Evaluating the role of common risk variation in the recurrence risk of schizophrenia in multiplex schizophrenia families

Abstract

Multiplex families have higher recurrence risk of schizophrenia compared to the families of sporadic cases, but the source of this increased recurrence risk is unknown. We used schizophrenia genome-wide association study data (N = 156,509) to construct polygenic risk scores (PRS) in 1005 individuals from 257 multiplex schizophrenia families, 2114 ancestry-matched sporadic cases, and 2205 population controls, to evaluate whether increased PRS can explain the higher recurrence risk of schizophrenia in multiplex families compared to ancestry-matched sporadic cases. Using mixed-effects logistic regression with family structure modeled as a random effect, we show that SCZ PRS in familial cases does not differ significantly from sporadic cases either with, or without family history (FH) of psychotic disorders (All sporadic cases p = 0.90, FH+ cases p = 0.88, FH- cases p = 0.82). These results indicate that increased burden of common schizophrenia risk variation as indexed by current SCZ PRS, is unlikely to account for the higher recurrence risk of schizophrenia in multiplex families. In the absence of elevated PRS, segregation of rare risk variation or environmental influences unique to the families may explain the increased familial recurrence risk. These findings also further validate a genetically influenced psychosis spectrum, as shown by a continuous increase of common SCZ risk variation burden from unaffected relatives to schizophrenia cases in multiplex families. Finally, these results suggest that common risk variation loading are unlikely to be predictive of schizophrenia recurrence risk in the families of index probands, and additional components of genetic risk must be identified and included in order to improve recurrence risk prediction.

Fig. 1. Mean Leave-N-Out PGC3-SCZ PRS for each of the diagnostic categories in the ISHDSF sample, sporadic SCZ cases and ancestry-matched population controls.

A subset of sporadic cases with family history (FH) information are further divided into FH+ (green bar) and FH− (red bar) categories. Unaffected relatives (dark orange bar) are distinct from population controls (black bar) as they represent unaffected individuals in the families. Familial SCZ cases are represented in the narrow spectrum category as described in the methods section. Error bars represent the standard error of the observed mean. X axis shows each of the diagnostic categories. Y axis shows the mean normalized Z-score for PGC3-SCZ.

Fig. 2. Comparison of PRS between the ISHDSF diagnostic categories and sporadic cases, versus population controls.

All analyses follow the hypothesis that ISHDSF members and sporadic cases have a higher PRS compared to population controls. All PRS have been normalized using Z-score standardization prior to obtaining odds ratios. The plots show odds ratios (OR, filled circles) with 95% confidence intervals (CI) for each category compared to population controls. X axis represents the odds ratios. Left side of Y axis represents each of the categories used for comparison versus population controls. Right side of the Y axis represents the p-values after multiple testing correction.