Exploring the Possible Impact of Unbalanced Open-Label Drop-In of Glucose-Lowering Medications on EXSCEL Outcomes

Background: EXSCEL (Exenatide Study of Cardiovascular Event Lowering) assessed the impact of once-weekly exenatide 2 mg versus placebo in patients with type 2 diabetes mellitus, while aiming for glycemic equipoise. Consequently, greater drop-in of open-label glucose-lowering medications occurred in the placebo group. Accordingly, we explored the potential effects of their unbalanced use on major adverse cardiovascular events (MACE), defined as cardiovascular death, nonfatal myocardial infarction or nonfatal stroke, and all-cause mortality (ACM), given that some of these agents are cardioprotective. Methods: Cox hazard models were performed by randomized treatment for drug classes where >5% open-label drop-in glucose-lowering medication occurred, and for glucagon-like peptide-1 receptor agonists (GLP-1 RAs; 3.0%) using three methodologies: drop-in visit right censoring, inverse probability for treatment weighting (IPTW), and applying drug class risk reductions. Results: Baseline glucose-lowering medications for the 14 752 EXSCEL participants (73.1% with previous cardiovascular disease) did not differ between treatment groups. During median 3.2 years follow-up, open-label drop-in occurred in 33.4% of participants, more frequently with placebo than exenatide (38.1% versus 28.8%), with metformin (6.1% versus 4.9%), sulfonylurea (8.7% versus 6.9%), dipeptidyl peptidase-4 inhibitors (10.6% versus 7.5%), SGLT-2i (10.3% versus 8.1%), GLP-1 RA (3.4% versus 2.4%), and insulin (13.8% versus 9.4%). The MACE effect size was not altered meaningfully by right censoring, but the favorable HR for exenatide became nominally significant in the sulfonylurea and any glucose-lowering medication groups, while the ACM HR and p-values were essentially unchanged. IPTW decreased the MACE HR from 0.91 (P=0.061) to 0.85 (P=0.008) and the ACM HR from 0.86 (P=0.016) to 0.81 (P=0.012). Application of literature-derived risk reductions showed no meaningful changes in MACE or ACM HRs or P values, although simulations of substantially greater use of drop-in cardioprotective glucose-lowering agents demonstrated blunting of signal detection. Conclusions: EXSCEL-observed HRs for MACE and ACM remained robust after right censoring or application of literature-derived risk reductions, but the exenatide versus placebo MACE effect size and statistical significance were increased by IPTW. Effects of open-label drop-in cardioprotective medications need to be considered carefully when designing, conducting, and analyzing cardiovascular outcome trials of glucose-lowering agents under the premise of glycemic equipoise. Registration:URL: https://www.clinicaltrials.gov; Unique identifier: NCT01144338.