Cumulative exposure and dynamic changes of the C-reactive protein–triglyceride–glucose index in relation to cardiometabolic multimorbidity risk among individuals with metabolic dysfunction-associated steatotic liver disease: a nationwide prospective cohort study

Metabolic dysfunction-associated steatotic liver disease (MASLD) carries a predisposition to cardiometabolic multimorbidity (CMM). The C-reactive protein-triglyceride-glucose index (CTI) captures inflammatory and metabolic derangements in a single metric, though whether its cumulative exposure or longitudinal change patterns predict CMM in MASLD has not been investigated. We therefore examined how cumulative CTI, CTI change cluster, and single-timepoint indices (CTI and TyG) relate to incident CMM in middle-aged and older Chinese adults with MASLD. We drew on 2,165 MASLD participants from the China Health and Retirement Longitudinal Study (2011–2020). CTI was derived at baseline (2012) and follow-up (2015); cumulative CTI was computed as the product of mean CTI and exposure duration. Three change phenotypes were identified through K-means clustering. Cox regression, restricted cubic splines, and Kaplan–Meier methods were applied to assess associations. Time-dependent AUC, net reclassification improvement (NRI), and integrated discrimination improvement (IDI) served to compare predictive utility. During a mean follow-up of 102.2 months, 350 participants developed CMM (16.2%). After full adjustment, each unit increase in cumulative CTI corresponded to a 13% higher CMM risk (HR 1.13, 95% CI 1.08–1.19). Relative to the low-stable phenotype, moderate-ascending (HR 1.37, 95% CI 1.07–1.76) and high-ascending (HR 2.02, 95% CI 1.50–2.71) phenotypes showed stepwise risk elevation. The highest cumulative CTI tertile carried 60% excess risk versus the lowest (HR 1.60, 95% CI 1.22–2.10). Baseline CTI predicted CMM in both continuous and categorical analyses; TyG was significant as a continuous measure (HR 1.28, 95% CI 1.08–1.50, P = 0.003) but lost graded significance when categorized into tertiles ( P for trend = 0.088). Dose–response curves were linear across all three exposures. Time-dependent AUC analyses demonstrated cumulative CTI superiority at all evaluated time points (48, 84, and 108 months). Cumulative CTI provided the relatively greatest incremental predictive value (NRI 0.175, P = 0.003; IDI 0.011, P < 0.001), outperforming both CTI and TyG. Among individuals with MASLD, cumulative CTI and dynamic CTI change cluster were independently associated with incident CMM and exhibited superior predictive performance compared with single-timepoint measures. These findings suggest that longitudinal metabolic-inflammatory monitoring may enhance CMM risk stratification in this population.