Drug-coated balloon angioplasty in coronary artery disease: state of the art review Authors: Panagiotis Theofilis, Aggelos Papanikolaou, Kyriakos Dimitriadis, Paschalis Karakasis, Panagiotis Iliakis, Konstantinos Aznaouridis, Konstantinos Tsioufis, Dimitris Tousoulis Published: 2026 ABSTRACT AbstractDrug-coated balloons (DCBs) provide a “leave-nothing-behind” revascularization strategy, delivering antiproliferative therapy without permanent implants. This review presents evidence on DCB performance across coronary indications. In in-stent restenosis and de novo disease, DCBs are associated with higher rates of repeat revascularization compared to drug-eluting stents (DES). In small vessel disease, DCBs achieve comparable or sometimes superior outcomes. In complex lesions or high-risk patients, dedicated studies demonstrate feasibility and signals of non-inferiority.Subject terms: Cardiology, Diseases, Medical research Introduction Percutaneous coronary intervention (PCI) has undergone transformative advancements since its inception, evolving from plain old balloon angioplasty (POBA) to bare metal stents (BMS), and more recently to drug-eluting stents (DES)1. While DES have markedly reduced restenosis rates, they remain associated with complications such as stent thrombosis, impaired vasomotion, delayed healing, and the need for prolonged dual antiplatelet therapy (DAPT)1. These limitations have driven the resurgence of interest in drug-coated balloons (DCBs), which deliver antiproliferative medication directly to the vessel wall without leaving a permanent implant. By avoiding metallic scaffolding, DCBs offer a “leave nothing behind” strategy that may preserve vascular physiology, reduce long-term complications, and allow shorter antiplatelet regimens in selected patients1. Their use has expanded beyond in-stent restenosis (ISR) to include small vessel disease (SVD), high bleeding risk (HBR) patients, bifurcation lesions, and even selected acute coronary syndrome (ACS) cases2. However, the clinical performance of DCB angioplasty varies substantially across these indications. While robust randomized evidence supports its use in some settings, results are inconsistent in others, and in several complex scenarios, data remain limited or observational. Consequently, current literature is fragmented, often presenting indication-specific findings without a unifying framework to guide clinical decision-making. In indications supported by high-quality randomized evidence, such as ISR and de novo disease, DCB angioplasty does not consistently outperform DES and may be associated with higher repeat revascularization rates. In SVD, DCBs achieve outcomes comparable to or in some analyzes superior to- DES, supporting their role as a preferred stent-sparing strategy. In complex or high-risk clinical and anatomical settings, DCB use demonstrates feasibility and signals of non-inferiority, but outcomes remain highly dependent on patient selection, lesion preparation, and procedural expertise. This review, therefore, evaluates the available evidence in relation to these propositions, providing a structured framework for understanding the role of DCBs in contemporary PCI. Literature search strategy A structured literature search was performed using PubMed, Cochrane Library, and Google Scholar. Keywords included “drug-coated balloon”, “drug-eluting stent”, “coronary artery disease”, “in-stent restenosis”, “small vessel disease”, and “acute coronary syndrome”. Priority was given to randomized controlled trials, meta-analyzes, and large prospective registries. Additional relevant studies were identified through manual reference screening. DCB technology DCBs represent a novel modality for local antiproliferative drug delivery during PCI without permanent implant placement (see Table 1). A DCB comprises a semi-compliant balloon catheter coated with an antiproliferative agent, typically paclitaxel or sirolimus, and an excipient that facilitates drug adherence during transit and transfer to the vessel wall during balloon inflation2,3. Table 1Comparative features of contemporary drug-coated balloon (DCB) platformsDCB platformDrug (Dosage)Excipient / CarrierDelivery mechanismRCT studied inSeQuent Please NEOPaclitaxel (3 µg/mm²)IopromideHydrophilic matrixISAR-DESIRE 348BASKET-SMALL 215PEPCAD China SVD49PEPCAD NSTEMI29IN.PACT FalconPaclitaxel (3.5 µg/mm²)UreaCrystalline urea coatingBELLO50Elutax SVPaclitaxel (2 µg/mm²)Shellac + Vitamin EStabilized resin coatingPICCOLETO II4PrevailPaclitaxel (3.5 μg/mm2)UreaPrevail Global program51DIORPaclitaxel (3 μg/mm2)ShellacValentines II52RestorePaclitaxel (3 µg/mm²)ShellacShellac matrixRESTORE ISR China53RESTORE SVD China54Agent DCBPaclitaxel (2 µg/mm²)Acetyl-tri-butyl CitratePolymer-free rapid transferAgent IDE55Support CPaclitaxel (3 µg/mm²)Butyryl-tri-hexyl CitrateDissolve DCBPaclitaxel (3 µg/mm²)Mid-chain triglycerideProlonged surface availabilityDissolve ISR56Dissolve SVD57Pantera LuxPaclitaxel (3 µg/mm²)Butyryl-tri-hexyl CitrateREVELATION26EssentialPaclitaxel (3 μg/mm2)Organic esterDanubioPaclitaxel (2.5 μg/mm2)Butyryl-tri-hexyl CitrateDEBSIDE58Genoss PCBPaclitaxel (3 μg/mm2)Shellac + Vitamin EEnhanced transfer and retentionGenoss PCB study59MagicTouch SCBSirolimus (1.25 µg/mm²)Phospholipid nanocarrierNanoparticle dispersionTRANSFORM I21Selution SLRSirolimus (1 µg/mm²)PLGA micro-reservoirsControlled elution ( ~ 90 days)SELUTION4ISR10SELUTION DeNovo25SeQuent SCBSirolimus (4 μg/mm2)SIBLINT-ISR9Biolimus BCBBiolimus A9 (1 µg/mm²)Polymer-free matrixRapid-release lipophilicREFORM60 Comparative features of contemporary drug-coated balloon (DCB) platforms ISAR-DESIRE 348 BASKET-SMALL 215 PEPCAD China SVD49 PEPCAD NSTEMI29 RESTORE ISR China53 RESTORE SVD China54 Dissolve ISR56 Dissolve SVD57 SELUTION4ISR10 SELUTION DeNovo25 Drug properties and pharmacokinetics Paclitaxel remains the most commonly used drug due to its high lipophilicity, rapid tissue uptake, and long intracellular retention. These properties allow for effective drug delivery with a single, short-duration balloon inflation (30–60 s) producing sustained inhibition of smooth muscle cell proliferation and neointimal formation1. Sirolimus, a cytostatic mTOR inhibitor, offers a broader therapeutic window and a more favorable vascular healing profile. However, its lower lipophilicity and slower tissue uptake initially limited its feasibility. Recent advances in carrier technologies, such as micro-reservoirs, polymer-free matrices, and encapsulated delivery systems, have enabled the successful development of sirolimus-coated balloons (SCBs), such as MagicTouch and Selution SLR3. Excipients The excipient, or drug carrier, plays a critical role in maintaining drug stability, minimizing premature wash-off, and ensuring homogeneous transfer to the arterial wall. Common excipients include iopromide, urea, shellac, citrate esters, and polyethylene glycol. These substances influence coating durability, drug dispersion, and downstream particle embolization risk. Mechanistic insights Preclinical research has provided crucial insights into the biological differences between DCB technologies. In a comparative rabbit model, a study assessed two PCBs and one SCB using standardized injury protocols3. The PCBs tested were SeQuent Please and In. Pact Falcon, while the SCB was MagicTouch. Histological analysis at 24 h revealed that PCBs achieved significantly deeper medial smooth muscle cell loss compared to the SCB, suggesting more effective initial drug penetration3. However, this came at a potential cost, as PCBs also exhibited greater downstream drug deposition, indicating a higher risk of distal embolization and systemic drug dispersion3. In contrast, the SCB demonstrated a more confined and uniform drug distribution, but with less extensive medial smooth muscle cell loss, possibly due to sirolimus’ slower cellular uptake and less lipophilic nature3. These findings underscore the importance of drug kinetics, tissue binding affinity, and coating stability in DCB design. While PCBs may exert stronger early antiproliferative effects, SCBs could offer a more controlled and biologically favorable healing environment. This balance between efficacy and vascular safety should be carefully considered when interpreting clinical trial results and selecting DCB platforms for specific lesion subsets. Lesion preparation Principles and objectives of lesion preparation The success of DCB angioplasty is heavily dependent on optimal lesion preparation, which is essential for both drug uptake and minimizing the risk of acute complications (see Fig. 1). Unlike stent-based strategies, where mechanical scaffolding compensates for inadequate lesion preparation, DCB-only PCI demands meticulous attention to vessel compliance, plaque modification, and residual stenosis2,3. The primary objectives of lesion preparation are to achieve full balloon expansion, restore unobstructed coronary flow (TIMI 3), and minimize residual stenosis to <30% with no significant dissection (NHLBI type C or higher). Bailout stenting is reserved for complications such as flow-limiting dissection, major recoil, or persistent severe stenosis. However, proper preparation and DCB technique can keep bailout rates below 10% in contemporary practice, as demonstrated in trials like PICCOLETO II4. Drug-coated balloon (DCB) angioplasty workflow. The process begins with lesion preparation using non-compliant, scoring, or cutting balloons to achieve adequate plaque modification and facilitate effective drug penetration. Following predilatation, operators must confirm optimal angiographic results, including residual stenosis <30%, TIMI 3 flow, and absence of flow-limiting dissection. Intravascular imaging (IVUS/OCT) may be used to optimize vessel sizing and lesion assessment. Once adequate preparation is achieved, DCB angioplasty is performed to deliver antiproliferative therapy without permanent implantation. In cases of suboptimal results, bailout stenting may be required. Lesion modification techniques Achievement of an adequate result often requires predilatation with non-compliant balloons (NCB) to optimize plaque fracture and reduce elastic recoil1. Dedicated tools such as scoring, cutting, and high-pressure balloons can facilitate safe expansion in fibrotic or calcified lesions while minimizing deep dissections. An experimental rabbit iliac artery model showed that cutting balloon pretreatment followed by DCB angioplasty resulted in deeper and more circumferential medial smooth muscle cell loss compared with DCB alone5. This effect was associated with greater internal elastic lamina disruption and more homogeneous drug distribution5. The ongoing NATURE trial is a multicenter, randomized study evaluating cutting balloon versus standard balloon predilatation prior to DCB-only PCI in de novo coronary lesions of normal-sized vessels6. It will provide the first prospective evidence on whether cutting balloons optimize lesion preparation and improve procedural outcomes in DCB-based PCI beyond ISR. Role of intravascular imaging The use of intravascular imaging, particularly intravascular ultrasound (IVUS) and optical coherence tomography (OCT), is gaining prominence in DCB PCI. Imaging provides a detailed assessment of plaque morphology, calcium burden, lesion length, and vessel sizing, all critical for selecting the appropriate balloon size and confirming optimal lesion preparation. IVUS-guided DCB PCI has been shown to achieve greater minimal lumen diameter (MLD) and lower late lumen loss (LLL) compared to angiography-guided procedures7. It also reduces the likelihood of geographic miss and underexpansion, both associated with target lesion failure (TLF)7. Incorporating imaging into DCB workflows enhances procedural precision and patient selection, aligning with the evolving paradigm of individualized PCI strategies. Clinical applications of DCBs The clinical application of DCB angioplasty in contemporary PCI is highly context-dependent, varying according to lesion characteristics, clinical presentation, and procedural factors (see Fig. 2). Accordingly, a structured evaluation of the available evidence is required to guide appropriate patient and lesion selection. The key randomized trials, registries, and meta-analyzes supporting the use of DCB are summarized in Table 2 for ISR and Table 3 for the remaining clinical scenarios. Integrated decision-making algorithm for drug-coated balloon (DCB) versus drug-eluting stent (DES) in percutaneous coronary intervention (PCI). Stepwise algorithm guiding device selection in patients undergoing PCI for coronary artery disease. Patients are first stratified by clinical context (stable CAD vs. ACS), where high thrombus burden or unstable plaques favor DES. Lesion and patient characteristics then guide strategy: DCB is favored in small vessel disease (≤2.75–3.0 mm) and high bleeding risk (HBR), while DES remains the default in de novo CAD; ISR, bifurcation, and CTO require lesion-specific or hybrid approaches. Intravascular imaging (IVUS/OCT) is recommended to optimize vessel sizing and lesion preparation. A key procedural checkpoint determines DCB eligibility: residual stenosis <30%, TIMI 3 flow, and no flow-limiting dissection. If these criteria are met, DCB angioplasty may be performed; otherwise, DES implantation is recommended. Hybrid strategies (DCB + DES) may be used in selected cases. Color key: green = DCB-favored, red = DES-favored, yellow = conditional/hybrid strategies. Table 2Summary of contemporary randomized trials and meta-analysis assessing DCBs in the setting of In-Stent RestenosisStudyComparatorSample sizeFollow-upPrimary endpointMain findingsISAR-DESIRE 348PCB vs PES vs. POBA40210 yearsMACEMACE: 72% (POBA) vs. 55.9% (PCB) vs. 62.4% (PES) (p < 0.001)DAEDALUS8DCB vs DES19583 yearsTLR, MACETLR: 20.3% (DCB) vs 13.4% (DES), HR 1.58MACE: 9.5% (DCB) vs. 13.3% (DES), HR 0.69RESTORE ISR China53Restore (PCB) vs SeQuent Please (PCB)2409 monthsIn-segment LLLLLL: 0.38 ± 0.50 mm (Restore) vs 0.35 ± 0.47 mm (SeQuent) (p non-inferiority = 0.02)SIBLINT-ISR9Sequent (SCB) vs SeQuent Please (PCB)2589 monthsIn-segment LLLLLL: 0.37 mm (SCB) vs 0.30 mm (PCB), difference 0.07 mm (95% CI: −0.05 to 0.19)TIS 261MagicTouch (SCB) vs SeQuent Please (PCB)14512 monthsIn-segment LLL, Binary ISR, MACEΔLLL: −0.024 (failed to establish non-inferiority)Binary ISR: 31.6% (SCB) vs. 30.4% (PCB)MACE: 31% for bothDissolve ISR56Dissolve (PCB) vs SeQuent Please (PCB)2609 monthsIn-segment LLLLLL: 0.50 mm (Dissolve) vs 0.47 mm (SeQuent Please)TLF: 17.5% vs 10.7% (p = 0.12)Binary restenosis: 23.4% vs 16.4% (p = 0.19)REFORM60Biolimus (BCB) vs SeQuent Please (PCB)2026 monthsIn-segment % diameter stenosis% diameter stenosis: 43.3% (Biolimus) vs. 31.4% (SeQuent Please)TLF: 23.7% (Biolimus) vs. 17.1% (SeQuent Please)Failed to establish noninferioritySELUTION4ISRSelution SLR (SCB) vs Usual care (POBA/DES)39012 monthsTLFTLF: 16.2% (SCB) vs. 14.5% (usual care) - noninferior TLF: 14.2% (SCB) vs. 9% (DES) - failed to establish noninferiorityPCB paclitaxel-coated balloon, PES paclitaxel-eluting stent, POBA plain old balloon angioplasty, DCB drug-coated balloon, DES drug-eluting stent, SCB sirolimus-coated balloon, ISR in-stent restenosis, BCB biolimus-coated balloon, MACE major adverse cardiovascular events, TLR target lesion revascularization, LLL late lumen loss, HR hazard ratio, CI confidence interval, TLF target lesion failure. Summary of contemporary randomized trials and meta-analysis assessing DCBs in the setting of In-Stent Restenosis TLR: 20.3% (DCB) vs 13.4% (DES), HR 1.58 MACE: 9.5% (DCB) vs. 13.3% (DES), HR 0.69 ΔLLL: −0.024 (failed to establish non-inferiority) Binary ISR: 31.6% (SCB) vs. 30.4% (PCB) MACE: 31% for both LLL: 0.50 mm (Dissolve) vs 0.47 mm (SeQuent Please) TLF: 17.5% vs 10.7% (p = 0.12) Binary restenosis: 23.4% vs 16.4% (p = 0.19) % diameter stenosis: 43.3% (Biolimus) vs. 31.4% (SeQuent Please) TLF: 23.7% (Biolimus) vs. 17.1% (SeQuent Please) Failed to establish noninferiority PCB paclitaxel-coated balloon, PES paclitaxel-eluting stent, POBA plain old balloon angioplasty, DCB drug-coated balloon, DES drug-eluting stent, SCB sirolimus-coated balloon, ISR in-stent restenosis, BCB biolimus-coated balloon, MACE major adverse cardiovascular events, TLR target lesion revascularization, LLL late lumen loss, HR hazard ratio, CI confidence interval, TLF target lesion failure. Table 3Summary of major studies comparing long-term efficacy of DCBs across main PCI indicationsStudyDesignTreatment strategiesSample sizeBailout DESIVI useFollow-upFindingsEvidenceRecommendationSVD BASKET-SMALL 215,16RCTPCB vs. DES7585.1%NR3 yearsTLF: DCB 15% vs. DES 15% (non-inferiority achieved)Major bleeding: PCB 2% vs. DES 4%HighConsider use PICCOLETO II4,19RCTPCB vs. DES2326.7%10%3 yearsTLF: DCB 10.8% vs. DES 20.8% (achieved superiority) ANDROMEDA20IPD MAPCB vs. DES11547.6%NR3 yearsTLF: DCB 14.7% vs. DES 17.6% (NS)De Novo CAD REC-CAGEFREE I23,24RCTPCB vs. DES2,2729.4%10.3%3 yearsTLF: DCB 8.2% vs. DES 5.0% (did not achieve non-inferiority)BARC type 3 or 5 bleeding: DCB 1.4% vs. DES 2.6%HighUse within the scope of the existing evidence - further research required Selution DeNovo25RCTSCB vs. DES3,34120.7%17.3%1 yearTLF: DCB 5.3% vs. DES 4.4% (non-inferiority achieved)ACS REVELATION26,27RCTPCB vs. DES12018%NR2 yearsTLF: DCB 5.4% vs. DES 1.9% (NS)Low-ModerateUse only in experienced centers - further research required PEPCAD NSTEMI29RCTPCB vs. BMS/DES21015%NR9 monthsTLF: DCB 3.8% vs. BMS/DES 6.6% (non-inferiority achieved) Merinopoulos et al. 28CohortDCB vs. DES1,139N/ANR3 yearsNo difference in all-cause mortality, cardiovascular mortality, MI, stroke, unplanned TLR, major bleedingBifurcations DCB-BIF30RCTDCB vs. NCB7843.6%27%1 yearTLF: DCB 7.2% vs. NCB 12.5% (superiority)LowUse only in experienced centers - further research required SPACIOUS31RCTPCB vs. SCB2290%0%1 yearTLF: PCB 6.1% vs. SCB 2.6% (NS)CTO PROGRESS-CTO38RegistryDCB only2202.2%39%9 monthsMACE: 7.7%, all-cause mortality 0%, TVR 7.7%, MI 0.7%LowUse only in experienced centers - further research required CTO-DENOVO39RegistryDCB only3097.2%16.8%6 monthsTLF: 3.5%, target-vessel MI: 0.3%, TLR: 3.2%HBR DEBUT41RCTPCB vs. BMS2082%NR9 monthsTLF: PCB 1% vs. BMS 14% (superiority)LowUse only in experienced centers - further research requiredTLF is a composite of (cardiovascular) death, (target-vessel) MI, or TLR.RCT randomized control trial, PCB: paclitaxel-coated balloon, DES drug eluting stent, NR not reported, TLF target-lesion failure, NS non-significant, SCB sirolimus-coated balloon, BMS bare metal stent, N/A not applicable, MI myocardial infarction, TLR target lesion revascularization, NCB non-compliant balloon, MACE major adverse cardiovascular events. Summary of major studies comparing long-term efficacy of DCBs across main PCI indications TLF: DCB 15% vs. DES 15% (non-inferiority achieved) Major bleeding: PCB 2% vs. DES 4% TLF: DCB 8.2% vs. DES 5.0% (did not achieve non-inferiority) BARC type 3 or 5 bleeding: DCB 1.4% vs. DES 2.6% TLF is a composite of (cardiovascular) death, (target-vessel) MI, or TLR. RCT randomized control trial, PCB: paclitaxel-coated balloon, DES drug eluting stent, NR not reported, TLF target-lesion failure, NS non-significant, SCB sirolimus-coated balloon, BMS bare metal stent, N/A not applicable, MI myocardial infarction, TLR target lesion revascularization, NCB non-compliant balloon, MACE major adverse cardiovascular events. In-stent restenosis (ISR) Despite the significant reduction in restenosis afforded by second-generation DES, ISR continues to be a major cause of PCI failure, especially in high-risk populations such as those with diabetes, long lesions, and overlapping stents. Management strategies for ISR must balance the need for effective neointimal suppression against the risks associated with additional metal layers and prolonged DAPT. DCB angioplasty has become an alternative treatment option for both BMS- and DES-related ISR. Its non-scaffold nature avoids further metallic layering and may preserve vessel physiology. Thus, this section examines whether DCBs can provide an effective stent-sparing strategy in the treatment of ISR. Several studies have assessed the role of DCBs in this setting (Table 2). The largest evidence comes from the DAEDALUS individual patient-level meta-analysis, which pooled 10 randomized controlled trials (RCTs) and over 1900 patients with ISR8. In BMS-ISR, DCBs and DES showed comparable efficacy and safety8. In DES-ISR, DES were more effective in reducing target lesion revascularization (TLR, 13.4% versus. 20.3%), but DCBs demonstrated a numerical advantage in safety outcomes (9.5% versus. 13.3%), particularly regarding late thrombotic events and myocardial infarction8. A closer look at recent randomized data further refines our understanding of how DCBs perform in ISR, particularly when comparing newer coating technologies or specific populations. The SIBLINT-ISR trial was a multicenter RCT comparing a novel SCB (SeQuent SCB) to a PCB for coronary ISR9. Notably, this SCB has a higher sirolimus dose (4 μg/mm²) than other SCBs. In 258 patients with 285 lesions, SCB was non-inferior to PCB in 9-month LLL (0.37 mm versus 0.30 mm, Δ = 0.07 mm, 95% CI −0.05 to 0.19) and showed comparable 12-month TLF and binary restenosis rates9. Lastly, SELUTION4ISR was a multicenter, randomized, single-blind trial comparing PCI with SELUTION SLR to standard of care (POBA-20% or DES-80%) in 390 patients with ISR (DES 90%)10. Among the study’s strengths were the mandatory use of intracoronary imaging to assess vessel size and confirm adequate stent expansion, the need for optimal lesion preparation (<30% stenosis) before randomization, and the frequent use of specialty balloons10. Regarding patient characteristics, the majority had single-layer ISR, with 22% having a double-layer ISR. Ultimately, at 1-year follow-up, this study demonstrates that SCBs are non-inferior to a mixed “real-world” strategy but remain inferior to DES in single-layer ISR, with differences driven primarily by repeat revascularization rather than safety endpoints10. Bleeding rates did not differ statistically across treatment arms, despite a shorter DAPT duration in the SCB arm10. The ongoing 5-year follow-up is expected to shed more light on the comparison of those strategies in ISR. Regarding multilayer DES ISR, a prespecified subgroup analysis of the AGENT IDE trial concerned patients with multilayer DES-ISR randomized to a low-dose PCB or POBA after successful lesion preparation11. DCB treatment significantly reduced TLF versus POBA, driven by lower ischemia-driven revascularization and target-vessel myocardial infarction, with no stent thrombosis events after DCB11. In view of the above, DCB angioplasty could represent a valuable stent-sparing strategy for ISR, particularly in selected settings (BMS-ISR, multilayer-ISR). However, heterogeneity in study design, device technology, and comparator strategies, lack of long-term data, as well as moderate signals of superiority of DES in DES-ISR, particularly for intermediate-term TLR reduction, temper the overall strength of support. It should be noted that existing guideline recommendations support the use of DES in ISR according to the existing evidence base (Class 1, Level of Evidence A)12,13. Small vessel disease SVD remains a well-recognized limitation in contemporary PCI. Despite the presence of various definitions (typically ≤2.75–3.00 mm in diameter), these arteries are more prone to restenosis and adverse events due to increased metal-to-artery ratio, poor stent expansion, and challenging anatomy. Emerging evidence also suggests that DCBs may offer superior preservation of endothelial function compared to DES in this setting. In a prospective, randomized trial by Kawai et al., 42 patients with de novo SVD were treated with either a PCB or a bioabsorbable polymer everolimus-eluting stent (EES) after successful predilation14. At 8-month follow-up, coronary vasomotion was assessed via acetylcholine and nitroglycerine challenge. The DCB group exhibited significantly better endothelium-dependent vasomotor responses in the distal segments, with markedly less vasoconstriction after acetylcholine infusion, suggesting superior preservation of vascular physiology. Therefore, this section examines whether DCB PCI can achieve comparable, or even superior, clinical outcomes to DES, particularly in light of the potential benefits of avoiding permanent metallic scaffolds. The BASKET-SMALL 2 trial was a pivotal randomized study that enrolled 758 patients with native SVD (<3 mm), comparing PCB (SeQuent Please) to second-generation DES (mostly EES). At 12 months, the primary composite outcome of cardiac death, nonfatal myocardial infarction, and target vessel revascularization (TVR) was statistically non-inferior in the DCB arm (7.3% versus. 7.5%, HR 0.97)15. Long-term follow-up confirmed sustained efficacy and safety up to 3 years, with no significant differences in TLR, death, or major adverse cardiovascular events (MACE)16. Importantly, subgroup analyzes in patients with chronic kidney disease and HBR suggested potential bleeding reduction with DCB use17,18. PICCOLETO II, another landmark RCT, compared Elutax SV (a PCB) with EES in 232 patients with de novo SVD. DCB was associated with significantly lower LLL (0.17 mm versus 0.29 mm, p = 0.01) and comparable TLR and MACE4. Three-year follow-up confirmed non-inferiority in clinical endpoints and continued superiority in angiographic parameters19. While limited by modest sample size, the trial set a precedent for routine DCB use in SVD. A major milestone in the evaluation of DCBs for SVD came with the publication of the ANDROMEDA study, an individual patient data meta-analysis of four RCTs (BELLO, BASKET-SMALL 2, PICCOLETO II, RESTORE SVD China)20. This included 1154 patients with 1360 lesions and aimed to clarify the long-term comparative efficacy of PCB versus DES in small vessels (≤3.0 mm in diameter). At 3-year follow-up, MACE occurred significantly less often in the PCB group than in the DES group (HR 0.67, 95% CI 0.47–0.96), driven mainly by reduced myocardial infarction and TVR rates. Notably, TLF was not significantly different between the groups (HR 0.77, 95% CI 0.49–1.21), affirming the non-inferiority of PCB angioplasty in terms of efficacy. A crucial strength of ANDROMEDA lies in its inclusion of long-term data, moving beyond the traditional 6–9-month angiographic endpoints that previously limited interpretation. Additionally, the authors addressed the heterogeneity between included trials, especially the more neutral BASKET-SMALL 2 effect, through stratified one- and two-stage analyzes. Even with variations in DES generations (some patients received first-generation DES), sensitivity analyzes confirmed the robustness of the findings. The inclusion of reconstructed time-to-event data from RESTORE SVD China, which directly compared a shellac-based PCB to a zotarolimus-eluting stent, further strengthened the conclusions. That study itself showed angiographic non-inferiority and equivalent 24-month clinical outcomes, reinforcing that modern PCBs perform reliably against second-generation DES. Crucially, ANDROMEDA confirmed that no increase in all-cause or cardiac mortality was observed with PCB treatment over 3 years, disproving persistent concerns from earlier paclitaxel safety controversies in peripheral artery disease. In sum, ANDROMEDA represents the most comprehensive clinical validation to date of DCB use in SVD, highlighting the clinical safety and angiographic efficacy of PCBs in this anatomy. While prior evidence has firmly established the efficacy of PCBs in SVD, the role of SCBs remains less defined. The TRANSFORM I trial was the first prospective, multicenter, randomized study to directly compare an SCB (MagicTouch) with a PCB (SeQuent Please Neo) in de novo SVD21. The study included 121 patients with angiographically defined SVD (reference vessel diameter <2.75 mm) and used OCT-based balloon sizing for precision. The primary endpoint, 6-month angiographic net lumen gain, failed to show non-inferiority of SCB versus PCB (0.25 mm versus 0.48 mm, p for noninferiority=0.173), with a net gain difference of −0.23 mm (95% CI −0.37 to −0.09)21. SCB was associated with higher rates of binary restenosis (32.8% versus. 12.5%, p = 0.012), lower angiographic fractional flow reserve (0.86 versus 0.91, p = 0.026), and more frequent vessels with functionally significant stenosis (fractional flow reserve (FFR) ≤ 0.80)21. Interestingly, SCB exhibited lower LLL (0.00 mm versus 0.32 mm), suggesting less proliferative response but also lower acute gain, which may have contributed to its underperformance21. These findings highlight the complexity of SCB pharmacokinetics stated previously. Despite the trial’s meticulous design, including core lab analysis, OCT guidance, and prespecified sensitivity analyzes, the SCB device could not demonstrate equivalence to PCB in de novo SVD. However, several caveats apply. The trial was relatively small and geographically limited, which may affect external validity. Moreover, its findings, particularly the dissociation between low LLL and higher restenosis, highlight potential challenges in achieving sufficient acute lumen gain with SCB. This underscores the need for optimized lesion preparation protocols and possibly longer drug transfer times. Importantly, SCB technology is heterogeneous, and results with MagicTouch may not extrapolate to other platforms. Ongoing trials with SCB, such as TRANSFORM II22 and SELUTION 4 De Novo Small Vessel IDE (NCT05946629), are better powered and will help define whether these early limitations are technological, procedural, or intrinsic to the sirolimus delivery paradigm. Overall, the evidence base in SVD is supported by multiple randomized controlled trials with consistent long-term follow-up, although studies are largely open-label and intravascular imaging use is variable. The strength of evidence is therefore high. In this setting, DCB angioplasty represents a well-supported stent-sparing strategy, although further data are needed to define the role of newer device platforms and optimize procedural standardization. De novo lesions The use of DCB in de novo coronary artery disease (CAD) is primarily motivated by the potential to avoid permanent metallic scaffolds, thereby reducing the risks of late thrombosis, impaired vasomotion, and prolonged DAPT, particularly in anatomically or clinically selected patients. Recent evidence has expanded its application to this broader. This section aims to demonstrate that in de novo CAD, the clinical performance of DCB angioplasty is variable and does not consistently surpass DES. The REC-CAGEFREE I trial represents the largest randomized comparison of a DCB-first strategy (with provisional stenting) versus routine DES implantation in non-complex de novo lesions, irrespective of vessel diameter23. In this open-label, multicenter study of 2272 patients, the DCB strategy failed to meet non-inferiority for the device-oriented composite endpoint at 24 months, with this difference persisting at 3 years, although the absolute gap narrowed over time23,24. In contrast, the open-label SELUTION DeNovo trial, which randomized 3341 patients with de novo CAD to either DCB (Selution SLR) with provisional DES or upfront DES, met the non-inferiority margin for TVF (4.4% in DES, 5.3% in DCB) at 1 year25. Both studies included a wide range of patients and lesions, but excluded patients with left main CAD, chronic total occlusion (CTO), and ISR, among others24,25. Importantly, large lesions > 3.0 mm (46–50% in REC-CAGEFREE I, 63–67% in Selution DeNovo) and LAD lesions (~50%) constituted a significant proportion in both trials24,25. However, their results diverged, underscoring the importance of procedural and methodological differences. First, the trial design differed substantially. REC-CAGEFREE I mandated successful lesion preparation prior to randomization, whereas SELUTION DeNovo randomized patients before lesion preparation23–25. This may partly explain the higher rate of provisional DES implantation in the SCB arm (20.7%) compared with rescue stenting in the PCB arm (9.4%)23–25. Consistently, the use of specialty balloons (cutting/scoring) was more frequent and balanced between groups in REC-CAGEFREE I, but relatively limited in SELUTION DeNovo (28.5% in DCB vs. 7.9% in DES)23–25. Second, differences in patient and lesion profiles may have influenced outcomes. REC-CAGEFREE I included a higher proportion of acute coronary syndromes (>50%), including STEMI, whereas SELUTION DeNovo enrolled fewer ACS patients (<35%) and excluded STEMI and unstable NSTEMI23–25. Proximal LAD lesions were also more frequent in REC-CAGEFREE I (~26–28% vs. 18-19%), suggesting a potentially higher-risk anatomical substrate23–25. Conversely, SELUTION DeNovo included a substantial proportion of moderately or severely calcified lesions (>20%), in which exploratory subgroup analyzes suggested a more favorable performance of SCB25. Intracoronary imaging was not routinely used, but was rarer in REC-CAGEFREE I (9.5–11.2% vs. 15.8–18.8%)23–25. Despite these differences, a consistent pattern emerges across both trials. The primary endpoint differences were largely driven by higher rates of repeat revascularization in the DCB arms, whereas hard clinical outcomes, including cardiovascular death, TV-myocardial infarction, and stroke, were similar between treatment strategies. However, the available data in REC-CAGEFREE I indicates that TVR were in most cases elective (DCB: 8.0% vs. DES: 4.8%), with no difference in urgent revascularization (DCB: 1.0% vs. DES: 1.2%)24. Additionally, major bleeding events were numerically lower with DCB therapy, likely reflecting reduced intensity and duration of antiplatelet therapy exposure during follow-up24. Taken together, these findings suggest that DCB angioplasty may represent a stent-sparing strategy in selected de novo coronary lesions, achieving comparable safety but at the cost of higher rates of (elective) repeat revascularization. The strength of evidence can be considered moderate/high, as large randomized trials are available but yield conflicting results that appear influenced by various factors. Consequently, the role of DCB in de novo CAD remains context-dependent, and further long-term data with contemporary device platforms are required to better define its place relative to DES. Acute coronary syndromes The role of DCBs in ACS remains an area of growing interest but also of clinical caution. Traditionally, stent-based strategies have dominated the management of ACS due to their capacity to immediately seal ruptured plaques and restore vessel integrity. However, DCBs offer theoretical advantages in this setting: reduced foreign body burden, preserved vasomotion, shorter DAPT duration, and potentially lower late adverse events. Still, concerns persist regarding thrombus load, lesion instability, and the risk of acute recoil or dissection in plaque-rich lesions. This section aims to demonstrate that in ACS, DCB angioplasty is feasible in selected cases but highly dependent on lesion preparation and patient selection. The REVELATION trial was a prospective, randomized, single-center study that compared PCB angioplasty with DES in 120 patients presenting with STEMI26. Patients with non-severely calcified culprit lesions and residual stenosis <50% after predilatation were randomized. The primary endpoint was FFR at 9 months26. Results demonstrated non-inferiority of DCB to DES, with mean FFR values of 0.92 ± 0.05 in the DCB group and 0.91 ± 0.06 in the DES group (p = 0.27)26. At 2-year follow-up, MACE occurred in 5.4% of the DCB group and 1.9% of the DES group (p = 0.34), indicating comparable safety and efficacy in selected patients27. Further strengthening the case for DCB in STEMI is a recent real-world, propensity score-matched cohort study by Merinopoulos et al., which analyzed outcomes in 1139 patients undergoing primary PCI for STEMI28. Among them, 452 patients received DCB-only angioplasty, while 687 were treated with second-generation DES. The study excluded patients requiring bailout stenting, cardiogenic shock, or cardiac arrest, allowing a cleaner comparison of intentionally stentless versus stented strategies. After a median follow-up of over 3 years, all-cause mortality and net adverse cardiac events, including cardiovascular death, ACS, stroke, major bleeding, and unplanned TLR, were statistically similar between the two groups28. In multivariable analysis, predictors of mortality were age, frailty, and comorbidities such as heart failure, but not the type of device used28. Even in propensity score-matched cohorts, no difference in mortality or safety outcomes was observed28. Notably, the DCB cohort included more patients with bifurcations and true bifurcation lesions, suggesting that DCB may have been selectively used in technically challenging anatomies where avoiding metal was advantageous28. Certain limitations should be acknowledged. The study was retrospective, single-center, and non-randomized, with potential for operator bias in choosing DCB versus DES. The high institutional expertise with DCBs, where over 100 STEMI patients annually were treated without stents, may not be representative of broader practice. The learning curve and case selection could significantly influence outcomes. Moreover, while the majority of vessels were >3 mm, the applicability to small vessel STEMI remains unclear. For non-STEMI (NSTEMI) patients, the PEPCAD NSTEMI trial evaluated DCB versus DES in 210 patients29. The study found that DCB treatment was non-inferior to stenting concerning LLL and clinical outcomes, suggesting that DCBs can be a viable alternative in selected NSTEMI cases29. The available evidence in ACS is derived from small randomized trials and observational studies, with inherent limitations including lack of blinding, modest sample sizes, and selective patient populations, as well as incomplete representation of high-risk lesions. The strength of evidence is low to moderate. While DCB angioplasty appears feasible in selected cases, its clinical value remains conditional on careful patient selection and optimal lesion preparation, requiring confirmation in larger randomized trials. Specifically, successful outcomes depend on complete thrombus resolution, optimal lesion preparation, and the absence of major dissections. Thus, operators must exercise caution in thrombus-laden, flow-limiting, or unstable plaques, where the risks of recoil and abrupt closure are unacceptably high. Therefore, in this context, DCB may serve best in a selective “lesion-specific” role: small vessel culprit lesions in NSTEMI, side branch (SB) in bifurcations, or when full stent deployment is technically challenging or undesirable. Importantly, newer generation SCBs have not yet been adequately studied in ACS, and their slower tissue uptake may limit their use in urgent or thrombus-heavy situations. DCB in bifurcation PCI Coronary bifurcation lesions account for approximately 15–20% of all PCIs and remain among the most technically challenging anatomical subsets. The mainstay of treatment is provisional stenting of the main vessel (MV) with optional SB intervention, especially when the SB is significantly compromised post-MV stenting. However, SB interventions with conventional balloon angioplasty or stenting carry risks of dissection, restenosis, or stent thrombosis, complications amplified by metallic overlap and vessel distortion. This section aims to demonstrate that in bifurcation PCI, DCB angioplasty, particularly for SB treatment, offers a stent-sparing approach with promising but still evolving evidence. The application of DCBs for the SB during provisional bifurcation PCI could potentially minimize the need for additional stent layers while delivering antiproliferative drugs to reduce restenosis. DCBs have demonstrated procedural feasibility and growing clinical support in this setting. The DCB-BIF trial, a multicenter RCT, compared DCB angioplasty to NCB inflation in the SB following MV stenting in true bifurcation lesions30. In 784 patients, DCB treatment significantly reduced the 1-year MACE rate compared to NCB (7.2% versus. 12.5%, HR 0.56, p = 0.013), primarily driven by fewer myocardial infarctions30. Importantly, crossover to two-stent strategy and stent thrombosis rates were similar, indicating procedural safety30. Additional support comes from the SPACIOUS trial, which compared a SCB to a PCB for SB treatment31. In 230 patients with non-left main bifurcation lesions, SCB was non-inferior to PCB in angiographic outcomes, suggesting that SCBs may be viable in this domain31. However, results remain preliminary, and differences in drug pharmacokinetics between sirolimus and paclitaxel call for cautious interpretation. A recent systematic review and meta-analysis pooled 11 studies (including 2 RCTs) comparing DCB and DES in bifurcation lesions32. It found no significant difference in MACE or TLR but did show lower LLL with DCBs (mean difference: −0.27 mm, p < 0.0001)32. The benefit was driven predominantly by de novo lesions, underscoring DCB’s potential in bifurcation PCI32. In addition to the growing evidence supporting DCBs in bifurcation PCI compared to DES, a meta-analysis by Rocchetti et al. evaluated five studies comparing DCB to NCB for SB intervention following main vessel provisional stenting33. DCB use was associated with a significant reduction in both MACE (OR 0.48, 95% CI 0.34–0.68) and myocardial infarction (OR 0.39, 95% CI 0.25–0.62), with no significant differences in TLR/TVR33. Multiple trials are ongoing to refine this strategy. The OCVC-BIF trial is assessing whether DCB use in the SB after kissing balloon inflation (KBI) provides superior restenosis prevention versus conventional balloon angioplasty34. Simultaneously, the Hybrid DEB and the PROVISION-DEB study compare a combined approach of MV stenting with SB DCB to a two-stent strategy, aiming to establish whether a simplified hybrid approach can match the outcomes of more complex techniques35,36. Evidence in bifurcation PCI focuses mainly on side branch treatment and is characterized by heterogeneity in procedural techniques and limited long-term follow-up. The overall strength of evidence is therefore low. While DCB angioplasty appears to reduce restenosis and avoid additional metal layers in the side branch, its incremental clinical value over contemporary provisional stenting strategies remains uncertain and requires confirmation in adequately powered trials with standardized methodologies. DCB in chronic total occlusion PCI CTO PCI frequently mandates implantation of multiple, long DES, which increases the risk of restenosis, late thrombosis, loss of vasomotion, and the need for prolonged DAPT. DCBs offer the possibility of local antiproliferative drug delivery while limiting or even avoiding additional metal, a particularly appealing concept in small, distal, or negatively remodeled segments beyond the occlusion and in in-stent CTOs. This section aims to demonstrate that in CTO PCI, DCB-based strategies can reduce metal burden in selected lesions, although their effectiveness remains dependent on careful selection and procedural optimization. Three contemporary registries have now characterized real-world use of DCBs in CTO PCI. In ERCTO, among 40,449 CTO PCIs performed at 184 centers between 2016 and 2023, DCBs were used in 6.2% of cases, with uptake rising from 3.4% to 14.9% over the study period37. Most DCB cases followed a hybrid DCB + DES approach, with a minority treated by DCB only. DCBs were preferentially used in small-caliber and distal vessels and in in-stent CTOs. After propensity score matching, DCB use was associated with a substantial reduction in implanted DES length (44 versus 58 mm), lower contrast volume, and less frequent use of complex bifurcation two-stent techniques, without any excess of periprocedural myocardial infarction or revascularization37. Complementary insights come from the PROGRESS-CTO registry, which analyzed 12,146 patients with successful CTO crossing between 2018 and 202538. DCBs were used in 3.7% of procedures, again with marked temporal growth, split almost evenly between DCB-only and hybrid strategies38. Lesions selected for DCB-only treatment were less complex, with lower J-CTO scores, less calcification and tortuosity, smaller reference diameters, and a high prevalence of in-stent CTO or diffuse distal disease38. Common indications were ISR CTO, distal vessel disease, and SB treatment. Technical success and in-hospital MACE were similar between DCB and DES cases, and DCB use was not independently associated with lower success or higher early risk38. On a median 9-month follow-up, patients treated with a DCB-only strategy experienced low event rates (MACE: 7.7%, All-cause mortality 0%, TVR 7.7%, MI 0.7%)38. Recent evidence further supports the feasibility of a DCB-only strategy in de novo CTO lesions. In the multicenter international CTO-DENOVO registry, 309 patients with 312 successfully recanalized CTO lesions were treated with a DCB-only approach across 42 centers39. At 6-month follow-up, the rate of target lesion failure was low (3.5%), with no cardiac deaths and very low rates of target vessel myocardial infarction (0.3%) and clinically driven target lesion revascularization (3.2%)39. Among patients undergoing angiographic follow-up, late lumen loss was minimal and nearly half of treated vessels demonstrated positive vessel remodeling, supporting the concept of vascular restoration after a stent-free strategy39. Restenosis occurred in approximately one quarter of lesions and was associated with higher residual stenosis immediately after PCI, underscoring the critical importance of optimal lesion preparation before DCB deployment39. To conclude, current data in CTO PCI are derived predominantly from observational registries, with no randomized comparisons and limited long-term follow-up, introducing substantial selection bias. The strength of evidence is low. Although DCB-based strategies are feasible in selected lesions (distal, small, or stent-failure-prone segments), their clinical value remains uncertain and requires validation in prospective randomized studies. Until such data are available, DCB-based approaches should be considered in carefully selected lesions with optimal lesion preparation and preserved distal vessel quality. High bleeding risk The management of patients at HBR undergoing PCI remains a clinical challenge, particularly in balancing thrombotic protection with hemorrhagic safety40. The use of DCBs, by eliminating the need for a permanent metallic scaffold, offers a compelling strategy to minimize both thrombogenic risk and the duration of antiplatelet therapy. This approach has garnered increasing interest for its potential to reduce bleeding events in vulnerable populations. This section aims to demonstrate that in patients at HBR, DCB-only strategies may allow reduction of antiplatelet therapy without compromising safety. The DEBUT trial was the first RCT to evaluate a DCB-only strategy in patients with de novo CAD and HBR41. Conducted across five Finnish centers, the trial compared PCBs with BMS in 208 patients. At 9 months, the primary endpoint of MACE occurred in only 1% of the DCB group compared to 14% in the stent group, demonstrating both non-inferiority and statistical superiority (p < 0.00001 for non-inferiority, p = 0.00034 for superiority)41. Importantly, all patients received only 1 month of DAPT, and no acute vessel closures occurred in the DCB group, underscoring the safety of the strategy in a high-risk cohort41. In addition, another observational study has shown that a 1-month DAPT strategy following elective DCB angioplasty for stable CAD is associated with excellent clinical outcomes, including the absence of stent thrombosis, cardiac death, or TLR at 6 months42. Further supportive evidence has emerged from real-world cohorts concerning the upfront use of single antiplatelet therapy (SAPT). In a Finnish single-center registry, 172 patients treated with a DCB-only approach were discharged on SAPT, most commonly aspirin or clopidogrel43. The majority fulfilled ARC-HBR criteria. At 12 months, the composite endpoint of cardiovascular death, myocardial infarction, and TLR was 1.4% in stable CAD and 7.1% in ACS, with all-cause mortality of 4.1% and 12.1%, respectively43. No acute or subacute vessel closures were reported43. Another registry analysis involving 107 patients treated with SAPT after DCB angioplasty for de novo lesions, often due to active bleeding or HBR, found no significant difference in MACE compared to patients on DAPT44. However, a significant reduction in BARC 2–5 bleeding events was observed in the SAPT group (6% versus. 9%, p = 0.04)44. Patients with chronic kidney disease (CKD) represent an important subgroup with HBR. In a propensity score-matched analysis including 830 patients with CKD undergoing PCI, a DCB-based strategy was associated with a significantly lower incidence of major bleeding at 2 years compared with a drug-eluting stent-only approach (2.3% vs. 5.9%, HR 0.41, 95% CI 0.19–0.92), while rates of MACE were comparable between the two strategies45. Notably, no cases of target lesion thrombosis were observed in the DCB-treated group. Within the DCB cohort, a DCB-only strategy was associated with an even lower risk of major bleeding compared with a hybrid strategy combining DCB and DES implantation (0.4% vs. 5.9%), while ischemic outcomes remained similar45. These data collectively suggest that a DCB-only approach permits substantial reduction in the intensity and duration of antithrombotic therapy without compromising procedural safety or long-term efficacy. One-month DAPT appears sufficient in stable CAD and may be further reduced to SAPT in selected patients with very HBR. Notably, even in ACS patients, a stepwise DAPT de-escalation strategy (aspirin + ticagrelor for 1 month, then ticagrelor monotherapy for 5 months, followed by aspirin monotherapy) after DCB angioplasty could be utilized to reduce bleeding risk46. However, the current evidence base is of low quality, limited by small sample sizes and lack of long-term follow-up, among others, mandating broader validation in larger and more diverse populations. Cost-effectiveness considerations Limited data are available regarding the cost-effectiveness of DCB angioplasty compared with DES. A contemporary real-world cost analysis including over 1900 patients with de novo CAD demonstrated no significant difference in overall costs between DCB-only and DES-based strategies, with comparable per-patient costs and slightly lower per-lesion costs in the DCB group47. Importantly, this analysis incorporated procedural costs, hospital stay, and antiplatelet therapy, and showed consistent results irrespective of the assumed duration of DAPT47. These findings suggest that potential savings related to reduced stent use and shorter DAPT may be offset by other procedural factors, and that cost considerations alone should not drive device selection47. Further health economic analyses, particularly across different healthcare systems and clinical indications, are warranted. Conclusion and future directions DCBs have evolved from a niche therapy for in-stent restenosis to a versatile platform with expanding roles in de novo coronary artery disease, small vessel interventions, and high bleeding risk scenarios. Their ability to deliver antiproliferative therapy without permanent implants offers a compelling advantage in preserving vascular physiology and minimizing long-term complications. While paclitaxel-based technologies have historically dominated the field, recent advancements in sirolimus formulations, leveraging micro-reservoirs, nanocarriers, and crystalline coatings, signal a potential paradigm shift, offering more favorable vascular healing profiles and broader therapeutic windows. However, significant questions remain. The absence of a class effect among DCBs underscores the need for rigorous, head-to-head trials that consider not only drug type but excipient design, tissue pharmacokinetics, and clinical endpoints. Greater emphasis on optimal lesion preparation, integration of intravascular imaging, and real-world evidence from high-risk populations will be pivotal in optimizing outcomes. Looking forward, innovations in drug delivery, including combination therapy, multi-drug coatings, and bioresorbable platforms, may expand the frontiers of DCB application. As we move toward a more personalized, anatomy-guided approach to PCI, DCB technology stands at the forefront of this transition, poised to redefine interventional strategy with efficacy, safety, and simplicity. Author contributions P.T. wrote the initial draft of the manuscript. A.P., K.D., P.K., P.I., N.P., K.A., K.T., D.T. revised the manuscript. All authors have read, reviewed, and approved the final version of the manuscript. Data availability No datasets were generated or analysed during the current study. Competing interests Panagiotis Theofilis acted solely as a Guest Editor at npj Cardiovascular Health. Panagiotis Theofilis was not involved in the journal’s review of, or decisions related to, this manuscript. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.