Large-scale registries can provide valuable complementary data to randomized controlled trials (RCT) for the postmarketing evaluation of coronary stents, but their scientific relevance remains debated.

We sought to compare the evidence on the performance of a single coronary stent platform generated by the RCT for its regulatory approval and a well-conducted international registry. Patients treated with the Ultimaster coronary stent in the CENTURY II (CII-UM) trial (n = 551) were compared to patients in the real-world e-ULTIMASTER (e-UM) registry (n = 35,389). All major events were adjudicated by an independent clinical event committee in both studies. Propensity weighted analysis was used to balance baseline and procedural differences between the 2 populations.

Coronary artery disease was more complex in e-UM compared to CII-UM, including more acute coronary syndromes, multivessel disease, left main, arterial, or venous grafts, and chronic total occlusions (P < .005 for all). At one-year follow-up and after excluding periprocedural myocardial infarction (MI) there was no statistically significant difference between CII-UM and e-UM regarding all-cause death (hazard ratio [HR] 0.55, 95% confidence interval [CI] 0.26-1.20, P = .14), cardiac death (HR 0.71, 95% CI 0.29-1.72, P = .45), target lesion failure (HR 1.18, 95% CI 0.78-1.78, P = .44), and target vessel MI (HR 0.76, 95% CI 0.24-2.38, P = .63). However, target vessel revascularization rate was significantly higher in CII-UM than in e-UM, HR 1.78, 95% CI 1.23-2.56, P = .002.

A well-conducted large-scale registry can provide valuable complementary evidence to RCTs on the postmarket performance of new coronary stents, across a wider range of uses and various geographic areas.