Objective: To compare placebo responses in neuropathic pain syndromes.

Design: Systematic literature review and meta-analysis.

Setting and patients: Randomized placebo-controlled trials assessing pain intensity or pain relief in any neuropathic pain syndrome published since 1995 with ≥5days follow-up.

Interventions: Placebo response.

Outcome measures: Pain intensity and responder rates (proportion reporting ≥50% pain relief). Meta-regression models were built.

Results: Ninety-four studies (N=5,317) were included in the pain intensity analysis; 47 studies (N=3,087) were included in the responder analysis. After controlling for potential confounders (e.g., subject characteristics, study design characteristics), the placebo response was found to be large and varied with the pain syndrome. Compared with diabetic neuropathic/polyneuropathic pain (DPN), the placebo response for a decline in pain intensity and responder rate was smaller in trials that assessed central pain and postherpetic neuralgia (PHN) and larger in trials that assessed HIV pain. The model-predicted mean decrease (95% confidence interval [CI]) from baseline in pain intensity (0-10 scale) was as follows: DPN, 1.45 (1.35 to 1.55); PHN, 1.16 (1.03 to 1.29); central pain, 0.44 (-0.41 to 1.30); HIV pain, 1.82 (1.51 to 2.12). The predicted responder rates (95% CI) were as follows: DPN, 20% (14.6 to 25.8); PHN, 11.5% (8.4 to 14.5); central pain, 7.2% (2.1 to 12.3); HIV pain, 42.8% (34.9 to 50.7). The type of treatment in the active arm also influenced the placebo response.

Conclusions: Placebo response is influenced by the pain syndrome evaluated. These differences should be considered when evaluating novel compounds for the treatment of neuropathic pain conditions.