Clinical Outcomes of Drug-Coated Balloon in Coronary Patients with and without Diabetes Mellitus: A Multicenter, Propensity Score Study

Background Relative to nondiabetic patients, percutaneous coronary intervention (PCI) in patients with diabetes mellitus (DM) is associated with inferior clinical outcomes. We aimed to evaluate the outcomes of drug-coated balloon (DCB) in diabetic versus nondiabetic patients. Methods and Results In this observational, prospective, multicenter study, we compared the outcomes of patients with and without DM after undergoing PCI with DCBs. Target lesion failure (TLF) was analyzed as primary endpoint. Secondary endpoints were the rates of target lesion revascularization (TLR), major adverse cardiovascular events (MACE), cardiac death, myocardial infarction (MI), and any revascularization. Propensity score matching was used to assemble a cohort of patients with similar baseline characteristics. Among 2,306 eligible patients, 578 with DM and 578 without DM had similar propensity scores and were included in the analyses. During follow-up (366 ± 46 days), compared with DM patients, patients without DM were associated with a lower yearly incidence of TLF (2.77% vs. 5.36%; OR, 1.991; 95% CI, 1.077 to 3.681; P = 0.025) and TLR (1.90% vs. 4.15%; OR, 2.233; 95% CI, 1.083 to 4.602; P = 0.026). No significant differences were observed with regards to rates of MACE (OR: 1.580, 95% CI: 0.912-2.735; P = 0.100), cardiac death (OR: 1.608, 95% CI: 0.523-4.946; P = 0.403), MI (OR: 4.042, 95% CI: 0.855-19.117; P = 0.057), and any revascularization (OR: 1.534, 95% CI: 0.983-2.393; P = 0.058). Conclusions Diabetic patients experience higher TLF and TLR rates following DCB angioplasty without substantial increase in the risk of MACE, cardiac death, MI, or revascularization.


Introduction
Patients with diabetes mellitus (DM) are at increased risk of coronary artery disease (CAD). DM patients often present with a combination of diffuse coronary lesions and small vessel disease, indicating difficult stent delivery and high restenosis rates postpercutaneous coronary intervention (PCI). Regarding coronary revascularization, recent technological advances have expanded PCI use to more complex lesions [1,2], especially drug eluting stents (DES), which have markedly reduced the rate of restenosis and repeat revascularization [3,4]. However, CAD morbidity and mortality in DM patients remain high, even with current use of DES [5,6].
Drug-coated balloon (DCB) is effective in treating instent restenosis (ISR) [7,8]. Various clinical studies have revealed its effectiveness against de novo coronary lesions [9][10][11], especially in small vessel disease. Differences in study results are mainly due to differences in study approach, which may use the "DCB-only" or "hybrid" strategy. It also preserves access for future coronary artery bypass grafting (CABG).
To date, no studies have investigated the use of DCB angioplasty-only in diabetic patients with coronary artery disease. Here, we evaluated the outcomes of DCBs in diabetic versus nondiabetic patients.

Study Population.
Patients were retrospectively enrolled at 3 Chinese centers from July 2014 to December 2019. Only patients with coronary vessel lesions sized between 2.0 and 4.0 mm, whether de novo or instent restenosis, and were eligible for the study. The exclusion criteria were ≥ type C dissection or >30% residual stenosis after lesion preparation, simultaneous ISR treatment and de novo lesions, revascularization within 1 month prior to index procedure, and unstable hemodynamics or cardiogenic shock ( Figure 1).
Patients were given 300 mg aspirin before intervention, unless they were on long-term aspirin treatment. Clopidogrel or ticagrelor was administered at loading doses of 600 and 180 mg, respectively. Patients who underwent DCB only were put on dual antiplatelet therapy (DAPT) for at least 1 month after the procedure. Those receiving DCB and stent implantation received DAPT in accordance with established guidelines [12,13]  hypersensitivity to DAPT, heparin, paclitaxel, or limus; women with childbearing potential; and those with a life expectancy below a year were excluded. Ethical approval for the study was granted by the First Affiliated Hospital of Zhengzhou University Institutional Review Board/Ethics Committee. All patients gave written informed consent. Data were collected using a common electronic case report form.

Study
Procedure. During intervention, special emphasis was paid to adequate lesion preparation prior to DCB treatment as recommended in the latest expert consensus [14]. Predilatation with semicompliant balloon, noncompliant balloon, scoring balloon, or cutting balloon with a balloonto-vessel ratio of 0.8-1.0 was mandatory (0.8-1.0 for de novo lesions and 1.0 for ISR lesions). Next, DCB angioplasty was done only in the absence of a major, flow-limiting dissection (≥ type C according to the NHLBI classification [15] and where residual stenosis was ≤30% based on at least 2 perpendicular angiographic views). In this study, DCB had a paclitaxel/iopromide matrix coating (SeQuent™ Please, B. Braun, Melsungen, Germany). To avoid geographic mismatch, DCB catheter length was designed to exceed that of the target lesion by at least 2 mm. DCB diameters were adapted to reference vessel diameters with a balloon-tovessel ratio of 0.8-1.0. Recommended inflation time was at least 30 seconds at >7 bars. New-generation DESs were implanted if DCB-only outcomes were unsatisfactory due to severe residual stenosis or dissections.

Clinical Endpoints and
Definitions. The study's primary outcome was one-year target lesion failure (TLF), a composite of cardiac death, target vessel myocardial infarction (MI), and target lesion revascularization (TLR). Various secondary outcomes were evaluated, including the rates of TLR, major adverse cardiovascular events (MACE, defined as the composite outcome of cardiac death, myocardial infarction (MI), and target vessel revascularization (TVR)), cardiac death, MI, and repeat revascularization (including PCI and CABG). Cardiac death is defined as death resulting from cardiovascular causes. And undetermined cause of death is defined as a death not attributable to any other category because of the absence of any relevant source documents. Such deaths will be classified as cardiovascular for end point determination [16,17]. MI was defined by typical clinical symptoms, ECG changes, and/or elevated cardiac troponin values with at least one value above the 99 th percentile upper reference limit (type 4b or 4c MI, except for perioperative MI) [18]. Patients' follow-up was by telephone or outpatient visits at 12 months.

Statistical
Analysis. Data analysis was done using R statistics packages (http://www.r-project.org) and Empower (R) (http://www.empowerstats.com, X & Y Solutions Inc.). Categorical variables are presented as frequencies (percentages) and continuous variables as mean ± SD. Comparisons between patients with and without DM were done using Fisher's exact test for each variable. Mann-Whitney Wilcox nonparametric tests were used for continuous variables. Given the baseline characteristic differences between eligible participants in the 2 groups of the observational study, a 1 : 1 propensity score matching (PSM) was used to select patients with comparable baseline data. After evaluation of covariates that were clinically and/or statistically associated with the treatment group and removal of repeatedly defined or collinear variables, including baseline characteristics, risk factors,

Sensitivity Analysis.
In the overall population, TLF exhibited good consistency before and after PSM. The incidence rates of TLR

Discussion
In this observational study, we evaluated the outcomes of PCI with DCB in diabetic versus nondiabetic patients treated for CAD at three participating centers. Our data suggest that relative to nondiabetic patients, diabetic patients treated with PCI with DCB exhibit higher incidence of TLF and TLR than nondiabetic patients. These findings enhance our current understanding of the safety and efficacy of DCB in all comers in contemporary clinical practice. However, the incidence rates of MACE, cardiac death, and any revascularization in the 2 groups were similar, suggesting that DCB angioplasty can serve as the default treatment option for such patients. Because our results are based on matching propensity scores, our findings are unlikely to result from negative confounding. Moreover, reliability of these findings was validated via sensitivity analysis methods like subgroup analysis.
CAD in the presence of DM has unique characteristics. The higher risk of ISR in patients with DM is secondary to complex pathophysiological mechanism, including endothelial dysfunction, serious vascular inflammation, high activated platelet levels, and higher levels of advanced end products of glycosylation [19,20]. ISR risk within firstgeneration DES is higher due to sustained drug release and the inflammatory effects of the polymer [21]. On how to explain the increased risk of stent undersizing or underexpansion and subsequent increase in stent-related complications, a past study found that when compared to non-DM patients, diabetic patients treated with DES have greater residual plaque burden throughout the reference segment [22].
Development of second-generation DES with biocompatible polymers and thinner struts has improved diabetic patients' outcomes upon PCI [23]. However, diabetes remains an independent predictor of major adverse events     Journal of Diabetes Research [24]. In a recent study of 1,919 patients who received PCI with 2 different new-generation DES, diabetic patients had higher target lesion failure (TLF) risk (cardiac death, target vessel myocardial infarction, or ischemia-driven TLR) relative to non-DM patients (7.8% vs. 4.2%, P = 0:002), mainly due to a higher TLR rate (4.5% vs. 2.0%, P = 0:002) [6]. In diabetic patients, DCB has various advantages over DES, including even distribution of the antiproliferative drug along the vessel wall, resulting in better positive remodeling [25,26]. In addition, local inflammatory reactions in diabetic patients are often severe, and DCB releases higher drug concentrations in the short-term, which is more conducive to inhibit inflammatory reactions. Although the newgeneration DES uses more biocompatible or absorbable polymer, the continuous stimulation of the stent struts may cause the local inflammation to persist.
To our knowledge, no studies have investigated the use of DCB-only strategy in diabetic patients with CAD. The  Journal of Diabetes Research PEPCAD IV DM study randomly allocated 84 diabetic patients with significantly stenosed native coronary arteries to a paclitaxel-coated PTCA-balloon SeQuent™ Please, followed by deployment of the cobalt-chromium stent Coro-flex™ Blue treatment group or to a paclitaxel-eluting stent Taxus™ Liberté treatment group [11]. At 9 months angiographic follow-up, the primary outcome of mean insegment late lumen loss was 0:37 ± 0:59 mm in the DCB + bare metal stents (BMS) group vs. 0:35 ± 0:63 mm in the DES group (P = 0:89). MACE rates were also similar in both groups ( , and LLL (mean difference, -0.13 mm, 95% CI: -0.41-0.14, P = 0:34) were similar to those of DES. Finally, it is inferred that in diabetic patients with de novo coronary lesions undergoing PCI, DCBs are associated with similar outcomes relative to first-generation DES, with a signal toward potential benefit in lowering target lesion revascularization. In the early stage, BMS implantation after DCB angioplasty was to avoid safety problems caused by excessive local antiproliferative drug concentration. However, the safety of DES implantation at the same location of DCB angioplasty has been verified in those patients with bailout DES deployment [29,30]. What needs to be pointed out is that whether it is the combination of DES + DCB or DES only, the incidence rates of ISR, stent thrombosis, and other adverse events that are lower than these in combination of DCB and BMS. In addition, the control groups of these studies are all 1st-generation DES. Although antiproliferative drug of DCB and 1st-generation DES is the same, as far as current clinical practice is concerned, the significance of such a comparison is relatively small.
Here, relative to non-DM patients, PCI with DCB in diabetic patients was associated with higher TLF risk before and after PSM. And binary logistic regression analysis of TLF in the overall population revealed that DM increased TLF risk. Most importantly, the incidences of TLR, MACE, cardiac death, and any revascularization in DCB-treated diabetic patients were much lower relative to DES-treated patients reported in past studies. If the answer is yes, the choice between DCB and DES, even PCI and CABG, should be made a great change.
4.1. Limitation. Because this was a nonrandomized, observational study, it suffers from potential selection and ascertainment bias despite our robust propensity score matching. Additionally, as part of PCI strategy, 32.6% of the patients received DES implantation in a different coronary artery in the same operation, which may affect the occurrence of clinical events. Lastly, this study only evaluated DCB application in patients with coronary heart disease and diabetes and could not be compared to DES over a same period. Thus, a prospective randomized clinical trial comparing the use of DCB and DES in diabetic coronary heart disease patients is highly warranted as it may have important clinical significance.

Conclusion
In conclusion, our findings suggest that relative to non-DM patients, DM patients experience higher TLF and TLR rates upon DCB angioplasty. However, there was no substantial increase in the risk of MACE, cardiac death, MI, and any revascularization attributable to DM.

Data Availability
The data sets generated during and/or analysed during the current study are not publicly available but are available from the corresponding author on reasonable request.