Abstract
KEY POINTS: Slow hypernatremia correction rates (≤0.50 mEq/L per hour) are associated with lower in-hospital mortality than fast rates (>0.50 mEq/L per hour). Fast hypernatremia correction is associated with lower odds of discharge to nursing facilities or hospice than slow correction. Hypernatremia severity, age, and kidney function do not modify associations between correction rates, hospital mortality, and disposition. BACKGROUND: Current recommendations for limiting hypernatremia correction rates to avoid cerebral edema in adults are supported by limited evidence. We explored the associations between rate of hypernatremia correction in hospitalized adults, in-hospital mortality, and discharge disposition. METHODS: Using a large, multicenter database, we analyzed 37,913 hospitalized adults with hypernatremia on admission. For the primary analysis, hypernatremia correction rates were categorized as slow (≤0.50 mEq/L per hour) or fast (>0.50 mEq/L per hour). Propensity score (PS) weighting and PS stratification were used to adjust for confounders. In secondary analyses, the results were stratified by initial sodium concentration, age, and initial eGFR. In a sensitivity analysis, correction rates were categorized as <0.40, 0.40–0.60, or >0.60 mEq/L per hour. RESULTS: Most (89%) patients experienced slow hypernatremia correction. In PS-weighted analyses, slow correction was associated with overall lower in-hospital mortality (adjusted odds ratio [aOR], 0.63; 95% confidence interval [CI], 0.59 to 0.67) but higher odds of discharge to hospice (aOR, 1.57; 95% CI, 1.38 to 1.78) or nursing facilities (aOR, 1.60; 95% CI, 1.52 to 1.69) than fast (reference) correction rates. After stratification by initial hypernatremia severity, age, and kidney function at admission, the associations between slow versus fast correction, in-hospital mortality, and discharge disposition were largely preserved without clear signals of effect modification by subgroup. When categorizing hypernatremia correction rates into three groups, <0.40 versus >0.60 mEq/L per hour (analogous to slow versus fast, respectively) continued to be independently associated with lower in-hospital mortality but higher rates of discharge to nursing facilities or hospice. CONCLUSIONS: In this analysis, the rate of hypernatremia correction was independently associated with opposing effects on survival and a favorable discharge disposition. Our findings suggest that balancing the risks and benefits of different dysnatremia correction rates should consider not only mortality but also patient-centered outcomes such as discharge disposition.