Christopher J Broyd, James Cockburn, Joao Martins, Timothy Bagnall, Mohammed Abouelasaad, Paul Bassett, Jessica Parker, Jeremy P Langrish, Jonathan Rawlins, Bernard Prendergast, Jonathan Hinton, Simon Eccleshall, David Hildick-Smith
Primary percutaneous coronary intervention (PCI) is the established gold-standard treatment for ST-elevation myocardial infarction (STEMI), but stent implantation confers an appreciable long-term failure rate driven by impaired vasomotion, neoatherosclerosis and late stent thrombosis. Drug-coated balloons (DCB) allow a leave-nothing-behind strategy that may optimise short- and long-term outcomes in this high-risk population, but no adequately powered randomised trial has compared a DCB-only strategy with drug-eluting stents (DES) in STEMI using an all-comers design with randomisation prior to diagnostic angiography. The STEMI DCB trial is an investigator-initiated, multi-centre, prospective, open-label, 1:1 randomised non-inferiority trial conducted across 15 United Kingdom sites (ISRCTN16341009), recruiting 700 patients presenting for primary angioplasty who will be randomised to a DCB-only strategy (SeQuent Please Neo) or DES prior to diagnostic angiography. An emergency consent process, followed by formal written consent before discharge, allows near-complete enrolment of this acute population. The primary endpoint is Target Vessel Failure (TVF), a composite of cardiovascular death and target vessel myocardial infarction or revascularisation, assessed for non-inferiority at 1 year with superiority testing if non-inferiority is met; patients will be followed to 8 years. By randomising before coronary anatomy is known, the STEMI DCB trial avoids the selection bias inherent in trials that randomise after angiography and restoration of flow, providing a true all-comers comparison of these strategies. This trial will provide the first adequately powered evidence on whether a DCB-only strategy is non-inferior to DES for STEMI, with potential to materially change primary PCI practice worldwide.