Recent reports supporting the use of lithium carbonate as an adjunct to tricyclic antidepressants for the treatment of refractory depression have not utilized standardized tricyclic antidepressant therapy, nor have they addressed the efficacy of lithium augmentation in a geriatric population. A 3-week open trial was added to the medication regimen of 15 elderly depressed inpatients who had already failed 4 weeks of therapeutic levels of nortriptyline. Treatment response was determined by the 17-item Hamilton Rating Scale for Depression (HAM-D). Two of 15 partial responders before lithium augmentation became complete responders. Of the remaining 13 "nonresponders" before lithium augmentation, one had a complete response, 7 had a partial response and 5 remained nonresponders. Although there was a mean HAM-D change of 8.3 points after lithium augmentation (24.7 +/- 5.9 to 16.4 +/- 6.8, p less than .001), when considering the previously reported similar efficiency of extended nonaugmented nortriptyline, these data do not strongly support lithium augmentation in elderly subjects who fail to respond after 4 weeks of nortriptyline. Further study is needed to determine what role, if any, lithium augmentation should play in the treatment of geriatric depression.