Development and validation of a prediction model for hypernatremia in patients with severe burns

Hypernatremia is a prevalent and serious electrolyte disturbance in patients with severe burns, associated with significantly increased mortality and prolonged ICU stays. Early identification of high-risk patients could enable timely preventive interventions; however, practical tools for individualized risk stratification are currently lacking. This study aimed to develop and validate a nomogram and a web-based calculator for the early prediction of hypernatremia in patients with severe burns. A total of 592 patients with burns covering ≥30% total body surface area were retrospectively included and randomly assigned to a training cohort (n = 415) and an internal validation cohort (n = 177). Six independent predictors were identified: TBSA, mechanical ventilation, blood urea nitrogen ≥10 mmol/L, sepsis, glucose ≥9 mmol/L, and inhalation injury. The nomogram achieved an apparent C-index of 0.929 and a bootstrap-corrected C-index of 0.923 in the training cohort, and a C-index of 0.953 in the internal validation cohort. Using dual cutoffs of 90 and 135 points, the low cutoff yielded negative predictive values of 96.4% and 98.1%, and the high cutoff yielded positive predictive values of 75.4% and 65.3% in the training and internal validation cohorts, respectively. The nomogram and web-based calculator provide accurate, individualized risk prediction, facilitating early identification of high-risk individuals and supporting preventive strategies.