Predictors of response to lithium augmentation in tricyclic antidepressant-resistant depression

Background: The efficacy of lithium augmentation in therapy-resistant depre ssion has been shown in a series of well-designed, placebo-controlled studi es. However, little is known about the predictors of a good response to thi s treatment strategy. Methods: We retrospectively examined the predictive v alue of 20 demographic, clinical, biochemical and endocrinological variable s using a two-step logistic regression. Seventy-one in-patients with depres sion refractory to tricyclic antidepressants had received lithium augmentat ion as part of a standardised treatment protocol. Results: Within 4 weeks 3 7 patients (52%) responded to lithium augmentation. Five variables with pre dictive value were found. Responders were more severely depressed according to Bech-Rafaelsen Melancholia Scale. The duration of their index episode w as shorter. Triiodothyronine serum levels were lower and neuroleptic co-med ication and co-diagnosis of personality disorder were less frequent. Limita tions: This was an open, retrospective study. Conclusions: Severity of depr ession is a predictor of response to lithium augmentation. This result conf licts with recent studies but is similar to results found in studies of oth er a pharmacological antidepressant strategies. (C) 2001 Elsevier Science B .V. All rights reserved.