Impact of Baseline Bleeding Risk on Efficacy and Safety of Ticagrelor versus Clopidogrel in Chinese Patients with Acute Coronary Syndrome Undergoing Percutaneous Coronary Intervention.

BACKGROUND: There was still conflict on the antithrombotic advantage of ticagrelor versus clopidogrel among East Asian population with acute coronary syndrome (ACS). We considered that the baseline bleeding risk might be an undetected key factor that significantly affected the efficacy of ticagrelor.
METHODS: A total of 20,816 serial patients who underwent percutaneous coronary intervention (PCI) from October 2011 to August 2014 in the General Hospital of Shenyang Military Region were enrolled in the present study. Patients receiving ticagrelor or clopidogrel were further subdivided according to basic bleeding risk. The primary outcome was net adverse clinical events (NACEs) defined as major adverse cardiac or cerebral events (MACCE, including all-cause death, myocardial infarction, ischemia-driven target vessel revascularization, or stroke) and any bleeding during 1-year follow-up. Comparison between ticagrelor and clopidogrel was adjusted by propensity score matching (PSM).
RESULTS: Among the 20,816 eligible PCI patients who were included in this study, there were 1578 and 779 patients in the clopidogrel and ticagrelor groups, respectively, after PSM, their clinical parameters were well matched. Patients receiving ticagrelor showed comparable NACE risk compared with those treated by clopidogrel (5.3% vs. 5.1%, P = 0.842). Furthermore, ticagrelor might reduce the MACCE risk in patients with low bleeding risk but increase MACCE in patients with moderate-to-high bleeding potential (ticagrelor vs. clopidogrel, low bleeding risk: 2.5% vs. 4.9%, P = 0.022; moderate-to-high bleeding risk: 4.8% vs. 3.0%, P = 0.225; interaction P = 0.021), with vast differences in all bleeding (low bleeding risk: 1.5% vs. 0.8%, P = 0.210; moderate-to-high bleeding risk: 4.8% vs. 3.0%, P = 0.002; interaction P = 0.296).
CONCLUSION: Among real-world Chinese patients with ACS treated by PCI, ticagrelor only showed superior efficacy in patients with low bleeding risk but lost its advantage in patients with moderate-to-high bleeding potential.

INTRODUCTION

Ticagrelor was strongly recommended for patients with acute coronary syndrome (ACS),[12] mainly according to the results of the PLATO trial.[34] Consistent results were also reported in real-world studies such as the PRACTICAL study from the SWEDEHEART registry.[5] However, there was also conflicting evidence on the efficacy of ticagrelor among the East Asian population with ACS. The PHILO study[6] recruited 817 East Asian patients with ACS and found no improvement; however, a numerical increase in ischemic risk (using a definition that is consistent with PLATO) in patients receiving ticagrelor compared with clopidogrel. The most recent meta-analysis,[7] which focused on the population of East Asia, also demonstrated no advantage of ticagrelor on antithrombotic efficiency. Whereas, two recent real-world studies reported that ticagrelor could provide a marginally or significantly favorable antithrombotic efficacy in patients with ACS[8] or acute myocardial infarction (AMI),[9] both with similar major bleeding rates compared with clopidogrel. These contradictory results have caused considerable confusion regarding the efficacy of ticagrelor, providing the impetus to formulate better antiplatelet strategies.

Bleeding is an independent risk factor for ischemic events.[1011] There is concern that compared with Western patients, East Asian patients suffer more bleeding with the “standard” antiplatelet therapy, which was also called the “East Asian paradox.”[12] In the PLATO trial, the increases in non-CABG-related major bleeding was approximately 170% versus 19% among East Asian patients versus the whole population.[313] Another real-world study from China also showed that bleeding could be lowered by approximately 40% after switching from ticagrelor to clopidogrel.[14] Therefore, the potential bleeding risk of East Asian patients might be closely related to the efficacy of ticagrelor, and the risk greatly changed when ticagrelor was applied from Phase III randomized clinical trials (RCTs) to the real-world clinical practice. However, the impact of baseline bleeding risk on the efficacy of ticagrelor versus clopidogrel has never been reported in the East Asian population.

The present study was a single-center, large-sample, pragmatic clinical trial, which divided the ACS patients by low- and moderate-to-high baseline bleeding risk according to the CRUSADE score.[15] We aim to clarify the influence of baseline bleeding risk on the efficacy advantage of ticagrelor and to provide more evidence for the optimal use of ticagrelor in the East Asian population.

METHODS

Ethical approval

The present study was approved by the Ethics Committee of the General Hospital of Shenyang Military Region. Since the present study was a post hoc analysis of exist data, there was no informed consent.

Study population and design

Utilizing information in the preregistered database, a total of 20,816 serial ACS patients from October 2011 to August 2014 in the General Hospital of Shenyang Military Region were enrolled in the present study. There were no special exclusion criteria. Patients were divided into two groups based on drug administration at discharge. The baseline bleeding risk was judged according to the CRUSADE score and was categorized into low bleeding risk (which contain CRUSADE risk levels from extremely low-to-low) and moderate-to-high bleeding risk (which contain CRUSADE risk levels from moderate to extremely high). Comparisons between ticagrelor and clopidogrel, as well as the subgroup analysis of different bleeding risk, were both adjusted by the propensity score matching (PSM).

Outcomes and definitions

The primary outcome was net adverse clinical events (NACEs) defined as major adverse cardiac or cerebral events (MACCE, a composite of all-cause death, myocardial infarction, ischemia-driven target vessel revascularization, or stroke) and any bleeding as defined by the Bleeding Academic Research Consortium (BARC) definition (Grades 1–5)[16] during 1-year follow up. Major secondary end-points were MACCE and any bleeding at 1 year. CRUSADE levels were defined as extremely low, low, moderate, high, and extremely high, corresponding to the CRUSADE score ranges of 1–20, 21–30, 31–40, 41–50, and 51–91, respectively.[15]

Statistical analysis

Categorical variables were summarized by frequencies and percentages and compared using the Chi-square or Fisher's exact test. Continuous data are summarized by the means and standard deviations, and the data were compared using an independent t-test. PSM was performed to minimize the differences in the baseline characteristics between the treatments of ticagrelor and clopidogrel. Unbalanced baseline factors were adjusted using multivariate logistic regression analysis, conditional on covariates such as sex, smoking status, ACS spectrum, previous ischemic stroke, estimated glomerular filtration rate (eGFR) <60 ml/min, use of glycoprotein IIb/IIIa inhibitor (GPI), and the presence of multivessel disease. The PSM was used to match patients administered ticagrelor and those given clopidogrel in a 1:2 ratio. All statistical analyses were performed with SPSS version 20.0 (IBM Corporation, Armonk, New York, USA), and P < 0.05 was considered statistically significant.

RESULTS

Clinical characteristics of the study population

A total of 20,816 patients with ACS undergoing percutaneous coronary intervention (PCI) were included in this study, with 779 patients in the group treated with ticagrelor. The clinical characteristics are summarized in Table 1. Compared with clopidogrel, ticagrelor was more likely to be administered in patients with poorer medical history, such as smoking (57.3% vs. 47.8%, P < 0.001) and prior stroke (8.9% vs. 6.9%, P = 0.034), and ACS categories of higher risk, such as ST elevation myocardial infarction (STEMI) (30.3% vs. 18.0%, P < 0.001) or non-ST elevation myocardial infarction (NSTEMI) (12.2% vs. 8.8%, P = 0.001). Patients assigned to ticagrelor treatment seem to have better profiles of renal function (eGFR: 124.38 ± 36.04 vs. 102.08 ± 34.23 ml/min, P < 0.001; eGFR <60 ml/min: 2.4% vs. 6.0%, P < 0.001).

Table 1

Baseline characteristics of the study population before and after PSM

Characteristics	Before PSM				After PSM
	Ticagrelor (n = 779)	Clopidogrel (n = 20,037)	Statistical values	P	Ticagrelor (n = 779)	Clopidogrel (n = 1558)	Statistical values	P
Female, n (%)	225 (28.9)	5961 (29.7)	0.270*	0.604	225 (28.9)	441 (28.3)	0.085	0.771
Age (years), mean ± SD	60.54 ± 10.53	61.30 ± 10.64	1.951†	0.051	60.54 ± 10.53	60.97 ± 10.54	0.918	0.359
Medical history, n (%)
Current smoker	446 (57.3)	9578 (47.8)	26.829*	<0.001	446 (57.3)	900 (57.8)	0.056	0.813
Hypertension	451 (57.9)	11,398 (56.9)	0.312*	0.576	451 (57.9)	849 (54.5)	2.435	0.119
Diabetes	192 (24.6)	4734 (23.6)	0.432*	0.463	192 (24.6)	371 (23.8)	0.198	0.657
Prior, n (%)
Myocardial infarction	121 (15.5)	3535 (17.6)	2.305*	0.129	121 (15.5)	269 (17.3)	1.122	0.290
Stroke	69 (8.9)	1380 (6.9)	4.494*	0.034	69 (8.9)	130 (8.3)	0.176	0.675
ACS categories, n (%)
Unstable angina	448 (57.5)	14,657 (73.1)	92.132*	<0.001	448 (57.5)	896 (57.5)	<0.001	1.000
NSTEMI	95 (12.2)	1767 (8.8)	10.496*	0.001	95 (12.2)	194 (12.5)	0.032	0.859
STEMI	236 (30.3)	3613 (18.0)	74.826*	<0.001	236 (30.3)	468 (30.0)	0.016	0.899
eGFR, mean ± SD	124.38 ± 36.04	102.08 ± 34.23	−17.512†	<0.001	124.38 ± 36.04	104.78 ± 32.90	−12.552	<0.001
<60 ml/min, n (%)	19 (2.4)	1207 (6.0)	17.385*	<0.001	19 (2.4)	38 (2.4)	<0.001	1.000
Platelet, mean ± SD	213.03 ± 54.00	205.75 ± 51.58	−3.799†	<0.001	213.03 ± 54.00	204.84 ± 50.69	−3.475	0.001
<100 × 109/L, n (%)	5 (0.6)	118 (0.6)	–	0.809	5 (0.6)	8 (0.5)	–	0.770
Hemoglobin, mean ± SD	139.90 ± 16.56	138.30 ± 15.79	−2.735†	0.006	139.90 ± 16.56	139.11 ± 15.05	−1.109	0.267
Anemia, n (%)	46 (5.9)	1175 (5.9)	0.002*	0.962	46 (5.9)	85 (5.5)	0.198	0.656
Medications before PCI, n (%)
Aspirin	776 (99.6)	19559 (97.6)	13.294*	<0.001	776 (99.6)	1534 (98.5)	6.070	0.014
Loading dose regimen	698 (89.6)	18263 (91.1)	2.203*	0.138	698 (89.6)	1442 (92.6)	5.865	0.015
GPI use, n (%)	273 (35.0)	4794 (23.9)	50.340*	<0.001	273 (35.0)	543 (34.9)	0.008	0.927
Radial access, n (%)	744 (95.5)	18,005 (89.5)	26.748*	<0.001	744 (95.5)	1492 (95.8)	0.083	0.774
Multivessel disease, n (%)	476 (61.1)	10,829 (54.0)	15.057*	<0.001	476 (61.1)	956 (61.4)	0.014	0.904
RVD (mm), mean ± SD	3.00 ± 0.37	2.99 ± 0.36	0.651†	0.515	3.00 ± 0.37	2.95 ± 0.45	0.612	0.618
Other medication, n (%)
Statins	749 (96.1)	17,008 (84.9)	75.920*	<0.001	749 (96.1)	1502 (96.4)	0.756	0.756
β-blocker	588 (75.5)	15,916 (79.4)	7.130*	0.008	588 (75.5)	1174 (75.4)	0.005	0.946

*χ2 values; -: Fisher exact test; †t values. Anemia was defined as hemoglobin <120 g/L for men or <11 g/L for women. PSM: Propensity score matching; ACS: Acute coronary syndrome; NSTEMI: Non-ST elevation myocardial infarction; STEMI: ST elevation myocardial infarction; SD: Standard deviation; eGFR: Estimated glomerular filtration rate; PCI: Percutaneous coronary inervention; GPI: Glycoprotein IIb/IIIa inhibitor; RVD: Reference vessel diameter.

We found that GPI was more frequently used in patients with ticagrelor (35.0% vs. 23.9%, P < 0.001), which was complicated with a higher incidence of multivessel disease (61.1% vs. 54.0%, P < 0.001). Nonetheless, the number of stents, along with the reference vessel diameter, were both comparable between the two groups. Radial access was more commonly applied in ticagrelor-treated patients (95.5% vs. 85.5%, P < 0.001). For other medications in the ticagrelor group, more patients received statins (96.1% vs. 84.9%, P < 0.001), while a fewer were treated with β-blockers (75.5% vs. 79.4%, P = 0.008).

There were 1578 and 779 patients in the clopidogrel and ticagrelor groups after PSM, respectively, and the clinical characteristics were well matched, with the exception of a higher level of eGFR in the ticagrelor group. However, the proportion of patients with an eGFR below 60 ml/min was comparable after PSM between the two groups. Although there was a significant difference in aspirin use and the loading dose, both were comparable between the two treatment groups.

One-year clinical outcomes in patients treated with ticagrelor versus clopidogrel

As shown in Table 2, among the unmatched population, the primary end-point of NACE was higher in patients in the ticagrelor group (5.3% vs. 4.1%, P = 0.116). Ticagrelor treatment did not significantly influence the risk of the composite endpoints of MACCE (ticagrelor vs. clopidogrel: 3.1% vs. 3.2%, P = 0.842) or its individual components (P > 0.05). Whereas ticagrelor significantly increased major bleeding risk defined as BARC3–5 (0.8% vs. 0.3%, P = 0.037), as well as BARC2–5 (1.0% vs. 0.4%, P = 1.022) and all bleeding (2.3% vs. 1.0%, P < 0.001).

Table 2

The 1-year clinical outcomes of the study population before and after PSM, n (%)

Outcomes	Before PSM				After PSM
	Ticagrelor (n = 779)	Clopidogrel (n = 20,037)	χ2 values	P	Ticagrelor (n = 779)	Clopidogrel (n = 1558)	χ2 values	P
NACE	41 (5.3)	825 (4.1)	2.469	0.116	41 (5.3)	79 (5.1)	0.040	0.842
MACCE	24 (3.1)	643 (3.2)	0.040	0.842	24 (3.1)	63 (4.0)	1.343	0.246
Death	7 (0.9)	242 (1.2)	0.606	0.436	7 (0.9)	19 (1.2)	0.486	0.486
MI	5 (0.6)	92 (0.5)	0.540	0.463	5 (0.6)	15 (1.0)	0.630	0.427
Stroke	1 (0.1)	30 (0.1)	–	1.000	1 (0.1)	4 (0.3)	–	0.670
TVR	13 (1.7)	325 (1.6)	0.010	0.919	13 (1.7)	29 (1.9)	0.109	0.741
All bleeding	18 (2.3)	195 (1.0)	13.244	<0.001	18 (2.3)	16 (1.0)	5.969	0.015
BARC2–5	8 (1.0)	84 (0.4)	–	0.022	8 (1.0)	6 (0.4)	–	0.084
BARC3–5	6 (0.8)	60 (0.3)	–	0.037	6 (0.8)	3 (0.2)	–	0.068

–: Fisher exact test. PSM: Propensity score matching; NACE: Net adverse clinical event; MACCE: Major adverse cardiac or cerebral events; MI: Myocardial infarction; TVR: Target vessel revascularization; BARC: Bleeding Academic Research Consortium.

Different from the results in the unmatched population, the incidence of NACE was comparable in the groups treated with ticagrelor versus clopidogrel after PSM (5.3% vs. 5.1%, P = 0.842). Reductions in the occurrence of MACCE were found in matched patients in the ticagrelor group versus the clopidogrel group (3.1% vs. 4.0%, P = 0.246), mainly due to different degrees of decline in each end-point of ischemia. The safety profile of ticagrelor in comparison to clopidogrel remained after PSM not only for all bleeding (2.3% vs. 1.0%, P = 0.015) but also for the higher trend of bleeding that requires medical therapy (BARC2–5, 1.0% vs. 0.4%, P = 0.084) and for major bleeding (BARC3–5, 0.8% vs. 0.2%, P = 0.068).

Impact of CRUSADE bleeding risk on clinical outcomes of ticagrelor versus clopidogrel adjusted by propensity score matching

Surprising outcomes were found on the efficacy of ticagrelor versus clopidogrel between patients with low- and moderate-to-high baseline bleeding risk. After PSM, there were 1578 patients treated with clopidogrel matched to the 779 patients treated with ticagrelor (2:1). We first noticed that ticagrelor was less frequently administered in patients with low bleeding potential (189/779, 24.3%) whereas the use of clopidogrel accounted for nearly half (706/1578, 44.7%). As demonstrated in Table 3, although with no significance, ticagrelor was found to reduce the number of NACE in patients with a CRUSADE score below 30 (4.1% vs. 5.8%, P = 0.152); however, ticagrelor markedly increased the incidence of NACE in those patients with a CRUSADE score above 30 (9.0% vs. 4.2%, P = 0.009). Our results showed that among patients with low bleeding risk, ticagrelor obviously reduced MACCE risk (2.5% vs. 4.9%, P = 0.022). Meanwhile, ticagrelor only slightly increased all bleeding, as well as bleeding with BARC2o reduce, with no statistical significance (P > 0.05). However, in patients of moderate-to-high bleeding risk, ticagrelor not only does not reduce the risk of MACCE but also causes an increase in the occurrence of MACCE (4.8% vs. 3.0%, P = 0.225), with an extreme increase in all bleeding (4.8% vs. 3.0%, P = 0.006) and bleeding of BARC2–5 (3.2% vs. 0.7%, P = 0.015) and BARC3–5 (2.1% vs. 0.4%, P = 0.040). Time-to-event curves for the first occurrence of NACE, MACCE, and bleeding outcomes in the subgroups of patients with low- and moderate-to-high bleeding risk during the 1-year follow-up period are shown in Figures 1–3 with an interaction analysis.

Table 3

Impact of CRUSADE bleeding risk on clinical outcome of ticagrelor versus clopidogrel adjusted by PSM, n (%)

Outcomes	Moderate-to-high bleeding risk				Low bleeding risk
	Ticagrelor (n = 189)	Clopidogrel (n = 706)	χ2 values	P	Ticagrelor (n = 590)	Clopidogrel (n = 852)	χ2 values	P
NACE	17 (9.0)	30 (4.2)	6.748	0.009	24 (4.1)	49 (5.8)	2.055	0.152
MACCE	9 (4.8)	21 (3.0)	1.470	0.225	15 (2.5)	42 (4.9)	5.233	0.022
Death	3 (1.6)	8 (1.1)	–	0.708	4 (0.7)	11 (1.3)	1.273	0.259
MI	1 (0.5)	8 (1.1)	–	0.693	4 (0.7)	7 (0.8)	–	1.000
Stroke	1 (0.5)	1 (0.1)	–	0.378	0 (0.0)	3 (0.4)	–	0.274
TVR	4 (2.1)	6 (0.8)	–	0.232	9 (1.5)	23 (2.7)	2.215	0.137
All bleeding	9 (4.8)	9 (1.3)	–	0.006	9 (1.5)	7 (0.8)	1.574	0.210
BARC2-5	6 (3.2)	5 (0.7)	–	0.015	2 (0.3)	1 (0.1)	–	0.571
BARC3-5	4 (2.1)	3 (0.4)	–	0.040	2 (0.3)	0 (0.0)	–	0.167

–: Fisher exact test. PSM: Propensity score matching; NACE: Net adverse clinical event; MACCE: Major adverse cardiac or cerebral events; MI: Myocardial infarction; TVR: Target vessel revascularization; BARC: Bleeding Academic Research Consortium.

Figure 1

Impact of bleeding risk on NACE among patients receiving ticagrelor versus clopidogrel. (a and b) Time-to-event curves for the NACE through 1-year follow-up in patients with low (a) or moderate-to-high bleeding risk (b); (c) Subgroup analysis on the net efficacy of ticagrelor versus clopidogrel in patients with different bleeding risk. NACE: Net adverse clinical event.

Figure 3

Impact of bleeding risk on all bleeding among patients receiving ticagrelor versus clopidogrel. (a and b) Time-to-event curves for the all bleeding through 1-year follow-up in patients with low (a) or moderate-to-high bleeding risk (b); (c) Subgroup analysis on the safety of ticagrelor versus clopidogrel in patients with different bleeding risk.

Impact of bleeding risk on MACCE among patients receiving ticagrelor vs. clopidogrel. (a and b) Time-to-event curves for the MACCE through 1-year follow-up in patients with low (a) or moderate-to-high bleeding risk (b); (c) Subgroup analysis on the ischemic efficacy of ticagrelor versus clopidogrel in patients with different bleeding risk. MACCE: Major adverse cardiac or cerebral events.

DISCUSSION

The current study observed the efficacy of ticagrelor versus clopidogrel among real-world Chinese patients with ACS undergoing PCI and the subgroups of patients with different baseline bleeding risks according to the CRUSADE score. We found that compared with clopidogrel, ticagrelor brought no antithrombotic benefit in the whole ACS population or in those with moderate-to-high baseline bleeding risk but markedly reduced the ischemic events in patients with low baseline bleeding risk. The baseline bleeding potential obviously affected the efficacy of ticagrelor vs. clopidogrel.

Unlike in Western countries, evidence for the efficacy of ticagrelor versus clopidogrel was conflicting in the population of East Asia. A randomized trial[6] and meta-analysis[7] demonstrated that ticagrelor was no longer superior over clopidogrel in efficiency, with increased major bleeding. Whereas, two observational studies reported that ticagrelor offered significantly or marginally better anti-ischemic protection than clopidogrel in ACS[8] or AMI[9] patients, both without an increase in the rate of major bleeding. The above evidence implied that some critical factor might be closely involved in the efficacy of ticagrelor, which was special in East Asian patients and was changed when ticagrelor was applied from Phase III RCTs to the real-world clinical practice.

Different from clopidogrel, ticagrelor retained its potent platelet inhibition in the East Asian population with no resistance phenomenon reported. Our previous PK/PD trial[17] found that even among patients with an eGFR below 60 ml/min, ticagrelor could inhibit platelet aggregation by 80% (30 patients in the ticagrelor group, all from China mainland). Therefore, the conflicting advantages of ticagrelor were unlikely due to the difference in antiplatelet activity. Substantial evidence has demonstrated that East Asian patients were much more susceptible to bleeding under antiplatelet therapy. It was reported that following the same does of ticagrelor, the degree of platelet inhibition was obviously higher in East Asian patients compared Caucasian.[18] What is worse, even on the same level of on-treatment platelet reactivity, East Asian patients were also more likely to have bleeding events.[19] There is no doubt that bleeding was closely related to the occurrence of ischemic events, directly or indirectly.[1011] Therefore, it might be the potential bleeding risk responsible for the conflicting results between Western and East Asia, as well as between RCTs and real-world studies among East Asian patients. However, few studies have focused on the impact of baseline bleeding risk on the efficacy of ticagrelor versus clopidogrel.

The present real-world study in mainland China found that the antithrombotic efficiency of ticagrelor was similar in comparison to clopidogrel in the whole ACS population, with an obvious or marginal increase in overall bleeding and major bleeding. We then further divided the patients into low- and moderate-to-high bleeding risk according to the CRUSADE score. First, we noticed that ticagrelor was more frequently administered in patients with low bleeding risk (75.7%), reflecting general concerns about the bleeding risk of ticagrelor and obvious selection bias for patients with low bleeding potential in the real-world clinical practice. We then surprisingly found that when compared with clopidogrel, ticagrelor was markedly superior in reducing MACCE in patients with low bleeding risk, with comparable major bleeding and slightly higher overall bleeding, which were quite similar with the results of the two observational studies in Taiwan, China.[89] What should also be mentioned is that even in the PRACTICAL study, bleeding that required admission was significantly 20% higher in the ticagrelor group. Hence, if there is no selective bias, East Asians who are more likely to have bleeding may not be the same or even have less major bleeding when using ticagrelor compared with clopidogrel.

Our results also showed that ticagrelor was associated with a worse occurrence of MACCE among patients with moderate-to-high bleeding risk, implying that a high baseline bleeding risk might attenuate the efficacy of ticagrelor in real-world patients of East Asia. What could also be shown is that the results of this subgroup were the primary reason for the comparable efficacy outcomes of ticagrelor versus clopidogrel in the whole ACS population. We believe that this finding might also be valuable to explain the outcomes of RCTs from East Asia, because, in an RCT, patients at moderate-to-high bleeding risk could not be excluded from the ticagrelor arm, as long as the general inclusion/exclusion criteria were met.

We noticed that in the following versions of ESC or ACC guidelines, ticagrelor was strongly recommended over clopidogrel in NSTE-ACS patients at moderate-to-high risk of ischemic events and in STEMI patients, as long as there is no contradiction.[12] Correspondingly, our results demonstrated that there should be more consideration of the bleeding potentiality, in addition to ischemic risk, when judging whether ticagrelor was superior in an individual patient from East Asia. Just as the DAPT score is used to guide whether to extend the duration of dual-antibiotic therapy,[20] the SYNTAX score II is used to assist in determining the method of revascularization;[21] perhaps for the East Asian population, there should also be an assessment method that combines ischemia and hemorrhage. More clinical trials focusing on a selection of P2Y12 antagonist and duration of the double antiplatelet therapy are still needed to provide critical evidence supporting the decision of optimal antiplatelet regimens, especially among East Asia patients with both high risks at ischemic and bleeding events.

Limitation

First, although ticagrelor was recommended in patients with STEMI, or with NSTE-ACS at moderate-to-high risk of ischemic events,[12] patients at low risk of ischemic events might be included in this study because of its nature of real-world study and without specific exclusion criteria. Therefore, the efficacy of ticagrelor might be underestimated. Second, the sample size was reduced after PSM, which leads to the loss of statistical significance in some bleeding results while the data remained unchanged. Third, the present study was observational; although PSM was performed, there might be some potential prognostic factors not included in the database and the matching. In addition, since prasugrel is not available in China, only ticagrelor and clopidogrel were observed in the present study. In addition, high combination rate of GPI (35% after PSM) was observed in the present study. Although we found no interaction impact of GPI on the bleeding outcomes of ticagrelor versus clopidogrel (P = 0.307 for interaction; data not shown), the “bail-out” use of GPI should be more highlighted in the real-world clinical practice. Finally, the study data were also inevitable on selection and (or) treatment bias, as well as underestimated event rate. Temporary discontinuation or switching of ticagrelor because of slight or mild bleeding (BARC1–2) might not be completely recorded in the current observational study, which might partially explain why we failed to discover a direct causation between hemorrhage and ischemic events in the subgroup of moderate-to-high bleeding risk in the current database (data not shown), which should be further confirmed in the further trials.

In conclusion, among real-world Chinese patients with ACS treated by PCI, the MACCE rate in the ticagrelor group was numerically lower than that in the clopidogrel group and affected by the baseline bleeding risk of patients. Ticagrelor showed superior efficacy in patients with low bleeding risk but loss its advantage in those with moderate-to-high bleeding risk.

Financial support and sponsorship

This work was sponsored by a grant from the National Key Research and Development Project (No. 2016YFC1301300).

Conflicts of interest

There are no conflicts of interest.