Background An individual’s healthcare costs can provide an objective measure of overall morbidity. Quantifying the impact of genetic variation on healthcare costs can inform public health policies and help identify genetic determinants of mortality.

Methods We estimated inpatient, outpatient, primary care and prescription-related costs, from the perspective of healthcare systems, for a total of 1.3 million individuals across 8 different healthcare systems. The median per-patient annual healthcare cost (expressed in 2019 prices) had substantial range across countries, from €483/year in the Estonian Biobank to €458,455/year in the Mass General Brigham Biobank.

A GWAS meta-analysis identified 27 and 106 significant, independent loci for inpatient and medication costs, respectively. Effects were overall modest, the most significant inpatient signal (rs6938239 near ILRUN) was associated with ~1% increased cost per year, per allele. Following stratified analyses, 33% of inpatient loci had sex-specific effects while 22% had an age-specific effect. E.g. homozygous carriers of variant rs62236881 in the 3’ UTR of ZNRF3 had 9% higher costs in females compared with 4% in males, potentially due to an increased risk of breast cancer.

Results Despite differences in tariff structure and cost reimbursement across healthcare systems, we observed strong genetic correlations between inpatient studies (median rg=0.88) and 95% of significant loci did not show heterogeneity across studies. PheWAS indicated that significant loci may impact healthcare costs via either common risk factors (e.g. BMI) or chronic diseases risk (e.g. asthma, depression).

Leveraging whole-exome sequencing data from the UK Biobank (UKB), we estimated the cost for putative loss of function (pLOF) burdens for 83 clinically relevant genes. The largest effects were observed for BRCA1/2, MSH2 and APC. Annual inpatient costs were ~1.8 times higher in individuals carrying one pLOF in BRCA2 than in non-carriers. We estimated that the total annual inpatient costs attributed to BRCA2 pLOF mutations amongst those aged 40-69 in the UK would be about €26 million.

Polygenic scores (PGS) for various risk factors and diseases were linked to higher healthcare costs. In UKB, the top 10% with the highest BMI PGS had a median annual cost of €414, compared to €329 for those in the middle 40-60%. Using polygenic weights from GenCost data, PGS for ~11,000 individuals in Genomics England showed that the top 10% incurred ~€250 more than those in the middle 40-60%.

Conclusion Genetic variation across individuals influences healthcare expenditure. Our findings are applicable across various healthcare systems and help better understanding global morbidity.