Estimated glucose disposal rate outperforms other insulin resistance surrogate indices in predicting cardiometabolic disease and major adverse cardiovascular events in metabolic dysfunction-associated steatotic liver disease: insights from UK Biobank.

backgroundMetabolic dysfunction-associated steatotic liver disease (MASLD) has rapidly evolved into a pressing global health issue, exacting a formidable and escalating toll on public disease burden. Insulin resistance (IR) serves as a shared pathophysiological pathway for MASLD and cardiometabolic dysfunction. However, the associations between IR surrogate indices, cardiometabolic disease (CMD) and major adverse cardiovascular events (MACE) in MASLD, particularly the comparative predictive performance of these indicators, have yet to be fully elucidated.

methodsThis large-scale prospective study incorporated 127,195 and 149,402 individuals with MASLD from UK Biobank in CMD and MACE cohort, respectively. Kaplan-Meier analysis was employed to estimate CMD and MACE risks across different IR indices quartiles. Cox regression model and restricted cubic splines (RCS) curve were performed to investigate the associations between IR-related indices, CMD and MACE in MASLD, with threshold effect analysis detecting potential inflection points upon observed nonlinear relationships. Additionally, receiver operating characteristic (ROC) analysis and Harrell's C-index, along with net reclassification index (NRI) and integrated discrimination improvement (IDI), were utilized to compare the predictive capability of estimated glucose disposal rate (eGDR) and other IR indices. Furthermore, subgroup and sensitivity analysis were conducted to validate the robustness of primary findings.

resultsDuring a median follow-up time of 13.67 and 13.92 years, 29,089 CMD (22.87%) and 15,445 (10.34%) MACE occurred, respectively. Kaplan-Meier curves indicated the risks of CMD and MACE increased progressively in parallel with descending eGDR quartiles (log-rank test, P < 0.001). After multivariable adjustment, eGDR was significantly associated with incident CMD (per 1 SD, HR = 0.70, 95% confidence interval [CI] 0.69-0.72) and MACE (per 1 SD, HR = 0.82, 95% CI 0.80-0.84), which held true across subgroup and sensitivity analyses. RCS analysis revealed the nonlinear associations between eGDR, CMD and MACE, while turning points were identified at 5.377 and 5.215. Notably, eGDR outperformed other IR surrogate indices in predictions for CMD and MACE overall, as corroborated by all evaluation metrics (AUC, Harrell's C-index, NRI and IDI).

conclusionseGDR demonstrated significant relationships with CMD and MACE in MASLD, yielding comparatively better predictive performance across diverse IR surrogate indices, which suggested its potential clinical utility.