Making sense of phantom limb pain

Abstract

Phantom limb pain (PLP) impacts the majority of individuals who undergo limb amputation. The PLP experience is highly heterogenous in its quality, intensity, frequency and severity. This heterogeneity, combined with the low prevalence of amputation in the general population, has made it difficult to accumulate reliable data on PLP. Consequently, we lack consensus on PLP mechanisms, as well as effective treatment options. However, the wealth of new PLP research, over the past decade, provides a unique opportunity to re-evaluate some of the core assumptions underlying what we know about PLP and the rationale behind PLP treatments. The goal of this review is to help generate consensus in the field on how best to research PLP, from phenomenology to treatment. We highlight conceptual and methodological challenges in studying PLP, which have hindered progress on the topic and spawned disagreement in the field, and offer potential solutions to overcome these challenges. Our hope is that a constructive evaluation of the foundational knowledge underlying PLP research practices will enable more informed decisions when testing the efficacy of existing interventions and will guide the development of the next generation of PLP treatments.

Keywords: BRAIN MAPPING; COGNITION; MOTOR CONTROL; PAIN; SOMATISATION DISORDER.

Figure 1

Phantom limb sensations. (A) Diagram of upper limb amputee with a phantom arm. Spectrum of phantom sensations from non-painful (B) to painful (C). (D) Residual limb pain (RLP) and PLP illustrated in an upper limb amputee. RLP is perceived to originate from the stump/residual limb, whereas PLP is pain perceived to originate from the phantom limb. (E) Cross-sectional surveys on PLP frequency (data adapted from figure 2 in Diers et al with permission) and (F) average intensity (data adapted from figure 2 in Diers et al with permission). Illustrations and remade figures are original.

Figure 2

Factors positively associated with phantom limb pain (PLP). Most common factors associated with PLP identified in the meta-analysis by Limakatso et al. In selecting these factors, we chose to report those that met the following two criteria: reported in at least two studies and have a positive association that was moderate to very strong in strength. Data adapted from Table 3 in Limakatso et al., 2020 with permission.

Figure 3

Time course of phantom limb pain (PLP) and residual limb pain (RLP). To characterise the time course of postamputation pain, we identified all longitudinal, peer-reviewed studies that report prevalence rates and average pain intensity values for PLP and RLP at multiple time points. In the top row, based on PLP data from prospective longitudinal studies, the number of individuals who report PLP and the intensity scores remains relatively constant with time. In the bottom row, based on RLP data from the longitudinal surveys, there is a large decrease in the number of individuals reporting RLP and in the intensity of RLP experienced in the first 6 months after amputation. Black dots represent individual timepoints. Colours depict individual studies; n values report the amputee sample sizes of each study. Patient drop-off across timepoints is not reflected in the starting sample sizes listed after references. The data used to generate these plots and a description for how the data were acquired are available at OSF (https://osf.io/3u8b4/).

Figure 4

Therapies and treatments for phantom limb pain (PLP). (A) A recent survey of 37 PLP clinicians, 23 pain physicians, 11 neurosurgeons, 2 anaesthetists and 1 rehabilitation physician, across 30 UK hospitals, revealed that most clinicians are primarily using pharmacological treatments for chronic PLP patients, as well as a combination of non-pharmacological and brain stimulation therapies. Data adapted from Table 27 in Corbett et al . (B) A global Delphi review of PLP treatments used by PLP experts, three anaesthetists, three physiatrists, two psychologists, two neurologists, eight physiotherapists, one nurse and one occupational therapist, revealed that most consensus treatments were non-pharmacological. Graded motor imagery (GMI); transcutaneous electrical nerve stimulation (TENS); transcranial magnetic stimulation (TMS). Data adapted from Table 2 in Limakatso and Parker with permission.

Figure 5

Recent studies of mirror therapy (MT) for PLP show a lack of efficacy when compared with various control therapies. (A) Randomised control trial (RCT) comparing MT to control treatments (combined data from three amputees assigned to a covered MT treatment and three amputees assigned to perform mental visualisation exercises). Data adapted from figure 2 in Finn et al with permission. (B) RCT comparing MT to phantom exercises. Data adapted from figure 2 in Anaforoğlu Külünkoğlu et al with permission. (C) RCT comparing MT to tactile therapy and a combination therapy treatment which included a serial combination of mirror and tactile therapy. Y-axis reflects decease in magnitude of VAS pain intensity before and after 4 weeks of therapy. Data adapted from table 2 in Ol et al with permission. (D) A randomised, blinded, sham-controlled, 2×2 factorial clinical trial comparing transcranial direct current stimulation (tDCS) and MT. Participants were placed into 1 of the 4 treatment groups: (1) active tDCS and active MT, (2) active tDCS and covered MT, (3) sham tDCS and active MT and (4) sham tDCS and covered MT. Data adapted from figure 3 in Gundez et al with permission. PLP, phantom limb pain; VAS, Visual Analogue Scale. (A, B) Scale is 0–100 mm. (C, D) Scale is 0–10 mm.

Figure 6

Challenges, limitations and recommendations for phantom limb pain (PLP) research. The methodological challenges with studying PLP have hindered the accumulation of reliable information on PLP. For each of these, we provide specific examples for how they might impact the existing literature and recommendations that could be adopted to help overcome each challenge. CONSORT, Consolidated Standards of Reporting Trials; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; STROBE, Strengthening the Reporting of Observational Studies in Epidemiology.