Predicting major adverse limb events in individuals with type 2 diabetes: Insights from the EXSCEL trial.

Abstract

Although models exist to predict amputation among people with type 2 diabetes with foot ulceration or infection, we aimed to develop a prediction model for a broader range of major adverse limb events (MALE)-including gangrene, revascularization and amputation-among individuals with type 2 diabetes. In a post-hoc analysis of data from the Exenatide Study of Cardiovascular Event Lowering (EXSCEL) trial, we compared participants who experienced MALE with those who did not. A multivariable model was constructed and translated into a risk score. Among the 14,752 participants with type 2 diabetes in EXSCEL, 3.6% experienced MALE. Characteristics associated with increased risk of MALE were peripheral artery disease (PAD) (HRadj 4.83, 95% CI: 3.94-5.92), prior foot ulcer (HRadj 2.16, 95% CI: 1.63-2.87), prior amputation (HRadj 2.00, 95% CI: 1.53-2.64), current smoking (HRadj 2.00, 95% CI: 1.54-2.61), insulin use (HRadj 1.86, 95% CI: 1.52-2.27), coronary artery disease (HRadj 1.67, 95% CI: 1.38-2.03) and male sex (HRadj 1.64, 95% CI: 1.31-2.06). Cerebrovascular disease, former smoking, age, glycated haemoglobin, race and neuropathy were also associated significantly with MALE after adjustment. A risk score ranging from 6 to 96 points was constructed, with a C-statistic of 0.822 (95% CI: 0.803-0.841). The majority of MALE occurred among participants with PAD, but participants without a history of PAD also experienced MALE. A risk score with good performance was generated. Although it requires validation in an external dataset, this risk score may be valuable in identifying patients requiring more intensive care and closer follow-up.