The association between gastrointestinal symptoms and headache is frequently unrecognized. The aim of the present study was to determine the prevalence of migraine in dyspeptic outpatients referred for upper gastrointestinal endoscopy. Patients aged 18 - 55 years undergoing upper gastrointestinal endoscopy for dyspeptic symptoms in three endoscopic units were recruited consecutively. All of the patients were given a validated questionnaire on headache symptoms in order to determine the prevalence of migraine. Patients were divided into four groups (ulcer-like dyspepsia, reflux-like dyspepsia, dysmotility-like dyspepsia, only nausea and/or vomiting). Age-matched blood donors were given the same questionnaires and served as controls. A total of 378 patients (mean age 40 +/- 11, 52 % men) and 310 controls (mean age 39 +/- 11, 56 % men) were enrolled. No differences were observed between the two groups in the prevalence of migraine (15 % vs. 11 %; P = 0.12). A higher prevalence of migraine was found among women in both groups (P < 0.006). In patients with reflux-like and ulcer-like dyspepsia, the prevalence of migraine did not differ from that in the control individuals (8 % and 7 %, respectively), whereas a higher prevalence of migraine was noted in patients with dysmotility-like dyspepsia (23 %; P < 0.02 vs. controls, those with ulcer-like dyspepsia and those with reflux-like dyspepsia) and in patients with nausea and/or vomiting alone (53 %; P < 0.002 vs. all other groups). The multivariate analysis confirmed that the symptom pattern and sex were the only variables independently associated with migraine. A diagnosis of migraine should be considered in young patients referred for upper gastrointestinal endoscopy due to nausea and/or vomiting or for dysmotility-like dyspepsia.