critically electrophysiological guillain-barrecritical illness meantsignificant involvement...

Journal of Neurology, Neurosurgery, and Psychiatry 1986;49:563-573

Critically ill polyneuropathy: electrophysiologicalstudies and differentiation from Guillain-BarresyndromeCHARLES F BOLTON, DEBORAH A LAVERTY, JOHN D BROWN,NORBERT J WITT, ANGELIKA F HAHN, WILLIAM J SIBBALD

From the Department of Clinical Neurological Sciences and the Critical Care/Trauma Unit, Victoria Hospital,.The University of Western Ontario, London, Ontario, Canada

SUMMARY A polyneuropathy of varying severity has been observed in association with sepsis andcritical illness in 15 patients. Since clinical evaluation is often difficult, electrophysiological studiesprovided definitive evidence for polyneuropathy. These revealed reductions in the amplitudes ofcompound muscle and sensory nerve action potentials, the most marked abnormality. Near-nerverecordings confirmed such reductions for sensory fibres. Needle electromyography revealed signs ofdenervation of limb muscles. Phrenic nerve conduction and needle electromyographic studies ofchest wall muscles suggested that the polyneuropathy partially explained difficulties in weaningpatients from the ventilator, an early clinical sign. No defect in neuromuscular transmission was

demonstrated, despite the use of aminoglycoside antibiotics in some patients. In those who survivedthe critical illness, clinical and electrophysiological improvement occurred. The 15 critically illpolyneuropathy patients were compared with 16 Guillain-Barre syndrome patients observed duringthe same period. The analysis showed that the two polyneuropathies are likely to be separate entitiesthat can be distinguished in most instances by the predisposing illness, electrophysiological featuresand cerebrospinal fluid results.

Critical illness arises when sepsis and multi-organdysfunction complicate the course of a primary illnessor injury. The pathophysiology, incidence and clinicalfeatures of the dysfunction of the various organs areonly now being clarified, and despite modern methodsof management the mortality rate is at least 60%.'

Little is known of the effects on the nervous system.Varying degrees ofcoma occur in approximately 10%of patients.' We have recently described the clinicaland pathological features of 12 cases of "septicencephalopathy".2 Attention to the peripheral ner-vous system has been limited to studies of metabolicdisturbances of skeletal muscle, a potential cause of

Presented in part at the 7th International Congress of Electro-myography, Munich, Germany, 9-13 October, 1983, and at the XIXCanadian Congress of Neurological Sciences, Edmonton, Canada,27-30 June, 1984.

Address for reprint requests: Dr CF Bolton, Department of ClinicalNeurological Sciences, Victoria Hospital, PO Box 5375, London,Ontario, Canada N6A 4G5.

Received 10 May 1985 and in revised form 28 July 1985.Accepted 3 August 1985

respiratory muscle fatigue.36 This may be one of themajor reasons for ventilatory assistance during thecourse of critical illness.The peripheral nerves are also affected. In the last

seven years we have observed a polyneuropathy, attimes severe, which has complicated the course of crit-ical illness. We have named it critically ill poly-neuropathy (CIP). Unexplained difficulty in weaningthe patient from the ventilator, reduced deep tendonreflexes and weakening of spontaneous orreflex-induced limb movements were early clinicalsigns.`- In the setting of critical illness these signswere difficult to elicit, but electrophysiologicalinvestigation provided the definitive evidence of poly-neuropathy.

This report, therefore, documents the electro-physiological findings in 15 such patients, empha-sising their role in arriving at an initial diagnosis andof documenting the subsequent course of the poly-neuropathy. It also compares these patients with 16patients who had Guillain-Barre syndrome and wereinvestigated and treated in our hospital during thesame period, since it is important to distinguish these

563

Protected by copyright.

on March 11, 2021 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.49.5.563 on 1 May 1986. D

ownloaded from

http://jnnp.bmj.com/

Table I Clinical data on septic and critically ill patients

Patient Age Sex Primary illness Severity of Course of Outcome of(yr) polyneuropathy polyneuropathy patients

1 56 F Lactic acidosis Severe Improved Recovered2 56 F Empyema Severe Improved Recovered3 66 M Obstructive lung Moderate Improved Death

disease4 58 F Pneumonia Severe Improved Death5 49 M Acoustic neuroma Moderate Improved Recovered

surgery6 66 M Obstructive lung Mild Unchanged Death

disease7 72 M Obstructive lung Mild Unchanged Death

disease8 74 M Aortic aneurysm Mild Improved Death

surgery9 65 F Pneumonia Severe Improved Recovered10 75 F Hip surgery Moderate Worsened DeathI1 72 F Pancreatitis Mild Improved Recovered12 67 F Obstructive lung Mild Worsened Death

disease13 57 M Bacterial endocarditis Severe Worsened Death14 83 F Endometrial cancer Mild Unchanged Death15 62 F Septic hip joint Mild Improved Recovered

Table 2 Clinical data on Guillain-Barre syndrome patients

Patient Age Sex Primary or predisposing Complicating illnesses Severity of Outcome of(yr) illness polyneuropathy polyneuropathy

1 74 F Asian influenza None Mild Improvedinnoculation

2 61 M Upper respiratory Respiratory and Severe Improvedinfection autonomic failure

Pneumonia3 25 M Acute stomatitis None Mild Improved4 61 M Upper respiratory Respiratory and Severe Improved

infection autonomic failure5 44 F Upper respiratory None Moderate Improved

infection6 42 F Upper respiratory None Moderate Improved

infection7 58 M Upper respiratory Urinary retention Moderate Improved

infection8 59 F Influenza innoculation Respiratory and Severe Improved

autonomic failure,pulmonary embolus,syndrome of inappropriateantidiuretic hormone release

9 19 M None None Mild Improved10 90 M Influenza Respiratory and Severe Improved

autonomic failure,pneumonia, multipleorgan failure

I1 28 F None None Moderate Improved12 61 M Influenza None Moderate Improved13 35 M Upper respiratory Respiratory and Severe Improved

infection autonomic failure,pneumonia, encephalo-pathy, pulmonary emboli

14 42 M Gastrointestinal None Moderate Improvedinfection

15 49 F None Respiratory and Severe Improvedautonomic failurePneumonia, sepsis

16 27 M Upper respiratory None Severe Improvedinfection

564 Bolton, Laverty, Brown, Witt, Hahn, Sibbald



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Critically ill polyneuropathy: electrophysiological studies and differentiation from Guillain-Barre syndrome 565

Median nerve conduction studies

F wave latency

I

0 12 31. 56

Compound thenar muscleAP latency

Compound digital nerveA P latency

41x.lj11

0 1 2 3 5 6Time (nonths)

I

I I I I I

0 1 2 3 4 5 6

Fig 1 The maximum speed ofimpulse conduction (mean - SD) in CIP (0) and Guillain-Barre syndrome (0) patients.This remained near normal in CIP patients, but was considerably slowed in Guillain-Barre syndrome patients, roughlycorrelating with the course ofthe polyneuropathy. Note markedprolongation ofcompound thenar muscle AP latency,suggesting predominant demyelination ofdistal motorfibres. Partly because oflarge standard deviations, F wave latencies at

zero and one month were the only values that were statistically different between the two groups (p < 005). (AP means actionpotential. Months refers to time afterfirst electrophysiological study).

two neuropathies, each potentially occurring as com-

plications of infection, surgery, and trauma.

Methods

Between 1977 and 1983, 15 CIP and 16 Guillain-Barr6 syn-drome patients (tables 1 and 2) were observed at VictoriaHospital. Both types of polyneuropathies were identifiedand graded as: mild (electrophysiological abnormalities

Table 3 Control median nerve conduction results (mean + SD,

present with normal or equivocally abnormal neurologicalsigns); moderate (muscle weakness of moderate severity andreduced or absent deep tendon reflexes); severe (severe mus-cle weakness, absent deep tendon reflexes and signs of severedenervation (grade 3 to 4 spontaneous activity of needleelectrode study of muscle). All severe Guillain-Barre syn-drome patients required assisted ventilation. The electro-physiological criteria for abnormality were: reductionsbelow two standard deviations in the amplitude of coni-pound action potentials from either muscle or sensory nerve

CIP (15 subjects) Guillain-Barre syndrome(16 subjects)

Motor conduction velocity (Elbow-Wrist) 55-3 + 3-7 m/s 58-8 + 2-7 m/sMotor distal latency 3-6 + 0-8 ms 3-7 + 0-6 msThenar compound muscle action potential amplitude 10-0 + 2-6 mV 11-4 + 3-3 mVSensory conduction velocity (Elbow-Wrist):

Females 56-6 + 5-4 m/s 60 8 + 4 5 m/sMales 57-1 ± 30 m/s 62-9 + 5-6 m/s

Sensory distal latency:Females 2 5 + 0-3 ms 2-6 + OS msMales 25 + 02ms 25 + 0-4ms

Digital compound nerve action potential amplitude:Females 21-3 + 9-6MV 34-7 + 12-6 1VMales 18-7 + 118pV 24-9 + 105pV

F wave latency 28-2 + 2-0 ms 27-9 + 2-3 msHand temperatures 32 1 + 0 90C 32 1 + 0 90C

3001

.5

0o



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


566

Compound thencirmuscle

125 AP aimplitude

, r10020

° 750

50-

25-

T X

0 1 2 3 4 5 6 0 1 2 3 4 5 6Time (months)

Fig 2 Mean amplitudes ( + SD) ofcompound actionpotentials (AP) in CIP (0) and Guillain-Barre syndrome(0) patients. Values were more depressed in Guillain-Barresyndrome patients, but returned towards normal in bothgroups at 6 months.

and abnormal spontaneous activity in muscles of at least twolimbs.The CIP patients were identified according to the above

criteria for polyneuropathy and that they were bothsuffering from sepsis and critically ill. Sepsis was defined as

septicaemia or a focus of infection with systemic effects, andcritical illness meant significant involvement of two or more

major organs. The first four of the CIP patients were

reported previously.9The Guillain-Barre syndrome patients were selected dur-

ing the same period on the basis that they were all adults(over 18 years of age), that the polyneuropathy had run a

monophasic course (since all CIP patients had run a

monophasic course) and that they conformed to Asbury'scriteria for the diagnosis of Guillain-Barre syndrome.10

In both groups, initial clinical and EMG examinationswere performed within one month of admission to hospital.It was possible to perform follow-up examination at approx-imately one month and 6 months after the first examination,on seven CIP and ten Guillain-Barre syndrome patients.Patients were lost to follow-up because of death of CIPpatients, early recovery and discharge of Guillain-Barre syn-drome patients, or transfer to another hospital.

Electrophysiological tests were performed on both groups

according to standard procedures using surface electrodesfor nerve conduction studies and concentric needle elec-trodes for electromyography.9 Surface temperatures were

measured on the palm and dorsum of the foot. Septicpatients had limb temperatures that were higher than con-

trol subjects or Guillain-Barre syndrome patients. Thus,each median nerve conduction value was corrected to a pal-mar skin temperature of 32 1°C (a control value for our lab-oratory), using correction factors previously established inour laboratory appropriate to either control subjects or per-sons with a polyneuropathy. 112 The correction factors perIC were: F wave latency, normal: 0 64, neuropathy:1-52 ms; CMAP latency, normal: 0-28, neuropathy: 0-23 ms;

Bolton, Laverty, Brown, Witt, Hahn, Sibbald

CSAP latency, normal: 0-09, neuropathy: 010Ims; CMAPamplitude, normal: 025, neuropathy: 0 11 mV; CSAPamplitude, normal: 1-96, neuropathy: 0 84 pV. All values intable 3 and figs 1 and 2 are temperature-corrected. A controlsubject was chosen who matched each CIP or Guillain-Barresyndrome patient for age and sex. All patients' nerve con-duction values were expressed as a percentage of the controlvalue. The analysis of digital CSAP amplitudes was per-formed separately for each sex, since these amplitudes arenormally greater in females than males."3 The lower ampli-tudes for digital compound nerve action potentials inGuillain-Barre syndrome are likely due to their younger agecompared to CIP patients (mean of 48 vs 65 years).

After the above data had been assembled, the CIP andGuillain-Barre syndrome groups were compared graphicallyand by the two sample t test of significance.

Needle electromyographic studies were carried out inselected proximal and distal muscles of both upper andlower limbs and from the external oblique and externalintercostal muscles at the fifth interspace. Abnormal sponta-neous activity was graded on a scale of0 (normal) to 4. Brieftrains of positive waves or fibrillation potentials induced byneedle movement or occurring in only one area of the musclewere simply described and not given a grade. More definitiveactivity was graded as: grade 1: trains of positive waves orfibrillation potentials in at least two areas of the muscle;grade 3: trains of such waves in all areas of the muscle; grade4: abundant such activity in all areas of muscle. The amountof motor unit potential activity (number of units recruitedby voluntary contraction) was graded on a scale of 4 (nor-mal) to 0 (absent). Initial and serial needle electrode studieswere invariably performed on the vastus lateralis muscle,and the findings were recorded on tape.

Other electrophysiological studies were carried out in afurther 10 CIP patients studied in 1983 and 1984; five hadevidence of polyneuropathy. Tests for defects in neu-romuscular transmission on these patients were made bystimulating either the median or ulnar nerve at the wrist andrecording from the thenar or hypothenar eminences,respectively. Trains of six supramaximal stimuli were deliv-ered at rates of 3 and 20 Hz. The patients were too ill tocontract voluntarily their muscles maximally, so effects afterexercise could not be evaluated. Near-nerve needle electroderecordings from the sural nerve were also carried out. Forthese studies, the sural nerve was stimulated at the calf andthe needle electrodes were inserted in a bipolar arrangementnear the sural nerve at the ankle. The responses to 64 stimuliwere averaged on a Nicolet 1170 averager.The cerebrospinal fluid was examined in all Guillain-

Barre syndrome patients and 12 CIP patients within onemonth of onset of the polyneuropathy. It was measured by aDupont automatic clinical analyser. The cell count was per-formed manually using a counting chamber.

Results

General observations on critically ill patientsThe mean age of the patients was 65 (49-83) years,nine females and six males. They presented with avariety of primary illnesses (table 1) which, withinhours, deteriorated to a state of critical illness



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Critically ill polyneuropathy: electrophysiological studies and differentiation from Guillain-Barre' syndrome 567

2Q4V

4ms

9.1.4V4ms

Fig 3 Sural nerve recordings in a 64-year-old-man who hada moderately severe critically illpolyneuropathy. Withsurface electrodes at the ankle and stimulating electrodes atmid-calf, no CSAP was recorded after three stimuli (upperfigure) or after 64 stimuli had been averaged (middlefigure).However, a low voltage CSAP (4), 3-4 i V, at maximalconduction velocity of38-8 mls, was demonstrated with a nearnerve recording of64 averaged responses (lowerfigure),consistent with partial axonal degeneration ofthe sural nerve.(The low voltage, later response in the middlefigure arosefrom muscle, not sural nerve;foot skin temperature was31 9°C).

requiring early ventilatory assistance and admissionto the critical care unit. All were septic and a varietyof organisms were cultured from various sites in thebody, including the blood. At least two major organsystems were involved but none of the patients whorecovered had severe or persisting dysfunction ofeither liver or kidney. Intensive treatment was given,including nasogastric tube feedings and total par-enteral nutrition. Episodes of septic shock, occurringin most patients, were managed by pressor drugs andblood transfusions. Multiple antibiotics were givenbut no patient consistently received any one type orgroup of antibiotics and most received both penicillinand aminoglycosides. The serum albumin and abso-lute lymphocyte counts were depressed in all patients.Careful analysis of our records has not shown anyparticular antibiotics to be responsible for the poly-neuropathy or that the institution of total parenteralnutrition was associated with recovery. Nine of the 14

patients ulimately died but in none of these patientsdid polyneuropathy appear to be a major reason fordeath.

Clinicalfeatures ofcritically ill polyneuropathyThe polyneuropathy usually began within one monthof admission to the critical care unit. Difficulty wean-ing the patient from the ventilator as the critical ill-ness was beginning to subside, the development ofweakness of the limbs and reduced deep tendonreflexes, were early clinical signs. However, in mild or

o F.

0D

Fig 4 Selected examples ofneedle electrodefindings inskeletal muscle in the acute phase ofcritically illpolyneuropathy. Positive waves (a), at times profuse (b),andfibrillation potentials (c) in resting muscle, and motorunit potentials (f) on voluntary activation which weredecreased in nwnber, slightly polyphasic, but ofrelativelynormal size, are allfindings that are consistent with recentdenervation. The reasonfor the high incidence ofmyotonic(d) or complex repetitive discharges (e) (seven patients) isspeculative. (Note thefrequently changing composition ofthecomplex repetitive discharge, an atypicalfeature).



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



moderate cases, these signs were often equivocal. Incases of severe neuropathy, limb movements werevery weak or absent, but head, facial, tongue and jawmovements were relatively preserved (except for someweakness of facial muscles in three cases). Sensorysymptoms and signs were milder and, when testable,consisted of varying degrees of stocking and gloveloss to all modalities. Autonomic dysfunction was notevident on routine clinical examination. All signswere most marked in the legs and distal parts of thelimbs. The polyneuropathy was severe in five, mod-erate in three, and mild in seven patients.

Clinical and electrophysiological follow-up in eightshowed improvement in the polyneuropathy. Ofthose who failed to improve, the polyneuropathy wasunchanged in three while it worsened in two. Themortality rate in the last group was highest, six ofseven patients dying. In the group whose poly-neuropathy improved, only three of eight patientsdied.

Recovery occurred over a matter of months. It waslongest for those who had the most severe neu-ropathies, and in these patients, mild, residual clinicalsigns remained.

K-

I4mV

2 3 4Time (months)

Fig 5 Concentric needle electrode activity ( + SD) in thequadriceps muscle ofCIP (0) and Guillain-Barre syndrome(0) patients. The results were more abnormal in distal andlower limb muscles. The abnormal spontaneous activityconsisted mainly ofpositive sharp waves andfibrillationpotentials, 7 CIP patients also exhibiting complex repetitivedischarges and myotonic discharges. The magnitude of theabnormalities correlated with the degree ofdenervation andreinnervation ofmuscle. Note the relative lack ofdenervationofmuscle in Guillain-Barre syndrome initially, even thoughfew motor units were recruited on voluntary contraction,consistent with predominant demyelination.

Fig 6 Repetitive ulnar nerve stimulation studies, recordingfrom hypothenar muscles, in a 43-year-old man, septic andcritically ill, on aminoglycoside antibiotics, but with noclinical or electrophysiological evidence ofpolyneuropathy.Upperfigure shows no change in hypothenar CMAP atstimuli of3 Hz. Lowerfigure, stimuli at 20 Hz, shows anapparent incrementing response, but this is"pseudofacilitation, " since the CMAP area did not change.Thus, no defect in neuromuscular transmission wasdemonstrated.

568

4r

._-> 3u

a0

c 2.-

cIn0

CL

0 1010

4-1-C

01

I.o

0

E

40

00 1

I [I J

5 6

- r

.-N,I



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



Electrophysiological features of critically illpolyneuropathyIn comparing CIP patients to controls, the mostmarked and the only statistically significant (p <0-02) nerve conduction abnormality was a reductionin the amplitudes ofCMAP and CSAP (figs 1 and 2),which was greatest for muscle (fig2). Measurementsof motor and sensory conduction velocities, distallatencies, and F wave latencies were only mildlyabnormal (fig 1).The pattern of results for common peroneal and

sural nerve conduction studies were similar to mediannerve, except the reductions in CMAP and CSAPamplitudes were even more marked; the CSAP wasabsent from the sural nerves of seven patients andfrom the CMAP of the EDB muscle in three patients.To find out if the CSAP amplitudes were absent as aresult of tissue oedema increasing the distancebetween surface recording electrode and the under-lying nerve, near nerve recordings were carried out ina further 10 critically ill patients, five of whom hadpolyneuropathy. The response was absent with sur-face electrodes in seven patients but was recorded atlow voltage but normal latency with averaged, nearnerve recordings in four of these patients (fig 3) ampli-tude 3-3 (26-3 9) pV, but was ,still absent in threepatients. We concluded these seven patients probablyhad axonal degeneration of their sural nerves, eventhough just five met the criteria (in Methods) for poly-neuropathy.

Phrenic nerve conduction studies in 10 of the 15patients revealed the response over the diaphragmwas absent in three, all of whom had a severe poly-neuropathy. In the remaining patients, the latencieswere not significantly prolonged, as one would expectin a purely axonal degeneration. In our experience,the amplitude of the CMAP from the diaphragm isnormally too variable to be usefully interpreted.The most striking abnormalities were observed on

needle electrode study of skeletal muscle (figs 4 and5). These studies showed typical findings of acutedenervation. In resting muscle, fibrillation potentialsand positive waves were present, being both numer-ous and continuous in severe denervation. Unusualspontaneous activity had a high incidence, myotonicdischarges being recorded in five patients and com-plex repetitive discharges in two patients. Needle elec-tromyography was performed unilaterally on theexternal oblique and external intercostal muscles in 12of 15 patients. Positive waves and fibrillation poten-tials were recorded in four of these patients, all havinga severe polyneuropathy.

Repetitive nerve stimulation studies on a further 10patients failed to reveal a defect in neuromusculartransmission, although five of these patients, all onaminoglycoside antibiotics, exhibited the phenom-

enon of "pseudofacilitation" at rapid rates of stimu-lation (fig 6).

Thus, the above findings point to a primary degen-eration of motor and sensory peripheral nerves sup-plying the limbs and respiratory system.

Follow-up electrophysiological studies, performedat least once in 11 patients, were commensurate withclinical observations regarding the course of the poly-neuropathy (table 1). Patients who improved whileshowing little change in the initially minimal derange-ments of speed of impulse conduction, demonstrateda rise in conduction velocities, a decrease in distallatencies, and in particular a rise in the amplitudes ofcompound muscle and nerve action potential ampli-tudes at six months (fig 2). Needle electrode studiesrevealed a gradual disappearance of abnormal spon-taneous activity and a reappearance of motor unitpotentials (fig 5), many of which were initially poly-phasic, consistent with reinnervation of muscle. Thetime course of this reinnervation and the clinicalrecovery, itself, was unusually rapid, suggesting thedistal parts of peripheral nerves had been predom-inantly affected and, hence, the distance for rein-nervation was not great.

Neuropathological findings in critically illpolyneuropathyNecropsy studies of the peripheral and central ner-vous systems were carried out in Patients 3, 4, 5, 9, 13,and 14, the first three of these patients having beenpreviously reported.9 The results for the last threepatients were similar. The findings were consistentwith a moderate to severe, primary, axonal poly-neuropathy affecting sensory and motor fibres, partic-ularly distal fibres. No significant inflammatorychange was seen. In Patients 3 and 4, there was evi-dence of nerve fibre regeneration. Muscles showedacute denervation atrophy, but Patient 3 showed fea-tures suggestive of an additional primary musclechange. However, the central nervous system wasspared, indicating this condition was confined to theperipheral nervous system.

Comparison of critically ill polyneuropathy andGuillain-Barre syndromeThe clinical data on both groups are shown in tables1 and 2. It can be seen that they were similar in sexand severity of polyneuropathy. However, theGuillain-Barre syndrome patients tended to be youn-ger (mean of48 vs 65 years). However, in CIP patientsthe polyneuropathy was noted at the peak of criticalillness and sepsis, in most instances within a month ofadmission to the critical care unit. In Guillain-Barresyndrome patients there were the usual predisposingillnesses (influenza innoculation in two, transientinfection in 11, and no predisposing cause in three)



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


570

which had disappeared before onset of poly-neuropathy. In CIP there were many complicating ill-nesses which began before the onset of poly-neuropathy, consistent with sepsis and critical illness,whereas in Guillain-Barre syndrome, complicationssuch as respiratory and autonomic failure, pneu-monia, occurred after the polyneuropathy had devel-oped.The severity of the polyneuropathy was similar in

the two types. However, it improved in all Guillain-Barre syndrome patients, but in CIP patients wasimproved in six, was unchanged in four andprogressed in three.Necropsy of six CIP patients revealed a primary

axonal degeneration of motor and sensory fibres withsparing of the central nervous system. None of theGuillain-Barre syndrome patients died or had nerveor muscle biopsy.

In electrophysiological studies, the various mea-surements of the speed of impulse conduction (F wavelatency, compound thenar muscle action potentiallatency, and compound digital nerve action potentiallatency (fig 1)) revealed values for CIP that were nearnormal, consistent with a purely axonal degeneration,and there was little change in these values in follow-up. However, Guillain-Barre syndrome patients, ini-tially and particularly at one month, showed consid-erable slowing of impulse conduction in motor andsensory fibres. This tendency was most marked in dis-tal motor fibres, as exemplified by the markedly pro-longed compound thenar muscle action potentiallatency. During the six months of follow-up, theGuillain-Barre syndrome abnormalities tended toreturn towards normal. The amplitudes ofcompoundthenar muscle and compound digital nerve actionpotentials were considerably reduced initially in bothtypes of polyneuropathy (fig 2), although to a greaterdegree in Guillain-Barre syndrome, and more so forcompound muscle action potentials. In follow-up,there was partial resolution in both types of poly-neuropathies. A comparison of the amplitude andshape of compound action potentials on proximalstimulation as compared with distal stimulationrevealed no evidence of conduction block or actionpotential dispersion in critically ill polyneuropathy,whereas these were present in eight of 16 Guillain-Barre syndrome patients.At the time of the initial study, needle electro-

myographic studies in CIP patients revealed relativelyabundant abnormal spontaneous activity, consistentwith denervation atrophy of muscle (fig 5). Suchactivity was minimal in Guillain-Barre syndromepatients, consistent with a mainly demyelinating poly-neuropathy. At one month, there was more electro-physiological evidence of denervation atrophy ofmuscle in the Guillain-Barre syndrome patients, indi-


cating a degree of associated axonal degeneration.The number of motor unit potentials firing onattempted voluntary contraction was markedlydepressed in both neuropathies, but with recovery inboth the number of motor unit potentials recruitablewith voluntary contraction progressively improved.The cerebrospinal fluid revealed only mild

elevations in CIP patients, mean 45 3 + 34-4, butmuch more marked elevations in Guillain-Barre syn-drome patients, mean 109 1 + 79-1 mg/dl. Thedifference of the means was statistically significantusing a two sample t test (p = 0 025).

Discussion

Critically illpolyneuropathyBetween 1977 and 1981, we observed five patientswho suffered from this unusual polyneuropathy.9 Justas they were recovering from their critical illness,there was an inexplicable difficulty in weaning themfrom the ventilator. Deep tendon reflexes previouslypresent were now reduced, and the limbs appearedflaccid and weak. While the face was mildly weak,muscles of the eyes, tongue, and jaw were relativelystrong. Electrophysiological studies were consistentwith a primary axonal degeneration of motor andsensory fibres and acute denervation of muscle.Follow-up revealed early improvement, althoughthree patients later died of causes presumablyunrelated to the polyneuropathy. Necropsy discloseda primary axonal degeneration of peripheral nerveswith sparing of the central nervous system.

Detailed analysis of the records of these fivepatients failed to uncover the aetiology.9 There wasno evidence of heavy metal poisoning, porphyria,viral infection, or specific vitamin deficiency. Allpatients received antibiotics, often in variouscombinations, and all received aminoglycosides, butblood levels of this group of antibiotics were notsignificantly elevated and a search of the literature hasfailed to reveal polyneuropathy as a complication ofsuch treatment. Depressed levels of blood albuminand lymphocytes and observed improvement after theinstitution of total parenteral nutrition suggest, butdo not prove, nonspecific nutritional deficiency as thecause. Since all patients were septic, bacterial toxinwas a possible aetiological factor. However, the onlyspecific toxins known to affect the peripheral nervoussystem are Legionella pneumophilia and Coryne-bacterium diphtheriae, neither of which were culturedin these patients. Nonetheless, the nonspecific effectsof sepsis, which appear to be widespread and as yetpoorly understood,34 may have had an effect on theperipheral nervous system. Finally, Guillain-Barresyndrome has been a well-documented complication



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Critically ill polyneuropathy. electrophysiological studies and differentiation from Guillain-Barre syndrome 571

of severe infection and trauma.'4 However, theunremarkable cerebrospinal fluid protein levels (seeResults), the electrophysiological and necropsyfindings of a primary, axonal degeneration ofperipheral nerve, and the lack of significantinflammatory change in peripheral nerves, all argueagainst this diagnosis.

Preliminary reports from two other centresdescribe severe neuromuscular disease associatedwith critical illness. Roelofs et al5 reported fourpatients who developed a severe, reversible, purelymotor, neuropathy in association with sepsis andsevere respiratory insufficiency. Rivner et al6described four patients who developed a similarmotor neuropathy, presumably as a complication ofsevere, hypotensive shock.The detection of 10 new patients between 1981 and

1983 in our hospital resulted from increased clinicalawareness and, particularly, more frequent orderingof electrophysiological studies.78 Critically ill patientswho were noted to have difficulty in weaning from theventilator and an unexplained weakness of limbmuscles with relative sparing of the face on voluntaryor reflex induced movements, were suspected ofhaving a polyneuropathy. However, as alreadyemphasised, these signs were often equivocal.

All patients were transported to the EMGlaboratory while still on a ventilator and receivingvarious forms of supportive care. Nerve conductionstudies revealed findings typical of a primary axonaldegeneration of motor and sensory fibres. Peripheralsensory nerve and muscle compound action potentialamplitudes were considerably decreased and thevelocity of impulse conduction was normal or mildlyslowed. The amplitude reductions were not due tolimb oedema (which would move the surfacerecording electrode away from the nerve,17 since nearnerve recordings have shown amplitudes in suchpatients remain low or absent) (fig 3). Needleelectromyography provided definitive evidence ofdenervation of muscle. There were numerousfibrillation potentials and positive sharp waves inproximal and distal limb, and intercostal muscles.The latter, coupled with abnormalities of phrenicnerve conduction, strongly implicated denervation ofrespiratory muscles, both chest wall and diaphragm,as the cause of difficulty in weaning from theventilator. Repetitive nerve conduction studies failedto reveal a defect in neuromuscular transmission,even in patients who had received aminoglycosides(fig 6).

In addition to signs of acute denervation of muscle,needle electrode studies revealed unusualspontaneous activity, and myotonic and complexrepetitive discharges in the muscles of seven patients.None of these patients had clinical evidence of either

myotonia or muscle stiffness. Both types of dischargeshave been recorded in a wide variety of both primarymuscle diseases and states of chronic denervation.18This finding, in conjunction with a marked loweringin the amplitude of the compound muscle actionpotential, suggests a predominant involvement ofmotor axons. However, sensory fibres are alsoinvolved. The CSAP was significantly lower thancontrol values (fig 1) and morphological examinationof purely sensory nerves has shown a primary axonaldegeneration.9

Follow-up electrophysiological studies wereconsistent with peripheral nerve axonal regeneration.This regeneration occurred within a few weeks in themilder cases, but even in severe polyneuropathy theperiod of recovery was earlier than one would haveexpected if there had been marked proximal axonaldegeneration. These observations suggest that thedegeneration and regeneration involved distal nervefibres predominantly.

Thus, electrophysiological findings have proven tobe of crucial importance in detecting and followingthe course of the polyneuropathy which complicatesthe course of critical illness. One explanation for apuzzling deterioration, or failure to improve, in apatient's condition, has been provided and with suchknowledge the patient can be managed moreintelligently. Intensive treatment should continue,and should include ventilatory assistance asnecessary, as well as the institution or continued useof total parenteral nutrition. Active physiotherapyshould be prescribed to prevent joint contracturesand pressure sores, which are particularly prone todevelop in critically ill patients who also have apolyneuropathy. Follow-up electrophysiologicalstudies will indicate whether or not thepolyneuropathy is improving. Moreover, with theability of electrophysiological studies to clearlyidentify this type of polyneuropathy, it may bepossible in the future, with more intensive prospectiveinvestigations, to discover the precise aetiology.

Comparison of critically ill polyneuropathy andGuillain-Barre syndromeOn purely clinical grounds it was difficult to dis-tinguish CIP from Guillain-Barre syndrome. In both,the polyneuropathy ran a monophasic course, theonset being relatively acute, with subsequentimprovement occurring in most instances. The clin-ical features were also similar since there was evidenceof muscle weakness in all four limbs, occasionallyinvolving facial muscles and frequently involving themuscles of respiration; the deep tendon reflexesbecame depressed or absent, and there was some evi-dence of distal sensory impairment. The predisposingcauses were, however, different. CIP invariably



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


572

occurred at the peak of critical illness and sepsis, butin Guillain-Barre syndrome there was a brief periodof recovery following a relatively minor illness orinnoculation. Finally, the prognosis was poorer inCIP, three patients failing to show improvement inthe polyneuropathy, whereas all Guillain-Barre syn-drome patients improved.

Electrophysiological differences also distinguishedthe two types of polyneuropathies. In CIP these fea-tures were typical of a primary axonal degenerationof motor and sensory nerves:19 a relative preservationof the speed of impulse conduction, reduction in theamplitudes of muscle and sensory compound actionpotential amplitudes, no evidence of conductionblock, and abundant spontaneous activity on needleelectromyography of muscle, consistent withdenervation atrophy. Alternatively, Guillain-Barresyndrome patients showed considerable prolongationof impulse conduction, particularly distally, evidenceof dispersed compound action potentials or of con-

duction block, and much less abnormal spontaneousactivity in muscle, as one would see in a predom-inantly demyelinating polyneuropathy.10 19 20 Thecerebrospinal fluid protein showed statisticallysignificant higher levels in Guillain-Barre syndromethan in critically ill polyneuropathy, a further methodof distinguishing the two polyneuropathies.21 There-fore, these two polyneuropathies are almost certainlyseparate entities that can be distinguished from one

another by the above studies in most cases. However,the relatively large standard deviations in electro-physiological and cerebrospinal fluid results in bothgroups indicates significant variation and there maybe difficulties in distinguishing between the two neu-

ropathies in some instances. However, patients whodevelop polyneuropathy, mild or severe, at the peakof critical illness or sepsis, have electrophysiologicalfeatures typical of a primary degeneration of motorand sensory fibres, and have a relatively normal cere-

brospinal fluid protein, likely have CIP. While theaetiology of CIP remains uncertain, a purely axonaldegeneration is more suggestive of a toxic or meta-bolic disturbance22 - 24 than of inflammatory or

immune-mediated demyelination.It was instructive to compare the two neuropathies

from the point of view of electrophysiology. In crit-ically ill polyneuropathy, motor and sensory fibreslikely degenerate in a random manner, some fasterconducting fibres being preserved, accounting for a

relatively normal speed of maximum impulse conduc-tion (fig 1). However, the fall-out of these fibrescauses reduction in muscle and sensory compoundaction potential amplitude (fig 2), with no evidence ofdispersion of the compound action potential, or ofconduction block. This pattern, as already noted ischaracteristic of a primary axonal degeneration.9


The electrophysiological findings in Guillain-Barresyndrome are consistent with a primary segmentaldemyelination in which even the fastest conductingfibres are involved and there is slowing of the max-imum rate of impulse conduction. Disproportionateslowing of conduction in some fibres results in a dis-persion of the compound action potential. Severedemyelination results in conduction block in whichcompound action potential areas are larger on distalstimulation than on proximal stimulation. One expla-nation for the reduced CMAP and SNAP amplitudes(fig2) is the possibility that the demyelination pre-dominantly affected the distal portions of the nervefibres, beyond the conventional point of distal stimu-lation.25 This possibility is supported by the fact that,initially, needle electromyography revealed little evi-dence of denervation atrophy of muscle (fig 5), includ-ing distal muscles such as the first dorsal interosseusof the hand and the extensor digitorum brevis of thefoot. The increased spontaneous activity at onemonth indicated a degree of axonal damage haddeveloped. Indeed, in three of our Guillain-Barr6 syn-drome patients, impulse conduction was not particu-larly slowed and spontaneous activity was moreabundant in muscle, indicating the main pathologywas axonal degeneration.

This pattern of nerve conduction and needle elec-trode abnormality in Guillain-Barre syndrome is con-sistent with observations by others. McLeod20observed similar marked prolongation of distal motorlatency and reductions in CSAP amplitudes. The pro-longed F wave latencies were similar to Kimura'spatients.26 Brown and Feasby25 performed severaltechniques and analyses which allowed them to con-clude that reduced CMAP amplitude was due to thecombination of conduction block and denervationatrophy. By stimulating beyond the conventionalpoint of distal stimulation, that is the median nerve inthe palm and peroneal nerve in the foot, they wereable to produce a further rise in amplitude, stronglysuggesting the presence of a distal block.

In both CIP and Guillain-Barre syndrome, electro-physiological abnormalities gradually returnedtowards normal after the first month, consistent withaxonal regeneration and segmental demyelination,respectively. In CIP, the degree of clinical and electro-physiological recovery was surprisingly rapid, sug-gesting axonal degeneration and regeneration werepredominantly distal. In both groups, residual clinicaland electrophysiological abnormalities were presentat 6 months, indicating recovery had been incomplete.

We thank personnel in the Critical Care/TraumaUnit, Mrs Wilma Koopman, Neuromuscular ClinicCo-ordinator, and Miss Betsy Toth, Secretary. Dr NJ



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



Witt is supported by an Alberta Heritage FoundationFellowship.

References

Pine RW, Wertz MJ, Lennard ES, Dellinger EP, CarricoCJ, Minshew BH. Determinants of organ malfunctionor death in patients with intra-abdominal sepsis. ArchSurg 1983;118:242-9.

2Jackson AC, Bolton CF, Young GB, Gilbert JJ. Theencephalopathy of sepsis. Ann Neurol 1983;14: 141.

3 Beisel WR. Mediators of fever and muscle proteolysis.N Eng J Med 1983;308:586-7.

4Clowes GH, George BC, Villee CA, Saravis CA. Muscleproteolysis induced by a circulating peptide in patientswith sepsis or trauma. N Eng J Med 1983;308:545-52.

5Gertz J, Hedenstierna G, Hellens G, Wahren J. Musclemetabolism in patients with chronic obstructive lungdisease and acute respiratory failure. Clin Sci Mol Med1977;52:395.

6Roussos C, Macklem PT. The respiratory muscles. N EngJ Med 1982;307:786-97.

7Bolton CF, Brown JD, Sibbald WJ. The electro-physiologic investigation of respiratory paralysis in crit-ically ill patients. Neurology (Cleveland) 1983;33:3:186.

8Bolton CF, Brown JD, Sibbald WJ. Polyneuropathy asso-ciated with major systemic illness. Can J Neurol Sci1984;11:2:348.

9 Bolton CF, Gilbert JJ, Hahn AF, Sibbald WJ. Poly-neuropathy in critically ill patients. J Neurol NeurosurgPsychiatry 1984;47: 1223-31.

Asbury AK. Diagnostic considerations in Guillain-Barresyndrome. Ann Neurol 198 1;9(suppl): 1-5.

Bolton CF, Sawa GM, Carter K. The effects of tem-perature on human compound action potentials. J Neu-rol Neurosurg Psychiatry 198 1;44:407-13.

12Bolton CF, Carter K, Koval JJ. Temperature effects onconduction studies of normal and abnormal nerve.Muscle Nerve 1982;5:S145-7.

1Bolton CF, Carter K. Human sensory nerve compoundaction potential amplitude. Variation with sex andfinger circumference. J Neurol Neurosurg Psychiatry1980;43:925-8.

14 Amason BGW. Acute inflammatory poly-radiculoneuropathies. In: Dyck PJ, Thomas PK, Lam-bert EH, Bunge R. Peripheral Neuropathy. Philadel-phia: WB Saunders, 1984:2051-5.

s Roelofs RI, Cerra F, Bielka N, Rosenberg L, Canton OH,Delaney J. Prolonged respiratory insufficiency due toacute motor neuropathy: a new syndrome? Neurology(Cleveland) 1983;33(suppl):240.

16Rivner MH, Kim S, Greenberg M, Swift TR. Reversiblegeneralized paresis following hypotension: a new neu-rological entity. Neurology (Cleveland)1983;33(suppl): 164.

7Bolton CF. Factors affecting the amplitude ofhuman sen-sory compound action potentials. American Associationof Electromyography and Electrodiagnosis. Min-imonograph #17, October 1981.

'8Kimura J. Electrodiagnosis in Diseases of Nerve andMuscle: Principles and Practice. Philadelphia: FADavis, 1983:274-7.

9 Brown WF. The Physiological and Technical Basis ofElec-tromyography. Boston: Butterworth Publishers,1984:xi-xiv.

20McLeod JG. Electrophysiological studies in the Guillain-Barre syndrome. Ann Neurol 1981 ;9(suppl):20-7.

21Amason BGW. Acute inflammatory demyelinating poly-radiculoneuropathies. Chapt. 90. In: Dyck et al, eds.Peripheral Neuropathy. Philadelphia: WB SaundersCompany, 1984:2059.

"Spencer PS, Schaumburg HH. Pathobiology of neu-rotoxic axonal degeneration. In: Waxman SC, ed. Phys-iology and Pathobiology of Axons. New York: RavenPress, 1978:265-82.

23Schaumburg HH, Spencer PS. Toxic neuropathies. Neu-rology (Minneap) 1979;29:429-31.

24Erbsloh F, Abel M. Deficiency neuropathies. In: VinkenP, Bruyn GW, eds. Handbook of Clinical Neurology.New York: North Holland, 1970;7:558-663.

2S Brown WF, Feasby TE. Conduction block anddenervation in Guillain-Barre polyneuropathy. Brain1984;107:219-39.

26 Kimura J. The F-wave. Chapter 17. In: Electrodiagnosis inDisease of Nerve and Muscle: Principles and Practice.Philadelphia: FA Davis, 1983:364-9.



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


critically electrophysiological guillain-barrecritical illness meantsignificant involvement...

Documents