Author + information
- Received September 12, 2018
- Revision received November 12, 2018
- Accepted December 3, 2018
- Published online May 6, 2019.
- Yasushi Ueki, MDa,
- Alexios Karagiannis, PhDb,
- Christian Zanchin, MDa,
- Thomas Zanchin, MDa,
- Stefan Stortecky, MDa,
- Konstantinos C. Koskinas, MD, MSca,
- George C.M. Siontis, MD, PhDa,
- Fabien Praz, MDa,
- Tatsuhiko Otsuka, MDa,
- Lukas Hunziker, MDa,
- Dik Heg, PhDb,
- Aris Moschovitis, MDa,
- Christian Seiler, MDa,
- Michael Billinger, MDa,
- Thomas Pilgrim, MDa,
- Marco Valgimigli, MD, PhDa,
- Stephan Windecker, MDa and
- Lorenz Räber, MD, PhDa,∗ ()
- aDepartment of Cardiology, Bern University Hospital, Bern, Switzerland
- bInstitute of Social and Preventive Medicine and Clinical Trials Unit, University of Bern, Bern, Switzerland
- ↵∗Address for correspondence:
Prof. Lorenz Räber, Department of Cardiology, Bern University Hospital, Freiburgstrasse 18, 3010 Bern, Switzerland.
Objectives This study sought to validate European Society of Cardiology guideline-endorsed high-risk features of stent-related recurrent ischemic events for the prediction of ischemic and bleeding outcomes including a stratification according to the PRECISE-DAPT score estimated bleeding risk.
Background The 2017 European Society of Cardiology–focused update on dual-antiplatelet therapy endorsed high-risk features of stent-related recurrent ischemic events. Because patients with high ischemic risk also have an increased bleeding risk, appropriate risk stratification for ischemic and bleeding events is crucial.
Methods Between January 2009 and December 2015, a total of 10,236 consecutive patients undergoing clinically indicated percutaneous coronary intervention were prospectively included in the Bern PCI Registry. Guideline-endorsed high-risk features were retrospectively assessed. The primary ischemic endpoint was device-oriented composite endpoint (DOCE) (cardiac death, target-vessel myocardial infarction, and target lesion revascularization) at 1 year, and the primary bleeding endpoint was Bleeding Academic Research Consortium (BARC) 3–5 at 1 year.
Results A total of 5,323 (52.0%) patients had at least 1 high-risk feature. Among patients with high-risk features, DOCE (12.3% vs. 5.5%; p < 0.001) and BARC 3–5 bleeding (4.9% vs. 2.2%; p < 0.001) occurred more frequently compared with those without. There was a graded risk increase for DOCE (0: 5.5%; 1 to 2: 11.3%; and ≥3: 16.7%; p < 0.001) and BARC 3–5 bleeding (0: 2.2%; 1 to 2: 4.5%; and ≥3: 6.6%; p < 0.001) as the number of high-risk features increased. High-PRECISE-DAPT score (≥25) was associated with an increased risk of DOCE and BARC 3–5 bleeding, irrespective of number of high-risk features.
Conclusions The European Society of Cardiology guideline-endorsed high-risk features were associated with increased ischemic and bleeding risks following percutaneous coronary intervention in routine clinical practice. (CARDIOBASE Bern PCI Registry; NCT02241291)
Based on recent advances in coronary intervention techniques and devices, percutaneous coronary interventions (PCI) are increasingly performed in patients with complex coronary artery disease (CAD). The complexity of PCI itself has been acknowledged as a determinant for future ischemic events, especially stent-related adverse outcomes (1,2). Various patient and lesion characteristics defining complex PCI have been identified in previous studies (1–3). Recently, the 2017 European Society of Cardiology (ESC) focused update on dual-antiplatelet therapy (DAPT) proposed high-risk features for stent-related recurrent ischemic events including prior stent thrombosis (ST) on adequate antiplatelet therapy, stenting of the last remaining patent coronary artery, diffuse multivessel disease especially in patients with diabetes, chronic kidney disease (CKD), at least 3 stents implanted, at least 3 lesions treated, bifurcation with 2 stents implanted, total stent length >60 mm, and treatment of chronic total occlusion (CTO) (4). Although these high-risk features were mainly derived from previously published studies (1,5–9), they have not been validated comprehensively in a real-world PCI population for the prediction of clinical outcomes.
Complex PCI is frequently associated with intensification or prolongation of DAPT in clinical practice, reciprocating the findings of pooled analyses of randomized controlled trials reporting a lower ischemic risk among patients undergoing complex PCI and prolonged (≥1 year) DAPT (1). Because patients with complex CAD have an increased risk for both bleeding and ischemic events (10), decisions on intensification or prolongation of DAPT are often challenging. Recently, the PRECISE-DAPT score, a 5-item scoring system to estimate the bleeding risk, was published and endorsed with a Class IIb (Level of Evidence: A) recommendation by the 2017 ESC guidelines (4,11). In the derivation and validation study of PRECISE-DAPT score, only patients without high PRECISE-DAPT score (i.e., ≥25) showed a significant ischemic benefit from long DAPT (12 to 24 months) (11). However, no studies were conducted to evaluate the utility of PRECISE-DAPT score among patients undergoing complex PCI to predict their ischemic and bleeding risk.
We therefore aimed to validate the ESC guideline-endorsed high-risk features not only for ischemic outcomes but also for bleeding endpoints with a stratification according to the baseline PRECISE-DAPT score-estimated bleeding risk in a large unselected PCI cohort.
All consecutive patients undergoing PCI at Bern University Hospital, Switzerland, between January 2009 and December 2015 were prospectively enrolled into the Bern PCI Registry (NCT02241291). For the present study, the following patient populations were excluded: undergoing balloon angioplasty alone, implantation of bioresorbable scaffolds, and not available for guideline-endorsed high-risk features at index PCI. The registry was approved by the institutional ethics committee. All patients provided written informed consent.
PCI was performed according to current guidelines (12). Heparin (initial bolus of 100 IU per kg body weight) was used for procedural anticoagulation with the aim to maintain an activated clotting time >250 ms. The periprocedural use of glycoprotein IIb/IIIa inhibitors was left to the discretion of the operator. DAPT consisting of acetylsalicylic acid and a P2Y12 inhibitor was initiated before, at the time, or immediately after the procedure. Prasugrel was routinely used in patients presenting with ST-segment elevation myocardial infarction (MI) as of September 2009, and ticagrelor was routinely used in patients with non-ST-segment elevation acute coronary syndrome as of November 2011. Most patients with stable CAD received clopidogrel. The routinely recommended DAPT duration was 12 months (13).
Patients were systematically and prospectively followed throughout 1 year to assess death, MI, stroke, revascularization, ST, bleeding complications, rehospitalization, and medical treatment. A health questionnaire was sent to all living patients with questions on rehospitalization and adverse events, followed by telephone contact in case of missing response. General practitioners and referring cardiologists were contacted as necessary for additional information. For patients treated for adverse events at other medical institutions, external medical records, discharge letters, and coronary angiography documentation were systematically collected and reviewed.
Clinical endpoints and definitions
A clinical event committee consisting of 2 cardiologists (and a third referee in case of disagreement) adjudicated all events against original source documents. The primary ischemic (stent-related) endpoint was the device-oriented composite endpoint (DOCE), defined as a composite of cardiac death, target-vessel (TV) MI, and target lesion revascularization (TLR). The primary bleeding endpoint was bleeding defined as Bleeding Academic Research Consortium (BARC) 3–5 (14). Secondary endpoints were all-cause death, cardiac death, any MI, TV-MI, any repeat revascularization, TLR, target vessel revascularization, non–TV revascularization, definite/probable ST, stroke, any bleeding, and BARC 2–5. Cardiac death was defined as any death caused by an immediate cardiac cause, procedure-related mortality, and death of unknown cause. MI was defined according to the modified historical definition (15). ST was classified according to the Academic Research Consortium criteria (16). Stroke was defined as rapid development of clinical signs of focal or global disturbance of cerebral function lasting >24 h with imaging evidence of acute, clinically relevant brain lesion.
Guideline-endorsed high-risk features in the present study and definitions were as follows: diffuse (defined as lesion length ≥20 mm) multivessel disease in patients with diabetes, CKD (defined as estimated glomerular filtration rate <60 ml/min/1.73 m2), at least 3 stents implanted, at least 3 lesions treated, bifurcation with 2 stents implanted, total stent length >60 mm (per lesion), and treatment of CTO.
Continuous variables were summarized as mean ± SD or median (interquartile range) and compared with Student’s t-test or the Mann-Whitney U test. Binary and categorical variables were calculated as frequencies (percentages) and were compared with the chi-square test or Fisher exact test. Kaplan-Meier cumulative event curves were constructed for time-to-event variables and compared using the log-rank test. Cox regression analysis was performed to build a separate multivariate model to calculate individual predicted probabilities for DOCE and BARC 3–5, and to test the prognostic significance of high-risk features and each component for study endpoints. For the former, a backward selection method was used to select variables with a p value <0.05 to stay in the final model. Correlation between both risks derived from these models was assessed by the Spearman correlation coefficient. For the latter, high-risk features as a categorical variable (0, 1 to 2, and ≥3) and each component were adjusted by clinically important variables reported by previous studies. For ischemic endpoints, age, female sex, current smoker, hypertension, peripheral artery disease, previous MI, previous revascularization (PCI and/or coronary artery bypass graft), left ventricular ejection fraction, stent type (bare-metal stent, first-generation drug-eluting stent, new-generation drug-eluting stent), mean stent diameter, and potent P2Y12 use at discharge; and for BARC 3–5 bleeding, age, female sex, body mass index, prior bleeding, history of malignancy, anemia, potent P2Y12 use at discharge, and DAPT and oral anticoagulant use at discharge were entered into a multivariate model (9,11,17,18). The p values were 2-tailed, and <0.05 was considered statistically significant in all analyses. Statistical analyses were performed with R version 3.4.1 (R Foundation for Statistical Computing, Vienna, Austria).
Of 12,100 patients enrolled into the Bern PCI Registry between January 2009 and December 2015, a total of 10,236 patients were analyzed for the present study with complete follow-up available for 9,508 (92.8%) patients at 1 year. Patients were excluded in cases of balloon angioplasty without stent placement (n = 472) or implantation of bioresorbable scaffolds (n = 42) and if ESC guideline-endorsed features were not available (n = 1,350).
Clinical and procedural characteristics and medication status at discharge are summarized in Tables 1, 2, and 3⇓⇓⇓. Patients with at least 1 high-risk feature (n = 5,323) were older and more commonly female; had more coronary risk factors, comorbidities, and stable CAD as an indication for PCI; and had higher PRECISE-DAPT and PARIS scores compared with those without. Among patients with high-risk features, PCI was more frequently performed on the left main, saphenous vein bypass grafts, and thrombus-containing lesions. New-generation drug-eluting stents were used in 92.7% of patients, with a higher frequency in the complex group. The use of potent P2Y12 was less frequent in patients with high-risk features. The overlap of guideline-endorsed high-risk features is summarized in Online Table S1. As expected, ≥3 stents implanted and ≥3 lesions treated frequently overlapped with other features.
Clinical outcomes at 1 year are summarized in Table 4. Median (interquartile range) days to DOCE and BARC 3–5 bleeding were 21 (1 to 178) and 29 (2 to 150) days, respectively. Patients with high-risk features had a higher risk of DOCE (12.3% vs. 5.5%; hazard ratio [HR]: 2.35; 95% confidence interval [CI]: 2.04 to 2.71; p < 0.001) and BARC 3–5 bleeding (4.9% vs. 2.2%; HR: 2.34; 95% CI: 1.87 to 2.93; p < 0.001) compared with those without. Findings were consistent for the early (0 to 30 days) and late (30 to 365 days) time periods and among patients with stable CAD and acute coronary syndrome (Online Tables S2 and S3). There was a gradual risk increase for DOCE (0: 5.5%; 1 to 2: 11.3%; and ≥3: 16.7%; p < 0.001) and BARC 3–5 bleeding (0: 2.2%; 1 to 2: 4.5%; and ≥3: 6.6%; p < 0.001) as a function of the number of high-risk features (Figure 1).
The incidence of DOCE, BARC 3–5, and all individual endpoints are presented in Figure 2 as a function of high-risk features and bleeding risk. Patients with high (≥25) PRECISE-DAPT score showed a consistently higher incidence of DOCE, cardiac death, definite or probable ST, and BARC 3–5 compared with those with low (<25) PRECISE-DAPT score across all high-risk feature categories. The incidence of TLR and TV-MI also increased as a function of the number of high-risk features but without notable difference between high and low PRECISE-DAPT score.
Multivariable models to calculate individual predicted probabilities for DOCE and BARC 3–5 had an acceptable discriminative power with Harrel’s C index 0.66 and 0.69, respectively. Figure 3 shows the relationship between predicted probabilities of DOCE and BARC 3–5 within individual patients. There were weak correlations between probabilities of DOCE and BARC 3–5 bleeding in low-PRECISE-DAPT score group (high-risk features 0: 0.08; 1 to 2: 0.20; ≥3: 0.12), whereas no correlations were observed in high-PRECISE-DAPT score group (high-risk features 0: −0.16; 1 to 2: 0.05; ≥3: −0.03). Figure 4 shows the distribution of the absolute difference of predicted probabilities of DOCE and BARC 3–5. The percentage of patients with higher probabilities of bleeding than that of DOCE was greater in the high-PRECISE-DAPT score group than the low-PRECISE-DAPT score group (overall: 11.3% vs. 1.5%). The distribution shifted wider in high-PRECISE-DAPT score group, and also as the number of high-risk features increased.
Cox regression analysis
Table 5 shows unadjusted and adjusted risks for each endpoint according to the number of high-risk features and PRECISE-DAPT score. The number of high-risk features was an independent predictor for DOCE, cardiac death, TV-MI, TLR, definite or probable ST, and BARC 3–5. These findings were consistent in each high- and low-PRECISE-DAPT score group. Figure 5 and Online Figure S1 summarize an unadjusted and adjusted risk of each component of high-risk features. Independent predictors for DOCE were CKD (adjusted HR: 1.46; 95% CI: 1.20 to 1.77; p < 0.001), ≥3 lesions treated (adjusted HR: 1.31; 95% CI: 1.06 to 1.63; p = 0.013), and ≥3 stents implanted (adjusted HR: 1.55; 95% CI: 1.27 to 1.89; p = 0.003). CKD was also an independent predictor for BARC 3–5 (adjusted HR: 1.60; 95% CI: 1.19 to 2.15; p = 0.002). CTO was not significantly associated with any endpoints.
To the best of our knowledge, the present study is the first to validate the ESC guideline-endorsed high risk of stent-driven recurrent ischemic events in a large unselected PCI cohort. There are 3 main findings of the present study. First, patients with guideline-endorsed high-risk features had a higher risk of ischemic (DOCE) and bleeding (BARC 3–5) events. The risks increased proportionally with the number of high-risk features. These findings were consistently observed during the early (<30 days) and late (30 to 365 days) periods as well as in the different clinical presentation (stable CAD and acute coronary syndrome). Second, patients fulfilling high-risk bleeding criteria (i.e., PRECISE-DAPT score ≥25) compared with those without carried a higher risk of DOCE and bleeding BARC 3–5, irrespective of the number of high-risk features present. Third, the predictive value of the individual guideline-endorsed complexity criteria varies for ischemia and bleeding events. Although CKD was an independent predictor of both DOCE and BARC 3–5, ≥3 stents implanted, and ≥3 lesions treated were independent predictors for DOCE but not for BARC 3–5.
Similar to the findings from recent studies using randomized trial data, this large PCI cohort reflecting a real-world setting showed that procedural complexity as assessed by guideline-endorsed high-risk features was consistently associated with ischemic events (DOCE), and the risk increased with the number of complexity criteria present (1,2). This was not only the case for the combined ischemia endpoint (DOCE) but also for all individual endpoints including cardiac death, TV-MI, TLR, and definite and probable ST.
Complex PCI criteria were also associated with a higher bleeding risk, and this finding was consistent for the early (0 to 30 days; i.e., periprocedural bleeding) and late time periods (30 to 365 days) following PCI. The results need to be interpreted against the background of a routinely recommended 12-month DAPT duration in most PCI patients and in the absence of a routine score-based assessment of the bleeding risk at baseline.
The relationship between bleeding and complex PCI features was expected as demonstrated by the proportional increase in PRECISE-DAPT score (no high-risk features: 15.4 ± 9.6; 1 to 2: 24.9 ± 14.4; ≥3: 28.3 ± 14.5; p < 0.001). Matteau et al. (10) reported that there was a strong correlation between predicted probabilities of ischemic (composite of cardiovascular death, MI, and definite or probable ST) and bleeding events, although most patients undergoing PCI generally had a greater risk of ischemic than bleeding events. In contrast to our findings, recent large studies reported that PCI complexity may not be associated with an increased risk of bleeding (1,2). This may be explained at least in part by the exclusion of high-risk patients included into randomized trials and the difference of definition of complex PCI.
The selection of adequate intensity and duration of DAPT is a challenging clinical decision, especially in patients with both high ischemic and bleeding risks. Interestingly, patients with high compared with those with low bleeding risk consistently showed a higher risk of DOCE, BARC 3–5, and cardiac death independent of the number of high-risk features present. For outcomes more specifically related to stents including TV-MI and TLR, risks increased in parallel with the complexity, but they were not consistently higher for low versus high bleeding risk categories. Because Genereux et al. (19) reported that post-discharge bleeding was strongly associated with mortality (bleeding with transfusion: HR: 4.71; 95% CI: 1.76 to 8.03) (bleeding without transfusion: HR: 5.27; 95% CI: 3.32 to 8.35), to a greater degree than post-discharge MI (HR: 1.92; 95% CI: 1.18 to 3.12), to select “high-risk bleeding patients” is essential when determining the intensity and duration of DAPT. Costa et al. (11) reported that long-term (12 to 24 months) DAPT was associated with no ischemic benefit but an increased risk of bleeding among patients with high PRECISE-DAPT score group, and a similar finding was observed among patients with low DAPT score (<2) in the DAPT study (17). There is a need for further prospective studies to ascertain the appropriate duration and intensity of DAPT among patients with high ischemic and bleeding risks in view of the considerably increased risk for events approaching 20% for DOCE. Furthermore, as shown in this study, an excess variation in patients with both high PCI complexity and bleeding risk implied the difficulty of accurate prediction in this heterogenous cohort. A more individualized approach and efforts to handle modifiable factors (e.g., avoid unnecessary long or side branch stenting, minimize duration of triple therapy, and risk factor control) may be needed to improve outcomes in this “dilemmatic” population.
The present study demonstrated guideline-endorsed high-risk features were independent predictors for ischemic and bleeding events among other established risk factors, making any decisions on intensity and duration of DAPT challenging. To investigate which individual component of high-risk features allows discrimination between ischemic and bleeding events, we further performed a fully adjusted analysis of all individual features. We found that CKD was the only independent predictor for both DOCE and BARC 3–5, and ≥3 lesions treated or ≥3 stents implanted were the only independent predictors for DOCE, but not bleeding. No independent association was found for CTO or bifurcation with 2 stents implanted.
First, the single-center design of this study may limit the generalizability of our findings. Second, not every guideline-endorsed high-risk feature was captured in the present study, although “prior ST on adequate antiplatelet therapy” and “stenting of the last remaining patent coronary artery” seem to be exceedingly rare. Third, DAPT duration and intensity cannot be considered because of the nature of this observational study. Finally, we included cardiac death in the primary ischemic composite endpoint because this event may occur as the immediate consequence of such ischemic events as MI or ST without detecting those as distinct events. Conversely, bleeding precursing death should be noted in most instances.
The guideline-endorsed ischemic high-risk features per se were associated with an increased bleeding risk and an ischemic risk, making the decision of intensity and duration of DAPT challenging. Individual high-risk features, such as number of stents or lesions, which are independent predictors for ischemic events but not for bleeding, may be helpful to guide prolongation of DAPT or use of potent P2Y12, whereas others seem to be less helpful (i.e., CKD). Although further studies are required to ascertain the appropriate DAPT regimen among patients with both high ischemic and bleeding risk, physicians should pay more effort to mitigate modifiable ischemic and bleeding risk factors to improve outcomes in this dilemmatic cohort.
WHAT IS KNOWN? Patients undergoing complex PCI have an increased risk of ischemic events and receive more benefit from a longer period of DAPT compared with shorter DAPT.
WHAT IS NEW? ESC guideline-endorsed high-risk features are associated with not only ischemic but also bleeding events. High PRECISE-DAPT score (≥25) was associated with an increased risk of DOCE and BARC 3–5 bleeding, irrespective of number of high-risk features.
WHAT IS NEXT? Further studies are required to test the appropriate intensity and duration of DAPT among the dilemmatic patients carrying both increased bleeding and ischemic risk.
Prof. Pilgrim has received research grants to the institution from Biotronik, Symetis/Boston Scientific, and Edwards Lifesciences; and speaker fees from Biotronik and Boston Scientific. Prof. Valgimigli has received research grants to the institution from Terumo, Medicure, Abbott, and AstraZeneca; and honorarium fees from Abbott, Chiesi, Bayer, Daiichi Sankyo, Amgen, Terumo, AstraZeneca, Alvimedica, and Biosensors. Prof. Windecker has received research grants to the institution from Abbott, Amgen, Bayer, Boston Scientific, Biotronik, Edwards Lifesciences, Medtronic, Terumo, and St. Jude Medical. Prof. Räber has received research grants to the institution from Abbott Vascular, Sanofi, and Regeneron. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- Bleeding Academic Research Consortium
- coronary artery disease
- confidence interval
- chronic kidney disease
- chronic total occlusion
- dual-antiplatelet therapy
- device-oriented composite endpoint
- European Society of Cardiology
- hazard ratio
- myocardial infarction
- percutaneous coronary intervention
- stent thrombosis
- target lesion revascularization
- target vessel
- Received September 12, 2018.
- Revision received November 12, 2018.
- Accepted December 3, 2018.
- 2019 American College of Cardiology Foundation
- Giustino G.,
- Chieffo A.,
- Palmerini T.,
- et al.
- Yeh R.W.,
- Kereiakes D.J.,
- Steg P.G.,
- et al.
- Chandrasekhar J.,
- Baber U.,
- Sartori S.,
- et al.
- Valgimigli M.,
- Bueno H.,
- Byrne R.A.,
- et al.
- Suh J.,
- Park D.W.,
- Lee J.Y.,
- et al.
- van Werkum J.W.,
- Heestermans A.A.,
- Zomer A.C.,
- et al.
- Brilakis E.S.,
- Banerjee S.,
- Karmpaliotis D.,
- et al.
- Baber U.,
- Mehran R.,
- Giustino G.,
- et al.
- Costa F.,
- van Klaveren D.,
- James S.,
- et al.
- Windecker S.,
- Kolh P.,
- Alfonso F.,
- et al.
- Mehran R.,
- Rao S.V.,
- Bhatt D.L.,
- et al.
- Cutlip D.E.,
- Windecker S.,
- Mehran R.,
- et al.
- Genereux P.,
- Giustino G.,
- Witzenbichler B.,
- et al.