Neuromuscular Respiratory Failure in Guillain-Barre Syndrome: Evaluation of Clinical and Electrodiagnostic Predictors

Original Article Neuromuscular Respiratory Failure in Guillain-Barre Syndrome: Evaluation of Clinical and Electrodiagnostic Predictors Uma Sundar, Elizabeth Abraham, A Gharat, ME Yeolekar, Trupti Trivedi, N Dwivedi Abstract Guillain- Barre Syndrome (GBS) has an unpredictable clinical course with up to 30% of patients requiring assisted ventilation during the course of their illness. Successful management mandates anticipation, prompt recognition and optimal treatment of neuromuscular respiratory failure in GBS. Aims : To identify clinical and electrodiagnostic predictors of neuromuscular respiratory paralysis in GBS. Materials and Methods : Fourty six patients of GBS were studied over a 6 year period, the study being 2 year retrospective and 4 year prospective. Clinical and electrodiagnostic data were compared between ventilated (28) and non-ventilated (18) patients. The clinical parameters assessed were median age, gender, antecedent infection, prior lung disease, time to peak disability, bifacial weakness, upper limb weakness, bulbar paralysis, neck weakness and autonomic dysfunction. Electrodiagnostic studies included motor nerve conduction studies in 11 ventilated and 13 non-ventilated patients, done prior to maximum disability in each group. Multiple logistic regression analysis was used to compare the two groups. Results : Comparing the clinical data in the ventilated and non-ventilated groups, early peak disability, autonomic dysfunction and bulbar weakness predicted the onset of respiratory paralysis. Age, gender, neck or bifacial weakness, upper limb paralysis, or preceding infection did not influence the development of neuromuscular respiratory weakness. Electrodiagnostic testing revealed abnormal H reflex and F waves to be the commonest abnormality in either group. Although data was not sufficient for statistical analysis, the presence of markedly attenuated Compound Muscle Action Potentials, inexcitable motor nerves and denervation changes on the electromyography, was commoner in the ventilated group. Thirty six patients received treatment with either plasmapheresis (12 ) or intravenous immunoglobulin (24). Overall mortality was 5, all 5 patients being on assisted ventilation. Conclusion : Early progression to peak disability, bulbar dysfunction and autonomic instability predicted the development of neuromuscular respiratory paralysis in GBS. Early electrodiagnostic studies in this series suggest axonopathic GBS as a predictor of respiratory paralysis, a finding that needs to be evaluated with sufficient data to permit statistical analysis. INTRODUCTION Guillain-Barre syndrome (GBS) is the commonest cause of acute flaccid quadriparesis. It is an acute inflammatory demyelinating polyradiculoneuropathy with an incidence of 0.6-1.5/100,000, diagnosis being based on a set of defined clinical and laboratory criteria. 1 GBS has an unpredictable course, with a mortality of 5-7%. Poor prognostic factors are advanced age, autonomic dysfunction, rapid progression, neuromuscular Department of Medicine- Neuromedicine Subdivision and Medical ICU, Lokmanya Tilak Municipal Medical College and Hospital, Sion, Mumbai. Received : 23.1.2004; Revised : 18.7.2005; Accepted : 29.7.2005 respiratory failure, and EMG findings of low Compound Muscle Action Potential (CMAP) amplitude with early denervation. 2,3 Up to 30% of patients develop a rapidly progressive disease with respiratory muscle weakness requiring mechanical ventilation. 4,5 As compared to the West, reports from India seem to indicate increased mortality, and overall a more fulminant form of the disease. 6,7 It is essential to identify early signs of neuromuscular respiratory failure, so as to enable optimum ventilatory management, and effect early ICU transfer. Aims of the present study was to identify clinical and electrodiagnostic predictors of neuromuscular respiratory failure in GBS. 764 www.japi.org JAPI VOL. 53 SEPTEMBER 2005

MATERIALS AND METHODS The study was conducted at a tertiary care hospital, over a 6 year period, being a 2-year retrospective and 4- year prospective study. Of 82 patients admitted with a diagnosis of acute flaccid paralysis during this period, 46 satisfied the Asbury and Cornblath criteria for GBS diagnosis 1 28 of these went on to require mechanical ventilation. Clinical and demographic data were compared between the ventilated (28) and nonventilated (18) groups, end point in time being taken as time of intubation in the ventilated group, and as maximum motor disability, as per Hughes score, in the non- ventilated group. Parameters compared (List 1) were median age, gender, antecedent respiratory or gastrointestinal infection, prior lung disease such as asthma or COPD, time to peak disability, bifacial weakness, upper limb weakness, bulbar paralysis (dysphagia, dysarthria, poor cough, palatal weakness), neck muscle weakness and autonomic dysfunction (resting tachycardia, labile blood pressure or new-onset hypertension). Cerebrospinal fluid (CSF) studies (routine biochemistry and cell count) prior to maximum disability were available in all nonventilated and 16 ventilated patients. Electrodiagnostic studies done prior to maximum disability were available in 11 ventilated and 13 nonventilated patients. Standard nerve conduction velocity studies were done, using surface recording electrodes. Motor nerve conduction studies included recording from abductor pollices brevis on median nerve stimulation, from extensor digitorum brevis on peroneal nerve stimulation, from abductor digiti minimi on ulnar nerve stimulation and from abductor hallucis on tibial nerve stimulation. These evaluated F waves (absence, latency, chronodispersion) in multiple motor nerves and H reflex (amplitude, latency) on stimulation of Posterior Tibial nerve. Motor nerve conduction studies included assessment of CMAP amplitude, distal motor latency (DML) and motor nerve conduction velocity (MNCV) along with assessment for temporal dispersion and conduction block. Antidromic studies on median, ulnar, superficial radial and sural nerve yielded sensory nerve conduction studies including sensory nerve action potential (SNAP) and sensory nerve conduction velocity (SNCV). Standard criteria, in comparison with standards for the particular laboratory, were applied to label a particular value as abnormal. Distal motor latency was prolonged if it was more than 150% of upper limit of normal. Motor conduction velocity was slowed if it was less than 70% of lower limit of normal and F wave latency was prolonged if more than 150% of upper limit of normal. Conduction block implied a 30% drop in CMAP amplitude on proximal stimulation as compared to distal stimulation, and temporal dispersion implied a 20% increase in CMAP dispersion on proximal stimulation, both these parameters indicating a proximal conduction block. Electromyography was available in 16 patients. NCV/EMG data were compared between 11 ventilated and 13 non-ventilated patients, the study having been done before maximum disability set in, in each of these two groups. Statistical analysis The 2 groups of ventilated and non-ventilated GBS were compared, using Fisher exact test for categorical variables. Logistic regression analysis with a background elimination process of non-significant variables was used, to assess multivariate predictors of mechanical ventilation. Statistical analysis of comparative data pertaining to electrodiagnostic studies was not possible due to insufficient sample size. RESULTS Of 82 patients with an admission diagnosis of acute flaccid paralysis, 46 satisfied criteria for GBS. Of the remaining 36, 24 had hypokalemic periodic paralysis, four had acute transverse myelitis, six compressive cervical myelopathy in spinal shock, and two had dumb rabies. On comparison of data in ventilated and nonventilated groups (Table 1), the following factors did not predict requirement for mechanical ventilation- age, gender, preceding infection, bifacial or neck weakness, and early upper limb paralysis. Time to peak disability was significantly shorter in the ventilated group (33 hours) as compared to non-ventilated group (6 days). The presence of bulbar palsy (18/28 in ventilated vs. 2/ 18 in non-ventilated) and autonomic dysfunction (14/ 28 ventilated vs. 1/18 non-ventilated) were significantly commoner in the ventilated group. Multiple logistic regression analysis on these data using mechanical ventilation as the independent variable, and the clinical and demographic data mentioned above as dependent variables, showed early peak disability, bulbar dysfunction, and autonomic instability to be Table 1: Demographic and clinical variables in ventilated and non-ventilated GBS patients Variable Ventilated Non-ventilated patients (28) patients (18) No. (%) No. (%) Age range (yrs) 14-63 median 36 17-72 median 46 Gender M: F 19:9 13:5 Antecedent lung disease 2 ( 8.23) 1(5.5) Preceding infection 15 ( 53.6) 9 (50) Time to peak disability (median.) 33 hrs 6 days * Bulbar weakness 18 ( 64.5) 2 (10.1)* Bifacial weakness 13 (46.3) 5 (27.7) Neck weakness 14 (50) 2 (10.1) Early upper limb weakness (<2 days) 11 (45.83) 7 (38.8) Autonomic dysfunction 14 (50) 1(5.5)* * P value < 0.05 (significant) JAPI VOL. 53 SEPTEMBER 2005 www.japi.org 765

independent predictors of mechanical ventilation in GBS. Data was collected on the timing of intubation. Sixteen of 28 intubations were done during daytime, and 12 at night. Twelve of 28 intubations were elective- this was judged on the basis of presence of serial respiratory monitoring including single breath count, chest expansion and ABG studies, prior to the decision to intubate. Electrophysiological studies done prior to maximum disability were available in 11 ventilated and 13 nonventilated patients (Table 2). Abnormal H-reflex, and F wave abnormality (absent or prolonged latency) were the commonest abnormalities in either group. H reflex was performed in nine ventilated patients, being abnormal in seven (absent in six, and low amplitude in one). In non-ventilated patients, it was done in 11 patients, seven being abnormal (absent in all 7). F waves were judged to be abnormal if one or more motor nerve studies showed an abnormality (absent, or prolonged latency). F waves were abnormal in 10/11 ventilated, and 12/13 non-ventilated patients respectively. Marked reduction in CMAP amplitude in more than one motor nerve tested was seen in 6/11 ventilated patients and in 2/13 non-ventilated patients. In the ventilated group, two patients had two or more inexcitable nerves. Temporal dispersion (4/11 ventilated vs. 4/13 nonventilated), Conduction block (1/11 ventilated vs. 3/13 non-ventilated) and significant MNCV slowing (3/11 ventilated vs. 5/13 non-ventilated) were infrequent in Table 2 : Early Electrodiagnostic findings in 24 GBS patients (ventilated vs. non-ventilated) prior to maximum disability NCV/EMG variable Ventilated Non-ventilate patients (11) patients (13) H Reflex abnormality 7/9 7/11 F Wave abnormality 1/11 12/13 SNAP abnormality 7/11 8/13 Markedly low CMAP in 1 or more nerves 6/11 2/13 Inexcitable nerves-1 or more 2 Temporal dispersion 4/11 4/13 Conduction Block 1/11 3/13 MNCV < 70% Normal 4/11 5/13 Denervation on EMG in 2 or more muscles 2/11 NCV= Nerve conduction velocity. EMG= Electromyography. SNAP= Sensory nerve action potential. CMAP= Compound muscle action potential. MNCV= Motor nerve conduction velocity both groups. Denervation changes on the EMG in more than one muscle tested were seen in 5/11 ventilated patients. Sensory nerve action potentials (either upper limb or lower limb) were abnormal in 7/11 ventilated and 8/13 non-ventilated patients. Prolongation of distal motor latency in two or more nerves was seen in 5/11 ventilated and 6/13 non-ventilated patients. Therapy with either plasmapheresis or intravenous immunoglobulin (Iv IgG) was instituted as per standard criteria for treatment initiation, in 12 and 24 patients respectively (Table 3). Seven patients in plasmapheresis group, and 18 in Iv IgG group recovered completely at discharge, average time to improvement being 9 days after 1 st cycle and 8 days after 1 st dose respectively. Significant residual deficit at discharge (mean of 24 days after admission) was present in five patients from the Plasmapheresis group, and nine from the Iv IgG group. Mortality was 1/12 and 2/24 in the Plasmapheresis and IvIgG group respectively. Two patients who had received plasmapheresis worsened in the 1 st week, and were given additional IvIgG course, after which one made a partial recovery by discharge. One patient worsened on 10 th day after institution of IvIgG therapy, and was given a second course of IvIgG, alternating with methyl prednisolone 500 mg. per day, following which he made a complete recovery. Total mortality was 5/46 patients, three belonging to the treated group, and two to the untreated group. All five patients were on assisted ventilation, three of them satisfying the criteria for ventilator-associated pneumonia. 2/5 had additional autonomic dysfunction, one each from treated and untreated group, which could have contributed to the terminal cardiorespiratory arrest. DISCUSSION Neuromuscular respiratory failure is one of the major factors influencing morbidity and mortality in GBS. The respiratory failure in GBS results from a combination of factors- tongue, pharyngeal and laryngeal weakness causing poor clearing of oral secretions, and diaphragmatic-intercostal weakness causing progressive type2 respiratory failure, leading to atelectasis and progressive hypoxia. 8 Successful management of GBS mandates anticipation of respiratory failure and timely intervention. Ropper and Kehne s established criteria for intubation include bulbar weakness, vital capacity <15ml/kg, and po 2 on room air <70mm Hg. 9 Lawn et al proposed the 20-30-40 rule, whereby intubation was considered if Table 3 : Treatment modalities and outcome in 46 GBS patients Treatment modality No. Recovered Avg. time to Mortality Residual deficit improvement (days) at discharge Plasmapheresis(12) 7 9 2/12 5/12 IV IgG (24) 18 8 3/24 9/24 766 www.japi.org JAPI VOL. 53 SEPTEMBER 2005

the vital capacity, maximum inspiratory pressure and maximum expiratory pressure fell below 20ml/kg, 30ml/ kg, and 40ml/kg respectively. 10 These cut-off points mandated a regular 8-hourly monitoring of respiratory parameters. Our study attempted to identify clinical features that would predict respiratory failure, with additional data such as single breath count, and blood gas determination of pco 2 and po 2 levels. These are simple bedside tests available in most hospitals. In our study, age and gender did not influence development of respiratory failure. Bifacial weakness could contribute to impaired swallowing and hence respiratory morbidity, as lip closure facilitates swallowing mechanism. However, facial as well as neck weakness were not independent predictors of respiratory failure in our study. Prior infection with campylobacter jejuni carries a worse prognosis, due to higher incidence of development of axonopathic variant of GBS. 11 Although a total of 24 patients had prior infection (16 gastrointestinal, and 8 respiratory), no significant difference was seen between the ventilated and non-ventilated groups. Stool or serological testing for C. jejuni was not done in this study. Autonomic dysfunction, rapid disease progression to maximum debility, and bulbar weakness were all independent predictors of neuromuscular respiratory weakness, in our study. Bulbar weakness is directly related to requirement for intubation, due to lack of oral and pharyngeal protective mechanisms in these patients, and qualifies as one of the indications for elective early intubation in patients with GBS. Autonomic dysfunction, besides presaging a worse form of GBS, due to involvement of cardiac and vagal nerves in addition to the motor nerves, directly effects need for intubation, probably via depression of the autonomically innervated respiratory drive. Autonomic dysfunction, per se, would be sufficient indication to transfer the patient to an intensive care setting for serial monitoring. 2,3 Twelve of the total of 28 intubations were clearly elective, with sufficient respiratory data prior to intubation to justify a decision to electively intubate. Of the 16 patients who did not fit into the electively intubated group, 12 were retrospectively studied, and insufficient data recording cannot be ruled out in them. The remaining four were prospectively studied-they were ward intubations, and all were intubations done at night. These were patients with only mild bulbar weakness and no autonomic dysfunction, and hence were monitored in the ward. Poor pulmonary mechanics at night, compounded by prolonged interval of time between monitoring at night, could have led to the intubation being non-elective in them. Among the three patients who were on ventilator and subsequently died, only one had been electively intubated. In a recent study pertaining to correlation of morbidity and mortality in GBS with intubation data (emergent or elective), Wijdicks et al retrospectively analysed 20 years GBS data, comparing 36 elective intubations with six emergency intubations during this period. No difference in mortality, pulmonary morbidy or duration of ventilation assistance was noticed, although one patient suffered anoxic encephalopathy following emergency intubation for respiratory arrest. 12 Electrophysiological testing prior to maximum disability was compared in the two groups. These were done at an average of 28 hours and six days after onset of weakness in the ventilated and non- ventilated groups respectively. The commonest finding in the two groups was a radiculopathy, as evidenced by impaired H reflex and F waves in the motor nerves. Markedly low CMAP amplitudes were seen in 6/11 ventilated patients. This could represent either severe distal demyelination or axon loss. Only 2/13 patients in the non-ventilated group had this finding. Evaluation of motor nerves by quantitation of CMAP amplitudes is one of the most important early studies in GBS. In our study, a low CMAP amplitude on distal stimulation, which could indicate either distal demyelination or axonopathy, was overall a much commoner finding (14/24) than temporal dispersion or conduction block, each of which indicates more proximal segmental demyelination. Three studies have shown that a markedly low distal CMAP indicates poor prognosis in GBS. Cornblath et al, from pooled data related to the North American Guillain-Barre Plasmapheresis trial, found a CMAP amplitude less than 20% of lower limit of normal, to be associated with a poor prognosis 13. Miller et al studied 65 GBS patients at a mean of 18 days after onset of weakness and found poor prognosis to be related to CMAP amplitude less than 10% of lower limit of normal. 14 Feasby et al studied the median, ulnar, and peroneal nerves of 25 patients with GBS prospectively and concluded that CMAP amplitudes of 10% of control mean for at least one nerve had a poorer outcome. 15 The presence of inexcitable motor nerves is associated with a poor outcome (Feasby 1986, Miller 1982, 1988). 14,15 In Brown, Feasby et al s data on detailed serial physiological observation of the function of distal motor nerves, three patients with rapidly progressive GBS were studied. They showed progressively increasing conduction block, leading to total loss of excitability in the terminal segments of motor nerves, developing within a few days of onset of paralysis. Additionally, subsequent appearance of fibrillation activity with very delayed recovery from the paralysis, and finally, biopsy confirmation, proved the presence of axonopathy in these patients. 16 In the present series, six of 11 ventilated patients, and two of 13 nonventilated patients, had markedly attenuated CMAP amplitudes in more than one motor nerve. In addition, in the ventilated group, two patients had more than one inexcitable motor nerve, this finding being absent in the non-ventilated group. Denervation changes in more than JAPI VOL. 53 SEPTEMBER 2005 www.japi.org 767

one muscle on EMG testing, were seen in five of 11 ventilated patients, and none of the non-ventilated group. Distal motor latency was prolonged in five ventilated, and six non-ventilated patients. Among the five patients in the ventilated group with prolonged distal latency, three also had markedly low CMAP amplitude. Low CMAP amplitude on distal stimulation is not necessarily a sign of axonopathy, as distal demyelination could give a similar finding. However, the two patients with inexcitable nerves, and the five patients with denervation on EMG, all belonged to the group of six patients with low CMAP amplitudes, in the ventilated group, making a case for Axonopathic GBS in these patients stronger. A combination of Axonopathic and Demyelinating GBS was seen in two patients in the ventilated group. A follow-up EMG in all our patients with low CMAP amplitudes might have confirmed the presence of axonopathy by showing diffuse denervation changes, but this was not uniformly available in all the patients, due to some being ventilator dependent. The presence of demyelination and axonopathy together in GBS has been well documented by Brown et al. This combination, in their view, could represent co-existent pathologies; alternatively, axonal degeneration could be secondary to severe demyelination, the so called bystander effect. 16 An axon loss variant of GBS occurs in 11% of patients, and is associated with rapid onset and progression. It may show a slower recovery from the weakness. 17 Occasionally, the axonal variant of GBS may be mimicked by secondary axonal damage in a primary demyelinating GBS. Although 36 of 46 GBS patients received therapy with either plasmapheresis or Iv IgG, the independent effect of therapy on the need for intubation could not be assessed, as a significant number of the patients were started on treatment after maximum disability had set in. Average time for starting treatment was 56 hours in the ventilated group, which was beyond the time to maximum disability -i.e., intubation, in this group. In the non-ventilated group, average time to starting treatment was 4 days. The effect of treatment on overall mortality and residual deficit at discharge, however, could be compared. Mortality in the treated group (3/ 36) and in the untreated group (2/10} was not statistically different. All five patients who expired were on mechanical ventilation. Three of them had evidence of ventilator associated pneumonia with methicillinresistant Staphylococcus aureus. Autonomic dysfunction, manifesting as cardiac arrhythmia and labile blood pressure, was an additional compounding factor in the other two. Early institution of therapy in GBS, with either plasmapheresis or IvIgG, has been shown to shorten the course of GBS, reduce the time required to receive mechanical ventilation, and lessen the overall severity of disease. 18-20 REFERENCES 1. Asbury AK, Cornblath DR. Assessment of current diagnostic criteria for Guillain- Barre Syndrome. Ann Neurol 1990;27(suppl):S21-S24. 2. McKhann GM, Griffin JW, Cornblath DR, Mellits ED, Fisher RS, Quaskey SA, for the Guillain-Barre Syndrome Study Group. Plasmapheresis and Guillain-Barre syndrome: analysis of prognostic factors and the effect of plasmapheresis. Ann Neurol 1988;23:347-53. 3. Winer JB, Hughes RA, Osmond C. A prospective study of acute idiopathic neuropathy, clinical features and their prognostic value. J Neurol Neurosurg Pshychiatry 1988;51:605-12. 4. Teitelbaum JS, Borel CO. respiratory dysfunction in Guillain- Barre syndrome. Clin Chest Med 1994;15:742-6. 5. Gracy DR, McMichan JC, Diverte MB, Howard FM. Respiratory failure in Guillain Barre syndrome. A 6-year experience. Mayo Clinic Proceedings 1982;57:742-6. 6. Gnanamuthu C, Ray D. outcome of patients with fulminant Guillain-Barre syndrome on mechanical ventilatory support. Indian J Chest Dis Allied Sci 1992;34:65-72. 7. Taly AB, Gupta SK, Vasanth, et al. Critically ill Guillain- Barre syndrome. J Assoc Physicians India 1994;42:871-4. 8. Angelika FH. Editorial: The challenge of respiratory dysfunction in Guillain-Barre syndrome. Arch Neurol 2001;58:871-2. 9. Ropper AH, Kehne SM. Guillain-Barre syndrome : management of respiratory failure. Neurology 1985;35:1662-5. 10. Jeremy H Rees, Sara ES, Norman AG, Richard ACH. Campylobacter jejuni infections and Guillain-Barre syndrome. N Eng J Med 1995;333:1374-79. 11. Lawn ND, Fletcher DD, Henderson RD, Wolter TD, Wijdicks EFM. Anticipating mechanical ventilation in Guillain-Barre syndrome. Arch Neurol. 2001;58:893-8. 12. Wijdricks EF, Henderson RD, McDolland RL. Emergency intubation in Guillain-Barre syndrome. Arch Neurol 2003;69:947-8. 13. Cornblath DR, Mellits ED, Griffin JW, McKhann GM, Albers JW, Miller RG, Feasby TE, Quaskey S and the Guillain-Barre Study Group. Motor conduction studies in Guillain-Barre syndrome: description and prognostic values. Ann Neurol 1988;23:354-58. 14. Miller RG, Peterson GW, Daube JR, Albers JW. Prognostic value of electrodiagnosis in Guillain-Barre syndrome. Muscle Nerve 1988,11:769-74. 15. Feasby TE, Gilbert JJ, Brown WF, Bolten CF, Hahn AF, Koopman WF, Zochodne DW. An acute axonal form of Guillain-Barre polyneuropathy. Brain 1986;109:1115-26. 16. William FB, Thomas EF, Angelika FH. Electrophsiological changes in acute axonal form of Guillain-Barre syndrome. Muscle Nerve 1993;16:200-5. 17. Brown WF, Feasby TE, Hahn AF. Electrophsiological changes in the acute axonal form of Guillain-Barre syndrome. Muscle Nerve 1993;16:200-05. 18. The Guillain-Barre Syndrome Study Group. Plasmapheresis and acute Guillain-Barre syndrome. Neurology 1985;35:1096-1104. 19. French Cooperative Group on Plasma Exchange in Guillain- Barre Syndrome. Efficiency of plasma exchange in Guillain- Barre syndrome: role of replacement fluids. Ann Neurol 1987;22:753-61. 20. Van der Meche FG, Schmitz PIM. For the Dutch Guillain- Barre Study Group. A randomized trial comparing intravenous immune globulin and plasma exchange in Guillain-Barre syndrome. N Engl J Med 1992;326:1123-29. 768 www.japi.org JAPI VOL. 53 SEPTEMBER 2005