Serum Anion Gap Level Predicts All-Cause Mortality in Septic Patients: A Retrospective Study Based on the MIMIC III Database.

Purpose: Sepsis is a significant threat in the intensive care unit (ICU) worldwide because it has high morbidity and mortality rates. Early recognition and diagnosis of sepsis are essential for the prevention of adverse outcomes. The present study aimed to quantitatively assess the association between serum anion gap (AG) levels and 30- and 90-day all-cause mortality among sepsis patients.

Methods: Clinical data of patients diagnosed with sepsis were extracted from the Medical Information Mart for Intensive Care III (MIMIC III) database. Kaplan-Meier curves and Cox proportional hazards models were used to evaluate the association between serum AG levels and all-cause mortality. A receiver operating characteristic (ROC) curve was drawn to quantify the efficacy of using the serum AG level to predict all-cause mortality.

Results: A total of 3811 patients were included in the study. The Kaplan-Meier curves showed that patients with higher serum AG levels had a shorter survival time than those with lower levels. Serum AG levels were found to be highly effective in predicting all-cause mortality secondary to sepsis (30-day: AUROC = 0.703; 90-day: AUROC = 0.696). The Cox regression model further indicated that the serum AG level was an independent risk factor for 30- and 90-day mortality in sepsis (HR 3.44, 95% CI 2.97-3.99 for 30-day; HR 3.17, 95% CI 2.76-3.65 for 90-day, P < 0.001 for both).

Conclusions: High serum AG may be considered as an alternative parameter for predicting the death risk in sepsis when other variables are not immediately available. Prospective large-scale studies are needed to support its predictive value in the clinic.