To evaluate the impact of phosphate supplementation therapy on the in-hospital mortality in septic patients with hypophosphatemia, and to identify the optimal serum phosphorus threshold for sepsis management.

The relevant clinical data of patients with sepsis were collected from the Medical Information Mart for Intensive Care Ⅳ (MIMIC-Ⅳ) database and Electronic Intensive Care Unit-Collaborative Research Database. The logistic regression model was used to assess the effect of phosphate supplementation therapy on in-hospital mortality in septic patients with hypophosphatemia. The blood phosphorus level was divided into six intervals, including significant hypophosphatemia (< 15 mg/L), mild hypophosphatemia (15-25 mg/L), normal low value (> 25-35 mg/L), normal high value (> 35-45 mg/L), mild hyperphosphatemia (> 45-55 mg/L), and significant hyperphosphatemia (> 55 mg/L). The proportion of exposure time of septic patients in each blood phosphorus interval was calculated by the time-weighted method, and the optimal blood phosphorus threshold was determined by the logistic regression model.

A total of 29 729 patients with sepsis were included, 250 856 blood phosphorus test data were analyzed, and 4 849 patients developed hypophosphatemia (serum phosphorus < 27 mg/L). Logistic regression analysis showed that septic patients with hypophosphatemia who received phosphate supplementation were significantly associated with a higher in-hospital mortality rate [odds ratio (OR) = 1.509, 95% confidence interval (CI) (1.203, 1.895), P < 0.001]. In the subgroups excluding patients with acute kidney injury (AKI) and/or refeeding syndrome (RFS), this association remained significant (all P < 0.05). There were 24 051 septic patients who had at least one recorded blood phosphorus test every 24 hours during the ICU period. Logistic regression analysis showed that septic patients with mild hypophosphemia [OR = 0.574, 95%CI (0.473, 0.697), P < 0.001] and those in the normal low value range [OR = 0.411, 95%CI (0.352, 0.480), P < 0.001] had a significantly reduced risk of in-hospital mortality.

Phosphate supplementation therapy cannot improve the prognosis of septic patients with hypophosphatemia. The optimal serum phosphorus level for septic patients is 15-35 mg/L, which is lower than the traditional reference range.