Low Frequency-Repetitive Nerve Stimulation and Factors Affecting Decremental Response in Amyotrophic Lateral Sclerosis: A Retrospective Study of 449 Cases

A number of studies have demonstrated that decremental response to low frequency repetitive nerve stimulation (LF-RNS) is frequently observed in amyotrophic lateral sclerosis (ALS). However, due to the small sample size involved in previous studies, large discrepancies exist about the positivity rates of LF-RNS tests and factors affecting decremental response. This retrospective study of 449 cases, the largest sample size ever reported, shows that the overall positivity rate of LF-RNS is 49.9%. 3Hz RNS delivered to the accessory nerve has the best sensitivity and highest positivity rate. It obviously increases in response to upper limb onset, disease progression rate < 0.5 score/month, denite ALS and electromyography positive(EMG(+)) in sternocleidomastoid muscle. There is a linear increase in the decrement percentage of CMAP amplitude at 3Hz RNS delivered to the accessory nerve in response to longer disease duration, longer MUP duration and greater MUP duration increment percentage. These ndings substantially advance the understanding of RNS results in ALS patients and effectively instruct clinical application. and random errors, we collected and analyzed an adequately large sample size of case data to improve the generalizability of the results. The present study shows that 3Hz RNS delivered to the accessory nerve has the highest diagnostic sensitivity. The overall positivity rate of 3Hz RNS (decremental response exceeding 10%) for the 449 patients is 49.9%, while that for the 364 clinical denite ALS patients reaches 52.5%. As for the range of decremental CMAP amplitude at RNS test, very few studies have reported the exact values of the range in addition to mean and standard deviation (Sun et al., 2018; Tomoko et al., Zheng et al., 2017a; Hu et al., 2018). The largest decrement of CMAP amplitude that has ever been reported was 50% in our previous research (Sun et al., 2018) while this study shows that it can reach 55% score/month, denite ALS and EMG(+) in sternocleidomastoid muscle. Perfect negative linear correlations exist between either disease duration or MUP duration and decremental percentage of CMAP amplitude. The regression models indicate that the neuromuscular junction


Introduction
Amyotrophic lateral sclerosis (ALS), a fatal motor neuron disease, causes progressive muscle atrophy and weakness in bulbar, limbs and trunks. It also affects sensory system, autonomic nerve system and senior cortex functions, leading to non-motor symptoms such as paresthesias, pain, cognitive impairment, affective Therefore, we collected the records of 449 ALS cases, the largest sample size that has ever been reported, from the Neurology Department of the First Medical Center, Chinese PLA General Hospital and performed a retrospective study of RNS involved in the cases with an aim to determine the positivity rate at LF-RNS, the best stimulus position and frequency, and to quantify the relationship of decremental CMAP amplitudes to clinical indicators and motor unit potential (MUP) duration, so as to substantially advance the understanding of LF-RNS results in ALS patients and effectively instruct clinical application.

Methods
The subjects of the study were 449 patients diagnosed with ALS based on the revised EI Escorial (Brook et al., 2000) at Neurology Department of the First Medical Center, Chinese PLA General Hospital from June 2016 to September 2020 and their clinical and neuroelectrophysiological data were collected with certainty that none of the subjects had other NMJ diseases. This study was approved by the Medical Ethics Committee of Chinese PLA General Hospital. All enrolled subjects signed informed consents by themselves or their legal guardians.
All the patients were classi ed into four diagnostic levels (clinically de nite, clinically probable, clinically probable laboratory-supported and clinically possible) according to the revised EI Escorial criteria. Their data included gender, age, disease duration, onset site, the results of needle electromyography (EMG) and LF-RNS, etc. The patients were scored according to the scales of ALSFRS-R (Cedarbaum et al., 1999), and the rate of disease progression (ΔFS) was calculated by the following formula: ΔFS = [48 -ALSFRS-R score]/duration (month) (Labra et al., 2016). The increasing values of △FS indicate the faster rates of the disease progression.
The electrodiagnostic studies, including RNS and EMG, were performed on a Keypoint workstation(31A06, Alpine Biomed ApS, Denmark). Skin temperature over the examined muscle was maintained at 32℃ or above throughout the entire measurement. Surface electrodes were used to record the belly-tendon compound muscle action potential (CMAP). LF-RNS was performed in the following muscles: deltoid for the axillary nerve (n = 65), trapezius for the accessory nerve (n = 446), abductor digiti minimi (ADM) for the ulnar nerve(n = 308), tibialis anterior(TIB) for the common peroneal nerve(n = 73). A low frequency of 3Hz train of the 10 stimuli was delivered to the nerves and recorded. The peak-to-peak CMAP amplitude decrement was measured by the decremental percentage of the forth CMAP as compared to the rst CMAP amplitude. Based on the conventional criterion, a decremental response of 10% or greater in 1Hz or 3Hz were considered positive, in accordance with the suggestions of the American Academy of Emergency Medicine Quality Assurance Committee. With reference to the normative values as set up by EMG laboratory of Peking Union Medical College Hospital, MUP mean duration increment percentage was calculated by the following formula: mean duration increment percentage (%) = (measured mean duration-normal mean duration)/normal mean duration. Measured mean duration indicates the mean value of 20 MUP durations. According to our EMG department criteria, MUP duration exceeding 20%, with or without abnormal spontaneous potential, is considered as chronic denervation potential, i.e. EMG positive result.
MS-Excel 2016 and MATLABR2019b software (multifunction, trial free version) were employed for statistical analysis. Measurement statistics satisfying the normal distribution was expressed as mean ± SD, while nonnormal distribution data was expressed as mean ± SD and median. Enumeration data were presented as the number of cases and percentage. Chi-square test was used for comparison between different enumerative groups.
Least squares method and F-test were used for simple linear regression. P < 0.05 was considered statistically signi cant.

LF-RNS positivity rate
The distributions of decremental responses on four detected nerves are summarized in Table 1. Results show: the positivity rate of 3Hz is signi cantly higher than that of 1Hz; 3Hz RNS on the accessory nerve, followed by that on the axillary nerve (40.0%), has the highest positivity rate (47.3%) and it covers all positive cases stimulated by 1Hz. The positivity rates of both common peroneal nerve and ulnar nerve are lower than 8%. As it is, the accessory nerve is the best ideal site for detection and 3Hz is the best ideal frequency. Statistic analysis also shows the distributions of positivity rates at different diagnostic levels (Table 2). Overall, the positivity rate of 3Hz RNS test for all diagnostic levels(n = 449) in at least one muscles is 49.9%(n = 224), and that of clinically de nite ALS(n = 364) at 3Hz RNS reaches 52.5%, both of which are signi cantly higher than that in other three diagnostic levels(n = 85, positivity rate = 38.8%)(Chi-square, p = 0.05).  (Table 3).  Table 4. No substantial difference is found in either between male and female patients, or among patients with an onset age of ≥ 60 and those < 60, nor is any difference found in other age brackets measured. No substantial difference is found among patients with ALSFRS-R at initial visit ≥ 40 and those < 40,, nor was any difference found in other ALSFRS-R score groups.
In contrast, the 3Hz RNS positivity rate of the patients with disease duration exceeding 6 months at initial visit(n = 339) is signi cantly higher than that of whose within 6 months (n = 107). The 3Hz RNS positivity rate of the patients with upper limb onset (n = 188, positivity rate = 60.6%) is signi cantly higher than that of those with lower limb onset (n = 158, positivity rate = 32.9%) or bulbar onset (n = 100, positivity rate = 45.0%). The 3Hz RNS positivity rate of the patients with ΔFS<0.5score/month(n = 56.5%) is signi cantly higher than that of those with ΔFS > 0.5score(n = 88, positivity rate = 40.2%). The positivity rate of de nite ALS (n = 362, positivity rate = 49.4%) is signi cantly higher than that of patients at other three diagnostic levels (n = 84, positivity rate = 37.6%). The positivity rate of patients with EMG (+) in sternocleidomastoid muscle (n = 183, positivity rate = 60.7%) is signi cantly higher than that of patients with EMG (-) (n = 141, positivity rate = 30.5%). In short, the positivity rate of 3Hz RNS delivered to accessory nerve obviously increases in response to disease duration > 6 months, upper limb onset, ΔFS<0.5 score/month, de nite ALS and EMG(+) in sternocleidomastoid muscle.

The relationship between decremental response at 3Hz RNS delivered to accessory nerve and disease duration at initial visit
Scatter plot is used to display disease duration and the decremental percentage of CMAP amplitude at 3Hz RNS test (Fig. 1). A linear regression analysis indicates that the decremental response is linearly associated with disease duration (Y = 8.890-0.126X, F = 12.8, P = 0.000).
3.5 The relationship of decremental response at 3Hz RNS delivered to accessory nerve respectively to MUP duration and to duration increment percentage in sternocleidomastoid muscle 324 patients underwent 3Hz RNS test on the accessory nerve along with needle EMG in the sternocleidomastoid muscle. Scatter plots are used to display the correlation between the decremental range of CMAP amplitude and percentage have signi cantly negative linear correlations with decremental response of 3Hz RNS delivered to the accessory nerve, suggesting that the more serious neurogenic damage is detected by EMG, the more obvious will be the decremental response. Besides, as can be seen from the regression model, MUP duration is more responsive than the duration increment percentage to the decremental response. . The largest decrement of CMAP amplitude that has ever been reported was 50% in our previous research (Sun et al., 2018) while this study shows that it can reach 55% (Table 1). in at least one muscle was observed in 41.3%. However, our study of the 449 cases shows that in 15.5%, 7.6% and 4.0% of the cases RNS decremental responses exceed 20%, 25% and 30% respectively. As far as we can see, the implementation of the current Eelectrophysiological Diagnostic Criteria will result in the misdiagnosis of more than 15.5% of ALS patients. Then, what percentage of RNS decremental response can serve as a criterion to exclude ALS still warrants further investigations.

According to the Electrophysiological
In addition to RNS test results in ALS patients mentioned above, the results of our statistical analysis also provide robust evidence for the correlations between RNS decremental responses and ALS clinical manifestations. For the rst time, a perfect negative linear correlation between disease duratrion and 3Hz RNS decremental responses is established by regression analysis and a signi cantly higher positivity rate in patients with ΔFS < 0.5 score/month(n = 88) is shown by Chi square. As can be drawn from the ΔFS formula, the longer the disease duration is, the lower the ΔFS will be. Taking the ΔFS formula and the above two results in account, our study indicates that the NMJ dysfunction is deteriorating as the disease progresses, especially after 6 months.
As for the relationship between RNS decremental responses and EMG results, we are the rst to establish a regression model (Fig. 2) and compare the sensitivities of two indicators (i.e. MUP duration and duration increment percentage) to RNS test. Our study shows that MUP duration rather than the range of duration increment is more suitable to be used in future studies. The regression model in Fig. 2 con rms the correlation between the impairment of motor neuron and NMJ dysfunction from the perspective of clinical neuroelectrophysiology for the rst time. By two intersecting lines, namely x = 20 and y=-10, the coordinate plane is divided into four quadrant (Fig. 2b). It's worth noting that scatters in the third quadrant satisfying the requirement of RNS(+) and EMG(-) seems to indicates that in some cases, RNS test is more sensitive than EMG and NMJ dysfunction is prior to the loss of motor neuron. Whether this electrophysiological phenomenon can serve as electrophysiological evidence for "dying-back" hypothesis (Michal et al., 2011) still needs further investigations.
This study is not free of limitations. Firstly, the positivity rates of clinically probable cases (n = 46), clinically probable laboratory-supported cases (n = 20) and clinically possible cases (n = 18) are 30.4%, 35.0% and 61.1%, respectively. However, the sample size of the cases at the three levels is relatively small in comparison with that of de nite ALS. The difference in positivity rate among these three diagnostic levels has very little meaning. On the one hand, our neurology department is one of the best national medical research institution which mostly receives patients suffering rare and serious diseases and most patients at the early stage of ALS just see a doctor near their hometown as their muscle weakness is not obvious. On the other hand, people are not well informed of ALS knowledge, which makes patients at early stage likely to be missed or even misdiagnosed. That's why we think multicenter retrospective studies are needed in the future. Secondly, present statistic analysis shows the positivity rate of 3Hz RNS test delivered to the accessory nerve in upper-limb-onset group is 60.6%, leading all the others and followed by bulbar group (45.0%) and lower limb group (32.9%), which is in line with some previous study . The probable reason for this obviously high positivity rate in the upper-limb-onset group may be ALS's preference for spreading around its onset site. However, as to whether the positivity rate on the common peroneal nerve at 3Hz RNS might be higher in patients with lower-limb-onset than that in patients with other onset sites, our statistic analysis can hardly lead to a conclusion because of the low positivity rate (3.8%) of RNS test applied to the common peroneal nerve and the lack of positive cases (n = 3). Thirdly, as can be seen from the Fig. 2b, some cases are found with RNS (+) and EMG (-). So some questions remain unanswered. What's the proportion of these cases in ALS population? Is disease duration of these cases shorter than others? How is their diagnostic levels? To answer them, we'll keep on collecting data and trying to illuminate these questions in future study.
In conclusion, this is a retrospective study based on 449 cases, with the largest sample size ever reported. It discovers that the overall positivity rate of RNS test is 49.9%. CMAP decremental response at 3Hz RNS applied to the accessory nerve has the best sensitivity and highest positivity rate (52.5% in de nite ALS). It obviously increases in response to upper limb onset, ΔFS<0.5 score/month, de nite ALS and EMG(+) in sternocleidomastoid muscle. Perfect negative linear correlations exist between either disease duration or MUP duration and decremental percentage of CMAP amplitude. The regression models indicate that the neuromuscular junction (NMJ) dysfunction is deteriorating as the disease progresses; the more serious neurogenic damage is detected by EMG, the more obvious will be the decremental response. Cases which are found with RNS (+) and EMG (-) may enlighten the pathogenic mechanisms from the perspective of clinical neuroelectrophysiology.

Declarations
Con ict of interest: There are no con icts of interest to declare.

Ethical publication statement
The patient has provided consent for the use of the clinical information. This study has been approved by the Chinese PLA General Hospital Ethical Review Committee and has therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.

Funding
National Natural Science Foundation of China (Grant No. 81671278).

Disclosure statement
The authors disclose no con icts of interest.
CRediT authorship contribution statement Jinghong Zhang had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Xusheng Huang and Jinghong Zhang contributed equally to this work. The relationship between disease duration and the decremental amplitude of CMAP at 3HzRNS delivered to accessory nerve