1 Introduction

Since the first description of the axonal form of Guillain–Barré syndrome (GBS) by Feasby et al. (1986), the subsequent classification of GBS has mainly transformed to acute inflammatory demyelinating polyradiculoneuropathy (AIDP), acute motor axonal neuropathy (AMAN), and acute motor sensory axonal neuropathy (AMSAN) (Feasby et al. 1993; Griffin et al. 1996; Hughes et al. 1999; McKhann et al. 1993).

Ho et al. (1997) reported a 64-year-old woman who was diagnosed as AMAN and improved quickly following plasmapheresis. This process indicated the reversible potential of some AMAN patients. Kuwabara, Asahina, et al. (1998) and Kuwabara, Yuki, et al. (1998) proposed the term reversible conduction failure (RCF) in their articles. RCF can promptly recover after immune-modulated treatment, which is due to injury located in nodal membrane of the axon or through detachment of terminal myelin loops (Kuwabara, Asahina, et al. 1998; van Doorn et al. 2023). The put forward of RCF makes the differential diagnosis of GBS subtypes more complicated.

Early recognition of different subtypes is important for determining the intervention and prognosis of these patients. If the conduction failure of axonal dysfunction was not resolved rapidly, they would progress to secondary Wallerian-like axonal degeneration. Currently, RCF can only be detected by serial nerve conduction studies. How can we distinguish the conduction failure as soon as possible? This study tries to analyze the initial clinical and electrophysiological characteristics of different subtypes of GBS diagnosed based on serial electrophysiological examinations. We hope that these features can provide help for the early identification of different subtypes of GBS.

2 Methods

3 Results

3.1 The Diagnostic Consistency Based on the Initial and Serial Electrophysiological Studies

This study included a total of 47 eligible GBS patients, comprising 19 AIDPs, 18 axonal degenerations (12 AMAN and 6 AMSAN), and 10 RCFs (6 AMAN and 4 AMSAN), based on serial electrophysiological studies using Uncini's criteria. The diagnosis of these 47 patients was 21 AIDP, 18 axonal, and 8 equivocal based solely on the initial electrophysiological examination with Rajabally's criteria. Figure 1 shows the relationship of diagnostic classification based on the initial and serial electrophysiological studies. As we can see in Figure 1, the diagnostic consistency of initial and serial electrophysiological studies in the AIDP and axonal degeneration groups was 89% and 83%, respectively. However, in the RCF group, 40%, 30%, and 30% of patients were diagnosed as AIDP, axonal, and equivocal at the initial study, respectively.

3.3 Quantitative Comparison of Initial Electrophysiological Parameters Among AIDP, Axonal Degeneration, and RCF Groups

As shown in Table 2, the occurrence rate of CB was 47% and 60%, and abnormal temporal dispersion was 32% and 0% in the AIDP and RCF groups, respectively. Although CB and abnormal temporal dispersion did not show a significant difference between the two groups, the RCF group had significantly shorter DML in median, tibial, and fibular nerves, lower CMAP amplitude in ulnar nerve, faster CMAP velocity in median nerve, shorter F_min, F_mean, and F_max in tibial nerve, and lower F% in ulnar nerve than the AIDP group.

TABLE 2. Comparison of initial electrophysiological parameters between acute inflammatory demyelinating polyradiculoneuropathy (AIDP), axonal degeneration, and reversible conduction failure (RCF).

	AIDP (n = 19)	Axonal degeneration (n = 18)	RCF (n = 10)	p	RCF vs. AIDP	RCF vs. axonal degeneration	AIDP vs. axonal degeneration
					p	p	p
Time from onset, d, median (IQR)	11.0 (10.0, 14.0)	10.5 (8.0, 14.0)	10.0 (8.0, 11.5)	0.106	/	/	/
DML, ms, median (IQR)
Median nerve	7.1 (4.6, 9.0)	3.6 (3.3, 4.0)	3.7 (3.6, 4.5)	<0.001	0.001	0.866	<0.001
Ulnar nerve	3.4 (2.9, 5.4)	2.6 (2.3, 3.1)	3.0 (2.7, 3.2)	0.001	0.277	0.377	<0.001
Tibial nerve	5.2 (4.7, 7.2)	3.9 (3.6, 4.7)	4.5 (3.9, 5.0)	<0.001	0.017	0.752	<0.001
Fibular nerve	6.0 (4.3, 7.9)	3.5 (3.1, 3.7)	3.8 (3.3, 4.0)	<0.001	0.002	1.000	<0.001
CMAP amplitude, mV, median (IQR)
Median nerve	7.6 (5.2, 13.2)	4.9 (1.5, 9.3)	5.8 (2.6,11.5)	0.035	0.350	1.000	0.035
Ulnar nerve	11.5 (8.1, 13.0)	2.5 (0.5, 8.9)	6.4 (1.9, 10.1)	<0.001	0.047	1.000	<0.001
Tibial nerve	4.7 (2.6, 9.6)	3.2 (0.2, 8.9)	3.6 (1.0, 8.5)	0.253	/	/	/
Fibular nerve	4.0 (1.6, 6.1)	3.7 (0.9, 4.6)	3.7 (1.4, 8.2)	0.511	/	/	/
CMAP velocity 1, m/s, median (IQR)
Median nerve	46.8 (36.2, 53.1)	53.6 (51.0, 56.0)	55.5 (51.6, 59.6)	<0.001	0.001	0.707	0.008
Ulnar nerve	51.1 (45.0, 57.6)	56.4 (52.8, 60.3)	55.7 (50.4, 62.8)	0.053	/	/	/
Tibial nerve	36.1 (28.5, 40.7)	42.6 (38.9, 46.9)	41.5 (37.3, 43.8)	0.001	0.072	1.000	0.001
Fibular nerve	37.8 (27.4, 43.4)	43.6 (41.0, 45.8)	42.8 (38.0, 45.5)	0.012	0.209	1.000	0.014
CMAP velocity 2, m/s, median (IQR)
Median nerve	17.5 (11.8, 30.8)	38.9 (36.2, 46.8)	37.2 (31.0, 44.8)	<0.001	0.003	1.000	<0.001
Ulnar nerve	45.6 (30.8, 54.7)	52.0 (49.6, 57.0)	52.9 (47.2, 57.2)	0.040	0.189	1.000	0.058
Fmin, ms, median (IQR)
Ulnar nerve	31.1 (27.2, 34.6)	26.9 (25.3, 28.4)	29.9 (26.9, 32.5)	0.010	1.000	0.237	0.009
Tibial nerve	57.9 (54.9, 68.5)	53.5 (50.3, 54.9)	52.8 (51.3, 54.5)	0.001	0.010	1.000	0.001
Fmean, ms, median (IQR)
Ulnar nerve	31.9 (28.0, 34.6)	26.9 (25.3, 28.4)	29.9 (26.9, 32.5)	0.031	1.000	0.303	0.031
Tibial nerve	58.7 (56.1, 71.6)	54.2 (51.2, 56.7)	52.5 (52.3, 55.0)	<0.001	0.011	1.000	0.001
Fmax, ms, median (IQR)
Ulnar nerve	32.5 (29.3, 36.5)	28.8 (27.1, 31.0)	31.2 (29.3, 36.6)	0.031	1.000	0.332	0.030
Tibial nerve	62.1 (58.4, 73.3)	55.6 (53.1, 57.7)	55.9 (53.3, 57.8)	<0.001	0.018	1.000	0.001
F%, %, median (IQR)
Ulnar nerve	100.0 (100.0, 100.0)	65.0 (43.8, 100.0)	20.0 (0.0, 91.0)	0.005	0.006	0.495	0.135
Tibial nerve	100.0 (0.0, 100.0)	85.0 (0.0, 100.0)	97.5 (20.0, 100.0)	0.469	/	/	/
Conduction block, n (%)	9 (47%)	0 (0%)	6 (60%)	0.001	/	<0.05	<0.05
Abnormal temporal dispersion, n (%)	6 (32%)	0 (0%)	0 (0%)	0.006	/	/	<0.05

• Note: CMAP velocity 1 included velocity from elbow to wrist for median nerve, from elbow below to wrist for ulnar nerve, from popliteal fossa to ankle for tibial nerve, and from fibular head to ankle for fibular nerve. CMAP velocity 2 included velocity from wrist to palm for median nerve and from elbow above to elbow below for ulnar nerve.
• Abbreviations: CMAP, compound muscle action potential; DML, distal motor latency; F%, F-wave persistence; F_max, maximum F-wave latency; F_mean, mean F-wave latency; F_min, minimum F-wave latency; IQR, interquartile range.

The RCF and axonal degeneration groups showed similarities in most electrophysiological parameters, except for CB, which occurred in 60% of patients in the RCF group but not in the latter.

The AIDP group had significantly longer DML in median, ulnar, tibial, and fibular nerves, higher CMAP amplitude in median and ulnar nerves, slower CMAP velocity in median, tibial, and fibular nerves, and longer F_min, F_mean, and F_max in ulnar and tibial nerves than the axonal degeneration group. The occurrence rate of abnormal temporal dispersion also showed significant differences between two groups.

Figure 2 presents the initial and second electrophysiological findings in the median nerve of two RCF patients. One patient had CB (Figure 2A), whereas the other showed a slightly prolonged DML without CB (Figure 2C). Both patients recovered rapidly after intravenous immunoglobulin therapy during the second examination (Figure 2B,D).

4 Discussion

The present study investigated the initial clinical and electrophysiological characteristics of AIDP, axonal degeneration, and RCF subtypes of GBS diagnosed based on serial electrophysiological examinations. The different diagnostic classifications based on the initial and serial electrophysiological findings in our study indicated that early identification of RCF may be the greatest difficulty in the differential diagnosis of GBS. Therefore, we further explored the differences between RCF and AIDP and axonal degeneration to uncover clues for the early identification of RCF.

The differential diagnosis of RCF and AIDP is one of the difficulties. Our study indicated that AIDP patients usually had higher CSF protein than RCF patients. High CSF protein level has been reported to be associated with an early severe disease course and a demyelinating subtype (Al-Hakem et al. 2023). The greater elevation of CSF protein in AIDP may originate from the more extensive blood–nerve barrier disruption and more intense inflammatory response (Al-Hakem et al. 2023; Yuki and Hartung 2012). Although we could not establish an appropriate diagnostic cutoff due to a small sample size, we could see the overall trend of DML, F_min, F_mean, and F_max in RCF patients showed normal or slightly prolonged, rather than apparently prolonged in AIDP patients. Uncini et al. (2017) and Uncini and Kuwabara (2012) recorded in their articles that transient CB/slowing in intermediate and distal nerve segments, mimicking demyelination but without the development of abnormal temporal dispersion is an important feature of RCF. The slightly prolonged DML in RCF patients mostly came from conduction slowing in the distal segment. Moreover, the reduction of F% mainly resulted from conduction failure along the nerve. Because tibial nerve contains more motor axons than ulnar nerve, it is easy to elicit F-waves and not easy to appear F% decrease for the former. In addition, the F% of the tibial nerve is higher than that of the ulnar nerve in normal people (Pan et al. 2013).

Differentiating between the RCF subtype and axonal degeneration subtype was another challenge in GBS diagnosis. Patients with RCF exhibited better motor function during hospitalization and had a favorable prognosis, as assessed by the Hughes functional grades at admission and discharge, as well as the variation from admission to discharge. In contrast, patients with axonal degeneration subtype typically experienced poor outcomes. This finding is consistent with the results reported in the study by Niu et al. (2018), where “AMAN with CBs” had a higher reduction of Hughes functional grade at 1 month than “AMAN without CBs.” The RCF and axonal degeneration groups shared similar electrophysiological features, except for CB, which was present in the former but not the latter. Due to the limited clinical and electrophysiological evidence available for distinguishing between RCF and axonal degeneration, early identification of RCF from axonal degeneration remains challenging, and further research is warranted.

The differential diagnosis between AIDP and axonal degeneration was relatively straightforward compared to distinguishing RCF from AIDP and axonal degeneration. Although demyelination is predominant in the former and axonal dysfunction in the latter, some patients exhibit both demyelination and axonal degeneration, making it challenging to differentiate (van Doorn et al. 2023). In comparison to axonal degeneration, the CSF protein levels of AIDP patients were significantly elevated. Additionally, prolonged DML and F-wave latencies, along with decreased CMAP velocity, remained prominent features of AIDP.

Previous studies reported RCF was significantly associated with IgG antibodies to GM1, GalNAc-GD1a, GD1b, or complex of LM1/GA1 (Créange et al. 2014; Ogawa et al. 2013; Shahrizaila et al. 2014). Antibodies bound to gangliosides at the nodes of Ranvier, which activated complement, disrupted sodium-channel clusters, and resulted in “axonal CB.” As “axonal CB” brought confusion in the electrodiagnostic diagnosis of GBS, some researchers proposed the term “nodal CB” (Oh 2021). This kind of “CB” can promptly recover after immune-modulated treatment without prominent pathological changes (Kaida 2019; Kuwabara and Yuki 2013). Conversely, if the block is not resolved quickly, they will develop axonal degeneration. Therefore, early identification of the RCF subtype can guide treatment in time. Some studies reported anti-ganglioside antibodies play an important role in the different patterns of evolution, RCF, or axonal degeneration (Kuwabara, Asahina, et al. 1998). However, a study conducted by Kokubun et al. (2010) did not find a difference in anti-ganglioside antibody profiles between patients with and without RCF. We also did not reveal a significant difference of ganglioside antibodies in CSF or blood among the three groups in our series.

In summary, we concluded the following points regarding three subtypes in the initial examinations. The early identification of RCF from AIDP had relatively obvious features, including slightly elevated CSF protein levels and normal or slightly prolonged DML and F-wave latencies, rather than significantly elevated and prolonged as in AIDP. The early identification of RCF from axonal degeneration remains challenging. One differentiator that may be helpful was that the motor function of the RCF showed better than in the latter.

Author Contributions

Shuo Yang: writing–original draft, methodology, formal analysis, data curation, writing–review and editing. Na Chen: methodology, data curation. Lei Zhang: methodology, data curation. Ying Wang: methodology, data curation. Lin Chen: data curation. Fan Jian: data curation. Zaiqiang Zhang: project administration, supervision. Yilong Wang: supervision, project administration. Hua Pan: writing–review and editing, supervision, validation, project administration.

Ethics Statement

We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Conflicts of Interest

The authors declare no conflicts of interest.