A multicentre double blind, randomised, placebo controlled 7-week study evaluated the efficacy and safety of gabapentin 1800 or 2400 mg/day in treating postherpetic neuralgia. Three hundred and thirty-four men and women aged at least 18 years (mean 73) received gabapentin 1800 or 2400 mg daily or placebo in three divided doses with a forced titration schedule. The primary outcome measure was change in average daily pain diary score (baseline week v final week). Secondary outcomes included mean weekly sleep interference score; Short Form-McGill Pain Questionnaire (SF-MPQ); Clinician and Patient Global Impression of Change (CGIC/PGIC); Short Form-36 Health Survey (SF-36). From week 1, pain scores showed a significantly greater improvement with gabapentin: the final difference v baseline was -34.5% for the 1800 mg dose, -34.4% for the 2400 mg dose compared with -15.7% for the placebo group. The difference vs. placebo was 18.8% for the 1800 mg dose (95% confidence interval 10.9-26.8%; P<0.01) and 18.7% for the 2400 mg dose (10.7-26.7%; P<0.01). Sleep interference diaries showed a similar pattern. There were significant differences in favour of gabapentin for number of patients reporting >50% reduction in their pain intensity, in the CGIC and PGIC, in the sensory and total scores of the SF-MPQ (both doses), in the visual analogue scale of pain of the SF-MPQ (2400 mg only) and in the vitality, bodily pain and mental health domains of the SF-36. Overall gabapentin was well tolerated. The most common adverse events were dizziness and somnolence, particularly during the titration phase. Thus, this study confirms the role of gabapentin as an efficacious and well-tolerated treatment for postherpetic neuralgia.