Prognostic Value of Serum Albumin Level in Critically Ill Patients: Observational Data From Large Intensive Care Unit Databases

BackgroundDecreased serum albumin level (SAL) is associated with adverse clinical outcomes. We designed the present study to further assess the prognostic value of SAL in critically ill patients based on data from large intensive care unit (ICU) databases. MethodsThis retrospective cohort study recruited 18,353 patients from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. Restricted cubic splines (RCS) were performed to visualize the association of SAL at admission with ICU and hospital mortalities. The prognostic value of SAL was analyzed using logistic regression models and receiver operating characteristic (ROC) curves in overall patients and subgroups. ResultsRestricted cubic splines revealed rapid increasing risks in ICU and hospital mortalities when SAL declined to below 30 g/l. Patients with SAL <30 g/l (n = 6,069) had higher ICU (13.7% vs. 6.4%, p < 0.001) and hospital (23.9% vs. 10.7%, p < 0.001) mortalities than those with SAL >= 30 g/l. Multivariable logistic regression model revealed that SAL <30 g/l independently correlated with higher risks of both ICU (odds ratio [OR]: 1.20, 95% confidence interval [CI]: 1.07-1.36) and hospital (OR: 1.51, 95% CI: 1.37-1.66) mortalities. However, the association diminished in patients with cirrhosis (OR: 1.16, 95% CI: 0.91-1.49 for ICU mortality; OR: 1.21, 95% CI: 1.00-1.48 for hospital mortality). ROC curves revealed a poor performance of SAL in predicting mortalities, both in overall patients and in those with cirrhosis. ConclusionsDecreased SAL is associated with increased risk of mortality. However, it possesses low sensitivity and specificity for outcome prediction in critically ill patients, especially in those with cirrhosis.