Role of the Serum Creatinine to Albumin Ratio in the Evaluation short- and long-term all-cause mortality of Patients with Aortic Valve Replacement: A Retrospective Cohort Study

Background: The identification of novel biomarkers has significantly enhanced prognostic capabilities in the context of cardiovascular diseases. Among these emerging markers, the Serum creatinine-to-albumin ratio (CAR) has garnered increasing attention as a potential prognostic indicator across a variety of clinical settings. To our knowledge, the association between short- and long-term all-cause mortality in patients with aortic valve replacement (AVR) and the CAR has not been investigated. This study discusses the role of CAR in the evaluation of patients with AVR. Methods: We performed a retrospective analysis of 700 patients who underwent AVR and whose data were extracted from the MIMIC-IV database. The main purpose is to evaluate all-cause mortality in different periods. We extracted demographic baseline data, vital signs, laboratory tests, and other relevant information from the MIMIC-IV database. Machine learning techniques were employed to select features based on the 28-Day all-cause mortality outcome of the patients. The X-tile software was used to determine the optimal threshold for the CAR. Cox regression analyses were used to investigate the relationship between the CAR and all-cause mortality. Additionally, ROC curve analysis was conducted to evaluate the predictive performance of different indicators for the outcome. Additionally, subgroup analyses were conducted. Results: Our analysis of 700 patients from the MIMIC-IV database who underwent aortic valve replacement revealed that the CAR is a significant predictor of 1-year all-cause mortality. The CAR ideal threshold, determined by X-tile software, was 0.43. LASSO regression, identified CAR as one of the important features in mortality prediction models. Restricted cubic spline analysis demonstrated a significant nonlinear association between the CAR and both 28-Day, 90-Day and 1-year mortality. Cox regression analysis confirmed a dose-dependent increase in all the periods mortality risk with the higher CAR groups. Kaplan-Meier survival analysis showed the lowest survival probability in the higher CAR group. ROC curves indicated that the CAR had a higher AUC for the prediction of 1-year mortality (AUC 0.655) than the other indicators did. These results suggest that the CAR is a robust and independent predictor of mortality in critically ill patients with AVR. Conclusions: Our findings suggest that the CAR holds significant promise as a prognostic marker for 1-year mortality in patients undergoing AVR. It can serve as a tool for risk stratification and prognostic assessment in AVR patients.