SUMMARY: The purpose of this retrospective, follow-up study was to characterise the use of antidepressant medication in a defined migraine population and evaluate the determinants thereof. Data was obtained from the PHARMO-RLS prescription databaseOur migraine population (2,517 people) included patients having commenced specific migraine drugs, ergotamine or sumatriptan, for the first time from January 1 1992 to December 31 1998. The corresponding date was termed the ‘index date’. Non-migraine patients, those not having used any medication specific for migraine, were selected and equally matched (n=2,517). The cumulative incidence of initiating antidepressant treatment was estimated during two-year observation periods prior to and after the index date. Several demographic and comedication characteristics were assessed as potential determinants of antidepressant drug use within the migraine population. Other determinants included usage patterns (“therapeutic intensity”) of ergotamine and sumatriptan, defined as the absolute number of Defined Daily Doses (DDDs) dispensed per patient during one year prior to initiation of antidepressant therapy. A total of 300 migraine patients (11.9%) and 213 non-migraine patients (8.5%) had initiated antidepressant treatment in the two-year period prior to or in the two-year period after the index date (RR adj 1.4; 95% CI 1.2-1.7). The cumulative incidence of initiation of antidepressant treatment for the migraine population was 3.0% per year prior to and 3.2%per year after the initiation of specific migraine analgesia. The concomitant use of benzodiazepines (RR adj 4.7; 95% CI 3.5-6.3), migraine prophylactic medication (RR adj 2.1; 95% CI 1.6-2.8) and heavy therapeutic intensity use of specific migraine analgesia, defined as >/=150 DDDs per year were highly predictive of antidepressant drug use within the migraine population.In conclusion, compared to the non-migraine population, the initiation of antidepressant treatment was only slightly higher in the migraine population. A number of determinants within the latter were found to be strongly associated with antidepressant drug use, the nature of which most likely reflects an increased severity of migraine whereby therapeutic needs are higher.