We report open-trial antidepressant response in 16 inpatients with mixed symptoms of depression and cognitive impairment, compared to 16 elderly depressives without cognitive impairment. Criteria for adequate treatment specified a steady-state plasma nortriptyline level of 50-150 ng/ml for 4 consecutive weeks or a minimum of six treatments with electroconvulsive therapy. Ten of 16 mixed-symptom patients showed a drop in Hamilton depression ratings greater than or equal to 50% during treatment. Similarly, Blessed dementia ratings declined significantly; the % change in Blessed dementia ratings was significantly correlated with improvement in Hamilton depression ratings. By contrast, Folstein mini-mental state scores did not change significantly during treatment. Six of 16 (37.5%) patients showed resolution of cognitive impairment with adequate treatment of depression. Mixed-symptom patients diagnosed as suffering from major depression (with cognitive impairment) showed more robust pre-post treatment differences, particularly in Hamilton, Folstein, and Blessed dementia scores, than did mixed patients diagnosed as having primary degenerative dementia (with depression). In cognitively intact elderly depressives, the mean % change in Hamilton ratings was 72% (4.3), not significantly different from mean % change in mixed-symptom patients (57.4 +/- 29.9). The proportion of intact depressives showing a reduction greater than or equal to 50% in Hamilton depression ratings was significantly greater (93.8%) than in the mixed group (62.5%). In both groups, 81.3% of patients (13 of 16 in each cell) had a final Hamilton rating less than or equal to 10. These data suggest that elderly patients with mixed depression and cognitive impairment respond to treatment similar to that used in cognitively intact elderly depressives. A controlled study of antidepressant treatment in mixed-symptom patients is warranted.