Ten RCTs were included (n=2,902 patients). All were double-blinded. Quality scores ranged from 12 to 15.
Topiramate for migraine (six RCTs, n=1,723 patients): Topiramate was significantly more likely than placebo to cause the following adverse effects: paraesthesia (risk ratio of 5.8 for 50 mg, risk ratio of 8.4 for 100 mg, risk ratio of 8.5 for 200 mg); fatigue (risk ratio of 1.7 for 200 mg); nausea (risk ratio of 1.7 for 200 mg); weight loss (risk ratio of 6.1 for 100 mg, risk ratio of 6.8 for 200 mg); somnolence (risk ratio of 3.1 for 200 mg); taste alteration (risk ratio of 11.3 for 50 mg, risk ratio of 4.7 for 100 mg, risk ratio of 9 for 200 mg); decreased appetite (risk ratio for 100 mg of 2.4, risk ratio for 200 mg of 2.3); concentration difficulty (risk ratio of 4.4 for 200 mg); and memory impairment (risk ratio of 2.8 for 100 mg, risk ratio of 4.9 for 200 mg). Drop-out rates due to adverse effects were 17% for 50 mg, 21% for 100 mg and 29% for 200 mg. The numbers needed to harm were 6 (95% confidence interval (CI): 4.68 to 8.25) for 50 mg, 4.8 (95% CI: 4.04 to 5.83) for 100 mg and 3.5 (95% CI: 3.08 to 4.05) for 200 mg.
Topiramate for epilepsy (four RCTs, n=1,179 patients): Drop-out rates due to adverse effects were 7% for 50 mg, 17% for 100 mg and 28% for 200mg. The numbers needed to harm were 15 (95% CI: 10.79 to 24.38) for 50 mg, 5.9 (95% CI: 4.56 to 8.15) for 100 mg and 3.6 (95% CI: 2.95 to 4.67) for 200 mg.
Topiramate for migraine versus topiramate for epilepsy: Cognitive complaints and taste alteration were found only in migraine patients. Behavioural adverse drug reactions, upper respiratory tract infections and headache were reported only by epilepsy patients. When data were aggregated, paraesthesia was significantly more common in the migraine trials at all three doses (risk ratio of 2.5 (95% CI 1.66 to 3.77) for 50 mg; risk ratio of 2.7 (95% CI: 1.8 to 3.97) for 100mg; risk ratio of 3.0 (95% CI: 1.95 to 4.56) for 200 mg). This finding was confirmed by the secondary analysis. The overall numbers needed to harm for the 50 mg dose were 2.5 times higher for epilepsy than for migraine (95% CI: 2.03 to 2.98). Other outcomes did not differ significantly between the groups when both methods of analysis were taken into account.