Consistent results have been seen out to two years in the AGENT investigational device exemption (IDE) trial, the first US randomised trial investigating the safety and efficacy of the use of a drug-coated balloon (DCB) for the treatment of coronary in-stent restenosis (ISR).
Jeffrey Moses (Columbia University Medical Center, New York and St Francis Heart Center, Roslyn, USA) presented the trial’s two-year findings at the 2025 Cardiovascular Research Technologies (CRT) meeting (8–11 March, Washington DC, USA), where he commented that the data support the use of the Agent DCB (Boston Scientific) for the treatment of coronary ISR.
Agent, a paclitaxel-coated balloon, is the first DCB to gain approval from the US Food and Drug Administration (FDA) for the treatment of coronary ISR, which is estimated to make up around 10% of percutaneous coronary intervention (PCI) procedures.
The AGENT IDE trial, a prospective, randomised, multicentre, superiority trial enrolled a total of 600 patients across 40 US sites, who were randomised 2:1 to undergo either DCB-PCI with the Agent device or plain balloon angioplasty. Moses described the trial population as being a “varied group”, with an average age of 68 years, 26% women and 25% non-white. Importantly, he said, over 40% of patients had multi-layer stents.
At one year, the trial demonstrated that Agent outperformed conventional balloon angioplasty against a primary endpoint of target lesion failure (TLF)—defined as a composite of target lesion revascularisation, target vessel myocardial infarction (MI) or cardiac death—at 12 months, with the investigatory device demonstrating statistical superiority to uncoated balloon angioplasty (17.9% vs. 28.7%).
Detailing the two-year results at CRT 2025, Moses reported that TLF stood at 27% for the Agent DCB compared to 34% for balloon angioplasty, with a hazard ratio (HR) of 0.73. Analysis of recurrent events showed that patients had nearly twice as high a likelihood of a subsequent target lesion revascularisation (TLR) event in the plain balloon angioplasty group compared to DCB.
“To look at it another way, if you look at the first event, it is 28% vs. 19%, but there is also a two-thirds reduction in subsequent events in patients with multiple TLRs,” said Moses.
In terms of other endpoints, mortality was similar, myocardial infarction (MI) trended lower, and target vessel revascularisation (TVR) was 22% for the Agent DCB vs. 31.4% for the plain balloon group, whilst target vessel failure (TVF) was 28% for Agent, compared to 36% in the comparator group.
Subgroup analysis showed that TLF occurred more often in patients with multiple layers of existing stents (35.4% in the Agent arm and 44% in the balloon angioplasty arm) compared to those with a single layer of stent (20.7% for Agent and 26.6% for balloon angioplasty).
“AGENT IDE is the first randomised trial conducted in the USA examining the efficacy and safety of DCBs, it was superior to conventional balloon angioplasty for the target lesion endpoint at one year, and consistent results were observed on the two-year follow-up with a significantly reduced instance of target lesion failure, with 37% reduction,” said Moses in his concluding remarks.
“The recurrent TLR analysis shows significantly more repeat events with placebo balloon angioplasty. There were no definite or probable thromboses that occurred with the Agent balloon, so these data support the use of the Agent DCB for treatment of coronary ISR.”
Moses was asked by session moderator Gregg W Stone (Icahn School of Medicine at Mount Sinai, New York, USA) how to view these data when deciding whether to use a DCB or a drug-eluting stent (DES), the current standard of care, and whether there are certain scenarios that may favour the DCB.
“The easy one is multi-layer stenting. Most of us are quite hesitant to add a third layer. Smaller vessels also, [it is] a matter of real estate, and I also see complex bifurcation restenoses which are pretty common. Those are the sweet spots,” commented Moses.
“As far as the first restenosis in a large vessel, we will have randomised data to help guide us in the future. It is a fielder’s choice at this point and you can make arguments on both sides, recent meta-analyses are literally all over the map so I can’t give you guidance on that, we will wait for the real randomised data.”
