Thiazide diuretics are associated with an increased risk of hypokalemia. However, pharmacogenetic markers of thiazide-induced changes in serum potassium are not well studied. The aim of this study was to investigate possible predictors of serum potassium changes after thiazide treatment. Nondiabetic hypertensive patients with a systolic blood pressure of ⩾140 or a diastolic blood pressure of ⩾90 mm Hg were enrolled in our study. After 2 weeks of lifestyle modification and diet instruction, patients with persistently elevated blood pressure were given 50 mg of hydrochlorothiazide every morning for 2 weeks. Twenty single-nucleotide polymorphism (SNP) markers were selected from two candidate genes, SLC12A3 and WNK1. Serum potassium levels were checked before and after hydrochlorothiazide treatment. A total of 75 patients eventually qualified for enrollment in our study. They received 50 mg of hydrochlorothiazide every morning for 2 weeks. Six SNPs in WNK1 (rs11064524, rs4980973, rs12581940, rs880054, rs953361, and rs10849582) were correlated with decreases in serum potassium. None of the SLC12A3 polymorphisms were correlated with decreases in serum potassium. After Bonferroni's correction, only rs4980973 was correlated with decreases in serum potassium (corrected P=0.014). Multivariate stepwise linear regression analysis revealed that the changes in serum potassium levels were independently associated with the baseline potassium level (β=-0.587, 95% confidence interval=-0.875--0.299, P=0.0001) and WNK1 rs4980973 (A/A and A/G vs. G/G, β=-0.418, 95% confidence interval=-0.598--0.237, P=0.00002). In conclusion, the baseline potassium level and the WNK1 rs4980973 polymorphism were independent predictors of decreases in serum potassium after short-term hydrochlorothiazide treatment in nondiabetic hypertensive patients.