Prognostic Value of the Albumin-Adjusted Anion Gap for Short-Term Mortality in Pulmonary Embolism: A Retrospective Analysis Using the MIMIC-IV Database

Background Pulmonary embolism (PE) is a major cause of mortality with rising global incidence. The anion gap (AG) is commonly used to assess metabolic status but is influenced by hypoalbuminemia. The albumin-corrected anion gap (ACAG), which accounts for serum albumin levels, may offer more accurate prognostic insight. However, its role in PE prognosis remains unclear. Methods This retrospective study used the MIMIC-IV database to analyze 714 ICU patients with confirmed PE, stratified by 30-day survival status. Clinical variables, including ACAG (calculated as AG + 2.5 × [4.0 − albumin, g/dL]), were extracted. Associations between ACAG and clinical outcomes were evaluated using multivariable logistic regression, restricted cubic spline (RCS) modeling, and receiver operating characteristic (ROC) analysis. Results Among 714 patients, the 30-day mortality rate was 15.97%. Non-survivors had higher ACAG levels (20.71 vs. 17.39 mmol/L, P < 0.001) and lower albumin levels (2.8 vs. 3.2 g/dL, P < 0.001) than survivors. Each 1 mmol/L increase in ACAG was associated with a 16% increase in in-hospital mortality risk (adjusted OR = 1.16, P < 0.001). Patients in the highest ACAG quartile had a 3.44-fold higher mortality risk than those in the lowest quartile (P = 0.020). ACAG showed superior prognostic performance for 30-day mortality (AUC = 0.75) compared to AG (AUC = 0.70) and albumin (AUC = 0.35). RCS modeling identified 17.71 mmol/L as a risk threshold (P = 0.032), with stronger predictive value observed in patients with concurrent sepsis (P < 0.05). Conclusions ACAG is independently associated with increased disease severity and mortality in PE. It outperforms AG and albumin in predicting short-term outcomes and may serve as a clinically useful prognostic marker. Further prospective studies are warranted to validate its utility.