Rheumatic heart disease (RHD) remains a leading cause of morbidity and premature mortality in developing countries. Recently, large-scale genetic studies have begun to provide much needed insight into the underlying pathogenesis.
We undertook a genome-wide association study of RHD susceptibility in 1163 individuals of Indian ancestry recruited in Fiji (n=309) and Northern India (n=854). Patients with incident or prevalent RHD were recruited as cases, while members of the general population were recruited as controls. We genotyped all individuals at ~250,000 variants using the Illumina HumanCore platform before estimating the genotypes of an additional ~5,000,000 variants by imputation. We used linear mixed models to analyse genotyped and imputed variants before combining association statistics from the two datasets using fixed-effects meta-analysis.
We observed a single signal at genome-wide significance, located in the human leukocyte antigen (HLA) class III region (OR=1.91, P=6.8 x 10-9). Using conditional analyses, we demonstrated this was comprised of two independent signals, the first spanning HLA class I (HLA-B) and the second HLA class II (HLA-DQA1/B1). While the IGHV4-61 signal was apparent, as previously reported in the Fijian Indian data (P=0.006), the signal was essentially absent in the Northern Indian data (P=0.25), leaving only negligible signal in the pooled analysis (OR=1.21, P=0.084).
By combining new data from India with previously reported data from Fiji, we provide the first insight into genetic architecture of RHD in South Asians with a promising finding in the HLA locus. We also highlight the challenge of heterogeneity between populations.