Patients scheduled for PCI were administered acetylsalicylic acid (100mg) and clopidogrel (75mg) daily for at least 3 days before the procedure. In cases of unstable angina or urgent PCI within 48h of admission, a 300mg clopidogrel loading dose was administered before the procedure. For all patients, a bolus injection of heparin (100 IU/kg) was administered after sheath insertion, and was titrated to maintain an activated clotting time of >250s throughout the procedure. Because of the observational nature of this study, the type of DES (sirolimus-, paclitaxel-, zotarolimus-, or everolimus-eluting stents) used was at the surgeon's discretion. After stent implantation, a high-pressure balloon inflation procedure was performed to achieve satisfactory angiographic results with < 25% residual stenosis by visual estimate. All patients received 100mg acetylsalicylic acid and 75mg clopidogrel per day for ≥12 months. Continuation of DAPT and other cardiac medications were prescribed at the physician's discretion.

On the basis of baseline diagnostic angiograms, each coronary lesion producing ≥ 50% diameter stenosis in vessels ≥ 1.5mm was scored separately. Thereafter, these scores were added together to provide the overall SXscore, which was calculated using the algorithm available on the SYNTAX website.10 The SXscores of patients were independently assessed by 2 experienced interventional cardiologists blinded to the clinical data, who had experience in calculating the SXscores of >100 patients before assisting in our study. The κ value for interobserver variability that was used to estimate the SXscores was 0.75, while the κ value for the intraobserver variability was 0.86. If there was disagreement regarding the SXscores, the average of the values from the 2 readers was used as the final value.11

The primary endpoint was the incidence of major bleeding within 2 years after PCI. Based on the Bleeding Academic Consortium (BARC) criteria, major bleeding was defined as BARC types 3 or 5. In brief, type 3a bleeding events included overt bleeding plus a decline in hemoglobin (3g/dL to < 5g/dL) and any transfusion with overt bleeding. Type 3b bleeding events included overt bleeding plus a decline in hemoglobin of ≥ 5g/dL, cardiac tamponade, bleeding requiring surgical intervention for control, and bleeding requiring intravenous vasoactive agents. Type 3c bleeding events included intracranial hemorrhage, subcategories confirmed by autopsy, imaging, or lumbar puncture, and intraocular bleeding compromising vision. Type 5a bleeding events included probable fatal bleeding suspected clinically but with no autopsy or imaging confirmation. Type 5b bleeding events included definite fatal bleeding, overt bleeding, and autopsy or imaging confirmation.12 Moreover, bleeding events were evaluated by the Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes with Early Implementation of the American College of Cardiology/American Heart Association Guidelines (CRUSADE) bleeding score.13,14 The CRUSADE score was calculated based on baseline patient characteristics (female sex, diabetes, prior vascular disease, heart rate, systolic blood pressure, congestive heart failure, hematocrit value, and estimated creatinine clearance). The secondary endpoint was definite stent thrombosis, which was diagnosed as an acute coronary syndrome with angiographic documentation of either target vessel occlusion or thrombus, within or adjacent to the previously stented segment, applying the BARC definition.15 Cardiac death was considered as any fatal event not attributed to a noncardiac cause. Patients were followed up for ≥ 2 years after implantation.

Data are presented as mean (standard deviation) or as percentages. Differences between groups were assessed using unpaired 2-tailed Student t tests for continuous variables and chi-square tests for categorical variables. Survival analysis was performed using the Kaplan–Meier method and differences were assessed using the log rank test. Univariate and multivariate Cox proportional hazards models were used to assess independent correlates of major bleeding and stent thrombosis. All variables in Table 1 were entered into the models. Continuous variables were transformed into binary data with “1” representing the presence of an assumed risk factor and “0” as otherwise. For example, age was dichotomized as elderly or not elderly (> 75 years vs ≤ 75 years) and ejection fraction was dichotomized as low or normal (< 40% vs ≥ 40%). Moreover, to avoid arbitrariness, we used the median value of each continuous factor (body mass index, stent length, lesion length, etc.) as the cutoff point for this division. With regard to the SXscores, CRUSADE scores, heart rate, and glomerular filtration rate, the unit of measurement was designated as 10. Eventually, adjustment was performed to calculate hazard ratios (HRs) and 95% confidence intervals (95%CIs) with only clinically relevant variables that had P values of <.05 in the univariate analysis. Area under the receiver operating characteristic curve analysis was performed to determine the ability of the SXscore and CRUSADE score to distinguish between patients with and without major bleeding. A C statistic value > 0.7 was considered to have acceptable discriminatory capacity. All analyses were performed using the IBM-SPSS statistics version 19 and R package. A P value of < .05 was considered to indicate statistical significance.

Table 1 shows the baseline clinical characteristics in the bleeding (n=47) and no bleeding (n=675) groups. The patients in the bleeding group were older, more often women, and were more likely to have a Killip class ≥ 2, multiple coronary vessel disease, and chronic kidney disease than patients in the no bleeding group. Body mass index, hematocrit value, and left ventricular ejection fraction in patients in the bleeding group were significantly lower than those in the no bleeding group. Furthermore, both the SXscores and CRUSADE scores were significantly higher in the bleeding group than in the no bleeding group. With the exception of transradial intervention, there was no significant difference in procedural factors between the groups.

Complete clinical follow-up data were available in 93% patients at 1 year and in 81% at 2 years. A total of 613 patients (84.9%) showed prolonged DAPT for up to 2 years following PCI. As shown in Table 2, patients were divided into 3 groups based on the original SYNTAX trial: low (≤ 22; n=484), intermediate (23-32; n=128), and high (≥33; n=110). Major bleeding according to BARC criteria was reported in 47 patients (6.5%; type 3a, n=18; type 3b, n=23; type 3c, n=2; type 5a, n=4; type 5b=0). With regard to the timing of bleeding complications, 23 patients (49%) experienced major bleeding within 30 days of the procedure, and the remaining 24 patients (51%) had major bleeding between 30 days and 2 years (Table 3). With regard to DAPT at the time of major bleeding, all 47 patients with major bleeding were receiving DAPT. On the other hand, stent thrombosis occurred in 12 patients (1.7%) within 2 years; of these, 3 were involved in antiplatelet therapy cessation. The mortality rates were 0.8% at 30 days and 4.6% at 2 years after PCI, respectively (Table 3). Figure 1 shows the survival curves for the patients who were free from major bleeding among the 3 SXscore tertiles. The 2-year incidence of major bleeding was 20.9% (n=23) among patients with high SXscores, 7.8% (n=10) among patients with intermediate SXscores, and 2.9% (n=14) among patients with low SXscores (Table 3). Similarly, stent thrombosis occurred more frequently among patients with high SXscores (6.4%; n=7) than those with intermediate (3.1%; n=4) or low (0.2%; n=1) SXscores (P < .0001). The area under the curve analysis also showed that the SXscore (C statistic = 0.751; 95%CI, 0.674-0.828; P < .0001) and the CRUSADE score (C statistic = 0.835; 95%CI, 0.774-0.897; P < .0001) had an adequate discriminatory capacity to distinguish between patients with and without major bleeding.

Figure 1.

Kaplan-Meier curves depicting the survival from major bleeding according to the SYNergy between percutaneous coronary intervention with TAXus and cardiac surgery scores. Patients were divided into the following groups according to the original SYNergy between percutaneous coronary intervention with TAXus and cardiac surgery trial: low (≤ 22), intermediate (23-32), and high (≥ 33).

(0.11MB).

The Cox proportional hazards model was used to identify independent predictors of 2-year major bleeding and stent thrombosis. Baseline variables listed in Table 1 that differed between groups with a P value of < .05 were entered into the model. As shown in Table 4, model 1 with clinically relevant variables and the SXscore identified the SXscore as an independent correlate of 2-year major bleeding (HR = 2.19; 95%CI, 1.54-3.11; P < .0001 for the SXscore tertile). Also, model 2, which was the addition of the CRUSADE score to model 1, demonstrated that both the SXscore and the CRUSADE score were independent predictors of major bleeding.

The predictive value of the SXscore and the CRUSADE score with regard to major bleeding was assessed by calculating the adjusted area under the receiver operating characteristic curve in multivariate models 1 and 2. The adjusted areas under the curve of models 1 and 2 demonstrated adequate discriminatory capacities to distinguish between patients with and without major bleeding (model 1: C = 0.812; model 2: C = 0.890) (Figure 2). However, model 2, which added the CRUSADE score, significantly improved the C statistic for predicting major bleeding compared with model 1 (z test=−257, P=.01).

Figure 2.

Adjusted receiver operating characteristic curves for predicting 2-year major bleeding by model 1 and model 2. 95%CI, 95% confidence interval. Model 1: clinically relevant variables plus SYNergy between percutaneous coronary intervention with TAXus and cardiac surgery score. Model 2: model 1 plus Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of the American College of Cardiology/American Heart Association Guidelines score.

(0.13MB).

Independent predictors of 2-year stent thrombosis were the SXscore (tertile; HR=4.39; 95%CI, 2.02-9.50; P=.0002) and severe calcified lesion (HR=3.97; 95%CI, 1.26-12.52, P=.02). According to the presence of major bleeding, the event-free survival curves in Figure 3A and 3B indicate that patients with major bleeding had a significantly higher incidence of cardiac death (14.9% vs 1.0%; P<.0001) and of all-cause death (23.4% vs 2.5%) at 2 years.

Figure 3.

A: Kaplan-Meier survival curves for cardiac death in patients with and without major bleeding; B: Kaplan-Meier survival curves for all-cause death in patients with and without major bleeding.

(0.2MB).

Regarding the risk of thrombotic events such as stent thrombosis and myocardial infarction, registries and observational studies have demonstrated that DAPT is not required beyond 6 months.17,18 The PRODIGY trial also demonstrated that the extended use of DAPT was associated with a higher rate of bleeding events and blood transfusions, without the protective effects of reducing the risk of thrombotic events.19 We observed that the incidence of stent thrombosis was very low among patients with low and intermediate SXscores. Therefore, such low-risk patients should be switched from DAPT to antiplatelet monotherapy, including low-dose acetylsalicylic acid, as soon as possible to avoid bleeding complications. On the other hand, patients with a high SXscore in the present study had significantly higher incidences of both major bleeding and stent thrombosis. In other words, in PCI in patients with a high SXscore, we may still face the unresolvable dilemma of whether to continue DAPT beyond 12 months after PCI.20 Therefore, the SYNTAX trial results along with the present findings suggest that patients with high scores and adequate distal vessel beds may be more appropriate candidates for coronary artery bypass grafts, with the exception of those with a high surgical risk and feasible PCI. Currently, many randomized clinical trials examining the optimal duration of DAPT following DES implantations are ongoing.21–23 To determine whether the duration of DAPT needs to continue beyond 6 or 12 months following PCI, it will be necessary to wait for the results of these trials.

The present study has several limitations. First, it was a single center study with a small sample size. However, to the best of our knowledge, no other studies have evaluated major bleeding following DES implantation using the BARC criteria and SXscores. The results may have been influenced by confounding factors. For example, decisions regarding DAPT duration may have differed among the attending physicians. Notably, in our study, many patients (84.9%) were treated with DAPT for up to 2 years after implantation; this number seems to be high. The reason might have been related to a concern about the development of very late stent thrombosis with first generation DES use. Furthermore, we did not use a glycoprotein IIb/IIIa inhibitor, prasugrel, and bivalirudin. Although 8% of patients in our study were on triple anticoagulation therapy, including warfarin, there was no difference in bleeding complications. Perhaps the expected effect of warfarin was decreased when the patients received DAPT.