Association of C-reactive protein-triglycerides-glucose index-waist-to-height ratio with mortality in metabolic dysfunction-associated steatotic liver disease (MASLD): evidence from two cohort studies

Abstract Background MASLD significantly increases the risk of cardiovascular events and all-cause mortality (ACM). The impact of the C-reactive protein-triglyceride-glucose index (CTI) and its derivatives on mortality in the MASLD population remains unclear. This study aimed to investigate their associations with mortality in the MASLD population and to compare their predictive value. Methods This cohort study was conducted in adult population with MASLD from the NHANES. Multivariable Cox regression was employed to assess the associations of CTI and its derived indices with the risks of ACM and cardiovascular mortality (CVM). Further analyses included Kaplan-Meier curves, smooth curve fitting, receiver operating characteristic (ROC) analysis, and exploratory statistical decomposition of the observed association (neutrophil-to-lymphocyte ratio, NLR, and systemic inflammation response index, SIRI). Subgroup and sensitivity analyses were performed to verify the robustness of the findings. Results A total of 5296 participants with MASLD were included in the study, with a median follow-up duration of 154 months. During this period, 1286 ACM and 410 CVM were recorded. The derivative index of CTI (CTI-waist-to-height ratio, CTI-WHtR) exhibited the strongest association with mortality risk compared to CTI alone. After adjusting for multiple confounding factors, elevated CTI-WHtR was significantly associated with an increased risk of ACM (HR = 1.79 per 1-unit increase, 95% CI 1.57–2.04, P < 0.0001; HR = 1.43 per 1-SD increase, 95% CI 1.32–1.55, P < 0.0001) and CVM (HR = 1.90 per 1-unit increase, 95% CI 1.50–2.40, P < 0.0001; HR = 1.48 per 1-SD increase, 95% CI 1.28–1.71, P < 0.0001). Kaplan-Meier curves indicated that higher CTI-WHtR levels were associated with reduced survival rates ( P < 0.0001). Smooth curve fitting revealed a nonlinear relationship between the CTI-WHtR index and mortality risk in the MASLD population. Exploratory statistical decomposition of the observed association showed that NLR and SIRI statistically accounted for 15.08% and 16.59% of the observed association between CTI-WHtR and ACM, and 15.98% and 16.60% of the association with CVM, respectively (all P < 0.0001). The cross-cohort analyses were replicated using representative data from the CHARLS. Conclusions The positive association between CTI-WHtR and mortality risk in the MASLD population was robust in both cohorts, but its independent discriminatory performance was generally weak, and its relative discriminatory performance varied across populations, lacking a stable cross-population advantage.