Clinical Data Identify Psychoses Subgroups

By Marilynn Larkin

February 20, 2020

NEW YORK (Reuters Health) - Subgroups of patients with schizophrenia and other affective disorders were identified through data-driven clustering, a technique that analyzes clinical data to identify common denominators among patients.

"The study shows that we can use computers to help us to rethink how people with established psychosis symptoms are diagnosed," Dr. Dominic Dwyer of Maximilian University of Munich told Reuters Health by email. "The study may lead to the definition of more specific psychosis categories that increase the precision of psychiatric care and research."

"The findings...showed me a different way to think about individuals who had been suffering from psychosis for many years," he noted. "When we re-categorized people, it was exciting to see that this helped us to detect hidden illness trajectories and genetic associations."

Dr. Dwyer and colleagues collected data from an ongoing, multisite study that started in 2012 across 18 sites in Germany and Austria. The study included a referred sample of 1,223 individuals - 765 in the discovery sample (mean age, 43; 45% women) and 458 in the validation sample - with DSM-IV diagnoses of schizophrenia, bipolar affective disorder (I/II), schizoaffective disorder, schizophreniform disorder, and brief psychotic disorder.

Illness courses were examined over 18 months and polygenic risk scores for schizophrenia, bipolar disorder, major depressive disorder, and educational achievement were assessed.

Clustering methods were used to separate out 188 variables measuring demographic characteristics, clinical history, symptoms, functioning, and cognition; of these, 45 variables were identified by supervised learning techniques as the most reliably discriminative.

As reported in JAMA Psychiatry, five psychosis subgroups were identified: affective (252 participants), suicidal (44), depressive (131), high-functioning (252), and severe (86).

Illness courses were described according to psychosis symptoms, depression symptoms, global functioning, and quality of life.

The depressive and severe psychosis subgroups demonstrated the lowest functioning, and illness courses that included partial recovery followed by recurrence of severe illness. A message for clinicians, Dr. Dwyer said, is that "subgroups of individuals with depression and severe psychosis who partially remit should be monitored to detect a functional relapse at 12-18 months."

Polygenic risk scores for educational attainment differed significantly among the subgroups, but diagnostic polygenic risks did not.

The results were largely replicated in the validation cohort, except for the suicidal subgroup, due to missing data. Also, in contrast to the discovery sample, there was a comparative increase in the effect size of subgroup differences using the schizophrenia polygenic risk score, according to the authors.

"Next steps," Dr. Dwyer said, "are to further investigate the groups in terms of associations with medication and treatment, to look at any other biological differences between the groups, and to validate the results in another sample."

Dr. Tiffany Herlands, a psychologist at ColumbiaDoctors and assistant professor of medical psychology at Columbia University Irving Medical Center in New York City, commented in an email to Reuters Health, "The study does not alter current clinical practice, but contributes to efforts that may one day redefine the diagnostic system in psychiatry, which may have practical clinical implications."

"Despite the heterogeneity of psychotic disorders and the challenge of establishing a more accurate diagnostic nomenclature...there are current treatments which successfully support psychiatric recovery," she said. "Providing comprehensive treatment that targets cognitive functioning, social functioning, intrinsic motivation, independent living skills and supports and includes families has proven to be a successful model to improve functioning and quality of life in persons with psychotic disorders."

Dr. Megan Schabbing, Medical Director, OhioHealth Psychiatric Emergency Services in Columbus, Ohio, also commented in an email to Reuters Health, "The identification of psychosis subgroups with distinct genetic, clinical and prognostic features shows promise in better informing and guiding the clinical management of a patient with psychotic illness."

However, she added, "Further research must be done before we can rely on this approach to dictate...clinical care for patients with psychosis and other brain disorders."

SOURCE: JAMA Psychiatry, online February 12, 2020.