Drug-Coated Balloon Angioplasty in Patients Undergoing Complex Percutaneous Coronary Intervention

Background There are limited clinical data on drug-coated balloon (DCB)-based percutaneous coronary intervention (PCI) compared with drug-eluting stent (DES)-only PCI in patients with complex coronary artery lesions. Objectives The goal of the current study was to investigate the efficacy of DCB in patients undergoing PCI for complex coronary artery lesions. Methods From an institutional registry of patients with de novo complex coronary artery lesions, 126 patients treated with DCB-based PCI were compared with 234 propensity score–matched patients treated with DES-only PCI. Complex coronary artery lesions were defined as the presence of at least 1 of the following: bifurcation, chronic total occlusion, unprotected left main disease, long lesion ≥38 mm, multivessel disease, lesion requiring ≥3 devices, or severe calcification. The primary endpoint was target vessel failure (TVF) at 2 years, a composite of cardiac death, target vessel–related myocardial infarction, and target vessel revascularization. Results Baseline characteristics were comparable between the 2 groups. DCB-based PCI showed a comparable risk of TVF vs DES-based PCI (7.6% vs 8.1%; HR: 0.81; 95% CI: 0.33-1.99; P = 0.638). The risks of cardiac death (5.0% vs 5.7%; HR: 0.78; 95% CI: 0.24-2.49), target vessel–related myocardial infarction (0.9% vs 1.3%; HR: 2.65; 95% CI: 0.26-27.06), and target vessel revascularization (3.5% vs 2.0%; HR: 1.30; 95% CI: 0.30-5.67) were also comparable between the 2 groups. Conclusions DCB-based PCI showed comparable risks of TVF vs those of DES-only PCI in patients with complex coronary artery lesions. DCB might be considered as a suitable alternative device to DES in patients undergoing complex PCI. (Long-term Outcomes and Prognostic Factors in Patient Undergoing CABG or PCI; NCT03870815)

P ercutaneous coronary intervention (PCI) of complex coronary artery lesions accounts for >30% of all contemporary PCI procedures. 1 Despite the use of new-generation drug-eluting stents (DES), the risk of adverse clinical events is still higher in patients undergoing complex PCI than in those undergoing noncomplex PCI. 2,3In addition to a higher burden of cardiovascular risk factors and residual coronary atherosclerosis, procedural factors, including longer and multiple stents, are associated with an increased risk of adverse clinical events due to a higher risk of suboptimal stent implantation or delayed endothelization after complex PCI compared with noncomplex PCI. 4 Drug-coated balloons (DCBs) were developed under the concept of angioplasty without implantation of metal and designed as a semi-compliant balloon coated with antiproliferative agents that are delivered to the target vessel wall during inflation. 5nsidering that DES implantation is inherently associated with a 2% annual risk of late stent-related events, use of DCB has been considered as an alternative strategy that could reduce the extent and number of DES used during PCI.Nevertheless, clinical evidence of DCB use in the treatment of coronary artery disease (CAD) has mostly been confined to instent restenosis (ISR).7][8] However, the results are conflicting, and evidence regarding DCB use in various complex coronary artery lesions remains limited.
The goal of the current study was to investigate the efficacy of DCB-based PCI in patients undergoing PCI for various complex coronary artery lesions.

METHODS
This was a retrospective analysis of a prospective institutional registry from a tertiary referral center in the Republic of Korea.Patients were selected from  Unfractionated heparin or low-molecular-weight heparin was used for procedural anticoagulation according to standard protocol.
In the DCB-based PCI group, all DCBs used were paclitaxel-coated balloons, and procedures were performed according to international DCB consensus. 11,12The target lesion was routinely predilated by using an optimal-sized balloon with a balloon-to-reference vessel ratio of 0.8 to 1.0.Use of specialty balloons, including scoring, cutting, or noncompliant balloons, was allowed per operator discretion during pre-dilatation.If there was a flowlimiting dissection or >30% residual stenosis after lesion preparation, bail-out stent implantation was recommended rather than using DCB.After Study flow is presented.A total of 1,940 patients who had a complex coronary artery lesion were included from a prospective institutional registry.Among them, 324 patients with in-stent restenosis were excluded, leaving 138 patients with drug-coated balloon (DCB)-based percutaneous coronary intervention (PCI) and 1,478 patients with drug-eluting stent (DES)-only PCI.Propensity score matching with a 1:2 ratio was used for adjustment of baseline characteristics differences, leaving 126 patients with DCB-based PCI and 234 patients with DES-only PCI for analysis.MI ¼ myocardial infarction.Procedure-related MI was not included as a clinical event in the current analysis.Spontaneous MI was defined according to the third universal definition for MI. 13 Death of unknown cause was assumed to be cardiac related according to the definitions of the Academic Research Consortiums. 14    Among the total target lesions (n ¼ 630), 70.6% were type B2/C lesions, 2.1% were culprit lesions of ACS, 20.1% were bifurcation lesions, 45.6% were diffuse, 2.2% were thrombotic, 13.2% were CTO, 12.0% were ostial lesions, and 12.6% had significant calcification.  2 and 3.
The 2-year risk of TVF was comparable between the 2 groups across various complex coronary artery lesions (Figure 4).The treatment effect of DCB-based PCI was also comparable with DES-only PCI across various clinical characteristics without significant CTO ¼ chronic total occlusion; IABP ¼ intra-aortic balloon pump; PCPS ¼ percutaneous cardiopulmonary support; other abbreviations as in Table 1.interaction (Supplemental Figure 1).In a multivariable analysis of the original population, the independent predictors for TVF were diabetes mellitus, chronic kidney disease, and LM disease but not DCBbased PCI (HR: 1.18, 95% CI: 0.59-2.36;P ¼ 0.643) (Supplemental Table 4).Sensitivity analyses using inverse probability-weighted and propensity scores also showed comparable risk of TVF between the 2 groups (Supplemental Table 5).

DISCUSSION
In the current propensity score-matched cohort study, DCB-based PCI showed comparable risk of  1 and 3. DCB-based PCI without leaving polymer and metallic struts in the coronary artery is an attractive alternative strategy to DES-based PCI, as foreign materials could be a source of neointimal hyperplasia, neoatherosclerosis, and thrombosis resulting in stentrelated adverse events. 15,16However, concerns for acute vessel closure or restenosis due to recoil have hampered the widespread use of DCB, especially for de novo CAD. 17 European practice guidelines give a Class I recommendation for the use of DCB in the treatment of ISR but acknowledge that there are no convincing data to support DCB-based PCI for de novo CAD. 18[21] Because PCI of complex coronary lesions tends to require more stents as well as longer stents, which can be associated with increased risk of stent undersizing, late malapposition, ISR, and stent thrombosis, 22,23 DCB-based PCI could be beneficial for reducing stent burden and stent-related adverse events.However, previous data are limited to several observational studies conducted in specific subsets of complex coronary lesions, and their results are conflicting. 7,8,24,25Shin et al 8 showed that the DCB group had a lower rate of 2-year major adverse cardiovascular events than the DES group (3.9% vs 11.0%) in multivessel disease.Similarly, Gitto et al 25 reported that DCB-based PCI had a lower incidence of target lesion failure than DES in do novo lesions on the left anterior descending artery.In contrast, 3-year event rates were similar between the DCB and DES groups with diffuse coronary lesions (TLR: 7.3% vs 8.3%, respectively) 7 and CTO lesions (major adverse cardiovascular events: 12.0% vs 11.8%). 24In the current It should be noted that there were some differences in procedural characteristics.First, DCB-based Considering that the risk of definite or probable stent thrombosis was higher after complex PCI than noncomplex PCI, 2 DCB-based PCI might be a reasonable option in patients with high bleeding risk and complex coronary lesions. 29 The current study evaluated clinical outcomes of drug-coated balloon (DCB)-based percutaneous coronary intervention (PCI) compared with drug-eluting stent (DES)-only PCI in patients undergoing PCI for various complex coronary lesions.DCBs were more frequently used for chronic total occlusion (CTO), multivessel coronary artery disease, and lesions requiring multiple devices.DCB-based PCI showed comparable risk of target vessel failure (TVF) at 2 years with DES-only PCI in patients with de novo complex coronary artery lesions.The comparable risk of TVF was consistently observed in various subgroups of complex coronary artery lesions.
was used less frequently for heavily calcified lesions and unprotected LM disease.Considering the higher possibility of suboptimal luminal gain and need for prolonged balloon inflation for delivery of antiproliferative drugs during DCB-based PCI, these subsets of lesions might not be ideal targets for DCB-based PCI. 17

2
institutional registries of Samsung Medical Center (Seoul, Republic of Korea), a registry of PCI in which DCBs have been used, enrolling from January 2016 to December 2019, and another registry of DES use, enrolling from January 2012 to December 2015.Previous reports were published using part of the DES registry. 9,10A total of 1,940 patients who had a complex coronary artery lesion and underwent PCI were selected.Complex coronary artery lesions were defined as 1 of the following: 1) bifurcation lesions with side branch diameter $2.5 mm; 2) chronic total occlusions (CTOs) with an occlusion duration $3 months; 3) unprotected left main (LM) disease; 4) long lesions (used stent or DCB length $38 mm; 5) multivessel PCI ($2 major epicardial coronary vessels treated at 1 PCI session); 6) multiple devices used ($3 DCBs or stents per patient); 7) ISR lesion; or 8) severely calcified lesion (requiring a rotablation).For the purposes of the current study, 324 patients with ISR were excluded.Patients whose PCI involved use of DCB (thus, both DCB-only and DCB and DES hybrid PCI, including bail-out DES implantation due to suboptimal DCB angioplasty) were classified into the DCB-based PCI group (n ¼ 138) and the others into the DES-only PCI group (n ¼ 1,478).To adjust for the significant differences of baseline characteristics between the 2 groups, 126 patients with DCB-based PCI were matched with 234 patients with DES-only PCI using propensity scores (Figure 1).The study protocol was approved and the requirement for informed consent was waived by the Institutional Review Board of Samsung Medical Center.The current study was conducted according to the principles of the Declaration of Helsinki.The A B B R E V I A T I O N S A N D A C R O N Y M S ACS = acute coronary syndrome(s) CTO = chronic total occlusion DCB = drug-coated balloon DES = drug-eluting stent(s) LM = left main MI = myocardial infarction PCI = percutaneous coronary intervention TLR = target lesion revascularization TVF = target vessel failure

FIGURE 1
FIGURE 1 Study Flow Both target vessel revascularization and target lesion revascularization were clinically driven.They were defined as revascularization at the previously treated segment from 5 mm proximal to the stent to 5 mm distal to the stent with $70% diameter stenosis and at least 1 of the following: 1) recurrence of angina; 2) positive noninvasive test result; or 3) positive invasive physiological test result.Definite, probable, and possible stent thromboses were defined according to the definitions of the Academic Research Consortiums.The mortality data for patients lost to follow-up were confirmed by National Death Records.All clinical events were adjudicated by expert interventional cardiologists blinded to treatment strategy.STATISTICAL ANALYSIS.All data were analyzed on a per-patient basis.Vessels with the most severe stenosis were selected as the representative vessel in patients with multivessel PCI at the time of the index procedure.Additional lesion-level analysis of target lesion characteristics was performed to compare lesion complexity between the 2 groups.All discrete and categorical variables are presented as numbers and relative frequencies (percentages).Continuous variables are expressed as mean AE SD or median (IQR) according to their distribution, checked by using the Kolmogorov-Smirnov test and visual inspection of a Q-Q plot.Categorical variables were compared by using the chi-square test, and continuous variables were compared by using Student's t-test or Mann-Whitney U test according to their distribution.To adjust for uneven distribution of baseline characteristics according to treatment strategies, 1:2 propensity score matching was performed using a caliper width of 0.1 between the 2 groups.Propensity score was calculated by using logistic regression, with variables including age, sex, hypertension, diabetes mellitus, hyperlipidemia, current smoker, chronic kidney disease, previous MI, previous PCI, previous coronary artery bypass graft, ACS, extent of disease, and multivessel disease.The covariate balance after propensity score matching was measured by calculating the absolute standard mean differences.Absolute standard mean differences were within 10% across all matched covariables, suggesting balance achievement between the 2 groups (Supplemental Patients with DCB-based PCI had higher proportions of CTO and ostial lesions than those with DES-only PCI.Conversely, the DCB-based PCI group had a lower proportion of unprotected LM disease than the DES-only PCI group.Although baseline stenosis severity and the total length of target lesions were comparable between the 2 groups, post-PCI diameter stenosis was higher in the DCB-based PCI group than in the DES-only PCI group.Table 3 summarizes procedural characteristics in the matched population.Compared with the DESonly PCI group, DCB-based PCI was more frequently performed in multivessel CAD and CTO.Conversely, unprotected LM disease was treated more with DES-only PCI (Figure 2).The mean number of devices used (2.6 AE 1.3 vs 1.9 AE 0.9) was higher, but mean diameter (2.62 AE 0.32 mm vs 3.00 AE 0.42 mm) and total length of devices used (43.8AE 29.0 mm vs 48.8 AE 26.1 mm) were lower in the DCB-based group than in the DES-only group.Comparisons of target vessel and procedural characteristics in the original population are provided in Supplemental Tables

FIGURE 2
FIGURE 2 Proportion of Complex Coronary Artery Lesions in the Study Population

FIGURE 3
FIGURE 3 Comparison of Target Vessel Failure Between DCB-Based vs DES-Based PCI study, DCB-based PCI showed similar rates of 2-year TVF compared with DES-only PCI in various complex coronary artery lesions, including bifurcation lesion, CTO, unprotected LM disease, long lesion, multivessel CAD, lesion requiring $3 devices, or severely calcified lesion.

FIGURE 4
FIGURE 4 Comparison of Target Vessel Failure According to Lesion Characteristics Conversely, there was a notable propensity of DCB use toward specific types of lesions such as CTO, multivessel CAD, and ostial lesions.Such lesions would be suitable for treatment with DCB because of the presence of collaterals and the benefit of DCB reducing the number of total stents, which is likely to be relatively high in CTO and multivessel disease.Ostial lesions could be treated both by DCB or DES, but DCB might aid in treating the region with size discrepancy between the proximal and distal segment.Even taking these procedural differences into account, comparable prognosis of DCB-based PCI with DES-only PCI was consistent across various subsets of complex coronary lesions without significant interaction, and DCB-based PCI was not an independent predictor for 2-year TVF in multivariable analysis.In addition, sensitivity analyses of the current study support the comparable efficacy between DCB-based PCI and DES-only PCI.The composition of treated lesion types in each group was significantly different, reflecting daily practice with DCB, and multivariable adjustments for LM disease, CTO, and multivessel PCI did not change the result.The current results imply that DCB-based PCI could be a reasonable treatment strategy to reduce stent burden and stent-related adverse events without concern for increased risk of TVF. 30 Further study is warranted to confirm these findings and to identify lesion types that would maximize the clinical efficacy of DCB.Ongoing trials such as DCB-HBR (Drug-Coated Balloon in Patients With High Bleeding Risk; NCT05221931) and REVERSE (Drug-Coated Balloon vs. Drug-Eluting Stent for Clinical Outcomes in Patients With Large Coronary Artery Disease; NCT05846893) will provide more evidence of DCB-based PCI for the treatment of de novo CAD.STUDY LIMITATIONS.First, because this was a nonrandomized observational study, inherent limitation of selection bias should be considered.Although we used multiple adjustment methods, there were several differences, especially in type of target lesion characteristics.Second, the current results cannot be applied to unprotected LM disease and severely calcified lesions.Third, the current study evaluated 2 years of follow-up data, and longer-term efficacy data of DCB-based PCI are thus not available.Fourth, only paclitaxel-coated DCBs were used in this study, and the results therefore might not be applicable for sirolimus-coated DCBs, as a class effect for DCB cannot be confidently assumed. 18Fifth, because the 2 study groups were selected from the various registries of different enrollment periods, variations in practice and other procedure-related factors could have acted as a bias.However, DES of the same generation were used during the total enrollment period, and multiple adjustments have been performed to compensate other discrepancies, such as exclusive inclusion of complex coronary lesion and propensity score matching.Sixth, DCB-based PCIs were heterogeneous, including DCB-only PCI, DCB and DES hybrid PCI, and bail-out DES implantation due to suboptimal DCB angioplasty.Seventh, the data related to bleeding, such as duration of post-PCI antiplatelet agents used and occurrence of severe bleeding complications, were not recorded in the registry, making analysis about the effect of DCB on bleeding impossible in this study.Eighth, some procedural data were not available, such as type of DCB and DES used, type of dissection observed in PCI, and the degree of residual stenosis leading to bailout stenting.CONCLUSIONS DCB-based PCI showed comparable risks of TVF vs DES-only PCI in patients with complex coronary artery lesions.This study suggests that DCB might be considered as a suitable alternative device to DES in patients undergoing complex PCI.Further study is needed to validate the current results.FUNDING SUPPORT AND AUTHOR DISCLOSURES Dr J.M. Lee has received an institutional research grant from Abbott Vascular, Boston Scientific, Philips Volcano, Terumo Corporation, Zoll Medical, and Donga-ST.Dr Hahn has received an institutional research grant from the National Evidence-based Healthcare Collaborating Agency, Ministry of Health & Welfare of the Republic of Korea, Abbott Vascular, Biosensors, Boston Scientific, Daiichi-Sankyo, Donga-ST, Hanmi Pharmaceutical, and Medtronic Inc. Dr Gwon has received an institutional research grant from Boston Scientific, Genoss, and Medtronic Inc.All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Gwangmyeong, Republic of Korea; h Gyeongsang National University School of Medicine, Gyeongsang National University Hospital, Jinju, Republic of Korea; i Uijeongbu St. Mary's Hospital, The Catholic University of Korea, Seoul, Republic of Korea; j Seoul St.
l Inha University Hospital, Incheon, Republic of Korea; m Division of Cardiovascular Medicine, Department of Internal Medicine, Seoul National University Bundang Hospital, Seongnam, Republic of Korea; n Department of Internal Medicine, College of Medicine, Chungbuk National University, Cheongju, Republic of Korea; o Yonsei University Wonju College of Medicine, Wonju Severance Christian Hospital, Wonju, Republic of Korea; and the p Department of Medicine, Inje University Ilsan Paik Hospital, Goyang, Republic of Korea.*DrsJoh and Kwon equally contributed to this study as first authors.The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors' institutions and Food and Drug Administration guidelines, including patient consent where appropriate.For more information, visit the Author Center.Manuscript received January 22, 2024; revised manuscript received April 5, 2024, accepted April 16, 2024.institutionalcardiovascularcatheterizationdatabase of Samsung Medical Center is registered at Clinical-Trials.gov (NCT03870815).given24hoursbefore PCI or 600 mg was given on the day of PCI, unless patients were already taking clopidogrel for at least 5 days.A loading dose of prasugrel 60 mg or ticagrelor 180 mg was given before PCI in selected patients with acute coronary syndrome (ACS), unless patients were already receiving potent P2Y12inhibitor therapy for at least 7 days.

Table 1 )
. In addition, inverse probability-weighted, propensity score-adjusted, and propensity scorestratified Cox proportional hazards regression analyses were conducted for sensitivity analysis in the original population.The cumulative incidence of clinical events is presented as Kaplan-Meier estimates and were compared by using a log-rank test.Stratified multivariable Cox proportional hazards regression with matched pairs as strata was used to calculate adjusted HRs and 95% CIs to compare the risk of clinical events between the matched groups.The assumption of proportionality was assessed by using Schoenfeld residuals and graphically by the log-log plot.The adjusted covariables were age, sex, diabetes mellitus, clinical presentation, target vessel location, multivessel PCI, PCI of CTO lesion, PCI of unprotected LM disease, and concomitant use of aspirin and a P2Y 12 inhibitor.Subgroup analysis of the primary outcome was performed according to clinical and lesion characteristics of interest between the 2 groups.The interaction between treatment effect and the RESULTS BASELINE CHARACTERISTICS.Table 1 presents baseline characteristics of the original and propensity score-matched populations.Before matching, pawith ACS, 294 patients (81.7%) had multivessel disease, and 50 patients (13.9%) had unprotected LM disease.TARGET VESSEL, TARGET LESION, AND PROCEDURAL CHARACTERISTICS.

Table 2
presents target vessel and lesion characteristics in the matched population.

TABLE 2
Comparison of Target Vessel and Lesion Characteristics According to Treatment Strategy in Matched Population Values are n (%) or mean AE SD unless otherwise indicated.ACC ¼ American College of Cardiology; ACS ¼ acute coronary syndrome; AHA ¼ American Heart Association; LAD ¼ left anterior descending; LCX ¼ left circumflex; RCA ¼ right coronary artery; other abbreviations as in Table1.

TABLE 3
Comparison of Procedural Characteristics According to Treatment Strategy in the Matched Population Values are mean AE SD or n (%).

TABLE 4
Comparison of Clinical Outcomes at 2 Years According to Treatment Strategy in Matched Population Adjusted covariables were age, male, diabetes mellitus, clinical presentation, target vessel location, multi-vessel PCI, PCI of CTO lesion, PCI of unprotected LM disease, and concomitant use of aspirin and P2Y12 inhibitor.
a NA ¼ not applicable; other abbreviations as in Tables Third, DCB-based PCI CENTRAL ILLUSTRATION Efficacy of Drug-Coated Balloon and Drug-Eluting Stent in Percutaneous Coronary