Over a period of time, Guillain–Barre syndrome (GBS) has diversified to various subforms, including acute inflammatory demyelinating polyradiculoneuropathy (AIDP), acute motor axonal neuropathy (AMAN), acute motor sensory axonal neuropathy (AMSAN), Miller–Fisher variant, pure sensory and pure autonomic varieties. The different manifestations of GBS may be due to different triggering factors in different geographical areas and genetic susceptibility. A higher prevalence of the AMAN variety has been reported from China.1 Neurophysiological changes have a crucial role in the diagnosis of GBS. Its sensitivity is however heavily depended on the demyelinating nature of the peripheral nerves. To improve the diagnostic yield, various neurophysiological criteria have been proposed. The diagnostic yield of these criteria ranges between 21% and 72% in clinically suspected GBS.2^–7 No study from India has evaluated the sensitivity of all of these criteria in GBS patients; one study evaluated three of these criteria and found a low sensitivity.8 Here we report the sensitivity of various neurophysiological criteria in GBS patients and their correlation with clinical presentation and outcome.

SUBJECTS AND METHODS

Consecutive patients with GBS admitted between 2004 and 2006 were included in the study. GBS was diagnosed on the basis of clinical and CSF criteria, and hypokalaemia or hyperkalaemia and porphyria were excluded. The clinical criteria included: rapidly progressive quadriparesis with or without cranial nerve palsy, reduced or absent reflexes, reduced or normal sensations, absence of fever and persistent bowel and bladder involvement and progression not exceeding 4 weeks. CSF criteria included raised protein with less than 10 lymphocytes/mm3. Two patients with pure sensory GBS did not have weakness. None of the patients had positive HIV serology. Two patients with Miller–Fisher syndrome were excluded as this study evaluated the sensitivity of nerve conduction studies (NCS) in the diagnosis of GBS using conduction of limb nerve. Patients underwent to a detailed clinical evaluation, including history of preceding illness. Presence of cranial nerve palsy was recorded. Muscle power was assessed using the Medical Research Council (MRC) scale, and muscle tone, tendon reflexes, sensations and cerebellar signs were noted. The presence of autonomic dysfunction was also noted. Disability was graded according to a scale from 0 to 10 scale9 as follows:

• Grade 0 = normal.

• Grade 1 = no disability, minor sensory signs or areflexia.

• Grade 2 = mild disability; ambulatory for >200 m; mild weakness in one or more limbs and sensory impairment.

• Grade 3 = moderate disability; ambulatory for >50 m without stick; moderate weakness MRC grade 4 and sensory impairment.

• Grade 4 = severe disability; able to walk >10 m with support of stick; motor weakness MRC grade 4 and sensory impairment.

• Grade 5 = requires support to walk 5 m; marked motor and sensory signs.

• Grade 6 = cannot walk 5 m, able to stand unsupported and able to transfer to wheelchair, able to feed independently.

• Grade 7 = bed ridden, severe quadriparesis; maximum strength MRC grade 3.

• Grade 8 = respirator and/or severe quadriparesis; maximum strength MRC grade 2.

• Grade 9 = respirator and quadriplegia.

• Grade 10 = dead.

Full blood count, haemoglobin, red blood cell indices, erythrocyte sedimentation rate at the first hour, fasting blood sugar, serum sodium, potassium, bilirubin, albumin, transaminases, creatinine and creatine phosphokinase, HIV serology, urine porphobilinogen, electrocardiogram and radiograph of the chest were carried out in all patients. Lumbar CSF was analysed in all patients, including sugar, protein and cells. Patients with a progressive clinical course or immobility due to weakness were treated with intravenous immunoglobulin (IVIg) infusion 400 mg/kg/day for 5 days. Patients were followed-up at 3 months and outcome was defined as complete (independent for activities of daily living), partial (partially dependent for activities of daily living), poor (bedridden) and death.10

Neurophysiological examination

All patients underwent NCS of at least six motor (median, ulnar, common peroneal bilaterally) and sensory (median, ulnar and sural bilaterally) nerves and F wave studies using standard techniques. The parameters included in the motor NCS were distal latency, conduction velocity, distal compound muscle action potential (CMAP) amplitude, conduction block (>20% drop in proximal CMAP to distal) and minimal F wave latency. Conduction velocity and conduction block were measured in the forearm (wrist to elbow) and leg segments (ankle to fibular head). In the sensory conduction, nerve conduction velocity and sensory nerve action potential were measured. For defining abnormality, mean (2.5 SD) of our laboratory control value was considered.11

The sensitivity of different neurophysiological diagnostic criteria (table 1) in the diagnosis of GBS was evaluated.12 For statistical analysis, duration of illness was categorised into 1 week, 2 weeks and more than 2 weeks; severity of illness was classified as ⩽3, 4–6 and >6 in disability grade; and number of nerves with abnormal distal motor latency, conduction velocity, conduction block, reduced CMAP amplitude and minimal F latency. The NCS criteria were correlated with preceding viral illness, duration of illness, disability grading scale, type of GBS and 3 month outcome. Outcome was also correlated with the above mentioned clinical and neurophysiological features. Statistical analysis was performed using SPSS V.10 software using the χ2 or Fisher’s exact test and was considered significant if the two sided p value was <0.05.

RESULTS

Fifty-one patients with GBS were managed during the study period and were aged 6–70 years (mean 35.41 (17.44) years); six were children less than 13 years of age and there were 11 females. Thirty-five patients had a history of preceding illness which included symptoms of upper respiratory tract infections in six, diarrhoea in 13 and fever in 16 patients. The peak disability grade ranged between 3 and 9 (mean 6.78 (1.38)). Thirty-four patients were bedridden and of them five needed ventilatory support. Autonomic dysfunction in the form of tachycardia, hypertension and absence of R-R variation with respiration were present in four patients. Sensory deficit in the form of touch, joint position and vibration impairment was present in 18 patients.

NCS were performed in all patients on the day of admission; within the first week in 25, within 2 weeks in 16 and in 10 patients NCS were done after 2 weeks. In three patients, motor nerve conduction studies were normal. Median, ulnar and peroneal NCS were carried out in 102 nerves each. Ulnar NCS was unrecordable in 9.8% and peroneal in 18.6% of nerves. Distal latency was prolonged in the median nerve in 43.1%, in the ulnar nerve in 56.9% and in the common peroneal nerve in 58.8%. Conduction velocity was slowed in the median in 48%, in the ulnar in 37.3% and in the common peroneal in 21.6% of nerves. CMAP amplitude was reduced in the median in 46.1%, in the ulnar in 55.9% and in the common peroneal in 13.7% of nerves. Conduction block was found in 105 (25.7%) nerves; median 34.3%, ulnar 41.2% and peroneal 27.5%. F responses were unrecordable on median stimulation in 35.3%, ulnar stimulation in 51% and common peroneal stimulation in 60.8% of nerves. Minimal F latency of the median nerves was prolonged in 38.2%, ulnar in 31.4% and common peroneal in 3.9% (table 2).

Neurophysiological studies fulfilled the diagnostic criteria of AIDP in accordance with Albers and colleagues2 (set 1) in 45 (88.2%) patients, Cornblath3 (set 2) in 20 (39.2%), Ho and colleagues4 (set 3) in 44 (86.3%), the Dutch GBS study group5 (set 4) in 42 (82.4%), the Italian GBS study group6 (set 5) in 33 (64.7%) and Albers and Kelly7 (set 6) in 25 (49%). The sensitivities were significantly different among the groups. The diagnostic yield of sets 1, 3 and 4 was higher compared with sets 2, 5 and 6 in all three clinical types of GBS (fig 1). The sensitivity of these criteria revealed a marginal increase from 88% in the first week to 90% in the third week in sets 1 and 3, whereas it was much more increased in set 2 (32% to 60%) and set 6 (48 to 60%). A similar sensitivity pattern of various neurophysiological criteria was observed with outcome but not with severity of illness.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Diagnostic yield of various neurophysiological criteria in Guillain–Barre syndrome and its various clinical subtypes.

The correlation of various neurophysiological criteria with type of GBS, duration of illness and severity of illness are presented in table 3. Sensory nerve conduction study of the median, ulnar and sural nerves on both sides (102 nerves each) revealed unrecordable results in the median nerve in 32.4%, in the ulnar nerve in 37.3% and in the sural nerve in 24.5%. Sensory conduction velocity was slow in the ulnar in 12.7%, in the median in 42.2% and in the sural in 20.6% of nerves. In 7 (13.7%) patients, median sensory conductions were abnormal but both sural conductions were normal. Following the criteria of Ho and colleagues,4 patients with GBS could be classified as AIDP (44 (86.3%)), AMAN (4 (7.8%)) and AMSAN (3 (5.9%)).

Twenty-seven patients were treated with IVIg. One patient died as a result of cardiac arrest and two were lost to follow-up. At 3 months, 11 (22.4%) patients improved completely, 28 (57.1%) partially and 9 (18.4%) patients were bedridden or wheelchair bound. Outcome correlated with CMAP amplitude (p = 0.0001), motor conduction velocity (p = 0.002) and severity of illness (p = 0.03) but not with distal latency (p = 0.43), conduction block (p = 0.12), F wave abnormality (p = 0.22), clinical presentation (p = 0.71), type of GBS as per Ho and colleagues4 (p = 0.60) or IVIg treatment (p = 0.16). Of three patients with AMSAN, one had complete recovery who had unrecordable ulnar and peroneal motor nerve conduction on both sides. Two patients had partial recovery; one had unrecordable peroneal motor nerve conduction on both sides and the other had an unrecordable left ulnar motor nerve.

DISCUSSION

The sensitivity of various neurophysiological criteria in the diagnosis of GBS in our study ranged between 39.2% and 88.2%, with higher sensitivity of set 1 (88.2%), set 3 (86.3%) and set 4 (82.4%). In a study in 180 patients with GBS, set 1 criteria revealed a sensitivity of 87%.2 Similar sensitivity (85%) has been reported by Meulstee et al using set 45; however, sensitivity decreased to 5–46% when sets 1, 2 and 6 were applied in the same patient population.5 Neurophysiological study in 38 Chinese patients with GBS revealed features of demyelination in 32% and AMAN in 55% using set 3.4 Using the same criteria4 in our patients, a diagnosis of demyelinating neuropathy was made in 86.3% of patients. The higher sensitivity of set 3 in the diagnosis of demyelinating neuropathy in our patients may be because of differences in patient population as the majority of the Chinese patients with GBS were of the AMAN variety whereas the majority of our patients had AIDP.

The sensitivity of various sets of neurophysiological criteria in our study was irrespective of the clinical types of GBS (ie, pure motor, sensorimotor and pure sensory). A similar diagnostic yield was reported in a previous study in 43 patients.13 From India, a lower sensitivity of neurophysiological criteria has been reported in which Alber et al’s criteria were fulfilled in 74.2% of patients, Albers and Kelly’s criteria in 40.9% and Cornblath’s criteria in 30.3%. The lower sensitivity in this study may because of the lower number of nerves studied or different patient populations.8 The difference in sensitivity of different neurophysiological criteria in the same patient population may be attributed to requirement of a number of demyelinating features, number of nerves with demyelinating features and definition of conduction block. In sets 1, 3 and 4, only one demyelinating feature (prolonged distal latency, conduction block, dispersion, slowed conduction velocity, prolonged minimal F latency) in two nerves was required to fulfil the criteria2 5 7 whereas sets 2, 5 and 6 required two to three of these demyelinating features in at least one to two nerves.3 4 6

Neurophysiological criteria in different types of GBS may have different sensitivities. In AMAN, NCS reveals inexcitable or increased threshold, reduced compound muscle action potentials, normal sensory and motor conduction velocity and absence of conduction block or marked prolongation of minimal F latency.14 In pure sensory GBS, however, NCS reveals motor conduction abnormalities. In a study of eight patients with pure sensory GBS, seven had conduction abnormalities in motor nerves.15 In our study, although clinically the majority of patients (60.8%) had pure motor GBS, on NCS, 86.3% of these patients revealed features of demyelination. All pure sensory varieties had a demyelinating nature. After clinical and NCS, only 4 (7.8%) patients had AMAN and 3 (5.9%) the AMSAN variety of GBS. The higher prevalence of AMAN has been reported from China and has been attributed to Campylobacter jejuni.1 5 16 There is a paucity of studies evaluating the diagnostic yield of various NCS criteria in these patients.

Evaluation of NCS in all four limbs improves the diagnostic yield and helps to classify GBS patients. Lack of this approach may erroneously label these patients as AMAN. The sensitivity of NCS criteria increases with duration of illness.7 In the Dutch GBS study, the sensitivity of the electrodiagnostic test was 60% in identifying demyelinating neuropathy at the first test (median 6 days), which increased to 72% at the third test (median 34 days).7 This phenomenon has also been seen in our study. Sensitivity from the first to the third week increased by 2% in sets 1 and 3. In set 2, sensitivity doubled in the third week compared with the first (32–60%). Overall sensitivity, however, remained higher in sets 1, 3 and 4 at various time periods.

The influence of time scale, severity of disease and outcome on various diagnostic criteria, however, has not been reported in previous studies.13 NCS criteria were not related to severity of disease or outcome. This may be because of the multifocal nature of the disease and bulbar and respiratory involvement which are the major causes of morbidity and mortality and are not evaluated in normal nerve conduction studies. None of the neurophysiological parameters or criteria was related to outcome except CMAP amplitude and conduction velocity. Low CMAP amplitude has also been reported to correlate with poor outcome in GBS.17 18

The present study revealed a wide variability in sensitivity of different NCS criteria, ranging from 39.2% to 88.2% in Indian patients with GBS. Patients were classified as AIDP (86.3%), AMAN (7.8%) and AMSAN (5.9%) using set 3 criteria and their outcome was related to severity of illness and CMAP amplitude.