A non-linear association between AST/ALT ratio and 28-day mortality in critically ill elderly: evidence from a multicenter study

Abstract

Abstract With the global aging population, the proportion of elderly patients in ICU continues to rise. The ratio of aspartate aminotransferase to alanine aminotransferase (AST/ALT) has been linked to unfavorable outcomes in several diseases, but its value in critically ill elderly patients remains unclear. Based on the eICU Collaborative Research Database, this study included 22,361 critically ill elderly patients aged ≥ 65 years. The connection between the AST/ALT ratio and 28-day all-cause mortality was assessed using multivariate Cox regression analysis, and generalized additive models were employed to investigate non-linear relationships. A mediation analysis was undertaken to investigate the mediating impact of white blood cell count (WBC). Following a comprehensive adjustment for confounding factors, the AST/ALT ratio showed a non-linear association with 28-day mortality, with an inflection point of 2.161. When the ratio was < 2.161, Every additional unit was linked to a 43.9% greater risk of mortality (HR 1.439, P < 0.0001); when the ratio was > 2.161, this association was not statistically significant. According to Kaplan–Meier analysis, a significantly lower survival probability was observed in patients with higher AST/ALT ratios. Mediation analysis revealed that WBC partially mediated the link between AST/ALT ratio and mortality, accounting for 4.17% of the total effect. In critically ill elderly patients, the AST/ALT ratio is independently linked to 28-day all-cause mortality, demonstrating a non-linear relationship. WBC is an important mediator of this association.