Repetitive after-discharges are more common in acquired demyelinating polyneuropathies

Abstract

Repetitive after-discharges (RAD) following compound muscle action potential (CMAP) have been increasingly observed during F-wave studies in the diagnosis of demyelinating polyneuropathy. RAD was defined as repetitive potentials occurring within 500 ms after the initial CMAP at 1 mV/div sensitivity. EMG traces from patients with demyelinating polyneuropathy (n = 91), sensorimotor axonal polyneuropathy (n = 119) and normal subjects (n = 84) were reviewed to identify RAD during F-wave studies. The temporal relationship between RAD and the mixed nerve silent period (MNSP) was also assessed. RAD was detected in 90.1% of patients with demyelinating polyneuropathy, 29.4% of patients with sensorimotor axonal polyneuropathy, and 2.3% of normal subjects. Among patients with hereditary demyelinating polyneuropathy, RAD was present in 19.4%. In 84.7% of traces with obtainable F-waves, the initial RAD components appeared during the first period of MNSP. RAD may serve as a useful electrophysiological indicator of demyelinating polyneuropathy and may help distinguish it from sensorimotor axonal polyneuropathy and healthy individuals. It may also contribute to differentiating hereditary from acquired demyelinating polyneuropathies. The occurrence of RAD during the first period of MNSP, when voluntary motor activity is suppressed, suggests that these signals are not due to incomplete relaxation or voluntary discharges.

Introduction

Nerve conduction study (NCS) findings play a well-validated role in diagnosing demyelinating polyneuropathies. Common features in demyelinating polyneuropathies include slowed conduction velocity, motor conduction block, prolonged distal and F-wave latencies, and the presence of A-waves. In addition to these, it has been reported that “repetitive after-discharges (RAD)” may also be observed in demyelinating polyneuropathies^1,2. RAD refers to self-sustained discharges occurring after the compound muscle action potential (CMAP), due to the repetitive or continuous firing following stimulus cessation³. While RAD has been more commonly documented in conditions such as peripheral nerve hyperexcitability syndromes^4,5, myasthenia gravis⁶, and motor neuron disease⁷, the knowledge that it may be seen in peripheral neuropathies has remained somewhat overlooked in the literature.

RAD has long been observed during F-wave studies of some patients with electrophysiological diagnosis of demyelinating polyneuropathies in our electromyography (EMG) laboratory. This study aims to investigate the frequency and characteristics of RAD in demyelinating polyneuropathies and to improve our understanding about its pathophysiological mechanisms. For this purpose, we also examined the temporal relationship between RAD and mixed nerve silent period (MNSP) that is the relative electromyographic silent period and it is interrupted at certain latencies when the electrical stimulation is applied to the mixed nerve innervating a tonically contracting muscle.

Methods

Hacettepe University EMG Laboratory database was searched using the terms “demyelinating polyneuropathy”, “axonal polyneuropathy” and “electrophysiologically normal” for the years 2003–2023. All traces consistent with demyelinating and axonal polyneuropathies, based on the criteria of Oh et al.⁸, as well as all electrophysiologically normal traces, were pooled. Among patients diagnosed with demyelinating polyneuropathy, only diagnostic-stage EDX studies from those with symptoms lasting ≤ 2 months and no prior treatment were included, while F-wave studies of nerves with comorbidities such as radiculopathy or entrapment neuropathy were excluded due to their potential association with RAD. Subjects coded as “electrophysiologically normal” were individuals who had been referred to the EMG laboratory for the evaluation of suspected radiculopathy, entrapment neuropathy, or polyneuropathy, but whose test results were normal and who had no additional systemic or neurological diagnoses. F-wave studies of all traces were blindly reviewed by two specialists (H.U.U and F.G.Y) who were unaware of both the clinical and electrophysiological diagnoses.

NCSs were performed utilizing the Keypoint EMG machine (Medtronic A/S, Copenhagen, Denmark). Motor NCSs were performed with percutaneous supramaximal nerve stimulation in all subjects on the median, ulnar, peroneal, and tibial nerves by placing an active surface Ag–AgCl disc electrode over the muscle belly and the reference electrode over the tendon with a filter setting of 20 Hz–10 kHz. F-waves were elicited with at least 10 supramaximal stimulations at 2 Hz. Minimum F-response latencies were marked.

RAD was defined as repetitive and individually identifiable potentials within 500 ms following initial CMAP at 1 mV/div sensitivity⁴. In our routine NCSs, to rule out artifacts, stimulations with the same intensity were repeated with the stimulator removed from skin. The relationship between the frequency of these signals and environmentally prevalent 50-Hz signal was also evaluated.

RAD was defined as self-sustained, repetitive motor unit like potentials occurring shortly after the CMAP and persisting beyond the stimulus. In contrast, A-waves are characterized by fixed latency, amplitude, and morphology with each electrical stimulus and typically lack sustained repetitive activity^9,10. Based on these differences RAD and A-waves were distinguished from each other. RAD was further distinguished from F-waves by its earlier onset and closely spaced repetitive discharge pattern. Traces containing RAD during F-wave studies were identified and analyzed based on these electrophysiological criteria.

To evaluate the temporal relationship between RAD and MNSP, traces containing both RAD and F-waves were identified. The first period of the established MNSP was determined as the interval between the end of the M response and the minimum F-wave latency, in accordance with established neurophysiological definitions. This interval was manually marked on the EMG software using the same time scale in all recordings. In traces where RAD preceded the F-wave, the latency of the first repetitive after-discharge signal following the M response was noted to determine whether it occurred within the defined MNSP time window.

Data analyses were performed using Statistical Package for the Social Sciences (SPSS) version 25 for Windows (IBM, Armonk, NY, USA). For continuously variable data, mean ± standard deviation (SD) or median, minimum (min) and maximum (max) values were used. Group differences in RAD frequency were evaluated using the chi-square test. Fisher’s exact test was used for pairwise comparisons when appropriate. A two-sided p value < 0.05 was considered statistically significant.

Ethical approval and consent

All methods were carried out in accordance with the Declaration of Helsinki and relevant regulations. The study protocol was approved by the Hacettepe University Health Sciences Research Ethics Committee (Approval No: SBA 23/385). Permission to access and analyze data from the Hacettepe University EMG Laboratory database was obtained from the Chair of the Department of Neurology at Hacettepe University. As this was a retrospective study using archived data from routine EMG examinations, the requirement for additional informed consent for this study was waived by the Hacettepe University Health Sciences Research Ethics Committee. However, written informed consent had been obtained from all participants at the time of their EMG procedures as part of the routine clinical practice.

Results

Demographic and clinical features

Between 2003 and 2023, 91 patients with polyneuropathy showing varying degrees of demyelinating features were included. RAD was present in 82 of 91 patients (90.1%). A total of 82 patients with both RAD and demyelinating polyneuropathies were analyzed. Of these, 41 were female. The age of the patients ranged from 17 to 84 years (mean, 48.7 years) (Table 1).

Table 1.

Data on all patients with RAD.

	Mean (± SD)	Median	Min	Max
Age of patients with RAD	48.7 (± 18.1)	51	18	84
Frequency of stereotypical RAD (Hz)	235 (± 159.03)	100	66	500
RAD_FCL of all extremities	22.7 (± 12.7)	19.2	3.8	73.7
RAD_FCL of upper extremities	25.6 (± 15.2)	20	3.8	73.7
RAD_FCL of lower extremities	20.2 (± 9.8)	18.7	3.85	62.8

*SD, Standard deviation; Min, Minimum; Max, Maximum; RAD_FCL, Repetitive after-discharges first component latency.

Seventeen patients were diagnosed with chronic inflammatory demyelinating polyneuropathy (CIDP) (Fig. 1) and 16 with Guillain-Barré Syndrome (GBS) (Figs. 2 and 3). Other relatively less frequent diagnoses were diabetes mellitus (DM) (n = 7), multifocal acquired demyelinating sensory and motor neuropathy (MADSAM) (n = 6), multifocal motor neuropathy (MMN) (n = 3) and Charcot-Marie-Tooth disease (CMT). RAD was detected in 7 of 36 patients with a definite or possible hereditary etiology (19.4%). The diagnosis distribution of all patients is presented in the Table 2. Figure 4 consists of traces of patients with mercury intoxication and cytosine arabinoside toxicity.

Fig. 1.

Repetitive after-discharges in the peroneal (a) and tibial (b) F-wave studies of a patient with CIDP.

Fig. 2.

Repetitive after-discharges in the left (a) and right (b) tibial F-wave studies of a patient with GBS.

Fig. 3.

Repetitive after-discharges in the median (a) and peroneal (b) F-wave studies of a patient with GBS.

Table 2.

Patients with RAD whose diagnoses could be clarified during follow-ups.

Diagnoses of patients	n
CIDP	17
GBS	16
DM	7
MADSAM	6
MMN	3
CMT	3
Both CIDP and DM	3
POEMS	2
Multiple myeloma	1
DD-CKD	1
Both DM and DD-CKD	1
ARA-C toxicity	1
Mercury intoxication	1
Nodo-paranodopathy	1
Anti-MAG peripheral neuropathy	1
HNPP	1
NF-2	1
Malignancy	1

*n, Number; GBS, Guillain–Barré Syndrome; CIDP, Chronic inflammatory demyelinating polyneuropathy; DM, Diabetes mellitus; MADSAM, Multifocal acquired demyelinating sensory and motor neuropathy; MMN, multifocal motor neuropathy; CMT, Charcot-Marie-Tooth disease; POEMS, Polyneuropathy, Organomegaly, Endocrinopathy, Monoclonal plasma cell disorder, Skin changes; DD-CKD, Dialysis-dependent chronic kidney disease; ARA-C, cytosine arabinoside; HNPP, Hereditary neuropathy with liability to pressure palsy; NF-2, Neurofibromatosis type 2.

Fig. 4.

Repetitive after-discharges in the peroneal F-wave study of a patient with cytosine arabinoside toxicity (a) and in the median F-wave study of a patient with mercury intoxication (b).

Electrophysiological data

In patients with demyelinating polyneuropathy, RAD was most frequently recorded from the tibial nerve (n = 48 nerves), followed by the ulnar (n = 42 nerves), peroneal (n = 34 nerves), and median nerves (n = 31 nerves). F-waves were absent in 102 of 161 RAD-positive recordings. The mean (± SD) first component latency of RAD (RAD_FCL) in all nerves was 22.7 (± 12.7) ms overall, 25.6 (± 15.2) ms in upper extremity nerves (median and ulnar nerves) and 20.2 (± 9.8) ms in lower extremity nerves (tibial and peroneal nerves) (Table 1). In 52.3% (n = 44) of patients, RAD_FCL in at least one motor nerve fell within the first period of the MNSP. More specifically, 84.7% (50/59) of traces with both RAD and F-waves showed RAD onset during the first period of MNSP.

RAD patterns were irregular in most patients; however 8 exhibited stereotypical discharges in at least one nerve. Frequencies of these discharges ranged from 66 to 500 Hz (Table 1), typically resembling one motor unit morphology.

Simple A-waves were detected in 45 nerves, 19 of which were upper extremity nerves. Complex A-waves, defined as more than three phases, desynchronization, or > 2 discrete biphasic potentials⁹, were found in 10 nerves across 7 patients, and were present in 3 patients with GBS, 2 patients with CIDP and 1 patient each with CMT and nodo/paranodopathy. RAD co-occurred with simple A-waves in 32 nerves and with complex A-waves in 5 nerves.

RAD was not detected in 86 of 119 (72.3%) patients with sensorimotor axonal polyneuropathy. Of the 35 patients with RAD in irregular pattern detected in at least one motor nerve study, radiculopathy was associated in 8 and median nerve compression at the wrist in 2. In 14 of the remaining 25 patients, RAD was detected in peroneal F-wave study recorded, which could not be associated with any electrophysiological diagnosis. F-wave studies in which RAD was detected in the other 9 patients can be listed as follows: 3 tibial, 1 median, 1 ulnar, 1 median and ulnar, 2 tibial and peroneal, 1 ulnar, peroneal and tibial.

As for normal subjects, peroneal RAD was detected in only 2 of 84 (2.3%). Lumbosacral magnetic resonance imaging (MRI) examinations performed on these two participants 4 and 6 years before the EMGs due to similar symptoms requiring EMG were normal. Thus, RAD in these two participants could not be associated with any known condition.

RAD frequency differed significantly among demyelinating, axonal, and control groups (χ²(2) = 151.04, p < 0.001). Pairwise comparisons demonstrated that RAD was significantly more frequent in the demyelinating group compared with both the axonal and control groups (Fisher’s exact test, p < 0.001 for both). Additionally, RAD frequency was significantly higher in the axonal group than in controls (Fisher’s exact test, p < 0.001).

Discussion

We frequently observed RAD during F-wave studies in patients with demyelinating polyneuropathy, mostly of acquired type, while it was less common in patients with sensorimotor axonal polyneuropathy. Among those with RAD detected without any obvious electrophysiological diagnosis, the majority were the peroneal and the tibial nerves. And finally, there was peroneal RAD in only 2.3% of the normal subjects.

We consider RAD to be biologically originated signals, not artifacts, aligning with Pascoe et al.¹¹. Given the significant association between RAD and demyelinating polyneuropathies in our series, the relatively few publications in the literature on RAD suggest that this finding may have been overlooked. Initially, peripheral nerve hyperexcitability syndrome was defined with the occurrence of after-discharges after CMAP^4,12. On the other hand, Verdru et al. suggested that these after-discharges are nonspecific and could be seen in motor neuron disease and lumbosacral radiculopathy⁷. Furthermore, it was pointed out that these discharges could also be observed in the conditions such as myasthenia gravis⁶, and peripheral neuropathies including GBS^1,2,13.

Serra et al. concluded that one of the important mechanisms underlying RAD in the demyelinating neuropathies is the ephaptic re-excitation of the axon, resulting in a hyperexcitable axon sprout¹. This supports the neuronal origin of RAD. While RAD might resemble exaggerated F-waves¹⁴ due to anterior horn cell excitability^15,16, earlier occurrence of RAD after CMAPs (min 3.8 ms), short inter-discharge intervals (min 2 ms), and stereotypical waveforms argue against this the exaggerated F-wave hypothesis.

MNSP that occurs when electrical stimulation is applied to the mixed nerve innervating a tonically contracting muscle is interrupted three period at appropriate latencies¹⁷. Although multiple physiological mechanisms have been suggested to contribute to the latter periods of MNSP, it is agreed that first period (the silent part between the end of CMAP and onset of the F-wave or H reflex) results from the collision in motor axons and represents a relatively absolute silent period. In 84.7% of combined RAD-F-wave traces, the first component of RAD was in the first period of MNSP. Therefore, the emergence of RAD in the first silent period suggests that these signals do not arise from voluntarily activated motor unit discharges due to incomplete muscle relaxation.

When examining the potential relationship between RAD and ion channel dysfunction, particularly in the context of stereotypical RAD, persistent sodium currents and imbalances between sodium and leak conductances were proposed as the mechanisms facilitating repetitive firing in demyelinated axons^18–20. A simulation study suggested that increased sodium conductance relative to leak conductance may predispose axons to after-discharges²⁰. However, as ion channel function is not directly assessed in the present study, these considerations remain theoretical and should be regarded as a possible mechanistic framework rather than a demonstrated explanation.

When we consider the underlying mechanisms above, we may speculate that ectopic discharges may be caused by a more local pathology such as ion channel disruption, causing stereotypical RAD, while ephaptic excitations may be caused by a more widespread demyelinating injury, resulting in irregular RAD. However, this distinction cannot be confirmed by the current data.

It is of interest that RAD became more conspicuous in some of our patients, as we increased the stimulus intensity; and in some patients, RAD did not exist at low-intensity stimuli, but appeared at relatively high-intensity stimuli. This observation may be compatible with an ephaptic mechanism, although it cannot be considered definitive evidence. Stereotypical waveforms, as seen in some patients, likely represent one or a few motor unit potentials. By increasing the stimulus intensity, nerve fibers with higher thresholds which are associated with smaller motor units, may begin to fire and RAD at these fibers could be record. However, the possibility of a distant trigger zone should not be disregarded. Serra et al. already concluded that after-discharges appear only after maximal M waves, when their electric field is very strong¹.

In the study of Serra et al., it was stated that RAD detected in a patient with GBS, could be recorded even with surface electrodes, since they consist of large amplitude complex potentials resulting from synchronized discharges of several muscle fibers¹. We recorded RAD using surface electrodes. Other patients with demyelinating polyneuropathy in whom we could not detect RAD with the surface electrodes may have had low amplitude-RAD. If we had used a needle electrode for recording, we could have detected RAD. Perhaps the RAD detected by needle electrode may represent bistable membrane-associated stimulus-induced myokymia^21,22.

Typically, the pattern of demyelination differentiate hereditary (uniform demyelination) form from acquired (segmental demyelination) neuropathies²³. However, some exceptions, such as presence of abnormal temporal dispersion²⁴ and motor conduction blocks in hereditary demyelinating neuropathies²⁵, could make the distinction difficult. Given that the distinguishing between hereditary and acquired demyelinating polyneuropathies is of paramount importance^26,27 especially for treatability²⁴, RAD may represent a potentially informative electrophysiological finding in this context. However, this observation should be interpreted cautiously, especially considering the limited number and heterogeneity of hereditary cases in the present study.

RAD in axonal polyneuropathies may hypothetically reflect early radiculopathy or entrapment causing focal demyelination not yet detectable electrophysiologically. However, this interpretation remains speculative, as the present study was not designed to evaluate structural correlates. Furthermore, the animal experiments have shown that a small proportion of primary afferent neurons in intact dorsal root ganglia fire spontaneously and the resulting after-discharges are significantly increased in chronic sciatic neuropathy²⁸. Whether a similar mechanism exists in motor fibers remains unclear. Therefore, the predominance of stimulation-evoked RAD in the sciatic nerve (peroneal and tibial nerves) should be interpreted as a hypothetical observation rather than direct evidence of causality. Similarly, although the two normal subjects with peroneal RAD had previously normal lumbosacral MRIs, the absence of simultaneous imaging prevented definitive conclusions regarding subclinical radiculopathy.

Limitations

A limitation of this study is that subtle operator-dependent differences over the extended study period cannot be entirely excluded and it may represent a potential source of variability. Another limitation of this study is the relatively small sample size and genetic heterogeneity of the hereditary demyelinating neuropathy group; this may limit the power of comparisons between hereditary and acquired cases. Although prolongation of the F-wave latencies and the MNSP, which are related to the same conduction pathways, is expected to show a correlation in demyelinating polyneuropathies, the possibility that prolonged minimum F-wave latencies may extend the initial period of the MNSP should be taken into consideration. This may potentially influence the temporal classification of RAD and represents another limitation of the study. Therefore, the relationship between RAD and the MNSP should be interpreted with caution.

Conclusion

For a long time, we have frequently been observing RAD in patients with electrophysiologically diagnosed demyelinating polyneuropathy in our laboratory. It is thought that RAD may serve as a supportive electrophysiological finding in the assessment of demyelinating polyneuropathy and help differentiate it from axonal neuropathies and healthy controls. Its potential role in distinguishing hereditary and acquired forms needs to be validated in larger, genetically well-defined cohorts. Similar to the indirect discharges included in the EAN/PNS guidelines for GBS²⁹, future studies may clarify whether RAD has a potential adjunctive role in diagnostic evaluation for acquired demyelinating polyneuropathies.

Author contributions

H.U.U, Ç.M.T. and F.G.Y. devised the conceptualization and methodology. H.U.U and F.G.Y performed the data curation. H.U.U and Ç.M.T. were involved in data analysis, validation, and visualization of the work. H.U.U drafted the manuscript, and all authors worked on the review and editing of the manuscript.

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.