OBJECTIVES: Randomized controlled trials of prophylactic treatments for migraine focus on the effects in an ideal situation and underestimate the impact of non-adherence, which in this study was examined in a natural setting.

MATERIALS AND METHODS: A sample of 174 adult migraineurs with a current prescription of pharmacological prophylaxis were consecutively recruited at a specialist clinic. Logistic regression analysis was performed to analyse the association between adherence (self-reported with the Medication Adherence Report Scale) and number of years with migraine, frequency of attacks, number of days with migraine per month, attack duration, presence of cardinal features, mean intensity of pain, and recovery between attacks.

RESULTS: One third (35%) were non-adherent. Neither demographic characteristics nor any of the disease specific variables were significantly associated with adherence.

CONCLUSION: Characteristics of the disease per se did not predict non-adherence which was also observed among patients with severe migraine. The full benefit of drugs cannot be realized at currently achieved levels of adherence.