Prevalence of postamputation pain and its subtypes: a meta-analysis with meta-regression

The inconsistent use of standardized approaches for classifying postamputation pain (PAP) has been a barrier to establishing its prevalence. The primary objective of this systematic review and meta-analysis is to determine the prevalence of nontraumatic lower-extremity PAP using an established taxonomy. The secondary objective is to determine the prevalence of PAP subtypes, including phantom limb pain and residual limb pain (RLP). An a priori protocol was registered, and a database search was conducted by a reference librarian. Randomized trials and uncontrolled studies were eligible for inclusion. The risk of bias was assessed using a tool developed for uncontrolled studies. A total of 2679 studies were screened, and 13 studies met inclusion criteria (n = 1063). The sources of risk of bias included selection bias and, to a lesser extent, whether the outcome was adequately ascertained. Two studies reported the prevalence of PAP and the pooled prevalence was 61% (95% confidence interval [CI], 33%-86%) with high heterogeneity (I = 93%). Thirteen studies reported the prevalence of phantom limb pain and the pooled prevalence was 53% (95% CI, 40%-66%) with high heterogeneity (I = 93%). Eight studies reported the prevalence of RLP and the pooled prevalence was 32% (95% CI 24%-41%) with high heterogeneity (I = 76%). Clinical subtypes of RLP were not reported. The prevalence of PAP is high in patients with nontraumatic lower-extremity amputations. Ongoing research that uses a taxonomy for PAP is needed to fully delineate the prevalence of PAP subtypes.