Individuals with cervical spinal cord injury (SCI) experience significant function impairment of their upper extremities, with subsequent challenges in activities of daily living (1). SCI predisposes survivors to muscle atrophy, contracture, and increased fat infiltration, which might be one of the causes of limited physical activity (2, 3). Muscle atrophy can impact the specific fiber type and is frequently accompanied by a slow muscle fiber to fast type fiber shift (4). The muscle changes may be also related to muscle disuse and higher levels of body fat leading to obesity and metabolic changes (5, 6). In addition, Biering-Sørensen and colleagues reviewed that there were muscle contractile changes besides the fiber-type transformation after SCI (7).

Various electrophysiological methods have been used to examine the paralyzed and spastic muscles after SCI. Abnormal electromyography (EMG) from paralyzed muscles of SCI subjects commonly include fibrillation and sharp waves (8, 9), spontaneous motor unit activity, motor unit loss, and enlarged motor unit action potentials (10–12), as a result of degeneration of spinal motoneurons and peripheral neuromuscular deterioration (13). These findings also provide evidence of alterations in muscle architecture and components after SCI. Bioelectrical impedance analysis (BIA) has been used to evaluate fat-free mass and total body water of SCI subjects (14). The whole body impedance results of BIA are limited in this application because selectivity is lost from individual muscles from paralyzed parts of body segments. In addition, one assumption of BIA method is that tissue-specific resistivity is constant for all body segments. However, tissue-specific resistivity has been shown to vary among body segments because of differences in tissue composition, hydration levels, and electrolyte concentrations (15). The results of whole body BIA could be impacted by the changes in the subject’s body gestures and joint angle position during the measurement, which further affects the repeatability and reliability of the results (16).

As a non-invasive bioimpedance-based technique, electrical impedance myography (EIM) is generally used to detect and quantify muscle health by sending high-frequency, low-intensity current into a discrete region of muscle tissue and measuring the consequent voltage (16). Three mainly used EIM parameters (17) include [1] resistance (R), representing the resistivity to current flow in the extracellular and intracellular fluids; [2] reactance (X), indicating how the current flow is affected by cell membranes and by the various fascia of the body; and [3] phase angle (θ), which is defined as θ = arctan (X/R). EIM has been used to reveal the changes in diseased muscles, such as in myopathies (18), radiculopathies (19), spinal muscular atrophy (SMA) (20, 21), Duchenne muscular dystrophy (DMD) (22), and amyotrophic lateral sclerosis (ALS) (23, 24). For example, the EIM findings from SMA demonstrated its usefulness of accurately categorizing patients between type 2 and type 3 (20). A longitudinal SMA study could also detect virtually static or no growth of active muscle maturation (21). In addition, EIM values were found to be correlated with standard ALS severity (25). Another important application of EIM is to investigate the anisotropy of muscle tissue which can be disturbed after pathological changes in neuromuscular diseases (26–28), since skeletal muscle is electrically anisotropic and current flow goes more easily in the longitudinal direction (parallel to the muscle fibers) than in the transverse direction. Neuromuscular diseases such as ALS might have impact on the anisotropy characteristics followed from muscle atrophy and fiber disorganization (29).

Electrical impedance myography has not yet been applied in assessing muscle health and anisotropy changes after SCI. The objectives of this study are to apply multifrequency local EIM technology to detect muscle impedance changes in individuals with SCI by comparing to age-matched healthy control subjects. To the best of our knowledge, this study represents the first effort of using EIM to examine SCI subjects, and the findings can provide useful information to understand architectural changes in muscles paralyzed by SCI.

We have demonstrated that, in SCI, multifrequency EIM patterns differ substantially from those in healthy control subjects at biceps brachii, especially the reactance and phase angle. The reactance of the SCI group was significantly lower than that from healthy controls at 50 kHz (Figure 2B). Reactance is related to the obstruction to current flow produced by the presence of capacitors (17). The smaller reactance in SCI is similar to previous EIM studies in aging, myopathy, and ALS (29–31). The decrease of reactance might be related to the disruption of cellular membrane integrity and injury (32). In addition, in animal study of ALS rat, the reduction of reactance was consistent with the ALS progression and also significantly correlated with the change in motor unit number estimation (MUNE) (33). Ahad and Rutkove also found the decreased reactance in animal study of sciatic injury (34). A tenable physiological interpretation of this condition is that muscle cell membranes have deteriorated severely in their ability to control the ionic flow (35) or the sarcolemma of individual myocytes are reduced in size and the membrane’s ability to momentarily store electrical charge is decreased (34). By using the sciatic crush rat model, Ahad and Rutkove correlated the EIM data and standard neurophysiologic parameters (such as nerve conduction, MUNE, and needle EMG) and found parallel changes in EIM to nerve motor conduction parameters (36). Therefore, EIM changes could be used to document the muscle cell membrane property changes after SCI as demonstrated in other populations. Rehabilitation interventions used to improve cellular membrane integrity or increase muscle myocyte size and number could be evaluated by EIM techniques that may help with the design of targeted therapies and/or outcome measures.

We found a significant smaller phase angle value in the SCI group at 50 and 100 kHz (Figure 2C). Phase angle represents the time shift of a sinusoid wave when passing through the muscle, which was previously used as a major outcome measure of EIM since it is less affected by the architecture and shape of the muscle compared with R or X (37). The smaller phase angle in the SCI group is in line with the data from a multicenter study of DMD children (38) and ALS subjects (23). Reductions in cell size and the presence of more connective tissue in the muscle also reduce the EIM phase angle (16). Another possible explanation might be that muscle disuse such as SCI causes type 1 fiber atrophy with a slow-to-fast fiber-type shift (2). Some muscles may undergo disuse myocyte atrophy and other muscles may be impacted by reinnervation (11). In addition, a smaller 50/200 kHz phase ratio was observed in this study (Figure 3), which may reflect the inherent property of muscle since electrical material properties of fat and muscle are distinct. Fat contributes similarly to the measured data at both 50 and 200 kHz, whereas muscle significantly contributing at 50 kHz and only minimally at 200 kHz. Thus by taking a ratio, it is possible to cancel out the effect of fat (22). Therefore, phase angle changes might be related to muscle cell atrophy and changes in fiber type in preserved paretic muscle of SCI survivors, which may reflect the subsequent effects of disuse, denervation, and/or reinnervation observed in electrophysiological studies (11, 39).

We did not reveal significant difference of resistance between the healthy control subjects and SCI subjects at selected frequencies, although the SCI group showed a larger trend of resistance. By using ultrasound, Gorgey and colleagues documented muscle atrophy of the wrist extensors that was 20–35% smaller in persons with tetraplegia compared with healthy controls (40). However, the impact of muscle size on impedance has remained a question and needs further discussion. For example, Rutkove and coworkers found the alterations in the EIM data in normal children were not simply an effect of increasing muscle mass and limb girth but rather a change in the structure and composition of the muscle itself (21). Another possible explanation for the resistance difference might be due to changes in subcutaneous fat thickness.

Significant decrease of anisotropy (ratio of transverse and longitudinal) of R in SCI at all three frequency values was found, when compared to those from healthy controls (Figure 4). This is similar to previous results from advanced ALS patients (28) and also similar to data from myopathy (29). Electrical anisotropy of skeletal muscle represents the inherit muscle fiber geometry within the muscle and the electrical current flowing more readily along muscle fibers than passing across them, since current perpendicular to fibers needs to go through more myocytes (26). After SCI, the survivors may suffer the loss of muscle fibers and associated fatty infiltration of tissue, leading to a general reduction in anisotropy since the muscle fibers are replaced by isotropic tissues such as inflammatory cells, connective tissues, and fat (29). Reduction of anisotropy was also observed in bovine muscle when it undergoes tissue disruption (27). However, neurological diseases may not always lead to decreases in anisotropy. For example, ALS patients in the early stage demonstrated an increased reactance anisotropy that may be due to reinnervated muscle fibers (28). Some metabolic myopathies might only cause subtle changes in muscle fiber size and structure and are associated with minimal deposition of endomysial connective tissue or fat, which would limit the ability of EIM to detect abnormalities. In addition, it is possible that unusual anisotropy results are simply due to the variation of direction of fiber arrangements within the muscle, while the fiber properties themselves are not changed (17). With further advances in measurement techniques, e.g., muscular ultrasound imaging, it may be possible to help understand the mechanism of alterations of electrical anisotropy and put this distinctive property of muscle to effective practical use and to differentiate the muscle changes from neurogenic and myopathic diseases.

In this study, both the EIM and electrical anisotropy showed a frequency dependency. Studies have shown that utilizing multifrequency measures may be more sensitive to disease status and progression over time (21, 22). With the frequency changes in the injecting current, the relative weights of resistive (fluid) and reactive (membranes) contributions to the total impedance are shifted, i.e., the cell membranes are modeled like the capacitor in an electrical circuit that very high frequencies make nearly no contribution of reactance to impedance (37). The changes in the frequency spectrum of the EIM may be due, in part, to reductions in myocyte size and number. This will lead to a decrease in the reactance, because in advanced chronic disease, myocytes are replaced with connective tissue and fat, and these changes in muscle composition will distort the normal spectra. Our results showed that the healthy control subjects had a larger log(freq)-R slope than the SCI subjects (Figure 5), which indicates that with increased frequencies the resistance of healthy control muscles decreased precipitously compared to paralyzed muscles. This finding is similar to a previous EIM clinical study on SMA children (20) in which they also observed a relatively small slope in the SMA group compared to the age-matched control group which may be caused by muscle fibrosis and fatty infiltration. The findings of this study support that using multifrequency EIM may potentially be clinically useful for extracting muscle-specific parameters for SCI survivors (16).

There are some cautions to be taken when interpreting the results of this study. We used the localized EIM technique that has been validated with high reproducibility (41). The probe size fits the biceps brachii muscle and the distance between current and voltage electrodes when calculating anisotropy is 1.4 cm. The optimal electrode distance between injecting current and measuring voltage to use is likely dependent on details of the skin–subcutaneous fat layer thickness of measured muscle (28, 42). With increasing fat and adipose tissue, a greater inter-electrode distance may be required to allow the current to go through the subcutaneous fat layer and then to effectively measure the anisotropy. This may be a plausible explanation why the anisotropy in this study is much lower than that measured on bare bovine muscles (27). A future dedicated study assessing the relationship between inter-electrode distance, subcutaneous fat layer thickness (could be measured from ultrasound), and the measured EIM changes in SCI will be useful. In addition, we referred to previous published studies that demonstrated the localized EIM’s good reliability and repeatability on human subject (35, 41). It may be better to have body mass index of SCI subjects in the future to correlate with EIM that can further verify the body size or limb girth effects. Second, this EIM comparison of patients with neurologically intact control subjects is considered as only the first step to demonstrate the feasibility of applying EIM in muscle health evaluation and to identify parameters that can represent clinically evident changes in muscle tissue of SCI survivors. An attempt was made to correlate the X, phase angle, 50/200 kHz phase ratio, AR-resistance, and log(freq)-R slope to SCI clinical characteristics, such as time since injury and motor score. We only find statistically significant relationship between X and time since injury (p = 0.035, Table 2). Therefore, the changes in EIM alone do not fully explain the mechanism of complex neuromuscular changes after SCI. The major finding of this EIM study is the difference between paralyzed and matched muscles, and EIM could provide a convenient approach to quantifying paralyzed muscle changes. The combination of other imaging techniques (i.e., musculoskeletal ultrasound), electrophysiological techniques such as MUNE (25) and quantitative motor unit action potential analysis is warranted to provide a comprehensive picture to aid in the understanding of denervation, reinnervation, and/or other neuromuscular alterations due to SCI.

In summary, we demonstrated that it is feasible to apply EIM in muscle health evaluation in persons with cervical SCI. Our results revealed reductions in reactance and phase angle in SCI group compared to healthy control which might be related to the changes in muscle inherent and structural property alterations after injury. In addition, we also observed that the AR of resistance was smaller which might be related to loss of muscle fibers and fat infiltration. These findings might provide novel information for understanding the muscle changes after SCI. The multifrequency EIM technique can further strengthen EIM as a potential tool for the non-invasive assessment of paralyzed muscles.

All procedures of the study were performed in accordance with the Declaration of Helsinki, and this study was approved by the Committee for the Protection of Human Subjects (CPHS) of University of Texas Health Science Center at Houston and TIRR Memorial Hermann (Houston, TX, USA). All the subjects gave their written consent (or witnessed verbal consent if unable to write) before the experimental procedures.

LL performed experiment, data analysis and interpretation, and wrote the first draft of the manuscript. AS provided clinical support for the study including SCI subject recruitment and evaluation, helped data analysis and interpretation, and revised manuscript. HS and XL provided technical support for the study, helped with experiment and data analysis, and revised manuscript. PZ oversaw different aspects of the study.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.