Objective The objective of this study was to investigate the usefulness of muscle ultrasound in evaluating dissociated small hand muscle atrophy, termed ‘split hand’, and its feasibility in the diagnosis of amyotrophic lateral sclerosis (ALS).
Methods Forty-four patients with ALS, 18 normal subjects and 9 patients with other neuromuscular disorders were included in this study. The hand muscles were divided into three regions, the median-innervated lateral hand muscle group (ML), the ulnar-innervated lateral hand muscle (UL) and the ulnar-innervated medial hand muscle (UM), and the muscle echo intensity (EI) and compound muscle action potential (CMAP) were measured. We calculated the split hand index (SHI) using muscle EI (SHImEI) and CMAP (SHICMAP) for comparison among groups. The SHI was derived by dividing muscle EI (or CMAP) measured at the ML and UL by that measured at the UM.
Results The SHImEI was significantly higher in patients with ALS (51.7±28.3) than in normal controls (29.7±9.9) and disease controls with other neuromuscular disorders (36.5±7.3; P<0.001), particularly in upper limb-onset ALS (66.5±34.0; P<0.001). Receiver operating characteristic curve analysis indicated that the SHImEI had significantly better diagnostic accuracy than the SHICMAP.
Conclusions The SHImEI was more sensitive in evaluating dissociated small hand muscle atrophy compared with the SHICMAP and may be a reliable diagnostic marker for differentiating ALS from other neuromuscular disorders and healthy controls.
Statistics from Altmetric.com
Amyotrophic lateral sclerosis (ALS) is a particularly devastating neurodegenerative disease that destroys the upper and lower motor neurons and leads to the inability to walk, talk, breathe or blink.1–3 Because of its poor prognosis, ALS diagnosis can be traumatic for patients, and clinicians should therefore exercise caution when diagnosing this disorder. However, accurate diagnosis of ALS is difficult, particularly in its early stages, because of the absence of a definitive diagnostic test and the heterogeneity of clinical phenotypes. Thus, an early and specific diagnostic marker for ALS is needed. In ALS, hand muscle atrophy preferentially affects the lateral hand, including the abductor pollicis brevis (APB) and first dorsal interosseous (FDI) muscles, with relative sparing of the medial hand muscles (the abductor digiti minimi, ADM).1 4 This is termed the split hand phenomenon and has been suggested to be an early and characteristic clinical feature of ALS.
Recent studies have attempted to use the split hand phenomenon in the diagnosis of ALS through compound muscle action potentials (CMAPs) recorded after electrical stimulation of nerves.5–7 However, split hand parameters based on CMAPs (such as the split hand index (SHI), APB:ADM ratio and FDI:ADM ratio) have some limitations. These parameters may not be available in patients with advanced ALS who have significant intrinsic hand muscle atrophy, resulting in CMAPs that are unobtainable or too small to measure accurately. In addition, CMAP amplitude is influenced by compensatory collateral reinnervation of denervated muscle fibres by the remaining motor neurons.8 In other words, muscle fibres are actually denervated, but the CMAP amplitude does not completely decrease due to compensatory reinnervation, resulting in less reliable split hand parameters in relatively early-stage ALS.
Recent studies reported that ultrasound (US) could help detect lower motor neuron involvement by evaluating echo intensity (EI), fasciculation and muscle volume in ALS.9–16 Muscle EI can be measured even in muscles where the CMAP cannot be obtained due to significant muscle atrophy. Muscle EI is less affected by compensatory reinnervation than CMAP due to its relationship with intramuscular fatty deposition and fibrous changes after denervation,17 18 which are not affected by compensatory reinnervation.13
The objective of this study was to investigate the usefulness of muscle EI in evaluating dissociated small hand muscle atrophy, termed ‘split hand’, and its feasibility for a reliable image marker in the differential diagnosis of ALS.
Patients diagnosed with ALS were recruited from a cohort at the ALS clinic in two university-affiliated hospitals between August 2015 and October 2016. The diagnosis of ALS was established based on the revised El Escorial criteria using clinical and electrophysiological data.19 We excluded patients with ALS complicated by other neuromuscular disorders based on sensory symptoms suggesting peripheral neuropathy or cervical radiculopathy; cervical spine problems, such as cervical spondylosis or herniated nucleus pulposus, on spine MRI; and/or abnormal electrodiagnostic findings other than motor neuron disease. Subjects with hand muscle weakness caused by other neuromuscular disorders were enrolled as disease controls. Healthy subjects who were recruited through the community and confirmed to have no abnormal findings on a nerve conduction study (NCS) were enrolled as normal controls. All subjects were required to provide written informed consent.
Clinical evaluation for ALS
Demographic data, including age, sex, height and weight, medical history, and disease duration were obtained from the subjects. The Amyotrophic Lateral Sclerosis Functional Rating Scale-Revised (ALSFRS-R) was used to assess disease severity and functional disability. Based on the ALSFRS-R subscores, we determined the stage of patients with ALS using the ALS Milano-Torino staging system.20 We also measured the Medical Research Council (MRC) sum score. The total MRC sum score was derived by assessing the following muscle groups bilaterally: shoulder abduction, elbow flexion, wrist dorsiflexion, thumb abduction, hip flexion, knee extension, ankle dorsiflexion and great toe dorsiflexion, yielding a maximal score of 80. A summated score of three items of the ALSFRS-R such as writing, eating and dressing was also used as a variable for hand function in each subject. Clinical evaluation was performed on the same day as the US examination.
Muscle EI of the hand muscles was measured on both hands by an expert (HYS) according to the method described in a previous study21 using B-mode US equipment (HD15 System, Philips Ultrasound, Bothell, Washington, USA) with a 5–12 MHz transducer and the following equipment settings: 50 dB gain, 56 dB dynamic range and 3 cm depth. During the measurements, subjects were asked to fully relax in the supine position with their forearm in full supination.
Muscle EIs at three regions of interest (ROIs) were measured as follows: the median-innervated lateral hand muscle group (ML; region including the APB, opponens pollicis and superficial head of the flexor pollicis brevis), the ulnar-innervated lateral hand muscle (UL; FDI) and the ulnar-innervated medial hand muscle (UM; ADM). The ROI of the ML muscle group was determined in the same manner as described in a previous study: (1) the line connecting the upper tip of the first metacarpal bone and the flexor pollicis longus tendon was the inferior border; (2) the line along the superior part of the ML muscle fascia was the superior border; (3) a vertical line passing the upper tip of the first metacarpal bone was the left border; and (4) a vertical line passing the upper tip of the flexor pollicis longus tendon was the right border.21 Cross-sectional images of the UL and UM clearly showed the hyperechoic epimysium, and the ROIs of the UL and UM muscles were determined by tracing just inside of the epimysium, respectively. The EIs of each ROI were measured at the point where the EIs were brightest by adjusting the transducer perpendicular to each ROI.
Greyscale image analysis was used to determine muscle EI within each ROI. A greyscale image consists of achromatic colour pixels with different brightness values ranging from 0 (pure black) to 255 (pure white). Muscle EI was determined as mean pixel brightness in each ROI, which was automatically calculated by software built into the US equipment (HD15 System, Philips Ultrasound; QLab ROI quantification). Muscle US images were acquired three times for each region, and the mean muscle EI value of the three images was used in the final analyses.
We calculated the SHI using muscle EIs from the three ROIs. The SHI derived from muscle EI was calculated by multiplying muscle EI measured over the ML and UL muscles and dividing this product by muscle EI recorded over the UM muscle as follows (figure 1):
SHImEI=MLmEI × ULmEI/UMmEI
Nerve conduction studies
NCSs were performed using standard electrodiagnostic equipment (Viking IV; Nicolet Biomedical, Madison, Wisconsin, USA). The median and ulnar nerves were stimulated electrically at the wrist, and the resultant baseline-to-peak CMAP amplitude (mV) was recorded at the APB, FDI and ADM muscles using disc electrodes positioned in a belly tendon arrangement. The SHI using the amplitude of the CMAPs was calculated by multiplying the CMAP parameters recorded over the APB and FDI muscles and dividing this product by the CMAP parameters recorded over the ADM muscle as follows5:
SHICMAP=APBCMAP × FDICMAP/ADMCMAP
Differences in demographics among patients with ALS, patients with other neuromuscular disorders as disease controls and normal healthy control subjects were analysed. In addition, we compared the SHI calculated using muscle EI with those calculated using the amplitude of the CMAP. Patients with ALS were further divided into three subgroups based on the ALSFRS-R score (≥41, mild; 36–40, moderate; and ≤35, severe) and onset region (bulbar-onset group, upper limb-onset group and lower limb-onset group). Analysis of variance and Χ2 methods were used for statistical analysis. Post-hoc analysis of significant differences among groups was performed using Tukey’s honest significance test. To determine cut-off values with optimal sensitivity and specificity, receiver operating characteristic (ROC) curve analysis of the SHImEI and SHICMAP was performed by comparing the ALS patient group with the healthy and disease control groups. Spearman correlation analysis was performed to assess the correlation between SHICMAP and SHImEI, and to evaluate the relationship of SHI with MRC sum score, ALSFRS-R and ALSFRS-R hand subscore. P values <0.05 were considered statistically significant. Statistical analyses were performed using SPSS V.21.0.
Clinical characteristics of the subjects
Forty-four patients (26 men, mean age 61.7±7.0) with ALS, 9 (5 men, mean age 62.9±7.8) with other neuromuscular disorders and 18 normal subjects (9 men, mean age 71.8±4.1) were recruited. Of the 44 patients with ALS, 9 were classified as definite, 29 as probable, 5 as laboratory-supported probable and 1 as possible ALS according to the El Escorial criteria. The mean ALSFRS-R score at the time of testing was 36.0±9.4, the mean MRC sum score was 66.1±11.4 and the mean disease duration was 18.6±16.3 months. Among the 44 patients with ALS, most patients were in stage 0 (n=37) or stage 1 (n=4; 93.2%); 2.3% (n=1) were in stage 3 and 4.5% (n=2) were in stage 4. The disease controls included in our study were diagnosed with cervical radiculopathy (n=2), chronic inflammatory demyelinating polyneuropathy (n=1), carpal tunnel syndrome (n=1), multifocal motor neuropathy with conduction block (n=1), spinobulbar muscular atrophy (n=1), myotonic dystrophy type 1 (n=1) and polymyositis (n=2).
Comparison of the SHI
The SHImEI was significantly higher in patients with ALS (51.7±28.3) than in normal controls (29.7±9.9) and disease controls with other neuromuscular disorders (36.5±7.3; P<0.001). Patients with ALS also had a significantly lower SHICMAP (4.6±3.6) than normal subjects and patients with other neuromuscular disorders (P<0.001; table 1). Comparison among the ALS subgroups according to disease severity using the ALSFRS-R revealed significant differences in the SHImEI (P=0.002), but no statistically significant difference in the SHICMAP (table 2). Post-hoc analysis revealed that the SHImEI was significantly higher in the severe ALS group (67.4±32.0) than in the mild (42.5±21.3) and moderate ALS groups (47.3±25.9). In addition, the increase in the SHImEI was most prominent in patients with ALS with upper limb-onset disease (66.5±34.0) compared with patients with ALS with bulbar-onset (42.4±18.5) or lower limb-onset (40.9±17.1, P<0.001; figure 2). However, the SHICMAP showed no significant differences among subgroups according to ALS onset pattern.
Diagnostic utility of the SHI
To determine the diagnostic utility of the SHI, analysis of ROC curves in the entire ALS cohort was performed. ROC curve analysis showed that the SHImEI had significantly better diagnostic accuracy (area under the curve (AUC)=0.76) than the SHICMAP (AUC=0.12) as a useful diagnostic marker of ALS. An SHImEI value of 42.3 differentiated patients with ALS from healthy and disease controls with 52.6% sensitivity and 80.4% specificity. In upper limb-onset ALS, in particular, the diagnostic accuracy of the SHImEI was significantly increased (AUC=0.84; figure 3). A cut-off SHImEI value of 47.9 differentiated the upper limb-onset ALS group from the other ALS groups, healthy subjects and disease controls with 73.1% sensitivity and 84.4% specificity.
Correlation of the SHI with other clinical variables
SHImEI was significantly and negatively correlated with SHICMAP (r=–0.516, P<0.001), MRC sum score (r=–0.469, P=0.001), ALSFRS-R (r=–0.464, P=0.001) and ALSFRS-R hand subscore (r=–0.610, P<0.001). In addition, SHICMAP showed a positive correlation with MRC sum score (r=0.367, P=0.023), ALSFRS-R (r=0.456, P=0.004) and ALSFRS-R hand subscore (r=0.639, P<0.001).
The present study explored the utility of lateral to medial hand muscle EI ratio as a novel US split hand marker for the diagnosis of ALS. Using muscle US, we found a significantly higher SHImEI in patients with ALS compared with healthy controls and patients with other neuromuscular disorders. We also observed that the SHImEI was significantly different between patients with severe and mild-to-moderate ALS according to the ALSFRS-R score, with a higher value in severe patients. Moreover, ROC curve analysis demonstrated that the SHImEI had significantly better diagnostic accuracy than the SHICMAP as a useful diagnostic marker of ALS, particularly in upper limb-onset disease.
In this study, we found that the SHImEI is more powerful in differentiating patients with ALS from controls than the SHICMAP. Although there were significant decreases in the SHICMAP, these reductions were not as striking as the changes in the SHImEI. ROC curve analysis showed that the SHImEI (AUC=0.76) had significantly better diagnostic accuracy than the SHICMAP (AUC=0.12) as a useful diagnostic marker of ALS. In addition, comparison among the ALS subgroups according to disease severity using the ALSFRS-R revealed significant differences in the SHImEI, but no statistically significant differences in the SHICMAP. A possible explanation of this finding is that compensatory reinnervation may temporarily preserve CMAP amplitude, while reinnervation does not reverse increases in EI.16 EI elevation results from altered muscle architecture and fatty and fibrous infiltration of muscle. When disease progression is slow, adequate reinnervation due to collateral sprouting from unaffected nerve fibres may prevent decreases in CMAP amplitude. Assessment of the EI may therefore be more useful than CMAP amplitude for detecting subclinically affected regions. The other possible reason for the observed differences in SHICMAP and SHImEI is the summation of CMAP of the median innervated thenar muscle with the ulnar innervated thenar muscle evoked by volume conduction during nerve stimulation. Avoiding the inclusion of the volume conduction response in the CMAP amplitude is critical when determining the diagnostic value of the SHI. Therefore, the SHImEI may be a better diagnostic marker than the SHICMAP, particularly in upper limb-onset ALS.
The diagnostic value of the SHICMAP in our study was lower than that reported in a previous study, which reported 74% sensitivity and 80% specificity in differentiating ALS from other neuromuscular diseases.5 A recent study using the CMAP amplitude and motor unit number index also demonstrated the high sensitivity (85%) and specificity (79.5%) of the SHICMAP in differentiating ALS from healthy controls.22 The discrepancy between previous studies and our results may be related to variation in the inclusion criteria used. One previous study only included patients with ALS affecting the upper limbs5; however, we included patients with ALS according to the revised El Escorial criteria, regardless of upper limb involvement.
One interesting point of this study is that the raw muscle EI of individual hand muscles could not be used to differentiate patients with ALS from mimic groups, whereas SHImEI could distinguish these groups (table 1). This suggests that the relative difference in atrophy between medial and lateral hand muscles is much more important than the severity of atrophy in any individual hand muscle, which confirms the general perception that the split hand phenomenon is characteristic of ALS.
Previous studies have suggested that muscle US is a useful tool for detecting fasciculations, particularly in the tongue, and thus substantially increases the diagnostic sensitivity for ALS.9–12 In our study, we did not evaluate the presence of fasciculations on US, but considering our findings together with these previous reports, the combination of US measurement of SHImEI and fasciculations may result in accurate and sensitive ALS diagnosis, especially for bulbar-onset and upper limb-onset disease. Future studies are needed to identify the diagnostic value of combined use of SHImEI and fasciculation on US for ALS.
This study has some potential limitations that should be considered. The first limitation is the small number of disease controls and severe ALS cases. The ALS mimic group was too small and too heterogeneous to enable statistical analysis of the specificity of muscle US. Future studies with larger disease control groups using a prospective, blinded design are needed to establish the true diagnostic value of US examination. In addition, independent validation of the cut-off value in a larger cohort will be needed prior to application of SHImEI in clinical practice. The normal healthy control group was older compared with the ALS and disease control groups. Muscle EI increases with age, but the patient group showed a significantly increased EI compared with the normal healthy control group despite this age difference. Use of the SHI as the ratio of muscle EIs might minimise the influence of individual confounding factors including age compared with a single muscle EI measurement. Therefore, the older age of the healthy group might not have affected the determination that SHI based on muscle US is a reliable tool for assessing patients with ALS. Muscle US and NCS were conducted by an investigator who was not blinded to patient diagnosis, and this in turn could potentially influence the results. However, it would be difficult to effectively blind the investigator because patients with ALS are more likely to show fasciculations on US. In addition, because muscle EI was calculated later using the QLab software after hand muscle US, we do not believe that this affected the results significantly.
In conclusion, the SHI of muscle EI was more sensitive in the evaluation of dissociated small hand muscle atrophy compared with parameters based on CMAP and was a reliable marker for the differentiation of patients with ALS from those with other neuromuscular disorders as well as healthy controls. The major practical conclusion of the study is the suggestion to include SHI US parameters into the diagnostic evaluation of ALS, especially for upper limb-onset disease.
HYS and JP contributed equally.
SHK and B-JK contributed equally.
Contributors B-JK and SHK conceived and designed the research. HYS performed the experiments. HYS, JP, YHK, K-WO and B-JK analysed the data. HYS, B-JK and SHK drafted the manuscript.
Funding This article was supported by grants from the Technology Innovation Program, which is funded by the Ministry of Trade, Industry and Energy (MI, Korea) (grant no 100499743, ‘Establishing a medical device development open platform, as a hub for accelerating close firm-hospital communication’), and from the KDA and the Korean Health Technology R&D Project, Ministry of Health, Welfare and Family Affairs, Republic of Korea (grant no HI16C2131).
Competing interests None declared.
Patient consent Obtained.
Ethics approval The institutional review board of the participating centres approved this study (#HYI-10-01e3 and #AN15198-004).
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The data used and/or analysed during the current study are available from the corresponding author on reasonable request.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.