The design and enrollment characteristics of 2 multicenter registries (IRIS-MAIN and KOMATE) have been published previously.7,8 Briefly, the IRIS-MAIN is a nonrandomized, multinational, observational registry gathering data on consecutive patients with unprotected LMCA disease. Another study population was derived from the KOMATE registry, which includes 8 coronary intervention centers in Korea. Both data sources had an “all-comers” design to evaluate the characteristics, treatments, and clinical outcomes of patients with LMCA disease in the real-world setting. Among merged individual-level data, we only included patients with LMCA who were treated with second-generation DES with available accurate information on postprocedural DAPT duration. The exclusion criteria were minimal; patients who underwent bare-metal stent (BMS) or first-generation DES. In addition, we excluded patients in whom we could not accurately judge the effectiveness of DAPT (patients with follow-up loss within 30 days, patients with in-hospital events within 30 days). A flow diagram of the current analysis is shown in figure 1. This study was approved by the institutional review board of each participating center for merged data use and all patients provided written informed consent. The study was performed in accordance with the Declaration of Helsinki and was approved by the research ethics committee of each participating center, and written informed consent was obtained by all participants.

Figure 1.

Study flow chart. BMS, bare-metal stent; DAPT, dual antiplatelet therapy; DES, drug-eluting stent; IRIS-MAIN, Interventional Cardiology Research Incorporation Society-Left Main Revascularization Study; KOMATE, Korean Multicenter Angioplasty Team Study.

(0.25MB).

All PCI procedures were conducted in accordance with local guidelines using standard techniques. Intraprocedural anticoagulation was maintained with unfractionated or low molecular weight heparin to achieve an activated clotting time of 250 to 300 s. Other procedural factors, such as access location, stent strategy, stent technique, and use of intravascular ultrasound, were left to the operator's discretion. Although the duration of DAPT (aspirin plus P2Y12 inhibitors [clopidogrel, ticagrelor, or prasugrel]) was usually recommended according to the current practice guidelines,4 the final duration was left to the treating physician's discretion with consideration of the patient's clinical and procedural characteristics and other comorbid medical conditions. The DAPT duration criterion was allowed up to 2 months in addition to the date of use, considering the time of the outpatient department visit. All clinical, angiographic, procedural, and outcome data were collected using a web-based reporting system. To identify the status of antiplatelet therapy, the dates and duration of prescribed antiplatelet agents were obtained from the electronic prescribing system of each hospital. Additional information was obtained by further inquiry into medical records or telephone contact, if necessary.

The efficacy outcome of the study was major adverse cardiovascular events (MACE), defined as a composite of cardiac death, fatal or nonfatal acute myocardial infarction, and stent thrombosis events. The secondary outcomes were all-cause mortality and target vessel revascularization. Fatal or nonfatal acute myocardial infarction was defined as an increase in the creatine kinase-myocardial band or troponin level to the 99th percentile of the upper limit of normal with ischemic symptoms or electrocardiographic findings indicative of ischemia not related to the index procedure (ie, procedural myocardial infarction was disregarded). Stent thrombosis was defined as definite stent thrombosis according to the Academic Research Consortium definition.9 The safety outcome was major bleeding. Bleeding events were defined as minor or major bleeding as per the Thrombolysis in Myocardial Infarction (TIMI) bleeding criteria.10 All clinical outcomes were independently adjudicated by an independent group of clinicians who were unaware of DAPT duration and types of DES.

Continuous variables are reported as mean± standard deviation and were analyzed using 1-way ANOVA. Categorical variables are reported as frequencies (percentages) and were analyzed using chi-square tests. Survival curves were prepared using Kaplan-Meier analysis and analyzed using log-rank tests. Cox proportional hazard regression analysis was used to identify independent predictors of primary endpoints and to estimate hazard ratios (HR) and 95% confidence intervals (95%CI) for clinical outcomes. To minimize confounding and residual selection bias in observational treatment comparisons, a propensity score weighting method was applied to control imbalances in various baseline characteristics across different groups of DAPT duration. For multigroup comparisons, multiple propensity scores were estimated using the Toolkit for Weighting and Analysis of Nonequivalent Groups (TWANG) method, and the corresponding inverse probabilities of treatment weight (the reciprocals of the propensity scores) were estimated by using generalized boosted models through an iterative estimation procedure.11 To calculate the propensity score, key clinical, anatomic, and procedural characteristics, such as age, sex, hypertension, diabetes mellitus, dyslipidemia, smoking, ACS, chronic kidney disease, multivessel disease, stent strategy, previous PCI, and intravascular ultrasound, were included. The balance of the pretreatment covariates was assessed, and significant improvement in baseline was achieved after weighting. Furthermore, the performance of this propensity model was confirmed by comparing the distributions of standardized mean differences of covariates and propensity scores between these groups before and after inverse probabilities of treatment weight.

Cox regression models with test for interaction were used to evaluate the consistency of treatment effects in multiple subgroups. The statistical analysis was performed using IBM SPSS version 23.0 (IBM, United States) and R software version 3.5.2 (R Project for Statistical Computing).

From July 2006 to August 2017, 1827 patients with LMCA lesions treated by PCI with second-generation DES and with valid data on DAPT duration were included in the final analytic datasets (figure 1). The practice pattern of DAPT duration in these patients is shown in figure 2. The median DAPT duration was 398 (interquartile range [IQR], 360-730) days for the entire population. According to DAPT duration, patients were categorized into 4 groups: DAPT <6 months (n=273), 6 to 12 months (n=477), 1224 months (n=637), and ≥ 24 months (n=440). The median DAPT durations were 99 (IQR, 36.7-180) days in DAPT <6 months, 365 (IQR, 341-365) days in DAPT 6 to 12 months, 523.5 (IQR, 397-730) days in DAPT 12 to 24 months, and 1095 (IQR, 1004-1316.5) days in DAPT ≥ 24 months, respectively. Since we enrolled patients for a long period of time (2006-2017), we investigated the pattern of DAPT duration according to the 2 different periods (2006-2011 and 2012-2017). The median duration of DAPT was longer in the 2006 to 2011 period (IQR, 497days) than in the 2012 to 2017 period (IQR 381 days). The pattern of DAPT between the 2 periods was the same as shown in figure 1 of the supplementary data.

Figure 2.

Distribution of participants according to DAPT duration. Values are presented as No. (%). DAPT, dual antiplatelet therapy.

(0.11MB).

The baseline clinical characteristics of the study population are shown in table 1. The clinical characteristics were similar between the 4 groups, except for dyslipidemia and history of previous PCI. DAPT duration tended to be significantly shorter in patients with stable angina than in those with ACS. The DAPT and procedural characteristics of the study population are shown in table 2. Regarding P2Y12 inhibitors, most patients received clopidogrel (94.6%), and ticagrelor and prasugrel was used in 4.1% and 1.3% of the patients, respectively.

The percentage of multivessel disease was the lowest in the DAPT <6 months group, and the highest in the DAPT ≥ 24 months group. The extent of CAD was similar between the 4 groups. Distal bifurcation involvement was lowest in the DAPT <6 months group and highest in the DAPT ≥ 24 months group. The average stent diameter and postdilatation balloon diameter were largest in the DAPT ≥ 24 months group. Postdilatation balloon pressure was the highest in the DAPT ≥ 24 months group. Patients who underwent PCI with intravascular ultrasound were significantly more likely to receive DAPT for ≥ 24 months.

The median length of follow-up among all patients was 47.9 [36.7-60.8) months. The number of patients lost to follow-up was 73 out of 1827 (3.99%). First, we fit Cox proportional-hazards log-linear models adjusted by baseline and procedural parameters with thin-plate spline curves to DAPT duration (figure 3). When we analyzed the risk of MACE and major bleeding according to DAPT duration as a continuous variable, risk of MACE progressively decreased, but the risk of major bleeding increased with>12 months of DAPT. We investigated Cox proportional-hazards log-linear models with thin-plate spline curves to DAPT duration according to stent strategies and with or without bifurcation lesions. Regardless of these factors, the results remained consistent (figure 2 of the supplementary data).

Figure 3.

Spline curves. Duration-response relationships between DAPT duration and MACE (A), and between DAPT duration and major bleeding (B) after PS weighting tested by a log-linear model with thin-plate spline curves. DAPT, dual antiplatelet therapy; HR, hazard ratio; MACE, major adverse cardiovascular event; PS, propensity score.

(0.06MB).

The cumulative incidences of clinical outcomes according to the categories of DAPT duration are shown in table 3 and in figure 2 of the supplementary data. The cumulative rates of MACE occurred more frequently in the DAPT <6 months group than in other DAPT groups (lowest for the DAPT ≥ 24 months [1.6%] and highest for the DAPT <6 months [5.1%]). There was no difference in the cumulative incidence of major bleeding (lowest for the DAPT <6 months [1.1%] and highest for DAPT ≥ 24 months [3.6%]).

The distributional balance of propensity scores according to DAPT duration before and after weighting is shown in figure 4 of the supplementary data. The adjusted risks for adverse clinical events according to the different categories of DAPT duration after application of multiple treatment propensity score weighting are shown in figure 4. With DAPT 12 to 24 months as the reference group, adjusted hazard ratios for MACE were significantly higher for DAPT <6 months and DAPT 6 to 12 months than for DAPT 12 to 24 months (HR, 4.51; 95%CI, 2.96-6.88 and HR 1.92; 95%CI, 1.23-3.00). Major bleeding events tended to be more likely in the ≥ 24 month group than in the other shorter DAPT groups. Multivariate analysis revealed that diabetes mellitus and chronic kidney disease were independent predictors of MACE (table 1 of the supplementary data).

Figure 4.

Central illustration. Comparison between outcomes of different DAPT durations in propensity score analyses. Adjusted HR. CI, confidence interval; HR, hazard ratio; MACE, major adverse cardiovascular events; PS, propensity score.

(0.22MB).

There are several limitations of our study. First, although multiple propensity score treatment analysis was performed, this study was an observational, nonrandomized study; therefore, it might be vulnerable to inherent limitations, including selection bias and unmeasured confounders. Thus, the overall observed findings should be interpretated as provisional and hypothesis-generating only. These findings should be confirmed or refuted through large randomized clinical trials. Second, owing to the limited number of clinical events, our study was underpowered to detect clinically relevant differences regarding hard clinical endpoints including death, stent thrombosis, or major bleeding. Third, we were not able to classify bleeding endpoints to other classifications such as the International Society on Thrombosis and Hemostasis (ISTH) or Bleeding Risk Estimation and the new Bleeding Academic Research Consortium (BARC) classification due to the limited information of our multicenter observational data. Fourth, we could not systematically measure detailed information on atherosclerotic burden and complexity, such as SYNTAX score. Fifth, the study population was enrolled over a wide period (2006-2017). This may have introduced some heterogeneity in the study population due to changes in clinical practice and improvements in PCI technology. Sixth, we did not include data on oral anticoagulation agents. Finally, in the current study, potent P2Y12 inhibitors, such as prasugrel and ticagrelor, were less frequently used. In a recent study, novel P2Y12 inhibitor monotherapy was shown to reduce major bleeding events without increasing ischemic events in patients with complex PCI.21 This concept should be further tested in patients with complex PCI including LMCA disease.