Study on the correlation between SPISE-IR, METs-IR, TyG and short-term adverse cardiovascular events in patients with heart failure and diabetes mellitus

Objective Heart failure (HF) patients with type 2 diabetes mellitus (T2DM) are at high risk of short-term major adverse cardiovascular events (MACE). Insulin resistance (IR) is a key shared mechanism, yet simple prognostic markers remain limited. This multicenter retrospective cohort study compared three IR surrogate indices (SPISE-IR, METs-IR, and TyG) for predicting 18-month post-discharge MACE in hospitalized HF-T2DM patients, aiming to identify practical tools for risk stratification. Methods A total of 477 hospitalized HF-T2DM patients from two tertiary hospitals in Zhejiang (2018–2024) were enrolled according to unified criteria. Baseline clinical, laboratory, echocardiographic, and medication data were collected to calculate SPISE-IR, METs-IR, and TyG. The primary outcome was 18-month post-discharge composite MACE. Associations were evaluated using Cox regression and RCS, while predictive performance was assessed by ROC and PR curves, DCA, NRI, IDI, and subgroup analyses. Results Of 477 patients, 191 (40.04%) experienced MACE over 18 months, including 3 cardiovascular deaths. Patients with MACE were older, had lower LVEF, worse renal function, and higher NT-proBNP (all p < 0.05). On multivariable analysis, SPISE-IR (HR 9.31, 95% CI 5.96–14.55, p < 0.001) and METs-IR (HR 1.07, 95% CI 1.05–1.09, p < 0.001) were associated with MACE, whereas TyG was not (HR 1.19, 95% CI 0.95–1.49, p = 0.127). SPISE-IR and METs-IR showed significant nonlinear dose–response relationships and yielded AUCs of 0.790 and 0.801, respectively, versus 0.539 for TyG. Higher SPISE-IR and METs-IR quartiles were associated with progressively higher cumulative event risk (log-rank p < 0.001), whereas no significant between-quartile difference was observed for TyG. Adding either index to the baseline model significantly improved NRI and IDI (all p < 0.05); SPISE-IR showed numerically larger NRI and IDI improvements than METs-IR. Subgroup analyses showed a consistent SPISE-IR association, whereas METs-IR exhibited effect modification by age, sex, smoking status, and hypertension; TyG remained largely non-significant. Heart failure rehospitalization accounted for 138 of 191 MACE (72.3%). After exclusion of 24 recurrent-angina events, SPISE-IR (HR 10.83, 95% CI 6.71–17.45) and METs-IR (HR 1.07, 95% CI 1.06–1.09) remained associated with the strict MACE endpoint (both p < 0.001). After 1,000 bootstrap resamples, optimism-corrected AUCs were 0.818 for the clinical + SPISE-IR model and 0.806 for the clinical + METs-IR model. In center-based external validation, their AUCs were 0.804 and 0.806, respectively. Conclusion SPISE-IR and METs-IR are associated with 18-month MACE risk in hospitalized HF-T2DM patients and show better predictive performance than TyG. These indices may provide additional value for risk stratification in this population.