Daily Electrical Muscle Stimulation Enhances Functional Recovery Following Nerve Transection and Repair in Rats

Abstract

Background. Incomplete recovery following surgical reconstruction of damaged peripheral nerves is common. Electrical muscle stimulation (EMS) to improve functional outcomes has not been effective in previous studies. Objective. To evaluate the efficacy of a new, clinically translatable EMS paradigm over a 3-month period following nerve transection and immediate repair. Methods. Rats were divided into 6 groups based on treatment (EMS or no treatment) and duration (1, 2, or 3 months). A tibial nerve transection injury was immediately repaired with 2 epineurial sutures. The right gastrocnemius muscle in all rats was implanted with intramuscular electrodes. In the EMS group, the muscle was electrically stimulated with 600 contractions per day, 5 days a week. Terminal measurements were made after 1, 2, or 3 months. Rats in the 3-month group were assessed weekly using skilled and overground locomotion tests. Neuromuscular junction reinnervation patterns were also examined. Results. Muscles that received daily EMS had significantly greater numbers of reinnervated motor units with smaller average motor unit sizes. The majority of muscle endplates were reinnervated by a single axon arising from a nerve trunk with significantly fewer numbers of terminal sprouts in the EMS group, the numbers being small. Muscle mass and force were unchanged but EMS improved behavioral outcomes. Conclusions. Our results demonstrated that EMS using a moderate stimulation paradigm immediately following nerve transection and repair enhances electrophysiological and behavioral recovery.

Introduction

Recovery from peripheral nerve injuries is rarely complete. The more proximal the injury, the less likely a patient will fully recover. ¹ Currently, the gold standard for treating peripheral nerve injuries is surgical repair. ² Microsurgical techniques have advanced such that we have now reached the limits of optimizing recovery following nerve transection with only surgical means.

Electrical stimulation of proximal nerve stumps prior to repair accelerates nerve outgrowth across the injury site^3-7 but may still leave distal muscles denervated for long periods resulting in their atrophy. Others have shown that electrical stimulation of muscle reverses or diminishes muscle atrophy and modifies fiber type characteristics^8-12 but none have demonstrated that the stimulation improves functional or behavioral outcomes. Previous studies have only examined axonal sprouting after partial denervation demonstrating chronic electrical stimulation (12-24 hours per day) ¹³ or chronic treadmill running^14,15 reduced sprouting, in particular terminal sprouting. Other studies have shown that electrical stimulation of muscle provides no functional benefit with stimulation being applied infrequently or with subthreshold amplitudes being used that could not elicit strong contractions.^16,17

Tam and Gordon ¹⁸ suggested that perhaps a more moderate activity paradigm is needed to obtain optimal results. In our previous work, we investigated a clinically translatable electrical stimulation paradigm ¹⁹ . This paradigm consists of daily muscle stimulation for one hour (5 days per week) with the idea that patients may be able to perform this treatment at home using small surface muscle stimulation devices. Our early work showed that this paradigm enhanced motor unit numbers following long term muscle denervation and subsequent nerve repair. ²⁰ In a separate study, we showed a residual effect of the therapy following 1 month of electrical stimulation after nerve injury and immediate repair with examination 2 months after the cessation of stimulation. ²¹ More recently, we investigated the effects of stimulating over a 2-week period. Even at this early time point, nerve outgrowth and muscle reinnervation were both significantly improved in stimulated muscles. ²²

Based on our previous work, we hypothesized that extending the duration of electrical muscle stimulation throughout the reinnervation period may enhance functional recovery. In addition, we also examined how reinnervation is affected by visualizing if axons reconnect to the muscle endplate directly from a nerve trunk or through collateral sprouting.

Methods

Animals

Experiments were performed on 29 adult male Thy-1 GFP (green fluorescent protein) transgenic rats ²³ weighing between 300 and 400 g. These rats express GFP in neurons and their axons throughout the peripheral nervous system, thus permitting their direct visualization. Animals were housed in temperature controlled rooms and given standard rat chow and water ad libitum. A standard 12-hour light cycle was maintained in the housing room. Experimental procedures were approved by The Hospital for Sick Children Laboratory Animal Services and performed according to guidelines set out by the Canadian Council on Animal Care.

Surgical Procedures

All animals were anesthetized using 5% isofluorane for induction and 2% isofluorane to maintain anesthesia. The right hindlimb of each animal was prepared for surgery using aseptic technique. A lateral incision was made below the femur to expose the sciatic nerve. The tibial nerve was then isolated and transected 10 to 13 mm proximal to the entry of the nerve branches to the gastrocnemius muscle. Two 9-0 epineurial sutures were used to oppose the proximal and distal nerve stumps for repair. Thereafter, the right gastrocnemius muscle was implanted with 2 intramuscular stimulating electrodes (Cooner Wire, AS 631) and the wires anchored to the muscle using nonabsorbable suture. ¹⁹ Briefly, the cathode was implanted in the belly of the muscle (through both heads of the gastrocnemius) with the anode placed 4 to 5 mm distal within the same muscle. Rats were randomly divided into 2 groups: those receiving electrical muscle stimulation of the gastrocnemius (immediate repair + electrical muscle stimulation [REP + EMS]) and rats without EMS (immediate repair [REP]). Animals that were to receive daily EMS had the proximal ends of the intramuscular stimulating wires threaded subcutaneously and externalized in the neck as described previously. ¹⁹ Electrodes in the control group (REP) were not externalized.

Electrical muscle stimulation began 2 days after surgery to allow recovery from the procedure, and proceeded for 1 month (n = 5), 2 months (n = 5), or 3 months (n = 5). Control animals (REP) also followed the same time course: 1 month (n = 4), 2 months (n = 5), and 3 months (n = 5).

Electrical Muscle Stimulation

The stimulation paradigm consisted of daily 1-hour sessions producing 600 contractions in 1 hour. Each contraction was 400 ms in duration, which consisted of a 100 Hz pulse train with biphasic 200- to 400-µs duration pulses. Electrical pulse amplitudes were adjusted throughout the 1-hour period to ensure that vigorous contractions were elicited, as noted visually.^9,10,24,25 All animals were awake throughout the stimulation sessions and placed in custom-designed restrainers with their right hindlimb hanging freely to ensure that no loading effects on muscle recovery were taking place. Up to 5 rats were stimulated at one time using a custom-designed multichannel electrical stimulator. ¹⁹

Functional Behavioral Measurements

Animals that were followed over a 3-month period were assessed once per week using walking track and tapered beam measurements. These functional measurements quantified functional recovery of overground and skilled locomotion, respectively.

Walking Track Analysis of Overground Locomotion

To determine recovery of flat-surface locomotion we used the tibial function index (TFI) as calculated by Bain et al. ²⁶ Briefly, the hind paws of all rats were dipped in finger paint and allowed to walk across a white sheet of paper approximately 1 meter in length.^27-29 Measurements included the experimental print length (EPL), the normal print length (NPL), experimental toe spread (ETS), normal toe spread (NTS), experimental intermediary toe spread (EIT), and the normal intermediary toe spread (NIT). Based on these measurements the TFI was calculated using the following equation ²⁶ :

$$-37.2\left(\mathrm{EPL}-\mathrm{NPL}\right)/\mathrm{NPL}+104.4\left(\mathrm{ETS}-\mathrm{NTS}\right)/\mathrm{NTS}+45.6\left(\mathrm{EIT}-\mathrm{NIT}\right)/\mathrm{NIT}-8.8$$

Tapered Beam Analysis

We performed skilled locomotion analysis using the tapered beam test in order to examine functional recovery after tibial nerve transection and surgical repair with and without electrical stimulation of denervated muscles.^27,28 For 2 weeks prior to the initial surgeries, animals were trained to traverse an elevated ledged tapered beam (80 cm long, 5 cm wide). ²⁷ Following tibial nerve transection and repair, animals were assessed once per week using the tapered beam with each run across the beam being video-recorded for future analysis. A total of 3 to 5 satisfactory runs were used for final measurements and video analysis was completed by a trained observer who was blinded to whether or not the denervated muscles in the rats were subjected to daily EMS. The number of times an animal slipped off the beam with their affected hindlimb was counted to quantify sensorimotor deficits. Hind paws that slipped onto the ledge were assigned a score of 1 (full slip). Paws that touched the side of the beam were assigned a score of 0.5 (half-slip). A slip ratio was then calculated as the number of slips divided by the total number of steps taken across the beam as previously.^27,28,30

Motor Unit Number Estimation

Following the end of each time period (1, 2, or 3 months) animals were anesthetized and functional measurements performed. Motor unit numbers were estimated using a modified incremental technique. ³¹ Briefly, 2 monopolar needle recording electrodes were placed in the gastrocnemius muscle and connected to a custom biopotential amplifier to record band-pass filtered (20-500 Hz) electromyographic (EMG) responses. A pair of hook electrodes was placed proximal to the tibial nerve coaptation site and connected to a Grass stimulator (model S9). Electrical pulses of 50-µs duration were used to stimulate the nerve and custom written LabVIEW software automatically characterized unique EMG responses using a 3-level wavelet decomposition scheme. ³² Amplitudes were increased slowly until 18 to 20 unique responses were collected. The number of motor units was then estimated by the ratio of the average motor unit size (calculated as the area under the rectified EMG response) and the size of the maximal M-wave response.

Force Measurements

Whole muscle isometric contractile force was measured by placing 2 monopolar stimulating needle electrodes in the gastrocnemius muscle. The calcaneal tendon was severed with the soleus and plantaris muscles removed from the calcaneus. The remaining gastrocnemius muscle and calcaneus assembly was then attached to a force transducer (Grass model FT03) that was connected to a custom amplifier. The knee was fixated and muscle length adjusted to produce maximal twitch force. Stimulation pulses 1 ms in duration were elicited using a stimulator (Grass technologies model SD9) and isometric twitch and tetanic contractions (100 Hz frequency) were recorded.^28,29 Contraction and half-relaxation times were measured from the twitch force profiles.

Muscle Mass

After the completion of the motor unit number estimation and force measurements, animals were deeply anesthetized with Euthanyl and transcardially perfused with physiological saline followed by cold 4% paraformaldehyde. Gastrocnemius muscles were then excised and weighed. The medial head of the gastrocnemius was dissected and placed in a solution of 30% sucrose in 4% paraformaldehyde for 5 days for cryoprotection.

Immunohistochemistry

Medial gastrocnemius muscles from rats of the REP and REP + EMS groups of 3-month duration were embedded in optimal cutting temperature embedding medium and frozen in a liquid nitrogen–cooled stainless steel bowl. Samples were then placed in a cryostat at −20°C and serial sections (80-100 µm thick) were cut and placed onto glass slides. Sections were then immunostained with rhodamine-bungarotoxin (α-BTX) conjugated to Alexa Fluor-594 (1:400, Invitrogen) and incubated overnight at 4°C to label the acetylcholine receptors in the postsynaptic membrane. Slides were then imaged using a confocal microscope (Olympus IX81) and z-stacks created (50-60 µm thick with a 2 µm slice thickness) for analysis of reinnervation.

Quantitative Analysis of Muscle Reinnervation

A minimum of 300 random endplates were examined (mean = 427) in each muscle using ImageJ software (National Institutes of Health). Each muscle endplate was characterized into 1 of 5 categories: normal reinnervated (one axon per muscle endplate arising from a nerve trunk), reinnervated by a nodal sprout, reinnervated by a terminal sprout, denervated, or reinnervated but unknown. Endplates that were characterized as reinnervated but unknown had colocalization of both green (axon) and red (endplate) fluorescence but the connecting axon was not entirely visible due to a sectioning artifact. The number of counted endplates in each category was then expressed as a percentage of the total number of sampled endplates for each rat.

Statistical Analysis

Differences in motor unit number estimation and muscle forces between nonstimulated and electrically stimulated muscles were compared at each time point using an unpaired 2-tailed t test. Quantification of the pattern of reinnervation for each of the characterized categories was also compared using an unpaired 2-tailed t test. For behavioral test analysis a repeated-measures 2-way analysis of variance was used. Contraction and half-relaxation times were compared using a 2-way analysis of variance with a Tukey correction for multiple comparisons. Statistical significance was defined as P < .05. All results are presented as mean ± standard error of the mean (SEM).

Results

Motor Unit Number and Size

Gastrocnemius muscles that were electrically stimulated with 400-ms duration 100-Hz trains for 1 hour daily had a significantly greater number of motor units (MUs) at each time point (Figure 1A). Starting at 1 month following tibial nerve transection and surgical repair, more tibial nerves reinnervated the stimulated muscles with the MU numbers already being significantly greater as compared with nonstimulated muscles (37 ± 6 vs 22 ± 2). The difference was greater at 2 months (104 ± 19 vs 57 ± 9). By 3 months after injury and repair, MU numbers in the stimulated muscles (142 ± 25) were the same as those of normal uninjured control muscles (155 ± 32), the differences not being significant, in contrast to those of nonstimulated muscles ( 81 ± 7).

Figure 1.

Motor unit number and size. (A) Motor unit numbers estimated using the incremental technique were significantly greater in stimulated muscle at 1, 2, and 3 months post–nerve injury and repair compared with nonstimulated muscle. (B) Mean values of the motor unit size measured as the integrated electromyography (EMG) response were significantly smaller in stimulated muscles at 1 and 2 months post–injury and repair compared with nonstimulated muscle. The hashed line and shaded region represent the mean value and standard error of normal uninjured muscles. REP, immediate nerve repair; REP + EMS, immediate nerve repair + electrical muscle stimulation; *P < .05.

The MU size (the area under the average rectified EMG response in each muscle) was significantly smaller in the stimulated muscles as compared with nonstimulated muscles at 1 month (0.15 ± 0.03 vs 0.27 ± 0.01 mV ms) and 2 months (0.26 ± 0.02 vs 0.50 ± 0.05 mV ms) after transection and surgical repair (Figure 1B). At 3 months after injury and repair, the average MU sizes were significantly larger than uninjured control muscles and not significantly different from one another. MU enlargement proceeded more rapidly in the nonstimulated muscles to reach an asymptote by 2 months with the enlargement proceeding more slowly in the stimulated muscles.

Skilled Locomotion and Walking Track Analysis

Skilled locomotion and sensorimotor deficits were assessed using a tapered beam test. Baseline measurements made prior to nerve injury resulted in very low slip ratios (approximately 9%, Figure 2A). Although, at 1-week postinjury, slip ratios in both groups were significantly lower than uninjured measurements (mean ratio of 2%), video analysis demonstrated gait impairment in all operated rats whose tibial nerve was transected and surgically repaired: The injured limbs were kept close to the midline of the body and slid posteriorly to maintain forward motion. In so doing, the experimental hindlimb did not slip off the beam. We quantified the number of slides each animal made throughout the three month recovery period. One and two weeks post tibial nerve injury and repair animals slid significantly more than at baseline (Figure 2B). This accounted for the significant reduction of the slip ratio. Starting at week 4, a separation in the slip ratio was evident between groups, with stimulated animals having significantly better outcomes than nonstimulated animals at 6, 7, 10, and 12 weeks post–injury and repair.

Figure 2.

Functional behavioral analysis. (A) The slip ratio, the number of times the operated hindlimb slipped off the tapered beam expressed as a ratio of the total number of steps across the beam, and (B) the number of slides recorded during walking across the beam are plotted as a function of time after tibial nerve transection and repair. (C) Tibial functional index calculated during overground walking is similarly plotted as a function of time after tibial nerve transection and repair. Details are provided in the text. In all cases, functional recovery, as assessed by tapered beam analysis but not by the tibial functional index in overground locomotions, was significantly better in the animals that received daily muscle stimulation (REP + EMS) as compared with those that did not: The slip ratio stabilized while it continued to climb in the rats whose gastrocnemius muscles were stimulated versus those whose muscles were not (A); the number of slides was less in the first few weeks after injury in the rats with muscle stimulation (B). REP, immediate nerve repair; REP + EMS, immediate repair + electrical muscle stimulation; *P < .05.

The tibial function index (TFI) was used to analyze overground locomotion based on walking track values. One week following nerve injury, both groups of animals were considered completely impaired with values ranging from −88 to −100 on the TFI scale. Over the course of 3 months, animals did not recover back to baseline measurements, consistent with previous reports. ²⁹ However, animals that were stimulated had significantly lower TFI scores 10 weeks after injury and repair (Figure 2C).

Muscle Mass and Contractile Forces

Muscle mass (expressed as a percentage of uninjured contralateral muscles) was not affected by electrical stimulation of the muscle, the mass being the same for stimulated and nonstimulated muscles (Figure 3A). Muscle weights did not reach uninjured values 3 months post–injury and repair (approximately 80% of contralateral control limb).

Whole muscle twitch and tetanic contractile forces were elicited by direct stimulation of the gastrocnemius muscle to provide a measure of the force capacity of the muscle regardless of the state of reinnervation 1 to 3 months after tibial nerve transection and repair. Typical twitch and tetanic isometric contractions recorded at three months post–nerve injury and repair, are smaller and slower than contralateral uninjured control muscles (Figure 3B and D). Mean muscle forces mirrored muscle weight with no significant differences between groups at any of the 3 time points (Figure 3C and E). The contractile forces had not reached control (uninjured) levels at 3 months.

Figure 3.

Muscle weights and forces. (A) Perfused gastrocnemius muscle wet weight, expressed as a percentage of contralateral control muscle is compared as a function of time after tibial nerve transection and repair. Maximum isometric twitch (B) and tetanic (D) contractions of a normally innervated contralateral control muscle and muscles that were and were not electrically stimulated for 3 months after tibial nerve injury and repair. Mean values are compared as a function of time and shown in (C) and (E). Mean values for stimulated muscles were higher than nonstimulated muscles but not significantly different. Whole muscle twitch (C) and tetanic forces (E) expressed as a percentage of contralateral control muscles, showed no significant differences between groups at any time point for the twitch contractions but significantly higher tetanic forces at 2 months. REP, immediate nerve repair; REP + EMS, immediate nerve repair with electrical muscle stimulation; *P < .05.

Contraction times, defined as the time to reach peak twitch force from resting force, were significantly shorter in the rats in which the denervated muscles were chronically stimulated as compared to those that were not (Figure 4A). Overall, both the stimulated and nonstimulated muscles contracted more slowly with longer contraction times than normal control muscles in the first 2 months (Figure 4A). By the third month, contraction times in the stimulated and nonstimulated muscles were the same with that of the stimulated muscles being longer than uninjured control muscles.

Figure 4.

Muscle contractile dynamics. (A) Contraction and (B) half-relaxation times of maximal twitch force. Electrically stimulated muscles were consistently slower (shorter contraction time and longer half-relaxation times) than normal and their contraction speed was not affected by the stimulation. REP, immediate nerve repair; REP + EMS, immediate nerve repair with electrical muscle stimulation; *P < .05.

The twitch contractions of nonstimulated muscles relaxed significantly more slowly in both the stimulated and nonstimulated muscles after tibial nerve transection and surgical repair as compared with normal control muscles: The half-relaxation time, defined as the time taken from the peak twitch force to half of that force was significantly longer in nonstimulated muscle at 1 and 2 months after nerve transection and surgical repair (Figure 4B).

Immunohistochemistry and Microscopy

To examine the extent of muscle reinnervation, we immunostained acetylcholine receptors at the muscle endplates with α-bungarotoxin and visualized them using confocal microscopy. Details of the characterization protocol are shown in Figure 5A. Muscle endplates in both groups were primarily reinnervated by single axons arising from a nearby nerve trunk (approximately 65% of all counted endplates) and not significantly different between groups (Figure 5B). Approximately 15% of all counted endplates were reinnervated but the source of the incoming axon was not visible due primarily to artifacts resulting from the muscle cryosectioning.

Figure 5.

Neuromuscular junction reinnervation. (A) Muscle endplate receptors were stained with α-bungarotoxin (shown in red) in Thy-1 GFP (green fluorescent protein) transgenic rats whose axons were fluorescent green: “normal” reinnervation—one axon arising from a nerve trunk (indicated by *), reinnervated by an unknown source of innervation (due to cryostat cutting artifact)—a muscle endplate without attached axon (indicated by ¥), a nodal sprout identified as an axon emerging from a node of Ranvier (indicated by #), a terminal sprout identified as an axon growing distal to the last node of Ranvier (indicated by §), denervated endplates are not shown in this example. (B) The mean number of muscle endplates as a percentage of all sampled endplates is compared for gastrocnemius muscles that were and were not electrically stimulated for 3 months. Sprouting is clearly a small component of reinnervation of muscle with a significant effect of electrical stimulation being discerned only for the terminal sprouts. REP, immediate nerve repair; REP + EMS, immediate nerve repair + electrical muscle stimulation; *P < .05.

Sprouting constituted a very small portion of overall muscle reinnervation. The proportion of endplates reinnervated by nodal sprouts, expressed as a percentage of all counted endplates, was approximately 6%. There was no significant difference in this nodal sprouting in the muscles that were and were not subjected to daily 1-hour stimulation. Terminal sprouting, being but a small fraction of the reinnervation of the denervated muscles, was significantly reduced by electrical muscle stimulation (2.8% ± 0.6% vs 5.8% ± 0.6%) compared with nonstimulated muscles, which is similar to what others have found with more chronic muscle stimulation paradigms.^13,33 Denervated endplates constituted 11% of all counted endplates and were not significantly different between groups.

Discussion

Electrical Muscle Stimulation Results in Greater Number of Functional Motor Connections

Our data demonstrate that daily muscle stimulation following nerve injury and immediate repair results in more regenerating nerves reinnervating denervated muscle to result in a greater number of functional motor units and enhanced functional recovery. We chose a stimulation paradigm that is clinically translatable with the goal to have patients perform daily stimulation sessions at home without the burden of frequent physiotherapy visits. We used a “moderate” paradigm consisting of 600 contractions in 1 hour carried out daily in light of previous literature suggesting that chronic stimulation or physical activity is detrimental to the reinnervation process.^13,14,18 Previous studies that investigated electrical stimulation and reinnervation of denervated muscle were limited to gross muscle parameters such as weight and force and did not investigate motor unit numbers or size.^34,35 Our results show that motor unit numbers are significantly increased in stimulated muscle starting at 1 month. Indeed, this evidence corroborates the findings of a preliminary study in which we stimulated muscles over a 2-week period following tibial nerve transection and immediate repair. ²² We found that motor unit numbers increased as early as 2 weeks, with motor unit numbers significantly increasing in stimulated muscle compared to nonstimulated and this difference is maintained throughout 3 months of stimulation. The reason why more motor nerves reinnervate electrically stimulated denervated muscles is as yet unknown but may be due to enhanced axon outgrowth at the site of nerve transection and repair and, in turn, to earlier functional reinnervation of target muscle. There is evidence of enhanced activity-dependent release of neurotrophic factors from innervated muscle.^36,37 Treadmill running or electrical stimulation of muscle following spinal cord injury or hindlimb unloading upregulate brain- and glial-derived neurotrophic factors, both potent neuron outgrowth and branching trophic factors.^37,38 In addition, previous work showed that electrical stimulation of denervated muscle improved the structure of denervated endplates. ³⁹

Electrical Muscle Stimulation Leads to Enhanced Recovery of Skilled Locomotion

Recovery of sensorimotor function is of vital importance to patients recovering from peripheral nerve injuries.^40-42 In the current study, behavioral recovery was assessed using a tapered beam test and walking track analysis. The walking track data did not demonstrate significant differences between groups except at 1 time point (10 weeks after injury) with neither group coming close to initial baseline outcomes. Although walking track analysis has been the most commonly used behavioral metric following nerve injury, certain limitations of the test such as the dragging of limbs and the inability to fully planter place paws, have caused its validity to be challenged.^43,44 Our experiment evaluated recovery from tibial nerve injury rather than from complete sciatic nerve injury where gross behavioral changes may be more readily visible. ⁴⁵ However, findings that electrical muscle stimulation increased the numbers of motor nerves that reinnervated muscle and resulted in greater skilled locomotion on the tapered beam, indicate that the walking track metric may not be sufficiently sensitive enough to capture the significantly enhanced recovery in this injury model paradigm. An alternative explanation to the results seen with the walking track analysis is that the functional benefits of our current protocol of electrically stimulating muscle may not have been robust enough to see differences between groups. Current research in the laboratory is looking at different stimulation paradigms on behavioral recovery.

In contrast to the walking track analysis, the tapered beam test clearly demonstrated improved functional recovery in the rats in which the denervated muscles were electrically stimulated. We are the first to utilize this test in a model of tibial nerve injury, the test having been used previously to assess sensorimotor deficits primarily after central nervous system injury models^46,47 and more recently after sciatic nerve injuries. ²⁷ . For the tibial nerve injury, the test was effective in quantitating the recovery of skilled locomotion with the caveat that early changes (1-2 weeks) were masked by initial spurious gait of caudal sliding of the injured hindlimb in the first weeks after nerve repair. This was an unexpected behavioral finding in these animals and has not been previously seen in sciatic nerve injured animals.^27,28,30 This presents a potential problem in evaluating behavioral recovery in this test during the early period following surgery (ie, the first 3 weeks). Future investigations must remain cognisant of this fact. However, sliding behavior displayed by these animals decreased dramatically at the 3 week evaluation time point with slides being no more frequent than prior to injury and the ratio stabilized to normal baseline levels in both groups. At the same time, the slip ratio in nonstimulated rats continued to increase to a significantly higher level. The rats whose muscles were stimulated daily recovered normal slip ratio values, performing better than nonstimulated animals from 6 weeks post–nerve injury and repair onward. The progressive increase in the slip ratio in the rats whose muscles were not stimulated indicated possible synkinesis (nerve misdirection) that has been documented after transection injuries and results in abnormal functional innervation. ⁴⁸ Electrical muscle stimulation may reduce synkinesis and promote more selective reinnervation.^49,50

While the present work focuses on motor reinnervation, there is also evidence that electrical muscle stimulation enhances the recovery of sensory fibers carrying mechanical information (muscle spindles and Golgi tendon organs) and metabolic information (unmyelinated afferents) both of which are important to overall functional recovery. ⁵¹ Others have also shown that daily electrical stimulation of muscle can lead to faster recovery of blink reflexes in a facial nerve injury model. ⁵² This supports our results in that faster recovery may be due to enhanced sensory and motor nerve outgrowth.

Gross Muscle Parameters Were Not Different Between Groups

Muscle weight and twitch and tetanic contractile forces, as traditional metrics of functional recovery, were higher in stimulated muscles but not significantly different compared with nonstimulated muscles at 1 to 3 months after nerve transection and repair. These results suggest that our stimulation paradigm may not be sufficiently adequate. We previously designed and tested this paradigm in completely denervated muscle with 1 month of stimulation significantly improving muscle weight and contractile force. ¹⁹ However, in an immediate repair scenario where stimulation is continued throughout the reinnervation period it may be necessary to adapt the paradigm as the muscle gradually becomes fully innervated. Muscle contractile dynamics were also different than uninjured control muscles further supporting the notion that the stimulation paradigm needs to be revised as the reinnervation process progresses. Appropriate electrical muscle stimulation paradigms can reverse these changes and maintain normal muscle fiber typing.^53-58

Reinnervation Is Not Impaired With Electrical Stimulation

In animal models of partial nerve injury, Tam et al ¹⁴ demonstrated motor unit enlargement by sprouting following extensive partial denervation, with the denervated muscle endplates being reinnervated by collateral sprouts with nodal sprouts being the most prevalent. Daily chronic treadmill activity significantly reduced sprouting, and in turn motor unit enlargement leading the authors to suggest a more moderate approach to rehabilitation following nerve injury. ¹⁴ Chronic electrical activation of partially denervated muscles reduced collateral sprouting, consistent with the effects of increased neuromuscular activity.^13,33 After complete rather than partial muscle denervation, specifically after nerve crush injury, the regenerative process resulted in partial reinnervation of a denervated muscle with evidence of extensive collateral sprouting. ¹³ Our current findings of only ~6% of endplates being reinnervated by terminal sprouts after a nerve transection injury and surgical repair do not concur with these findings. This small percentage of sprouts was reduced significantly by muscle electrical stimulation as in other studies.^13,15,33 Reinnervation by sprouts, either terminal or nodal, occurred in 12% in nonstimulated muscle and 9% in stimulated muscle, similar to what Gorio et al ⁵⁹ found in reinnervated extensor digitorum longus muscle. The small contribution of terminal sprouting to overall reinnervation is unlikely to have any clinical significance on functional outcomes since most (>60%) of the regenerating axons reinnervated the muscle in a normal fashion.

Conclusions

Recovery following peripheral nerve injuries is suboptimal. In our study, we assessed a clinically translatable electrical muscle stimulation paradigm and investigated its effects on reinnervation following nerve injury and repair. Brief intermittent electrical stimulation of muscle enhances functional recovery. In future studies, this paradigm may enhance recovery in humans following peripheral nerve injuries.