Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease characterized by the degeneration of upper and lower motor neurons, resulting in a wide range of motor and extra-motor system symptoms (1). The clinical presentation exhibits considerable heterogeneity among patients, with some presenting with limb muscle weakness indicative of spinal-onset disease and others manifesting dysarthria or dysphagia associated with bulbar-onset disease. Progressive muscle atrophy gradually develops in various muscle groups such as the hands, fingers, and legs (2, 3). However, this process of muscle atrophy is asymmetrical and lacks consistency in terms of speed and severity—a phenomenon referred to as dissociated muscle atrophy (4, 5). Clinically, due to the heterogeneity of ALS clinical phenotypes and a lack of special diagnostic markers in the early stages, there is an estimated delay of 14 months in diagnosing ALS (6). Therefore, this phenomenon of dissociated muscle atrophy may serve as a potential reliable diagnostic marker of ALS (7–9).

Among the observed muscle atrophy phenomena in ALS, the split-hand sign (SHS) is a well-established clinical manifestation characterized by pronounced dysfunction of the thenar group of intrinsic hand muscles, specifically the abductor pollicis brevis and first dorsal interosseous, compared to the hypothenar muscles (10–12). Furthermore, instances of the split-finger and split-leg phenomena have been documented in ALS. These clinical manifestations are characterized by more pronounced weakness and atrophy of the first flexor digitorum profundus muscle (FDP1) compared to the fourth flexor digitorum profundus muscle (FDP4) and FDP to the little finger (13), as well as heightened dysfunction of ankle plantar flexors relative to dorsiflexor muscles (2). While another study chose extensor digitorum brevis (EDB) and the abductor hallucis (AH) in ALS patients to suggest two muscles involved in the split-leg sign (14, 15).

The split-elbow sign (SES), which indicates preferential dysfunction of the biceps brachii compared to the triceps muscle, has been identified in recent studies (16, 17). Khalaf et al. observed a significant decrease in the medical research council (MRC) scores for the biceps brachii compared to those of the triceps (16). Furthermore, a study calculated the split-elbow index (SEI) using compound muscle action potential (CMAP) amplitude and demonstrated its ability to discriminate between individuals diagnosed with ALS and controls (17).

However, the assessment of the SES using MRC scores and CMAP amplitude in these studies has certain limitations. The use of MRC scores introduces subjectivity in evaluating muscle strength, potentially leading to biased results. Notably, a previous study investigating dissociated muscle atrophy of SES in ALS also used the MRC score and reported a higher score for the biceps muscle compared to the triceps muscle (18), which is inconsistent with the findings of Khalaf et al. (16).

The motor unit number index (MUNIX), a non-invasive electrophysiological measure, has demonstrated its utility as a biomarker for quantifying the functionality of lower motor neurons. In our previous study (41), we established reference values for MUNIX in five muscles, which can aid in monitoring the progression of neuromuscular diseases. Furthermore, MUNIX has been identified as a highly sensitive and effective marker for tracking the progressive loss of motor units in ALS (19–21). In addition, numerous studies also highlighted the potential of ultrasound in evaluating lower motor neuron involvement by quantifying echo intensity (EI) and fasciculation (22–24). Compared to the CMAP index alone, both MUNIX (20) and quantitative ultrasound indices, such as the echo intensity (EI) index (24), have exhibited superior sensitivity and accuracy when evaluating the severity of SHS in ALS. However, the quantified SEI derived from MUNIX and EI has not been investigated in SES. Therefore, this study aimed to investigate the characteristics of the SES using the SEI, including SEI_CMAP, SEI_MUNIX, and SEI_EI in ALS patients and explore their clinical application as biomarkers in ALS diagnosis.

In total, 70 patients who were diagnosed with ALS were enrolled in this study, consisting of 44 males and 26 females (Table 1). The mean disease duration was 14.54 ± 11.16 months, while the ALSFRS-R score averaged at 39.49 ± 6.29. King’s staging for ALS was 2.0 (1.0). In addition, a control group comprising 41 DCs and 40 HCs was recruited for comparison purposes. Age and gender distribution did not exhibit any significant differences among the ALS, DC, and HC groups (54.96 ± 10.5 vs. 52.41 ± 8.0 vs. 51.68 ± 10.74). Furthermore, based on disease severity as determined by ALSFRS-R scores, ALS patients were further categorized into three groups: “severe” (scores ≤35), “moderate” (36 ≤ scores ≤41), and “mild” (scores ≥42). Notably, it is worth mentioning that the severe group had a longer duration than both the moderate and mild groups.

Neurological disorders included in the DC group were as follows: Kennedy’s disease (2.5%), multifocal motor neuropathy (2.5%), facioscapulohumeral dystrophy (2.5%), myopathy (5.11%), myasthenia gravis (8.18%), cervical myelopathy (9.21%), and chronic inflammatory demyelinating polyradiculoneuropathy (13.35%).

In the ALS and DC groups, the score in MRC between biceps brachii and triceps brachii was assessed. It was not significantly different from ALS patients and DCs (X² = 3.95, p = 0.14) (Figure 1). Biceps brachii MRC scores were significantly lower than triceps brachii MRC scores in ALS patients (p < 0.0001). The neurophysiological details are succinctly presented in Tables 1, 2, encompassing CMAP amplitudes, MUNIX, and EI values recorded over the biceps brachii and triceps brachii among the various groups. ALS patients exhibited significantly lower CMAP amplitudes than HCs and DCs in the biceps brachii, while such change was observed only when compared to HCs in the triceps brachii (p < 0.05). ALS patients exhibited significantly lower MUNIX values than HCs and DCs both in the biceps brachii and triceps brachii (p < 0.05). In addition, ALS patients exhibited significantly higher EI values than HCs and DCs in the biceps brachii, but this change was observed only when compared to HCs in the triceps brachii (p < 0.05). Subsequently, we calculated SEI to quantify the degree of preferential weakness and atrophy in muscles. The SEI_MUNIX and SEI_CMAP were found to be significantly lower in ALS patients than those in DCs and HCs, and the SEI_EI was significantly higher than that in DCs and HCs (Table 1) (p < 0.05).

The subgroup analysis revealed a significantly lower level of SEI_MUNIX in different severity groups with ALS compared to DCs and HCs. Furthermore, SEI_MUNIX values decreased as the disease progressed, according to the ALSFRS-R score (p < 0.001). However, no significant differences were observed in SEI_CAMP and SEI_EI among different severity groups with ALS (p > 0.05) (Table 2). By comparing the median of the different violin plots, we could preliminarily determine whether there was a difference in the central trend between the different groups. Overall, the upper limb-onset group exhibited the lowest SEI_CMAP and SEI_MUNIX values, while the upper limb-onset group had the highest SEI_EI values than lower limb- and bulbar-onset groups (Figures 2A–C) (p < 0.05).

The ROC curve was used to conduct power analysis for SEI_CMAP, SEI_MUNIX, and SEI_EI. The area under curve (AUC) of SEI_CMAP was 0.74 (95% CI: 0.65–0.82), with a cut-off value of 58.7 distinguishing ALS patients from controls with a sensitivity of 87.7% and specificity of 54.3%. For SEI_MUNIX, the AUC was 0.85 (95% CI: 0.78–0.92), and the cut-off value of 60.6 differentiated ALS patients from controls with a sensitivity of 90.1% and specificity of 72.9%. As for SEI_EI, the AUC was determined as 0.82 (95% CI: 0.76–0.89), while the cut-off value of 113 differentiated ALS patients from controls with a sensitivity of 69.1% and specificity of 84.3% (Supplementary Figures S1a–d). In addition, we conducted diagnostic power comparisons among SEI_CMAP, SEI_MUNIX, and SEI_EI using R software package. The results revealed that there was a significant difference between SEI_CMAP and SEI_MUNIX (p < 0.05), indicating that SEI_MUNIX exhibited diagnostic power. However, the comparison between these curves for SEI_MUNIX and SEI_EI did not yield significant differences (p = 0.613). Furthermore, a comprehensive diagnosis strategy combining these methods resulted in an ROC curve with an AUC value of 0.90 (95% CI 0.84–0.96), sensitivity of 95%, and specificity of 81% (Figure 3A). SEI_EI might present better power for diagnosis in ALS patients and DCs (AUC 0.876, p = 0.012), while the cut-off value of 113 could differentiate ALS patients from controls with a sensitivity of 80.5% and specificity of 84.3% (Figure 3B).

Collectively, we conducted a cohort study to validate SES as a clinical feature in ALS and establish a comprehensive diagnostic approach. When compared to DCs and HCs, the SEI_EI increased, while SEI_CAMP and SEI_MUNIX decreased significantly in ALS patients, indicating quantified weakness of biceps brachii relative to triceps brachii using SEI measurements. In addition, we observed progressively lower SEI_MUNIX values in ALS patients with advancing disease stages defined by the ALSFRS-R. To assess diagnostic accuracy, an ROC curve was drawn to determine which method was more suitable for diagnosing ALS. Overall, this study provides support for the presence of dissociated muscle atrophy in ALS through electrodiagnostic and ultrasound methods combined with clinical diagnosis and suggest that SEI_MUNIX could serve as a predictor of early diagnosis and disease progression. SEI_EI might be helpful to distinguish ALS from other neurological disorders.

Prior to discussing the presence of the split-elbow sign and the utility of the SEI, it is important to note that the split-hand phenomenon was first identified and proposed (31, 32). SHS likely represents an early and highly specific clinical manifestation of ALS (7, 33). Regarding potential pathophysiological mechanisms, cortical motoneuronal hyperexcitability appears to be the most plausible cause of dissociated muscle atrophy (34). In addition, an another study suggested that the intrinsic hand muscles occupy a more prominent cortical somatotopic location, which may contribute to glutamate-mediated hyperexcitability through larger cortical projection neurons (35). As for SES, previous studies have consistently demonstrated that the triceps brachii muscle is relatively preserved compared with the biceps brachii muscle, as confirmed by various methodologies (16, 17). Transcranial magnetic stimulation experiments have revealed a higher influx of corticomotoneuronal inputs to the biceps brachii than to the triceps brachii (36). These suggest that corticomotoneuronal hyperexcitability may also contribute to SES. However, a study conducted in Asia reported inconsistent results for SES and indicated preferential weakness in elbow extensors over flexors (18). In contrast, our study found that biceps brachii MRC scores were significantly lower than triceps brachii MRC scores in ALS patients. In addition, biceps brachii MRC scores were not significantly lower in ALS patients than in DCs in our study, which is consistent with the study by Pavey et al. (17). This may be related to disease severity of ALS, and there is a certain subjectivity when using this assessment tool. Hence, our study applies electrodiagnostic and ultrasound examinations to reaffirm that SES is indeed a clinical feature of ALS. Our study further supports the notion of preferential dysfunction in the biceps brachii muscle. Considering that subjective estimation may introduce potential bias into MRC muscle score system assessments, it implies that more precise techniques are required alongside this scoring system.

To quantify the SHS as a potential diagnostic biomarker of ALS, the split-hand index was established (8). In line with this, the SEI, which is derived neurophysiologically from CAMP amplitudes and the MUNIX values recorded by the biceps and triceps brachii, may also serve as a novel diagnostic biomarker in ALS (37). However, it should be noted that compensatory collateral reinnervation by remaining motor neurons can potentially influence CMAP amplitude in denervated muscle fibers (24). Ultrasound imaging remains an effective method for measuring EI due to its minimal interference from compensatory reinnervation processes (38), thus suggesting an advantage for SEI_EI based on these considerations. Our study used SEI_CAMP, SEI_MUNIX, and SEI_EI measurements to find statistically significant differences in SEI between ALS patients and controls.

To assess the efficacy of the three methods in different stages of ALS, we categorized ALS patients into mild, moderate, and severe groups according to their ALSFRS-R scores. SEI_MUNIX revealed a significant distinction when compared among ALS,DCs, and HCs. Although the difference between HCs and DCs was significant, it was inevitable because DCs included patients with similar symptoms but not ALS. Therefore, the significant difference between the DC and ALS groups was more concerning. More importantly, a progressive decline in SEI_MUNIX values was observed as ALS patients progressed through the disease stages. In line with this finding, Kim et al. also reported a significant correlation between split-hand index MUNIX (SI_MUNIX) and ALSFRS-R (r = 0.33) (20). Hence, these results suggested that SEI_MUNIX could serve as an early diagnostic marker and an indicator of disease progression. While SEI_EI exhibited significantly higher values than controls at each stage of ALS, there were no notable differences in SEI_EI and SEI_CMAP among subgroups of different severity within the ALS population; indicating that these two measures may hold promise as differential diagnostic indices but are not suitable for assessing disease progression. However, further studies with larger sample sizes are warranted to validate these findings. Interestingly, SEI_CAMP and SEI_MUNIX have lower statistical significance in the upper limb-onset ALS patients, but SEI_EI is higher in the upper limb-onset group, which is consistent with the results of SHS by EI (24, 39). This alignment can be attributed to the fact that EI evaluates alterations in the muscular architecture and infiltrated muscular fat and fiber content (30).

In our study, we used the ROC curves for SEI_CAMP, SEI_MUNIX, and SEI_EI and compared the AUC for diagnostic power of the above methods. We found that SEI_MUNIX had significantly better performance than SEI_CAMP, but there was no significant difference between SEI_MUNIX and SEI_EI. Neuwirth et al. also found MUNIX as an electrophysiological marker was significantly better than CMAP for testing lower motor neuron loss (40). It was consistent that MUNIX and SEI_MUNIX  could be better potential markers than CMAP and SEI_CAMP. In addition, the ROC curve was used for three methods between the ALS group and DCs. There was a significant difference in SEI_EI compared to SEI_CMAP. A previous study found that EI and fasciculation quantified by muscle ultrasound were tested for sensitivity and specificity to distinguish ALS and non-ALS (39). Hence, for the efficiency verification of two methods, it might take more participants to observe whether there is a difference between SEI_MUNIX and SEI_EI. In addition, considering that the diagnosis of ALS requires a combination of clinical symptoms and electrophysiological strategies, we combined the three methods as a joint diagnostic strategy and compared their diagnostic accuracy and found that it seemed to be more significantly helpful as a biomarker.

Of course, some potential limitations in this study should be considered: First, the small number of ALS patients and controls is a possible limitation. Then, since our study was a retrospective cross-sectional design, it is necessary to use SEI_MUNIX and SEI_EI tracking ALS progression as a longitudinal study in the future. Third, the MUNIX depends on whether patients can complete the corresponding movements of the target muscles; hence, patients with poor strength or fit would bias the results. Further studies with larger, multi-center, disease control and prospective design are needed to clarify the diagnostic utility of SEI_MUNIX and SEI_EI in ALS.

The SES is significantly more common in ALS patients when compared to disease controls and healthy controls. It could be supported by the reduction of SEI_MUNIX and the elevation of the SEI_EI. Importantly, our study found that SEI_MUNIX could serve as an early diagnostic marker and an indicator of disease progression. Although the ROC of the SEI_MUNIX indicates that it is a great approach for the diagnosis of ALS, the combined methods are important for diagnosis.