Design: Prespecified subgroup of a randomised trial
Population: 522 high bleeding risk PCI patients, 67% on oral anticoagulants
Primary outcome: Target vessel failure, BARC 3–5 bleeding and net adverse events at 1 year
Effect: TVF 4.9% vs 7.9% (RD −3.05%, 95% CI −7.27 to 1.17); NACE 6.8% vs 10.7%

SELUTION DeNovo randomised 3,323 patients to a sirolimus-eluting balloon strategy with stenting only if needed, or routine drug-eluting stents. This prespecified analysis looked at the 522 who met Academic Research Consortium high bleeding risk criteria; two thirds left hospital on an oral anticoagulant.

At one year, target vessel failure occurred in 4.9% with the balloon strategy and 7.9% with stents, and major bleeding in 2.3% and 3.2%. Neither difference was statistically significant. About one in six balloon patients needed bailout stenting. All antiplatelet therapy had been stopped in 28% of balloon patients by six months, against 15% with stents.

The subgroup was not powered to show superiority or non-inferiority, so the result is consistent with a balloon strategy being as good, not proven to be. The practical attraction is the shorter antiplatelet exposure in patients already on anticoagulation, which is exactly the group in whom bleeding drives outcomes.

Assess bleeding risk with the Academic Research Consortium criteria before choosing a PCI strategy.
Consider a drug-coated balloon strategy in high bleeding risk patients where lesion preparation is satisfactory.
Plan the shortest antiplatelet course the device allows for patients also taking an anticoagulant.
Keep bailout stenting available; about one in six needed it.

Why it matters

Leaving less metal behind may let anticoagulated patients drop antiplatelet therapy sooner.

The statistics, in plain English

The confidence interval for the difference in target vessel failure runs from 7.3 points better to 1.2 points worse with balloons, so it includes no difference. A subgroup of 522 from a trial of 3,323 is too small to settle the question on its own.