Hyponatremia can be associated with substantial morbidity and mortality. Identification of elements associated with the development of symptomatic hyponatremia is important in determining preventive strategies. This research sought a comprehensive list of risk factors for severe hyponatremia by studying a large ambulatory population with data on demographics, medication use, and chronic diseases.


The study population was derived from enrollees in a large health maintenance organization in a vertically integrated health system. A search of computerized medical records from January 1, 1997, through November 30, 2007, was undertaken for health plan members with a serum sodium below 120 mEq/L Adults with prescription medication coverage by the health plan were considered eligible for inclusion. Persons were excluded if they developed hyponatremia while inpatients or had a serum glucose greater than 400 mg/dL at the time of measurement. Each case was matched with 19 health plan members with prescription benefit coverage and a normal outpatient serum sodium. Previously published articles on hyponatremia influenced the choice of variables to examine. Conditional logistic regression was used to examine bivariate and multivariate relationships.


A total of 316 patients with severe hyponatremia were each matched with 19 control health plan members for a total of 6320 subjects for analysis. Table 1 lists the results of bivariate analysis. Neither female sex nor antide‐pressant use was associated with hypona‐tremia. Table 2 lists the significant factors found in multivariate analysis. Patients taking antipsychotics had triple the risk, and those on thiazides increased their risk by 55%. Thyroid hormone use decreased risk by 86%. Chronic conditions like hypothy‐roidism increased risk by 370%, liver disease by 341%, alcoholism by 283%, lung cancer by 260%, and heart failure by 127%.

Table 1. Bivariate Analysis

Table 2: Multivariate Analysis


Severe hyponatremia is associated with a number of demographic markers, medications, and chronic disease states. Knowing the relative magnitude of predisposing factors might aid in the prevention of severe hyponatremia. Possibly, medications like thiazides and antipsychotics could be avoided in patients at high risk. Alternatively, closer monitoring of electrolytes in hyponatremia‐prone patients could detect changes in serum sodium before they become dangerous. Further research is needed to establish effective strategies for prevention of severe hyponatremia.

Author Disclosure:

G. Buran, none.