Association of four non-insulin-based insulin resistance surrogate markers with colorectal cancer risk: a large-scale prospective cohort study using the UK Biobank.

BACKGROUND: Colorectal cancer (CRC) remains a persistent global health challenge, with a rising incidence strongly linked to metabolic dysfunction. Insulin resistance (IR), a hallmark of metabolic dysregulation, is a well-established driver of tumorigenesis and cancer progression. However, the relative predictive value of distinct non-insulin-based surrogate markers of IR for CRC risk assessment remains poorly defined.

METHODS: This prospective study analyzed data from 385,206 participants in the UK Biobank. Non-insulin-based surrogate markers evaluated included the TyG index, TyG-BMI, the TG/HDL-C ratio, and METS-IR. Cox proportional hazards models estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for incident CRC. Restricted cubic splines (RCS) assessed dose-response relationships. Incremental predictive value and clinical utility were evaluated via C-index increments and decision curve analysis (DCA). Robustness was verified through extensive subgroup and sensitivity analyses.

RESULTS: Over a median follow-up of 13.8 years, 5,175 incident CRC cases were documented. All four IR surrogates were independently and positively associated with CRC risk (all P < 0.001). In the fully adjusted model, comparing extreme quartiles (Q4 vs. Q1), the HRs (95% CIs) were 1.16 (1.07-1.26) for the TyG index, 1.20 (1.11-1.31) for TyG-BMI, 1.09 (1.01-1.19) for the TG/HDL-C ratio, and 1.21 (1.11-1.32) for METS-IR. RCS revealed significant non-linear relationships for the TyG index and METS-IR, whereas linear trends were observed for the remaining markers in the fully adjusted model. Notably, adding METS-IR and TyG-BMI to the base model yielded the most significant improvements in discrimination (both C-index increments = 0.0015; P < 0.001). Furthermore, DCA confirmed sustained net clinical benefits across practical thresholds. Subgroup analyses revealed significant interactions between the IR surrogate markers and age, sex, and smoking status (P-interaction ≤ 0.05). Multiple sensitivity analyses, including excluding cases diagnosed within the first two years, confirmed the findings' robustness.

CONCLUSIONS: Elevated non-insulin-based IR surrogates are robust predictors of CRC risk. The BMI-integrated indices, METS-IR and TyG-BMI, demonstrated superior risk stratification performance. These low-cost, readily available tools may facilitate the early identification of high-risk individuals and inform evidence-based primary prevention strategies.