Predictors of 24-Hour Sodium Correction in Moderate to Severe Hyponatremia

Background: Moderate to severe hyponatremia is a common occurrence in acute-care settings and must be managed with caution due to the risk of complications from overly rapid correction. This study assessed the prevalence and predictors of 24-hour serum sodium change in a large U.S. health system cohort. Methods: Using an acute-care data warehouse (January 2011 to June 2024; 8.6 million encounters), we identified adult inpatients with an initial serum sodium level of ≤125 mEq/L who had a repeat sodium measurement within 24 hours (n = 6,393 patients). Data collected included demographics, comorbidities, and intravenous fluid use. Sodium correction after 24 hours (ΔNa) was classified as: Undercorrection: <6 mmol/L, Appropriate correction: 6-12 mmol/L, and Overcorrection: >12 mmol/L. Results: The median age of patients was 64 years; 54% were female and 84% were White. Appropriate sodium correction occurred in 36% of patients (n = 2,307), while 54% experienced overcorrection (n = 3,427), and 10% remained undercorrected (n = 659). Age and Charlson Comorbidity Index (CCI) were inversely associated with the rate of sodium correction, ΔNa -0.0729 mmol/L per year of age (p < 0.001) and ΔNa -0.1827 mmol/L per CCI point (p < 0.001). Lower initial sodium levels were predictive of less rapid correction, ΔNa -0.408 mmol/L per mmol/L (p < 0.001). Administration of normotonic and hypertonic IV fluids had no significant impact on sodium correction rates. In contrast, hypotonic fluid administration was weakly associated with a higher correction rate (ΔNa 0.001 mmol/L change per mL; p < 0.001), though this likely reflects correlation rather than causation. Multivariate linear regression yielded modest performance (Train R2= 0.3134; Test R2 = 0.3163; Train MSE = 22.39; Test MSE = 22.90). A nonlinear model using gradient boosting showed slightly better predictive performance (Train R2 = 0.5370; Test R2 = 0.3615; Train MSE = 15.0953; Test MSE = 21.3844). Conclusion: Inappropriate sodium correction remains a significant issue in inpatient care. Only 36% of patients achieved guideline-concordant sodium correction (6-12 mmol/L) within the first 24 hours. Predicting the rate of sodium correction remains challenging. These results underscore the need for advanced predictive models and standardized treatment protocols to enhance hyponatremia management.