The review included 35 reports of 31 studies (n=727). These included 6 controlled trials (4 placebo-controlled and 2 active-controlled), 15 case series, 7 case reports, 2 open-label trials and 1 retrospective review of charts.
Non-controlled studies (26 patient populations): 314 (79.5%) of the 395 patients experienced meaningful relief with gabapentin.
Controlled studies: the quality assessment on the 6 controlled trials found 2 high-quality and 2 low-quality placebo-controlled randomised controlled trials. Of the 2 active-controlled trials, one was high quality and one was low quality. In the placebo-controlled trials (n=4), the number of patients reporting moderate or excellent pain relief showed statistical homogeneity (chi-squared 7.06, d.f.=3) and resulted in a relative benefit of 2.5 (95% CI: 1.9, 3.4).
The visual analog scale (VAS) results (n=2) gave a weighted VAS mean difference of -11.1 mm (95% CI: -13.2, -11.1; chi-squared 3.56, d.f.=1). The short form of the McGill Pain Questionnaire (SF-MPQ) (n=2) reported a weighted final SF-MPQ mean difference of -5.89 (95% CI: -6.20, -5.59; chi-squared 0.13, d.f.=1). The patients' global impression of change results (n=2) gave a relative benefit of 2.44 (95% CI: 1.8, 3.31; chi-squared 4.76, d.f.=1). The clinicians' global impression of change results (n=2) gave a relative benefit of 2.65 (95% CI: 1.86, 3.78; chi-squared 0.78, d.f.=1). The SF-36 Quality of Life questionnaire was measured (n=2), but the results were not pooled. Both studies reported an improvement in the SF-36 for the gabapentin group in comparison with placebo.
There were no differences between data analysed separately for 'high- certainty' versus 'low-certainty' answers for the questions pertaining to (1) the diagnosis of neuropathic pain, and (2) measurable outcomes. The sensitivity analysis showed that the results obtained for the number of patients reporting moderate or excellent pain relief were consistent and independent of the methodological quality of the included studies.
Side-effects (n=29): fewer and less severe side-effects were reported in the uncontrolled studies.