Blood Urea Nitrogen-to-Albumin Ratio in Predicting Long-Term Mortality in Patients Following Coronary Artery Bypass Grafting: An Analysis of the MIMIC-III Database

Abstract

Background: This study examined the role of blood urea nitrogen-to-albumin ratio (BAR) in predicting long-term mortality in patients undergoing coronary artery bypass grafting (CABG).

Methods: In this retrospective cohort study, patients undergoing CABG were enrolled from the Medical Information Mart for Intensive Care III (MIMIC III) database. Patients were divided into the three groups according to the optimal cutoff values of BAR determined by X-tile software. The survival curve was constructed by the Kaplan-Meier method and multivariate Cox regression analysis was performed to explore the independent prognostic factors of 1- and 4-year mortality after CABG. The receiver operating characteristic (ROC) curves and the areas under the ROC curves (AUCs) were calculated to estimate the accuracy of BAR in predicting the outcomes. Subgroup analyses were also carried out.

Results: A total of 1,462 patients at 4-year follow-up were included, of which 933, 293, and 236 patients were categorized into the group 1 (≤ 6.45 mg/g), group 2 (>6.45 and ≤ 10.23 mg/g), and group 3 (>10.23 mg/g), respectively. Non-survivors showed an increased level of BAR at both 1- (p < 0.001) and 4-year (p < 0.001) follow-up compared with the survivors. The patients with a higher BAR had a higher risk of 1- and 4-year mortality following CABG (33.05 vs. 14.33 vs. 5.14%, p < 0.001 and 52.97 vs. 30.72 vs. 13.08%, p < 0.001, respectively). Cox proportional hazards regression model suggested a higher BAR as an independent risk factor of 1-year mortality (HR 3.904; 95% CI 2.559-5.956; P < 0.001) and 4-year mortality (HR 2.895; 95% CI 2.138-3.921; P < 0.001) after adjusting for confounders. Besides, the receiver operating characteristic (ROC) curves showed the better predictive ability of BAR compared to other grading scores at both 1- (0.7383, 95% CI: 0.6966-0.7800) and 4-year mortality (0.7189, 95% CI: 0.6872-0.7506). Subgroup analysis demonstrated no heterogeneous results of BAR in 4-year mortality in particular groups of patient.

Conclusion: This report provided evidence of an independent association between 1- and 4-year mortality after CABG and BAR. A higher BAR was associated with a higher risk of long-term mortality and could serve as a prognostic predictor in patients following CABG.

Keywords: MIMIC III database; albumin; blood urea nitrogen; coronary artery bypass grafting; mortality.

Figure 1

Flow diagram of study patient selection. ICU, intensive care unit; CABG, coronary artery bypass grafting; BUN, blood urea nitrogen.

Figure 2

Comparison of BAR levels between survivors and nonsurvivors at 1- (A) and 4-year (B) follow-up. BAR, blood urea nitrogen-to-albumin ratio.

Figure 3

Relationship between BAR and all-cause mortality in patients after CABG: (A) the mortality rate in each endpoint according to BAR levels, (B) The Kaplan–Meier survival curve of survival probability in patients with different BAR levels. P-value was calculated by log-rank test and indicated in the plot. BAR, blood urea nitrogen-to-albumin ratio; CABG, coronary artery bypass grafting.

Figure 4

The receiver operating characteristic curves of the predictive value of BAR, APS III, SOFA, and SIRS for 1- (A) and 4-year (B) all-cause mortality in patients after CABG. BAR, blood urea nitrogen-to-albumin ratio; APS III, acute physiology score III; SOFA, sequential organ failure assessment; SIRS, systemic inflammatory response syndrome; CABG, coronary artery bypass grafting.

Figure 5

The predictive value of prognostic models for long-term mortality in CABG patients. ROC curve for 1- (A) and 4-year (B) mortality. (C) Time-AUC curves of model 2 and BAR. CABG, coronary artery bypass grafting; ROC, receiver operating characteristic curve; AUC, area under the ROC curve; BAR, blood urea nitrogen-to-albumin ratio.