Short-term adherence to migraine prophylaxis is poor, with about one-half of patients stopping their medication within a year, according to the results of a systematic review (Hepp et al. J Manag Care Pharm 2014;20:22-33).

The study reviewed 33 prospective and retrospective observational studies on adherence and persistence to three of the most commonly used migraine prophylactic drugs: propranolol, amitriptyline and topiramate. At the time of the literature search, propranolol and amitriptyline but not topiramate were recommended first-line agents for migraine prophylaxis in guidelines produced by the U.S. Headache Consortium of medical societies (Ramadan et al. Free full text at http://tools.aan.com/professionals/practice/pdfs/gl0090.pdf).

Data were analysed for the rates of adherence and persistence. Adherence referred to how well a patient took a medication as prescribed (timing, dose, frequency), as determined by pill counts, the medication possession ratio (MPR), or other method. Persistence was the amount of time that a patient remained on a medication, as determined by trial discontinuation rates, patient questionnaires, or other method.

In observational studies, the rate of adherence ranged from 41-95% at two months, 21-80% at six months, and 35-56% at 12 months.

The rate of persistence ranged from 41-88% at two months, 19-79% at six months, and 7-55% at 12 months. For randomized controlled trial data, the pooled persistence rates at 12-26 weeks were 77% for propranolol, 55% for amitriptyline, and 57% for topiramate. The most common reason for treatment discontinuation was adverse effects.

A previous study of patient preferences reported that adverse effects commonly associated with discontinuation of migraine prophylactic drugs were cognitive effects (topiramate), somnolence (propranolol, amitriptyline) and weight gain (divalproex) (Kowacs et al. Headache 2009;49:1022-1027).

Comment
Dr. Daniel Selchen: This interesting analysis of multiple studies suggests that both adherence and persistence are poor with migraine prophylactic agents, particularly beyond six months. What it does not do, however, is sort out to what extent the adherence and persistence issues relate to efficacy failure as opposed to other factors. Not surprisingly, the short-term adherence rates appear to be much better for a beta-blocker then for a tricyclic or an anticonvulsant – which is almost certainly related to side effects.