W10. NEURODEVELOPMENTAL DISORDER COPY NUMBER VARIANTS AND MULTIMORBIDITY OF INTERNALISING AND CARDIOMETABOLIC DISORDERS ACROSS LIFESPAN USING DANISH POPULATION-BASED BIOBANKS LINKED TO NATION-WIDE HEALTH AND SOCIAL REGISTERS

In a population-based sample of young Danes we found no association between carriage of neurodevelopmental disorder copy number variations (NDD-CNVs) and increased risk of depression. However, some recent evidence indicates that NDD-CNV-associated risk of internalising disorders (INT) in the UK Biobank is higher among older female subjects with comorbid cardiometabolic disorders (CMD) compared to those without CMD. We want to investigate whether risk of depression associated with NDD-CNVs is mediated through lifestyle-associated complications across lifespan, in a sex-dependent manner. We leverage on two existing Danish biobanks, the Copenhagen Hospital Biobank (N = 227,672) and the Danish Blood Donor Study (N = 100,872). CHB includes samples from patients admitted to hospitals for diagnostic and treatment purposes, with median at age 75; while DBDS includes healthy volunteer blood donors, with median age at 50. Both biobanks are linked with a comprehensive collection of nation-wide registers, which includes hospital diagnoses, drug prescriptions, education, and income. This grants us unprecedented opportunity to study a wide range of disease outcomes and ancillary variables in association with genetic exposures such as NDD-CNVs. We conduct a sex-stratified analysis of the association between NDD-CNVs and risk of internalising disorders, with and without comorbid CMD, adjusting for age, separately for DBDS and CHB. We defined 3 disease groups, INT only, CMD only and comorbid INT-CMD. We investigate whether 24 single locus CNVs with a population prevalence of > 0.015%, are associated with these 3 disease groups. To increase statistical power, we also grouped together CNVs based on their LOEUF scores, a measure that relates to the constraint of genes contained within a CNV. We categorized CNVs with LOEUF score of less than 10 as low constraint, between 10 and 25 as medium constraint, and greater than 25 as high constraint. We also use generalized estimating equation (GEE) to account for single- or multi-morbidity within individuals and estimate the population average risk. Additionally, multistate modelling is employed to investigate the longitudinal progression through different disease states across lifespan. There is very little evidence that any CNVs or LOEUF groups are associated with INT only, for both CHB and DBDS. However, more significant signals are found for CMD only and comorbid INT-CMD, mostly in DBDS rather than CHB, regardless of a considerable larger sample size for the latter. The high constraint LOEUF group is more significantly associated with CMD only and comorbid INT-CMD in DBDS, compared to CHB. The lack of evidence for the association between NDD-CNVs and INT replicated what we found in a previous study. The differences we observed in analyses on the CHB only and DBDS cohorts are likely due to the different profile of the participants in the cohorts, where in DBDS they are younger, healthier blood donors, while in CHB they are older adults having received treatment for unknown conditions.