Non-thiazide diuretics and hospitalization due to hyponatremia: a population-based case-control study.

Background Diuretics are often implicated in hyponatremia. While thiazides constitute one of the most common causes of hyponatremia, data on loop-diuretics and potassium-sparing agents are limited and partly conflicting. The objective of this investigation was to study the association between use of different types of non-thiazide diuretics and hospitalization due to hyponatremia. Design & methods This was a register-based case-control study on the adult Swedish population. By linking national registers, patients hospitalized with a principal diagnosis of hyponatremia (n=11,213) from 1 October 2005 through 31 December 2014 were compared with matched controls (n=44,801). Multivariable logistic regression, adjusted for multiple confounders, was used to analyze the association between use of diuretics and hyponatremia. In addition, newly initiated (≤90 days) and ongoing use was examined separately. Results Adjusted odds ratios (aORs) (95% CI) were 0.61 (0.57-0.66) for use of furosemide, 1.69 (1.54-1.86) for use of amiloride and 1.96 (1.78-2.18) for use of spironolactone and hospitalization due to hyponatremia. For newly initiated therapy, aORs ranged from 1.23 (1.04-1.47) for furosemide to 3.55 (2.75-4.61) for spironolactone. The aORs for ongoing use were 0.52 (0.47-0.57) for furosemide, 1.62 (1.47-1.79) for amiloride and 1.75 (1.56-1.98) for spironolactone. Conclusions Ongoing use of furosemide was inversely correlated with hospitalization due to hyponatremia, suggesting a protective effect. Consequently, if treatment with furosemide precedes development of hyponatremia by some time, other causes of hyponatremia should be sought. Spironolactone and amiloride may both contribute to hyponatremia; this effect is most prominent early in treatment.