ICOSR 2011—Predicting Outcome in Early Psychosis
As the final installment in our coverage of the 2011 International Congress on Schizophrenia Research (ICOSR), held 2-6 April in Colorado Springs, Colorado, we bring you this session summary from Young Investigator travel award winner Kristen A. Woodberry of Beth Israel Deaconess Medical Center and Harvard Medical School. Thanks so much Kristen and all the other rapporteurs!
16 December 2011. Presenters in the Wednesday, 6 April afternoon session, chaired by Kristin Cadenhead of the University of California, San Diego, reported on a number of new findings related to predicting the onset of psychosis and differential diagnosis in the context of complex psychiatric presentations. Cross-sectional, longitudinal, and collateral data were brought into play to highlight symptom and functional predictors of elevated or reduced risk and symptom trajectories over time.
A number of speakers reviewed studies of individuals at “clinical” or “ultra” high risk (CHR or UHR, respectively) for psychosis. Eva Velthorst of the University of Amsterdam, The Netherlands, discussed self-reported disability data from the European Prediction of Psychosis Study. Social disability, as measured by difficulty “getting along with people” on the Disability Assessment Schedule of the World Health Organization (WHO-DAS-II), contributed significantly to prediction of later psychosis in young people at clinical high risk based on the Structured Interview of Prodromal Syndromes (SIPS) or the Bonn Scale for the Assessment of Basic Symptoms (BSABS). In fact, the transition rate nearly doubled for those who displayed this social difficulty at baseline.
In the next talk, Barnaby Nelson of the University of Melbourne, Australia, reported on analyses of differential rates of transition to psychosis in the various subgroups of UHR. As predicted, conversion rates at six-month follow-up were highest for those with baseline brief limited intermittent psychotic syndrome (BLIPS, 13.9 percent), and lowest for those with a trait vulnerability (functional decline in the context of a family history of psychosis or schizotypal personality disorder), with the largest group with attenuated psychotic symptoms (APS), with and without coexisting trait vulnerability, having intermediate rates of conversion (8.7 and 9.4 percent, respectively). When these intermediate groups were combined, the BLIPS, APS, and trait vulnerability subgroups had significantly different rates of transition at six months, consistent with models of early and late prodromal populations, suggesting a further “closing in” on those at highest risk.
From an analysis of CHR non-converters assessed in the Center for the Assessment and Prevention of Prodromal States (CAPPS) program, Danielle Schlosser of the University of California, San Francisco (UCSF), proposed a refinement of the CHR criteria to limit exposure of false positive cases to risks associated with interventions. Non-converters included a group with remitted positive symptoms over two-year clinical follow-up (47 percent), and a group with stable positive symptoms without conversion. Those whose symptoms remitted had significantly lower initial positive (P) symptom scores on the Structured Interview of Prodromal Syndromes (SIPS) and better psychosocial functioning than both converters and non-converters who remained symptomatic. In fact, using a total P score of 7 at baseline as a cutoff provided a substantial increase in positive predictive power.
Rachel Loewy of UCSF reported that a self-reported history of childhood trauma might be associated with increased risk for transition to psychosis. Examining one possible mechanism for such a vulnerability-stress interaction, she examined salivary cortisol in a CHR sample. Results suggested normal dexamethasone suppression, and thus normal hypothalamic pituitary adrenal (HPA) axis functioning, but a slower return to baseline cortisol levels following the Trier Social Stress Test. Subsequent discussion raised the question of whether there might be a blunting of response to stress associated with prior exposure to trauma.
The remaining talks focused on issues of differential diagnosis and symptom trajectories over time. Comparing data from a sample of youth with first-episode schizophrenia (FES), CHR, and autism spectrum disorders (ASD) relative to a typically developing group (TYP), Bailey Seymour from the University of California, Davis, presented parent reports of internalizing and externalizing behaviors using the Behavioral Assessment System for Children (BASC)- Parent Rating Scale—Adolescent Version. All clinical groups rated high on withdrawal and showed no differences in mean level of anxiety, but the FES and CHR groups were rated higher on aggression and conduct problems. The CHR and ASD groups showed the most difficulty with hyperactivity symptoms. The atypicality scale of this measure applies to a broad range of symptoms and was actually highest for the ASD group. Specific item review for this scale is needed to prompt further inquiry about potential psychotic symptoms.
In a similar presentation highlighting diagnostic difficulties in children with complex presentations, Rachel Miller of the National Institute of Mental Health (NIMH) reported on longitudinal follow-up diagnoses in an NIMH sample referred for possible childhood-onset schizophrenia. Out of 202 hospitalized for a medication washout after provisional diagnosis with schizophrenia, 85 were given alternative diagnoses, and 35 of these (mean age 11 at admission) were available for follow-up (mean age 26). At follow-up, one (3 percent) met criteria for schizophrenia, two (6 percent) for schizoaffective disorder, seven (20 percent) for psychosis NOS, and 17 (48 percent) for bipolar disorder. The other 20 percent had mood, anxiety, or no diagnoses. Issues such as central processing disorders and emerging personality issues were discussed as components of the complexity of diagnosing this young and seriously disabled population.
Examining the role of symptom trajectories in the differential diagnosis of psychotic and mood disorders, Vina Goghari of the University of British Columbia, Canada, presented data on the longitudinal trajectory of hallucinations in schizophrenia, schizoaffective, and bipolar disorder with psychosis patients from the Chicago Follow-Up Study. In short, she found that the schizophrenia group had the highest overall rates of hallucinations, with schizoaffective patients having initial rates more similar to the schizophrenia group, but over time, looking more similar to bipolar patients. In contrast to some previous studies, she found that hallucinations did contribute to the prediction of later work and global functioning.
In the final talk of the session, Frank Van Dael of Maastricht University, The Netherlands, examined the provocative question, “Can obsessions drive you mad?” Reporting on subclinical obsessive-compulsive and subclinical psychotic symptoms at three time points in general population data from The Netherlands Mental Health Survey and Incidence Study (NEMESIS), his group found that OCD and psychosis phenotypes cluster together and predict each other over time. In fact, the presence of obsessive-compulsive symptoms at baseline predicted both the persistence of psychotic symptoms and the transition to a psychotic disorder at two years.
In sum, imminent risk for a transition to psychosis appears to be elevated in clinical high-risk populations with brief, limited, intermittent psychotic symptoms; difficulty getting along with other people; and possibly histories of trauma. In the general population, obsessive-compulsive symptoms increase risk for a later psychotic disorder. By contrast, those with a clinical high-risk syndrome with initially low total positive symptom scores on the SIPS and higher initial functioning are at significantly reduced risk. Somewhat surprisingly, higher scores on the atypicality scale of the commonly used BASC-Parent Report are more likely to be associated with developmental difficulties than psychosis, and vulnerability to stress does not necessarily imply abnormalities in the body’s stress response system. Not surprisingly, time is needed for diagnostic clarity in the most complex cases.—Kristen A. Woodberry.