Behavioral and Physical Treatments for Migraine Headache

Abstract

Objectives To identify and summarize evidence from controlled trials on the efficacy of behavioral and physical treatments for migraine. Search strategy A strategy combining the MeSH term “headache” (exploded) and a previously published strategy for identifying randomized controlled trials was used on the January 1966 to December 1996 MEDLINE database. Other computerized bibliographic databases, textbooks, and experts were also utilized. Selection criteria English-language controlled trials involving patients with migraine in which at least one treatment offered was a behavioral or physical treatment were selected. Data collection and analysis Measures of headache index and headache frequency reported as group means (and standard deviations) were used to calculate standardized mean differences (or effect sizes). Where similar trials provided data, meta-analysis of efficacy measures was performed. The number of patients obtaining at least a 50% reduction in headache index, frequency, or severity was recorded and used to calculate odds ratios. Main results Behavioral treatments for migraine have a consistent body of research indicating efficacy. Summary effect sizes from a meta-analysis of 18 trials suggest that relaxation training, thermal biofeedback combined with relaxation training, electromyographic (EMG) biofeedback, and cognitive-behavioral therapy are all modestly effective in treating migraine when compared to a wait-list control. Thermal biofeedback alone or combined with cognitive-behavioral therapy yielded similar effect sizes that failed to reach statistical significance. Physical treatments have been less often studied. Six small trials of acupuncture yielded mixed results. Other physical treatments for which controlled trials have been reported include transcutaneous electrical nerve stimulation (TENS) (2 trials), cervical mobilization and manipulation (1 trial), occlusal adjustment (1 trial), and hyperbaric oxygen (1 trial). Conclusions Each of the behavioral therapies considered has modest efficacy for migraine. There is little information about which patients will benefit from particular behavioral approaches; the choice among them may, for the present, depend more on availability and acceptability than on data about efficacy. There are insufficient data about any of the physical treatments to draw conclusions about their efficacy.