Genome-Wide Association Study of Apparent Treatment Resistant Hypertension in the Million Veteran Program

Background: Apparent treatment resistant hypertension (ATRH) is common. 9-16% of individuals in different populations are affected, and it is associated with poor outcomes compared with controlled hypertension, including increased risk of ESRD, cardiovascular morbidity and mortality. Methods: We performed a large GWAS of ATRH in US veterans enrolled in the Million Veteran Program. ATRH was defined as treated with 3 different antihypertensive drugs including a thiazide with BP above goal (SBP ≥ 140 mm Hg and/or diastolic BP ≥ 90 mm Hg) or the use of 4 or more antihypertensive drugs regardless of BP control. Controls were patients with controlled hypertension. All participants had eGFR ≥60 ml/min/1.73m2. The analysis included 26,902 cases (27% AA, 73% EA) and 81,187 controls (18% AA, 82% EA). Association analyses were performed with logistic regression with an additive model adjusting for age, sex, body mass index, and 10 principal components of ancestry in GWAS data imputed using 1000 genomes project haplotypes. Analyses were stratified by ancestry (EA and AA) and inverse variance-weighted fixed-effects meta-analyses were carried out using METAL. Results: Twenty one loci reached genome wide significance in the transethnic meta-analysis: CASZ1 (2.46E-11), WNT2B(2.84E-09), KCNK3 (2.78E-15) CACNA1D (2.14E-11), ENPEP (6.38E-10), FBN2(1.05E-11), HTR4(3.88E-09), TRIM36 (1.89E-09), RSPO3(3.17E-10), SLC22A7 (1.25E-09), HOTTIP (8.50E-17), HOXA10-HOXA9 (8.40E-11), CYP11B2-GML (4.84E-09), GATA4 (1.03E-11), MAPKAP1(2.10E-08) LSP1 (3.09E-10), TBX3(1.39E-09), RXFP2 (4.27E-24) and KLF5 (1.33E-11). CASZ1 was previously reported associated with ATRH in REGARDS and CACNA1D has also been reported in candidate gene studies. The remaining associated loci are novel for ATRH. All these loci have been previously associated with BP or hypertension. We also performed a phenome-wide association analysis for each sentinel variant and three loci were associated with hyperaldosteronism and hypopotassemia. Conclusions: These results indicate that RH genetic susceptibility likely arises at known BP loci. Also our study results suggests an underlying physiology related to mineralocorticoid physiology. Funding: Veterans Affairs Support