This study is the first double-blind placebo-controlled trial of fluoxetine for chronic daily headache (CDH) and migraine. After a one month single-blind baseline on placebo, subjects with CDH (n = 64) and migraine (n = 58) were randomly assigned to a three month trial of fluoxetine (20 mg) or an identical placebo. Fluoxetine and placebo were increased to 40 mg in the second month, depending on patient response. Patients kept daily headache records, and completed 100 mm visual analogue scales (VAS) of headache and mood each month. For the group of CDH patients on fluoxetine, overall headache status (VAS) after three months compared to the end of the single-blind placebo baseline improved a mean of 50% vs. 11% for those receiving the double-blind placebo (P = .029), with 47% vs. 23% improving at least 50% (P = .097, n.s.). Fluoxetine patients showed significant improvement in monthly mood ratings compared to placebo (.001 by the end of the study), and modest but significant improvement in daily records of headache frequency (P = .019) but not pain severity. Significant mood improvements preceded improvement in headache, reaching significance by the end of the second month on fluoxetine (P = .013), while headache improvement emerged only during the third month (P = .001). Double-blind investigator judgement identified more headache improvement in fluoxetine than placebo recipients (40% vs. 22%, P = .032).(ABSTRACT TRUNCATED AT 250 WORDS)