The prognostic value of the direct bilirubin to albumin ratio

The direct bilirubin-to-albumin ratio (DBAR) has emerged as a reliable prognostic marker for assessing 28-day mortality in critically ill patients with cirrhosis, as demonstrated by a study utilizing the MIMIC-IV database. Cirrhosis patients in ICUs face high mortality rates, and traditional models like MELD and Child-Pugh have limitations due to complexity and subjective parameters. DBAR, which combines two key liver function markers—direct bilirubin (reflecting hepatocyte injury) and albumin (indicating liver synthetic ability)—offers a simpler, objective, and accessible alternative.

The study analyzed data from 509 adult cirrhotic patients and found that elevated DBAR values were strongly associated with worse outcomes. Patients with DBAR ≥ 4 were classified as high-risk and demonstrated significantly lower survival rates (56.9% vs. 18.4% mortality within 28 days). Multivariate Cox regression analysis confirmed DBAR as an independent predictor of mortality (HR 1.16), alongside age, lactate, INR, and vasoactive medication use. Kaplan-Meier survival curves and nonlinear restricted cubic spline analysis further validated DBAR’s prognostic accuracy.

DBAR showed good predictive performance (AUC 0.702), comparable to the MELD score (AUC 0.744) and superior to albumin alone (AUC 0.549). Machine learning ranked DBAR among the top predictors of cirrhosis prognosis, alongside lactate and BUN. External validation using the eICU-CRD database confirmed its reproducibility (AUC ≈ 0.71). DBAR’s predictive effect was consistent across subgroups, though variations were noted in patients with ascites and hepatorenal syndrome.

Clinically, elevated DBAR signals severe hepatic dysfunction, inflammation, and malnutrition, making it a practical and cost-effective biomarker for bedside mortality risk assessment in ICU settings.