Thiazide-Associated Hyponatremia

A Population-Based Study

Eline M. Rodenburg, MD; Ewout J. Hoorn, MD, PhD; Rikje Ruiter, MD, PhD; Jan J. Lous, MSc; Albert Hofman, MD, PhD; André G. Uitterlinden, PhD; Bruno H. Stricker, PhD; Loes E. Visser, PharmD, PhD


Am J Kidney Dis. 2013;62(1):67-72. 

In This Article


The study population consisted of 13,325 participants, of whom 59% were women. Mean age was 63.1 ± 13.9 (SD) years (Table 1). Before the start of the study period, 1,346 participants had ever used a thiazide (10.1%), of which most used hydrochlorothiazide, often in combination with a potassium-sparing agent. During the start of the study, there were 718 prevalent users of thiazides (5%). An additional 2,738 participants started on thiazide therapy after the start of the study period, on average 6 years after the start of the study period. During follow-up, 522 participants developed hyponatremia. Of these, 32.4% were exposed to thiazide diuretics at the time of hyponatremia.

Thiazide exposure was associated with an almost 5 times higher risk of hyponatremia than nonexposure (HR, 4.95; 95% CI, 4.12–5.96). In women, this risk was slightly, but not significantly, higher than in men (HRs of 5.31 [95% CI, 4.32–6.53] and 3.71 [95% CI, 2.38–5.77], respectively; P for interaction = 0.8; Table 2). Exposure with a minimum dose of 1 DDD was associated with a significantly higher risk of hyponatremia than exposure <1 DDD (HRs of 5.72 [95% CI, 4.67–7.00] and 3.42 [95% CI, 2.46–4.77], respectively). The mean DDD did not differ significantly (P = 0.2) between men (0.85 ± 0.38) and women (0.86 ± 0.35).

Across the age strata, the risk of hyponatremia due to thiazide exposure decreased with older age (Table 2). The lowest age category was not analyzed because it contained only 1 exposed case. Regarding BMI, the risk of thiazide-associated hyponatremia was highest in the lowest stratum: almost 8 times higher in the exposed than in the nonexposed group (HR, 7.91; 95% CI, 5.31–11.79). Risk was lowest in the highest BMI stratum (HR, 2.98; 95% CI, 2.05–4.32). In the eGFR strata, the highest risk of hyponatremia with thiazide exposure was seen within the highest eGFR quartile (suggesting better kidney function), but CIs overlapped (P = 0.2). Both age and BMI significantly modified thiazide effect (P = 0.01 and P = 0.001, respectively). The further adjusted model showed comparable results (Table 2). The risk of thiazide-associated hyponatremia was not influenced significantly by adjusting for these covariables.

Testing for proportionality showed there was significant interaction with follow-up time (P = 0.007). This was caused mainly by the category with the shortest follow-up and probably was caused by "depletion of susceptibles." This is a common phenomenon in a closed cohort study of adverse drug reactions because those with an adverse event often will stop the drug or change to another drug with the same indication.

Sensitivity analyses showed slightly lower risks, but results were comparable to those of our main analyses (data not shown).

Of the 522 cases of hyponatremia that occurred during the study period, 80 were moderate to severe (serum sodium <130 mmol/L). Mild, moderate, and severe hyponatremia were analyzed separately. The risk of mild hyponatremia was more than 4.5 times higher in individuals exposed to thiazides than in nonexposed individuals (HR, 4.66; 95% CI, 3.80–5.71). The risks of moderate (n = 32) and severe hyponatremia (n = 48) were comparable (HRs of 7.85 [95% CI, 5.30–11.63] and 8.28 [95% CI, 4.69–14.61], respectively).