Forty-four RCTs (n=3,399) were included. Sample sizes ranged from 16 to 375 (mean 77). Mean dropout rate was 18% (range zero to 52%). Fifteen studies were considered to be at low risk of bias.
Antidepressants: Eighteen studies compared antidepressants to placebo and 81% of these comparisons showed no statistically significant differences between trial arms on headache. Two of the four studies that had a low risk of bias and adequate power and patient selection showed statistically significant benefits for antidepressants on headache improvement, but not frequency. Twelve studies compared different types or dosages of antidepressants; all except two showed no statistically significant differences between treatment groups and neither of these trials was considered to be high quality. Amitriptyline showed no statistically significant differences from placebo in four out of five trials. Comparisons of amitriptyline with other medications or non-pharmacological treatments showed no statistically significant differences with the sole exception of a benefit in reduced headache index compared to biofeedback.
Muscle relaxants: Three studies found no statistically significant differences between muscle relaxants and placebo. One study showed no difference between different doses of the same agent.
Benzodiazepines: Two of three studies that compared benzodiazepines provided data and neither showed a statistically significant benefit. One very small study found a benefit of biofeedback over benzodiazepines at follow-up.
Vasodilator agents: There was conflicting evidence on the efficacy of vasodilators versus placebo and limited evidence on their negative impact on depression compared with placebo or biofeedback.
5-HT receptor agonists: One study showed no statistically significant differences between buspirone and amitriptyline.
Limited results were reported for other types of drugs.