Amitriptyline versus gabapentin for the management of neuropathic pain: a systematic review and Bayesian meta-analysis
Abstract
Background: Amitriptyline and gabapentin are first-line drugs for neuropathic pain, yet a systematic head-to-head comparison between them is lacking. This meta-analysis synthesizes randomized controlled trials with direct comparison between these drugs, evaluating analgesic efficacy and safety profile. Objective: To compare, using a Bayesian statistical model, potential differences in treatment outcomes between Amitriptyline and Gabapentin in patients with neuropathic pain. Methods: Bayesian meta-analysis of randomized controlled trials (RCTs) comparing Gabapentin (intervention) and Amitriptyline (comparator) for the management of neuropathic pain. Searches were conducted in PubMed, Embase, and Cochrane databases with no language or date restrictions, in March 2026. Outcomes assessed were pain reduction, measured by the standardized mean difference (SMD; Hedges' g) given the heterogeneity of scales used, and the occurrence of any adverse events, measured by Risk Ratio (RR). A random-effects model was implemented in R (brms/Stan packages) with weakly informative priors: Normal(0,1) for the mean effect and Half-Normal(0; 0.5) for heterogeneity (tau). Convergence was assessed using R-hat diagnostics, and sensitivity analyses with vaguer priors were performed to verify the robustness of conclusions. Risk of bias was assessed using the RoB 2.0 tool. Results: Four RCTs were included in the statistical analysis, comprising 141 patients. Regarding pain reduction, no statistically significant difference was found between treatments across the different scales used (SMD = 0.063; 95% CrI: [−0.476, 0.589]), with only a 38% probability of Gabapentin achieving greater pain reduction than Amitriptyline (SMD < 0). The robustness of this finding was supported by the similarity between the original and vaguer prior estimates (SMD = 0.046; 95% CrI: [−0.761, 0.810]), with moderate heterogeneity (tau = 0.404). Regarding adverse events, no clinically relevant difference was found between treatments (RR = 0.743; 95% CrI: [0.347, 1.509]), a finding corroborated by the consistency with the vaguer prior estimate (RR = 0.653; 95% CrI: [0.201, 1.619]) and moderate heterogeneity (tau = 0.465). Nevertheless, the Bayesian probability of interest indicated an 81.4% chance that Gabapentin is associated with fewer adverse events than Amitriptyline (RR < 1). Risk of bias assessment using RoB 2.0 classified one study as high risk, two as having some concerns, and one as low risk. Conclusion: It is not possible to establish the superiority of either therapy for neuropathic pain management, as both present a similar profile regarding pain reduction and adverse event rates. Additional well-designed RCTs are needed to strengthen the certainty of evidence for these findings.