Utility of Serial Nerve Conduction Studies in the Electrodiagnosis of Guillain–Barre Syndrome
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.314529
Source of Support: None, Conflict of Interest: None
Keywords: GBS, reversible conduction failure, serial nerve conduction studiesKey Message: Serial nerve conduction studies are mandatory for an accurate electrodiagnosis of GBS
Guillain–Barre syndrome (GBS) is the most common cause of acute flaccid paralysis worldwide. It can be classified electrophysiologically into demyelinating and axonal subtypes. Acute inflammatory demyelinating polyneuropathy (AIDP) is the demyelinating subtype, while acute motor axonal (AMAN) and acute motor-sensory axonal neuropathy (AMSAN) are the axonal subtypes. It is difficult to differentiate between them clinically and nerve conduction studies (NCS) can play an important role.
Determining the electrophysiological subtype is useful as it can give insights into the underlying pathophysiology. This has both therapeutic and prognostic significance. Various electrodiagnostic criteria sets have been proposed to differentiate between the demyelinating and axonal subtypes.[2–4]
There has been a controversial debate as to whether the subtypes can be diagnosed accurately based on a single study, with one group of authors advocating a “serial studies approach”[6–8] and another group rebutting this and advocating a “one study approach”.,,
There is a lack of studies universally as well as from India looking at the utility of serial NCS in the management of patients with GBS. Hence this retrospective analysis was done to determine if serial NCS changes the electrodiagnostic classification, and to identify additional parameters that may support the electrodiagnosis in early stages.
We retrospectively reviewed our institutional database of adult patients aged 15 to 80 years, admitted with a diagnosis of GBS between August 2015 and July 2017 at Christian Medical College, Vellore. Patients who had at least two serial electrophysiological recordings were selected. The study variables were retrieved from a prospectively maintained database.
All the patients had NCS done on eight motor and eight sensory nerves using conventional techniques. Parameters measured included: distal motor latency (DML), sensory latencies, motor and sensory conduction velocities (CV), compound action potentials (CMAP), F-wave latencies (FL), and sensory action potentials (SNAPs). The skin temperature was maintained over 32°C. The value of each variable was compared with the upper and lower limits of normal for our laboratory. For DML, motor CV, F wave latency, CMAP, and SNAP amplitudes, the upper and lower limits of normal were defined as the mean ± 2.5 SD of control values of our laboratory. Sural sparing pattern was defined as sural SNAP >15microV and ulnar SNAP <2.5 SD of LLN. Facial nerve conduction, blink reflex, phrenic nerve conduction, and electromyography data were analyzed if available. Facial nerve dysfunction was defined as facial CMAP less than 4 mV. Abnormal blink reflex was defined as marked delayed or absence of all potentials of the blink response. Phrenic nerve dysfunction was defined as either phrenic CMAP <0.4 mV or phrenic latency >9 ms.
We applied the various published criteria [Table 1] to the two sets of NCS and classified them into AIDP and axonal GBS. Temporal dispersion (TD) was not included for classification as we did not have the required data. Patients who did not satisfy the criteria for AIDP or axonal GBS were classified as equivocal. We also established the diagnostic shifts for each set of criteria, as a result of serial NCS.
Anti-ganglioside antibody testing
Serum anti-ganglioside antibodies were tested in serum by ELISA. Testing was done for GM1, GD1a, GD1b, GT1a, and GQ1b.
Patients were grouped according to electrodiagnosis at first study, electrodiagnosis at the second study, and anti-ganglioside antibody status. The statistical analysis was performed using IBM Corp. Released 2012. IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM Corp. Descriptive analysis was performed using the Chi-square test. Statistical significance was taken to be at the two-tailed 0.05 level.
There were a total of 40 patients with GBS, out of which 31 patients had serial NCS. There were 21 males and 10 females. The mean age of the cohort was 37.8 (SD: 17.8) years. The mean interval between the onsets of illness to first NCS was 1.6 (SD: 0.7, range 1-3) weeks and the second NCS was 3.5 (SD: 1.5, range 2-8) weeks. The mean interval between the serial NCS was 1.9 (range 1-6) weeks. Twenty-three patients (74%) were treated with plasma exchange while 8 patients (26%) were treated with IVIG. The clinical profile, anti-ganglioside antibody status, the requirement of assisted ventilation, and electrodiagnosis on serial NCS are shown in [Table 2].
Electrodiagnostic shifts on serial NCS
The electrodiagnosis of GBS on serial NCS according to the various published criteria are given in [Table 3]. According to Cornblath, Hadden, and Rajabally criteria, 5 patients (16.1%), 3 patients (9.6%), and 5 patients (16.1%) respectively changed classification. Out of the five patients that changed classification from AIDP to axonal/equivocal [Table 4], four were due to the resolution of reversible conduction failure (RCF) [Figure 1]. All the six patients that changed classification from axonal to AIDP/equivocal [Table 5] were due to misclassification. Four out of the six patients were misclassified as AIDP on the second study.
Additional parameters supporting electrodiagnosis
The relation between the GBS subtype and the other electrophysiological parameters were studied at the first NCS [Table 6]. Among them, the presence of a sural sparing pattern, facial nerve dysfunction, abnormal blink reflex, and phrenic nerve dysfunction was found to be statistically significant. The sural sparing pattern was more common in AIDP, ranging from 22.7% to 35.7% of patients with AIDP. It was statistically significant for Rajabally criteria. Facial nerve dysfunction was also more common in AIDP, ranging from 42.8% to 69.2% of patients. It was statistically significant for all three criteria. Abnormal blink reflex was also more common in AIDP, ranging from 45% to 75% of patients, respectively. It was statistically significant for Cornblath and Rajabally criteria. Three patients had an abnormal blink reflex with normal facial conductions. Phrenic nerve dysfunction was also more common in AIDP, ranging from 36.3% to 57.1% of patients. It was also statistically significant for Rajabally criteria.
Relations between GBS subtype and anti-ganglioside antibodies
Anti-ganglioside antibodies were tested in 26 out of the 31 patients. They were present in 46.1% of patients (12/26). Six patients were positive for the single type of anti-ganglioside antibody while the remaining six patients showed positivity to two types of anti-ganglioside antibodies. Antibody to GM1 was the most common subtype. They were more common in axonal GBS, present in 56.2%, 50%, and 53.8% of patients according to Cornblath, Hadden, and Rajabally criteria respectively at the second electrodiagnostic study. They were present in 2 patients with a shift in electrodiagnosis from AIDP and in 3 patients with a shift from axonal.
Accurate electrodiagnosis of GBS is important as it offers insight into the underlying pathophysiology. In axonal GBS, there is nodal and paranodal involvement due to antibody-mediated attack and complement activation resulting in conduction failure. This can manifest as a conduction block, conduction slowing, and isolated F wave absence. Classically, this results in axonal degeneration. However, in some patients, conduction failure resolves and axonal degeneration does not occur.[13–15] Thus, there are two possible patterns of evolution which can be seen in axonal GBS. The common pattern is a reduction of distal CMAP amplitudes which is due to axonal degeneration. In the other pattern, a rapid improvement in CMAP amplitudes and CV is seen due to the resolution of nodal conduction failure. Both patterns are not associated with an increase in TD. AIDP is characterized by segmental demyelination in both motor and sensory nerves, resulting in abnormalities of the F and H responses, conduction blocks, abnormal temporal dispersion. As the process of remyelination occurs, there is a slowing of CV with persistently increased TD. There has been evidence of complement activation leading to myelin disruption as well as secondary axonal degeneration in AIDP as well. Thus, the pathophysiology of GBS is dynamic and serial studies may allow a more accurate diagnosis of subtypes.
Considering the recent advances in the field of both pathophysiology and treatment of GBS, an early accurate electrodiagnosis becomes paramount in the management. IVIG could potentially be more effective in settings where there is more complement-mediated damage There are some preliminary reports which have suggested that for IVIG is better than PLEX for patients with AMAN. There is also a need to identify early the subtypes with a potentially refractory disease given the options of considering IVIG and eculizumab which could have a major impact on long-term morbidity. This emphasizes the importance of an early accurate electrodiagnosis in guiding early targeted treatment.
The patterns of recovery in axonal GBS and AIDP are also different, with axonal GBS generally taking much longer than AIDP. Hence, accurate electrodiagnosis of GBS has both therapeutic and prognostic implications.
The frequency of subtypes of GBS has substantial variability in different series. This may be due to differences in triggering factors, genetic susceptibility, and different antecedent infections. It can also be due to the different electrodiagnostic criteria used and whether it was based on a single study or serial NCS. In North America and Europe, AIDP is more common, comprising up to 90%.
In India, using the Cornblath criteria on a single NCS, Alexander et al. found 55.7% of patients to have AIDP while Kalita et al. found that 39.2% of patients had AIDP. Using the Cornblath criteria on serial NCS, we found that 45.2% had AIDP on the first NCS. On the second NCS, the number of patients with AIDP was reduced to 41.9%. Using Hadden's criteria, Uncini et al. found 67% of patients to have AIDP, 18% to have axonal GBS, and 15% to have equivocal electrodiagnosis on the first NCS. On the second NCS, the number of patients with AIDP and equivocal electrodiagnosis reduced to 58% and 4% respectively, while the number of patients with axonal GBS increased to 38%. Using Hadden's criteria, we found that 71% had AIDP and 21% had axonal GBS on the first NCS. On the second NCS, the number of patients with AIDP reduced to 64.5%, while the number of patients with axonal GBS remained the same at 21% and 6.5% had an equivocal electrodiagnosis.
In our study, 16.1%, 9.6%, and 16.1% of patients changed classification on serial NCS, using, Cornblath, Hadden, and Rajabally criteria, respectively. Uncini et al. found that 23.6% of patients changed subtype, using Hadden's criteria and the majority of the shifts were from AIDP and equivocal groups to axonal GBS. This was mainly due to the recognition of RCF by serial NCS. All patients who shifted to the axonal group also had anti-ganglioside antibodies. In our study, there were shifts in both directions. Historically, there has been no gold standard for GBS subtypes diagnosis and the final diagnosis can be considered as the reference diagnosis, as it was reached considering the whole electrophysiological history of the patient. For practical purposes, Uncini and Kuwabara have suggested that at the first study Rajabally criteria sets can be employed for an indicative subtype diagnosis. This is an observation from our study as well considering the comprehensiveness of Rajabally criteria.
The major reason for shifts from AIDP to axonal GBS was the resolution of RCF on serial NCS.[13–15] A single NCS can't distinguish between demyelinating conduction block and RCF and can misclassify patients with axonal GBS as having AIDP. RCF is an a posteriori diagnosis and can be identified only on serial NCS [Figure 1]. Electrodiagnostic criteria for RCF has been published recently.
The major reason for shifts from axonal GBS to AIDP was the misclassification of subtypes due to inherent flaws in the criteria. With the current criteria, there is a tendency for underdiagnoses of axonal GBS, primarily due to misclassification as AIDP., This is because both demyelination and nodal conduction failure can lead to prolonged DML, decreased CV, and conduction block. Thus, in the early stages of GBS, the distinction between demyelinating and axonal subtypes can be difficult in some patients. In our study, misclassification in these cases occurred probably because the second study was done earlier done at an average interval of 1.9 (range 1-6) weeks after the first study. Traditionally, the first NCS has been done at around 2 weeks after disease onset and the second study at a variable interval after the same. Uncini and Kuwabara have suggested at least two NCS in the first 4–6 weeks of the disease. Shahrizaila et al. have suggested that performing NCS at two time intervals, 1st NCS at admission and 2nd NCS at an interval of 3–8 weeks after disease onset can make an accurate electrodiagnosis of GBS.
A sural sparing pattern is seen only in AIDP and is a specific electrodiagnostic feature that can discriminate AIDP from axonal GBS. Derksen et al. studied 73 patients with GBS and 38 patients with GBS mimics and found that the most specific finding suggestive of AIDP was the sural sparing pattern. In our study also, we found that the sural sparing pattern was more common in AIDP.
The literature concerning the utility of blink reflex, facial nerve, and phrenic nerve conductions to differentiate between AIDP and axonal subtype is scant. Our study has also identified that the presence of facial nerve dysfunction, abnormal blink reflex, and phrenic nerve dysfunction helps to differentiate AIDP from axonal subtype in the early stages. Compared with AIDP, patients with axonal GBS have less frequent cranial-nerve involvement. Vucic et al. analyzed blink responses in 38 patients with AIDP within ten days of symptom onset and found abnormalities in 51% of patients, out of which 69% clinically exhibited facial weakness. They found that blink reflexes can be abnormal in some AIDP patients with clinically normal facial strength. Hence blink reflex should be done in all patients with suspected GBS. In our study, we found that facial dysfunction and abnormal blink reflex were more common in AIDP. This is probably due to demyelination in the facial and trigeminal nerves and can help to differentiate AIDP from axonal subtype in the early stages. In a study by Gourie-Devi and Ganapathy, phrenic nerve conduction abnormalities were observed in 64.3% of patients with GBS. Sen et al. studied phrenic nerve conduction in 64 patients with AIDP and found abnormal phrenic nerve conductions in 65.62% of patients. In our study also, abnormal phrenic nerve conduction was found to be more common in AIDP than axonal GBS. These parameters potentially have a greater relevance even compared to the conventionally described “sural sparing pattern” which was found only in a relatively smaller number. These indicate the presence of subclinical dysfunction in a non-length dependent pattern as described in demyelinating neuropathies.
Anti-ganglioside antibodies have an important role in the immunopathogenesis of GBS and can be directed against myelin or the axon. The most common antibodies that are identified are against GM1, GD1a, GD1b, GT1a, and GQ1b gangliosides. They are more common in axonal GBS than AIDP. Naik et al. tested 73 patients with GBS for anti-ganglioside antibodies and found that they were present in 45.6% with AIDP and 50% with axonal subtype. In our study, they were present in 46.1% with AIDP and 53.3% with axonal subtype. Anti-GT1b antibody was the most common antibody in their study, while in our study, anti-GM1 was the most common.
The strengths of the study include the incorporation of a standard protocol comprising NCS from all the four limbs with blink reflex, facial and phrenic conductions. All the studies were performed under the supervision of a core group of senior neurologists. Limitation of this study was the small sample size and retrospective design. However, it has given a good insight into the utility of serial NCS as the electrodiagnostic studies were comprehensive.
Thus, serial nerve conduction studies are mandatory for an accurate electrodiagnosis of GBS subtypes, which has both therapeutic and prognostic implications. Also, the use of additional parameters such as blink reflex, facial and phrenic nerve conduction, and anti-ganglioside antibodies may supplement routine NCS.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]