Remission in First-Episode Psychosis: Predictor Variables and Symptom Improvement Patterns

Robin Emsley, MD; Piet P. Oosthuizen, MD; Martin Kidd, PhD; Liezl Koen, MB; Dana J. H. Niehaus, MD; H. Jadri Turner, MB

J Clin Psychiatry 2006;67(11):1707-1712

Article Abstract

Background: Previous attempts to identify clinically useful predictors of treatment outcome in schizophrenia have been hampered by methodological inconsistencies, including a lack of standardized outcome measures. Recently proposed operationally defined criteria for remission provide an opportunity to readdress this topic.

Method: We applied the remission criteria to a sample of 57 subjects with first-episode psychosis (DSM-IV schizophrenia, schizoaffective disorder, or schizophreniform disorder), treated according to a fixed protocol in a prospective study. Subjects were recruited between April 1999 and January 2000 and were followed for 2 years. Various demographic, baseline clinical, and early-response variables were subjected to discriminant analysis for their ability to predict remission or nonremission. We also assessed the symptom improvement patterns over time and compared endpoint psychopathology in the remitters and nonremitters.

Results: A model incorporating neurologic soft signs, 6-week treatment response, duration of untreated psychosis, marital status, and Positive and Negative Syndrome Scale excited factor baseline score was able to correctly predict 89% of the remitters and 86% of the nonremitters. Symptom reduction at 6 weeks, including core psychotic symptoms, was significant in both groups (remitters, p < .0001; nonremitters, p < .0001), although reduction was substantially greater in the remission group (p = .004). Thereafter, the remission group continued to improve (p < .01), while the nonremitting group failed to do so (p = .55). Considerable overlap of endpoint symptoms was observed, and depressive symptom scores were similar in remitters and nonremitters.

Conclusion: A combination of demographic, baseline clinical, and acute treatment response variables may accurately predict treatment outcome. Persistent noncore psychotic symptoms in subjects meeting proposed remission criteria require further investigation.