Drug-coated balloons are exactly what the name suggests: angioplasty balloons coated with an antiproliferative drug, most commonly paclitaxel, that is transferred to the artery wall during a brief inflation. The drug suppresses the cellular overgrowth that causes arteries to renarrow, while the balloon itself leaves no permanent metallic scaffold behind. Drug-eluting stents, by contrast, combine a metallic mesh with a polymer that slowly releases a similar antiproliferative agent. Stents have dominated interventional cardiology since the second-generation devices proved themselves in landmark trials, and they remain the guideline-recommended treatment for newly diagnosed coronary lesions. DCBs, meanwhile, are already established and guideline-endorsed for treating in-stent restenosis, the renarrowing that occurs inside an existing stent, but their role in de novo disease, meaning lesions that have never been treated before, has remained an open question.

The appeal of a stentless strategy is considerable. By avoiding a permanent metallic implant, DCB angioplasty may preserve the artery’s natural vasomotor function, reduce the chronic inflammatory stimulus that a foreign body imposes, eliminate the risk of mechanical stent failure such as fracture or strut malapposition, and potentially allow shorter courses of dual antiplatelet therapy in selected patients. It also leaves future surgical or percutaneous revascularisation options unobstructed. In small vessels, where stent struts occupy a proportionally greater fraction of the lumen and raise the metal-to-artery ratio, these advantages are particularly relevant, and a recent patient-level meta-analysis found that DCBs actually reduced major adverse cardiovascular events compared with stents in small-vessel disease, a finding expected to influence future interventional guidelines.

What has been missing is high-quality randomised evidence in the broader, all-comer population of de novo coronary disease. That gap began to close with the publication of two large trials: REC-CAGEFREE I, which randomised 2272 patients in China across 43 centres, and SELUTION De Novo, a 3323-patient trial conducted across 62 centres in Europe and Asia. The new meta-analysis incorporates both, along with eight earlier trials, to produce the most contemporary synthesis available. The researchers systematically searched PubMed, Embase, and Scopus from inception to 30 May 2026, screening 2510 abstracts and ultimately including ten studies that compared DCB-only with DES-only percutaneous coronary intervention, required second-generation or newer stents in the comparator arm, and reported at least twelve months of follow-up.

Because trial-level definitions of major adverse cardiovascular events varied widely, the researchers analysed the individual components separately: cardiovascular death, myocardial infarction, and target lesion or target vessel revascularisation. The results were striking in their consistency. Cardiovascular death showed no significant difference between strategies, with an odds ratio of 1.06 and zero heterogeneity across trials. Myocardial infarction likewise showed no difference, with an odds ratio of 1.02. Critically, when the researchers excluded the four small-vessel-only trials to examine all-comer vessel sizes, the findings held: cardiovascular death yielded an odds ratio of 1.01 and myocardial infarction an odds ratio of 1.06, both statistically indistinguishable from no difference, with zero heterogeneity in every analysis.

Revascularisation told a more complicated story. Across all ten trials, there was no significant difference in target lesion or target vessel revascularisation between the two strategies, with an odds ratio of 1.39 and moderate heterogeneity. But when small-vessel-only studies were excluded, the picture changed: DCB-treated patients experienced significantly higher rates of repeat revascularisation, with an odds ratio of 1.84 and a p-value of 0.03. This finding, however, was heavily influenced by a single trial. When REC-CAGEFREE I was removed in a leave-one-out sensitivity analysis, the significant difference disappeared entirely. The researchers also noted that revascularisation as an endpoint is inherently vulnerable to variability in clinical judgement, patient preference, and local practice, and that a sensitivity analysis restricted to studies reporting target lesion revascularisation specifically produced consistent results, suggesting the findings were not driven by differences in endpoint definitions.

The divergence between the two largest trials deserves scrutiny. REC-CAGEFREE I, which used the Swide paclitaxel-coated balloon, failed to meet its non-inferiority criterion, driven largely by higher target vessel revascularisation in the DCB arm. SELUTION De Novo, which tested a sirolimus-eluting balloon, met non-inferiority on its intention-to-treat analysis, though the authors of the meta-analysis note that aspects of the non-inferiority margin have been debated and the secondary per-protocol analysis failed to meet the threshold. The two trials also differed fundamentally in design: REC-CAGEFREE I randomised patients after lesion preparation, making it a device trial, whereas SELUTION De Novo randomised before lesion preparation, making it a strategy trial. In DCB angioplasty, where procedural success depends heavily on lesion preparation and appropriate bailout stenting, this distinction matters considerably. Follow-up duration also differed, with REC-CAGEFREE I reporting three-year outcomes and SELUTION De Novo reporting one-year outcomes, leaving open the question of whether the favourable early results will persist.

The meta-analysis also raises a crucial point about device technology: a class effect cannot be assumed. Most of the included trials used paclitaxel-coated balloons, but the pharmacokinetic properties, coating technologies, and excipients differ substantially between devices. The Swide balloon used in REC-CAGEFREE I had limited published clinical evidence before that trial, and the device is not commercially available outside the trial programme. SELUTION SLR, the sirolimus-eluting balloon in the other large trial, uses a different drug and transfer mechanism altogether. Comparative data between sirolimus and paclitaxel-coated balloons remain limited, and early studies suggest broadly similar outcomes, but the authors caution that equivalence between platforms cannot currently be taken for granted. Notably, leave-one-out analyses excluding the sole sirolimus trial yielded consistent findings, suggesting the overall results were not driven by any single platform or antiproliferative agent.

The safety findings extended to thrombosis and to acute coronary syndrome populations. Target lesion thrombosis, reported in six studies covering 7326 patients, showed no statistical difference between strategies, though the point estimate numerically favoured DCB treatment with an odds ratio of 0.54. An exploratory analysis of ACS populations similarly found no significant difference in major adverse cardiovascular events, with an odds ratio of 1.07 and zero heterogeneity. Risk of bias assessment using the Cochrane RoB 2 tool classified seven studies as low risk across all domains, and GRADE assessment yielded moderate certainty for the cardiovascular death and myocardial infarction findings, though only low certainty for revascularisation outcomes. Sensitivity analyses excluding studies with bias concerns and stratifying by follow-up duration did not materially alter the core conclusions, although a time-dependent pattern emerged for revascularisation: the excess with DCB was significant at twelve months but no longer detectable at twenty-four months or beyond.

The authors are careful about the limits of their evidence. The analysis used aggregate trial-level data rather than individual patient records, limiting adjustment for baseline differences. Clinical and methodological heterogeneity was substantial, spanning stable disease and acute coronary syndromes, small vessels and all-comer populations, and twelve to sixty months of follow-up. The results may not generalise to complex lesion subsets such as left main stem disease or chronic total occlusions, which were excluded from the included trials. Several ongoing randomised trials, including TRANSFORM II, CAGE-FREE III, and the Reverse trial, will further define the role of DCB-only intervention across broader de novo coronary disease. For now, the message is measured: drug-coated balloons appear safe across any vessel size within the studied lesion subsets, but the higher revascularisation signal in larger vessels, and the meaningful differences between DCB platforms, mean that stentless angioplasty remains a strategy under active evaluation rather than a settled replacement for the stent.

Subject of Research: Comparison of drug-coated balloons versus drug-eluting stents for treating de novo coronary artery disease

Article Title: Drug coated balloons versus drug eluting stents for de novo coronary artery disease: a systematic review and meta-analysis of randomised controlled trials

Article References: Corballis, N., Natarajan, R., Merinopoulos, I., Tsampasian, V., Clark, A., Eccleshall, S., & Vassiliou, V. S. (2026). Drug coated balloons versus drug eluting stents for de novo coronary artery disease: a systematic review and meta-analysis of randomised controlled trials. eClinicalMedicine, Article 104204. https://doi.org/10.1016/j.eclinm.2026.104204

Image Credits: AI Generated

DOI: 10.1016/j.eclinm.2026.104204

Keywords: drug-coated balloons, drug-eluting stents, coronary artery disease, meta-analysis, randomised controlled trials, percutaneous coronary intervention, target lesion revascularisation, myocardial infarction, cardiovascular death, small vessel disease, interventional cardiology, Drug