Estimating the risk of cardiovascular outcomes and all‐cause mortality in individuals with type 2 diabetes: Validation of the UKPDS outcomes model using TECOS and EXSCEL data

TL;DR

This study validates a diabetes risk model using two clinical trials, finding it accurately predicts cardiovascular risks but underestimates outcomes in one trial.

Contribution

The study validates the UKPDS outcomes model for cardiovascular risk in type 2 diabetes using real-world trial data.

Findings

The UKPDS-OM2 model accurately simulated relative risks in both TECOS and EXSCEL trials.

The model underestimated primary outcomes by about one-third in the EXSCEL trial arms.

The model's accuracy was confirmed for cardiovascular death, MI, stroke, and all-cause mortality.

Abstract

To evaluate United Kingdom Prospective Diabetes Study Outcomes Model version 2 (UKPDS‐OM2) cardiovascular risk estimates for people with type 2 diabetes using TECOS and EXSCEL data.

We compared model‐simulated and TECOS and EXSCEL observed event rates for the composite outcome of cardiovascular death, myocardial infarction (MI) or stroke, each component, and all‐cause mortality. Risk factors analysed were age, sex, race/ethnicity, height, diabetes duration, atrial fibrillation, albuminuria, and baseline plus annual measures of smoking, HDL‐cholesterol, LDL‐cholesterol, weight, systolic blood pressure, HbA1c, heart rate, white cell count, haemoglobin, and estimated glomerular filtration rate. Other factors were prior ischemic heart disease, heart failure, amputation, blindness, kidney failure, MI, stroke, and diabetic foot ulcer.

Median follow‐up was 3.0 and 3.2 years for the 14 671 TECOS and 14 752 EXSCEL participants, respectively. The primary outcome occurred in TECOS for 839 (11.4%) and 851 (11.6%) sitagliptin and placebo group participants, respectively (hazard ratio [HR] 0.98, 95%CI 0.89–1.08), compared with UKPDS‐OM2 simulated events of 776 (10.6%) and 778 (10.6%), respectively (relative risk 1.00). The primary outcome occurred in EXSCEL for 839 (11.4%) and 905 (12.2%) once‐weekly exenatide and placebo group participants, respectively (HR 0.92, 95%CI 0.84–1.01), compared with UKPDS‐OM2–simulated events of 579 (7.9%) and 593 (8.0%), respectively (relative risk 0.98).

UKPDS‐OM2 accurately simulated relative risks between randomized groups in both trials. The proportion of participants with the primary outcome was accurately estimated in both TECOS arms but underestimated by around one‐third in both EXSCEL arms.

Figures