Sex- and Diabetes-Dependent Perioperative Model for End-Stage Liver Disease Trajectories Identify Distinct Hepatorenal Stress Phenotypes After Surgical Coronary Revascularization.

Background/

Objectives

Perioperative risk stratification in cardiac surgery is based mainly on static preoperative variables and therefore does not fully capture dynamic multiorgan responses to surgical stress. The Model for End-Stage Liver Disease (MELD) score, which integrates bilirubin, creatinine, and the international normalized ratio (INR), reflects hepatorenal function, but its perioperative dynamics remain insufficiently explored. This study aimed to characterize perioperative MELD trajectories in patients undergoing off-pump coronary artery bypass grafting (OPCAB) and to assess the influence of sex and diabetes mellitus on these changes and their clinical relevance.

Methods

This retrospective observational study included 111 patients undergoing elective OPCAB. MELD scores were assessed preoperatively (MELD0), on postoperative day 1 (MELD1), and on day 6 (MELD6). Dynamic indices of MELD change were calculated, including the early postoperative increase (ΔMELD01). The effects of sex and diabetes mellitus on MELD trajectories were analyzed using multivariable linear regression and generalized estimating equations. A high-surge phenotype was defined as the upper quartile of ΔMELD01.

Results

MELD increased significantly on postoperative day 1 and partially recovered by day 6 (p < 0.001). Female sex was independently associated with lower postoperative MELD values (β = -2.54, p < 0.001) and a smaller ΔMELD01, whereas diabetes mellitus was associated with a reduced MELD rise (β = -1.07, p = 0.028). Patients with a high-surge MELD phenotype had significantly longer hospitalization than those with a lower MELD response (12.8 ± 2.1 vs. 9.2 ± 1.2 days, p < 0.001).

Conclusions

Perioperative MELD trajectories reflect a dynamic hepatorenal stress response after OPCAB and may improve identification of clinically relevant physiological vulnerability.