The course of depressive symptoms in predicting relapse in schizophrenia: A double-blind, randomized comparison of olanzapine and risperidone

Background: Depressive symptoms are common during the course of schizophren ia and may carry prognostic relevance. Methods: From a 28-week prospective, double-blind, randomized study of olan zapine and risperidone, a post hoc evaluation of changes on the Positive an d Negative Syndrome Scale (PANSS) depression cluster (PDC) and the subseque nt risk of relapse were analyzed by logistic regression. Results: Olanzapine was associated with a significantly higher categorical rate of improvement on the PANSS depression cluster (greater than or equal to 7 points) (p < .05). Although the baseline severity of depressive sympto ms was not a significant predictor of relapse, the degree of acute (8-week) mood improvement on the PANSS depression cluster (but neither negative or positive symptom changes) was related to the probability of a subsequent ps ychotic relapse. Acute mood improvement with olanzapine was inversely relat ed to a nonsignificantly lower risk of relapse. However, an opposite and si gnificant relationship was observed among risperidone-treated subjects. Ris peridone-treated subjects with a greater degree of acute mood change were b oth 3.58 times more likely to relapse than their risperidone counterparts w ho had experienced less mood improvement (p = .008) and 8.55 times more lik ely than olanzapine-treated subjects who had had similar mood improvements (p = .001). Conclusions: These data suggest the underlying p pharmacologic differences between the two drugs may bestow different rates of longer-term mood stabil ization and relapse prevention. In a second series of analyses, worsening o n the PANSS depression cluster in the 4 weeks or less preceding a clinical relapse was a significant prodromal predictor of relapse among all subjects . As a whole, subjects with a worsening on the PDC demonstrated a 1.77 time s higher risk of a relapse during the subsequent 4 weeks (p = .001). Among this mood-worsening stratum, risperidone-treated patients were 3.51 times m ore likely to relapse in those next 4 weeks (p = .005) than their olanzapin e counterparts. Future comparative drug studies in this area will further c ontribute to our understanding of the pathophysiology of mood change and it s relationship to psychosis, including clinical relapse and how newer agent s may differ in their respective delivery of long-term treatment outcomes. Biol Psychiatry 1999;46:365-373 (C) 1999 Society of Biological Psychiatry.