Treatment outcome variation between depression symptom combinations in the STAR*D study

Background: In response to recent documentation of symptom and subtype heterogeneity in major depressive disorder, we report on exploratory analyses of the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) clinical-trial data to further describe heterogeneity in depression and test the hypothesis that citalopram treatment-outcome patterns differ as a function of depression symptom combinations. Methods: Combinatorial algorithms, latent profile analysis, and repeated-measures multivariate analysis of variance were employed to characterize heterogeneity and depression outcome-measure profile variability in the most prevalent symptom combinations with full data (26% of baseline and 13% of endpoint total sample). Results: Descriptive results suggest that substantial heterogeneity and moderate coherence characterize major depressive disorder; as in previous analyses, pairs of individuals sharing no symptoms in common were observed. Exploratory latent profile analysis indicated that different patterns of treatment outcome data exist among STAR*D participants. A small but significant interaction effect of symptom combination x outcome measure profile was observed for clinician-rated but not self-reported symptom combinations. Limitations: Factors moderating the generalizability of these findings include binary symptom measures, a short treatment period, and a smaller number of individuals per combination. Conclusions: These results provide evidence that citalopram treatment outcomes vary as a function of diagnostic combinations, thereby providing preliminary evidence that the substantial heterogeneity documented in depression symptom presentations may carry implications for prognosis and treatment outcome. At the level of descriptive phenomenology, these results appear to corroborate the claim that depression is not a homogenous syndrome. (C) 2016 Elsevier B.V. All rights reserved.