Treatment of late life depression : towards a better prognosis

Abstract

Major depression in the elderly is often associated with a poorer prognosis than in younger patients. More recent studies comparing cohorts of older depressed patients with their younger counterparts usually have failed to demonstrate any difference in prognosis, and a recent systematic review of comparative studies, controlling for confounding variables, found remission rates in patients in late life little different from those in midlife. In a systematic review, we included more than 100 double-blind RCTs in elderly depressed patients and conformed the good prognosis found in other reviews in the elderly. However, the majority of these studies are in relatively healthy, mild to moderately depressed outpatients, leaving doubts whether antidepressants are also efficacious in severely depressed inpatients with comorbid physical and mental disorders. In this thesis, 81 severely depressed inpatients, often with comorbid psychiatric and physical disorders (half of them had psychotic features) were included in different treatment studies and followed for a maximum of 3 years. All 81 patients participated in a double blind, randomised controlled trial (RCT) comparing venlafaxine with nortriptyline. These antidepressants did not differ in efficacy, and also both were tolerated well. The percentage of patients achieving remission (defined as a final score of 10 or less on the Montgomery ųberg Depression Rating Scale) was 32.1%, which is comparable with efficacy in studies with adult patients. Patients not responding to one of these antidepressants were asked to participate in a second RCT (n=29), comparing two antidepressants often prescribed to treatment-resistant patients, lithium and phenelzine. Lithium was significantly more effective than phenelzine, and again both were tolerated remarkably well. All patients not responding with one of these antidepressants were asked to participate in an open treatment study. Within 3 years of treatment, 78 of the 81 patients (96.3%) achieved a response (at least 50% reduction in MADRS score) and 68 patients (84%) a complete remission. This is comparable with other sequential treatment strategies in adult patients. We conclude from these studies that treatment in severe depressed elderly inpatients may be as effective as in younger adult patients. Using multivariate survival analysis, remission was predicted only by a lower MADRS score at baseline and a shorter duration of the index depression episode. Interestingly, many key predictors of depression onset, as mentioned above, did not predict final remission. We also studied the course of side effects in these patients, and found that the number and severity of side effects decreased over time, and that only a small minority of our patients could not tolerate the prescribed antidepressant. We have compared different models of predicting final remission based on the amount of change in depression severity in the first weeks of treatment. Although the majority of patients finally achieving a response, can be identified early during treatment, even the best prediction model proved to be of limited clinical utility. Finally, we have found that there is no difference in decrease early in treatment between somatic and psychological symptoms of depression.