Br HeartJ_ 1995;73:237-241

Effects of captopril and oxygen on sleep apnoea inpatients with mild to moderate congestive cardiacfailure

John T Walsh, Richard Andrews, Rowena Starling, Alan J Cowley, Ian D A Johnston,William J Kinnear

Divisions ofCardiovascular andRespiratory Medicine,University Hospital,NottinghamJT WalshR AndrewsA J CowleyI D A JohnsonW J KinnearDepartment ofNeurophysiology,University Hospital,NottinghamR StarlingCorrespondence to:Dr J T Walsh, Division ofCardiovascular Medicine,University Hospital,Nottingham NG7 2UH.Accepted for publication20 September 1994

AbstractObjectives-To determine the effects ofcaptopril and oxygen on sleep quality inpatients with mild to moderate cardiacfailure.Design-An open observational study.Patients-12 patients with New YorkHeart Association class II-III heart fail-ure were studied at baseline. 9 of thesepatients were then examined at the end of1 month of treatment with captopril; 9 ofthe patients were separately assessedduring a single night of supplementaryoxygen.Main outcome measures-Sleep patternsby polysomnography, overnight oxime-try, and subjective sleep assessmentusing visual analogue scores.

Results-Abnormal sleep was present inall baseline studies. Complete poly-somnograms after treatment with capto-pril were obtained in 8 patients. Lightsleep (stages 1 and 2) was reduced (mean(SEM) 61%(8)% to 48%(6)% actual sleeptime, P < 0.05) but slow wave (stages 3and 4) and REM (rapid eye movement)sleep increased (25%(6)% to 31%(5)%,14%(2)% to 21%(5)% actual sleep time,P < 0.05). Apnoeic episodes (242(59) to118(30), P < 0-05), desaturation events(171(60) to 73(37), P < 0.05), and arousals(33(5) to 18(3) P < 0.01) were reduced.Visual analogue scores of sleep qualityincreased 49(5) to 69(5), P < 0.01).Complete polysomnograms were ob-tained in 7 patients treated with oxygen.Light sleep duration was reduced (55%(7)% to 42%(5)% actual sleep time, P <

0 05) and slow wave sleep increased(30%(5)% to 38%(6)% actual sleep time,P < 0.05). REM sleep duration was notsignificantly different. Total arousals(33(6)% to 20(2) P < 0.05), desaturationevents (140(33) to 38(10), P < 0.01), andapnoeic episodes (212(53) to 157(33), P <

0.05) were reduced. Visual analoguescores of sleep quality were unchanged.Conclusions-Captopril and oxygen mayimprove sleep quality and reduce noctur-nal desaturation in patients with mild tomoderate cardiac failure. Improved sleepquality could explain the reduction indaytime symptoms seen after treatmentin patients with chronic heart failure.

(Br Heart3' 1995;73:237-241)

Keywords: captopril, oxygen treatment, sleep apnoea,congestive heart failure

Paroxysmal nocturnal dyspnoea commonlyoccurs in patients with congestive cardiac fail-ure. With the introduction of polysomnogra-phy, it is now recognised that even in theabsence of this classic symptom abnormalbreathing patterns during sleep are present inpatients with severe disease.' 2 Periodicbreathing, and hypopnoeic and apnoeicepisodes in these patients may be associatedwith hypoxia and can lead to arousal fromsleep. Disruption of normal sleep architecturemay explain some of the symptoms of heartfailure, such as lethargy, traditionally ascribedto poor cardiac output.

Previous studies have examined the effectsof oxygen, vasodilators, and sedation on sleepin end stage cardiac failure but there is littleinformation concerning those less severelyaffected.2-5 Angiotensin converting enzyme(ACE) inhibitors improve symptoms andprognosis in all grades of heart failure, butwhether they also improve the quality of sleepis unknown. We therefore studied sleep inpatients with mild to moderate heart failurebefore and after treatment with captopril.Patients were also assessed during a singlenight of supplementary oxygen. As there arefew data in the literature regarding sleepabnormalities in patients with milder degreesof heart failure an open observational assess-ment was considered a prerequisite to furtherdetailed study.

Patients and methodsTwelve patients (11 male and one female ofmean (SD) age 63(7) years) were recruitedfrom the outpatient department over a 12month period. All patients had mild to mod-erate cardiac failure (New York HeartAssociation class II-III) and were sympto-matic despite medication with frusemide (atleast 40 mg) or equivalent daily. All hadimpaired left ventricular function on echocar-diography as defined by an increase in leftventricular end diastolic volume greater than5-5 cm and reduced fractional shortening to< 30%. Cardiomegaly greater than 50% waspresent in all patients but none had clinical orradiological evidence of pulmonary oedema.Ischaemic heart disease was the cause ofsymptoms in eight patients, alcoholic car-diomyopathy in two and in the remaining twothe cause was unknown.

237

on July 26, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.73.3.237 on 1 M

arch 1995. Dow

nloaded from

Walsh, Andrews, Starling, Cowley, _rohnston, Kinnear

Patients were excluded if they had any ofthe following: (1) morbid obesity (> 115% ofideal body weight); (2) a history of significantrespiratory disease confirmed with abnormalpulmonary function tests; (3) daytimehypoxia (arterial oxygen pressure < 12 kPa);(4) significant neurological disease or primarysleep disorder; (5) contraindications to ACEinhibitor therapy; or (6) concurrent adminis-tration of /3 blockers, cardiac glycosides, vaso-dilators, or other medication known to altersleep architecture, including sedatives andantidepressants.

Patients were initially studied over a periodof three consecutive nights. All were admittedfor an adaptation night followed by a baselinenight of full polysomnography. During thethird night of assessment patients receivedsupplemental oxygen delivered by nasal can-nulas at a rate of 2 1/min. Captopril 75 mgdaily was then prescribed over a one monthperiod after which patients were studied for afourth night.

SLEEP ASSESSMENTAll patients retired to bed at their usual timesand the study was terminated at theirnormal time of rising. The MEDILOGMultiparameter Analysis System (OxfordMedical, Oxford) was used to record thepolysomnographs. Surface electrodes wereused to record the electroencephalogram,electro-oculogram, and mental electromyo-gram. Two further electrodes were placedover the anterior chest wall and respiratorymovements were measured by impedance.Heart rate and rhythm were monitored withstandard limb leads. Arterial oxygen concen-tration was measured throughout the nightwith a pulse oximeter (Biox 3700; Ohmeda,Boulder, Colorado) attached to the rightindex finger. Oral and nasal airflow weredetected with small thermistors held in placeby an adjustable head band and connected tothe multichannel recorder. The data wereanalysed using the MEDILOG 9000-IIIReplay and Display System and a SS90-IIISleep Stager (Oxford Medicals, Oxford).Standard criteria were used to define abnormalsleep patterns and sleep stage.6 Actual sleeptime was defined as the time spent in bedminus total wake time (as recorded from aactivity and increased frequency of the eletro-encephalogram). The following definitionswere used by the computer to classify abnormalbreathing patterns and arousal. Apnoea wasdefined as the absence of airflow for morethan 10 s. Hypopnoea was defined as areduction in the amplitude of respiratorymovement for more than 10 s to less than50% of the maximum amplitude recordedduring the preceding periodic breathingcycle. The apnoea-hypopnoea index was cal-culated by the computer system from thestored data.

Arousal was defined as awakening fromsleep for greater than 5 s, as evidenced by thesimultaneous occurrence of a activity on theelectroencephalogram, electromyogram acti-vation, and eye movements. Desaturation

events were defined as a reduction in oxygenconcentration of4% or greater.

Sleep stage, arousal, and apnoea classifica-tion by the computer were subsequentlychecked in all patients from inspection of theraw data traces by an independent trainedtechnician (RS).

CARDIAC OUTPUT AND RESPIRATORY GASMEASUREMENTSCardiac output and respiratory gas exchangewere measured before the first and third nightof polysomnography. All patients fasted fora minimum of 6 h before assessment.Respiratory gases were measured with a massspectrometer (VG Medicals, Cheshire) andcardiac output using the indirect Fick princi-ple with carbon dioxide as the indicator.Carbon dioxide production was calculatedfrom the minute ventilation and mixedexpired carbon dioxide concentration. Thepartial pressure of carbon dioxide in pul-monary venous blood was derived from endtidal carbon dioxide concentration and thepartial pressure in mixed venous blood wasmeasured after a rebreathing manoeuvre.These three variables were then used to solvethe Fick equation. This method is non-invasive and correlates with thermodilution.7All measurements were made with the patientstanding at rest.

SLEEP QUALITY AND SYMPTOM ASSESSMENTPatients were asked to assess their daytimesymptoms, including energy scores and sub-jective sleep quality, using a visual analoguescale after each night in hospital. Forms werecompleted in the absence of the investigators.All patients gave informed consent and thestudy was approved by the hospital ethicscommittee.

STATISTICAL ANALYSISData obtained from the sleep studies duringoxygen treatment and after 1 month of capto-pril were separately compared with those frombaseline. All variables are expressed as means(SEM) and were compared using a Wilcoxonsigned rank test. Observed differences wereconsidered significant at P < 0 05.

ResultsAbnormal sleep architecture and breathingpatterns were identified in all twelve baselinestudies according to standard criteria.8 Allpatients experienced frequent arousal in asso-ciation with episodes of apnoea, hypopnoea,and desaturation. These were predominantlyfound during light (stages 1 and 2) sleep.There was no evidence of obstructive sleepapnoea in any sleep recording. All patientswere treated with captopril but equipmentfailure and poor quality polysomnogramsreduced the final data available for analysis.Nine patients were studied after treatmentwith captopril and satisfactory polysomno-grams were obtained in eight. Before treat-ment with captopril six of these patientsand three others were also assessed during a

238

on July 26, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.73.3.237 on 1 M

arch 1995. Dow

nloaded from

Effects of captopril and oxygen on sleep apnoea in patients with mild to moderate congestive cardiac failure

* BaselineEl Captopriltreatment

300> 250

0200 -

0150a 100 _ *X5 50

0

c >z n

0L

4- <

Figure 1 Effect of captoprilon sleep apnoea anddesaturation. *P < 0 05;**P < 0 01 v baseline.Bars are means (SEM).

100 BaselineCaptopril

90 treatment

80

a 70 - T

on sle qult Tndeeg

OD 60O-o 50-400

Fu30

20-2

10

01Sleep Energyscore score

Figure 2 Effect of captoprilon sleep quality and energyscore. **J <0001 v

baseline. Bars are means(SEM).

* BaselineOxygen

a) 80 treatmentC300-

°0)

> 250

ux

-*- 2000

C100

50

00

c

0L

Figure 3 Effect of oxygentreatment on sleep apnoea

and desaturation.

*P< 0.0; **P < 0.0Jv baseline. Bars are means

(SEM).

100

Baseline90 El1 Oxygen

aD 80 treatment

o 70rIn

0D 60A0)

50 ~

C 40-

o7 30-

20

10

Sleep Energyscore score

Figure 4 Effect ofoxygentreatment on sleep qualityand energy score. Bars aremeans (SEM).

Table 1 Effect of captopril on sleep architecture andoxygen saturation

Captopril(75 mgdaily for

Baseline 1 month)Variable (n = 8) (n = 8)

Actual sleep time (min) 412 (29) 342 (22)Stages 1 and 2 (% actual sleep) 61 (8) 48 (6)*Stages 3 and 4 (% actual sleep) 25 (6) 31 (5)*REM sleep (% actual sleep) 14 (2) 21 (5)*No of arousals 33 (5) 18 (3)**Desaturation events 171 (60) 73 (37)*Apnoea/hypopnoea 242 (59) 118 (30)*Apnoea/hypopnoea/h 35 (7) 20 (5)*Minimum oxygen saturation (%) 83 (3) 85 (4)

Values are means (SEM). *P < 0 05; **P < 0-01 v baseline.REM, rapid eye movement.

night of supplemental oxygen. Satisfactorypolysomnograms were obtained in seven ofthe nine patients who received supplementaloxygen.EFFECT OF CAPTOPRIL ON SLEEPTable 1 gives baseline values and the effect ofcaptopril on sleep architecture in eightpatients. Actual sleep time did not signifi-cantly change but light sleep (stages 1 and 2)was significantly reduced. Slow wave sleep(stages 3 and 4) and the duration of rapid eyemovement associated with sleep increasedafter treatment. Apnoeic episodes, desatura-tion events, and arousal were all reduced(fig 1). Subjective sleep quality and daytimeenergy improved (fig 2).

EFFECT OF CAPTOPRIL ON CENTRALHAEMODYNAMICS AND RESPIRATORY GASEXCHANGEThere were no significant changes in cardiacoutput and oxygen uptake after treatment.End tidal carbon dioxide concentrations weresignificantly increased and daytime minuteventilation reduced (table 2).

EFFECT OF OXYGEN ON SLEEPTable 3 gives baseline values and the effect ofoxygen treatment on sleep architecture in

Table 2 Effect of captopril on cardiac output andrespiratoty gas exchange

Baseline CaptoprilVariable (n = 8) (n = 8)

Cardiac output (1/min) 4-8 (0-4) 5-2 (0-4)Oxygen uptake (mllkg/min) 4-5 (0-3) 4-4 (0-4)End tidal carbon dioxide (kPa) 4-36 (0-4) 4-66 (0-36)*Minute ventilation 16-2 (1-5) 13 (1-4)*

Values are means (SEM). *P < 0-05 v baseline.

Table 3 Effect of oxygen on sleep architecture and oxygensaturation

OxygenBaseline (I night only)

Vaiable (n = 7) (n = 7)

Actual sleep time (min) 394 (30) 399 (22)Stages 1 and 2 (% actual sleep) 55 (7) 42 (5)*Stages 3 and 4 (% actual sleep) 30 (5) 38 (6)*REM sleep (% actual sleep) 15 (3) 20 (2)No of arousals 33 (6) 20 (2)*Desaturation events 140 (33) 38 (10)**Apnoea/hypopnoea 212 (53) 157 (33)*Apnoea/hypopnoea/h 32 (7) 23 (4)*Minimum oxygen saturation (%) 79 (5) 87 (1)*

Values are means (SEM). *P < 0-05; **P < 0-01 v baseline.REM, rapid eye movement.

seven patients. Actual sleep time wasunchanged but light sleep (stages 1 and 2)was significantly reduced. Slow wave (stages 3and 4) sleep increased but the duration ofrapid eye movement associated with sleepwas similar to that of baseline values. As withcaptopril apnoeic episodes and desaturationevents were also reduced and there was a sig-nificant decrease in the number of arousals(fig 3). This improvement was not reflected inthe patients' subjective assessment of sleepquality (fig 4).

DiscussionOur observations suggest that captoprilimproves abnormalities of breathing duringsleep in patients with mild to moderate car-diac failure. Patients have fewer arousals andspend more time in sleep stages 3 and 4. Theyenjoy a better night's sleep and feel less lethar-gic throughout the day. Supplemental oxygenhad a similar effect on sleep stage and breath-ing abnormalities, although there was no sta-tistically significant improvement in sleepquality scores assessed by the patients. It isunlikely that these changes are explained byadaptation to the sleep laboratory as patientsreassessed after 1 month of taking captoprilwere admitted only for a single night.Furthermore, although the proportion of lightsleep can change with adaptation, there isconsiderable evidence demonstrating thatabnormalities of breathing are less subject tonightly variation.9-'2

Although patients had been told that theirsymptoms might improve no comments weremade regarding sleep quality and so webelieve their subjective response to be real.The use of visual analogue scales to assessdaytime symptoms and sleep quality is previ-ously described,'3 although more formal mea-sures of the quality of life and sleep latencytesting are perhaps more reliable. 14-16 Aplacebo response seems unlikely given theobjective evidence of improved sleep qualitybut firm conclusions from these data are lim-ited by the study design. A controlled doubleblind randomised trial is needed to confirmour findings.Most previous studies in heart failure have

focused on patients with severe symptoms.' 417Periodic breathing with Cheyne-Stokes respi-ration during sleep is well described.2Explanations for these abnormalities are prob-ably multifactorial and various mechanismshave been proposed,3 including a reduction inoxygen stores, increased hypoxic sensitivity,and delayed circulation time between thecarotid body and lung. Abnormal respiratorypatterns are also associated with increasedoscillation in arterial carbon dioxide pressure(Paco,) around the apnoeic threshold.'8 19This represents a critical level of arterial car-bon dioxide (CO2) below which apnoeaoccurs during sleep.

Cheyne-Stokes respiration is more com-mon in early sleep in patients with heart fail-ure and an increase in the duration of stages 1and 2 correlates closely with sleep disruption.The ventilatory response to hypoxia andhypercapnia is greater during this period and

239

on July 26, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.73.3.237 on 1 M

arch 1995. Dow

nloaded from

Walsh, Andrews, Starling, Cowley, _Johnston, Kinnear

respiratory control unstable.' Deeper sleepoccurs during the later stages (3 and 4) andCheyne-Stokes respiration often disappears.Cheyne-Stokes respiration is characterised byhyperpnoeic and apnoeic stages. Apnoeaoccurs after an increase in ventilation whichlowers the Paco2 below the apnoeic threshold.A subsequent increase in CO2 concentrationrestores ventilation and the cycle then contin-ues. Arousal in patients with heart failurecoincides with the hyperpnoeic stage, in con-trast to these with obstructive sleep apnoea inwhom arousal coincides with apnoea causingbreathing to resume.' The significance of thisdifference is unclear.The reason for the initiation of periodic

breathing in some individuals with heart fail-ure but not others is unknown. Daytimehyperventilation causing hypocapnia maydetermine which patients develop abnormalsleep architecture.'8 19 This is more commonin patients developing Cheyne-Stokes respira-tion and nocturnal periodic breathing. Theincrease in ventilation is thought to be causedby stimulation of pulmonary receptors byinterstitial oedema. In the patients treatedwith captopril in our study there was a smallbut significant decrease in daytime minuteventilation associated with improved sleepquality and reduction in breathing abnormali-ties. Daytime hypocapnia was also reduced asevidenced by the increase in Paco2, estimatedfrom end tidal CO2 concentration.

Abnormal breathing patterns are welldescribed in those with severe symptoms andit has been suggested that this relates to poorventricular function.2021 Alternatively, thiscould reflect the patients selected for study asthere has been little assessment of patientswith less severe ventricular dysfunction. Sleepabnormalities did not relate to resting cardiacoutput or cardiac index in our patients withmild to moderate heart failure.

Nocturnal hypoxia seen in our patients haspreviously been described in those with severesymptoms.' 34 Periodic breathing, however, isnot the only cause of desaturation duringsleep in patients with heart failure. Reducedlung compliance, poor ventricular function,and loss of effective lung volume when supineare additional contributory factors.' 22 Theclinical significance of oxygen saturationabnormalities has yet to be determined, par-ticularly as many patients have normal arterialoxygen concentrations during the day.Myocardial infarction,23 high grade arrhyth-mias,24 and systemic25 and pulmonary hyper-tension26 are all associated with nocturnalhypoxia. Supplemental oxygen significantlyimproves sleep quality and reduces nocturnaldesaturation in patients with severe heart fail-ure.3 We have demonstrated that breathingabnormalities and desaturation in milderdegrees of heart failure also respond to oxygentherapy. A placebo randomised study of theeffects of oxygen, however, would be neces-sary to confirm this finding. For reasons thatare unclear there was no significant improve-ment in the patients' subjective assessmentusing a visual analogue scale. One possible

explanation is that, although oxygen acutelycorrects sleep abnormalities, prolongedadministration is required before patientsthemselves appreciate any benefit. The long-term effects of oxygen supplementation in thiscontext are unknown.

There have been few studies of the effectsof drugs on sleep in patients with congestivecardiac failure. Improved breathing abnor-malities and oxygen saturation were reportedin six patients with decompensated severesymptoms administered standard medicaltreatment.4 Such management consisted ofintravenous diuretics, inotropes, and oxygenmaking it impossible to assess the individualtherapeutic effects. Benzodiazepines haverecently been shown to reduce nocturnalarousals and Cheyne-Stokes respiration insevere heart failure, apparently with littleeffect on arterial oxygen saturation.527 Thereare several possible mechanisms by whichcaptopril may improve sleep quality inpatients with heart failure. Improved cardiacfunction could reduce circulatory delay andstabilise ventilatory control. The reduction inminute ventilation and increase in Paco2 seenin the patients we studied could have a similareffect. Alternatively, captopril may attenuatethe adverse effects of hypoxia. Hypoxia isassociated with increased renin and cate-cholamine levels but the effects of neurohor-monal activation on central ventilatory controlare unknown.2026 Captopril may similarlyimprove cardiac function and respiratory gasexchange by reducing hypoxic vasoconstric-tion of systemic and pulmonary vessels.26The role of nasal continuous positive air-

way pressure in patients with heart failure hasyet to be defined, with conflicting data fromseveral small studies.28- 30 It is unclear whycontinuous positive airway pressure should beeffective in central apnoea, although it hasbeen suggested that by increasing functionalresidual capacity arterial oxygen and carbondioxide concentrations increase.2 This wouldstabilise the respiratory system to the effects ofhypocapnia, reducing both periodic breathingand Cheyne-Stokes respiration. Our data sug-gest that captopril has similar effects on sleep.

Poor sleep quality is likely to contribute tothe overwhelming fatigue seen in patients withall grades of congestive cardiac failure.Captopril appears to improve sleep qualitysubjectively and objectively in patients withless severe disease. This may in part explainthe benefits already described in patientstreated with ACE inhibitors in terms of symp-toms, quality of life, and prognosis.

1 Hanly PJ, Millar TW, Steljes DC, Baert R, Frais MA,Kryger MH. Respiration and abnormal sleep in patientswith congestive heart failure. Chest 1989;96:480-8.

2 Kryger MH. Sleep and heart failure. Eur RespirJ 1990;3:1103-4.

3 Hanly PJ, Millar TW, Steljes DG, Baert R, Frais MA,Kryger MH. The effect of oxygen on respiration andsleep in patients with congestive heart failure. Arm InternMed 1989;111:777-82.

4 Dark DS, Pingleton SK, Kerby GR, et al. Breathing pat-tern abnormalities and arterial oxygen desaturation duringsleep in the congestive heart failure syndrome. Chest1 987;91 :833-6.

5 Biberdorf DJ, Steens R, Millar TW, Kryger MH.

240

on July 26, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.73.3.237 on 1 M

arch 1995. Dow

nloaded from

Effects of captopril and oxygen on sleep apnoea in patients with mild to moderate congestive cardiac failure

Benzodiazepines in congestive heart failure: effects oftemazepam on arousability and Cheyne-Stokes respira-tion. Sleep 1993;16:529-38.

6 Kales A, Rechstaffen A, eds. A manual of standardisedterminology, techniques and scoring system for sleep stagesof human subjects. NIH publication no 204. Bethseda,MD: National Institute of Neurological Institute ofNeurological Disease and Blindness, 1968.

7 Cowley AJ, Stainer K, Murphy DT, Murphy J, HamptonJR. A non invasive method for measuring cardiac out-put: the effect of a Christmas lunch. Lancet 1986;ii:1422-4.

8 William RL, Karacan I, Hursch CJ. Electroencephalography(EEG) of human sleep: clinical applications. New York:Wiley, 1974.

9 Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S.The occurrence of sleep disordered breathing amongmiddle aged adults. N EnglJ Med 1993;328:1230-5.

10 Agnew HW Jr, Webb WB, Williams RL. The first nighteffect: an EEG study of sleep. Psychophysiology 1966;2:263-6.

11 Lord S, Sawyer B, O'Connell D, et al. Night to night vari-ability of disturbed breathing during sleep in an elderlycommunity sample. Sleep 1991;14:252-8.

12 Wittig RM, Roaker A, Zorick FJ, Roehers TA, CanwayWA, Roth T. Night to night consistency of apnoeas dur-ing sleep. Sleep 1991;14:383-5.

13 Morriss R, Sharpe M, Sharpely AL, Cowen PJ, Hawton K,Morris J. Abnormalities of sleep in patients with thechronic fatigue syndrome. BMJ 1993;306: 1161-4.

14 Thorpy MJ. The clinical use of the multiple sleep latencytest. Sleep 1992;15:268-76.

15 Greenberg GD, Watson RK, Depulta D. Neuro-psychological dysfunction in sleep apnoea. Sleep 1987;10:254-62.

16 Hunt SM, Mcewen J, McKenna SP. Perceived health: ageand sex comparisons in a community. J EpidemiolCommunity Health 1984;38: 156-60.

17 Findley U, Zwillich CW, Ancoli-Israel S, Kripke D, TisiG, Moser M. Cheyne-Stokes breathing during sleep inpatients with left ventricular failure. South Med J 1985;78:11-5.

18 Naughton M, Bernard D, Tam A, Rutherford R, BradleyTD. Role of hyperventilation in the pathogenesis of cen-tral sleep apnoeas in patients with congestive heart fail-ure. Am Rev Respir Dis 1993;148:330-8.

19 Hanly PJ, Zuberi N, Gray R. Pathogenesis of Cheyne-Stokes respiration in patients with congestive heart fail-ure. Chest 1993;104:1079-84.

20 Feld H, Priest S. A cyclic breathing pattern in patientswith poor left ventricular function and compensatedheart failure: a mild form of Cheyne-Stokes respiration?JAm Coil Cardiol 1993;21:971-4.

21 Gottlieb SS, Kessler P, Lee WH, Medina N, Yushak M,Packer M. What is the significance of Cheyne-Stokesrespiration in severe chronic heart failure? Jf Am CollCardiol 1986;7(suppl A):43A.

22 Webb P. Periodic breathing during sleep. I Appl Physiol1974;37:899-903.

23 Hung J, Whitford EG, Parsons RW, Hillman DR.Association of sleep apnoea with myocardial infarctionin men. Lancet 1990;336:261-4.

24 Cripps T, Rocker G, Stradling J. Nocturnal hypoxia andarrhythmias in patients with impaired left ventricularfunction. Br Heart J 1992;68:382-6.

25 Williams AJ, Houston D, Finberg S, Lam C, Kinney JL,Santiago S. Sleep apnoea syndrome and essential hyper-tension. Am I Cardiol 1985;55:1019-22.

26 Davies SW, Wedzicha JA. Hypoxia and the heart. Br HeartJ 1993;69:3-5.

27 Guilleminault C, Clerk A, Labanowski M, Simmons J,Stoohs R. Cardiac failure and benzodiazepines. Sleep1993;16:524-8.

28 Buckle P, Millar T, Kryger M. The effect of short termnasal CPAP on Cheyne-Stokes respiration in congestiveheart failure. Chest 1992;102:31-5.

29 Davies RJ, Harrington KJ, Ormerod OJ, Stradling JR.Nasal continuous positive airway pressure in chronicheart failure with sleep disordered breathing. Am RevRespir Dis 1993;147:630-4.

30 Takasaki Y, Off D, Popkin J, Rutherford R, Liu P, BradleyTD. Effect of nasal continuous positive airway pressureon sleep apnoea in congestive heart failure. Am RevRespirDis 1989;140:1578-84.

241

on July 26, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.73.3.237 on 1 M

arch 1995. Dow

nloaded from