Progressive kidney disease may not alter the association of hyponatremia with mortality.

Abstract

Hyponatremia is a common electrolyte disorder and a prognostic marker for mortality. We hypothesize that in advanced chronic or acute kidney disease, hyponatremia is not independently associated with mortality because of the contribution of kidney failure to its pathophysiology. Clinical Looking Glass, Montefiore's clinical database, was used to build a cohort of all patients hospitalized between January of 2009 and December of 2011. A chronic kidney disease (CKD) group and an acute kidney injury (AKI) group were defined based on GFR measurements during and before index hospitalization. Cox regression models assessed the hazard for death for those with community acquired hyponatremia as compared to those without hyponatremia, stratified by stage of kidney disease within each cohort. Forty-four thousand four hundred and seventy-six patients were studied. Forty six percent (46.2%) of subjects were in the CKD cohort and 53.8% were in the AKI cohort. Hyponatremic patients were older, and had a higher prevalence of CKD and AKI. A total of 7,934 subjects died (17.8%) during 22months of follow-up. In CKD and AKI cohorts, hyponatremia, age, race, illness severity and Charlson score were associated with mortality. Hyponatremia had similar hazard ratios (HR) across kidney disease stages despite loss of statistical significance in later stages due to smaller sample size. The association between community acquired hyponatremia and mortality showed consistent HRs across progressive stages of CKD and AKI suggesting that the contribution of tubular dysfunction to hyponatremia in advanced kidney disease does not alter this association.