Polygenic Risk Score and Psychiatric Family History in Autism Spectrum Disorder

Background: Both psychiatric family history and genetic liability are linked to autism spectrum disorder (ASD) and there is considerable familial and genetic overlap between ASD and specific psychiatric conditions.

Objectives: We performed a population-based integration of family history and genetic liability, using the ASD polygenic risk score (ASD-PRS; derived from the ASD iPSYCH-PGC GWAS), in ASD risk. We examined ASD-PRS variation associated with distinct psychiatric family histories and ASD risk from both psychiatric family history and ASD-PRS.

Methods: The iPSYCH study population comprised all Denmark-born singleton births, 1 May 1981-31 December 2005; resided in Denmark on their one-year birthday; and mother known. ASD cases comprised study persons with an ICD-10 ASD diagnosis reported through 2012 (n=16,146) to the Danish Psychiatric Central Research Registry (PCRR). iPSYCH controls comprised a random sample of 30,000 persons from the study population (2%). ASD cases and non-ASD controls in this study additionally comprised persons whose parents were born in Denmark (84% ASD cases, 84% controls) and, among these, had an ASD-PRS (91% ASD cases, 88% controls with Denmark-born parents). We defined eight non-overlapping family history categories based on PCRR diagnoses reported through 2016 in parents or full siblings: ASD; schizophrenia; developmental disorders or intellectual disability; ADHD; affective disorders; adult personality disorders; any other psychiatric diagnosis; no psychiatric diagnoses. We estimated least-squares means (LSMean) for PRS z-scores by psychiatric family history category and adjusted odds ratios (adjOR) for ASD by decile of PRS and by psychiatric family history category. All analyses were adjusted for the first 5 genetic ancestry principal components, year of birth, age at end of follow-up, sex, parental ages at birth, and having a sibling (Yes/No). ORs by PRS decile were additionally adjusted for psychiatric family history and ORs by psychiatric family history were additionally adjusted for PRS.

Results: Psychiatric family history was observed in 36% of ASD cases (n=12,346; 9% with ASD family history) versus 19% of non-ASD controls (n=21,904; 2% with ASD family history). ASD risk increased by PRS decile (10^th decile adjOR 2.21 (1.98-2.46)) and by each psychiatric family history category (highest risk, ASD family history (adjOR 6.71 (5.91-7.61); lowest risk, other psychiatric family history (adjOR 1.73 (1.58-1.90)). There was virtually no attenuation of ASD risk after adjusting PRS ORs for psychiatric family history or adjusting psychiatric family history ORs for PRS. After Bonferroni correction, among ASD cases the only significant pair-wise difference in PRS z-score LSMean by psychiatric family history was between cases with an ASD family history and no psychiatric family history. The PRS z-score LSMean of controls with an ASD family history was comparable to that of ASD cases, regardless of family history, and significantly higher than the control LSMean with no psychiatric family history.

Conclusions:

ASD polygenic burden and psychiatric family history appear to independently contribute to ASD risk, yet better understanding of factors contributing to ASD-PRS variation with psychiatric family history is warranted. Integration of individual genetic liability and psychiatric family history is critical for understanding ASD risk architecture.