Predicting High Excess Risk of Hyponatremia Among Thiazide Users

Background and Aims Hyponatremia is a potential serious adverse drug reaction to treatment with thiazide diuretics. This study aimed to determine whether individuals at high risk of developing thiazide-induced hyponatremia can be identified before treatment initiation. Methods A population-based cohort study was conducted in Denmark among individuals aged ≥40 years from 2014 to 2020. Moderate-to-severe hyponatremia (plasma sodium <130 mmol/L) within 120 days of treatment was compared in new users of thiazide or non-thiazide antihypertensive drugs. Using the causal forest method, models to predict individual-level risk of thiazide-induced hyponatremia were trained in a development cohort (n=185,699; 2014-2018) and validated in a separate cohort (n=75,030; 2019-2020). Results Individual-level excess risk could be parsimoniously described by a four-covariate model that included information on age and baseline plasma sodium, hemoglobin, and C-reactive protein levels with good calibration and concordance-for-benefit (0.66; 95% CI, 0.66-0.67) in the validation cohort. The average 120-day excess risk of hyponatremia among thiazide-treated patients was 1.8% (95% CI, 1.3% to 2.2%), with individual-level heterogeneity ranging from –1.6% to 15.9%. For the 10% of thiazide-treated with the highest excess risk of hyponatremia the average excess risk was 7.4% (95% CI, 4.4% to 10.5%). Reassigning this high-risk group to non-thiazide drugs would reduce the excess risk within the thiazide-treated population by 0.7% (95% CI, 0.4%-1.0%), corresponding to a 42% relative reduction. Conclusions The population-level burden of thiazide-induced hyponatremia can potentially be markedly reduced by identifying and prescribing alternative antihypertensive drugs to high-risk patient groups using a simple set of baseline information.