Abstract
BACKGROUND: The degree to which the triglyceride-glucose (TyG) index might interact with colorectal cancer (CRC) incidence and mortality is undefined. This systematic analysis was conducted through meta-analyses and large-scale databases. METHODS: A meta-analysis was conducted through database search up to April 1, 2025, and articles investigating the incidence of CRC with clearly reported TyG index values, focusing on their dose-response relationship, were included. To validate the findings, data from the 1999-2018 National Health and Nutrition Examination Survey (NHANES) were utilized to explore links of TyG levels with CRC-linked mortality. Analyses involving restricted cubic splines (RCSs), weighted logistic regression (WLR), and multivariate Cox proportional hazards models were performed. Receiver operating characteristic (ROC) analyses explored predictive potential. Subgroup evaluation was subsequently conducted for detecting the susceptible populations. RESULTS: Multivariate logistic regression showcased a greater CRC incidence among individuals in quartiles two, three, and four of TyG and TyG-body mass index (TyG-BMI) (95% confidence intervals (CIs): 1.14-1.5, 1.07-1.45; 1.39-1.83, 1.11-1.64; and 1.60-2.12, 1.19-1.93, respectively; P < 0.001). In contrast, multivariate Cox regression indicated a significant increase in CRC-related mortality only in the second quartile of TyG (95% CI: 1.22-7.47; P < 0.05). Analysis of the RCS curves demonstrated that the incidence of CRC displayed a nonlinear association with TyG (P = 0.045 for nonlinearity) but a positive linear association with TyG-BMI (P = 0.385 for nonlinearity). The TyG and associated indices did not exhibit any obvious dose-response association with CRC-related mortality (P > 0.05). ROC analyses exploring CRC risk revealed that TyG-BMI outperformed all indicators (area under the curve (AUC) = 0.71). Subgroup analysis revealed statistically significant links of CRC incidence with both TyG-BMI and female sex. CONCLUSION: TyG and TyG-BMI may function as dependable markers for predicting CRC likelihood, and TyG-BMI outperformed all other predictors considered herein. However, TyG and associated indices showed no significant interrelationships with the mortality of CRC.