The comorbidity of headache and depression is rarely studied in the elderly. Confounders were seldom controlled in previous studies. From August 1993 to March 1994, we conducted a door-to-door survey to investigate the relationship of headache and depression in a Chinese elderly population (age > or = 65 years old) in two townships of Kinmen, Taiwan. A total of 1421 participants (71%) out of 2003 eligible citizens completed five measurements: a structured headache interview, Geriatric Depression Scale-short form (GDS-S), a survey of chronic medical illness. Cognitive Abilities Screening Instrument and an evaluation of activities of daily living. Headache diagnoses were made according to the criteria of the International Headache Society (IHS), 1988. Depression was defined as a GDS-S score > or = 8. After adjustment for confounding, subjects with more frequent headaches, more severe headaches, diagnoses of IHS migraine or chronic tension-type headaches in the past year, or a lifetime history of any headache including migraine were more likely to be depressed. In addition, the most relevant headache-related predictors of depression were the presence of any reported lifetime headache (odds ratio (OR) = 1.8, P < 0.01) and headache frequency > or = 7 days/month in the past year (OR = 2.0, P = 0.01). This study provided evidence that headache is independently associated with depression in the elderly. A high comorbidity of depression was found in the elderly with IHS migraine or chronic tension-type headaches. Not only the headache profile in the past year but also that in their lifetime was important in predicting current depression in the elderly. 1