Targeted Muscle Reinnervation: A Brief History of a Promising Procedure for Effective Management of Amputation Pain

Each day, 411 people in the United States undergo a lower extremity amputation, accounting for 150 000 cases per year.¹ More than 2 million people in the United States are living without some or all of a lower limb.² The most common causes of amputation include uncontrolled diabetes, peripheral vascular disease, neuropathy, and trauma.^1,3 Up to 27.5% of patients who undergo major limb amputation experience at least 1 major postoperative complication; these complications include infection, need for revision amputation, and debilitating postoperative pain.⁴ 

Postoperative pain following lower extremity amputation is classified into 2 broad categories: PLP and RLP. PLP refers to pain experienced by the patient in the limb that has been removed. Although the exact cause of PLP has yet to be determined, it is believed to be caused by “aberrant peripheral nerve changes in addition to central nervous system cortical reorganizations that begin immediately from the time of amputation.”³ PLP is debilitating and often is exacerbated by prosthesis use, decreasing ambulation, and increasing morbidity.⁵  

The second type of amputee pain is RLP, that is, the pain in the remaining stump due to soft tissue inflammation, soft tissue infection, osteomyelitis, poor soft tissue coverage, or sensory neuroma.³ Sensory neuromas are believed to be a large contributor to RLP. During limb amputation, sensory nerves supplying the distal amputation site are severed, and the tissues they innervate are removed. Following surgery, these sensory nerves attempt to regenerate. However, after amputation the distal target of the nerve is no longer available. This leads to Wallerian degeneration of distal nerve ends as well as aberrant and uncontrolled axonal sprouting of the proximal nerve end, which can result in neuroma formation.^6,7 Traditional surgical options for neuroma removal include traction neurectomy, neuroma excision with burying into tissue (bone or muscle), neuroma excision with repair, neurolysis with coverage, and chemical or thermal ablation.^2,8 In addition, more recent studies have revealed potential benefits of local anesthetic therapies for the prevention and management of posttraumatic neuropathic pain.⁹ Poppler et al⁸ conducted a comparative meta-analysis of the efficacy of various surgical interventions for relief from painful neuromas. They reported that 77% of patients who underwent surgical intervention for neuroma experienced meaningful relief; however, there was no statistically significant difference based on the procedure used. Notably, that meta-analysis did not include TMR, a surgical technique that has been found to be effective in reducing postamputation pain. While efforts have been made to decrease the frequency of neuroma formation, including proximal nerve stump implantation or capping, no technique exists that is a proven method for neuroma prevention.¹⁰ 

TMR was initially developed to improve prosthesis control. Traditional prosthetic extremities rely on myoelectric signals from a single agonist/antagonist muscle pair, commonly limiting extremity movement to 1 joint. As prostheses continue to develop, increased myoelectric signals are required to operate multiple joints more intuitively in a single prosthesis. 

In the early 2000s, Gregory A. Dumanian, MD, introduced TMR as a novel procedure to increase myoelectric signals for an upper extremity prosthesis.¹¹ Following amputation, peripheral sensory nerves lose their distal target; however, the proximal end remains viable and able to generate a myoelectric signal. After identifying nearby motor branches to either multiheaded or segmentally innervated muscles, a single motor branch is severed, and the remaining sensory nerve is coapted to the distal motor branch. Thus, 2 myoelectric signals are generated where only 1 existed before. With proper documentation of nerve coaptations and sufficient spacing of signals during the procedure, prosthetists detect these myosignals through electromyographic technology and subsequently design each prosthesis accordingly.¹¹

Neurophysiology

Following the incidental discovery that TMR reduced postamputation pain, researchers theorized that TMR provided a transected sensory nerve “somewhere to go and something to do” as opposed to disorganized growth without a distal target.¹² Proof of concept was first performed in a rabbit model that involved primary upper extremity amputation, followed by use of a rectus abdominis rotational flap and TMR in a secondary procedure.¹³ Median, radial, and ulnar nerves were coapted to motor nerves of the rectus abdominis. At 10-week follow-up after sensory nerve coaptation to a motor nerve branch, these mixed sensory/motor nerves appeared morphologically similar to a motor nerve branch with decreased myelinated fiber counts and increased fascicle diameter, explaining the reduced incidence of neuroma formation following coaptation.¹³ 

Interestingly, the target muscle also transforms physiologically following TMR to more closely resemble the original muscle.¹¹ Chen et al¹⁴ sought to determine if TMR also affected the motor homunculus. Following high-density electroencephalography in 3 patients with upper limb amputations, quantitative evidence demonstrated re-mapping of motor function closer to original locations on the motor homunculus, thus reversing maladaptive cortical reorganization.^11,12,14 

While TMR was originally performed for increased upper extremity prosthesis control, the field has advanced; currently, the procedure is performed on any extremity at all levels of standard amputation. TMR can be performed either at the time of amputation or as a secondary procedure following the development of RLP and PLP. 

Following intubation, the patient should remain under general anesthesia without paralysis to allow donor motor nerve stimulation. Although not necessary, a tourniquet can be used. Tourniquet time should be limited to 30 to 60 minutes to prevent dampening of nerve signals. ^5,6,11,15 Preoperative considerations include maintaining stump length for prosthetic device control as well as adequate soft tissue coverage. If soft tissue coverage is insufficient, a myocutaneous flap may be necessary.¹¹ Chang et al³ determined that when TMR is performed in conjunction with the initial amputation, surgical time increased by a mean of 34.8 minutes ± 7.5 standard deviation. 

The TMR procedure can be simplified to 3 major steps: identification of major nerves in the anatomic region, identification of a motor nerve usually within the same fascial compartment, and neurorrhaphy. Examples of donor sensory nerves and their recipient muscle belly target are summarized in the Table.¹⁶ To assist with identification of a motor nerve, a handheld nerve stimulator is used to confirm the motor nerve end points in target muscles while limiting excess dissection. Once identified, end-to-end neurorrhaphy is performed using interrupted 7-0 or 8-0 nylon epineurial sutures. As shown in the Figure, size mismatch does not significantly affect the outcome of TMR; when present, surrounding muscle tissue can be wrapped around the coaptation with absorbable suture. ^7,12 An alternative means of wrapping the nerve coaptation includes the use of a nerve conduit, which can guide axonal regeneration.¹⁷

To date, a single RCT evaluating TMR has been published, consisting of 28 patients with major limb amputation who were prospectively enrolled in a surgical trial at Northwestern Memorial Hospital (Chicago, IL) and Walter Reed National Military Medical Center (Bethesda, MD).¹² Inclusion criteria for that study were as follows: amputation above the wrist or ankle, age 18 years or older, and no prior neuroma treatment for pain after amputation. Three patients underwent a transhumeral amputation, 1 patient underwent a transradial amputation, 10 patients underwent an above-knee amputation, and 16 underwent a below-knee amputation. A total of 86.7% of patients required amputation secondary to trauma, with the remaining 13.3% requiring amputation secondary to infection. Patients were randomized in the operating room and remained unaware of their procedure for 1 year. Patients in the control group underwent standard neuroma treatment, which included neuroma excision and muscle burying. The primary outcome studied was the change in self-reported preoperative and postoperative 11-point NRS pain score for both RLP and PLP at 1-year follow-up. Secondary outcomes included NRS pain score at the most recent follow-up, PROMIS pain scale score, and patient functionality, reported as OPUS with Rasch conversion score for upper extremity amputation and Neuro-QOL score for lower extremity amputation. 

Using a longitudinal mixed model analysis, Dumanian et al¹² determined that PLP was significantly improved in the TMR group compared with the control group (P = .03). RLP did not differ significantly between groups. However, 3 patients in the control group elected to undergo TMR after standard neuroma treatment failed to provide adequate relief, suggesting the efficacy of TMR over more traditional neuroma treatment modalities. According to PROMIS scale outcomes, 72% of patients who underwent TMR had minimal to no PLP at last follow-up and 40% of patients in the control group had minimal to no PLP postoperatively. No functionality studies were conducted on upper extremity amputations owing to the low representation in this study (n = 4). There was no significant difference in Neuro-QOL scores between treatment groups in patients who underwent lower extremity amputation.¹²

A research team comprising several of the same members who conducted the aforementioned RCT conducted an additional concurrent prospective study that included 33 patients who did not meet inclusion criteria for the RCT owing to prior neuroma surgery or who refused to undergo randomization.¹⁸ In the concurrent study, 58% of patients underwent an upper extremity amputation, while 42% underwent a lower extremity amputation. The etiology of amputation was trauma (91%), infection (6%), and ischemia (3%). The primary and secondary outcomes in this study were identical to those of the RCT. 

Miotin et al¹⁸ reported statistically significant results in NRS pain scores for RLP, with a decrease from a mean of 6.4 ± 2.6 to 3.6 ± 2.2 (mean difference, −2.7 [95% CI, −4.2 to −1.3]; P < .001), as well as for PLP, with a decrease from a mean of 6.0 ± 3.1 to 3.6 ± 2.9 (mean difference, −2.4 [95% CI, −3.8 to −0.9]; P < .001). Additionally, a value of 7 or greater on the NRS scale was deemed severe limb pain; following TMR, only 6% of patients experienced severe RLP and 15% experienced severe PLP. Mean PROMIS scores improved significantly for both RLP (53.4 ± 9.7 preoperatively to 44.4 ± 7.9 postoperatively) and PLP (49.3 ± 10.4 preoperatively to 43.2 ± 9.3 postoperatively) (P < .001). 

Miotin et al¹⁸ also documented significant improvements in functional outcomes, with mean OPUS scores with Rasch analysis increasing from 53.7 ± 3.4 at baseline to 56.4 ± 3.7 1 year postoperatively and mean Neuro-QOL scores increasing from 32.9 ± 1.5 to 35.2 ± 1.6 in the same period. A limitation to consider in this study is self-reporting bias in this patient population; 3 of the 33 patients refused to participate in the RCT owing to their preconceptions of TMR.

While these prospective studies documented the efficacy of TMR, the vast majority of patients in these studies underwent amputation secondary to trauma (83%¹⁸ and 85%¹²). Chang et al³ sought to determine if TMR was equally beneficial in patients with significant comorbidities in a retrospective review of 200 patients who underwent below-knee amputation. Indications for amputation included infection of a diabetic wound, arterial or venous ulcer, or other chronic wound; failed Charcot reconstruction; ischemic pain in the absence of a wound; and infected orthopedic hardware. Amputations in the setting of trauma or cancer resection were included in the study. Comorbidities analyzed were those included in the Charlson Comorbidity Index and included diabetes, peripheral vascular disease, prior myocardial infarction, congestive heart failure, prior cerebrovascular accident, end-stage renal disease requiring dialysis, and smoking status. The control group consisted of 100 patients who underwent traction neurectomy at the time of amputation; the treatment group consisted of 100 patients who underwent TMR at the time of amputation. No statistically significant demographic differences were observed between the 2 groups. Outcome measures included pain, opioid and neuroleptic medication usage, ability to ambulate with a prosthesis, surgical complications, and mortality. 

Significant decreases in pain were noted following TMR, with 71% of patients being pain-free at last follow-up (mean, 9.6 months), compared with 36% in the control group (mean, 18.5 months) (P < .01).³ Additionally, those with pain in the TMR cohort reported a lesser rating than their control group counterparts, with average pain ratings of 3.2 and 5.2, respectively. In the TMR cohort, 14% of patients reported RLP and 19% reported PLP; in contrast, in the control cohort 57% of patients reported RLP and 47% reported PLP (P < .01). Differences in use of opioids and neuroleptic medications were not statistically significant. A significant difference was noted in ability to ambulate with a prosthesis, with 90.9% of patients in the TMR cohort able to ambulate at the most recent follow-up compared with 70.5% in the control group (P < .01). Regarding surgical complications, fewer patients in the TMR group than in the control group developed wound infections requiring debridement and more proximal amputation (P = .02). The difference in 12-month mortality rate between the 2 groups was not significant.³  

Bowen et al¹⁹ conducted a retrospective record review of 22 patients who underwent TMR following below-knee amputation. At 18-month follow-up, none of these patients developed painful neuromas. Eighteen patients underwent a primary TMR procedure, defined as TMR at the time of amputation, and PLP was assessed in this cohort. An abrupt decline in PLP was noted in the primary TMR cohort, with 72% reporting PLP after 1 month, 19% reporting PLP at 3 months, and 13% reporting PLP at 6 months. The authors of that study attributed this marked decrease in PLP over 3 to 6 months to a timeline of reinnervation after noting voluntary muscle twitches at 3 months. They also noted that rates of symptomatic neuromas and PLP were significantly improved when compared with institutional controls.

Special considerations

TMR is not limited to major limb amputations. The procedure can be used anywhere a painful neuroma can develop. 

O’Brien et al²⁰ used TMR to treat chronic pain in women following mastectomy, also known as post-mastectomy pain syndrome. Anterior and lateral cutaneous nerves were identified near the sternal border and the midaxillary line, respectively, and were dissected proximally to the external intercostal muscles to gain length. A nerve stimulator was then used to identify nearby redundant motor nerve branches supplying the serratus anterior, pectoralis minor, or intercostal muscles, and cutaneous nerves were coapted to freshly transected motor nerves. In a small cohort of 11 patients, only 4 patients completed the Physical Well Being: Chest Scale of the BREAST-Q survey, which was used to monitor patient outcome. In this cohort, the average score was 77.5 out of 100, as opposed to a mean score of 71 in patients in a different study who did not undergo TMR. Overall, O’Brien et al²⁰ determined that TMR is a safe option that likely improves patient outcomes. 

Neuromas contribute greatly to postoperative pain in patients who require abdominal wall surgery, as well. In a retrospective study of 20 patients (27 nerves) treated surgically for chronic abdominal wall pain, Chappell et al²¹ noted that painful neuromas were most common in intercostal nerves, followed by ilioinguinal, genitofemoral, and iliohypogastric nerves. Eight patients underwent TMR of the ilioinguinal (7 patients) or genitofemoral (1 patient) nerve to a motor branch of the internal oblique muscle. Pain subjectively improved in 6 of these 8 patients with TMR alone. Of the other 2 patients, 1 patient experienced a new incisional pain and the other patient required a peripheral nerve stimulator for pain control.²¹ 

Unfortunately, neuroma formation is not an uncommon occurrence in hand surgery, and special considerations are made to prevent such formation in routine procedures. In the case of digital amputations, 6.6% to 7.8% of patients will develop a symptomatic neuroma. Thus, TMR also can be useful in digital amputations.²² In a case report by Daugherty et al,²³ a single patient reported 75% improvement in symptoms and discontinued use of narcotic pain mediation after TMR. TMR has also been used in headache surgery, with transection of the greater occipital nerve and lesser occipital nerve and subsequent coaptation to motor branches innervating the semispinalis capitis muscle.²⁴ 

Success rates

Success rates of TMR vary significantly in the literature, with minimal attention paid to failed procedures.²⁵ Other surgical techniques, such as regenerative peripheral nerve interface surgery, have been described as a treatment for RLP and PLP, with success rates similar to those of TMR.²⁶

While TMR is a promising technique, the procedure has limitations that should be considered by provider and patient alike. Adequate dissection of the donor and recipient nerves is crucial to success. Surgical time could be much longer for less experienced surgeons or more complicated dissections. TMR does not guarantee reduction in PLP or RLP. Additional RCTs are necessary to determine the best course of treatment in patients in whom TMR is unsuccessful. Finally, further verification of these nerve coaptations would be useful, including the study and quantification of motor evoked potentials and sensory evoked potentials.²⁷

The TMR technique has proved to be beneficial for both prosthesis control and pain reduction. TMR is beneficial in upper extremity amputation, most importantly in glenohumeral or transhumeral amputations, because it provides adequate myoelectric signals for complex upper extremity prosthesis use. It is imperative to establish realistic patient expectations and ensure that patients are committed to working alongside their prosthetist as well as with both physical and occupational therapists.⁸ Arguably more beneficial than traditional neuroma treatment, TMR has been proven to decrease PLP and may decrease RLP by removing an existing neuroma or preventing neuroma formation. Decreased postamputation pain improves quality of life and may result in decreased dependence on pharmacologic pain therapies. 

Data on TMR are limited, and more research is required to confirm the effectiveness of this technique. As more research becomes available, it is possible that TMR could become a mainstay of both neuroma prevention and treatment. While not all surgeons will perform this operation, patients may request information regarding TMR prior to scheduled amputation. Thus, a basic knowledge of the history, procedure, and benefits will facilitate discussion with patients and, it is hoped, will improve patient outcomes.