Clinical, cognitive, and premorbid functioning correlates of autistic traits in early psychosis: Findings from the BICEPS study
Nicolas Raymond (1), Megan Good (1,2), Kathryn E. Lewandowski (2,3), Lynn DeLisi (4), Elliot Gershon (5), Sarah Keedy (5), Jennifer E. McDowell (6), Brett Clementz (6), Godfrey Pearlson (7,8), Carol Tamminga (9), Matcheri S. Keshavan (1,2)
1. Department of Psychiatry, Beth Israel Deaconess Medical Center, Boston MA
2. Department of Psychiatry, Harvard Medical School, Boston MA
3. Division of Psychotic Disorders, McLean Hospital, Belmont MA
4. Psychiatry Research Office, Cambridge Health Alliance, Cambridge MA
5. Department of Psychiatry and Behavioral Neuroscience, University of Chicago, Chicago, IL
6. Department of Psychology, University of Georgia, Athens, GA
7. Department of Psychiatry, Yale University, New Haven, CT
8. Olin Neuropsychiatry Research Center, Institute of Living, Hartford HealthCare, Hartford, CT
9. Department of Psychiatry, UT Southwestern Medical Center, Dallas, TX
Background: Approximately 75% of genes associated with autism are also implicated in schizophrenia. Phenotypically, there is overlap between psychosis symptoms and autistic behaviors. However, research understanding how autistic traits impact outcomes of in early psychosis (EP) is lacking.
Methods: 104 participants (39.4% female, mean age (standard deviation)=24.04 (4.27) years, 39.4% identifying as a racial minority) with EP (<4 years from symptom onset) were enrolled and eligible. Participants completed the brief assessment of cognition in schizophrenia (BACS), positive and negative syndrome scale (PANSS), premorbid adjustment scale (PAS), and social responsiveness scale-2 (SRS-2). The SRS-2 quantifies autism traits, yielding total scores and two- (social communication/interaction (SCI), restricted/repetitive behavior (RRB)) and five-factor subscale scores (emotion recognition (ER), social avoidance, interpersonal relatedness, insistence on sameness, repetitive mannerisms (RM)). Spearman correlations were computed between the SRS-2 total and subscales with the BACS, PANSS, and PAS scores. The false discovery rate (FDR) corrected P-values.
Results: The PAS correlated significantly with the SRS-2 total (ρ=0.29, p=0.0056, FDR-p=0.025), SCI (ρ=0.29, p=0.0060, FDR-p=0.025), and ER scores (ρ=0.28, p=0.0078, FDR-p=0.027). The RM subscale correlated with the symbol coding (ρ=0.29, p=0.0080, FDR-p=0.027), verbal fluency (ρ=0.30, p=0.0069, FDR-p=0.027), and composite BACS scores (ρ=0.28, p=0.013, FDR-p=0.041), which the RBB score also correlated with (ρ=0.28, p=0.013, FDR-p=0.041). We found the SRS-2 total and all its subscales correlated significantly with PANSS total (ρ>0.27, p<0.0077, FDR-p<0.027) general (ρ>0.40, p<0.000048, FDR-p<0.00046) and positive scores (ρ>0.25, p<0.015, FDR-p<0.045). Additionally, the ER subscale correlated significantly with the PANSS negative subscale (ρ=0.27, p=0.0091, FDR-p=0.03). Other correlations were not significant (FDR-p>0.05).
Conclusions: Greater prevalence of autism-like traits in a sample of individuals with EP related to poorer premorbid functioning, more severe symptoms, but better cognition. While future work must evaluate the significance of baseline autism-like traits on longitudinal outcomes, results indicate the import of autistic traits to the presentation of EP.