. The activity of this enzyme is dependent on ACTH and
high intra-medullary concentrations of glucocorticoids [Wurtman 2002], with a number
of studies demonstrating an impaired adrenaline response to various stressors in the
setting of ACTH and glucocorticoid deficiency [Kvetnansky et al 1995, Jeong et al
2000, Wurtman 2002]. On the basis of this interaction, glucocorticoid deficiency (as
produced by metyrapone administration) might be expected to impair the SAM response
to stress.
On the other hand, MR antagonists (such as spironolactone) would be expected to
inhibit HPA axis negative feedback and result in increased cortisol secretion. This is
indeed the case in rodents [Bradbury et al 1994, Young et al 1998] and has been shown
in one human study. Young et al [1998] administered a high dose of spironolactone
144
(400 mg) to healthy volunteers and demonstrated a transient (< 2 hour) elevation of
cortisol. Two doses of spironolactone (400 mg 5 hours apart) produced a more
sustained cortisol elevation, but did not increase ACTH levels.
The effect of glucocorticoid deficiency and mineralocorticoid antagonism in normal
volunteers was assessed using the standard 35% CO2 challenge as previously described.
Metyrapone and spironolactone were chosen because of their known safety, selectivity
and ease of administration (both are available orally).
Nine (5 male) healthy volunteers mean age 22 years (range 20 – 39) who had not
participated in any previous CO2 related studies were recruited from within the
University of Bristol. Volunteers attended on four occasions one week apart. At each
visit they received a single dose of 35% CO2 according to the same protocol as
described above. For each visit volunteers were pre-treated on one occasion with both
spironolactone and metyrapone, once with metyrapone only, once with spironolactone
only and once with placebo only. Visits were randomised, and subjects were blinded to
the treatment condition on each occasion. Doses used were spironolactone 50 mg,
metyrapone 750 mg or placebo in an equivalent number of capsules. For each visit, a
dose was taken the evening before the test with a second dose the morning of the test.
The treatments were all well tolerated with no significant side effects from either drug
reported. All the tests were conducted between 1 and 3 pm and all subjects completed
all the tests. Baseline and test VC was recorded and all test breaths were within 80% of
the baseline breath. Cardiovascular measures were recorded as before with the
Dynamap monitor recording pulse rate and blood pressure every minute for 5 minutes
before and 5 minutes after CO2 exposure. Psychological responses were recorded using
the same visual analogue scales. An intravenous line for blood sampling was placed 30
145
minutes before testing commenced and samples for cortisol levels were taken at
baseline, 10, 20 and 30 minutes after exposure. Blood samples were not collected on 2
occasions (once during a spironolactone only visit and once during a
spironolactone/metyrapone visit) because of failed intravenous access at the time of
testing.
Repeated measures analysis of variance was used to determine the effects of CO2
exposure and the differences between the groups with paired t-test analysis used to
determine differences between single time points including baseline measures for the
placebo and active treatment visits for each subject.
4.4.2. Results
There was no difference in baseline psychological or cardiovascular parameters for any
of the visits. Baseline cortisol levels, shown in Figure 4.07, were significantly reduced
on the two occasions subjects took metyrapone. There was no independent or
additional effect of spironolactone on cortisol levels.
A single breath of 35% CO2 produced a significant increase from baseline in cortisol,
prolactin, SBP, anxiety, fear and breathlessness with a significant fall in HR (p < 0.05
for all). Cortisol responses to 35% CO2 were equivalent in the control, spironolactone
only and metyrapone only conditions, but there was no significant cortisol response in
the combined spironolactone/metyrapone condition (p = 0.03 control vs
spironolactone/metyrapone) (Figure 4.08). Psychological responses, including anxiety,
fear, breathlessness (shown in Figure 4.09) and various somatic symptoms (not shown)
146
were no different between any of the treatment or control groups. Similarly, as shown
in Figure 4.10, SBP and HR responses were equivalent in all groups.
147
Baseline Cortisol
C S M S/M
0
100
200
300
400
500
nmol
/l *† **‡
Figure 4.07. Baseline cortisol levels in the control (C), spironolactone (S), metyrapone (M) and combined (S/M) groups. *p=0.03 C v M; †p=0.006 C v SM **p=0.0008 S v M; ‡p=0.0005 S v SM
148
Cortisol response to 35% CO2
C S M S/M
-100
0
100
200
nmol
/l
Cortisol response to 35% CO2
C S M S/M
0
25
50
75
100
nmol
/l
Figure 4.08. Peak cortisol response to 35% CO2 in the control (C), spironolactone (S), metyrapone (M) and combined (S/M) groups. *p=0.03 C v SM; †p=0.009 S v M
149
Psychological responses to35%CO2
Anx
iety
Fear
Bre
athl
essn
ess
0
25
50
75CSMS/Mm
m
Figure 4.09. Psychological responses to 35% CO2 in the control (C), spironolactone (S), metyrapone (M) and combined (S/M) groups. Differences between groups are all non-significant.
150
SBP
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
100
110
120
130CSMSM
Time (mins)
mm
Hg
SBP
C S M S/M
-10
0
10
20
30
mm
Hg
HR
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
50
60
70
80CSMSM
Time (mins)
bpm
HR
C S M S/M
-50
-25
0
25
bpm
Figure 4.10. Cardiovascular responses to 35%CO2 in thecontrol (C), spironolactone (S), metyrapone (M) and combined (S/M) groups. Differences between groups are all non-significant.
151
4.4.3. Discussion
Two doses of spironolactone alone (50 mg 12 hours apart) had no effect on baseline
cortisol levels 4-5 hours after the second dose compared to placebo. Further,
spironolactone alone had no effect on CO2 stimulated cortisol levels. This may well
have been due to the smaller doses and longer intervals used in this study, however, as
has been seen previously baseline cortisol levels have not affected cortisol
responsiveness to a single breath of 35% CO2. Spironolactone alone also had no effect
on cardiovascular or psychological responses to CO2.
Metyrapone, a specific inhibitor of the 11β-hydroxylase enzyme, prevents cortisol
synthesis and reduces circulating cortisol levels [Laborie et al 2003]. Effects on other
aspects of the HPA and SAM axes are complex, with acute dosing likely to reduce
cortisol levels and possibly impair adrenergic adrenomedullary responses. Longer term
dosing, on the other hand, as has been shown by Laborie et al [2003] to produce
positive feedback with increased CRH neurotransmission and increased ACTH. This is
likely to enhance the effect of CRH on brainstem noradrenergic activity and increase
adrenal intra-medullary adrenaline production thereby enhancing the sympathetic
behavioural and cardiovascular responses to stress. Consistent with this, cortisol levels
in those individuals taking metyrapone were significantly reduced compared to the
placebo and spironolactone only visits. Cardiovascular and psychological responses,
however, were not affected by metyrapone pre-treatment. The likely explanation for
this is that whilst metyrapone reduced cortisol levels by about 50%, there was still
sufficient cortisol present to ensure normal glucocorticoid responsiveness, particularly
within the brain. Alternative possibilities include the effect of cortisol-like molecules
152
formed from other steroid biosynthetic pathways associated with the inhibition of 11β-
hydroxylase [Lamberts et al 1987].
The combination of spironolactone and metyrapone again had no effect on
psychological and cardiovascular responses, although this group did demonstrate a
reduced cortisol response to the CO2 challenge. Given both spironolactone alone and
metyrapone alone were each associated with significant cortisol response to the CO2
challenge, it is difficult to postulate a mechanism whereby the combination would
suppress a cortisol response. The anticipated enhanced cardiovascular response from
increased CRH action on sympathetic brainstem centres was not apparent, suggesting
the principal cardiovascular and psychological responses are driven from a primary
stimulation of brainstem noradrenergic centres and/or limbic centres with HPA
responses occurring as a secondary phenomenon.
153
4.5. Lactation study
4.5.1. Introduction and methods
Lactation, at least in animal models, represents a time of relative neurohormonal
hyporesponsiveness to a variety of stressors [Lightman et al 1997, Lightman et al 2001,
Heinrichs et al 2002]. Studies in rodents that have included both psychological
stressors such as noise [Windle et al 1997], conditioned footshock [Shanks et al 1997]
and restraint [Shanks et al 1999], as well as physical stressors such as intra-peritoneal
NaCl [Lightman and Young 1989] or lipopolysaccharide [Lightman 1993], have
demonstrated attenuated ACTH, cortisol, catecholamine, oxytocin and prolactin
responses to stress [Higuchi 1989, Lightman 1993, Altemus et al 1995, Neumann et al
1998, Heinrichs et al 2001, Lightman et al 2001, Russell 2001, Heinrichs et al 2002].
Further, behavioural responses to stress in rodents also appeared attenuated during
lactation [Altemus et al 1995, Heinrichs et al 2001, Lightman et al 2001, Russell 2001,
Heinrichs et al 2002]. This hyporesponsiveness was a generalised feature of lactation
and was not limited to selective times such as immediately following suckling [Windle
et al 1997, Heinrichs et al 2002]. It has been felt that neuroendocrine and behavioural
stress hyporesponsiveness during lactation is advantageous both for the mother and the
infant. For the mother, attenuated stress responsiveness is thought to protect against
distracting stimuli during feeding, conserve energy for suckling and rearing, improve
nutritional milk quality and promote protective immune activity [Lightman et al 1997,
Lightman et al 2001, Russell 2001, Heinrichs et al 2002]. For the infant it protects
against the potentially damaging effects of exposure to high glucocorticoid levels at a
time of vulnerability of the developing brain [Lightman et al 1997, Russell 2001,
Heinrichs et al 2002]. The neurobiological pathways mediating this reduced stress
154
reactivity are yet to be fully elucidated. The principal hormones involved in mammary
gland development, growth, milk production and ejection include oxytocin, prolactin,
estrogens, GH, thyroid hormone, ACTH and glucocorticoids, most of which also have
significant roles in behavioural modification [Lightman and Young 1989, Tucker 1994,
Lightman et al 1997, Neumann et al 2000, Lightman et al 2001, Heinrichs et al 2002]
and stress responsiveness. Oxytocin and prolactin, for example, are both increased
during lactation and are both inhibitors of the HPA axis [Heinrichs et al 2002].
Endogenous opiates are also modulators of HPA axis activity, inhibiting ACTH and
cortisol release. Hypothalamic opiate receptors have been shown to change during
lactation and are associated with reduced oxytocin and HPA activity [Carter and
Lightman 1987, Hammer and Bridges 1987]. It is likely that the complex interplay of
these and a number of other neurohormones, neurotransmitters as well as environmental
influences are involved in the plasticity of stress responses seen during lactation.
The effect of lactation on stress-response systems in humans is not well characterised
but appears more complex than the situation in rodents. The first study to be performed
in this area in humans examined neuroendocrine responses to physical stress (treadmill
exercise) in healthy women 7-18 weeks post-partum [Altemus et al 1995]. In that
study, breastfeeding began 60 minutes before the start of the exercise challenge and
basal ACTH and cortisol levels were the same for lactating women and non-lactating
controls. Basal prolactin levels were significantly higher in the lactating group. In
response to the exercise, ACTH and cortisol increased to a lesser degree in the lactation
group than in the control group, and prolactin levels fell steadily. Oxytocin levels were
unchanged. The authors concluded that HPA axis responses to physical stress were
suppressed in lactating women, and this effect was independent of oxytocin. A second
study by the same group [Altemus et al 2001] used a psychological stressor (Trier
155
Social Stress Test) in healthy women 6-24 weeks post-partum administered 60 minutes
after feeding. This study also examined bottlefeeding mothers as well as a non-lactation
control group. Baseline ACTH and cortisol levels were equivalent as were the
increases in ACTH, cortisol, blood pressure, heart rate and anxiety symptoms in all
groups. A similar lack of endocrine hyporesponsiveness to psychosocial stress was seen
in a study by Redwine et al [2001]. In this study breast or bottlefeeding within an hour
of being challenged (TSST) had no effect on adrenocortical responses.
It has been well established that suckling itself will suppress the HPA axis probably as a
result of the inhibitory actions of oxytocin and prolactin [Amico et al 1994]. Heinrichs
et al [2001] undertook a study using a similar psychological stressor (TSST) to
determine the effects of acute suckling on HPA responsiveness. The aim was to see if
there is a sensitive time period during lactation where the HPA axis may be
hyporesponsive as opposed to the apparent general hyporesponsiveness of the HPA axis
throughout lactation in rodents. Healthy women, 6-18 weeks post-partum either
breastfed or just held their infant for a 15 minute period with the onset within 30
minutes of receiving the TSST challenge. Baseline ACTH, cortisol, oxytocin, prolactin
and catecholamine levels were equivalent prior to exposure, however, HPA axis
responses were markedly diminished in the breastfeeding group. Prolactin levels in this
group also fell steadily throughout and following the stress. The authors conclude that
unlike rodents humans demonstrate a time sensitive period immediately following
suckling that is associated with significant blunting of HPA responses to psychological
stress. This blunting is not apparent if the stress exposure occurs about 1 hour after
suckling [Heinrichs et al 2001, Heinrichs et al 2002].
Twenty two healthy women participated in the study after giving written informed
consent. Potential breast and bottle-feeding subjects received information regarding the
156
study antenatally during a routine visit to the Antenatal Day Assessment Unit at St
Michael’s maternity Hospital, Bristol. Subjects were then recruited approximately 4
weeks after delivery. All subjects were paid for their participation. Subjects were asked
to attend once, where they all received a single breath of 35% CO2 according to the
same protocol as described above.
Fourteen subjects were 6-8 weeks post-partum and were in good general health
following uncomplicated term deliveries. Participants were non-smokers who were not
taking any regular medications. All post-partum participants completed an interview,
physical examination and the Edinburgh Post-Natal Depression score (see appendix 6)
[Cox et al 1987] prior to inclusion in the study. In addition to the standard inclusion
and exclusion criteria applied in previous studies, subjects were only included if they
had a low risk of post-natal depression based on the EPDS. Eight of the fourteen post-
partum women (age range 18 – 34) were exclusively breastfeeding (BF) their infants,
whilst the remaining six (age range 22-36) had been exclusively bottlefeeding (BO)
their infants from birth. Eight control (C) women (age range 22-27) were recruited
from within the University of Bristol. None of these subjects were, or had ever
previously been, pregnant.
All subjects attended at 11 am, and an antecubital intravenous line was placed 45
minutes before the CO2 exposure. Following insertion of the cannula, mothers fed
(either breast or bottle according to their standard practice) their infants for
approximately 15 minutes. Thirty minutes after arrival, the first of two baseline blood
samples were taken. Cardiovascular monitoring commenced and the second baseline
sample was taken after a further 10 minutes. The breath of 35% CO2 was then
administered. The time from cessation of feeding to receiving the CO2 was
157
approximately 20 to 30 minutes. Sampling continued for 30 minutes following
exposure. A nurse was present during the procedure to assist with caring for the infants.
Two-way, repeated measures, analysis of variance was used to determine between-
group differences and time effects, with two-tailed t-tests used to compare single time-
point data including baseline differences. Paired t-test analysis was used to determine
differences in maximum response from baseline within each group.
4.5.2. Results
One individual (from the bottlefeeding group) did not take an adequate breath of CO2
(VC < 80% of baseline) and data from this individual has been excluded from the
analysis. All other subjects completed the study without significant adverse effects.
Physical, hormonal and psychometric characteristics of all subjects at baseline are
presented in Table 4.02. There was no difference at baseline between the control and
bottlefeeding groups, however, as shown in Figure’s 4.11 and 4.12, baseline cortisol
levels in the breastfeeding group were significantly reduced compared with both the
control and bottlefeeding groups (p = 0.002 and p = 0.003 respectively). In addition,
baseline prolactin levels in the breastfeeding group were significantly higher compared
with both the control and bottlefeeding groups (p = 0.0003 and p = 0.003, respectively).
For all the groups combined, there was a significant increase in cortisol, SBP, and
subjective feelings of anxiety, fear and breathlessness (p < 0.05 for all). Further, there
was also a significant fall in heart rate (p < 0.05). For cortisol, two-way ANOVA
revealed a significant group effect between control (C) and breastfeeding (BF) (F(1.4) =
40.2, p < 0.0001) and between breastfeeding (BF) and bottlefeeding (BO) (F(1.4) = 26.3,
p < 0.0001). Post-hoc analysis indicated significant differences at all time points for C
158
and BF (p < 0.05) and at –5, 2 and 10 minutes for BF and BO (p < 0.05). Maximum
cortisol change from baseline, however, was equivalent in all three groups (Figure
4.11).
Two-way ANOVA also revealed a significant group effect between control and
breastfeeding (F(1.3) = 80.3, p < 0.0001) and between breastfeeding and bottlefeeding
(F(1.3) = 39.5, p < 0.0001) for prolactin responses. Post-hoc analysis indicated
significant differences at all time points for both C and BF (p < 0.01 for all) and for BF
and BO (p < 0.05 for all). There was no significant increase in prolactin from baseline
in the C or the BO group, however, there was sustained and marked fall in prolactin in
the BF group with a significant maximum decrease from baseline (p = 0.04) (Figure
4.12).
Baseline SBP and HR measures were no different between each group. Within each
group there was a significant increase in SBP and a significant fall in HR following CO2
exposure (p < 0.05 for maximum change from baseline for SBP and HR in all three
groups), as shown in Figure 4.13. Between groups, the change in cardiovascular
parameters over time and the maximum change in SBP and HR were equivalent.
Similarly, anxiety, fear, breathlessness and other somatic symptoms of fear were
equivalent at baseline. Within each group there was a significant increase in these
symptoms (p < 0.05 for all) that was equivalent between each group (Figure 4.14).
159
Table 4.02. Baseline hormonal, cardiovascular and psychological parameters for
control (C), breastfeeding (BF) and bottlefeeding (BO) women.
Control (n=8) Breastfeeding (n=8) Bottlefeeding (n=5)
Age (yr) 24.7+/-1.5 29.2+/-1.9 31.2+/-3.0
Cortisol (nmol/l) 300.8+/-25.9 191.1+/-12.8* 274.0+/-16.7
Prolactin (nmol/l) 138.0+/-17.6 2071+/-432.9† 107.0+/-15.1
SBP (mmHg) 108.8+/-4.9 110.0+/-4.2 114.2+/-4.0
HR (b/m) 67. 8+/-3.7 76.1+/-1.6 73.0+/-2.3
Anxiety (mm) 10.6+/-3.9 11.8+/-5.9 13.0+/-5.8
Fear (mm) 6.8+/-2.6 6.2+/-4.0 11.0+/-5.5
Breathlessness
(mm)
1.2+/-0.8 3.1+/-3.1 1.0+/-1.0
*p=0.002 C v BF and p=0.003 BF v BO
†p=0.0003 C v BF and p=0.003 BF v BO
160
Baseline cortisol
0
100
200
300
400controlbreastfeedingbottlefeeding
nmol
/l
Cortisol response to 35% CO 2
Con
trol
Brea
stfe
edin
g
Bottl
efee
ding
-100
0
100
200
300
nmol
/l
Absolute cortisol response to35%CO2
-10 0 10 20 30 40
100
200
300
400controlbreastfeedingbottlefeeding
Time (mins)
nmol
/l
**
Figure 4.11. Baseline (top panel), absolute (middle panel) and maximum cortisol (lower panel) responses to a single breath of 35% CO2 in control women (C; n=8), breastfeeding women (BF; n=8) and bottlefeeding (BO; n=5) women. *p=0.002: C v BF; †p=0.003: BF v BO; p=ns: C v BO. **p<0.05: C v BF all time-points and BF v BO at 0, 2, 10 mins. ‡p<0.05 ∆max cortisol within each group from baseline. Responses between each group were not significantly different.
161
Baseline prolactin
0
1000
2000
3000controlbreastfeedingbottlefeeding
nmol
/l
Prolactin response to 35% CO 2C
ontro
l
Bre
astfe
edin
g
Bot
tlefe
edin
g
-500
-250
0
250
nmol
/l
Percent prolactin response to35% CO2
-20
-10
0
10
20ControlBreastfeedingBottlefeeding
%
162
Figure 4.12. Baseline (top panel) and stimulated prolactin (middle andlower panels) responses to a single breath of 35% CO2 in control women (C; n=8), breastfeeding women (BF; n=8) and bottlefeeding (BO; n=5) women. Middle panel is absolute change, lower panel is percent change. *p=0.0003: C v BF †p=0.003: BF v BO p=ns: C v BO. **p=0.04 maximum absolute change in prolactin; ‡p<0.05 maximumSBP response to 35% CO 2
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
80
100
120
140
160controlbreastfeedingbottlefeeding
time
mm
Hg
SBPresponse to 35% CO2
Con
trol
Bre
astfe
edin
g
Bot
tlefe
edin
g
0
10
20
30
40
mm
HgHR response to 35% CO 2
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
50
60
70
80
90controlbreastfeedingbottlefeeding
Time (min)
b/m
HR response to 35%CO 2
Con
trol
Bre
astfe
edin
g
Bot
tlefe
edin
g
-30
-20
-10
0
10b/
m
Figure 4.13. SBP (top panels) and HR (lower panel) responses to a single breath of 35% CO2 in control (C), breastfeeding (BF) and bottlefeeding (BF) women. *p<0.05 ∆max SBP and ∆max HR. Responses between groups are all equivalent.
163
Psychological responses to
0
25
50
75
100ControlBreastfeedingBottlefeeding
%
Anx
iety
Fear
Bre
athl
essn
ess
Figure 4.14. Anxiety, fear and breathlessness responses to a single breath of35% CO2 in control (C), breastfeeding (BF) and bottlefeeding (BF) women. *p<0.05 ∆max anxiety, fear and breathlessness from baseline. Responses between groups are all equivalent.
35% CO2 * * *
164
4.5.3. Discussion
In this study, the CO2 challenge was timed to occur within 30 minutes of breastfeeding
so as to coincide with the expected HPA hyporesponsive period. Baseline cortisol
levels in the breastfeeding group were significantly lower than in the control or
bottlefeeding group consistent with the expected effect of suckling induced HPA axis
suppression. Baseline prolactin levels in this group were also significantly elevated
again consistent with the recent suckling. This is in contrast to the study by Heinrichs et
al [2001] where sampling after a similar time period from the completion of suckling
was not associated with significant differences in baseline cortisol or prolactin levels. It
is, however, consistent with the findings of Amico et al [1994] that did demonstrate
cortisol suppression 15 minutes after the completion of feeding.
In response to the CO2 challenge, the absolute (and percent) maximum increase in
cortisol from baseline was equivalent in all three groups. This suggests the ability of
the HPA axis to respond to this challenge is not blunted during the lactation phase and
is not affected by a suppressed baseline cortisol level. This would also be consistent
with the notion that acute CO2 exposure (single breath of 35% CO2) represents an
immediate threat to respiratory homeostasis and a diminished response to this stressor
would be inappropriate particularly for a nursing mother.
Consistent with previously mentioned studies in humans, prolactin levels fell steadily
following CO2 exposure in the breastfeeding group. Of note, however, there was only a
small increase in prolactin following CO2 in the control group (as opposed to prolactin
responses in earlier studies) and no significant change in prolactin from baseline in the
bottlefeeding group. The reason for this is yet to be determined.
165
As in the study by Altemus et al [2001] baseline SBP was equivalent in all three groups.
However, unlike the Altemus study that indicated bottlefeeders have a reduced cardiac
vagal tone associated with significantly increased basal heart rates, there was no
difference in HR in all three groups in this study. A single breath of 35% CO2 increased
SBP and reduced HR to a similar degree in all groups indicating that the SAM response
to 35% CO2 is not attenuated in lactation and is not influenced by background HPA axis
activity. Similarly, as was the case in the Altemus study [2001], baseline anxiety and
somatic symptoms of fear were no different between the groups and 35% CO2 caused
significant anxiety responses that were equivalent between the three groups. This
would again suggest that lactation has no effect on anxiety responses to 35% CO2 and
that these responses are also not dependent on basal cortisol levels. Similar findings
have been found with various other stressors applied during lactation in both human and
animal studies [Heinrichs et al 2001]. Indeed even in those studies where HPA axis
responses were suppressed, anxiety and emotional arousal responses were no different
in the breastfeeding groups compared with either control or bottle-feeding mothers
[Heinrichs et al 2001].
Overall, this study confirms that within 20-30 minutes of suckling, baseline cortisol
levels in breastfeeding mothers are suppressed, an effect that is probably mediated
through the inhibitory effects of prolactin and oxytocin on the PVN. However, unlike
the response to a psychological stressor, exposure to acute hypercapnia is not associated
with diminution of HPA, autonomic or psychological stress responses. This would be
compatible with the notion that suppression of stress responses may be advantageous to
a nursing mother faced with stressors that are not pertinent to survival, but when
physiological homeostasis is threatened survival demands a full neuroendocrine and
behavioural response.
166
4.6. Peripheral versus central autonomic nervous system effects
4.6.1. Introduction and methods
The role of the autonomic nervous system in the CO2 response was examined when the
opportunity to study an adult with the congenital central hypoventilation syndrome
(CCHS) presented itself. This syndrome (as described below) is associated with central
CO2 insensitivity and was compared to an individual who had peripheral pulmonary
denervation following dual lung-heart transplant for cystic fibrosis. Because of the
extreme rarity of the CCHS syndrome (approximately 200-300 living children and
young adults worldwide [American Thoracic Society 1999]) and the limited number of
heart-lung transplant patients well enough to undergo evaluation in Bristol, only one
individual with each condition was studied. As a result, formal statistical evaluation of
their responses has not been performed and the significance of their results should be
considered in this context.
One of the individuals studied was a 27 year old male who had had a dual lung-heart
transplant 3 years previously for cystic fibrosis and at the time of this study was well,
being maintained on anti-rejection therapy that included prednisolone 5 mg daily,
cyclosporin and azathioprine. His cystic fibrosis had been complicated by both
exocrine and endocrine pancreatic failure and he was on pancreatic enzyme
supplements and insulin, but had no evidence of diabetic complications and was
normotensive at the time of the study. The other participant was a 43 year old male
with CCHS. He was on non-invasive night-time ventilation and his condition had been
complicated by secondary polycythaemia that had required previous venesection. He
was on no regular medications and was normotensive at the time of the study.
167
4.6.2. Results
As shown in Figure 4.15, the lung transplant individual, as a result of his exogenous
glucocorticoid use, had a suppressed baseline cortisol with no change in his cortisol
level seen following CO2 exposure. He did, however, demonstrate a rise in prolactin to
a similar degree as that seen in healthy volunteers. The individual with CCHS showed
no change in his cortisol and only a small (5 nmol/l or 2.5%) increase in prolactin
following CO2 exposure.
The cardiovascular responses between the two individuals were markedly different.
Pulmonary denervation was associated with a slightly higher baseline heart rate, a
significant rise in SBP (+11 mmHg) in response to CO2 but no bradycardia (+2 b/m).
The cardiovascular responses over time and the peak change following CO2 exposure
are shown in Figure 4.16. In contrast, the individual with CCHS demonstrated no rise
in SBP (+0 mmHg) but he did show a marked bradycardic response (-12 b/m). Figure
4.17 illustrates the psychological response to CO2. Both individuals demonstrated mild
anticipatory anxiety, but whilst the lung transplant subject had a significant anxiety
response to the CO2 challenge equivalent to that seen previously in healthy volunteers,
the CCHS individual experienced no significant anxiety. Similarly, the lung transplant
individual also experienced a significant increase in the symptoms of fear, feeling hot,
blurred vision and dizziness to a degree that was very similar to the response of normal
individuals. The CCHS subject experienced none of these symptoms apart from some
very mild dizziness and breathlessness. The lung transplant subject experienced no
breathlessness.
168
Cortisol response to 35% CO2
Lung Tx CCHS0
100
200
300
400PrePost
nmol
/l
Prolactin response to 35% CO2
Lung Tx CCHS
150
175
200
225PrePost
nmol
/l
Figure 4.15. Cortisol (upper panel) and prolactin (lower panel) response to a single breath of 35% CO2 in an individual with cardiopulmonary denervation (lung Tx) and in an individual with central CO2 insensitivity (CCHS). Note the subject post heart-lung transplant was on exogenous glucocorticoids.
169
SBP
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
125
135
145
155Lung TxCCHS
Time (mins)
mm
Hg
SBP response to 35% CO2
Lung Tx CCHS
140
145
150
155
160PrePost
mm
Hg
HR
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
85
90
95
100
105Lung TxCCHS
Time (mins)
b/m
HR response to 35% CO 2
Lung Tx CCHS
85
90
95
100
105prepost
b/m
Figure 4.16. SBP (upper panel) and HR (lower panel) response to a single breath of 35% CO2 in an individual with cardiopulmonary denervation (lung Tx) and in an individual with central CO2 insensitivity (CCHS).
170
Anxiety
Lung Tx CCHS
0
25
50
75PrePost
mm
Fear
Lung Tx CCHS
0
25
50
75
mm
Breathlessness
Lung Tx CCHS
0
25
50
75
mm
Feeling hot
Lung Tx CCHS
0
25
50
75
mm
Blurred vision
Lung Tx CCHS
0
25
50
75
mm
Dizziness
Lung Tx CCHS
0
25
50
75
mm
Figure 4.17. Psychological (anxiety, fear, breathlessness and the common somatic symptoms of fear associated with CO2 exposure) response to a single breath of 35% CO2 in an individual with cardiopulmonary denervation (lung Tx) and in an individual with central CO2 insensitivity (CCHS).
171
4.6.3. Discussion
Idiopathic congenital central hypoventilation is a rare syndrome characterised by a
failure of the automatic control of breathing [Gozal 1998, American Thoracic Society
1999]. Typically, ventilation is most severely affected during non-rapid eye movement
(NREM) sleep, a time where automatic neural control of breathing predominates.
Severe hypoventilation usually requiring assisted ventilation occurs during sleep, with
lesser, albeit variable, degrees of hypoventilation during wakefulness. When awake
many patients can sustain normal ventilation without assistance, although some will
require continuous ventilatory assistance [Gozal 1998]. A similar condition has been
described following high cervical trauma or in people with cerebrovascular
malformations, infections or tumours in and around the brainstem [Severinhaus and
Mitchell 1962, Bower and Adkins 1980, Jensen et al 1988, Mukhopadhyay and
Wilkinson 1990, Weese-Mayer et al 1992, Gozal 1998, Weese-Mayer et al 2001] and a
related animal model has been developed by inducing lesions of the intermediate area of
the ventral medullary surface [Schlaefke et al 1979]. Despite this, the majority of
patients with idiopathic CCHS do not have discrete anatomical or neuropathological
lesions. There have only been two case reports of discrete anatomical lesions identified
in children with idiopathic CCHS, one with arcuate nucleus agenesis [Folgering et al
1979] and another with neuronal loss in the region of the reticular nuclei and the lower
cranial nerve nuclei [Weese-Mayer et al 2001]. More likely, the origin of the condition
is genetic, with about 20% of sufferers having associated Hirschbrung’s disease (a
neurocristopathy characterised by loss of parasympathetic intrinsic ganglion cells of the
hindgut and associated with mutation of the RET proto-oncogene) [Haddad et al 1978,
American Thoracic Society 1999]. There are a few case reports of families with the
condition either having affected siblings or having a common RET or other gene
172
mutation, although these are only the minority [Gozal 1998, American Thoracic Society
1999].
Associated autonomic dysfunction occurs relatively commonly, although to a variable
degree. Commonly reported features include; constipation, oesophageal dysmotility,
pupillary abnormalities, abnormalities of cardiac rhythm, decreased basal body
temperature and altered sweating [Severinhaus and Mitchell 1962, Weese-Mayer et al
1992, Commare et al 1993, Weese-Mayer et al 1993, Silvestri et al 2000]. Further,
neuropsychological studies have reported reduced pain perception [American Thoracic
Society 1999], reduced anxiety perception [Pine et al 1994] and a reduced sensation of
breathlessness with prolonged breathholding [American Thoracic Society 1999,
Spengler et al 2001]. Physiological evidence of intact peripheral chemoreceptor activity
with impaired central chemoreceptor function [Gozal et al 1993, Spengler et al 2001]
has led to the current hypothesis that the condition is principally related to impaired
central CO2 chemosensation along with abnormalities of central autonomic integration.
The origin of this abnormality is most likely within the brainstem resulting
predominantly in ventilatory but also other autonomic abnormalities [Gozal 1998,
American Thoracic Society 1999, Spengler et al 2001, Weese-Mayer et al 2001].
In contrast to the situation of central chemoreceptor and autonomic activity described in
CCHS, pulmonary denervation, as occurs following heart-lung transplant, represents a
model of peripheral autonomic denervation affecting both afferent and efferent
pulmonary autonomic pathways. Amphibians and reptiles possess CO2 sensitive
olfactory receptors that cause a dose-dependent decrease in breathing when stimulated
by CO2 [Coates 2001] and there is some evidence that mammals may possess similar
chemoreceptors that play a role in olfaction [Coates 2001]. Pulmonary CO2 sensitive
173
chemoreceptors have not been identified in humans [Lahiri and Forster 2003] and
whether chemoreceptors located within the laryngopharynx exist, or have a role in
mediating rapid responses to acute CO2 exposure in humans is not known.
Pulmonary stretch receptors and their vagal afferents are known to mediate important
ventilatory [Trachiotis et al 1994] and cardiovascular responses to breathing [Taha et al
1995]. Taha et al [1995], for example, have shown that in pulmonary denervated
individuals, respiratory sinus arrhythmia is significantly reduced compared to normal
individuals as a result of altered vagal feedback pathways.
It is important to remember that whilst these individuals represent interesting
physiological models of impaired autonomic control of hypercapnia, only single
individuals with each condition were tested because of the rarity of their respective
conditions. As such their responses are purely descriptive and formal statistical analysis
has not been used to examine differences in responses. Similarly, given the intra-
individual variability in responses in normal individuals, the responses seen in these two
individuals should not be extrapolated to include everyone with similar conditions.
The individual following lung transplantation was taking exogenous glucocorticoids
(prednisolone 5 mg daily) as part of his anti-rejection regime and this would explain
both the suppressed basal cortisol level and the absent change in cortisol following CO2
exposure. His prolactin response was similar to that seen for normal individuals. In
contrast the CCHS subject had normal baseline cortisol and prolactin levels, but failed
to show a response of either to CO2.
174
Consistent with the vagal denervation following lung transplantation, the post-transplant
subject had a higher resting heart rate, and failed to show any drop in heart rate
following CO2, although his SBP response was normal. In contrast, the CCHS subject
demonstrated a normal bradycardic response, but failed to show a rise in SBP. As
expected neither subject described much in the way of breathlessness following the
CO2, however, anxiety, fear and their associated somatic symptoms were experienced
normally by the lung transplant individual, but not at all by the CCHS individual.
Taken together, these results suggest that central chemoreceptor sensation and
integration is essential for generating a neurohormonal (cortisol and prolactin) and
psychological response to CO2. Further, central brainstem stimulation is necessary for a
SBP response. The bradycardia, on the other hand, is likely to be a direct vagal
response as a result of direct stimulation of either the nerve or its central nuclei. It had
been suggested by Tenney [1956] in early experiments of CO2 on isolated heart muscle
preparations that there is a direct effect of CO2/pH causing a decrease in the
spontaneous firing rate of cardiac myocytes and that this may have been an explanation
for the observed bradycardia. If this were the case, it would be expected that heart rate
slowing would still occur in the subject post heart-lung transplant. Vagal denervation
did, however, result in loss of the bradycardic response. In addition, at least in this one
individual with CCHS, stimulation of the vagus was preserved and thus appears to be
separate from other centres involved in mediating the integration and/or response to
CO2 exposure.
175
4.7. Conclusion
The results of the above experiments provide further evidence to suggest that vagally
mediated bradycardia is the first response to CO2 and this is mediated by direct
stimulation of the vagus nerve itself or of its nuclei located within the brainstem. In
addition, it appears that stimulation of the vagus occurs independently of the stimulation
of other brainstem sites responsible for the integration and output of other autonomic
responses to CO2. Stimulation of brainstem noradrenergic centres, particularly the
VLM and the LC amongst others, are likely to be responsible for activation of
descending sympathetic nervous system pathways generating the acute pressor
response. The origins of both the vagal and sympathetic responses are discussed more
fully in the next chapter.
The absence of CO2 chemoreceptors within the hypothalamus and pituitary, coupled
with experimental evidence showing a close correlation between the SAM response and
the cortisol response and the absence of a cortisol response in the subject with central
CO2 insensitivity, suggests HPA axis stimulation occurs as a result of projections from
brainstem noradrenergic centres to the PVN and/or other hypothalamic sites. Further,
the absence of a significant effect on SAM responses in the setting of a
pharmacologically (metyrapone) or physiologically (lactation) suppressed HPA axis
again suggests HPA axis stimulation is driven by brainstem sympathetic centres.
Similarly, SAM responses were not increased in the setting of a disinhibited HPA axis
(naltrexone), although this does not necessarily exclude a role for CRH projections to
the brainstem enhancing or modifying SAM responses to CO2 exposure.
176
Psychological responses also appear to require intact central chemoreceptor
mechanisms and are not dependent on baseline or subsequent cortisol responses. Indeed
this was also evident from studies of panic whereby some panicogens produce panic in
association with cortisol release whilst others produce the same emotional response
without any evidence of HPA axis activation [Sinha 1999]. This suggests the pathways
mediating the psychological response to CO2 operate separately to those that may
mediate the cortisol response. Whether the psychological response occurs as a result of
direct stimulation of limbic centres by CO2 or from indirect activation via the VLM and
LC is not certain. Evidence for a separate pathway, or at least a pathway with different
thresholds of activation, comes from the dose response study that indicated the
significantly lower doses required to produce a psychological response as compared to a
sympathetic (or cortisol) response. As discussed in Chapter 3, acute CO2 exposure may
well represent a ‘limbic insensitive’ threat [Herman and Cullinan 1997] with initial
brainstem activation whilst prolonged exposure to lower doses of CO2 may represent a
‘limbic sensitive’ threat requiring cortical processing first, with subsequent autonomic
and HPA activation.
Manipulating central serotonin, noradrenaline or opiate pathways in normal individuals
had no effect on either psychological or sympathetic responses although there is
considerable evidence that in susceptible individuals serotonin as well as other
neurotransmitters do have a modifying effect on psychological responses to CO2
exposure [Bertani et al 1997, Klaasen et al 1998, Ben Zion et al 1999, Miller et al 2000,
Schruers et al 2000, Perna et al 2002, Schruers et al 2002, Bertani et al 2003].
177
Figure 4.18 indicates schematically a hypothetical pathway that may mediate the
respective hormonal, cardiovascular and psychological response to a single breath of
35% CO2.
In the following chapter, another series of experiments has been performed that attempt
to further examine and characterise the potential mechanisms and pathways that may
mediate the response to acute CO2 exposure. In addition, these studies will also
examine some of the potential clinical implications an altered CO2 response system may
have for particular individuals.
178
Single breath of
35% CO2
Noradrenergic brainstem
Limbic centres
(CnA)
?
FmVrPsbHap(
Vagal centres/vagus nerve
centres(VLM, LC) ?
PVN Bradycardia
NA ACTH
Emotional arousal ↑SBP Cortisol
igure 4.18. Schematic diagram suggesting the principal pathways of CO2ediated HPA, cardiovascular and psychological responses (solid arrows). agal and noradrenergic brainstem centres are directly stimulated by CO2
esulting in bradycardia and an acute pressor response respectively. sychological responses are probably due to a combination of direct CO2timulation of a central fear circuit as well as indirect stimulation via rainstem NA centres. PA responses occur indirectly in response to brainstem NA stimulation and
lso possibly from the CnA. CRH neurones from the PVN also have the otential for augmenting sympathetic and psychological responses represented by hashed arrows), although this role is less clear.
179
CHAPTER 5
THE 35% CO2 MODEL: RESPONSES IN SPECIFIC
SUBPOPULATIONS – FURTHER MECHANISMS
AND POTENTIAL CLINICAL RELEVANCE
180
5.1. The role of the HPA axis
Earlier experiments had identified emotional arousal and autonomic activation as the
principal responses to CO2 inhalation, with HPA axis activation probably occurring in
response to stimulation of either or both of the former two centres. Autonomic
activation appeared to involve direct parasympathetic stimulation as well as central
sympathetic stimulation with the pressor response occurring as a result of noradrenaline
mediated peripheral vasoconstriction. The pressor response to stress is, however,
complex and the reported role cortisol plays in regulating this response [Harbuz and
Lightman 1992, Chrousos 1998, Sapolsky et al 2000, Habib et al 2001, Tsigos and
Chrousos 2002] varies in the literature depending in the nature of the stressor, the
individual within whom it is occurring and the methods used to measure the response
[Kvetnansky et al 1995, Sapolsky et al 2000]. Traditionally, glucocorticoids enhance
the cardiovascular response to stress by increasing blood pressure and cardiac output
[Sapolsky et al 2000]. Similarly, glucocorticoid deficiency is associated with
hypotension and impaired cardiac function, particularly under conditions of stress
where, if the increased cardiovascular demands are not met, an acute adrenocortical
crisis may occur that could be rapidly fatal [Stewart 2003].
The interaction between the HPA and SAM axes, as detailed previously, occurs on
several levels within the brain and the periphery. CRH neurones are known to project
to brainstem noradrenergic centres where they increase activity of sympathetic
autonomic centres [Jezova et al 1999, Habib et al 2001, Gammatopoulos and Chrousos
2002]. Cortisol itself has direct actions on cardiac and vascular tissues [Habib et al
2001], although, its most important function in enhancing the cardiovascular responses
181
to stress stems from its role in augmenting the activity of catecholamines and other
vasoconstrictors [Habib et al 2001, Stewart 2003].
Glucocorticoids produced in the adrenal cortex are transported via a portal capillary
route to the adrenal medulla where they are required for the survival and maintenance of
adrenomedullary chromaffin cells [Kvetnansky et al 1995, Habib et al 2001,
Zuckerman-Levin et al 2001, Stewart 2003]. In addition, expression of the enzyme
phenylethanolamine N-methyltransferase (PNMT), required for the conversion of
adrenomedullary noradrenaline to adrenaline, depends on glucocorticoids at the
transcriptional level [Kvetnansky et al 1995, Sapolsky et al 2000, Zuckerman-Levin et
al 2001, Stewart 2003]. Glucocorticoids also enhance catecholamine action by
increasing α-adrenergic receptor binding sensitivity and catecholamine-induced cyclic
AMP synthesis [Sapolsky et al 2000]. Finally, glucocorticoids also inhibit the
vasodilatory effects of prostaglandins [Sapolsky et al 2000].
On the other hand others have shown that adrenocortical deficiency (for example
Addison’s disease) is associated with compensatory noradrenaline overproduction with
an increase in the noradrenaline to adrenaline ratio and increased cardiovascular
responses to stress [Bornstein et al 1995, Kvetnansky et al 1995]. Further, using a
model of immobilisation stress Kvetnansky et al [1995] describe the predominant acute
effects of glucocorticoids on the SAM and sympathoneural response to stress as reduced
catecholamine release and turnover with suppression of sympathoadrenal activity. In
accordance with the theory of Munck et al [Munck and Naray-Fejes-Toth 1994,
Sapolsky et al 2000], these authors imply that glucocorticoids act to restrain and modify
the stress response to prevent the damaging effects of excessive exposure to stress
mediators. Further, they ensure the response is kept appropriate to the intensity and
182
duration of the stress being faced [Munck and Naray-Fejes-Toth 1994, Sapolsky et al
2000]. More than likely, the effect of glucocorticoids and the HPA axis on SAM and
sympathoneural responses to stress is likely to vary (enhanced in some cases, inhibited
in others) depending several factors. These may include: the nature of the
glucocorticoid abnormality (central vs peripheral deficiency or excess, or exogenous
exposure); the duration and severity of the hormone alteration; associated changes such
as mineralocorticoid deficiency; the specific characteristics of the stressors involved;
and the specific characteristics of the individual in whom they are occurring.
A study of isolated glucocorticoid deficiency due to inherited ACTH resistance in a
family of six individuals was studied by Zuckerman-Levin et al [2001] with the
hypothesis that this condition will be associated with impaired adrenaline production,
compensatory noradrenaline release and physiological changes reflecting these
alterations. Subjects were examined at rest and following three stressors – upright
posture, the cold pressor test and exercise. Results showed significantly reduced
adrenaline levels compared with matched controls with slightly higher noradrenaline
levels and a marked increase in the noradrenaline:adrenaline ratio under all conditions.
With this, the pulse rate response to upright posture in patients was increased (an effect
of noradrenaline), whilst the systolic blood pressure response to cold was impaired
(probably due to insufficient adrenaline). Finally the diastolic blood pressure response
to exercise was increased in the patients, also thought to be secondary to the effect of
the increased noradrenaline:adrenaline ratio [Zuckerman-Levin et al 2001].
Experiments performed in this thesis examining the effect of changes to the HPA axis
have not identified a specific effect of cortisol on the cardiovascular response to CO2.
Disinhibiting the axis with naltrexone raised baseline cortisol levels but did not affect
183
SBP or HR responses. Similarly lactating women who have lower baseline cortisol
levels also had equivalent responses to controls. In addition the administration of
metyrapone (to decrease cortisol synthesis) with or without mineralocorticoid inhibition
(with spironolactone) had no effect on the cardiovascular response to CO2. It is
possible, however, that in all these conditions, particularly those of glucocorticoid
deficiency, there was still enough circulating cortisol to facilitate the necessary SAM
and sympathoneural responses associated with this challenge. To further elucidate the
importance of glucocorticoids in the CO2 response, and to assess any clinical role
cortisol deficiency may have in modulating the SAM and sympathoneural response to
stress, the CO2 challenge was performed in patients with Addison’s disease (primary
adrenal failure, see below). The hypothesis being that combined glucocorticoid and
mineralocorticoid deficiency will result in impaired cardiovascular responses to the
challenge and potentially increased emotional arousal through impaired feedback and
increased CRH.
184
5.2. Addison’s disease
5.2.1. Introduction and methods
In western countries, primary adrenocortical failure, or Addison’s disease, is most
commonly due to autoimmune adrenalitis [Stewart 2003]. In this condition, adrenal
antibodies cause destruction of the adrenal cortex but leave the adrenal medulla
unaffected. In contrast to secondary (pituitary) adrenocortical insufficiency (or ACTH
resistance as in the study by Zuckerman-Levin et al [2001]), Addison’s disease is
associated with both glucocorticoid and mineralocorticoid deficiency and clinical
symptoms are due to their combined lack. Adrenal androgens are also deficient and
some clinical symptoms may be due to their deficiency [Stewart 2003]. The clinical
presentation is usually of an insidious onset with malaise, fatigue, anorexia, nausea,
vomiting, abdominal pain and symptoms of postural hypotension. Skin pigmentation
due to ACTH stimulation of the melanocortin-2 receptor distinguishes primary from
secondary adrenocortical failure. Less commonly, long standing Addison’s disease may
be associated with overt psychiatric symptoms including memory impairment,
depression or anorexia nervosa [Stewart 2003]. Acute adrenal insufficiency or
addisonian crisis is an important, life threatening state that usually occurs during times
of intercurrent illness or stress. It is characterised by hypotension and shock out of
proportion to the current illness and is often associated with nausea, vomiting, acute
abdominal pain, hypoglycaemia, fever and electrolyte disturbances [Stewart 2003].
Addison’s disease is treated with the replacement of both glucocorticoids and
mineralocorticoids. Glucocorticoids are given in divided doses (less in the afternoon) to
try and mimic the normal diurnal rhythm of cortisol. Additional glucocorticoids need to
185
be given during times of illness, accident, surgery or other significant (including
mental) stress. Mineralocorticoids are given once daily to prevent postural hypotension
and electrolyte imbalance. These traditional replacement strategies have been
considered sufficient to allow patients with Addison’s disease to live normally,
however, many patients with this condition continue to complain of persistent fatigue,
malaise, poor exercise tolerance, a reduced ability to perform normally at work and
reduced stress tolerance [Riedel et al 1993, Arlt et al 1999, Lovas et al 2002]. Lovas et
al [2002] used several fatigue, general health and vitality perception scales to compare a
population of patients with Addison’s disease to the general population and identified
the Addison’s group as being quite heterogenous in their responses. Whilst many
patients had normal responses, there was a substantial group that described markedly
reduced subjective feelings of general health and working ability.
Adrenal androgens are not commonly replaced in clinical practice, although there is
some evidence that their administration may improve subjective health status and
sexuality, particularly in females [Riedel et al 1993, Arlt et al 1999, Stewart 2003]. On
the whole, however, they have failed to significantly improve the feelings of fatigue and
impaired health held by many patients [Riedel et al 1993, Hunt et al 2000, Lovas et al
2002]. It is not clear why these subjective feelings of fatigue and ill-health are so
prominent in many patients, or why they often feel less able to manage mild day-to-day
stress. Refinements and modifications of glucocorticoid replacement schedules
(including both the type of glucocorticoid used and the timing of its administration) are
used extensively to try and improve feelings of fatigue and well-being in Addison’s
patients. This is based on the premise that fixed twice or three times a day dosing is
insufficient to account for the normal day-to-day fluctuations caused by mild to
moderate stress normally experienced by all individuals. One principal objective of
186
studying patients with Addison’s disease was to assess whether their response to a mild
stressor (CO2 exposure) varied according to the state of hormone replacement and
whether this could provide any insights into the potential reasons for their general
feelings of ill-health.
It has also been suggested that Addison’s disease is associated with an increased
prevalence of affective and anxiety symptoms [Musselman and Nemeroff 1996],
although in normally replaced individuals the literature supporting an increased
prevalence of anxiety disorders is scant.
Five (2 male) subjects with known Addison’s disease of at least 5 years duration were
recruited from a pool of Addison’s disease sufferers who attend the Endocrinology
clinic at the Bristol Royal Infirmary. The subjects had a mean age of 49.8 years (range
44–55 years) and were currently well controlled on glucocorticoid and
mineralocorticoid replacement. Three subjects also had autoimmune primary
hypothyroidism and all were clinically and biochemically euthyroid on thyroxine
replacement at the time of testing. Subjects were free of all other new medications in
the two weeks prior to testing as well as during the testing weeks. During the testing
weeks, all subjects remained well with no evidence of intercurrent illnesses. In a
randomised, placebo controlled, double blind fashion, subjects attended for testing on
two occasions one to three weeks apart. For 48 hours prior to one visit subjects
received hydrocortisone 20 mg and fludrocortisone 100 mcg in the morning with a
further 10 mg hydrocortisone in the afternoon instead of their usual glucocorticoid and
mineralocorticoid replacement. For the other visit, subjects received an equal number
of matched placebo containing capsules taken at the same times.
187
All the tests were conducted between 1 and 3 pm and all subjects completed all the
tests. The tests were performed using the same protocol as described for previous
experiments and included the recording of baseline and test VC. Cardiovascular
measures were recorded with the Dynamap monitor recording pulse rate and blood
pressure every minute for 5 minutes before and 5 minutes after CO2 exposure.
Psychological responses were recorded using the same visual analogue scales. An
intravenous line for blood sampling was placed 30 minutes before testing commenced
and samples for prolactin levels were taken at baseline, 10, 20 and 30 minutes after
exposure. Cortisol levels are not reported because of the use of exogenous
glucocorticoids as part of their replacement therapy.
Repeated measures analysis of variance was used to determine the effects of CO2
exposure with paired t-test analysis used to determine differences between single time
points including baseline measures for the placebo and active replacement visits for
each subject.
5.2.2. Results
Baseline prolactin levels were not significantly different between replacement and
placebo visits. CO2 stimulated increases in prolactin were small and did not reach
statistical significance in either condition. As shown in Table 5.1, compared with the
healthy volunteers who participated in the mineralocorticoid/glucocorticoid study,
baseline SBP and HR were non-significantly higher in the subjects with Addison’s
disease, although these subjects were also older (49.8 compared with 22.0 years). The
bradycardic response under both Addison’s replacement conditions was significant (p <
188
0.05), however, the fall was the same on both occasions (-9.0 +/- 3.4 bpm on
replacement; -10.6 +/- 3.3 bpm off replacement). There was a significant (p < 0.05)
pressor response for the visit when subjects were taking replacement hormones (+15.6
+/- 5.0 mmHg), but no significant response off replacement (+4.2 +/- 3.3 mmHg; p =
0.02) (Figure 5.01). As a whole, anxiety, fear, breathlessness and selected somatic
symptom increases were significant (p <0 .05), however, baseline and stimulated
anxiety, fear and breathlessness responses were no different between each visit.
Similarly, somatic symptom responses between each visit were equivalent (Figure
5.02).
189
Table 5.01. Baseline and stimulated SBP and HR responses in patients with Addison’s Disease both off and on standard replacement therapy (n=5). For comparison, responses in young healthy volunteers taking placebo, spironolactone, metyrapone or both spironolactone and metyrapone are presented (n=9).
Addison’s
On
Addison’s
Off
Control Spironolact-
one
Metyrapone Spironolact-
one /
Metyrapone
Baseline SBP
118.2+/-6.9
114.6+/4.5
110.3+/2.6
110.7+/2.9
108.4+/-2.8
109.6+/-2.1
∆max SBP 15.6+/-5.0 1.0+/-4.4* 10.2+/-3.4 13.8+/-2.7 11.8+/-2.9 11.5+/-3.1
HR 72.0+/-3.2 71.0+/-4.3 68.0+/-3.5 65.7+/-3.5 68+/-3.7 65.8+/-2.9
∆max HR -9.0+/-3.4 -10.6+/-3.3 -14.7+/-2.4 -9.1+/-3.2 -18.1+/-4.2 -16.0+/-3.5
190
SBP
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
105110115120125130135140145
On replacementOff replacement
Time (mins)
mm
Hg
SBP response to 35% CO2
On replacement Off replacement
-10
0
10
20
30
40
*
mm
Hg
PR
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
55
60
65
70
75
80On replacementOff replacement
Time (mins)
b/m
PR response to 35% CO2
On replacement Off replacement
-30
-20
-10
0
b/m
Figure 5.01. SBP (upper panel) and HR (lower panel) in response to a single breath of 35% CO2 in patients with Addison’s disease on and off standard gluciocorticoid and mineralocorticoid replacement. Left hand panels show responses over time and right hand panels show peak stimulated response. n=5. *p=0.02
191
B a s e lin e fe e lin g s o f a n x ie ty ,fe a r a n d b re a th le s s
Anxi
ety
Fear
Bre
athl
essn
ess
0
2 5
5 0
7 5On rep lac em entOff rep lac em ent
mm
An x ie ty , fe a r a n db re a th le s s n e s s re s p o n s e to
3 5 % C O 2
Anx
iety
Fear
Bre
athl
essn
ess
0
2 5
5 0
7 5
1 0 0On rep lac em entOff rep lac em ent
mm
S o m a tic s y m p to m re s p o n s e to3 5 % C O 2
Hea
rt ra
cing Hot
Blu
rred
vis
ion
Diz
zine
ss
0
2 5
5 0
7 5On rep lac em entOff rep lac em en t
mm
192
Figure 5.02. Baseline (upper panel) and stimulated (middle panel) anxiety, fear and breathlessness responses to a single breath of 35% CO2 in patients with Addison’s disease on and off standard gluciocorticoid and mineralocorticoid replacement. Lower panel indicates somatic symptom responses. n=5. Baseline and stimulated responses between groups are all equivalent.
5.2.3. Discussion
In this study using a single breath of 35% CO2 as a challenge, no difference in baseline
subjective anxiety or fear was identified in subjects on full replacement compared with
no replacement for 48 hours. Anxiety and fear scores were also equivalent to normal
healthy volunteers and to volunteers treated with either or both spironolactone and
metyrapone. Similarly, subjective anxiety and fear responses were equivalent in the
replaced and placebo conditions and was also no different to healthy controls or the
spironolactone and metyrapone conditions. This would suggest that baseline HPA axis
function does not affect psychological responding to the 35% CO2 challenge.
Patients with Addison’s disease on full glucocorticoid and mineralocorticoid
replacement were able to mount the typical pressor response to the 35% CO2 challenge
that was similar in magnitude to young healthy controls. They also showed the typical
bradycardic response. Without replacement, the pressor response was significantly
impaired, although the bradycardia was not significantly different to the replacement
condition or to controls. There are several potential mechanisms that would explain this
observation. Firstly, the renin-angiotensin system is an important mediator of the
response to cardiovascular stressors such as postural change and volume loss [Van de
Kar and Blair 1999]. Angiotensin II is an important vasoconstrictor and also acts to
augment SAM activity both centrally and peripherally [Reid 1992, Van de Kar and
Blair 1999]. Deficiency of angiotensin II or inhibition of its receptor centrally is
associated with reduced SAM and hormonal stress responses and a reduction in blood
pressure [Armando et al 2003]. Aldosterone, the principal mineralocorticoid secreted
by the adrenal cortex in response to renin-angiotensin stimulation, acts to increase
sodium and water reabsorption in the kidney, thereby increasing plasma volume and
193
blood pressure [Stewart 2003]. Deficiency of aldosterone is the principal reason for
symptoms of postural hypotension in untreated Addison’s disease and may explain the
reduced pressor response observed. Of note, in isolated glucocorticoid deficiency
[Zuckerman-Levin et al 2001] with an intact renin-aldosterone system, the
haemodynamic challenge of postural change was associated with a normal blood
pressure response suggesting glucocorticoids are not involved in this response. In this
same study [Zuckerman-Levin et al 2001], the pressor response to cold was impaired
suggesting an impaired SAM response to this challenge as a consequence of
glucocorticoid deficiency. This may represent an alternative explanation for the
impaired pressor response to CO2. As previously discussed, glucocorticoids play an
important part in enhancing the SAM response to stress both at a central (brainstem)
and peripheral level including maintaining adrenal medullary chromaffin cells and the
production of adrenal catecholamines [Bornstein et al 1995, Sapolsky et al 2000,
Laborie et al 2003]. Glucocorticoid deficiency may then result in impaired adrenaline
mediated responses to this challenge. CO2 challenge studies described earlier indicated
the pressor response is predominantly due to a noradrenaline mediated increase in total
peripheral resistance, however, these studies have not excluded the potential
contribution of adrenal catecholamine to this response. Importantly, whilst samples for
adrenaline and noradrenaline were taken and measured, inconsistencies in and problems
with the HPLC assay made their interpretation unreliable and results have not been
reported.
Earlier studies of this challenge had suggested the bradycardic response was mediated
independently of the pressor response most likely through direct stimulation of vagal
centres by CO2 itself. The bradycardia observed in patients with Addison’s was the
same regardless of replacement status and was similarly no different to healthy
194
volunteer controls and to those taking either spironolactone, metyrapone or both. This
would again reinforce the likelihood that the bradycardia is occurring independently of
the HPA axis and also the pathways involved in mediating the pressor response.
Prolactin release is also seen following stress exposure including haemodynamic
challenges [Van de Kar and Blair 1999], although its role is uncertain. In patients with
Addison’s disease, prolactin responses to 35% CO2 were not affected by hormone
replacement status, suggesting prolactin release occurs independently of HPA axis
activity.
195
5.3. The role of the autonomic nervous system
The principal effector arms of the stress response (the HPA and SAM axes) are
interconnected at several levels [Chrousos 1998, Bornstein and Chrousos 1999, Habib et
al 2001, Wurtman 2002]. In addition to components of the HPA axis being important in
the maintenance, regulation and enhancement of the SAM response to stress (as detailed
above), the SAM system plays an essential role in the regulation of the HPA response to
specific stressors. Projections of noradrenergic neurones from brainstem autonomic
centres to the hypothalamic PVN significantly influence CRH and HPA activity
[Kvetnansky et al 1995, Jezova et al 1999, Koob 1999]. Stimulation of these neurones
or administration of noradrenaline directly into the area of the PVN significantly
increases CRH mRNA production, CRH synthesis and release with a dose dependent
increase in ACTH and glucocorticoids [Itoi et al 1994, Kvetnansky et al 1995, Pacak et
al 1995]. Inhibition of these neurones results in the opposite effect [Kvetnansky et al
1995]. Importantly, however, significant stressor specificity exists such that the
presence of an HPA and SAM response as well as the intensity and duration of the
response will vary according to the nature of the stressor applied [Kvetnansky et al
1995, Pacak et al 1995]. Further, there is evidence to suggest that stressor-specific
activation of noradrenergic brainstem systems is an important contributor to the
differential regulation of HPA and SAM responses to specific stressors [Kvetnansky et
al 1995, Pacak et al 1995, Pacak et al 1998].
Activation of the SAM/sympathoneural system in response to stress exposure is
responsible for the early ‘fight or flight’ response. This response comprises immediate
stereotypical behaviour (such as freezing or fleeing) and activation of multiple
physiological processes designed to release energy stores, heighten alertness, increase
196
motor activity, protect against injury and inhibit non-essential functions [McCarty 1994,
Young and Landsberg 1998]. The peripheral components of this system, the
sympathetic and parasympathetic nervous systems, are essentially antagonistic, but act
in a functionally synergistic fashion to enhance their control. Activation of the
sympathetic system, for example, is usually accompanied by withdrawal of the
opposing parasympathetic inputs [Gilbey and Spyer 1993, Janig 1983, Elenkov et al
2000].
Studies of cardiorespiratory regulation in reptiles and mammals has led to the
identification of opposing vagally mediated responses and the subsequent description of
the likely way in which vagal pathways mediate heart rate responses to novelty and
various other stressors [Porges 1995]. This ‘Polyvagal Theory’ identifies two
anatomically and functionally discrete vagal systems [Porges 1995, Porges et al 2003].
The first, a phylogenetically more primitive system, originates from the Dorsal Motor
Nucleus of the vagus (DMNX) and is characterised by unmyelinated fibres that produce
heart rate slowing and neurogenic bradycardia. The second, an evolutionarily more
recent system, comprises myelinated fibres originating from the Nucleus Ambiguus.
This system, is responsible for producing respiratory sinus variation through tonic
inhibition of the sympathetic nervous system and therefore acting as a ‘brake’ on the
energy expensive and metabolically demanding ‘fight-or-flight’ sympathetic system
[Porges 1995, Porges et al 2003]. In addition to mediating heart rate responses, these
systems are also intimately involved with regulating the stereotypical emotional and
behavioural responses to stress through their relationship with other cranial nerves (for
example those involved with facial expression and vocalisation [Porges et al 2003]), as
well as other brainstem and higher centres [Porges 1995]. The theory postulates that
during times of stress, control of autonomic function shifts from the phylogenetically
197
newer to the older and more primitive systems [Porges et al 2003]. In other words,
when faced with a challenge, withdrawal of the nucleus ambiguus system would allow
increased activity of the sympathetic system necessary for meeting the metabolic
demands of that challenge, but would make the individual more susceptible to
stimulation of the DMNX system with significant bradycardia and even syncope. This
theory has thus been used to explain vasovagal syncope associated with fright [Porges
1995, Porges et al 2003], and with the heart rate changes observed in response to
hypoxia during foetal heart rate monitoring [Reed et al 1999].
The brainstem cardiorespiratory centres, particularly the more primitive centres, are
much more sensitive to CO2 than O2 [West 1974, Porges 1995], and it is possible that
the 35% CO2 challenge is directly stimulating the DMNX system thereby producing the
significant bradycardia. Syncope is prevented by sympathoneural activation with
peripheral vasoconstriction, increased total peripheral resistance and the subsequent
pressor response.
198
5.4. Experimental plan
Three clinical models of autonomic dysfunction have been studied in order to further
clarify some of the mechanisms that may regulate the cardiovascular and HPA
responses to 35% CO2, and to also evaluate the potential clinical relevance of an altered
response in these individuals. The groups chosen included: patients with diabetic
autonomic neuropathy (DAN) (a peripheral neuropathy with predominant
parasympathetic dysfunction); patients with pure autonomic failure (PAF) (a model of a
purely peripheral mixed sympathetic and parasympathetic neuropathy); and patients
with multiple systems atrophy (MSA) (representing a model of central autonomic
failure). Based on earlier experiments, it was hypothesised that peripheral
parasympathetic failure (DAN and PAF) would result in loss of the bradycardic
response but the pressor response would be preserved if sympathetic function was not
affected (that is, preserved in early DAN but not in PAF). Central autonomic failure
(MSA) would be associated with a loss of the pressor response but not of the
bradycardia. HPA and psychological responses would be preserved in peripheral (DAN
and PAF) but not central (MSA) autonomic failure syndromes. Finally an attempt was
made to determine whether the responses were sufficiently discrete and predictable
enough (based on the pathophysiology of the underlying condition), to use the CO2 test
clinically when investigating these conditions.
199
5.5. Diabetic autonomic neuropathy
5.5.1. Introduction and methods
Autonomic neuropathy is an important chronic complication of diabetes producing a
variety of clinical features and physiological changes related to the widespread
involvement of both the sympathetic and parasympathetic divisions of the autonomic
nervous system [Ward 1992, Forst et al 1996]. Damage to peripheral nerves is related
to glycaemic control and the frequency of clinically apparent neuropathy is related to
both the duration of diabetes and the adequacy of control [Ward 1992]. The exact
pathological mechanism is, however, unknown although popular theories include nerve
damage secondary to sorbitol accumulation or from ischaemia following microvascular
occlusion [Ward 1992]. The prevalence of autonomic neuropathy is highly variable
ranging from 0% - 93% depending on the diagnostic criteria used and the population
studied [Kempler et al 2002]. More recently, studies in type 1 diabetics have suggested
prevalence rates of between 20 and 36% [Stella et al 2000, Kempler et al 2002]. Many
patients with diabetes will demonstrate abnormalities of autonomic function on specific
testing without having any clinical symptoms [Ward 1992], however, once established
autonomic neuropathy is associated with a significant deterioration in quality of life and
increased morbidity and mortality [O'Brien et al 1991, Ward 1992, Ziegler 2001,
Wheeler et al 2002, Ziegler 2002]. The manifestations of autonomic neuropathy are
diverse involving numerous organ systems [Ziegler 2001]. Most commonly this
involves:
i. The cardiovascular system with resting tachycardia, reduced heart rate
variability, impaired heart rate and blood pressure responses to exercise and
200
postural changes, silent myocardial ischaemia and possibly an increased risk
of increased arrhythmias and sudden death.
ii. The respiratory system with impaired ventilatory responses to hypoxia and
hypercapnia [Tantucci et al 2001], and sleep apnoea.
iii. The gastrointestinal system with oesophageal dysmotility, gastroparesis,
diarrhoea and anorectal dysfunction.
iv. The genitourinary tract with bladder and erectile dysfunction.
Several studies have demonstrated an increased risk of premature death associated with
the presence of diabetic autonomic neuropathy [O'Brien et al 1991, Wheeler et al 2002,
Ziegler 2002]. O’Brien et al [1991], for example, demonstrated a 5-year mortality rate
that was more than 5-fold greater in type 1 diabetic patients with autonomic neuropathy
compared with those without neuropathy. The mechanism behind this increased
mortality is unclear, however, most hypotheses have implicated the increased
arrhythmogenic potential of cardiac autonomic neuropathy [Wheeler et al 2002]. The
parasympathetic nervous system (PNS) is typically involved earlier in diabetic
autonomic neuropathy [Edmonds and Watkins 1992]. This system provides tonic
inhibitory control over the heart, and the bradycardic response seen in the normal
response to the 35% CO2 challenge may be mediated by direct stimulation of the region
of the DMNX [Porges 1995]. In addition, withdrawal of the nucleus ambiguus
component of the vagal response allows sympathetic stimulation to predominate. It has
been postulated the withdrawal of the nucleus ambiguus component leaves the
individual susceptible to the effects of excessive neurogenic bradycardia if sympathetic
compensation is inadequate. However, it could also be postulated that parasympathetic
neuropathy removes both vagal components that would normally act to protect the
individual from excessive sympathetic stimulation increasing the risk of cardiac
201
tachyarrhythmias and resulting in an increased mortality associated with diabetic
autonomic neuropathy [Porges 1995, Reed et al 1999, Wheeler et al 2002, Porges et al
2003]. However, not all studies have shown an increase in cardiovascular (as opposed
to all-cause) mortality or an increased prevalence of non-fatal arrhythmias [Wheeler et
al 2002]. It has also been noted that most deaths appear to occur during times of acute
stress or during sleep [Tantucci et al 2001] and several studies have examined the
integrity of components of the stress response following various challenges.
In view of its immediate clinical relevance, the most commonly evaluated challenge has
been that of hypoglycaemia. A recent study by Meyer et al [1998] in longstanding
diabetics demonstrated impaired counter regulatory hormone production including
noradrenaline, adrenaline, cortisol and GH in all subjects, but catecholamine responses
were more severely impaired in those with autonomic neuropathy compared with those
without. Other studies that have incorporated such challenges as the valsalva
manoeuvre, postural changes, the cold pressor test and exercise [Forst et al 1996,
Granados et al 2000] have all indicated autonomic neuropathy is associated with low
baseline noradrenaline levels and impaired noradrenaline responses to stress. Both
baseline and stimulated adrenaline levels on the other hand, were more variable. Forst
et al [1996] examined catecholamine responses to a psychological stressor and
identified normal stimulated noradrenaline responses, but impaired adrenaline responses
in those with diabetic autonomic neuropathy compared with those without.
Haemodynamic differences, however, were not were not observed. Most studies
conclude that there is evidence of impaired adaptive responses to stress particularly
involving the SAM and sympathoneural systems in patients with diabetic autonomic
neuropathy, and that this may be important in the observed increased mortality
202
associated with this particular diabetic complication [Forst et al 1996, Meyer et al 1998,
Granados et al 2000, Wheeler et al 2002].
Another facet that has been investigated in subjects with diabetic autonomic neuropathy
in order to further understand their increased mortality has been the regulation of
breathing in these subjects. Several studies have indicated impaired peripheral
chemosensitivity and reduced hypoxic drive [Nishimura et al 1989, Tantucci et al
2001], although central chemosensitivity and responses to CO2 have been conflicting.
Whilst some studies have shown normal or increased ventilatory and haemodynamic
responses to hypercapnia [Nishimura et al 1989, Tantucci et al 2001], others have
shown impaired responses [Williams et al 1984, Tantucci et al 2001]. More recently,
Tantucci et al [2001] have demonstrated altered chemosensitivity to hypercapnia
depending on the extent and severity of the underlying autonomic dysfunction. Subjects
with autonomic neuropathy that included postural hypotension (reflecting both
sympathetic and parasympathetic dysfunction) were shown to have an increased
ventilatory and cerebrovascular reactivity responses to progressive hypercapnia, as
compared with those with autonomic neuropathy without postural hypotension
(parasympathetic dysfunction alone). This suggested central autonomic function was
impaired as a result of peripheral dysfunction of the autonomic nervous system and the
extent of this impairment reflected the balance between the activities of the sympathetic
and parasympathetic systems [Tantucci et al 2001].
Male subjects aged between 18 and 70 years with diabetes of at least 3 years duration
were recruited from the Bristol Royal Infirmary’s diabetes outpatient clinic. The first
12 subjects were recruited on the basis of a high clinical likelihood of the presence of
autonomic neuropathy (including the presence of peripheral neuropathy and/or
203
suggestive symptoms). No subject had previously had formal autonomic function
studies performed. These first subjects underwent detailed autonomic functioning
testing (as described below) followed immediately by a 35% CO2 challenge study using
the same protocol as already described. Autonomic function tests were not analysed
until after the completion of the CO2 test. Of the twelve, ten were identified as having
autonomic neuropathy (AN), whilst the remaining two were not and were included in
the control group (C). Subsequently, a further eight consecutive subjects were
recruited. Of these, one was identified with autonomic neuropathy and included in that
group and data from one was excluded because of an inadequate intake of CO2 during
the challenge (VC < 80% of baseline). As a result, there were 11 subjects with
autonomic neuropathy and 8 control subjects. As shown in Table 5.02, those with
autonomic neuropathy were significantly older (60.0 +/- 2.4 compared with 40.6 +/-
5.23 years, p = 0.002) and had had diabetes for longer (25.9 +/-4.0 compared with 12.4
+/- 3.6 years, p = 0.03). In addition, those with autonomic neuropathy were more likely
to have other microvascular complications of diabetes, and were more likely to be on
anti-hypertensive medication. The number of subjects with type 1 compared with type
2 diabetes within each group was not significantly different.
Standard exclusion criteria for the CO2 test as previously described were applied in this
study. Specifically, subjects were excluded if they had uncontrolled diabetes (defined
as an HbA1c ≥ 10%), uncontrolled hypertension, a history of recent angina or ischaemic
heart disease or any history of cerebrovascular disease. Other exclusion criteria
included age over 70 years, a history of asthma or a history of panic disorder or severe
anxiety. Patients were allowed to remain on their usual medications as long as there
had been no recent (within 8 weeks) change and as long as they were not taking
exogenous glucocorticoids, anti-depressants, anti-psychotics or anti-anxiety agents, β-2
204
agonists or centrally acting anti-hypertensives. Other anti-hypertensives were allowed,
and one subject (in the autonomic neuropathy group) was taking a β-blocker. Subjects
were asked to avoid consuming cigarettes, alcohol and caffeine for 24 hours prior to
testing.
CO2 testing was performed according to the same protocol as described for the
experiment involving subjects with Addison's disease and used the same cardiovascular
monitoring techniques, visual analogue scales, and blood sampling procedures.
Methods of statistical analysis were also the similar using repeated measures ANOVA
and unpaired t-test analysis.
On the day of testing, all subjects were asked to attend between 11 am and midday and
had their blood glucose level measured on arrival. Testing was only undertaken if blood
glucose levels were between 5 to 12 mmol/l. An intravenous line was placed in an
antecubital vein as before and subjects where asked to rest quietly on a bed prior to
autonomic function studies being performed. A 12-lead ECG was recorded and
autonomic function studies performed according to the method of Ewing and Clark
[1982]. This procedure involved determining the R-R interval from lead II of the ECG
and classifying subjects according to the following [Granados et al 2000]:
1. The supine position test determines the variation in sequential R-R intervals
after a one-minute rest (normal, 1.4 or more; borderline, 1.1 –1.3; abnormal,
1.0 or less).
2. The expiration-to-inspiration ratio (E/I) with the subject breathing at
maximal vital capacity with a regular rate of six breaths per minute. During
each expiration-inspiration cycle, the ratio of the longest to the shortest R-R
interval is calculated and the mean of six ratios taken as the E/I ratio
205
(normal, 15 beats/min or more; borderline, 11-14 beats/min; abnormal, 10
beats/min or less).
3. The Valsalva index is calculated from the longest R-R interval following a
Valsalva manoeuvre divided by the shortest R-R interval during the
manoeuvre (normal, 1.21 or more; abnormal, 1.20 or less).
4. Postural hypotension was determined by comparing supine SBP after a 5
minute rest with SBP after 1 minute in the standing position (normal, 10
mmHg or less; borderline, 11-29 mmHg; abnormal 30 mmHg or more).
Using these criteria, subjects were classified as having no autonomic neuropathy (all
four tests normal) or as definite autonomic neuropathy (first 3 tests all abnormal) with
or without postural hypotension (depending on the result of the fourth test). As shown
in Table 5.02, 3 (27%) of the 11 subjects with autonomic neuropathy also had postural
hypotension.
5.5.2. Results
Baseline cortisol and prolactin levels were equivalent in both groups (Figure 5.03). In
response to a single breath of 35% CO2 there was a significant increase in cortisol in
both groups (p < 0.05) with no difference in the maximum cortisol change between the
control (C) and autonomic neuropathy (AN) groups (Figure 5.03). The prolactin
response to 35% CO2 did not reach statistical significance in either group.
As shown in Table 5.02 and Figure 5.04, baseline blood pressure in the AN and C
groups was similar (135/74 and 137/70 mmHg, respectively) and whilst heart rates were
206
consistently higher in the AN group compared with the C group at baseline (76 +/- 3.5
and 69 +/- 3.2 b/m, respectively) and at all time points during the study, these
differences did not reach statistical significance. In response to 35% CO2, there was a
significant pressor response in both groups (+13.2 +/- 1.7 mmHg in the C and +12.3 +/-
3.7 mmHg in the groups respectively). However, whilst the C group demonstrated a
significant bradycardic response similar to that seen in normal healthy volunteers (-10.5
+/- 1.9 bpm), no subject with AN demonstrated a bradycardic response. The mean
maximum change in heart rate in the AN group was +2.7 +/- 1.3 bpm (p < 0.0001
compared with the C group) (Figure 5.04).
Baseline anxiety, fear and somatic symptoms of anxiety were minimal in all subjects
and did not differ between groups. In response to 35% CO2 there was a significant
increase in subjective feelings of anxiety, fear, breathlessness and in some specific
somatic sensations (p < 0.05). There was no difference in the change in these feelings
between the two groups (Figure 5.05) apart from a trend towards less breathlessness in
the AN group.
207
Table 5.02. Baseline characteristics of diabetic subjects with (n=11) and without (n=8) autonomic neuropathy
Control Autonomic
Neuropathy
p
Age (years)
40.6 +/- 5.3
60.0 +/- 2.4
0.002
Age range 24 – 36 40-69
Diabetes type
Type 1 75% 55%
Type 2 25% 45%
Duration of diabetes (years) 12.4 +/- 3.6 25.9 +/- 4.0 0.03
Range of duration 3 – 36 3 – 43
Complications
Diabetic retinopathy 38% 63% <0.05
Nephropathy (including
microalbuminuria)
0% 45% <0.05
Peripheral neuropathy 25% 91% <0.05
Cardiovascular parameters
Resting blood pressure (mmHg) 137 / 70 135 / 74 ns
Postural hypotension - 27%
Resting heart rate 69.3 +/- 3.2 76.4 +/- 3.5 ns
Antihypertensive medications <0.05
Nil 87.5% 45%
ACEI alone 12.5% 18%
ACEI plus other - 27%
AngIIAntagonist - 9%
β-blocker - 9% (1 subject)
208
Baseline cortisol
Control AN
0
100
200
300
400
500
nmol
/l
Cortisol response to 35% CO2
Control AN
-100
0
100
200
300
400
nmol
/lBaseline prolactin
Control AN
0
100
200
300
nmol
/l
Prolactin response to 35% CO 2
Control AN
-100
0
100
200
nmol
/l
Figure 5.03. Baseline and stimulated cortisol (upper panel) and prolactin (lower panel) responses to a single breath of 35% CO2 in diabetic subjects with (n=11) and without (n=8) autonomic neuropathy. Differences between groups are all non-significant.
209
SBP
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
125
150
175ControlAN
Time (mins)
mm
Hg
SBP
Control AN
-10
0
10
20
30
mm
Hg
HR
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
50
60
70
80
90ControlAN
Time (mins)
bpm
HR
Control AN
-30
-20
-10
0
10
20
bpm
*
Figure 5.04. SBP (upper panel) and HR (lower panel) responses to a single breath of 35% CO2 in diabetic subjects with (n=11) and without (n=8) autonomic neuropathy. *p<0.0001
210
Psychological responses to35% CO2
Anxiety Fear Breathlessness
0
25
50
75ControlAN
mm
Somatic symptom responsesto 35% CO2
Hea
rt B
eat
Feel
ing
Hot
Blu
rred
Vis
ion
Diz
zine
ss
0
25
50
75ControlAN
mm
Figure 5.05. Subjective anxiety, fear and breathlessness (upper panel) and somatic symptom (lower panel) responses to a single breath of 35% CO2 in diabetic subjects with (n=11) and without (n=8) autonomic neuropathy. Differences between groups are all non-significant.
211
5.5.3. Discussion
The most obvious abnormality between those with autonomic neuropathy compared to
those without was the failure to observe a bradycardic response in any of the AN +
subjects. In an earlier experiment involving a subject with pulmonary denervation, the
absence of bradycardia (with a normal pressor response) implied an intact vagus was
necessary for this response, whilst earlier studies showing the bradycardia occurring
prior to the pressor response implied this is not reflex bradycardia. The observation in
the diabetic subjects with autonomic neuropathy supports the notion that this is a direct
vagally mediated response that is lost in the presence of parasympathetic dysfunction.
A maintained bradycardia with no change in blood pressure in the Addison’s patients
off treatment also supports this concept.
Furthermore, all of these subjects demonstrated a pressor response that was no different
to the control group. As mentioned, it has been consistently shown that diabetic
autonomic neuropathy is associated with reduced basal and impaired stimulated
noradrenaline responses [Forst et al 1996, Granados et al 2000]. Since the pressor
response to 35% CO2 is thought to be predominantly mediated through noradrenaline
mediated vasoconstriction, it had been postulated that the pressor response might be
impaired in this group as well. Indeed, Forst et al [1996] had indicated declining
catecholamine responses to stress with increasing severity of autonomic dysfunction
(reflecting the increasing significance of sympathetic nervous system, involvement),
although no difference in haemodynamic responses were noted even in the severe group
with the most marked catecholamine difference. As with this study, Tantucci et al
[2001] demonstrated significantly higher resting heart rates in subjects with autonomic
neuropathy, although resting blood pressure was equivalent. These authors also showed
heart rate and blood pressure responses to progressive hypercapnia (breathing 7% CO2
212
according to the Read re-breathing method) were no different in control subjects as
compared with diabetic subjects with autonomic neuropathy, including when these were
sub-divided into those with or without postural hypotension. They did, however, show
differences in cerebrovascular reactivity in controls compared with autonomic
neuropathy subjects both with and without postural hypotension. In this study, 3 of the
11 AN+ subjects had postural hypotension >30mmHg. There was no difference in the
pressor response in these three individuals as compared with either the other AN+
subjects or the control group. However, it may be that this number was too small to
identify a real difference.
Baseline and stimulated cortisol responses were equivalent in those with and without
diabetic autonomic neuropathy. This would suggest the HPA axis to this challenge
remains intact and is independent of peripheral autonomic dysfunction. Studies of
uncontrolled diabetes in rats have indicated HPA axis dysregulation with increased
baseline CRH, ACTH and cortisol levels and impaired cortisol responses to applied
stress [Williams et al 1984, Chan et al 2002]. Interestingly, this dysregulation was
normalised with insulin treatment, although similarly impaired adrenaline responses
were not affected by the same treatment [Williams et al 1984]. It is likely that subjects
with acute hyper or hypoglycaemia will have alterations in HPA axis output as a
reflection of the glucose levels at the time and that subjects with persistent poorly
controlled blood glucose levels or frequent hypoglycaemia will have altered HPA axis
activity. In this study with treated control and autonomic neuropathy subjects matched
for glycaemic control and having relatively normal glucose levels at the time of testing
the effect of glucose levels on the HPA axis was minimised.
213
Cortisol responses to graded exercise have been shown to be impaired in diabetic
subjects with autonomic neuropathy compared with those without, although only at
maximum workload [Hilsted et al 1980 , Chan et al 2003]. The mechanism of this is
unclear, although it has been suggested that this is a function of impaired sympathetic
afferent impulses [Chan et al 2003]. Exercise at maximal workload is a complex
stressor that involves other neurohormonal pathways such as GH (which is also blunted
in those with diabetic autonomic neuropathy) and the HPA axis response that is almost
certainly mediated through different pathways to the CO2 challenge [Iranmanesh et al
1990].
Reduced circulating and stimulated prolactin levels have previously been demonstrated
in poorly controlled diabetic patients and laboratory animals with induced diabetes
[Iranmanesh et al 1990, Arroba et al 2003]. The mechanism producing this is not fully
understood, however, recent evidence suggests a reduction in lactotroph number within
the pituitary is in part responsible [Arroba et al 2003]. The consequences of this change
in prolactin secretion is also not fully understood, although one group [Hartmann and
Cregan 2001] has suggested diabetic women could have delayed or reduced breast milk
production as a result. Compared with results from non-diabetic healthy volunteers,
baseline prolactin levels were no different in those with diabetes regardless of their
neuropathy state. Apart from two diabetic control subjects, none of the other diabetic
subjects demonstrated a prolactin response to the CO2 challenge. This is in contrast to
the significant prolactin secretion seen in other experiments involving healthy non-
diabetic volunteers. This would be consistent with the already mentioned literature
regarding prolactin release in diabetics and, although at the time of testing blood sugar
control was not poor, most of the subjects tested have had diabetes for many years and
may have had periods of poor control in the past.
214
Finally, emotional arousal, as measured by subjective anxiety, fear, breathlessness and
somatic symptoms of fear where no different to those seen in healthy volunteers with
autonomic neuropathy having no additional effect on anxiety or other psychological
responses. This would appear to reinforce the notion that psychological responses to
35% CO2 appear to be mediated through pathways involving central chemoreceptors
that are probably stimulated independently, but that may subsequently influence
hypothalamic (HPA and prolactin) and brainstem (SAM) centres through reciprocal
connections.
215
5.6. Chronic autonomic failure syndromes
5.6.1. Introduction and methods
The autonomic failure syndromes are numerous in origin, most commonly occurring in
association with other known pathologies that also involve the autonomic nerves. These
‘secondary’ forms of autonomic neuropathy most commonly occur in association with
diabetes and other metabolic insults, amyloidosis, inflammatory disease, malignancy
and exposure to various medications or poisons. Other associations include a number of
inherited diseases, chronic infection and the aging process [Bannister and Mathias
1992a]. Chronic ‘primary’ autonomic neuropathy (or that not occurring in association
with other known pathologies) has traditionally been subdivided into three forms:
i. PAF,
ii. MSA,
iii. Autonomic failure with Parkinson’s disease (PD) [Bannister and Mathias
1992a].
Since the autonomic nervous system innervates all organ systems in the body, the
clinical manifestations of autonomic failure are widespread and varied. Early changes
are often subtle with few clues as to the underlying disease process due to the existence
and activation of multiple compensatory reflexes [Bannister and Mathias 1992b]. As
will be detailed further, the pathophysiology of PAF involves post-ganglionic fibres and
thus represents model of failure of the peripheral autonomic nervous system. In
contrast, MSA involves pre-ganglionic systems and therefore represents central
autonomic failure.
216
Postural hypotension is the most common and most dramatic clinical feature of all three
forms of chronic autonomic failure [Bannister and Mathias 1992b]. Other clinical
features include visual disturbance, defective sweating, cardiorespiratory abnormalities
and problems with urogenital function including bowel, bladder and sexual dysfunction
[Bannister and Oppenheimer 1972, Bannister and Mathias 1992b]. Movement disorders
are also apparent in MSA and PD forms [Bannister and Mathias 1992b, Parikh et al
2002]. Patients vary considerably in the degree of involvement of various components
of the autonomic nervous system. However, early in their presentation the clinical
features are strikingly similar and it can often take several years before it is clinically
apparent which form of the disease a particular patient may have. This distinction is
critical, as the prognosis of the various forms is very different [Bannister and Mathias
1992b]. PAF is typically characterised by a relatively benign course with predominant
postural hypotension, no movement disorder and an essentially normal life expectancy
[Bannister and Mathias 1992a, Bannister and Mathias 1992b]. MSA on the other hand
runs a debilitating, progressive and distressing course with increasing involvement of
autonomic systems, progressive motor weakness and rigidity, and is invariably fatal
with a life expectancy of less than 10 years [Bannister and Mathias 1992b, Parikh et al
2002].
In recent years, much has been learnt about the pathophysiology of these forms of
autonomic failure and several new tests have been developed to aid in their diagnosis.
As yet, however, no single test exists that can accurately distinguish peripheral from
central autonomic failure and the diagnosis remains a clinical one. MSA and PAF have
a distinct pathophysiological basis, and their detailed evaluation has provided an
enormous amount of valuable information on the mechanisms of autonomic function
217
and has led to the considered introduction of new and more effective therapies
[Bannister and Mathias 1992b].
The primary site of pathological involvement in PAF is the post-ganglionic sympathetic
neuron. This condition is characterised by reduced intra-neuronal noradrenaline
synthesis as well as reduced noradrenaline re-uptake from the synaptic cleft [Ziegler et
al 1977, Parikh et al 2002] and is reflected by lower plasma noradrenaline levels
compared to both normal individuals and those with MSA or PD [Bannister and
Mathias 1992b, Parikh et al 2002]. Consistent with this, histological studies of PAF
subjects show markedly reduced post-ganglionic nerve fibres with remaining nerve
endings showing abnormalities including Lewy bodies, distorted neurites and high
concentrations of lysosomes [Bannister and Mathias 1992b, Parikh et al 2002].
Functionally they are characterised by post-synaptic receptor hypersensitivity to
noradrenaline with associated baseline vasoconstriction and an increased TPR. Whilst
sympathetic failure and reduced baroreflex buffering capacity with postural hypotension
is the most significant clinical symptom [Bannister and Mathias 1992b, Shannon et al
2000, Parikh et al 2002], a large number of patients have associated supine hypertension
that is not mediated by increases in cardiac output and only partly due to residual
sympathetic tone mediated by post-synaptic hypersensitivity [Shannon et al 2000,
Parikh et al 2002].
Parasympathetic dysfunction in PAF has also been described [Parikh et al 2002], again
demonstrating a post-ganglionic defect with post-synaptic receptor hypersensitivity to a
variety of stimuli.
In contrast to the peripheral post-ganglionic defect seen in PAF, MSA is characterised
by features suggesting a central or preganglionic abnormality. Noradrenaline levels are
218
normal but they fail to respond to stimuli that require centrally mediated interpretation
or integration [Bannister and Mathias 1992b, Shannon et al 2000, Parikh et al 2002].
Histologically, the most striking abnormality is loss of cells from the intermediolateral
column of the spinal cord as well as other sites within the brainstem [Daniel 1992,
Parikh et al 2002]. Supine hypertension is seen more often in MSA subjects and is
more clearly related to residual post-ganglionic sympathetic tone [Polinsky et al 1991,
Parikh et al 2002], and is also characterised by orthostatic tachycardia [Parikh et al
2002]. Parasympathetic dysfunction has similarly been described in MSA, although
patients with MSA are difficult to distinguish from those with PAF based on
parasympathetic changes [Parikh et al 2002].
A study of the 35% CO2 challenge was therefore undertaken in patients with either
MSA or PAF to examine the role of the autonomic nervous system in regulating the
specific response to acute CO2 exposure, and to assess its potential to distinguish
between these two forms of autonomic failure.
Male and female patients with clinically established MSA or PAF were recruited from
the neurology outpatient clinics of the National Hospital for Neurology and
Neurosurgery, Queen Square, London and St Mary’s Hospital, Praed Street, London.
Control subjects were recruited from within the Autonomic Unit or were a well family
member of recruited patients. Nine MSA, 9 PAF and 5 control subjects were recruited.
MSA and PAF were defined using established diagnostic criteria [Gilman et al 1998,
Mathias and Bannister 2002]. All patients had documented sympathetic and
parasympathetic dysfunction with severe orthostatic hypotension. No subjects were on
anti-parkinsonian medication and all vasoactive medications were withdrawn the night
prior to the study. Control subjects were healthy with no active or chronic medical
219
illnesses and had not been on any medications apart from simple anaelgesia for the 2
weeks prior to the study. Exclusion criteria for this study included age over 70 years,
uncontrolled hypertension, a history of recent angina or ischaemic heart disease or any
history of cerebrovascular disease, asthma or a history of panic disorder or severe
anxiety. Subjects were asked to avoid consuming cigarettes, alcohol and caffeine for 24
hours prior to testing.
Subjects were asked to attend once, where they all received a single breath of 35% CO2
according to the protocol described previously. The ethics committees of the United
Bristol Healthcare Trust, the National Hospital for Neurology and Neurosurgery, and St
Mary’s Hospital all approved the study and all subjects provided written informed
consent prior to participation.
The study took place in a dedicated autonomic laboratory between 10 am and 1 pm.
After emptying the urinary bladder, the subject was seated and an intravenous line
(Venflon, Viggo Spectramed, Helingsborg, Sweden) was placed in an antecubital vein.
After a 10 minute rest, cardiovascular monitoring commenced. Continuous
measurement of beat-to-beat heart rate and blood pressure was made using a Portapres
II (TNO-TPD Biomedical Instruments, Amsterdam) device on the middle finger of the
right hand. Subsequent calculation of cardiac output (CO), stroke volume (SV) and
total peripheral resistance (TPR) using Modelflow analysis (Beatscope Software)
according to previously validated methods. Modelflow method uses a three-element
model of the aortic input impedance to compute flow from the pulsation of the arterial
pressure [Wesseling et al 1993, Langewouters et al 1998].
On the left side one Laser Doppler (Perimed) probe was taped over the pulp of the index
finger and one probe over the anterior forearm. Skin blood flow was measured by laser
220
Doppler flowmetry [Perimed Periflux 5000/5010 solid state diode laser set at 780 nm
with a 1 mW maximal power output at the probe tip]. Fibre optics permit the Laser
Doppler probe to shine a laser light directly at the skin surface. Light is reflected back to
the recording element of the probe from red blood cells in the skin capillaries. Light
reflected back to the probe has its wavelength altered by the Doppler shift caused by the
movement of the red blood cells relative to the probe. This altered wavelength is then
used to calculate an arbitrary perfusion rate (“Perfusion units”; PU), rather than absolute
flow, which is then visualised as a real-time trace. Cutaneous vasoconstriction will
therefore be shown as a reduction in perfusion units relative to baseline perfusion. The
forearm probe was used as a control to exclude movement or temperature artefact, as
forearm skin blood flow is under thermoregulatory rather than the “emotional”
sympathetic regulation present in the fingertip skin [Johnson et al 1995, Saad et al
2001].
One way analysis of variance was used to determine between-group differences with
two-tailed t-tests used to compare single time-point data including baseline differences.
Cardiovascular responses for a particular time point were calculated as the mean of +/-3
beats from that point.
For skin blood flow, from the individual perfusion unit data the calculation of %
reduction in skin blood flow following inhalation of air or CO2 was calculated as shown
in Figure 5.06. This allowed intra- and inter-groups comparisons. Results (both for skin
blood flow at baseline and reduction ratios after gasp of air or CO2) were compared by
ANOVA with Newman-Keuls post-hoc testing.
221
Figure 5.06. Calculation of % reduction in Skin Blood Flow (SBF) following inhalation of air or 35% CO2
222
5.6.2. Results
Anxiety symptoms were all mild and transient and the CO2 challenge was well tolerated
by all participants with no significant adverse effects recorded. In all cases, the test
breath was considered adequate being > 80% of the measured baseline VC.
A total of 18 patients (9 with PAF and 9 with MSA) and 5 control subjects were
studied. Baseline characteristics are given in Table 5.03. The PAF group comprised 5
males and 4 females with a mean age of 66.4 +/- 2.1 years. The MSA group comprised
4 males and 5 females with a mean age of 57.1 +/- 3.1 years whilst the control group
comprised 4 males and 1 female with a mean age of 47.0 +/- 6.8 years.
Cortisol, prolactin and salivary amylase were not significantly different between the
three groups at rest. Compared with both control and MSA subjects, PAF subjects had
significantly lower baseline noradrenaline levels (126.8 +/- 19.7 pg/ml compared with
288.4 +/- 31.3 and 246.9 +/- 13.1 pg/ml respectively, p < 0.001 for both). Baseline
subjective anxiety and somatic symptom VAS ratings were also not significantly
different between the three groups.
As demonstrated in Table 5.04 and Figure 5.07, in response to a single breath of 35%
CO2, systolic blood pressure increased significantly from baseline in control (+60.2 +/-
13.9 mmHg, p = 0.01), PAF (+26.8 +/- 3.2 mmHg, p < 0.001) and MSA (+18.3 +/- 2.7
mmHg, p = 0.002) subjects. The rise in SBP was, however, significantly greater in
control subjects compared with either PAF or MSA subjects (p < 0.01 and p < 0.001,
respectively). This peak pressor response occurred 32.4 +/- 2.1 seconds following the
CO2 breath in control. In contrast, the peak pressor response was significantly delayed
223
in both the PAF (152.4 +/- 23.9 secs) and MSA subjects (140.2 +/- 35.5 secs) (p = 0.03
and p = 0.04 compared with controls respectively). Diastolic blood pressure also
increased significantly in control (+21.8 +/- 5.0 mmHg, p = 0.01), PAF (+13.7 +/- 2.0
mmHg, p = 0.002) and MSA subjects (+10.5 +/- 2.7 mmHg, p = 0.004). There was a
similar, although non-significant trend towards a smaller rise in DBP in the PAF and
MSA groups compared with controls. Bradycardia was observed following CO2
exposure in 4 of the 5 controls, although peak change from baseline was not significant
(–9.1 +/- 3.5 bpm, p = 0.06). Bradycardia was only observed in 3 of 9 PAF and 2 of 9
MSA subjects.
Corresponding to the pressor response in control subjects, there was a non-significant
trend toward a rise in both SV and TPR with no change in cardiac output (Table 5.04).
MSA subjects, similarly showed no change in CO, SV or TPR whereas PAF subjects
demonstrated a significant increase in TPR from baseline (p = 0.04) that was
significantly greater than the peak response seen in control subjects (p = 0.03). As seen
in Table 5.04 and Figure 5.08, there was a marked increase in noradrenaline levels from
baseline in control subjects (+41.7 +/- 7.1 %, p < 0.0001). This increase was
significantly greater than the noradrenaline response seen in either PAF (+4.2 +/- 2.2 %)
or MSA (+8.0 +/- 2.4 %) subjects (p < 0.0001 for both). No significant change from
baseline was seen in salivary amylase or prolactin in any of the groups. There was a
trend towards a greater increase in cortisol in control (+35.2 +/- 3 0.6 %) and PAF
(+23.7 +/- 14.3 %) subjects compared with MSA subjects (+8.8 +/- 8.9 %) (Table 5.04
and Figure 5.08).
Figure 5.09 demonstrates the subjective psychological responses to the CO2 challenge.
Baseline psychological symptoms were similar in all three groups. Changes in anxiety,
224
fear and breathlessness VAS responses were not significantly different. Overall, MSA
subjects noted fewer somatic symptoms than either the control or PAF subjects. Of the
4 most commonly experienced symptoms (awareness of one’s heart beat; feeling hot;
having blurred vision; and dizziness), only 2 of the 9 MSA subjects experienced more
than one of these symptoms compared with all of the controls and 7 of the 9 PAF
subjects. Considered individually, there was only a non-significant trend towards lower
mean changes in VAS scores in the MSA group however.
Baseline systolic and diastolic blood pressures were the same in all three groups,
although MSA subjects had a higher resting heart rate (77 +/- 4.1 bpm compared with
60.9 +/- 3.9 bpm in controls, p = 0.02). PAF subjects had a lower resting cardiac output
(2.8 +/- 0.4 L/min) compared with both control (4.9 +/- 0.9 L/min, p = 0.02) and MSA
subjects (4.7 +/- 0.6 L/min, p = 0.02). In association with their lower CO, PAF subjects
also had a lower resting stroke volume and higher resting total peripheral resistance (p =
0.02 and p = 0.03 respectively) compared with controls.
Baseline skin blood flow was significantly lower in PAF compared to either controls (p
= 0.000006) or MSA (p = 0.026) (Table 5.05); MSA baseline skin blood flow was lower
than in controls (p = 0.002). Whilst all 3 groups showed reduction in skin blood flow
with inhalation of air, the reduction relative to baseline was less in PAF than controls (p
< 0.0006) or MSA (p < 0.04); (Table 5.05, Figure 5.10). MSA % reduction of skin
blood flow with air was less than seen in controls (p < 0.012). The time in seconds from
inspiratory gasp to onset of the skin vasomotor response was 3.1 +/- 0.3 for controls, 2.9
+/- 0.3 for MSA and 4.1 +/- 0.6 for PAF without significant differences between the
groups. The mean length of vasoconstriction following inhalation of air was 87.6 +/-
225
28.9 seconds for controls, 62.2 +/- 30.6 seconds for MSA without significant difference
between the groups. PAF showed vasodilatation rather than vasoconstriction.
After CO2 healthy controls showed a similar skin blood flow response to that seen with
inhalation of air over a similar time course (onset: 2.9 +/- 0.4 seconds after CO2;
duration: 91 seconds +/- 41.3 seconds). In PAF there was no reduction in skin blood
flow following CO2, but a striking transient increase in skin blood flow was observed in
PAF within 45 seconds (+/- 1.1 seconds) of CO2 inhalation (Figures 5.10 and 5.11).
This increase in skin blood flow was significantly greater than in controls (p < 0.00003)
or MSA (p < 0.00013). No increase in skin blood flow after CO2 inhalation was seen in
the controls, but a small increase was noted in 5/9 MSA subjects (+28.5% +/- 12.4),
which was less than that seen in PAF (p < 0.06).
226
Table 5.03. Baseline characteristics of subjects.
Control PAF MSA p
n
5 9 9
Age (years)
47 +/- 6.8 57.1 +/- 3.1 66.4 +/- 2.1
Cardiovascular parameters
SBP (mmHg)
122.5 +/- 12.3 128.8 +/- 13.6 132.7 +/- 8.7 -
DBP (mmHg)
66.9 +/- 3.8 75.7 +/- 4.3 76.1 +/- 7.2 -
HR (b/m)
60.92 +/- 3.925 66.8 +/- 2.6 77.0 +/- 4.1 0.02 (CvPAF)
CO (L/min)
4.9 +/- 0.9
2.8 +/- 0.4 4.7 +/- 0.6 0.04 (CvPAF) 0.02 (PAFvMSA)
SV (ml)
82.6 +/- 6.1
43.5 +/- 7.6 65.9 +/- 0.9 0.02 (CvPAF)
Skin Blood Flow (perfusion Units)
278.8 +/- 18.2 59.8 +/- 14.6 143.7 +/- 29.5 0.003 (CvPAF) 0.03 (PAFvMSA)
0.02 (CvMSA)
TPR (PRU)
1.3 +/- 0.2
2.2 +/- 0.2 1.5 +/- 0.3 0.03 (CvPAF)
Endocrine parameters
Cortisol (nmol/l)
439.8 +/- 75.8
347.1 +/- 52.0 403.7 +/- 53.0 -
Prolactin (mIU/l)
109.8 +/- 14.1 119.1 +/- 12.8 148.3 +/- 26.7 -
Salivary amylase (IU/l)
127643 +/- 20070
126686 +/- 35460
254867 +/- 92630
-
Noradrenaline (pg/ml)
288.4 +/- 31.3 126.8 +/- 19.7 246.9 +/- 13.1 <0.001 (CvPAF) <0.001 (PAF
vMSA)
227
Peak SBP response
control PAF MSA
0
25
50
75
mm
Hg
Time to peak SBP
control PAF MSA
0
100
200
secs
Peak DBP response
control PAF MSA
0
10
20
30
mm
Hg
Peak HR response
control PAF MSA
-15
-10
-5
0
b/m
Figure 5.07. Peak (A) and time to peak (B) systolic blood pressure; diastolic blood pressure (C) and heart rate (D) responses to a single breath of 35% CO2 in control, PAF and MSA subjects. The peak increase in SBP was significantly greater in controls compared to both PAF (p<0.01) and MSA (p<0.001) patients and the time to this peak was significantly shorter in the control subjects (p=0.03 and p=0.04 compared with PAF and MSA patients respectively).
228
Table 5.04. Peak cardiovascular and endocrine responses to CO2 exposure.
Control PAF MSA p Cardiovascular parameters
SBP (mmHg) ∆max 60.2+/-13.9 26.8+/-3.2 18.3+/-2.7 <0.01 (CvPAF)
%∆max 50.3+/-11.8 23.2+/-3.9 15.1+/-2.5 <0.001 (CvMSA) DBP (mmHg)
∆max 21.8+/-5.0 13.7+/-2.0 10.5+/-2.7 - %∆max 32.3+/-7.4 14.7+/-3.1 18.7+/-2.9
HR (b/min) ∆max -9.1+/-3.5 -4.3+/-2.3 -4.1+/-2.1 -
%∆max -16.1+/-6.9 -7.1+/-3.7 -6.4+/-3.2 CO (L/min)
∆max -0.6+/-0.5 -0.1+/-0.1 -0.1+/-0.1 - %∆max -6.2+/-9.5 -4.5+/-3.7 -3.3+/-4.0
SV (ml) ∆max 19.5+/-34.3 -1.7+/-1.4 0.6+/-2.4 -
%∆max 18.4+/-14.5 -0.5+/-5.7 2.0+/-4.8 TPR (PRU)
∆max 1.7+/-0.2 2.8+/-0.3 1.8+/-0.4 0.03 (CvPAF) %∆max 14.3+/-9.0 24.6+/-5.5 17.8+/-7.5
Endocrine parameters
Cortisol (nmol/l) ∆max 105.2+/-79.9 72.8+/-45.6 38.2+/-32.2 -
%∆max 35.2+/-30.6 23.7+/-14.3 8.8+/-8.9 Prolactin (mIU/l)
∆max 1.8+/-4.7 15.2+/-11.4 -5.4+/-9.3 - %∆max 1.7+/-4.0 7.0+/-5.4 -2.0+/-4.3
Salivary amylase (IU/l)
∆max 5171+/-12411 16651+/-13958
15551+/-5065 -
%∆max 3.7+/-9.1 15.2+/-12.5 15.2+/-13.8 Noradrenaline
(pg/ml)
∆max 124.2+/-19.5 5.2+/-2.8 20.0+/-6.4 <0.001 (CvPAF) %∆max 41.7+/-7.1 4.2+/-2.2 8.0+/-2.4 <0.001 (CvMSA)
229
Peak % cortisol response
control PAF MSA
0
25
50
75
%
Peak % prolactin response
control PAF MSA
-10
0
10
20
%Peak % salivary amylase
response
control PAF MSA
0
10
20
30
%
Peak % noradrenalineresponse
control PAF MSA
0
25
50
%
Figure 5.08. Peak percent cortisol, prolactin,
salivary amylase and noradrenaline levels following a single breath of 35% CO2.
230
Change in anxiety, fear andbreathlessness responses
Anx
iety
Fear
Bre
athl
essn
ess
0
25
50
75ControlPAFMSA
%
Somatic symptom responses
Hea
rt be
at
Hot
Blu
rred
vis
ion
Diz
zine
ss
-25
0
25
50
75
%
Figure 5.09. Percent change in anxiety, fear, breathlessness (top panel) and somatic symptom (lower panel) responses to 35% CO2 exposure.
231
Table 5.05. Changes in Skin Blood Flow in perfusion Units (+/- SE) after single breath of air or CO2
Baseline % Change in flow with
Room air
% Change in flow with
CO2
Controls 278.8 -83.8 ( 3.9) - 79.0 ( 5.6)
MSA 143.7 -52.7 ( 8.6) - 4.8 (15.9)
PAF 59.8 -31.7 (14.6) + 151.2 (40.2)
232
0
100
200
300
400
CONTROL R
0
100
200
300
400
8
0
100
200
300
400
Figure ControlPerfusio
AI
MSA
R
AI5 2 9 6 3 0 7 4 8 5 2 9 6 3 0 7 4 8 5 2 9 6 3 0 7 4 8 5 2 9 6
PAF
AIR
0 45 90 135 Time (in seconds) po
5.10. Skin Blood Flow changes follo Subject (top), PAF (middle) and Mn Units.
CO2
st
wSA
CO2
CO2
0 45 90 135 inhalation of air or CO2
ing inhalation of air or 35%CO2 in (lower). Y-axis shows skin blood flow in
233
-200
-100
0
100
200
300
Mea
n %
Cha
nge
in S
BF
follo
win
g in
hala
tion
AIR CO2 AIR CO2 AIR CO2
MSA (n = 9) PAF (n = 9) CONTROLS (n = 5)
Figure 5.11. Skin Blood Flow changes following inhalation of air or 35%CO2 in Control Subjects (gray), MSA (white) and PAF (black). Error bars = SE. Y-axis shows % change in SBF relative to baseline.
234
5.6.3. Discussion
In keeping with the known pathophysiology of MSA and PAF, we observed a
significant reduction in baseline noradrenaline levels in PAF subjects that was
associated with evidence of post-ganglionic hypersensitivity with vasoconstriction
(reduced skin blood flow and increased TPR) with a slightly reduced CO and SV
[Bannister and Mathias 1992a, Bannister and Mathias 1992b, Shannon et al 2000].
Also, as expected, MSA subjects had higher baseline heart rates although their other
cardiovascular parameters were no different to controls [Shannon et al 2000]. Apart
from noradrenaline, other hormone levels were no different between the three groups.
In control subjects, following a single breath of 35% CO2 there was the expected
increased in systolic and diastolic blood pressure associated with a marked increase in
plasma noradrenaline. Cardiac output remained unchanged and there was a trend
toward increased TPR. Both MSA and PAF subjects demonstrated a significant systolic
and diastolic pressor response following 35% CO2 exposure. However, these responses
were both smaller in magnitude, and occurred later than the responses observed in
normal controls. Whilst the magnitude of the responses were similar in MSA and PAF,
as is the case with supine hypertension in these subjects, it is likely they occurred
through different mechanisms.
Both MSA and PAF subjects showed a markedly reduced noradrenaline response
compared with controls, however, PAF subjects did increase their TPR compared to
both MSA and controls. It has been shown previously that resting noradrenaline levels
are reduced in PAF as are stimulated noradrenaline levels following various
haemodynamic and pharmacological challenges [Bannister and Mathias 1992a,
235
Bannister and Mathias 1992b, Shannon et al 2000]. Despite the histopathological
damage to peripheral autonomic nerves with the resultant impairment in noradrenaline
synthesis and release, sympathetically-mediated increases in blood pressure are seen in
PAF subjects, for example following the administration of the α-adrenoreceptor agonist
yohimbine [Shannon et al 2000]. This is most likely the result of post-ganglionic
receptor hypersensitivity to the small changes in noradrenaline seen in PAF subjects
[Bannister and Mathias 1992b, Shannon et al 2000], partly due to adrenoreceptor up-
regulation. This receptor hypersensitivity that increases the sensitivity of PAF patients
to the action of other pressor agents and vasodilators [Shannon et al 2000], is also
responsible for preserving, at least in part, the pressor response to CO2.
MSA, on the other hand, is characterised by features suggesting a central or
preganglionic abnormality. Noradrenaline levels are normal but they fail to respond to
stimuli that require centrally mediated interpretation or integration [Bannister and
Mathias 1992b, Shannon et al 2000, Parikh et al 2002]. The adreno-medullary response
to hypoglycaemia, for example, requires activation of central glucose receptors and an
efferent pre-ganglionic sympathetic arc [Shannon et al 2000]. This response is blunted
in both PAF and MSA, but can be restored in MSA subjects through the application of
post-ganglionic stimuli [Shannon et al 2000]. As mentioned, the pressor response to a
low dose CO2 challenge was lost in MSA subjects compared to controls presumably due
to impaired or absent responses of sympathetic autonomic nervous system structures in
the brainstem or cervical trunk [Braune et al 1997]. Supine hypertension, however, is
often seen in MSA subjects related to residual post-ganglionic sympathetic tone
[Shannon et al 2000, Parikh et al 2002] that occurs due to involvement of the
baroreceptor reflex arc with subsequent loss of baroreceptor-mediated buffering
[Shannon et al 2000] as well as post-ganglionic receptor hypersensitivity. This residual
236
post-ganglionic sympathetic tone is the likely mechanism producing the pressor
response observed in this study. In addition, it also likely contributes to the resting and
orthostatic tachycardia seen in MSA [Shannon et al 2000].
Only one MSA and 2 PAF subjects demonstrated a bradycardic response to the CO2 in
contrast to the control response. Again this is consistent with the parasympathetic
dysfunction previously described in both these conditions [Parikh et al 2002]. PAF
subjects behave in a manner similar to vagotomised subjects (such as the individual post
heart-lung transplant described in Chapter 4). MSA subjects differ from the individual
with CCHS described previously in that whilst both have central lesions, the process in
MSA is more diffuse with more severe autonomic involvement that is thus more likely
to involve parasympathetic centres that the more focal deficit that characterises CCHS
[Bannister and Mathias 1992b, Shannon et al 2000].
Along these lines, study by Braune et al [1997] evaluated the response of MSA patients
to a low dose (7%) CO2 challenge. MSA subjects failed to demonstrate the rise in
systolic blood pressure or heart rate seen in a comparison group of normal control
subjects again indicating the sympathetic response to CO2 is dependent on the presence
of intact preganglionic brainstem and cervical trunk centres. Based on this, our a priori
hypothesis was that PAF subjects would be expected to show impaired sympathetic and
parasympathetically mediated cardiovascular responses (as detailed above), but would
have preserved neurohormonal and psychological responses. Further, MSA subjects
would showed impaired autonomic cardiovascular responses, but would also show
impaired hormonal and psychological responses as these are centrally mediated.
Consistent with this hypothesis, endocrine and psychological responses in PAF subjects
were no different than in controls. MSA subjects on the other hand, showed a trend
towards lower cortisol responses compared with controls (8.8 +/- 8.9 % vs 35.2 +/- 30.6
237
% respectively). Similarly, there was an overall trend toward fewer somatic symptom
responses in MSA subjects compared with both control and PAF subjects.
Acute sympathetic denervation results in increased skin blood flow, however, chronic
denervation results in reduced skin blood flow [Barcroft and Walker 1949]. In this
study, subjects with PAF (which is a condition of chronic denervation), had
significantly lower skin blood flow readings compared with either MSA or control
subjects. Previous studies have shown that the inspiratory gasp reflex is associated with
a significant reduction in skin blood flow and that this effect is preserved in subjects
with MSA but impaired in those with PAF [Asahina et al 2003]. It is also known that
CO2 can cause cutaneous vasodilation (and an increase in skin blood flow) by non-
neural mechanisms [Bullard 1964, Ito et al 1989]. Following 35% CO2 control subjects
showed the expected reduction in skin blood flow following both an inspiratory gasp
(air breath) and a gasp plus CO2 (the 35% CO2 breath). MSA subjects showed a
blunting of the skin blood flow fall in response to both gasp and CO2 exposure, whilst
PAF subjects showed a blunted gasp response and a striking increase in skin blood flow
(or vasodilatory response) following 35% CO2 exposure. It seems likely that in normal
controls, the non-neurally mediated vasodilation produced by CO2 is masked by the
vasoconstriction produced by both the gasp reflex and the neurally mediated CO2-
induced noradrenaline release. In MSA subjects, the combination of a preserved gasp
reflex with a blunted neurally mediated CO2-induced vasoconstriction is still sufficient
to produce a vasoconstrictor response, albeit a smaller one as compared with controls.
PAF subjects on the other hand, have both blunted gasp reflex and an impaired neurally
mediated CO2 response that unmasks the usually hidden non-neurally mediated CO2-
induced vasodilatory response. This difference in skin blood flow following 35% CO2
exposure may proved a useful adjunct to the tests used to distinguish MSA and PAF.
238
With respect to the 35% CO2 model, these studies in patients with PAF (peripheral
autonomic failure) and MSA (central autonomic failure) add support to the proposed
underlying mechanisms that appear to drive the response to CO2. Overall, peripheral
autonomic failure is associated with an impaired parasympathetic response and
preserved neurohormonal and psychological responses similar to that observed in other
studies of peripheral denervation (heart-lung transplant and diabetic peripheral
neuropathy). This highlights the importance of central chemoreception in generating
these responses. However, the sympathetic involvement in PAF is also associated with
a blunted and delayed pressor response and an un-masking of the cutaneous
vasodilatory effect of CO2 that was previously hidden by the sympathetically-mediated
vasoconstrictor effect. MSA, on the other hand, is associated with impairments in both
the sympathetic and parasympathetic response, as well some in the neurohormonal and
psychological responses. The intact anxiety and fear responses seen in MSA subjects
suggests these occur directly from CO2 stimulation of limbic centres and independently
of brainstem sympathetic activity. However, the impaired somatic symptom responses
seen in both MSA and CCHS subjects suggests these may have more to do with
brainstem activation or at least communication pathways between limbic and brainstem
centres. Further, this also suggests the HPA response is likely secondary to activation
of brainstem sympathetic centres, rather then in response to projects from limbic
centres.
As detailed in Chapter 2, it had been reported elsewhere [Chatterton et al 1996] that
changes in salivary amylase reflects plasma catecholamine (particularly noradrenaline)
levels. In this study, however, there was no association between plasma noradrenaline
and salivary amylase either at rest or in response to the 35% CO2 challenge. At
baseline, PAF subjects had significantly lower noradrenaline levels but amylase levels
239
were no different to either controls or MSA subjects. Following 35% CO2 exposure,
noradrenaline levels increased significantly in controls (+41.7 +/- 7.1 %), but no such
increase in salivary amylase was observed (+3.7 +/- 9.1 %). Based on these results, and
until further studies are performed, salivary amylase does not seem to accurately predict
either resting or stimulated plasma noradrenaline.
240
5.7. Conclusion
The conditions studied in the above series of experiments were chosen as they represent
clinical models of basic pathophysiology. Their study provides useful information
regarding not only some of the potential mechanisms by which the response to CO2
occurs, but also about the disease itself – its pathophysiology, reasons for clinical
symptoms, potential for treatment and their efficacy as well as potential risks and
complications.
In Chapter 4 it was noted that a physiologically or pharmacologically suppressed HPA
axis had no influence over an individuals ability to mount a normal pressor response to
CO2, nor did it influence the bradycardic response or the psychological symptoms
experienced with 35% CO2 inhalation. However, in these situations baseline cortisol
levels were not completely suppressed and it was still possible that low levels of
circulating (and more particularly brain) glucocorticoid was sufficient to mediate the
central (brainstem and limbic) responses to the CO2 challenge. Subjects with Addison’s
disease, without glucocorticoid replacement, still demonstrated the same bradycardic
and psychological response to CO2 than when they were adequately replaced with
glucocorticoid. This would again suggest that the psychological response is
independent of HPA activity as is the bradycardia, which is most likely due to direct
vagal stimulation by CO2. Unreplaced subjects with Addison’s disease showed a
smaller pressor response than when replaced with both glucocorticoid and
mineralocorticoid. The difference was small and may be due to a reduced positive re-
enforcement effect of the HPA axis on brainstem sympathetic activity. However, it is
probably more likely that the lack of mineralocorticoid which, through the renin-
angiotensin system, plays an important role in maintenance of blood pressure in patients
241
with adrenocortical failure was responsible for this difference. The striking differences
in the pressor response seen in the later studies in people with autonomic failure
suggests the sympathetic system per se has a much greater role in determining the
pressor response to CO2 than does the HPA axis.
The study of patients with diabetic autonomic neuropathy has shed more light on the
origin of the bradycardia. The heart rate slowing in response to 35% CO2 exposure fits
well with the Polyvagal theory of Porges [Porges 1995]. The normal slowing following
a single breath of 35% CO2 is consistent with withdrawal of the NA component of the
vagal response. A maintained bradycardia with no change in blood pressure in the
Addison’s patients off treatment also supports this concept.
A similar response has been demonstrated in other situations of acute stress [Porges
1995]. Parasympathetic failure (as occurs in diabetic autonomic neuropathy) may result
in impaired activity of both components of the vagal system thereby allowing
sympathetic activity to predominate. This would be expected to result in higher resting
heart rates and a loss of the bradycardia with 35% CO2 exposure, as was seen in this
group. We would also postulate that this is a potential mechanism to explain the
increased cardiovascular mortality associated with diabetic autonomic neuropathy and
further studies are planned to evaluate this further (see below).
As mentioned, the sympathetic autonomic system appears much more important in
determining the overall response to CO2. Central autonomic failure (MSA) resulted in
a reduced pressor response, less frequently observed bradycardia, fewer somatic
symptoms of fear and a smaller cortisol response. As was also evident from the
individual with CCHS, brainstem sympathetic systems appear to be the principal site
242
regulating the response to 35% CO2. In PAF where peripheral autonomic systems are
involved, both the pressor and bradycardia were affected, but the HPA and
psychological systems were not. Given the previous evidence that acute fear (in the
form of panic attacks) can be precipitated by CO2 exposure without necessarily
producing an HPA response, it is again likely that the HPA response occurs
predominantly following activation of brainstem sympathetic systems. The vagal
response, on the other hand, occurs independently whilst the psychological response
occurs probably as a result of both direct activation as well as with some influence from
brainstem sympathetic systems.
Figure 4.18 can then be updated to include this current evidence to provide an
illustration of the likely pathways mediating the response to a single breath of 35% CO2,
as shown in Figure 5.12.
243
Single breath of
35% CO2
Noradrenergic brainstem
Limbic centres
FoabPsbH
Vagal centres/vagus nerve
centres(VLM, LC) (CnA)
PVN Bradycardia
NA ACTH
Emotional arousal ↑SBP Cortisol
igure 5.12. Revised schematic diagram suggesting the principal pathways f CO2 mediated HPA, cardiovascular and psychological responses (solid rrows). Vagal and noradrenergic brainstem centres are directly stimulated y CO2 resulting in bradycardia and an acute pressor response respectively. sychological responses occur due to a combination of direct CO2timulation of a central fear circuit as well as indirect stimulation via rainstem NA centres. PA responses occur indirectly in response to brainstem NA stimulation.
244
CHAPTER 6
SUMMARY, CONCLUSIONS AND FUTURE
DIRECTIONS
245
6.1. Summary
The three main objectives of this thesis were to:
i. Design, develop and evaluate the single breath 35% CO2 test as a potential model
of the stress response in humans,
ii. Understand the mechanisms that are involved in mediating the observed
physiological, endocrine and psychological responses to the test, and
iii. Investigate its potential for use in clinically important disease states.
The initial design of the test was based on an existing model from the psychiatric
literature that had been used and evaluated as a means of studying acute anxiety and
panic in susceptible patients. Detailed physiological and endocrine studies had not been
previously performed. The first studies carried out in this thesis evaluated different CO2
doses and confirmed the linear dose related increase in anxiety responses. The
sympathetic and HPA responses, on the other hand, appeared to be threshold dependent
with 35% CO2 producing the most consistent psychological, endocrine and
physiological response whilst being well tolerated with no significant adverse effect.
The neurohormonal response was limited to the HPA axis (ACTH and cortisol) and
prolactin with no release of other anterior pituitary hormones, vasopressin or renin seen.
The physiological cardiovascular response differed from that seen in studies of low dose
CO2 exposure. Low dose and steady state protocols of CO2 exposure, principally used
in the study of anxiety and ventilatory physiology, are associated with tachycardia and
systolic hypertension. A single breath of 35% CO2 is associated with a systolic pressor
response and bradycardia. Initial thoughts that the bradycardia was due to reflex
baroreceptor activity were later proved to be incorrect (see below). The psychological
246
response included general feelings of anxiety, fear and breathlessness and also invoked
a few very specific somatic symptoms.
Subsequent experiments were undertaken to assess the reliability and reproducibility of
the test and also to refine the technique of administration of the CO2. It was clear that
certain operator dependent factors were important particularly the specific instructions
for performing the test and the nature and detail of the symptoms they might expect. In
addition, refinements to the equipment used were made in reverting back to a mouth-
piece and nose clip arrangement rather than a nasal-oral face mask that was associated
with poor inspiratory efforts and limited responses.
With these refinements, the psychological and physiological responses were shown to
be robust and easily reproducible. HPA responses were modest and somewhat more
variable. They appear more easily influenced by a number of potentially confounding
factors such as the individuals level of anticipatory anxiety, the specifics of the
instructions received, their expectations based on details received before taking the
breath, and the performance of the test especially the adequacy of the inspiration.
As mentioned, the HPA axis response was modest and does not appear to be the
primary determinant of either the psychological or autonomic responses.
Administration of naltrexone to remove negative feedback inhibition of the HPA axis
and raise baseline cortisol levels had no effect on either psychological or autonomic
responses. Similarly, the administration of other neurotransmitters that may be
expected to acutely enhance the HPA response to stress (for example paroxetine), also
had no effect on other responses. Further, there was no effect from decreasing the
activity of the HPA axis. This was done by either evaluating subjects with
247
physiologically suppressed axes (lactating women), pathologically suppressed axes
(Addison’s disease) or through pharmacological manipulation with the administration of
metyrapone (decrease cortisol synthesis) or spironolactone (decrease activity of central
mineralocorticoid receptors). In contrast, subjects with deficits of central autonomic
function (MSA) demonstrated a reduced ability to mount a cortisol response to the CO2
challenge.
Clearly distinguishable pattern of sympathetic and parasympathetic responses emerged.
Initial impressions that the bradycardia was as a baroreceptor response to the increased
systolic blood pressure were discounted on the grounds that continuous cardiovascular
monitoring identified the bradycardia as occurring prior to the increase in blood
pressure. Further, vagotomy (heart-lung) transplantation and parasympathetic failure,
both central (MSA) and peripheral (PAF and DAN), abolished the bradycardia
regardless of whether the pressor response was preserved (vagotomy and diabetes) or
impaired (MSA and PAF). The sympathetic response was clearly centrally mediated,
being preserved in peripheral neuropathy (diabetes) and HPA axis impairment (as
above). The response, characterised by systolic and diastolic hypertension was
multifactorial in origin, however, in normal individuals appeared principally to be
associated with significant increases in noradrenaline and peripheral vasoconstriction.
Abnormalities of central CO2 chemoreception and/or integration (CCHS and MSA)
caused significant blunting of the pressor response. The response, however, was not
abolished in MSA subjects but was also not associated with an increase in
noradrenaline. Residual post-ganglionic sympathetic activity (also responsible for
supine hypertension in these subjects) is the likely mechanism by which this response
occurred. The response was also not abolished in subjects with peripheral autonomic
failure (PAF and diabetic peripheral neuropathy with orthostatic hypotension). In this
248
case, post-synaptic receptor hypersensitivity to small changes in noradrenaline is
presumed to be the likely mechanism producing this response.
Subjective anxiety and fear responses were not affected by changes in HPA axis
activity, peripheral neuropathy (diabetes and PAF) or by central autonomic dysfunction
(MSA). This suggests the anxiety component is mediated by direct CO2 stimulation of
limbic fear centres such as the central nucleus of the amygdala. The specific somatic
symptoms of fear typically associated with 35% CO2 exposure were blunted in patients
with abnormalities of brainstem CO2 chemosensation (CCHS) and/or autonomic
integration (MSA). This suggests that projections from brainstem autonomic centres to
limbic fear centres have at least some role to play in generating the overall response to
CO2.
From a clinical perspective, a study of women during the lactation phase has reinforced
the notion that maternal brain plasticity during pregnancy and post-partum alters stress
hormone responsiveness to provide a more advantageous environment for the well-
being of the infant. Previous studies of rodents had demonstrated reduced maternal
HPA responsiveness to stress during lactation [Lightman et al 1997], but in humans
similar suppression following psychological and exercise stress was limited to a specific
time following suckling [Altemus et al 1995, Heinrichs et al 2001]. Following 35%
CO2 exposure, on the other hand, no significant suppression of stress responses was
seen. This would indicate that suppression of stress responses may be advantageous to
a nursing mother faced with stressors that are not pertinent to survival, but when
physiological homeostasis is threatened survival demands a full neuroendocrine and
behavioural response. Failing to appropriately alter stress response patterns during this
time (for example by continuing to respond excessively to non-pertinent survival
249
threats, or alternatively, not responding sufficiently to pertinent threats) may predispose
these individuals to an increased risk of stress related illnesses (such as post partum
depression) during this period.
Another study demonstrated dissociative sympatho-vagal dysfunction in diabetic
autonomic neuropathy. This included preservation of the systolic pressor response
whilst the normally observed bradycardia was abolished. The initial bradycardia seen in
healthy controls could be explained by the Polyvagal theory of Porges [Porges 1995]
which distinguishes between the rhythmic beat-to-beat variations in heart rate (mediated
by phylogenetically newer myelinated fibres from the nucleus ambiguus) and the
marked bradycardia (mediated by more primitive unmyelinated fibres from the DMNX)
that is occasionally observed during acute stress. Active withdrawal of the nucleus
ambiguus component leaves the individual susceptible to the effects of excessive
neurogenic bradycardia if sympathetic compensation is inadequate and this is the likely
explanation for neurogenic syncope [Porges 1995, Reed et al 1999]. In those with
diabetic autonomic neuropathy the loss of both components of the vagal response could
result in unprotected excessive sympathetic activity and this could serve as a potential
explanation for the increased risk of cardiac tachyarrhythmias seen in these patients.
Finally, a study of individuals with autonomic failure syndromes has shed further light
on the presence of residual mechanisms whereby sympathetic tone is maintained despite
severe pathological involvement of either central (MSA) or peripheral (PAF) autonomic
systems. This residual sympathetic tone is an important contributing factor to the
observed supine hypertension many of these subjects display (particularly those with
MSA) and is also an important limiting factor in the treatment of their orthostatic
hypertension. Moreover, this study demonstrated a marked difference in the pattern of
250
skin blood flow changes following 35% CO2 exposure in PAF subjects compared with
MSA and control subjects. Control subjects normally display a significant reduction in
skin blood flow following CO2 that is likely due to the combination of the inspiratory
gasp reflex and the CO2 stimulated increase in noradrenaline that increases TPR. MSA
subjects show a reduction in skin blood flow that is significantly reduced in magnitude
compared with controls most likely from a loss of the CO2-induced sympathetic
stimulation. PAF subjects on the other hand show a completely different skin blood
flow pattern with a striking increase in flow. This probably results from the loss of the
vasoconstrictor effects of both the gasp reflex and the noradrenaline-mediated CO2
effect with an unmasking of a local, non-neural vasodilatory effect of CO2. This
difference could potentially serve as an important adjunct in the clinical differentiation
of these two disorders.
251
6.2. Conclusions and future directions
The 35% CO2 single breath test has shown itself to be a safe, reliable and reproducible
means of generating a general stress response in humans. With a few carefully
considered limitations, the test could be applied to a wide range of individuals without a
significant risk of harm. It is convenient, quick and simple to administer and doesn’t
require sophisticated monitoring. Further work is planned to further define the
limitations of the HPA axis response particularly with regard to potential confounding
variables such as anticipatory anxiety, age and gender differences and menstrual cycle
changes. Further reproducibility studies with larger subject numbers are also planned,
as are studies to compare this model to other established models of both psychological
(such as the TSST) and physical (such as exercise) stress.
Studies to determine whether the test can be performed entirely non-invasively with the
use of non-invasive cardiovascular monitors and salivary biochemical measures
(including salivary cortisol and amylase as surrogates for plasma cortisol and plasma
noradrenaline respectively) are also due to begin shortly.
The model has a large number of potential clinical applications both at an
epidemiological level and at an individual level. There is already the potential to use it
in the diagnosis and assessment of diabetic autonomic neuropathy and in the assessment
of autonomic failure syndromes including the syndrome of unexplained syncope, where
it may be more useful than the conventional tilt test. The model has potential for use in
a multitude of conditions that are characterised by or associated with abnormalities of
HPA or autonomic system dysfunction. These include depression, anxiety and affective
disorders, irritable bowel syndrome, obesity and the metabolic syndrome, chronic
252
fatigue, coronary artery disease, immune and inflammatory conditions and so on. It
could potentially server as a predictor of disease susceptibility, a marker of disease
occurrence or as a determinant of treatment success or outcome. A number of studies
using the model in some the above conditions have commenced or are being planned.
The stress response system has evolved as both an early warning system capable of
recognising potential or existing threats, and as a response system that can initiate and
drive the necessary processes required to escape or confront the threat. By its very
nature, the response is dynamic, beginning rapidly with brain and behavioural activation
followed quickly by physiological activation. These processes are characterised by
positive-feedback and feed forward loops that enhance and reinforce themselves as well
as recruiting other arms of the stress response. Slower acting hormone systems provide
checks and balances to the already active, but energy expensive systems, putting a brake
on the whole response to ensure it is kept appropriate to the type of stress faced, to its
intensity and duration, and to ensure the response is switched off when the threat has
been adequately dealt with. It is well established that impaired functioning of these
stress response systems is central to the pathophysiological pathways that underlie many
psychiatric, immune and physical disorders.
The challenge for the future is a need to develop tools that can accurately and
consistently measure the reactivity of these systems and the consequences of their
activation. It is essential to be able to do this in individuals, in order to predict their
vulnerability to illness, and in large populations, in order to measure the impact of stress
and the efficacy of any therapeutic intervention. Further, there is a need to expand our
current understanding of the mechanisms that produce stress-related pathophysiology in
order to design more effective therapies.
253
CHAPTER 7
REFERENCES
254
1. Agelink MW, Ullrich H, Baumann B, Strum S and Majewski T. Effects of reboxetine, a selective norepinephrine reuptake inhibitor, on sympathetic and parasympathetic outflow to the heart: preliminary data. Psychopharmacology 2002; 163(2):151-156
2. Aguilera and Rabadan-Diehl. Vasopressinergic regulation of the hypothalamic-pituitary-adrenal axis: implications for stress adaptation. Regul Pept 2000; 96(1-2):23-29
3. Al’Absi M, Bongard S, Buchanan T, Pincomb GA, Licinio J and Lovallo WR. Cardiovascular and neuroendocrine adjustment to public speaking and mental arithmetic stressors. Psychophysiol 1997; 34(3):266-275
4. Altemus M, Deuster PA, Galliven E, Carter CS and Gold PW. Suppression of hypothalmic-pituitary-adrenal axis responses to stress in lactating women. J Clin Endocrinol Metab 1995; 80(10):2954-2959
5. Altemus M, Redwine LS, Leong YM, Frye CA, Porges SW and Carter CS. Responses to laboratory psychosocial stress in post-partum women. Psychosom Med 2001; 63:814-821
6. Amico JA, Johnston JM and Vagnucci AH. Suckling-induced attenuation of plasma cortisol concentrations in postpartum lactating women. Endocrine Res 1994; 20(1):79-87
7. Andrews ZB and Grattan DR. Opioid receptor subtypes involved in the regulation of prolactin secretion during pregnancy and lactation. J Neuroendocrinol 2003; 15(3):227-236
8. Anton RF, Moak DH, Waid LR, Latham PK, Malcolm RJ and Dias JK. Naltrexone and cognitive behavioral therapy for the treatment of outpatient alcoholics: results of a placebo-controlled trial. Am J Psychiatry 1999; 156(11):1758-1764
9. Argyropoulos SV, Bailey JE, Hood SD, Kendrick AH, Rich A, Laszlo G, Lightman SL and Nutt DJ. Inhalation of 35% CO2 results in subjective fear and activates the HPA axis. Psychoneuroendocrinol 2002; 27:715-729
10. Arlt W, Callies F, van Vlijmen JC, Koehler I, Reincke M, Bidlingmaier M, Huebler D, Oettel M, Ernst M, Schulte HM and Allolio B. Dehydroepiandrosterone replacement in women with adrenal insufficiency. N Engl J Med 1999; 341(14):1013-1020
11. Armando I, Seltzer A, Bregonzio C and Saavedra JM. Stress and angiotensin II: novel therapeutic opportunities. Curr Drug Target CNS Neurol Disord 2003; 2(6):413-419
12. Arroba AI, Frago LM, Paneda C, Argente J and Chowen JA. The number of lactotrophs is reduced in the anterior pituitary of streptozotocin-induced diabetic rats. Diabetologia 2003; 46(5):634-638
13. Asahina M, Kikkawa Y, Suzuki A and Hattori T. Cutaneous sympathetic function in patients with multiple system atrophy. Clin Auton Res 2003; 13(2):91-95
14. Askanazi J, Silverberg PA, Foster RJ, Hyman AI, Milic-Emili J and Kinney JM. Effects of respiratory apparatus on breathing pattern. J Appl Physiol 1980; 48(4):577-580
15. Aston-Jones G, Rajkowski J and Cohen J. Role of locus coeruleus in attention and behavioral flexibility. Biol Psychiatry 1999; 46(9):1309-1320
16. Bailey J, Argyropoulos SV, Lightman SL and Nutt DJ. Does the brain noradrenaline network mediate the effects of the CO2 challenge. J Psychopharmacol 2002; 17:252-259
17. Bakshi VP and Kalin NH. Corticotropin releasing hormone and animal models of anxiety: Gene-environment interactions. Biol Psychiatry 2000; 48:1175-1198
18. Bannister R and Mathias CJ. Introduction and classification of autonomic disorders. In: Autonomic failure. A textbook of clinical disorders of the autonomic nervous system. (Bannister and Mathias eds.) 3rd Edition, 1992a; pp1-12, Oxford University Press, Oxford
19. Bannister R and Mathias CJ. Primary autonomic failure syndromes. In: Autonomic failure. A textbook of clinical disorders of the autonomic nervous system. (Bannister and Mathias eds.) 3rd Edition, 1992b; pp531-547, Oxford University Press, Oxford
20. Bannister R and Oppenheimer DR. Degenerative diseases of the nervous system and autonomic failure. Brain 1972; 95:457-474
255
21. Barcroft H and Walker AJ. Return of tone in blood-vessels of the upper limb after sympathectomy. Lancet 1949; 1035-1039
22. Barker EL and Blakely RD. Norepinephrine and serotonin transporters: molecular targets of antidepressant drugs. In Psychopharmacology: The Fourth Generation of Progress (Bloom FE and Kupfer DJ eds), 1995; Raven Press, New York
23. Battaglia M and Perna A. The 35% CO2 challenge in panic disorder: optimization by receiver operating characteristic (ROC) analysis. J Psychiatric Res 1995; 29:111–119
24. Battaglia M, Bertella S, Ogliari A, Bellodi L and Smeraldi E. Modulation by muscarinic antagonists of the response to carbon dioxide challenge in panic disorder. Arch Gen Psych 2001; 58:114-119
25. Bellodi L, Perna G, Caldirola D, Arancio C, Bertani A and Di Bella D. CO2-induced panic attacks: a twin study. Am J Psychiatry 1998; 155(9):1184-1188
26. Ben Zion IZ, Meiri G, Greenberg BD, Murphy DL and Benjamin J. Enhancement of CO2-induced anxiety in healthy volunteers with the serotonin antagonist metergoline. Am J Psychiatry 1999; 156:1635–1637
27. Benedict CR. Simultaneous measurement of urinary and plasma norepinephrine, epinephrine, dopamine, dihydroxyphenylalanine, and dihydroxyphenylacetic acid by coupled-column high-performance liquid chromatography on C8 and C18 stationary phases. J Chromatogr 1987; 385:369-375
28. Bertani A, Bellodi L, Bussi R, Caldirola D, Cucchi M and Perna G. The effect of one-week treatment with venlafaxine on 35% CO2 hyperreactivity in patients with panic disorder: an open study. J Clin Psychopharmacol 2003; 23(1):106-108
29. Bertani A, Perna G, Arancio C, Caldirola D and Bellodi L. Pharmacologic effect of imipramine, Paroxetine and sertraline on 35% carbon dioxide hypersensitivity in panic patients: a double-blind, random, placebo-controlled study. J Clin Psychopharmacol 1997; 43:612-615
30. Billingsley ML and Kubena RK. The effect of naloxone and picrotoxin on the seative and anticonflict effects of benzodiazepines. Life Sci 1978; 22:897-906
31. Biondi M and Picardi A. Clinical and biological aspects of bereavement and loss-induced depression: A reappraisal. Psychother Psychosom 1996; 65:229-245
32. Biondi M and Picardi A. Psychological stress and neuroendocrine function in humans: The last two decades of research. Psychother Psychosom 1999; 68:114-150
33. Boone JL and Christensen JF. Stress and disease. In: Behavioural medicine in primary care. (Feldman and Christensen eds), 1998; pp 265-76, Appleton and Lange, Stamford.
34. Borm K, Slawik M, Seiler L, Flohr F, Petrick M, Honegger J and Reincke M. Is the plasma ACTH concentration a reliable parameter in the insulin tolerance test? Eur J Endocrinol 2003; 149(6):535-541
35. Bornstein SR and Chrousos GP. Adrenocorticotropin (ACTH)- and non-ACTH-mediated regulation of the adrenal cortex: Neural and immune inputs. J Clin Endocrinol Metab 1999; 84(5):1729-1736
36. Bornstein SR, Breidert M, Ehrhart-Bornstein M, Kloos B, and Scherbaum WA. Plasma catecholamines in patients with Addison’s disease. Clin Endocrinol 1995; 42:215–218
37. Bourin M, Baker GB and Bradwejn J. Neurobiology of panic disorder. J Psychosom Res 1998; 44(1):163-180
38. Bower RJ and Adkins JC. Ondine's curse and neurocristopathy. Clin Pediatr 1980; 19(10):665-668
39. Bradbury MJ, Akana SFand Dallman MF. Roles of type I and II corticosteroid receptors in regulation of basal activity in the hypothalamo-pituitary-adrenal axis during the diurnal trough and the peak: evidence for a nonadditive effect of combined receptor occupation. Endocrinology 1994; 134(3):1286-1296
40. Bradwejn J, Koszycki D and Bourin M. Dose ranging study of the effects of cholecystokinin in healthy volunteers. J Psych Neurosci 1991a; 16(2):91-95
41. Bradwejn J, Koszycki D and Shriqui C. Enhanced sensitivity to cholecystokinin tetrapeptide in panic disorder. Clinical and behavioral findings. Arch Gen Psych1991b; 48(7):603-610
256
42. Brambilla F. Psychoneuroendocrinology: a science of the past or a new pathway for the future? E J Pharmacol 2000; 405:341-349
43. Brannan S, Liotti M, Egan G, Shade R, Madden L, Robillard R, Abplanalp B, Stofer K, Denton D and Fox P. Neuroimaging of cerebral activations and deactivations associated with hypercapnia and hunger for air. PNAS 2001; 98(4):2029-2034
44. Braune S, Hetzel A, Prasse A, Dohms K, Guschlbauer B and Lucking CH. Stimulation of sympathetic activity by carbon dioxide in patients with autonomic failure compared to normal subjects. Clin Auton Res 1997; 7(6):327-332
45. Brunson KL, Avishai-Eliner S, Hatalski CG and Baram TZ. Neurobiology of the stress response in early life: evolution of a concept and the role of corticotropin releasing hormone. Mol Psych 2001; 6:647-656
46. Bugajski J, Gadek-Michalska A, Olowska A, Borycz J, Glod R, and Bugajski AJ. Adrenergic regulation of the hypothalamic-pituitary-adrenal axis under basal and social stress conditions. J Physiol Pharmacol. 1995; 46(3):297-312
47. Bullard RW. Effects of carbon dioxide inhalation on sweating. J Appl Physiol 1964; 19:137-141
48. Bullinger M, Naber D, Pickar D, Cohen RM, Kalin NH, Pert A and Bunney WE Jr. Endocrine effects of the cold pressor test: relationships to subjective pain appraisal and coping. Psychiatry Res 1984; 12:227–233
49. Busch L, Sterin-Borda L and Borda E. Differences in the regulatory mechanism of amylase release by rat parotid and submandibular glands. Arch Oral Biol 2002; 47(10):717-722
50. Butelman ER and Kreek MJ. kappa-Opioid receptor agonist-induced prolactin release in primates is blocked by dopamine D(2)-like receptor agonists. Eur J Pharmacol2001; 423(2-3):243-249
51. Cacioppo JT, Malarkey WB, Kiecolt-Glaser JK, Uchino BN, Sgoutas-Emch SA, Sheridan JF, Berntson GG and Glaser R. 1995. Heterogeneity in neuroendocrine and immune responses to brief psychological stressors as a function of autonomic cardiac activation. Psychosom Med 57:154-164
52. Caldirola D, Perna G, Arancio C, Bertani A and Bellodi L. The 35% CO2 challenge test in patients with social phobia. Psychiatry Res 1997; 71(1):41-48
53. Carrasco GA and Van Der Kar LD. Neuroendocrine pharmacology of stress. E J Pharmacol 2003; 463:235-272
54. Carter DA and Lightman SL. Opioid control of oxytocin secretion: evidence of distinct regulatory actions of two opiate receptor types. Life Sci 1987; 40(23):2289-2296
55. Chan O, Chan S, Inouye K, Shum K, Matthews SG and Vranic M. Diabetes impairs hypothalamo-pituitary-adrenal (HPA) responses to hypoglycemia, and insulin treatment normalizes HPA but not epinephrine responses. Diabetes 2002; 51(6):1681-1689
56. Chan O, Inouye K, Riddell MC, Vranic M and Matthews SG. Diabetes and the hypothalamo-pituitary-adrenal (HPA) axis. Minerva Endocrinol 2003; 28(2):87-102
57. Chaouloff F. Serotonin, stress and corticoids. J Psychopharmol 2001; 4:139-151 58. Charney DS and Heninger GR. Alpha 2-adrenergic and opiate receptor blockade:
Synergistic effects on anxiety in healthy subjects. Arch Gen Psychiatry 1986; 43:1037-1041
59. Chatterton RT Jr, Vogelsong KM, Lu YC, Ellman AB and Hudgens GA. Salivary alpha-amylase as a measure of endogenous adrenergic activity. Clin Physiol 1996; 16(4):433-448
60. Chatterton RT Jr, Vogelsong KM, Lu YC and Hudgens GA. Hormonal responses to psychological stress in men preparing for skydiving. J Clin Endocrinol Metab 1997; 82(8):2503-2509
61. Checkley S. The neuroendocrinology of depression and chronic stress. Br Med Bull1996; 52:597-617
62. Chrousos GP and Gold PW. A healthy body in a healthy mind––and vice versa––the damaging power of "uncontrollable" stress. J Clin Endocrinol Metab 1998; 83:1842–1845
257
63. Chrousos GP. Stressors, stress, and neuroendocrine integration of the adaptive response: The 1997 Hans Seyle memorial lecture. Ann NY Acad Sci 1998; 851:311-335
64. Chrousos GP. The role of stress and the hypothalamic-pituitary-adrenal axis in the pathogenesis of the metabolic syndrome: neuro-endocrine and target tissue-related causes. Int J Obesity 2000; 25(S2):S50-S55
65. Coates EL. Olfactory CO2 chemoreceptors. Respir Physiol 2001; 129(1-2):219-229 66. Cohen ME and White PD. Life situations, emotions and neurocirculatory aesthenia.
Psychosom Med 1951; 13:335-357 67. Cohen S, Kamarck T and Mermelstein R. A global measure of perceived stress. J
Health Soc Behav 1983; 24(4):385-396 68. Commare MC, Francois B, Estournet B and Barois A. Ondine's curse: a discussion of
five cases. Neuropediatrics 1993; 24(6):313-318 69. Coplan JD, Moreau D, Chaput F, Martinez JA, Hoven CW, Mandell DJ, Gorman JM
and Pine DS. Salivary cortisol concentrations before and after carbon-dioxide inhalations in children. Biol Psych 2002; 51:326-333
70. Cox JL, Holden JM and Sagovsky R. Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. Br J Psych 1987; 150:782-786
71. Cross BA and Silver IA. Central activation of the sympathetico-adrenal system by hypoxia and hypercapnia. J Endocrinol 1962; 24:91-103
72. Cullen D H and Edgar E. Cardiovascular effects of carbon dioxide. Anaesthesiology 1974; 41(4):345-349
73. Daniel SE. The neuropatyhology and neurochemistry of multiple system atrophy. In: Autonomic failure. A textbook of clinical disorders of the autonomic nervous system. (Bannister and Mathias eds.) 3rd Edition, 1992; pp564-85, Oxford University Press, Oxford
74. Davy H. Researchers, chemical and philosophical: chiefly concerning nitrous oxide (facsimile of the original contribution, 1800). Butterworths, London. 1972
75. de Kloet RE. Brain corticosteroid receptor balance and homeostatic control. Front Neuroendocrinol 1991;12:95–164
76. de Vries WR, Bernards NT, de Rooij MH and Koppeschaar HP. Dynamic exercise discloses different time-related responses in stress hormones. Psychosom Med 2000; 62(6):866-872
77. DiBona GF and Kopp UC. Neural control of renal function. Physiol Rev 1997; 77(1):75-197
78. Dinan T. Serotonin and the regulation of hypothalamic-pituitary-adrenal axis function. Life Sci 1996; 58(20):1683-1694
79. Drolet G, Dumont EC, Gosselin I, Kinkead R, Laforest S and Trottier JF. Role of endogenous opioid system in the regulation of the stress response. Prog Neuropsychopharmacol Biol Psychiatry 2001; 25(4):729-741
80. Drury AN. The percentage of carbon dioxide in the alveolar air and the tolerance to accumulating carbon dioxide in cases of so-called ‘irritable-heart’ of soldiers. Heart 1919; 7:165-173
81. Durel LA, Kus LA, Anderson NB, McNeilly M, Llabre MM, Spitzer S, Saab PG, Efland J, Williams R and Schneiderman N. Patterns and stability of cardiovascular responses to variations in the cold pressor test. Psychophysiol 1993; 30(1):39-46
82. Edmonds ME and Watkins PJ. Clinical presentations of diabetic autonomic failure. In: Autonomic failure. A textbook of clinical disorders of the autonomic nervous system. (Bannister and Mathias eds.) 3rd Edition, 1992; pp 698-720, Oxford University Press, Oxford
83. Ehlert U, Gaab J and Heinrichs M. Psychoneuroendocrinological contributions to the aetiology of depression, posttraumatic stress disorder, and stress-related bodily disorders: the role of the hypothalamus-pituitary-adrenal axis. Biol Psychiatry 2001; 57:141-152
84. Elenkov IJ, Wilder RL, Chrousos GP and Sylvester V. The sympathetic nerve – An intergrative interface between two supersystems: The brain and the immune system. Pharm Rev 2000; 52:595-638
258
85. Ellingboe J, Mendelson JH and Kuehnle, JC. Effects of heroin and naltrexone on plasma prolactin levels in man. Pharmacol Biochem Behav 1980; 12:163–165
86. Eriksen HR, Olff M, Murison R and Urin H. The time dimension in stress responses: relevance for survival and health. Psych Res 1999; 85:39-50
87. Erturk E, Jaffe CA and Barkan AL. Evaluation of the integrity of the hypothalamic-pituitary-adrenal axis by insulin hypoglycemia test. J Clin Endocrinol Metab 1998; 83(7):2350-2354
88. Ewing DJ and Clark BF. Diagnosis and management of diabetic autonomic neuropathy. Br Med J 1982; 285:916
89. Folgering H, Kuyper F and Kille JF. Primary alveolar hypoventilation (Ondine's curse syndrome) in an infant without external arcuate nucleus. Case report. Bull Eur Physiopathol Respir 1979; 15(4):659-665
90. Forst T, Pfutzner A, Jahn C, Schmitz H, Lichtwald K, Beyer J and Lehnert H. Decreased sympatho-adrenal activity in diabetic patients with autonomic dysfunction following mental stress. J Auton Nerv Syst 1996; 61(1):31-36
91. Forster CD and MacDonald IA. The assay of the catecholamine content of small volumes of human plasma. Biomed Chromatogr 1999; 13(3):209-215
92. Frecska E, Arato M, Banki CM, Bagdy G, Perenyi A, Mohari K and Fekete MI. Diurnal variation in fentanyl-induced hormone responses and side effects. Neuropsychopharmacology 1988; 1:235–238
93. Fukudo S, Nomura T and Hongo M. Impact of corticotropin-releasing hormone on gastrointestinal motility and adrenocorticotropic hormone in normal controls and patients with irritable bowel syndrome. Gut 1998; 42(6):845-849
94. Gammatopoulos DK and Chrousos GP. Functional characteristics of CRH receptors and potential clinical applications of CRH-receptor antagonists. Trends Endocrinol Metab 2002; 13(10):436-444
95. Gerra G, Zaimovic A, Mascetti CG, Garini S, Zambelli U, Timpano M, Raggi MA and Brambilla F. Neuroendocrine responses to experimentally-induced psychological stress in healthy humans. Psychoneuroendocrinology 2001; 26:91-107
96. Gilbeau PM, Almirez RG, Holaday JW and Smith CG. Opioid effects on plasma concentrations of luteinizing hormone and prolactin in the adult male rhesus monkey. J Clin Endocrinol Metab 1985; 60:299–305
97. Gilbey P and Spyer P. Essential organization of the sympathetic nervous system. Bailliere's Clin Endocrinol Metab 1993; 319:413–420.
98. Gilman S, Low PA, Quinn N, et al Consensus Statement on the diagnosis of MSA. Clin Auton Res 1998; 8(6):359-62
99. Gold MS, Redmond Jr DE and Donabedian RK. The effects of opiate agonist and antagonist on serum prolactin in primates: possible role for endorphins in prolactin regulation. Endocrinology 1979; 105:284–289
100. Gold PW and Chrousos GP. Organization of the stress system and its dysregulation in melancholic and atypical depression: high vs low CRH/NE states. Mol Psych 2002; 7:254-275
101. Gold PW and Chrousos GP. The endocrinology of melancholic and atypical depression: relation to neurocircuitry and somatic consequences. Proc Assoc Am Physicians 1999; 111:22–34
102. Gold PW, Goodwin F and Chrousos GP. Clinical and biochemical manifestations of depression: Relationship to the neurobiology of stress (Part II). N Eng J Med 1998; 319:413-420
103. Goldstein DS and McEwan B. Allostasis, homeostasis and the nature of stress. Stress2002; 5(1):555-8
104. Gorman JM, Askanazi J, Liebowitz MR, Fyer AJ, Stein J, Kinney JM and Klein DF. Response to hyperventilation in a group of patients with panic disorder. Am J Psychiatry 1984; 141(7):857-861
105. Gorman JM, Fyer MR, Goetz R, Askanazi J, Liebowitz MR, Fyer AJ, Kinney J and Klein DF. Ventilatory physiology of patients with panic disorder. Arch Gen Psych1988; 45:31-39
259
106. Gorman JM, Kent JM, Sullivan GM and Coplan JD. Neuroanatomical hypothesis of panic disorder, revised. Am J Psych 2000; 157:493-505
107. Gorman JM, Papp LA, Coplan JD, Martinez JM, Lennon S, Goetz RR, Ross D and Klein DF. Anxiogenic effects of CO2 and hyperventilation in patients with panic disorder. Am J Psych 1994; 151(4):547-553
108. Gorman JM. Does the brain noradrenaline network mediate the effects of the CO2 challenge? J Psychopharmacol 2003; 17(3):265-266
109. Gozal D, Marcus CL, Shoseyov D and Keens TG. Peripheral chemoreceptor function in children with the congenital central hypoventilation syndrome. J Appl Physiol1993; 74(1):379-387
110. Gozal D. Congenital central hypoventilation syndrome: an update. Pediatr Pulmonol1998; 26(4):273-282
111. Granados G, Garay-Sevilla ME, Malacara JM, Wrobel-Zasada K, Rivera-Cisneros A. Plasma epinephrine and norepinephrine response to stimuli in autonomic neuropathy of type 2 diabetes mellitus. Acta Diabetol 2000; 37(2):55-60
112. Greiz E, de Loof C, Pols H, Zandbergen J and Lousberg H. Specific sensitivity of patients with panic attacks to carbon dioxide inhalation. Psych Res 1990b; 31:193-199
113. Greiz E, Zandbergen J, Pols H and de Loof C. Response to 35% CO2 as a marker of panic in severe anxiety. Am J Psych 1990a; 145:796-797
114. Griez E and Schreurs K. Mechanisms of CO2 challenges. J Psychopharmacol 2003; 17(3):260-262
115. Griez EJ, Lousberg H, van den Hout MA and van der Molen GM. CO2 vulnerability in panic disorder. Psychiatry Res 1987; 20(2):87-95
116. Gritz ER, Shiffman SM, Jarvik ME, Schlesinger J and Charuvastra VC. Naltrexone: physiological and psychological effects of single doses. Clin Pharmacol Ther 1976; 19(6):773-776
117. Habib KE, Gold PW and Chrousos GP. Neuroendocrinology of stress. Endocrinol Metab Clin North Am 2001; 30:695-728
118. Habib KE, Weld KP, Rice KC, Pushkas J, Champoux M, Listwak S, Webster EL, Atkinson AJ, Schulkin J, Contoreggi C, Chrousos GP, McCann SM, Suomi SJ, Higley JD and Gold PW. Oral administration of a corticotropin-releasing hormone receptor antagonist significantly attenuates behavioural, neuroendocrine, and autonomic responses to stress in primates. Proc Natl Acad Sci 2000; 97:6079–6084.
119. Haddad GG, Mazza NM, Defendini R, Blanc WA, Driscoll JM, Epstein MA, Epstein RA and Mellins RB. Congenital failure of automatic control of ventilation, gastrointestinal motility and heart rate. Medicine 1978; 57(6):517-26
120. Hammer RP Jr and Bridges RS. Preoptic area opioids and opiate receptors increase during pregnancy and decrease during lactation. Brain Res 1987; 420(1):48-56
121. Harbuz MS and Lightman SL. Stress and the hypothalamo-pituitary-adrenal axis: acute, chronic and immunological activation. J Endocrinology 1992; 134:327-339
122. Hardgrove M, Roth GM and Brown GE. The pressor reaction produced by inhalation of carbon dioxide; studies of patients with normal blood pressure and with hypertension. Ann Intern Med 1938; 12:482-492
123. Harrington PJ, Schmidt NB and Telch MJ. Prospective evaluation of panic potentiation following 35% CO2 challenge in nonclinical subjects. Am J Psych 1996; 153(6):823-825
124. Hartmann P and Cregan M. Lactogenesis and the effects of insulin-dependent diabetes mellitus and prematurity. J Nutr 2001; 131(11):3016-3020
125. Haxhiu MA, Tolentino-Silva F, Pete G, Kc P and Mack SO. Monoaminergic neurons, chemosensation and arousal. Respir Physiol 2001; 129:191-209
126. Heinrichs M, Meinlschmidt G, Neumann I, Wagner S, Kirschbaum C, Ehlert U and Hellhammer DH. Effects of suckling on hypothalamic-pituitary-adrenal axis responses to psychosocial stress in postpartum lactating women. J Clin Endocrinol Metab 2001; 86(10):4798-4804
127. Heinrichs M, Neumann I and Ehlert U. Lactation and stress: protective effects of breast-feeding in humans. Stress 2002; 5(3):195-203
260
128. Henry RA. Lu IL, Beightol LA, and Eckberg DL. Interactions between CO2 chemoreflexes and arterial baroreflexes. Am J Physiol Heart Circ Physiol 1998; 274(6):H2177-H2187
129. Herman JP and Cullinan WE. Neurocircuitry of stress: central control of the hypothalamo-pituitary-adrenocortical axis. TINS 1997; 20(2):78-84
130. Hersh D, Van Kirk JR and Kranzler HR. Naltrexone treatment of comorbid alcohol and cocaine use disorders. Psychopharmacology 1998; 139(1-2):44-52
131. Higuchi T, Negoro H and Arita J. Reduced responses of prolactin and catecholamine to stress in the lactating rat. J Endocrinol 1989; 122(2):495-498
132. Hilsted J, Galbo H and Christensen NJ. Impaired responses of catecholamines, growth hormone, and cortisol to graded exercise in diabetic autonomic neuropathy. Diabetes 1980; 29(4):257-262
133. Hojo H, Takai N, Nishikawa Y and Yamaguchi M. Assessment of psychological stress by measurement of salivary amylase level. (Abstract) 32nd Annual Meeting of the American Association of Dental Research, San Antonio, Texas, 2003
134. Holsboer F. The corticosteroid receptor hypothesis of depression. Neuropsychopharmacol 2000; 23(5):477-501
135. Hunt PJ, Gurnell EM, Huppert FA, Richards C, Prevost AT, Wass JA, Herbert J and Chatterjee VK. Improvement in mood and fatigue after dehydroepiandrosterone replacement in Addison's disease in a randomized, double blind trial. J Clin Endocrinol Metab 2000; 85(12):4650-4656
136. American Thoracic Society. Idiopathic congenital central hypoventilation syndrome: diagnosis and management. Am J Respir Crit Care Med 1999; 160(1):368-373
137. Imai Y, Ito S, Maruta K and Fujita K. Simultaneous determination of catecholamines and serotonin by liquid chromatography, after treatment with boric acid gel. Clin Chem 1988; 34(3):528-530
138. Inder WJ, Ellis MJ, Evans MJ and Donald RA. A comparison of the naloxone test with ovine CRH and insulin hypoglycaemia in the evaluation of the hypothalamic-pituitary-adrenal axis in normal man. Clin Endocrinol 1995; 43(4):425-431
139. Iranmanesh A, Veldhuis JD, Carlsen EC, Vaccaro VA, Booth RA Jr, Lizarralde G, Asplin CM and Evans WS. Attenuated pulsatile release of prolactin in men with insulin-dependent diabetes mellitus. J Clin Endocrinol Metab 1990; 71(1):73-78
140. Ito T, Moore JI and Koss MC. Topical application of CO2 increases skin blood flow. J Invest Derm 1989; 93:259-262
141. Itoi K, Suda T, Tozawa F, Dobashi I, Ohmori N, Sakai Y, Abe K and Demura H. Microinjection of norepinephrine into the paraventricular nucleus of the hypothalamus stimulates corticotropin-releasing factor gene expression in conscious rats. Endocrinology 1994; 135(5):2177-2182
142. Janig W. The autonomic nervous system. In: Human Physiology (Schmidt RF and Thews G eds),1983, Springer-Verlag, New York
143. Jensen TH, Hansen PB and Brodersen P. Ondine's curse in listeria monocytogenes brain stem encephalitis. Acta Neurol Scand 1988; 77(6):505-506
144. Jeong KH, Jacobson L, Pacak K, Widmaier EP, Goldstein DS and Majzoub JA. Impaired basal and restraint-induced epinephrine secretion in corticotropin-releasing hormone-deficient mice. Endocrinoogyl 2000; 141(3):1142-1150
145. Jessop DJ, Harbuz MS and Lightman SL. CRH in chronic inflammatory stress. Peptides 2001; 22:803-807
146. Jezova D, Ochedalski M, Glickman M, Kiss A and Aguilera G. Central corticotropin-releasing hormone receptors modulate hypothalamic-pituitary-adrenocortical and sympathoadrenal activity during stress. Neuroscience 1999; 94(3):797-802
147. Johnson JM, Pérgola PE, Liao FK, Kellogg DL, Jr and Crandall CG. Skin of the dorsal aspect of human hands and fingers possess an active vasodilator system. J Appl Physiol 1995; 78:948-954
148. Johnston TG, Kelly CB, Stevenson MR and Cooper SJ. Plasma norepinephrine and prediction of outcome in major depressive disorder. Biol Psychiatry 1999; 46(9):1253-1258
261
149. Jones DNC, Kortekaas R, Slade PD, Middlemiss DN and Hagan JJ. The behavioural effects of corticotropin releasing factor-related peptides in rats. Psychopharmacology1998; 138:124-132
150. Jorgensen LS, Christiansen P, Raundahl U, Ostgaard S, Christensen NJ, Fenger M and Flachs H. Autonomic response to an experimental psychological stressor in healthy subjects: measurement of sympathetic, parasympathetic, and pituitary-adrenal parameters: test-retest reliability. Scand J Clin Lab Invest 1990; 50(8):823-829
151. Kc P, Haxhiu MA, Ovid Troud C, Balan KV, Anderson WA and Mack SO. CO2-induced c-Fos expression in hypothalamic vasopressin containing neurons. Respir Physiol 2002; 129:289-296
152. Kempler P, Tesfaye S, Chaturvedi N, Stevens LK, Webb DJ, Eaton S, Kerenyi Z, Tamas G, Ward JD and Fuller JH; EURODIAB IDDM Complications Study Group. Autonomic neuropathy is associated with increased cardiovascular risk factors: the EURODIAB IDDM Complications Study. Diabetic Med 2002; 19(11): 900-909
153. Kety S and Schmidt C. J Clin Invest 1948; 27:485 154. Kiecolt-Glaser JK, McGuire L, Robles TF and Glaser R. Emotions, morbidity and
mortality. New perspectives from psychoneuroimmunology. Ann Rev Psychol 2002; 53:83-107
155. Kimber J, Mathias CJ, Lees AJ, Bleasdale-Barr K, Chang HS, Churchyard A and Watson L. Physiological, pharmacological and neurohormonal assessment of autonomic function in progressive supranuclear palsy. Brain 2000; 123(7):1422-1430
156. Kimber J, Watson L and Mathias CJ. Cardiovascular and neurohormonal responses to i. v. l-arginine in two groups with primary autonomic failure. J Neurol 2001; 248(12):1036-1041
157. King AC, Schluger J, Gunduz M, Borg L, Perret G, Ho A and Kreek MJ. Hypothalamic-Pituitary-Adrenocortical (HPA) Axis Response and Biotransformation of Oral Naltrexone. Preliminary Examination of Relationship to Family History of Alcoholism. Neuropsychopharmacology 2002; 26(6):7787-7788
158. Kirschbaum C, Prussner JC, Federenko I, Gaab J, Lintz D, Schommer N and Hellhammer DH. Persistent high cortisol responses to related psychological stress in a subpopulation of healthy men. Psychosom Med 1995; 57(5):468-474
159. Kirschbaum C, Strasburger CJ and Langkrar J. Attenuated cortisol response to psychological stress but not to CRH or ergometry in young habitual smokers. Pharmacol Biochem Behav 1993; 44(3):527-531
160. Klaasen T, Klumperbeek J, Deutz NEP, Van Praag HM and Griez E. Effects of tryptophan depletion on anxiety and on panic provoked by carbon dioxide challenge. Psychiatry Res 1998; 77:167-174
161. Klein D. False suffocation alarms, spontaneous panics, and related conditions. An integrative hypothesis. Arch Gen Psych 1993; 50:306-317
162. Knol BW. Stress and the hypothalamus-pituitary-testis system: A review. Vet Q 1991; 13:104-114
163. Koob GF. Corticotropin releasing factor, norepinephrine and stress. Biol Psychiatry1999; 46:1167-1180
164. Koszycki D, Bradwejn J and Bourin M. Comparison of the effects of cholecystokinin-tetrapeptide and carbon dioxide in health volunteers. Eur Neuropsychopharmacol1991; 1(2):137-141
165. Krantz DS and McCeney MK. Effects of psychological and social factors on organic disease: A critical assessment of research on coronary heart disease. Ann Rev Psychol2002; 53:341-369
166. Kreek MJ, Schluger J, Borg L, Gunduz M and Ho A. Dynorphin A1-13 causes elevation of serum levels of prolactin through an opioid receptor mechanism in humans: gender differences and implications for modulation of dopaminergic tone in the treatment of addictions. J Pharmacol Exp Ther 1999; 288(1):260-269
167. Krystal JH, Woods SW, Levesque M, Heninger C and Heninger G. The effects of carbon dioxide inhalation on plasma MHPG, plasma hormones, respiratory rate, and behaviour in the rhesus monkey. Life Sci 1989; 45:1657-1663
262
168. Kvetnansky R, Pacak K, Fukuhara K, Viskupic E, Hiremagalur B, Nankova B, Goldstein DS, Sabban EL and Kopin IJ. Sympathoadrenal system in stress. Interaction with the hypothalamic-pituitary-adrenocortical system. Ann N Y Acad Sci 1995; 771:131-158
169. La Verne AA. Rapid coma technique of carbon dioxide inhalation therapy. Dis Nervous Sys 1953; 14:141-144
170. Laborie C, van Camp G, Bernet F, Montel V and Dupouy JP. Metyrapone-Induced Glucocorticoid Depletion Modulates Tyrosine Hydroxylase and Phenylethanolamine N-Methyltransferase Gene Expression in the Rat Adrenal Gland by a Noncholinergic Transsynaptic Activation. J Neuroendocrinol 2003; 15(1):15-23
171. Lahiri S and Forster RE II. CO2/H(+) sensing: peripheral and central chemoreception. Int J Biochem Cell Biol 2003; 35(10):1413-1435
172. Lamberts SW, Bons EG, Bruining HA and de Jong FH. Differential effects of the imidazole derivatives etomidate, ketoconazole and miconazole and of metyrapone on the secretion of cortisol and its precursors by human adrenocortical cells. J Pharmacol Exp Ther 1987; 240(1):259-264
173. Lange M, Svendsen OL, Skakkebaek NE, Muller J, Juul A, Schmeiglow and Felt-Rasmussen U. An audit of the insulin-tolerance test in 255 patients with pituitary disease. Eur J Endocrinol 2002; 147:41-47
174. Langewouters GJ, Settels JJ, Roelandt R and Wesseling KH. Why use Finapres or Portapres rather than intra-arterial or intermittent non-invasive techniques of blood pressure measurement? Medical Engineering Technology 1998; 22:37-43
175. Lejuez CW, Forsyth JP and Eifert GH. Devices and methods for administering carbon dioxide-enriched air in experimental and clinical settings. J Behav Ther Exp Psychiatry 1998; 29(3):239-248
176. Leonard BE. Noradrenaline in basic models of depression. Eur Neuropsychopharmacol 1997; 7(S1):1-6
177. Levine S. Influence of psychological variables on the activity of the hypothalamic-pituitary-adrenal axis. Eur J Pharmacol 2000; 405:149-160
178. Lightman SL and Young WS III. Lactation inhibits stress-mediated secretion of corticosterone and oxytocin and hypothalamic accumulation of corticotropin-releasing factor and enkephalin messenger ribonucleic acids. Endocrinology 1989; 124(5):2358-2364
179. Lightman SL, Windle RJ, da Costa APC, Shanks N and Ingram C. Lactation: A physiological model of stress hyporesponsiveness of the neuroendocrine system. In: New fronteirs in stress research: Modulation of brain function. (Levy A, Grauer E, Ban-Nathan D and de Kloet ER eds), 1997; pp 59-71, Harwood, Amsterdam.
180. Lightman SL, Windle RJ, Wood SA, Kershaw YM, Shanks N and Ingram CD. Peripartum plasticity within the hypothalamo-pituitary-adrenal axis. Prog Brain Res2001; 133:111-129
181. Lightman SL. Alterations in hypothalamic-pituitary responsiveness during lactation. Ann NY Acad Sci 1993; 689:340-346
182. Linton EA, Tilders FJ, Hodgkinson S, Berkenbosch F, Vermes I and Lowry PJ. Stress-induced secretion of adrenocorticotropin in rats is inhibited by administration of antisera to ovine corticotropin-releasing factor and vasopressin. Endocrinology1985; 116(3):966-970
183. Liotti M, Brannan S, Egan G, Shade R, Madden L, Abplanalp B, Robillard R, Lancaster J, Zamarripa F, Fox P and Denton D. Brain responses associated with consciousness of breathlessness (air hunger). PNAS 2001; 98(4):2035-2040
184. Loubinoux I, Pariente J, Boulanouar K, Carel C, Manelfe C, Rascol O, Celsis P and Chollet F. A single dose of the serotonin neurotransmission agonist paroxetine enhances motor output: double-blind, placebo-controlled, fMRI study in healthy subjects. Neuroimage 2002; 15(1):26-36
185. Lovas K, Loge JH and Husebye ES. Subjective health status in Norwegian patients with Addison's disease. Clin Endocrinol 2002; 56(5):581-588
186. Lowry CA. Functional subsets of serotonergic neurones: implications for control of the hypothalamic-pituitary-adrenal axis. J Neuroendocrinol 2002; 14(11):911-923
263
187. Ma X, Levy A and Lightman SL. Emergence of an isolated arginine vasopressin (AVP) response to stress after repeated restraint: A study of both AVP and corticotropin-releasing hormone messenger ribonucleic acid (RNA) and heteronuclear RNA. Endocrinoogyl 1997; 138:4351-4357
188. Ma XM and Lightman SL. The arginine vasopressin and corticotrophin-releasing hormone gene transcription responses to varied frequencies of repeated stress in rats. J Physiol 1998; 510(2):605-614
189. Ma XM, Lightman SL and Aguilera G. Vasopressin and corticotrophin-releasing hormone gene responses to novel stress in rats adapted to repeated restraint. Endocrinology 1999; 140:3623-3632
190. Malarkey WB, Lipkus IM and Cacioppo JT. The dissociation of catecholamine and hypothalamic-pituitary-adrenal responses to daily stressors using dexamethasone. J Clin Endocrinol Metab 1995; 80(8):2458-2463
191. Mansi JA, Laforest S and Drolet G. Effect of stress exposure on the activation pattern of enkephalin-containing perikarya in the rat ventral medulla. J Neurochem 2000; 74(6):2568-2575
192. Mansour A, Fox CA, Akil H and Watson SJ. Opioid-receptor mRNA expression in the rat CNS: anatomical and functional implications. TINS 1995; 18(1):22-29
193. Mason JW. A review of psychoneuroendocrine research on the pituitary-adrenal cortical system. Psychosom Med 1968; 30:576-607
194. Mathe G. The need of a physiologic and pathophysiologic definition of stress. Biomed & Pharmacother 2000; 54:119-121
195. Mathias CJ and Bannister R Autonomic Failure - A Textbook of Clinical Disorders of the Autonomic Nervous System, Fourth Edition, 2002, Oxford University Press
196. McCarty R. Regulation of plasma catecholamine response to stress. Semin Neurosciences 1994; 6:197-204
197. McCaul ME, Wand GS, Stauffer R, Lee SM and Rohde CA. Naltrexone dampens ethanol-induced cardiovascular and hypothalamic- pituitary-adrenal axis activation. Neuropsychopharmacology 2001; 25(4):537-547
198. McCubbin JA, Surwit RS and Williams RB. Opioid dysfunction and risk for hypertension: Naloxone and blood pressure responses during different types of stress. Psychosom Med 1988; 50:8-14
199. McCubbin JA. Stress and endogenous opioids: behavioral and circulatory interactions. Biol Psychol 1993; 35(2):91-122
200. McEwen B. Introduction: Stress and the nervous system. Sem Neurosci 1994; 6:195-196
201. McEwen BS. Protective and damaging effects of stress mediators. N Eng J Med1998; 338(3):171-179
202. McEwen BS. Protective and damaging effects of stress mediators: the good and bad sides of the response to stress. Metabolism 2002; 51(6 S1):2-4
203. McEwen BS. The neurobiology of stress: from serendipity to clinical relevance. Brain Res 2000; 886:172-189
204. Meiri G, Ben-Zion I, Greenberg B, Murphy D and Benjamin J. Influence of the serotonin antagonist, metergoline, on the anxiogenic effects of carbon dioxide, and on heart rate and neuroendocrine measures, in healthy volunteers. Hum Psychopharmacol Clin Exp 2001; 16:237-245
205. Mello NK, Mendelson JH, Bree MP and Skupny A. Naltrexone effects on pituitary and gonadal hormones in male and female rhesus monkeys. Pharmacol Biochem. Behav 1989; 31:683–691
206. Meyer C, Grossmann R, Mitrakou A, Mahler R, Veneman T, Gerich J and BretzelRG. Effects of autonomic neuropathy on counterregulation and awareness of hypoglycemia in type 1 diabetic patients. Diabetes Care 1998; 21(11):1960-1966
207. Miller DB and O'Callaghan JP. Neuroendocrine aspects of the response to stress. Metabolism 2002; 51(6 S1):5-10
208. Miller HE, Deakin JF and Anderson IM. Effect of acute tryptophan depletion on CO2-induced anxiety in patients with panic disorder and normal volunteers. Br J Psychiatry2000; 176:182–188
264
209. Miller JA, Leckie BJ, Morton JJ, Jordan J and Tree M. A micro assay for active and total renin concentration in human plasma based on antibody trapping. Clinica Chimica Acta 1980; 101:5-15
210. Morrison WE, Haas EC, Shaffner DH, Garrett ES and Fackler JC. Noise, stress and annoyance in a pediatric intensive care unit. Crit Care Med 2003; 31(1):113-119
211. Mortimore C and Anderson IM. d-Fenfluramine in panic disorder: a dual role for 5-hydroxytryptamine. Psychopharmacology 2000; 149(3):251-258
212. Mukhopadhyay S and Wilkinson PW. Cerebral arteriovenous malformation, Ondine's curse and Hirschsprung's disease. Dev Med Child Neurol 1990; 32(12):1087-1089
213. Munck A and Naray-Fejes-Toth A. Glucocorticoids and stress: permissive and suppressive actions. Ann N Y Acad Sci 1994; 746:115-130
214. Musselman DL and Nemeroff CB. Depression and endocrine disorders: focus on the thyroid and adrenal system. Br J Psychiatry 1996; 30:123-128
215. Nardi A, Valenca A, Nascimento I, Mezzasalma M and Zin W. Double-blind acute clonazepam vs. placebo in carbon dioxide-induced panic attacks. Psych Res 2000; 94:179-184
216. Nater UM, Rohleder N, Gaab J, Berger S, Jud A, Kirschbaum C and Ehlert U. Human salivary alpha-amylase as an indicator in psychosocial stress reaction. Poster presented at the 5th International Congress of Neuroendocrinology, Bristol, UK 2002
217. Nattie E. CO2, brainstem chemoreception and breathing. Prog Neurobiol 1999; 59:299-331
218. Nechmad A, Maayan R, Spivak B, Ramadan E, Poyurovsky and M, Weizman A. Brain neurosteroid changes after paroxetine administration in mice. Eur Neuropsychopharmacol 2003; 13(5):327-332
219. Negrao AB, Deuster PA, Gold PW, Singh A and Chrousos GP. Individual reactivity and physiology of the stress response. Biomed Pharmacother 2000; 54:122-128
220. Neumann ID, Johnstone HA, Hatzinger M, Liebsch G, Shipston M, Russell JA, Landgraf R and Douglas AJ. Attenuated neuroendocrine responses to emotional and physical stressors in pregnant rats involve adenohypophysial changes. J Physiol 1998; 508:289-300
221. Neumann ID, Torner L and Wigger A. Brain oxytocin: differential inhibition of neuroendocrine stress responses and anxiety-related behaviour in virgin, pregnant and lactating rats. Neuroscience 2000; 95(2):567-575
222. Nishimura M, Miyamoto K, Suzuki A, Yamamoto H, Tsuji M, Kishi F and Kawakami Y. Ventilatory and heart rate responses to hypoxia and hypercapnia in patients with diabetes mellitus. Thorax 1989; 44(4):251-257
223. Noble RE. Diagnosis of stress. Metabolism 2002; 51(6 S1):37-39 224. Nutt DJ, Glue P, Lawson C and Wilson S. Flumazenil provocation of panic attacks.
Arch Gen Psych 1990; 47:917-925 225. Nye EJ, Grice JE, Hockings GI, Strakosch CR, Crosbie GV, Walters MM, Torpy DJ
and Jackson RV. The Insulin Hypoglycemia Test: Hypoglycemic Criteria and Reproducibility. J Neuroendocrinol 2001; 13(6):524-530
226. O’Connor TM, O’Halloran DJ and Shanahan F. The stress response and the hypothalamic-pituitary-adrenal axis: from molecule to melancholia. Q J Med 2000; 93:323-333
227. O'Brien IA, McFadden JP, Corrall RJ. The influence of autonomic neuropathy on mortality in insulin-dependent diabetes. Q J Med 1991; 79(290): 495-502
228. Olson GA, Olson RD and Kastin AJ. Endogenous opiates: 1995. Peptides 1996; 17(8):1421-1466
229. Oncken C, Van Kirk J and Kranzler HR. Adverse effects of oral naltrexone: analysis of data from two clinical trials. Psychopharmacology 2001; 154(4):397-402
230. O'Sullivan SE and Bell C. Training reduces autonomic cardiovascular responses to both exercise-dependent and -independent stimuli in humans. Auton Neurosci 2001; 91(1-2):76-84
231. Owens MJ and Nemeroff CB. Physiology and pharmacology of Corticotropinreleasing factor. Pharmacol Rev 1991; 43:425-473
265
232. Pacak K and Miklos P. Stressor specificity of central neuroendocrine responses: Implications for stress-related disorders. Endocrinol Rev 2001; 22(4):502-548
233. Pacak K, Miklos P, Yadid G, Kvetnansky R, Kopin I and Goldstein D. Heterogenous neurochemical responses to different stressors: a test of Selye’s doctrine of nonspecificity. Am J Physiol 1998; 44:R1247-R1255
234. Pacak K, Palkovits M, Kopin IJ and Goldstein DS. Stress-induced norepinephrine release in the hypothalamic paraventricular nucleus and pituitary-adrenocortical and sympathoadrenal activity: In vivo microdialysis studies. Fron Neuroendocrin 1995; 16:89-150
235. Papp LA, Martinez JM, Klein DF, Coplan JD, Norman RG, Cole R, de Jesus MJ, Ross D, Goetz R and Gorman JM. Respiratory psychophysiology of panic disorder: Three respiratory challenges in 98 subjects. Am J Psychiatry 1997; 154:1557-1565
236. Parikh SM, Diedrich A, Biaggioni I and Robertson D. The nature of the autonomic dysfunction in multiple system atrophy. J Neurol Sci 2002; 200(1-2):1-10
237. Pascualy M, Petrie EC, Brodkin K, Peskind ER, Wilkinson CW and Raskind MA. Hypothalamic pituitary adrenocortical and sympathetic nervous system responses to the cold pressor test in Alzheimer's disease. Biol Psychiatry 2000; 48(3):247-254
238. Pedersen AM, Bardow A, Jensen SB and Nauntofte B. Saliva and gastrointestinal functions of taste, mastication, swallowing and digestion. Oral Dis 2002; 8(3):117-129
239. Penttila J, Syvalahti E, Hinkka S, Kuusela T and Scheinin H. The effects of amitriptyline, citalopram and reboxetine on autonomic nervous system. A randomised placebo-controlled study on healthy volunteers. Psychopharmacology 2001; 154(4):343-349
240. Perna G, Barbini B, Cocchi S, Bertani A and Gasperini M. 35% CO2 challenge in panic and mood disorders. J Affective Disord 1995a; 33;189–194
241. Perna G, Battaglia M, Garberi A, Arancio C, Bertani A and Bellodi L. 35% CO2/65% O2 inhalation test in panic patients. Psych Res 1994; 52:159-171
242. Perna G, Bertani A, Arancio C, Ronchi P and Bellodi L. Laboratory response of patients with panic and obsessive–compulsive disorders to 35% CO2 challenges. Am J Psych 1995b; 152:85–89.
243. Perna G, Bertani A, Caldirola D, Gabriele A, Cocchi S and Bellodi L. Antipanic drug modulation of 35% CO2 hyperreactivity and short-term treatment outcome. J Clin Psychopharmacol 2002; 22(3):300-308
244. Perna G, Bertani A, Gabrielle A, Politi E and Bellodi L. Modification of 35% carbon dioxide hypersensitivity across 1week of treatment with clomipramine and fluvoxamine: a double-blind, randomised, placebo-controlled study. J Clin Psychopharmacol 1997; 17:173-178
245. Perna G, Cocchi S, Allevi L, Bussi R and Bellodi L. A long-term prospective evaluation of first-degree relatives of panic patients who underwent the 35% CO2 challenge. Biol Psychiatry 1999; 45(3):365-367
246. Perna G, Cocchi S, Bertani A, Arancio C and Bellodi L. Sensitivity to 35% CO2 in healthy first-degree relatives of patients with panic disorder. Am J Psychiatry 1995c; 152(4):623-625
247. Perna G, Gabriele A, Caldirola D and Bellodi L. Hypersensitivity to in halation of carbon dioxide and panic attacks. Psych Res 1995e; 152:1081-1083
248. Perna G, Gabriele A, Caldirola D and Bellodi L. Hypersensitivity to inhalation of carbon dioxide and panic attacks. Psych Res 1995d; 52:267–273
249. Peskind ER, Jensen CF, Pascualy M, Tsuang D, Cowley D, Martin DC, Wilkinson CW and Raskind MA. Sodium lactate and hypertonic sodium chloride induce equivalent panic incidence, panic symptoms, and hypernatremia in panic disorder. Biol Psychiatry 1998; 44(10):1007-1016
250. Petrides JS, Mueller GP, Kalogeras KT, Chrousos GP, Gold PW and Deuster PA. Exercise-induced activation of the hypothalamic-pituitary-adrenal axis: marked differences in the sensitivity to glucocorticoid suppression. J Clin Endocrinol Metab1994; 79(2):377-383
266
251. Pine DS, Weese-Mayer DE, Silvestri JM, Davies M, Whitaker AH and Klein DF. Anxiety and congenital central hypoventilation syndrome. Am J Psychiatry 1994; 151(6):864-870
252. Polinsky RJ, Brown RT, Curras MT, Baser SM, Baucom CE, Hooper DR and Marini AM. Central and peripheral effects of arecoline in patients with autonomic failure. J Neurol Neurosurg Psychiatry 1991; 54(9):807-812
253. Pols H, Griez E, Verburg K and van der Werf D. Yohimbine premedication and 35% CO2 vulnerability in healthy volunteers. Eur Arch Psychiatry Clin Neurosci 1994; 244(2):81-85
254. Pols H, Lousberg H, Zandbergen J and Griez E. Panic disorder patients show decrease in ventilatory response to CO2 after clomipramine treatment. Psych Res 1993; 47(3):295-296
255. Pols H, Verburg K, Hauzer R, Meijer J and Griez E. Alprazolam premedication and 35% carbon dioxide vulnerability in panic patients. Biol Psychiatry 1996b; 40(9):913-917
256. Pols H, Zandbergen J, de Loof C and Griez E. Attenuation of carbon dioxide-induced panic after clonazepam treatment. Acta Psychiatr Scand 1991; 84(6):585-586
257. Pols HJ, Hauzer RC, Meijer JA, Verburg K and Griez EJ. Fluvoxamine attenuates panic induced by 35% CO2 challenge. J Clin Psychiatry 1996a; 57(11):539-542
258. Pomara N, Shao B, Choi SJ, Tun H and Suckow RF. Sex-related differences in nortriptyline-induced side-effects among depressed patients. Prog Neuropsychopharmacol Biol Psychiatry 2001; 25(5):1035-1048
259. Porges SW, Riniolo TC, McBride T and Campbell B. Heart rate and respiration in reptiles: contrasts between a sit-and-wait predator and an intensive forager. Brain Cogn 2003; 52(1):88-96
260. Porges SW. Orienting in a defensive world: mammalian modifications of our evolutionary heritage. A Polyvagal Theory. Psychophysiology 1995; 32(4):301-318
261. Raggi MA, Sabbioni C, Casamenti G, Gerra G, Calonghi N and Masotti L. Determination of catecholamines in human plasma by high-performance liquid chromatography with electrochemical detection. J Chromatogr B Biomed Sci Appl1999; 730(2):201-211
262. Rammana Reddy SV, Yaksh TL, Anderson RE and Shundt Jr TM. Effect in cat locus coeruleus lesions on the response of cerebral blood flow and cardiac output to altered paCO2. Brain Res 1986; 365:278-288
263. Rapee RM, Brown TA, Antony MM and Barlow DH. Response to hyperventilation and inhalation of 5.5% carbon dioxide-enriched air across the DSM-III-R anxiety disorders. J Abnorm Psychol 1992; 101(3):538-552
264. Rassovsky Y and Kushner MG. Carbon dioxide in the study of panic disorder: issues of definition, methodology, and outcome. J Anxiety Disord. 2003; 17(1):1-32
265. Read DJC. A clinical method for assessing the ventilatory response to carbon dioxide. Australasian Ann Med 1967; 16:20-32
266. Redwine LS, Altemus M, Leong YM, and Carter CS. Lymphocyte responses to stress in postpartum women: relationship to vagal tone. Psychoneuroendocrinol 2001; 26:241-251
267. Reed SF, Ohel G, David R and Porges SW. A neural explanation of fetal heart rate patterns: A test of the polyvagal theory. Dev Psychobiol 1999; 35(2):108-118
268. Refshauge C, Kissinger PT, Dreiling R, Blank L, Freeman R and Adams RN. New high performance liquid chromatographic analysis of brain catecholamines. Life Sci1974; 14(2):311-322
269. Reichlin S. Neuroendocrinology in Williams Textbook of Endocrinology. 9th Edition (Wilson et al. ed), 1998, WB Saunders Co, Philadelphia
270. Reid IA. Interactions between ANG II, sympathetic nervous system and baroreceptor reflexes in regulation of blood pressure. Am J Physiol 1992; 262:763-778
271. Ressler KJ and Nemeroff CB. Role of norepinephrine in the pathophysiology and treatment of mood disorders. Biol Psychiatry 1999; 46:1219-1233
267
272. Richer SD, Schurmeyer TH, Schedlowski M, Hadicke A, Tewes U, Schmidt E and Wagner T. Time kinetics of the endocrine response to psychological stress. J Clin Endocrin Metab 1996; 81(5):1956-1960
273. Riedel M, Wiese A, Schurmeyer TH and Brabant G. Quality of life in patients with Addison's disease: effects of different cortisol replacement modes. Exp Clin Endocrinol 1993; 101(2):106-111
274. Rosen MI, Kosten TR and Kreek MJ. The effects of naltrexone maintenance on the response to yohimbine in healthy volunteers. Biol Psychiatry 1999; 45(12):1636-1645
275. Rosmond R, Dallman M and Björntorp P. Stress related cortisol secretion in men: Relationship with abdominal obesity, endocrine, metabolic and hemodynamic abnormalities. J Clin Endocrinol Metab 1998; 83:1853-1859
276. Rozanski A, Blumenthal JA and Kaplan J. Impact of psychological factors on the pathogenesis of cardiovascular disease and implications for therapy. Circulation1999; 99:2192-2217
277. Rubin P, Swezey S and Blaschke T. Naloxone lowers plasma-prolactin in man. Lancet1979; 1:1293
278. Russell JA, Douglas AJ and Ingram CD. Brain preparations for maternity-adaptive changes in behavioral and neuroendocrine systems during pregnancy and lactation. An overview. Prog Brain Res 2001; 133:1-38
279. Saad AR, Stephens DP, Bennettt LAT, Charkoudian N, Kosiba WA, Johnson JM. Influence of isometric exercise on blood flow and sweating in glabrous and nonglabrous human skin. J Appl Physiology 2001; 91:2487-2492
280. Sanderson WC and Wetzler S. Five percent carbon dioxide challenge: valid analogue and marker of panic disorder? Biol Psychiatry 1990; 27(7):689-701
281. Sapolsky RM, Romero LM and Munck AU. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocrine Rev 2000; 21(1):55-89
282. Sapolsky RM. Individual differences and the stress response. Semin Neurosciences1994; 6:261-269
283. Schlaefke ME, Kille JF and Loeschcke HH. Elimination of central chemosensitivity by coagulation of a bilateral area on the ventral medullary surface in awake cats. Pflugers Arch 1979; 378(3):231-241
284. Schruers K, Klaassen T, Pols H, Overbeek T, Deutz NE and Griez E. Effects of tryptophan depletion on carbon dioxide provoked panic in panic disorder patients. Psych Res 2000a; 93:179–187
285. Schruers K, Pols H, Overbeek T and Griez E. 10 microg CCK-4 premedication and 35% CO2 challenge in healthy volunteers. Prog Neuropsychopharmacol Biol Psychiatry 2000b; 24(3):409-418
286. Schruers K, van Diest R, Overbeek T and Griez E. Acute -5-hydroxytryptophan administration inhibits carbon dioxide-induced panic in panic disorder patients. Psych Res 2002; 113:237-243
287. Scott LV and Dinan TG. Vasopressin and the regulation of hypothalamo-pituitary-adrenal axis function: implications for the pathophysiology of depression. Life Sci 1998; 62(22):1985-1998
288. Sechzer PH, Egbert LD, Linde HW, Cooper DY, Dripps RD and Price HL. Effect of CO2 inhalation on arterial pressure, ECG and plasma catecholamines and 17-OH corticosteroids in normal man. J Appl Physiol 1960; 15:454-458
289. Selye H. The stress of life. New York: McGraw-Hill; 1956 290. Sephton SE, Sapolsky RA, Kraemer HC and Spiegel D. Diurnal cortisol rhythm as a
predictor of breast cancer survival. J Natl Cancer Inst 2000; 92(12):994-1000 291. Severinhaus JW and Mitchell RA. Ondine’s curse: Failure of respiratory centre
automaticity while awake. Clin Res 1962; 10:122 292. Sgoutas-Emch SA, Cacioppo JT, Uchino BN, Malarkey W, Pearl D, Kiecolt-Glaser J,
et al. The effects of an acute psychological stressor on cardiovascular, endocrine and cellular immune response: A prospective study of individuals high and low in heart rate reactivity. Psychophysiol 1994; 31:264-271
268
293. Shanks N, Kusnecov A, Pezzone M, Berkun J and Rabin BS. Lactation alters the effects of conditioned stress on immune function. Am J Physiol 1997; 272:R16-25
294. Shanks N, Windle RJ, Perks P, Wood S, Ingram CD and Lightman SL. The hypothalamic-pituitary-adrenal axis response to endotoxin is attenuated during lactation. J Neuroendocrinol 1999; 11(11):857-865
295. Shannon JR, Jordan J, Diedrich A, Pohar B, Black BK, Robertson D and Biaggioni I. Sympathetically mediated hypertension in autonomic failure. Circulation 2000; 101(23):2710-2715
296. Silvestri JM, Hanna BD, Volgman AS, Jones PJ, Barnes SD and Weese-Mayer DE. Cardiac rhythm disturbances among children with idiopathic congenital central hypoventilation syndrome. Pediatr Pulmonol 2000; 29(5):351-358
297. Singh A, Petrides JS, Gold PW, Chrousos GP and Deuster PA. Differential hypothalamic-pituitary-adrenal axis reactivity to psychological and physical stress. J Clin Endocrinol Metab 1999; 84(6):1944-1948
298. Sinha SS, Coplan JD, Pine DS, Martinez JA, Klein DF and Gorman JM. Panic induced by carbon dioxide and lack of hypothalamic-pituitary-adrenal axis activation. Psych Res 1999; 86:93-98
299. Smedes F, Kraak JC and Poppe H. Simple and fast solvent extraction system for selective and quantitative isolation of adrenaline, noradrenaline and dopamine from plasma and urine. J Chromatogr 1982; 231(1):25-39
300. Speigel D, Sephton SE, Terr AI and Stites DP. Effects of psychosocial treatment in prolonging cancer survival may be mediated by neuroimmune pathways. Ann NY Acad Sci 1998; 840:674-683
301. Spengler CM, Gozal D and Shea SA. Chemoreceptive mechanisms elucidated by studies of congenital central hypoventilation syndrome. Respir Physiol 2001; 129(1-2):247-255
302. Stahl SM. Essential Psychopharmacology. 1996; Cambridge University Press, New York.
303. Stahl SM. Mechanism of action of serotonin selective reuptake inhibitors. Serotonin receptors and pathways mediate therapeutic effects and side effects. J Affect Disord1998; 51:215–235
304. Stedman TL. Stedman’s Medical Dictionary (ed 25). 1990; pp1487, Williams and Wilkins, Baltimore
305. Stella P, Ellis D, Maser RE and Orchard TJ. Cardiovascular autonomic neuropathy (expiration and inspiration ratio) in type 1 diabetes. Incidence and predictors. J Diabetes Complications 2000; 14(1):1-6
306. Sternberg EM, Chrousos GP, Wilder RL and Gold PW. The stress response and the regulation of inflammatory disease. Ann Intern Med 1992b; 117(10): 854-866
307. Sternberg EM, Glowa JR, Smith MA, Calogero AE, Listwak SJ, Aksentijevich S, Chrousos GP, Wilder RL and Gold PW. Corticotropin releasing hormone related behavioral and neuroendocrine responses to stress in Lewis and Fischer rats. Brain Res 1992a; 570(1-2):54-60
308. Sternberg EM. Emotions and disease: from balance of humors to balance of molecules. Nat Med 1997; 3:264-267
309. Stewart PM. The adrenal cortex. In: Williams Textbook of Endocrinology (Larsen PR, Kronenberg HM, Melmed S and Polosky KS eds). 10th Edition 2003; WB Saunders Co. Philadelphia
310. Taha BH, Simon PM, Dempsey JA, Skatrud JB and Iber C. Respiratory sinus arrhythmia in humans: an obligatory role for vagal feedback from the lungs. J Appl Physiol 1995; 78(2):638-645
311. Tantucci C, Bottini P, Fiorani C, Dottorini ML, Santeusanio F, Provinciali L, Sorbini CA and Casucci G. Cerebrovascular reactivity and hypercapnic respiratory drive in diabetic autonomic neuropathy. J Appl Physiol 2001; 90(3): 889896
312. Templ E, Koeller M, Riedl M, Wagner O, Graninger W and Luger A. Anterior pituitary function in patients with newly diagnosed rheumatoid arthritis. Br J Rheumatol 1996; 35(4):350-356
269
313. Tenney SM. Sympatho-adrenal stimulation by carbon dioxide and the inhibitory effect of carbonic acid on epinephrine response J Physiol 1956; 187:341
314. Tenney SM. The effect of carbon dioxide on neurohumoral and endocrine mechanisms. Anesthesiology 1960; 21:674-685
315. Thews G. Pulmonary respiration. In: Human Physiology (Schmidt RF and Thews G eds), 1983, Springer-Verlag, New York
316. Thorner MO, Vance ML, Laws ER Jr, Horvath E and Kovacs K. The Anterior Pituitary in Williams Textbook of Endocrinology. (Wilson et al. ed), 9th Edition, 1998, WB Saunders Co, Philadelphia
317. Torner L, Toschi N, Nava G, Clapp C and Neumann ID. Increased hypothalamic expression of prolactin in lactation: involvement in behavioural and neuroendocrine stress responses. Eu J Neurosci 2002; 15:1381-1389
318. Torner L, Toschi N, Pohlinger A, Landgraf R and Neumann ID. Anxiolytic and anti-stress effects of brain prolactin: Improved efficacy of antisense targeting of the prolactin receptor by molecular modelling. J Neurosci 2001; 21(9):3207-3214
319. Torpy DJ, Grice JE, Hockings GI, Crosbie GV, Walters MM and Jackson RV. The effect of desipramine on basal and naloxone-stimulated cortisol secretion in humans: interaction of two drugs acting on noradrenergic control of adrenocorticotropin secretion. J Clin Endocrinol Metab 1995; 80(3):802-806
320. Trachiotis GD, Knight SR, Pohl MS, Patterson GA, Cooper JD and Trulock EP. Tidal volume and respiratory rate changes during CO2 rebreathing after lung transplantation. Ann Thorac Surg 1994; 58(6):1718-1720
321. Tsigos C and Chrousos GP. Hypothalamic-pituitary-adrenal axis, neuroendocrine factors and stress. J Psychosom Res 2002; 53:865-871
322. Tucker HA. Lactation and its control. In: The physiology of reproduction. (Knobil E and Neill JD eds), 1994, pp1065-1098, Raven Press, New York
323. Ursin H. The psychology in psychneuroendocrinology. Psychoneuroendocrinol 1998; 23(6):555-570
324. Vale WW, Spiess S, Rivier C and Rivier J. Characterization of a 41-residue ovine hypothalamic peptide that stimulates secretion of corticotropin and beta-endorphin. Science 1981; 213:1394–1397
325. Van de Kar LD and Blair ML. Forebrain pathways mediating stress-induced hormone secretion. Front Neuroendocrinol 1999; 20:1-48
326. Van den Hout MA, Griez E, Van der Molen GM and Lousberg H. Pulmonary carbon dioxide and panic-arousing sensations after 35% carbon dioxide inhalations. J Behav Ther Exp Psychiatry 1987; 18(1):19-23
327. Van den Hout MA and Griez E. Panic symptoms after inhalation of carbon dioxide. Br J Psychiatry 1984; 144:503-507
328. Van Duinen M, Schreurs K, Jaegers E, Maes M and Griez E. HPA-axis function following 35% CO2 inhalation in healthy volunteers. Prog Neuropsychopharmacol Biol Psychiatry 2004; 28(2):279-283
329. Van Eck MM, Nicolson NA, Berkhof H and Sulon J. Individual differences in cortisol responses to a laboratory speech task and their relationship to responses to stressful daily events. Biol Psychol 1996; 43(1):69-84
330. Vanltallie TB. Stress: A risk factor for serious illness. Metabolism 2002; 51(6):40-45 331. Velasco M, Gomez J, Blanco M and Rodriguez I. The cold pressor test:
pharmacological and therapeutic aspects. Am J Ther 1997; 4(1):34-38 332. Verberg K, Perna G and Griez EJL. A case study of the 35% CO2 challenge. In:
Anxiety Disorders (Griez EJL, Faravelli C, Nutt D and Zohar D eds) 2001, John Wiley and Sons Ltd
333. Verberg K, Pols H, de Leeuw M and Griez E. Reliability of the 35% carbon dioxide panic provocation challenge. Psych Res 1998; 78:207-214
334. Verburg K, Griez E, Meijer J and Pols H. Discrimination between panic disorder and generalized anxiety disorder by 35% carbon dioxide challenge. Am J Psychiatry 1995; 152(7):1081-1083
270
335. Volavka J, Mallya A, Bauman J, Pevnick J, Cho D, Reker D, James B and Dornbush R. Hormonal and other effects of naltrexone in normal men. In: Modulators, Mediators, and Specifiers in Brain Function. (Ehrlich et al eds) 1979, Plenum, New York
336. Vythilingam M, Anderson GM, Owens MJ, Halaszynski TM, Bremner JD, Carpenter LL, Heninger GR, Nemeroff CB and Charney DS. Cerebrospinal fluid corticotropin-releasing hormone in healthy humans: effects of yohimbine and naloxone. J Clin Endocrinol Metab 2000; 85(11):4138-4145
337. Wang W, Bradley SR and Richerson GB. Quantification of the response of rat medullary raphe neurones to independent changes in p(O) and P(CO2). J Physiol2002; 540:951-970
338. Wang W, Tiwari JK, Bradley SR, Zaykin RV and Richerson GB. Acidosis-stimulated neurons of the medullary raphe are serotonergic. J Neurophysiol 2001; 85(5):2224-2235
339. Ward JW. Diabetic neuropathy. In: International textbook of diabetes mellitus. (Alberti KGMM, Defronzo RA, Keen H and Zimmet P eds) 1992, John Wiley and Sons, UK
340. Washington LL, Gibson SJ and Helme RD. Age-related differences in the endogenous analgesic response to repeated cold water immersion in human volunteers. Pain 2000; 89(1):89-96
341. Webster J, Tonelli L and Sternberg EM. Neuroendocrine regulation of immunity. Annu Rev Immunol 2002; 20:125-613
342. Weese-Mayer DE, Silvestri JM, Huffman AD, Smok-Pearsall SM, Kowal MH, Maher BS, Cooper ME and Marazita ML. Case/control family study of autonomic nervous system dysfunction in idiopathic congenital central hypoventilation syndrome. Am J Med Genet 2001; 100(3):237-245
343. Weese-Mayer DE, Silvestri JM, Marazita ML and Hoo JJ. Congenital central hypoventilation syndrome: inheritance and relation to sudden infant death syndrome. Am J Med Genet 1993; 47(3):360-367
344. Weese-Mayer DE, Silvestri JM, Menzies LJ, Morrow-Kenny AS, Hunt CE and Hauptman SA. Congenital central hypoventilation syndrome: diagnosis, management, and long-term outcome in thirty-two children. J Pediatr 1992; 120(3):381-387
345. Wesseling, K. H., Jansen, J. R. C., Settels, J. J. and Schreuder, J. J. Computation of aortic flow from pressure in humans using a nonlinear, three-element model. J Appl Physiol 1993; 74:2566–2573
346. West JB. Respiratory Physiology. 1974, The Williams and Wilkins Co, Baltimore 347. Wheeler SG, Ahroni JH and Boyko EJ. Prospective study of autonomic neuropathy as
a predictor of mortality in patients with diabetes. Diabetes Res Clin Pract 2002; 58(2):131-138
348. Williams JG, Morris AI, Hayter RC and Ogilvie CM. Respiratory responses of diabetics to hypoxia, hypercapnia, and exercise. Thorax 1984; 39(7):529-534
349. Williams KL, Ko MC, Rice KC and Woods JH. Effect of opioid receptor antagonists on hypothalamic-pituitary-adrenal activity in rhesus monkeys. Psychoneuroendocrinol 2003; 28(4):513-528
350. Windle RJ, Wood S, Shanks N, Perks P, Conde GL, da Costa AP, Ingram CD and Lightman SL. Endocrine and behavioural responses to noise stress: comparison of virgin and lactating female rats during non-disrupted maternal activity. J Neuroendocrinol 1997; 9(6):407-414
351. Witzleb E. Functions of the vascular system. In: Human Physiology (Schmidt RF and Thews G eds) 1983, Springer-Verlag, New York
352. Wolpe J. Psychotherapy by reciprocal inhibition. 1958; Stanford University Press, Stanford
353. Woods SW, Charney DS, Delgado PL and Heninger GR. The effect of long-term imipramine treatment on carbon dioxide-induced anxiety in panic disorder patients. J Clin Psychiatry 1990; 51(12):505-7
354. Woods SW, Charney DS, Goodman WK and Heninger GR. Carbon dioxide induced anxiety. Arch Gen Psych 1988; 45:43-52
271
355. Wright RJ, Rodriguez M and Cohen S. Review of psychosocial stress and asthma: an integrated biopsychosocial approach. Thorax 1998; 53(12):1066-1074
356. Wurtman RJ. Stress and the adrenocortical control of epinephrine synthesis. Metabolism 2002; 51(6 S1):11-14
357. Xiao Y, Via D, Kyle R, Mackenzie CF and Burton P. Stress with simulated trauma management measured by salivary amylase. Anesthesiology 2000; 93(34):A-1226
358. Yehuda R, Teicher MH, Levengood RA, Trestman RL and Siever CJ. Circadian regulation of basal cortisol levels in posttraumatic stress disorder. Ann NY Acad Sci1994; 746:378-386
359. Yergani VK, Rao RK, Smitha RM, Pohl RB, Balon R and Srinivasan. Diminished chaos of heart rate time series in patients with major depression. Biol Psychiatry2001; 51:733-744
360. Young EA, Lopez JF, Murphey-Weinberg V, Watson SJ and Akil H. The role of mineralocorticoid receptors in hypothalamic-pituitary-adrenal axis regulation in humans. J Clin Endocrinol Metab 1998; 83(9):3339-3345
361. Young JB and Landsberg L. Catecholamines and the adrenal medulla. In: Williams Textbook of Endocrinology. 9th Edition (Wilson et al eds) 1998, WB Saunders Co, Philadelphia
362. Young WS III and Lightman SL. Chronic stress elevates enkephalin expression in the rat paraventricular and supraoptic nuclei. Brain Res Mol Brain Res 1992; 13(1-2):111-117
363. Zandbergen J, Pols H, De Loof C, Lousberg H and Griez E. Effect of hypercapnia and other disturbances in the acid-base-balance on panic disorder. Hillside J Clin Psychiatry 1989; 11(2):185-197
364. Zeigler DR, Cass WA and Herman JP. Excitatory influence of the locus coeruleus in hypothalamic-pituitary-adrenocrtical axis responses to stress. J Neuroendocrinol1999; 11:361-369
365. Ziegler D. Diabetic autonomic neuropathy. Cardiac sympathetic "dysinnervation," QT interval prolongation, and mortality. Clin Auton Res 2002; 12(5): 349-352
366. Ziegler D. Diagnosis and treatment of diabetic autonomic neuropathy. Curr Diab Rep 2001; 1(3): 216-227
367. Ziegler MG, Lake CR and Kopin IJ. The sympathetic-nervous-system defect in primary orthostatic hypotension. N Eng J Med 1977; 296:293-297
368. Zobel AW, Nickel T, Kunzel HE, Ackl N, Sonntag A, Ising M and Holsboer F. Effects of the high-affinity corticotropin-releasing hormone receptor 1 antagonist R121919 in major depression: the first 20 patients treated. Psychiatr Res 2000; 34(3):171-181
369. Zuckerman-Levin N, Tiosano D, Eisenhofer G, Bornstein S and Hochberg Z. The importance of adrenocortical glucocorticoids for adrenomedullary and physiological response to stress: A study in isolated glucocorticoid deficiency. J Clin Endocrinol Metab 2001; 86(12) 5920-5924
370. Zvolensky MJ, Feldner MT, Eifert GH and Stewart SH. Evaluating differential predictions of motional reactivity during repeated 20% carbon dioxide-enriched air challenge. Cognition Emotion 2001; 15(6):767-786
272