Background. The differential diagnosis of early dementia of the Alzheimer's type from depression in the elderly is often made difficult by the presence of significant memory impairment in depressed patients. The Delayed Word Recall test (DWR) was developed to facilitate the early diagnosis of Alzheimer's disease. The DWR involves: (a) repeated elaborate encoding of ten separate words; (b) a filled delay; (c) delayed free recall. A recognition memory test has also been recently developed. The available evidence suggests impressive sensitivity and specificity when the DWR has been used to separate patients with early Alzheimer's disease from very well matched controls.

Methods. In the present study, the DWR was evaluated with regard to its ability to separate a group of 50 patients with early Alzheimer's disease from 50 elderly patients with major depression in a between-subjects experimental design.

Results. For both free recall and recognition indices, the between-group overlap was large. Using recommended cut-off scores for the detection of Alzheimer's disease, 44% of the depressed patients would have been misclassified as demented based on their free recall scores, and 48% of the depressed patients would have been misclassified on the basis of their recognition scores.

Conclusion. We conclude that the DWR is not specific enough to clearly distinguish patients with early Alzheimer's disease from elderly patients with major depression.