Abstract
BACKGROUND: The Albumin-Bilirubin (ALBI) grade was initially used to assess liver reserve function in patients with cirrhosis and has since been applied in the prognostic evaluation of various diseases. This study explored the relationship between the ALBI grade and the prognosis of patients with pulmonary edema (PE). METHODS: We conducted a retrospective analysis with 1,562 ICU patients with pulmonary edema from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database. This study investigated the correlation between the ALBI grade and prognosis in patients diagnosed with acute or chronic pulmonary edema at different stages of the disease. Kaplan-Meier survival analysis, multivariate Cox regression, and ROC curves were used to evaluate the prognostic value of ALBI. Kaplan‒Meier survival curves and multivariate Cox proportional hazards regression models assessed the associations between ALBI grade and short-term and long-term prognoses. Receiver operating characteristic (ROC) analysis was used to evaluate the prognostic prediction ability, sensitivity, specificity, and area under the curve (AUC) of short-term and long-term ALBI grades of PE. Subgroup analyses were also conducted. The Boruta algorithm was used to assess the predictive ability of ALBI, and a predictive model was built using machine learning algorithms. Machine learning algorithms including Gradient Boosting Machine (GBM), Random Survival Forest (RSF), Lasso Cox Regression (Lasso_Cox), etc., were employed to build predictive models. RESULTS: This study included a cohort of 1,562 participants. Cox proportional hazards models revealed independent associations between ALBI grade and 30-, 90-, 180-day, and 1-year outcomes in PE patients before and after confounder adjustment. The survival curves indicated that patients with an ALBI grade of 3 (-1.39 to 3) had lower survival rates at 30, 90, and 180 days and 1 year. High ALBI grade (Grade 3) was independently associated with increased mortality at 30 days (HR = 47.4, 95% CI: 6.55-344), 90 days (HR = 8.56, 95% CI: 4.11-17.8), and 1 year (HR = 8.01, 95% CI: 4.58-14.0). The AUC values for ALBI in predicting 30-day and 1-year mortality were 0.741 and 0.700, respectively. Subgroup analysis revealed no significant interactions between ALBI grade and any subgroup. Elevated ALBI levels were associated with increased short-term and long-term all-cause mortality (ACM) in PE patients, suggesting that the ALBI grade may be an independent prognostic factor across disease stages. CONCLUSION: The ALBI grade is significantly associated with ACM and prognosis in PE patients, which provides a scientific basis for the development of precise treatment strategies.