Seventy-two trials that assessed fibromyalgia (mean number of participants 176) and 70 trials that assessed diabetic peripheral neuropathy (mean number of participants 200) were included in the review. The mean Jadad score for the fibromyalgia trials was 2.9 (standard deviation 1.8) and the mean score for the diabetic trials was 2.8 (standard deviation 1.7).
Fibromyalgia syndrome: Mean end of treatment pain scores were significantly higher than mean baseline pain scores in placebo groups (WMD 7.69, 95% CI 6.10 to 9.29; Ι²=71%) and in the fibromyalgia syndrome active drug group (WMD 17.11, 95% CI 16.41 to 17.90; Ι²=84%) on a pain scale of zero to 100.
Mean pain score was significantly higher in the active drug group (SMD 0.82, 95% CI 0.72 to 0.92; Ι²=77%) compared to fibromyalgia syndrome placebo groups (SMD 0.42, 95% CI 0.35 to 0.49; Ι²=53%).
Painful diabetic neuropathy: Mean end of treatment pain score was significantly higher than mean baseline pain score in painful diabetic neuropathy placebo groups (WMD 13.96, 95% CI 11.93 to 15.99; Ι²=85%) and in active drug groups (WMD 22.54, 95% CI 20.49 to 24.58; Ι²=85%) on a zero to 100 pain scale.
Mean pain score in the active drug group (SMD 1.14, 95% CI 1.02 to 1.25; Ι²=85%) was significantly higher than mean pain score in the painful diabetic peripheral neuropathy placebo groups (SMD 0.71, 95% CI 0.61 to 0.81; Ι²=81%).
No evidence of publication bias was found. Subgroup and sensitivity analyses yielded no significant results.
Meta-regression: Factors associated with pooled weighted mean difference of placebo on pain in both fibromyalgia syndrome and painful diabetic neuropathy trials were pain baseline, incremental year of study initiation and effect of active medication on pain. Pooled estimates of effects differed by number of study sites for fibromyalgia syndrome trials only.