Emerging neural modulation techniques for the management of phantom limb pain: Evidence from randomized controlled trials

Abstract

Phantom limb pain (PLP), a common sequela of amputation, affects up to 86% of amputees and significantly impairs quality of life. PLP is thought to stem from complex central and peripheral nervous system plasticity. Current treatments, including pharmacological and non-pharmacological approaches, have limited efficacy. Recently, extended reality technologies have emerged as promising tools for PLP management, leveraging immersive sensory input to modulate cortical reorganization. Of note, emerging neural modulation techniques also offer promising alternatives, including peripheral nerve stimulation, repetitive transcranial magnetic stimulation and transcranial direct current stimulation. These approaches demonstrate clinical efficacy in relieving pain, improving functional outcomes and reducing opioid usage. Future research could prioritize large-scale trials to validate the efficacy of nerve stimulation techniques and explore their integration with extended reality technologies for PLP.

Core Tip: Phantom limb pain remains a complex and challenging condition due to its multifactorial neuroplastic origins, with conventional therapies often failing to provide durable relief. Emerging non-pharmacological approaches in recent years include extended reality technologies and neural modulation strategies. The results from randomized controlled trials on peripheral nerve stimulation, repetitive transcranial magnetic stimulation, and transcranial direct current stimulation have demonstrated favorable effects on pain relief and functional improvement and opioid usage reduction. Therefore, these nerve stimulation strategies offer promising alternatives for phantom limb pain management.

TO THE EDITOR

Phantom limb pain (PLP) refers to a chronic or paroxysmal pain perceived in a limb or body part that has been amputated or is partially missing, occurring in individuals with limb loss. It is a common sequela of amputation that significantly impairs quality of life. Globally, approximately 57.7 million individuals live with limb loss[1], with projections expected to double by the year 2050[2]. The incidence of PLP following surgical or traumatic amputations is very high, with 27%-86% of amputees experiencing this phenomenon throughout their lifetime[3,4]. The discomfort associated with PLP is often severe, manifesting as throbbing, stabbing, or electric shock sensations, which often contribute to depression, sleep disturbance, and reduced physical/motor abilities[5]. Despite decades of research, the underlying mechanisms of PLP remain incompletely understood. Current evidence implicates a complex interplay between changes in the central and peripheral nervous systems along the neuroanatomical pathways disrupted by amputation[6].

Current treatments for PLP, either pharmacological or non-pharmacological treatments, are diverse but suboptimal. Pharmacological treatment options, including gabapentin, pregabalin, and opioids, provide partial relief but carry risks of tolerance, addiction, or systemic side effects[7]. Evidence surrounding the use of botulinum toxin and calcitonin has been predominantly inconclusive[8,9]. Numerous non-pharmacological treatments have also been applied in the treatment of PLP. Mirror therapy (MT) is perhaps one of the least expensive and most effective modalities, although some patients do not benefit from it[10]. Extended reality technology, encompassing virtual reality, augmented reality, and mixed reality, offers a promising advancement. A recent study of Gan et al[11] reviewed the research of extended reality technology in PLP treatment by describing the basics of extended reality technology and progress of different extended reality-based treatments. These findings suggest the promising prospect of this technology, and highlight the need for further validation through more long-term research and large-scale clinical trials. Given the limitations of existing behavioral therapies, such as variable response rates and suboptimal efficacy in a significant subset of patients, there is a pressing need to explore alternative or complementary approaches. Against this backdrop, emerging evidence highlights nerve stimulation technology targeting the peripheral nerve, prefrontal cortex, or motor cortex as a complementary strategy. Such stimulation provides input into the amputation zone and thus undoes the organisational changes after amputation. To preliminarily assess the evidence on the efficacy of nerve stimulation, PubMed and EMBASE databases were searched using the keywords ‘phantom limb pain’, ‘peripheral nerve stimulation’, ‘repetitive transcranial magnetic stimulation’, ‘transcranial direct current stimulation’, and ‘randomized controlled trial’ to screen relevant studies, and studies were included if they reported quantitative outcomes on ≥ 10 adult PLP patients. Clinical evidence from randomized controlled trials (RCTs) highlights their ability to reduce pain intensity, improve functional capacity, and decrease opioid reliance.

Electrical stimulation over the peripheral nerve

Peripheral nerve stimulation (PNS) has been reported to decrease pain and opioid requirements following amputation by evoking paresthesias to override pain signals, and inhibiting nociceptive transmission[12-14]. A randomized, double-blind, placebo-controlled trial by Gilmore et al[12] indicated that a significantly greater proportion of subjects receiving PNS (n = 7/12, 58%, P < 0.05) demonstrated ≥ 50% reductions in average postamputation pain during weeks 1-4 compared with subjects receiving placebo (n = 2/14, 14%, Table 1). The longer-term outcomes in the same cohort further demonstrated the sustained relief of chronic pain by a 60-day PNS treatment, with 67% (6/9) of PNS group achieving ≥ 50% reductions in average weekly pain at 12 months vs 0% (0/14) in placebo group at the end of the placebo period (P < 0.001) (Table 1)[13]. Building on this evidence of PNS’s effectiveness in chronic PLP, a small pilot RCT of Albright-Trainer et al[14] explored the feasibility of 60-day PNS to treat acute post-amputation pain. Likewise, the PNS group had significantly greater reductions in average PLP, residual limb pain, and daily opioid consumption than the placebo group (Table 1).

Table 1.

Summary of randomized controlled trials with nerve stimulation in phantom limb pain

Ref.	Design	Population	Group	Intervention	Stimulation location	Control	Outcome
Gilmore et al[12], 2019; Gilmore et al[13], 2019	Multicenter, double-blinded, RCT	Lower extremity amputees (n = 26)	I: n = 12; C: n = 14	8 weeks of PNS	Femoral and sciatic, with needle electrode 0.5-3 cm from nerve trunk	Sham stimulation for 4 weeks, followed by a crossover of additional 4 weeks PNS	Responders with ≥ 50% reductions in average pain: I: 58%, 7/12 (weeks 1-4) vs C: 14%, 2/14 (weeks 1-4); P < 0.05. I: 67%, 8/12 (weeks 5-8) vs C: 14%, 2/14 (weeks 1-4); P < 0.05. I: 67%, 6/9 (12 months) vs C: 0%, 0/14 (end of the placebo period); P < 0.001
Albright-Trainer et al[14], 2022	Single center, open label, RCT	Lower extremity amputees (n = 16)	I: n = 8; C: n = 8	Standard medical therapy in combination with 8 weeks of PNS	PNS leads implanted approximately 1-3 cm distant from the femoral and sciatic nerves	Standard medical therapy alone	Responders with ≥ 50% reductions in average pain: I: 100%, 5/5 vs C: 50%, 4/8 (8 weeks); I: 100%, 5/5 vs C: 86%, 6/7 (3 months). Opioid consumption: I > 60% decrease vs C > 200% increase (the end of week 8)
Kapural et al[16], 2024; Kapural et al[17], 2024	Multicenter, double-blinded, RCT	Unilateral lower-limb amputees (n = 170)	I: n = 85; C: n = 85	HFNB for day 28-365, with a 30-minute session/day	Cuff electrode wrapped around the damaged nerve, and approximately 1 cm from nerve terminus	Sham stimulation with sub-therapeutic ultra-low frequency for day 28-91, followed by a crossover of HFNB for day 91-365	Day 28-91 responders with ≥ 50% reductions in average pain: I: 24.7%, 21/85 vs C: 7.1%, 6/85; P < 0.01 (30 minutes post treatment); I: 48.1%, 37/77 vs C: 22.2%, 18/81; P < 0.001 (120 minutes post treatment). Opioid usage: I: 6.9 MED/day vs C: 3.6 MED/day reduction, not significant. Day 91-365 average NRS pain: By month 12, combined cohort = 2.3 ± 2.2 points (95%CI: 1.7-2.8; P < 0.0001), 30 minutes post treatment; 2.9 ± 2.4 points (95%CI: 2.2-3.6; P < 0.0001), 120 minutes post treatment. Opioid usage: Combined cohort: 6.7 ± 29.0 MED/day reduction from baseline to month 12 (P < 0.05)
Vats et al[19], 2024	Single center, double-blinded, RCT	Trauma amputees (n = 19)	I: n = 10; C: n = 9	10 sessions of rTMS given over 2 weeks	rTMS at the DLPFC contralateral to the amputation site. Surface electrodes on abductor pollicis brevis, ground on wrist	Sham stimulation	VAS: I: 6.50 (8.00-5.25) at baseline to 0.00 (0.75-0.00, P < 0.0001) at the end of the therapy, 0.00 (1.00-0.00, P < 0.001) at 15 days post treatment, 1.00 (2.00-0.00, P < 0.01) at 30-days post treatment, 0.50 (1.75-0.00, P < 0.01) at 60 days post treatment. C: No significant difference
Kikkert et al[22], 2019	Single center, double-blinded, RCT	Unilateral upper-limb amputees (n = 15)	I: n = 15; C: n = 15	4 consecutive tDCS sessions spaced at least 1 week apart	Anodal over S1/M1 missing hand cortex, cathodal over contralateral supraorbital area, sham electrodes on intact hand S1/M1 and supraorbital area	Sham stimulation	Percentage change of PLP ratings: I: -6.1, immediately after tDCS; I: -20.3, end of experimental session. C: +42.9, immediately after tDCS; C: +28.3, end of experimental session
Gunduz et al[23], 2021	Multicenter, double-blinded, 2 × 2 factorial, RCT	Unilateral traumatic lower limb amputees (n = 112)	Active tDCS/active MT: n = 29, sham tDCS/active MT: n = 28, active tDCS/covered MT: n = 28, sham tDCS/covered MT: n = 27	20 minutes tDCS stimulation a daily session for 10 days	The anodal electrode was placed over the M1 contralateral to the amputation side and the cathodal over the contralateral supraorbital area	Sham stimulation	VAS: No interaction between tDCS and MT groups (F = 1.90, NS). In the adjusted models, there was a main effect of active tDCS compared to sham tDCS (beta coefficient = -0.99, P < 0.05) on phantom pain. The overall effect size was 1.19 (95%CI: 0.90-1.47)

RCT: Randomized controlled trial; I: Intervention; C: Control; PNS: Peripheral nerve stimulation; MED: Morphine equivalent dose; NRS: Numerical rating scale; rTMS: Repetitive transcranial magnetic stimulation; DLPFC: Dorsolateral prefrontal cortex; VAS: Visual analog scale; tDCS: Transcranial direct current stimulation; S1: Primary somatosensory cortex; M1: Primary motor cortex; PLP: Phantom limb pain; MT: Mirror therapy; NS: Not significant.

High-frequency nerve block (HFNB) mimics local anesthetics through inhibiting voltage-gated sodium channels to block pain signal transmission[15]. A multicenter, double-blinded RCT named QUEST enrolled 180 unilateral lower-limb amputees to assess the efficacy and safety of peripheral HFNB for PLP treatment[16,17]. In this trial, 170 subjects were implanted and randomized to 3 months of HFNB or sham treatment. The primary endpoint showed 24.7% of the HFNB group achieved ≥ 50% pain relief at 30 minutes vs 7.1% in controls (P < 0.01), with responder rates rising to 46.8% vs 22.2% at 120 minutes (P < 0.001) (Table 1)[16]. HFNB treatment also significantly reduced average worst end-of-day pain by 22% (7.6 to 6.0) vs 12% (7.7 to 6.7) in controls (P < 0.05), and mean end-of-day pain by 32% (6.1 to 4.2) vs 12% (7.7 to 6.7) and 17% (5.9 to 4.9) in controls (P < 0.01) at 3 months, indicating lasting pain profile improvement. In addition, the HFNB group showed a trend toward reduced opioid use and comparable adverse event rates. Following the initial phase, sham-treated subjects crossed over to 12 months of on-demand HFNB[17]. By month 12, average pain scores dropped by 2.3 (30 minutes) and 2.9 (120 minutes) points (P < 0.0001), weekly pain days decreased by 3.5 (P < 0.001), daily opioid use fell by 6.7 morphine equivalents (P < 0.05), and quality of life improved by 2.7 points (P < 0.001) (Table 1). Device-related serious adverse events occurred in 8%, confirming HFNB’s sustained safety for chronic PLP. Collectively, QUEST demonstrated that HFNB provided sustained, on-demand relief of chronic PLP, reduced opioid dependency, and enhanced functional outcomes with acceptable safety.

Repetitive transcranial magnetic stimulation over the dorsolateral prefrontal cortex

The dorsolateral prefrontal cortex (DLPFC) plays a central role in pain processing and modulation, and repetitive transcranial magnetic stimulation (rTMS) of the DLPFC has been shown to induce synaptic plasticity, thereby providing sustained pain relief for chronic pain conditions of neuropathic origin[18]. The study of Vats et al[19] assessed the effect of DLPFC-targeting rTMS on the pain status of PLP. This RCT enrolled 19 traumatic amputees with PLP, randomizing them to real (n = 10) or sham (n = 9) rTMS groups. The real rTMS group received 10 sessions over 2 weeks, while the sham rTMS group underwent identical protocol but without active stimulation. The real rTMS group showed a significant reduction of visual analog scale score from baseline (6.50 ± 1.51) to 0.00 (0.75-0.00, P < 0.001) at treatment completion, with sustained effects at 60 days [0.50 (1.75-0.00), P < 0.01] (Table 1). In contrast, the sham group showed no significant visual analog scale changes. No adverse effects were reported in either group. These findings demonstrated that low-frequency rTMS targeting the DLPFC provided significant and sustained PLP relief for up to 60 days, supporting its potential as a noninvasive therapy for chronic PLP.

Transcranial direct current stimulation over the motor cortex

Maladaptive plasticity in the primary somatosensory (S1) and motor cortex (M1) is associated with sensory deafferentation following an amputation, and thus one of the contributors for excessive pain[20]. Targeting the M1 to modulate the dysfunctional sensorimotor circuits offers a new potential approach for PLP treatment. Transcranial direct current stimulation (tDCS), one of noninvasive brain stimulation techniques, has been used for direct M1 stimulation. It is thought to alleviate neuropathic pain through changing motor cortex thalamic connectivity, with thalamic pathway activation counteracting sensory afferent loss to enhance inhibitory pain networks[21]. Kikkert et al[22] conducted a within-participants, double-blind, and sham-controlled trial to evaluate the PLP relief via task-concurrent tDCS over the S1/M1 missing hand cortex. Seventeen unilateral upper-limb amputees received 20 minutes of anodal tDCS over the primary sensorimotor cortex (S1/M1) contralateral to the amputation while performing phantom hand movements, with sham tDCS applied over the intact hand’s cortex. Functional magnetic resonance imaging revealed that a single tDCS session significantly reduced PLP by 29.5% (Table 1), with effects lasting ≥ 1 week, and PLP relief was correlated with decreased S1/M1 activity post-stimulation. Another multicenter, randomized 2 × 2 factorial trial enrolled 112 traumatic lower-limb amputees to assess the effects of combined and alone tDCS and MT in PLP[23]. Participants were randomized to active or sham tDCS over contralateral M1, combined with active MT (mirrored movements) or covered MT (imagined movements). The primary outcome was PLP changes on the visual analogue scale at the end of interventions (4 weeks). Active tDCS alone reduced PLP significantly (beta coefficient = -0.99, P = 0.04, effect size = 1.36), whereas no interaction was found between tDCS and MT groups (F = 1.90, P = 0.13) (Table 1). tDCS was associated with increased intracortical inhibition (coefficient = 0.96, P = 0.02) and facilitation (coefficient = 2.03, P = 0.03) as well as a posterolateral shift of the center of gravity in the affected hemisphere. MT induced no motor cortex plasticity changes. The trial confirmed tDCS as an effective PLP therapy via M1 plasticity, with no synergistic benefit from MT.

Conclusion and prospect

PLP remains challenging due to its complex neuroplastic basis. Recent RCTs have suggested that neural modulation techniques (PNS, rTMS, and tDCS) could provide PLP reduction, function improvement, or opioid use reduction. Notably, current evidence is constrained by very small sample sizes, short follow-up periods, heterogeneity in outcome measures, and unclear mechanisms driving sustained pain relief. To strengthen the evidence base, large-scale, long-term trials are imperative to rigorously assess the efficacy and safety of these neural modulation approaches. Additionally, comparative effectiveness research between these techniques is also needed. While the integration of neural modulation with extended reality technologies represents a theoretically promising avenue to address both neural signaling and cortical reorganization in PLP, this hypothesis requires empirical testing in well-designed studies before definitive conclusions can be drawn.