Prognostic Models for Predicting Coronary Heart Disease Risk in Patients with Type 2 Diabetes Mellitus: A Systematic...

Area of Science:

• Cardiology
• Endocrinology
• Biostatistics

Background:

• Type 2 diabetes mellitus (T2DM) significantly elevates coronary heart disease (CHD) risk.
• Existing CHD prediction models for T2DM populations lack generalizability and transportability.
• Evaluating multivariable prognostic models is essential for accurate CHD risk assessment in T2DM.

Purpose of the Study:

• To systematically identify and evaluate multivariable prognostic models for predicting CHD in adults with T2DM.
• To assess the performance, heterogeneity, and applicability of existing CHD prediction models.
• To inform future development of reliable CHD prediction tools for T2DM patients.

Main Methods:

• PRISMA-guided systematic review and meta-analysis of prognostic models.
• Extraction of model characteristics and performance using CHARMS and TRIPOD-SRMA frameworks.
• Assessment of bias and applicability using the PROBAST tool.

Main Results:

• Thirteen studies included diverse models (clinical, imaging, omics-augmented).
• Pooled AUC was 0.69, indicating moderate discrimination but high heterogeneity (I²=97.4%).
• Machine learning and omics models showed promise but suffered from small sample sizes and limited validation.

Conclusions:

• Prognostic models for CHD in T2DM exhibit moderate-to-good discrimination but significant heterogeneity and miscalibration.
• Clinical utility hinges on external validation and local recalibration, especially for complex models.
• Future research requires standardized outcomes, calibration reporting, and multimodal, transportable models.