Regenerative Peripheral Nerve Interface Surgery to Treat Chronic Postamputation Pain: A Prospective Study in Major Lower Limb Amputation Patients

Abstract

Objective:

The objective was to assess the postsurgical outcomes of regenerative peripheral nerve interface (RPNI) surgery in a prospective cohort of major lower extremity amputation patients with chronic postamputation pain.

Background:

Chronic pain in lower limb amputation patients is commonly the result of neuroma formation after traumatic peripheral nerve injury. By implanting more proximal transected nerve ends into autologous free muscle grafts, RPNI surgery can treat postamputation pain by diminishing the development of neuromas. RPNI surgery in prior retrospective studies has been shown to mitigate postamputation pain.

Methods:

Twenty-two lower limb amputation patients with established chronic postamputation pain were recruited from 2 studies in this prospective study. All patients underwent RPNI surgery to treat identified symptomatic neuromas within the residual limb. Patient-reported outcome instruments were administered preoperatively and postoperatively at 1 week, 4 months, and 12 months to examine residual limb pain (McGill Pain Questionnaire, PROMIS Pain Intensity, and PROMIS Pain Interference), phantom limb pain (modified PROMIS Pain Intensity and Phantom Limb sensation questionnaire), psychosocial status (PHQ-9, GAD-7, and PCS), and functional (OPUS) outcomes.

Results:

RPNI surgery significantly improved residual limb pain. While phantom limb sensation improved significantly, phantom limb pain demonstrated a modest decrease. Psychosocial outcomes also improved significantly after RPNI surgery. Prosthetic use slightly increased, and patients did not experience loss of function.

Conclusions:

RPNI surgery leverages the processes of reinnervation to successfully treat residual limb pain and improve psychosocial outcomes in patients with chronic postamputation pain. Phantom limb pain may be more difficult to treat in chronic pain patients who have central sensitization at the time of surgery.

INTRODUCTION

Limb loss affects over 1.6 million people in the United States.¹ In addition to the functional limitations of missing a limb, the amputation population is often plagued by another sequela that prevents optimal rehabilitation.² Among them, pain is one of the top contributors to a decreased quality of life.³ Nearly all major limb amputation patients experience postamputation pain, and about 30% to 50% patients develop chronic postamputation pain.^4,5

Phantom limb pain (PLP) and residual limb pain (RLP) are 2 main types of chronic pain in amputation patients.⁵ PLP describes the perception of pain felt in the limb that is no longer physically present.⁶ In contrast, RLP describes pain that is contained within the residual limb and is often related to neuroma formation.⁵ Neuromas occur as a result of traumatic injuries when transected peripheral nerves attempt to regenerate and reinnervate end organ targets.⁷ However, as the distal limb is not present following an amputation, a neuroma will inevitably form from the regenerating axons, which aimlessly sprout and elongate without end organs to reinnervate.⁷ At least 1 in 7 patients will develop symptomatic neuromas following a below-knee amputation,⁸ but this incidence is likely underappreciated. Over time, increased nociceptive activity from neuromas can contribute to central sensitization, PLP, and other chronic pain pathologies.⁹

Regenerative peripheral nerve interface (RPNI) surgery is a promising method for treating posttraumatic neuropathic pain.^10–12 By suturing a denervated skeletal muscle graft to the end of a transected nerve, RPNI surgery provides a physiologic target for nerve regeneration. The formation of new neuromuscular junctions within the muscle graft reduces the number of aimless axons present in the distal end of the regenerating peripheral nerve, limiting the pathologic mechanism of neuroma formation. Previous work has successfully demonstrated the formation of new neuromuscular junctions through immunohistologic imaging of the RPNI.¹³ Animal models and retrospective clinical studies provided early evidence of its efficacy for treating chronic neuroma pain and PLP.^12,14 Other studies have examined the use of the RPNI as a prosthetic interface and have proven that these new neuromuscular junctions are physiologically active, allowing the transduction of efferent motor signals to a robotic hand and afferent signals back to the central nervous system (CNS).^13,15,16

This report presents the outcomes of a prospective clinical investigation using RPNI surgery for the treatment of chronic neuropathic pain in lower extremity amputation patients. The study utilizes various validated patient-reported outcome measures assessing for pain, psychosocial, and functional domains to evaluate the results of RPNI surgery.

METHODS

This study (HUM00120915) was approved by the University of Michigan and the Veteran’s Affair Institutional Review Board and was conducted in accordance with the ethical standards of the University of Michigan Committee on Human Experimentation.

Study Cohort

Patients were prospectively recruited from Michigan Medicine and the Veteran Affairs Ann Arbor Healthcare System during the enrollment period of 2018 to 2022. Potential study patients were identified through screening clinic schedules via the electronic health system records or directly from each participating surgeon’s office. Only adult patients (≥18 years old) with a history of unilateral below-knee-amputation or above-knee-amputation and documented symptomatic neuromas were eligible for this study. Patients were excluded if their amputation occurred within 6 months of enrollment as the short time frame would not allow for clear establishment of chronic postamputation pain and rehabilitation. Other exclusion criteria included previous nerve-modulating surgery, previous RPNI surgery, bilateral amputations, and segmental injuries. Eligible study candidates were directly recruited from the clinic setting. Study enrollment was halted during the COVID-19 pandemic due to institutionally mandated restrictions on clinical trial recruitment protocols. A power analysis was performed to determine adequate sample sizes to compare study groups using an alpha level of 0.05 with 80% statistical power.

A chart review of the electronic health system record was performed to collect clinical demographics and medical history for each enrolled patient. All study participants were then seen in clinic 1 to 4 weeks before their operation for the administration of preoperative survey instruments designed to collect baseline patient-reported outcome (PRO) data on RLP, PLP, psychosocial factors, and functional outcomes. A thorough physical examination and preoperative ultrasound imaging were performed to fully document neuroma pain quality and location. Patients then proceeded with RPNI surgery and were followed postoperatively for up to 1 year. Surgical variables, postoperative complications, and postoperative prosthetic use data were also collected. Repeat PRO surveys were administered postoperatively at 4- and 12-month time points (Supplemental Material 1, see http://links.lww.com/AOSO/A445).

RPNI Surgery

This technique has been previously described in the literature.¹⁷ Preoperatively, symptomatic neuromas were identified based on patient history, examination (eg, Tinel’s sign), and ultrasound. An incision was made either along the previous incision of the index amputation, or a new longitudinal incision was made away from the previous incision. Following identification and exposure of each neuroma, the distal bulb was sharply excised, and the nerve was cut back to length using a bread-loafing technique until healthy axonal architecture was visualized (Fig. 1). Autologous free skeletal muscle grafts were then harvested from donor muscle, commonly from the vastus lateralis in the proximal thigh through a separate incision. Care was taken to remove all connective tissue from the free skeletal muscle grafts to optimize muscle regeneration and reinnervation. Each free skeletal muscle graft was approximately 3 cm in length, 2 cm in width, and 0.5 cm in thickness. The proximal transected nerve end was placed within the middle portion of the free muscle graft and secured by suturing the epineurium to the muscle tissue with 6-0 nonabsorbable sutures. The free muscle graft was then wrapped circumferentially to fully encompass the nerve end and secured with suture. For larger nerves (eg, tibial, sciatic), intraneural dissection was performed to separate the nerve into component fascicles before performing RPNI surgery to each fascicle (Fig. 2). This is done to maximize the number of denervated muscle fibers presented to the regenerating axons. A deep soft tissue pocket was bluntly made where the RPNIs were then placed in a protected location away from the distal residual limb before closure. An individualized perioperative pain control plan, including the decision to perform either a preoperative or postoperative nerve block, was developed for each patient by the anesthesia service in conjunction with the surgical service. As each patient had a unique presentation of chronic postamputation pain, this allowed our study team to evaluate the effectiveness of RPNI surgery in the context of each patient’s individualized pain management plan.

FIGURE 1.

Excision of symptomatic neuroma. A, Symptomatic sciatic nerve neuroma on the end of a previously transected nerve. B, Neuroma excision is planned proximal to the bulb. C, Neuroma bulb is excised using a bread-loafing technique until healthy axonal architecture is visualized.

FIGURE 2.

RPNI surgery on sciatic nerve. A, Intraneural dissection to divide the sciatic nerve into 3 component fascicles in preparation for RPNI surgery. Separation of larger nerve into components optimizes the innervation ratio (denervated muscle fibers to regenerating axons). B, Each nerve is placed in the middle portion of the free muscle graft and secured by suturing the epineurium to muscle tissue. C, The muscle graft is wrapped around the end of the nerve and secured with monofilament suture. D, Final result with 3 separate RPNIs. The 3 RPNIs will be tucked into a deep pocket surrounded by vascularized soft tissue. Contact between RPNIs has not significantly diminished RPNI reinnervation in the authors’ experience.

Patient Reported Outcomes

Nine PRO surveys were used to quantify and track preoperative and postoperative RLP, PLP, psychosocial, and functional outcomes.

To quantify RLP, both the short-form McGill Pain Questionnaire (SF-MPQ) and the Patient-Reported Outcome Measurement Information System (PROMIS) were used. The SF-MPQ asks patients to rate the intensity of 15 different pain descriptors (eg, “throbbing” and “shooting”) on a 0 to 10 scale experienced in the past week.¹⁸ Through the use of these descriptors, the SF-MPQ parses out distinct constructs of pain.¹⁸ The SF-MPQ was chosen because of its ability to report pain through different nociceptive subscales and is thought to be more sensitive in quantifying changes in pain, especially if an intervention is thought to affect one aspect of pain more than others.¹⁸

PROMIS is a standardized set of item banks developed through the National Institutes of Health using massive datasets of self-reported patient data.¹⁹ Specifically, the PROMIS Pain Intensity, Pain Interference, and Neuropathic Pain Quality scales were developed for assessing pain and are commonly used in the amputation population.^20,21 The pain intensity scale asks patients to report current pain intensity, pain intensity on average, and pain intensity at the worst, whereas pain interference assesses how much pain interferes with different aspects of a patient’s life. Similar to the SF-MPQ, PROMIS Neuropathic Pain Quality scale asks patients to rate 5 specific pain descriptions that are commonly associated with neuropathic pain.²⁰ The incorporation of the PROMIS pain scales assesses the consequence of RLP on daily life and compares the pain intensity scores to the SF-MPQ.

PLP was assessed using a modified version of the PROMIS Pain Intensity scale where “phantom pain” was specifically used in the typical PROMIS questionnaire (Supplemental Material 2, see http://links.lww.com/AOSO/A446). The modified PROMIS Pain Intensity scale for PLP is used commonly in the amputation population to gauge the intensity of PLP and separates PLP from RLP.²¹ In addition to PLP, this study also sought to quantify phantom sensations in this patient population, which is defined as nonpainful sensation in the phantom limb.²² Given the lack of established PRO surveys to assess phantom limb sensation, the study team designed several ad hoc questions to measure the frequency and intensity of these sensations (Supplemental Material 2, see http://links.lww.com/AOSO/A446).

It was theorized that RPNI surgery and its ability to reduce chronic postamputation pain would likely change the presentation of mood disorders. For this purpose, the Patient Health Questionnaire-9 (PHQ-9) and Generalized Anxiety and Depression-7 (GAD-7) surveys were included to evaluate depression and anxiety symptom severities, respectively.^23,24 The PHQ-9 and GAD-7 assess symptoms over a 2-week period with higher values representing more severe symptoms.^23–26

In addition, psychosocial factors relate to pain in the amputation population through the concept of catastrophizing. Catastrophizing is defined as a thinking pattern that emphasizes the negative perspective of a situation.²⁷ It consists of 3 dimensions (rumination, magnification, and helplessness) of negative cognitive processes and is evaluated by the Pain Catastrophizing Scale (PCS).²⁷ Prior research has correlated PCS scores with pain intensity and disability.^28,29 Thus, the PCS was implemented in this study to measure the effect of RPNI surgery on this critical psychosocial factor.

Finally, the Orthotic and Prosthetic User’s Survey (OPUS) lower extremity function survey was included to assess the impact of RPNI surgery on functional outcomes.³⁰ During RPNI surgery, the free skeletal muscle grafts are typically harvested from the ipsilateral thigh, which then becomes a site for acute surgical pain. In patients with chronic postamputation pain, the experience of this new acute pain from the donor site can be quite uncomfortable, especially after central sensitization. Therefore, OPUS was included to assess if prosthetic use declined following surgery due to donor site morbidity after muscle graft harvest. Furthermore, OPUS outcomes can also assess whether improvements in pain would lead to improvements in prosthetic use after RPNI surgery.

Statistical Analysis

For all PROs, means and standard deviations were reported by assessment times. Summaries of net changes and percent changes from baseline were also reported at each follow-up time by subtracting individual postoperative values from preoperative values. Since chronic pain is defined as pain lasting longer than 3 months, improvement of chronic pain following RPNI surgery can be expected and measured after 3 months. Cross-sectional means for each outcome measure were graphically assessed over time, and longitudinal outcome data were modeled using a linear mixed-model. The model used outcomes assessed at baseline, 4 month, and 12 month as the response variable with indicators for months 4 and 12 as predictors. For each outcome measure, if no significant difference in changes from baseline was found at 4 and 12 months (as determined by no significant difference [P > 0.05] between the coefficients of the 2 time indicators), the model was refit with an indicator for the combined period of 4 and 12 months. The coefficient of the indicator for the combined 4- and 12-month period was reported as the time-averaged expected postoperative change in outcome following RPNI surgery.³¹ If the difference in means between 4 and 12 months was significant, then the coefficient of the 12-month indicators was reported as the expected postoperative change from baseline at 12 months.

RESULTS

A total of 22 amputation patients from the University of Michigan (86%) and VA (14%) with chronic pain were included. Baseline clinical characteristics are presented in Tables 1 and 2. The mean time from initial amputation was 13.5±10 years with trauma being the most common cause. About 77% had previously failed prior treatment for chronic pain, including nerve block injection (27%) and nerve stimulation (18%). Previous surgical interventions include excision of heterotopic ossification (14%) and neuroma excision alone (23%). Twenty-three percent of patients had either undergone no previous treatments or only nonsurgical therapies for pain. Seventy-three percent of patients underwent prosthetic rehabilitation before RPNI surgery.

TABLE 1.

Baseline Demographic and Clinical Characteristics of Patients with Painful Neuromas after Major Limb Amputation

Demographic and Clinical Characteristics	N = 22
Age, mean (SD, min–max)	53.3 (13.7, 27–76)
Male, n (%)	12 (54.55%)
Race, n (%)
Other	0 (0%)
Black or African American	0 (0%)
White	22 (100%)
BMI, mean (SD, min–max)	30.0 (7.7, 20–47.7)
Smoking status, n (%)
Never smoked	6 (27.3%)
Current smoker	3 (13.6%)
Previous smoker	13 (59.1%)
Education, n (%)
High school or some high school	6 (27.3%)
Some college, trade, or university	11 (50.0%)
≥College, trade, or university degree	5 (22.7%)
Hispanic or Latino*, n (%)	0 (0%)
Diabetes uncomplicated, n (%)	1 (4.55%)
Diabetes complicated, n (%)	1 (4.55%)
Elixhauser comorbidity index, mean (SD)	2.4 (1.9)
Site, n (%)
University of Michigan	19 (86.36%)
VA	3 (13.64%)
Marital status, n (%)
Married or living with significant other	12 (54.6%)
Widowed, separated, or divorced	5 (22.7%)
Single, never married	5 (22.7%)

^*One person is missing ethnicity data.

TABLE 2.

Baseline Clinical Characteristics of Patients with Painful Neuromas after Major Limb Amputation

Clinical Characteristics	N = 22
Cause of amputation, n (%)
Trauma	18 (81.8%)
Metastatic cancer	0 (0%)
Solid tumor cancer	2 (9.09%)
Peripheral vascular disorders	2 (9.09%)
Location of amputation of interest
Left below knee	9 (40.9%)
Right below knee	6 (27.3%)
Left above knee	3 (13.6%)
Right above knee	4 (18.2%)
Through knee amputation	0 (0%)
Previous pain treatment, n (%)
Nerve block injections	6 (27.3%)
Nerve stimulation (peripheral or spinal cord)	4 (18.2%)
Surgical removal of HO	3 (13.6%)
Neuroma excision	5 (22.7%)
No treatment or behavioral modification	5 (22.7%)
Uses orthotic (brace) or prosthetic device*, n (%)	16 (72.7%)

^*Based on self-reported 0 days per week and 0 hours per day use.

Surgical Variables

Surgical variables and complications are presented in Table 3. In total, 21 patients received 72 RPNIs. One patient was missing data only on the number of RPNI used in the surgery. Complications occurred in 7 patients, with delayed wound healing (18%) being the most common. Only 1 patient required reoperation due to unmasking of another neuroma after RPNI surgery.

TABLE 3.

Surgical Variables and Complications

Variables
No. of RPNIs, mean (SD, min–max)	3.4 (1.1, 2–6)
Location of RPNIs, n (% of patient)	Total = 21 patients
Sciatic	5 (23.8%)
Common peroneal	3 (14.3%)
Tibial	14 (66.7%)
Deep peroneal	13 (61.9%)
Superficial peroneal	14 (66.7%)
Saphenous	3 (14.3%)
Sural	7 (33.3%)
Total number of RPNIs	72
Complications, n (% of patients)	Total = 22 patients
Hematoma	2 (9.1%)
Delayed wound healing	4 (18.2%)
Recurrent delayed wound healing	2 (9.1%)
Postop wound infection	3 (13.6%)
Oral antibiotic treatment	2 (9.1%)
IV antibiotic treatment	1 (4.5%)
Wound infection requiring surgical intervention	0 (0%)
Seroma	2 (9.1%)
Reoperation	1 (4.5%)

Residual Limb Pain

Figure 3 depicts average SF-MPQ scores over time. Pain scores remained stable after 4-month postop, and no significant differences between the 4- and 12-month scores were found. Table 4 shows SF-MPQ and PROMIS pain scores at each assessment time. The last column in Table 4 represents the mean change from baseline in the nonacute postoperative time frame, averaged over 4 and 12 months, along with the corresponding 95% confidence interval. After RPNI surgery, patients reported significantly lower pain scores in all domains.

FIGURE 3.

Changes in McGill Pain Questionnaire response over time.

TABLE 4.

Pain Patient Reported Outcomes (PROs) at Baseline and at Post-RPNI Follow-Up Times in Patients with Painful Neuromas after Major Limb Amputation (N = 22)

PROs	Baseline	Time Since Enrollment			Change*	% Change
		1 Week	4 Months	12 Months
Residual limb pain
McGill pain questionnaire
Total Pain	4.6 (2.2)	3.9 (2.2)	3.2 (2.7)	3.6 (2.0)	−1.3*	28%‡
(range: 0–10)	Change†	0.8 (1.6)	1.4 (2.2)	1.4 (1.8)	(−2.1, −0.6)
Continuous pain score	5.1 (2.7)	4.8 (2.7)	3.4 (2.7)	4.1 (2.1)	−1.4*	27%‡
(range: 0–10)	Change†	0.4 (2.0)	1.6 (2.7)	1.4 (2.3)	(−2.4, −0.5)
Intermittent pain score	5.6 (2.4)	4.6 (2.6)	4.0 (3.2)	4.1 (2.2)	−1.7*	30%‡
(range: 0–10)	Change†	1.0 (2.1)	1.7 (2.7)	1.9 (2.3)	(−2.7, −0.8)
Neuropathic pain score	4.2 (2.2)	3.0 (1.6)	3.1 (2.7)	3.3 (1.9)	−1.0*	24%‡
(range: 0–10)	Change†	1.2 (1.7)	1.1 (2.0)	0.9 (1.9)	(−1.8, −0.2)
Affective pain score	3.1 (3.0)	2.7 (2.8)	2.1 (3.0)	2.2 (3.0)	−1.2*	39%‡
(range: 0–10)	Change†	0.6 (3.0)	1.1 (3.0)	1.9 (2.5)	(−2.2, −0.2)
PROMIS pain score
Intensity	10.6 (2.0)	10.5 (2.4)	8.6 (2.8)	9.8 (2.4)	−1.7*	16%‡
(range: 3–15)	Change†	0.4 (2.3)	2.0 (2.3)	1.4 (2.1)	(−2.6, −0.7)
Interference	23.7 (5.4)	NA	17.8 (6.6)	21.4 (7.6)	−4.6*	19%‡
(range: 6–30)	Change†	NA	6.0 (7.3)	3.6 (7.3)	(−7.5, −1.7)
Neuropathic quality	16.9 (4.0)	15.7 (4.3)	13.8 (5.7)	15.7 (4.7)	−2.3*	14%‡
(range: 5–25)	Change†	0.9 (4.5)	3.1 (6.1)	0.7 (5.4)	(−4.5, −0.1)
Phantom limb
Modified PROMIS pain score (for phantom limb pain)
Intensity	6.3 (2.7)	4.8 (3.3)	4.5 (3.0)	6.0 (3.0)	−0.2§	8%
(range: 3–15)	Change†	1.5 (2.1)	1.8 (2.7)	0.3 (3.3)	(−0.8, 0.5)
Phantom limb sensation (ad hoc)
Total	2.9 (1.0)	1.9 (1.2)	2.0 (1.0)	2.4 (1.0)	−0.5§	17%‡
(range: 0–4)	Change†	1.1 (0.8)	0.8 (1.0)	0.5 (0.5)	(−0.8, −0.2)

All cell values are mean (SD). Numbers of patients providing the data were 22, 19, 21, and 17 at baseline, week 1, month 4, and month 12, respectively.

^*Expected mean change from baseline averaged over months 4 and 12 and the corresponding 95% confidence interval in parentheses. Estimates are based on mixed model fit with data at baseline, month 4 and month 12 as responses, and time-averaged estimates are reported where no significant difference was seen between months 4 and 12.

^†Change from baseline; larger negative values imply greater improvement (eg, pain reduction).

^‡Statistically significant change from baseline at 0.05 level where percent changes are calculated as 100 × [time-averaged change/ baseline mean] or 100 × [change at 12 month/ baseline mean].

^§When outcomes were found to differ at month 4 versus month 12 based on the first model described above,^* mean change at 12 months from baseline at month is reported here.

Phantom Limb Pain

Fifty-five percent of patients reported baseline chronic phantom limb sensation and PLP. Table 4 shows changes in phantom limb sensation and pain scores. At 1-year postoperative follow-up, phantom limb sensation was significantly lower compared with preoperative measurements, whereas PLP was slightly, but not significantly lower than baseline.

Psychosocial Outcomes

Table 5 reports psychosocial outcomes as measured by the PCS, PHQ-9, and GAD-7. Pain catastrophizing, depression, and anxiety scores markedly decreased following RPNI surgery.

TABLE 5.

Psychosocial Patient Reported Outcomes (PROs) at Baseline and at Post-RPNI Follow-Up Times in Patients with Painful Neuromas after Major Limb Amputation (N = 22)

PROs	Baseline	Time Since Enrollment			Change*	% Change
		1 Week	4 Months	12 Months
Psychosocial factors
GAD-6 (anxiety)
Total	8.5 (4.7)	NA	5.3 (4.9)	4.3 (4.0)	−3.5*	41%‡
(range: 0–21)	Change†	NA	3.1 (5.1)	3.5 (3.9)	(−5.2, −1.7)
PHQ-9 (depression)
PHQ-9 Total	10.7 (6.1)	NA	7.5 (6.9)	7.9 (6.5)	−2.6*	24%‡
(range: 0–27)	Change†	NA	2.5 (7.5)	1.6 (5.0)	(−4.9, −0.2)
Pain catastrophizing scale
Total	25.4 (13.5)	15.9 (10.6)	13.9 (14.5)	12.2 (13.0)	−11.7*	46%‡
(range: 0–52)	Change†	9.5 (14.3)	10.8 (17.6)	11.1 (13.8)	(−17.0, −6.3)
Rumination	10.2 (5.0)	6.9 (4.0)	4.7 (5.1)	5.1 (4.8)	−5.2*	51%‡
(range: 0–16)	Change†	4.0 (4.6)	5.2 (6.3)	5 (5.2)	(−7.2, −3.1)
Magnification	4.9 (3.3)	2.9 (2.3)	2.6 (3.1)	1.7 (2.1)	−2.5*	51%‡
(range: 0–12)	Change†	2.0 (4.0)	2.1 (4.2)	2.6 (3.2)	(−3.8, −1.2)
Helplessness	11.5 (6.1)	6.9 (4.6)	6.6 (6.6)	7 (6.8)	−4.4*	38%‡
(range: 0–24)	Change†	5.0 (5.8)	4.7 (7.8)	2.9 (6.2)	(−6.9, −1.9)

All cell values are mean (SD). Numbers of patients providing the data were 22, 19, 21, and 17 at baseline, week 1, month 4, and month 12, respectively.

^*Expected mean change from baseline averaged over months 4 and 12 and the corresponding 95% confidence interval in parentheses. Estimates are based on mixed model fit with data at baseline, month 4, and month 12 as responses, and time-averaged estimates are reported where no significant difference was seen between months 4 and 12.

^†Change from baseline; larger negative values imply greater improvement (eg, pain reduction).

^‡Statistically significant change from baseline at 0.05 level where percent changes are calculated as 100 × (time-averaged change/baseline mean) or 100 × (change at 12 month/baseline mean).

§When outcomes were found to differ at month 4 versus 12 based on the first model described above,

^*mean change at 12 months from baseline at month is reported here.

Functional Outcomes

RPNI surgery did not lead to lasting functional deficits after the free skeletal muscle graft harvest. While postoperative values of OPUS scores were trending higher compared with baseline (Fig. 4), the increases were not significant (P = 0.7 and P = 0.4, respectively). The mean 12-month OPUS score (42.8) was higher compared with the baseline score (36.8) by about 16%, but this was not statistically significant based on the mixed model analysis.

FIGURE 4.

Preoperative and postoperative OPUS lower extremity function score.

DISCUSSION

Following limb amputation, there is a high incidence of both chronic RLP and PLP.⁵ The formation of neuroma and associated increase in nociceptive activity leads to central sensitization.³² In addition, following deafferentation injury, adjacent areas or dermatomes can occupy the deafferented area in the primary somatosensory cortex.³³ One study reported a large correlation between the degree of cortical reorganization and PLP in upper extremity amputees.³⁴ Central sensitization and cortical reorganization after deafferentation are thought to be due to factors related to the development of chronic pain.^9,32,33,35 Furthermore, afferents in a neuroma have dramatically reduced depolarization potentials, leading to more firing in the peripheral nervous system.⁹ The sensitized CNS experiences overall more afferent signaling from the peripheral nervous system due to reduced depolarization potentials, increased spontaneous neural activity, and expanding dorsal horn neuron receptive fields.⁹ Together, these factors make chronic pain more difficult to treat.

Treatment for chronic pain within the amputation population is typically centered on pharmacologic or surgical interventions.^36–38 Pharmacologic treatments including peripheral nerve blocks, ketamine, and opioids, have not shown sufficient long-term efficiency.^36,37 Similarly, traditional surgical interventions such as simple neuroma excision and nerve transposition also ignore the pathophysiology of neuroma formation, leading to high reoperation rates and pain recurrence after surgery.³⁸ Notably, the majority of our patients (77%) had failed prior treatments for chronic postamputation pain. In contrast to previous studies that excluded these patients,³⁹ we felt that it was important to include the treatment-resistant patients as it is a defining attribute of patients with major lower limb loss.

Despite the challenges in this difficult population, RPNI surgery successfully demonstrated efficacy in reducing RLP. In the literature, the concept of minimal clinical important difference (MCID) helps gauge if the statistically significant pain reduction is actually a clinically important reduction that should change current practices.⁴⁰ Pain reduction in this population is particularly hard to achieve compared with a control population, leading to lower differences on pain scales pre- and postintervention.⁴¹ Previous studies have identified an MCID of 15% for pain scales.⁴² In this context, RPNI surgery for chronic postamputation pain results in a noteworthy clinical reduction in the experience in pain and various pain subdomains. These results firmly establish RPNI surgery as a successful intervention for RLP reduction in the treatment-resistant, chronic postamputation population. Furthermore, RPNI surgery should be even more effective in patients with easier-to-treat pain based on these results.

Within the general amputation population, the prevalence of psychiatric conditions is high (32%–84%).⁴³ Specifically, rates of depression and anxiety are reported to range from 21% to 63% and 25% to 57%, respectively.⁴⁴ In addition, the presence of psychosocial factors has been linked to the development and persistence of chronic pain, and patients living with chronic pain also have a high propensity to develop mood disorders.⁴⁵ Our study population consisting of major lower limb amputation patients with chronic pain are at high risk for mood disorders. However, psychosocial outcomes in the chronic amputation pain population are rarely studied and are underappreciated when evaluating interventional outcomes.²¹ Therefore, a key strength of this study is the incorporation of psychosocial outcomes before and after RPNI surgery. It is crucial to study these psychosocial factors with any chronic pain intervention as they contribute to poor outcomes such as increased disability, higher unemployment rates, and higher healthcare utilization.^28,29 Furthermore, psychosocial diagnoses have been associated with persistent pain and poor response to treatment.³⁹ Baseline GAD-7 and PHQ-9 scores in this study population were significantly higher than the general population.^24,46 Baseline average PCS score was also clinically high.⁴⁷ Anxiety, depression, and catastrophizing symptoms all decreased significantly at 1 year postoperatively from RPNI surgery. MCID for anxiety and depression is cited to be about 20% and 41% for PCS^23,47,48 Thus, RPNI surgery not only improves pain outcomes but also the sequelae of pain including associated mood disorders and catastrophizing.

This study also sought to assess PLP and sensation. While RPNI surgery successfully reduced phantom limb sensation, there was only a modest improvement in PLP. As prior studies have demonstrated significant improvements in PLP following RPNI surgery, this finding could represent that this study was underpowered for examining PLP changes.^10,12 This study included 22 patients, whereas prior studies had 45 to 46 patients.^10,12 Nonetheless, chronic PLP patients likely have powerfully centralized pain. Patients in this study had a mean time from amputation of 13.5 years. It may be possible that peripheral nerve interventions such as RPNI surgery are less effective in the setting of years of maladaptive CNS changes. If this is the case, earlier interventions with RPNI surgery, including prophylactic placement at the time of amputation, could be key in preventing these CNS transformations.¹⁰ It is also possible that PLP and phantom sensations have somewhat different underlying mechanisms, and RPNI may be more effective in treating phantom sensations. Flor et al³³ found that persons with PLP demonstrated profound cortical reorganization, whereas persons with nonpainful sensations did not display significant cortical reorganization.

During RPNI surgery, free skeletal muscle grafts are harvested typically from the ipsilateral thigh. This differs from many cases of prophylactic RPNI surgery where muscle can be effectively harvested from the distal amputation part. Thus, acute surgical pain is expected at the proximal thigh donor site. In the lower extremity amputation population that is already experiencing chronic postamputation pain, there is always the potential of impaired functional outcomes as a direct result of this donor site pain. Thus, OPUS was included to assess the effects of RPNI muscle harvest on functional outcomes. Our results demonstrate that functional outcomes did not decline, dispelling the notion that acute pain at the donor site contributes to possible functional loss after RPNI surgery. Interestingly, it was surprising that we did not find a more robust improvement in prosthetic use following surgery, particularly after the clear improvement in pain and psychosocial outcomes. Others have also not found a direct connection between increased functional status and improvement in these domains.²¹ In reality, pain comprises only 1 of the identified factors impacting disability following amputation.⁴⁹ Socket comfort, level of amputation, early rehabilitation, psychological factors, and others have also been identified to impact functional status following amputation.⁵⁰

This study was limited by a relatively low number of patients as the enrollment period was considerably affected by the COVID-19 pandemic research hiatus and resulted in missing patient data at various postoperative time points. Accordingly, as chronic pain improvements are measured after 3 months, the decision to use a mixed model to assess mean change from baseline also helped preserve sample size and statistical comparisons. A large sample size would have increased the power of this study and may have enhanced the analysis of the surgical effects on PLP and prosthetic rehabilitation. Furthermore, another limitation is the use of self-reported outcomes, which are subjected to each individual study participant’s experiences and biases.

CONCLUSION

Chronic postamputation pain is a common sequela of major lower extremity limb loss. RPNI surgery harnesses the physiologic principles of nerve regeneration and muscle reinnervation to mitigate the development of symptomatic neuromas. In this prospective study, RPNI surgery significantly improved chronic pain as measured by various patient-reported outcome instruments. Furthermore, the alleviation of this chronic postamputation pain coincided with improvement in psychosocial outcomes. Despite anticipated acute donor site pain from the harvest of free skeletal muscle grafts, functional status did not diminish after RPNI surgery. Future large-scale prospective studies will elucidate the effect of RPNI surgery in patients with established centralized pain and also determine if successful RPNI surgery facilitates enhanced prosthetic use.

ACKNOWLEDGMENTS

J.C.L. drafted the manuscript and its revisions. C.A.K. and C.S.W.B. contributed to study design and data acquisition. J.B.H. was responsible for data management and cleaning. J.K. crafted many of the tables and graphs in the final manuscript. H.M.K. designed the statistical plan for analyzing the data and crafted many of the tables and graphs included in the final manuscript. R.S.R. was crucial for informing the study design and patient reported outcome metric selection. J.H.K. involved in the study design, implementation, and patient recruitment at the Veteran’s Affairs Hospital. M.J.T. was crucial for informing the study design and patient reported outcome metric selection and involved in the study design, implementation, and patient recruitment at the Veteran’s Affairs Hospital. M.E.G. was crucial for informing the study design and patient reported outcome metric selection and significantly helped with data analysis and interpretation. P.S.C. involved in study conception, design, implementation, and patient recruitment at the University of Michigan and significantly helped with data analysis and interpretation. S.W.P.K. involved in the study conception and design and significantly helped with data analysis and interpretation. T.A.K. conceptualized the study design, guided its implementation, and informed data analysis/interpretation.