Background/Purpose:  Virtually all gout patients have high levels of uric acid in the blood (hyperurincemia, HU), but approximately 80% of patients with HU will never develop gout.  There are a numbers of GWAS that identified risk alleles for hyperuricemia and/or gout using general population as comparison. This study examines genes that might influence the risk of gout among subjects with HU.

Methods: Subjects with HU (uric acid ≥6.5 mg/dl from lab data) and available GWAS data were requested from Partners HealthCare Biobank, a collaborative database that contains linked electronic medical records (EMR), survey data and genomic data for 15,000+ subjects. Gout was defined using a validated EMR algorithm with positive predict value of 0.95. Samples were genotyped from three Illumina MEGA chips.  We performed standard quality control (QC) procedures for each file, merged them into one analysis dataset, and applied standard QC again.  We restricted our genome-wide association study (GWAS) analysis to the Caucasian population.  Principal component analysis (PCA) was performed. Among subjects with HU, association between SNPs and gout was tested using logistic regression assuming a genetic additive model. The first 10 PCAs were utilized as covariates to control for any potential population stratification. We also looked at potential gout and HU SNPs with genome-wide significance among European ancestry from the GWAS Catalog (https://www.ebi.ac.uk/gwas/) .

Results: 3146 self-reported Caucasian subjects with HU were identified from the Biobank, with 467 (14.8%) were identified as having gout. The gout group was slightly older (mean ±SD, 72±11 vs 66±14, p<0.0001), more likely to be male (79% vs 56%, p<0.0001) compared to non-gout HU group.  793,514 SNPs passed QC and were utilized in the GWAS.  The most significant SNP was rs1481012 in ABCG2 gene with p value of 1.8 x 10^-6, which is a known risk factor for gout.  We identified 14 SNPs at p ~10^-6 (see Table); 12 were new SNPs and 2 had been previously identified as gout/HU risk SNPs from the GWAS catalog.  53 SNPs were identified at p~10^-5; only 1 is a previously identified gout/HU risk SNP.  623 SNPs were identified at p~10^-4; with only 1 as a known gout/HU SNP.  

Conclusion: We carried out a GWAS of gout in patients with HU, and found that some of the risk alleles for gout in the general population are also associated with gout in a population with HU. But, we found many possible new SNPs and some previously identified SNPS were null in the HU population. The null findings may be because of limited power, confounding, or true non associations. 

Table   Top 20 SNPs associated with gout compared to hyperuricemia