The dexamethasone suppression test (DST) was performed on 18 patients with presenile and senile dementia of the Alzheimer's type. Objective cognitive testing showed that 13 patients were mildly to moderately impaired and five were moderately to severely impaired. The Hamilton Depression Scale yielded normal results in all patients. The DST results were abnormal in only one of the mildly impaired patients but in four of the five moderately impaired patients. These data suggest that the DST may be a useful clinical tool in mildly impaired patients with Alzheimer's disease but is likely to be confounded by disease in moderately to severely impaired patients.