Abstract
Background
Ticagrelor reduces ischaemic end points in acute coronary syndromes. However, outcomes of ticagrelor versus clopidogrel in real‐world patients with acute coronary syndromes treated with percutaneous coronary intervention (PCI) remain unclear. We sought to examine whether treatment with ticagrelor is superior to clopidogrel in unselected patients with acute coronary syndromes treated with PCI.
Methods and Results
We used data from SCAAR (Swedish Coronary Angiography and Angioplasty Registry) for PCI performed in Västra Götaland County, Sweden. The database contains information about all PCI performed at 5 hospitals (∼20% of all data in SCAAR). All procedures between January 2005 and January 2015 for unstable angina/non‒ST‐segment‒elevation myocardial infarction and ST‐segment‒elevation myocardial infarction were included. We used instrumental variable 2‐stage least squares regression to adjust for confounders. The primary combined end point was mortality or stent thrombosis at 30 days, secondary end points were mortality at 30 days and 1‐year, stent thrombosis at 30 days, in‐hospital bleeding, in‐hospital neurologic complications and long‐term mortality. A total of 15 097 patients were included in the study of which 2929 (19.4%) were treated with ticagrelor. Treatment with ticagrelor was not associated with a lower risk for the primary end point (adjusted odds ratio [aOR], 1.20; 95% CI, 0.87–1.61; P=0.250). Estimated risk of death at 30 days (aOR, 1.18; 95% CI, 0.88–1.64; P=0.287) and at 1‐year (aOR, 1.28; 95% CI, 0.86–1.64; P=0.556) was not different between the groups. The risk of in‐hospital bleeding was higher with ticagrelor (aOR, 2.88; 95% CI, 1.53–5.44; P=0.001).
Conclusions
In this observational study, treatment with ticagrelor was not superior to clopidogrel in patients with acute coronary syndromes treated with PCI.
Keywords: acute coronary syndrome, clopidogrel, non‐ST‒segment‐elevation myocardial infarction, ST‐segment‒elevation myocardial infarction, P2Y12 receptor antagonists, percutaneous coronary intervention, ticagrelor, unstable angina
Subject Categories: Catheter-Based Coronary and Valvular Interventions, Percutaneous Coronary Intervention, Revascularization, Treatment, Pharmacology
Nonstandard Abbreviations and Acronyms
- HR
hazard ratio
- PCI
percutaneous coronary intervention
- PLATO
Study of Platelet Inhibition and Patient Outcomes
- SCAAR
Swedish Coronary Angiography and Angioplasty Registry
- STEMI
ST‐segment‒elevation myocardial infarction
- SWEDEHEART
The Swedish Web‐system for enhancement and development of evidence‐based care in heart disease evaluated according to recommended therapies
Clinical Perspective
What Is New?
Different from the results from PLATO (Study of Platelet Inhibition and Patient Outcomes), ticagrelor was not superior to clopidogrel in this observational all‐comer analysis of patients with acute coronary syndromes undergoing percutaneous coronary intervention in Western Sweden.
In the present analysis treatment with ticagrelor was not associated with any mortality benefits in comparison with clopidogrel while the incidence of in‐hospital bleeding was significantly increased.
What Are the Clinical Implications?
Benefits of ticagrelor, as demonstrated in the PLATO trial, may not be externally valid when applied to unselected patients with acute coronary syndromes undergoing percutaneous coronary intervention.
Current European and American guidelines commend ticagrelor over clopidogrel for treatment of patients with acute coronary syndrome (ACS). 1 , 2 These guidelines are based on the PLATO (Study of Platelet Inhibition and Patient Outcomes) trial which demonstrated that ticagrelor improved the composite end point of vascular death, non‐fatal myocardial infarction, or stroke. 3 In addition, it has been argued that ticagrelor also decreases all‐cause mortality in ACS. Since then, ticagrelor has been increasingly endorsed as the favored P2Y12 receptor antagonist worldwide. 1 , 2 , 4 However, since the publication of the PLATO trial 1 decade ago, observational studies in unselected populations with ACS treated with ticagrelor have generated conflicting results. While some post‐marketing observational reports confirmed the PLATO data, 5 , 6 , 7 other studies reported that ticagrelor is not superior to clopidogrel. 8 , 9 , 10 , 11 This discrepancy has revived the discussion about the external validity of the PLATO trial and constitute the rationale for our analysis. Thus, we sought to investigate whether treatment with ticagrelor is superior to clopidogrel in an unselected population of patients with ACS who were treated with PCI.
Methods
Database and Study Population
We used data from SCAAR (Swedish Coronary Angiography and Angioplasty Registry) database reported from 5 hospitals within the Västra Götaland County in western Sweden—representing ≈20% of the total SCAAR registry. Because of the sensitive nature of the data collected for this study, requests to access the data set from qualified researchers trained in human subject confidentiality protocols may be sent to the SCAAR working group via the corresponding author of this paper. All PCI procedures between 2005 and 2015 for unstable angina/non‒ST‐segment‒elevation myocardial infarction and ST‐segment‒elevation myocardial infarction (STEMI) were included. Unstable angina/non‒ST‐segment‒elevation myocardial infarction and STEMI were defined according to the diagnostic criteria established by the European Society of Cardiology. 1 SCAAR is a web‐based platform (established in 1992), containing >250 clinical, angiographic, and PCI‐related variables which document the entire interventional process (www.ucr.uu.se/swedeheart). Swedish Health Authorities sponsor the registry with no funding from commercial interest. SCAAR is an integral part of the larger SWEDEHEART (The Swedish Web‐system for enhancement and development of evidence‐based care in heart disease evaluated according to recommended therapies) registry platform. More detailed information about SWEDEHEART has been published elsewhere. 12 , 13 The SCAAR database merged with the national death registry—managed by the Swedish National Board of Health and Welfare—to obtain information about the vital status. The ethical committee at the University of Gothenburg has approved the study protocol. Informed consent was not required. Additional methods can be found in Data S1.
Statistical Analysis
Continuous variables are shown as mean±SD and categorical variables are presented as frequencies. Normal distribution was assessed by the Shapiro‐Wilk test as well as inspection of histograms. Linear regression was used to test the intergroup differences in continuous variables while logistic regression was used for categorical variables.
Missing data were imputed with the multiple imputation chain‐equation method 14 , 15 with 5 data sets. The calendar year, an indicator of missingness and an event indicator were included as regular variables. 16 Continuous variables were imputed by ordinary least‐squares multiple regression, binary variables by logistic regression, and categorical variables by multinomial logistic regression. The imputation procedure and subsequent analyses were done according to Rubin’s protocol 17 under the assumption that missing data are missing at random.
We based our primary statistical model on the instrumental variable analysis to reduce bias because of unmeasured and unknown confounders. This method is a post‐hoc analytic technique based on statistical principles similar to those used in the analysis of randomized controlled trials 18 , 19 , 20 (see supplement for more information). In addition to the instrumental variable, propensity score (as continuous variable) was entered into 2‐stage least squares regression. The score was calculated from the logistic model based on all variables presented in Tables 1 and 2 that were significantly different (P<0.05) between the 2 groups. We used multilevel Cox proportional‐hazards regression using the same instrumental variable estimator and propensity score as in the primary model based on 2‐stage least squares regressions to compare the groups long‐term mortality. Because the SCAAR registry is a hierarchical database with clustering of patients within hospitals and regions, we entered individual hospitals into the regression model as random effects variables. We assessed trends in 30‐days and 1‐year mortality over the study period by including the calendar year into the logistic regression as a continuous variable in addition to age and sex.
Table 1.
Patients’ Characteristics at Baseline
| Clopidogrel (n=12 168) | Ticagrelor (n=2929) | P Value | |
|---|---|---|---|
| Age, y | 0.829 | ||
| Mean (SD) | 67.30 (11.48) | 67.25 (11.64) | |
| Median (Q1, Q3) | 68.0 (59.0, 76.0) | 67.0 (59.0, 76.0) | |
| Age >75 y — no. (%) | 0.593 | ||
| Yes | 3613 (29.7) | 855 (29.2) | |
| Creatine clearance | 0.964 | ||
| Mean (SD) | 86.85 (34.9) | 86.81 (35.7) | |
| Median (Q1, Q3) | 82.8 (62.7, 106.9) | 83.7 (60.9, 106.8) | |
| BMI | <0.001 | ||
| Mean (SD) | 27.1 (5.2) | 27.6 (8.70) | |
| Median (Q1, Q3) | 26.5 (24.2, 29.3) | 26.7 (24.3, 29.7) | |
| Sex — no. (%) | 0.771 | ||
| Male | 8812 (72.4) | 2129 (72.7) | |
| Female | 3356 (27.6) | 800 (27.3) | |
| Diabetes mellitus — no. (%) | 0.192 | ||
| Yes | 1865 (15.3) | 481 (16.4) | |
| Missing | 140 (1.2) | 11 (0.4) | |
| Hypertension — no. (%) | 0.007 | ||
| Yes | 5606 (46.1) | 1446 (49.4) | |
| Missing | 363 (3.0) | 54 (1.8) | |
| Hyperlipidaemia — no. (%) | <0.001 | ||
| Yes | 4840 (41.2) | 964 (33.7) | |
| Missing | 418 (3.4) | 69 (2.6) | |
| Smoking — no. (%) | 0.058 | ||
| Never smoker | 4644 (38.2) | 1151 (39.3) | |
| Previous smoker | 3541 (29.1) | 853 (29.1) | |
| Active smoker | 2637 (21.7) | 723 (27.3) | |
| Missing | 1346 (11.1) | 202 (6.9) | |
| Previous myocardial infarction — no. (%) | 0.002 | ||
| Yes | 2275 (18.7) | 466 (15.9) | |
| Missing | 180 (1.5) | 46 (1.6) | |
| Previous CABG — no. (%) | 0.010 | ||
| Yes | 858 (7.1) | 163 (5.6) | |
| Missing | 4 (0.0) | 0 (0.0) | |
| Previous PCI — no. (%) | 0.030 | ||
| Yes | 1572 (12.9) | 335 (11.4) | |
| Indication for PCI — no. (%) | <0.001 | ||
| UA/NSTEMI | 6947 (57.1) | 1491 (50.9) | |
| STEMI | 5221 (42.9) | 1438 (49.1) | |
| Cardiogenic shock — no. (%) | 0.032 | ||
| Yes | 250 (2.9) | 64 (2.2) | |
| Missing | 3637 (29.9) | 2 (0.1) | |
| Procedure performed off‐hours/regular hours — no. (%) | <0.001 | ||
| Acute‐regular hours | 3094 (25.7) | 746 (25.5) | |
| Acute‐off‐hours | 3924 (32.6) | 1078 (36.9) | |
| Subacute‐regular hours | 4628 (38.0) | 800 (27.3) | |
| Subacute‐off‐hours | 388 (3.2) | 298 (10.2) | |
| Missing | 134 (1.1) | 7 (0.2) |
BMI indicates body mass index; CABG, coronary artery by‐pass surgery; NSTEMI, non–ST‐segment‐elevation myocardial infarction; PCI, percutaneous coronary intervention; STEMI, ST‐segment–elevation myocardial infarction; and UA, unstable angina.
Table 2.
Angiography and PCI
| Clopidogrel (n=12 168) | Ticagrelor (n=2929) | P Value | |
|---|---|---|---|
| Number of stents | <0.001 | ||
| Mean (SD) | 1.71 (0.98) | 1.88 (1.15) | |
| Median (Q1, Q3) | 1.0 (1.0, 2.0) | 2.0 (1.0, 2.0) | |
| Total stent length, mm | <0.001 | ||
| Mean (SD) | 30.83 (19.80) | 39.36 (27.97) | |
| Median (Q1, Q3) | 24.0 (18.0, 39.0) | 30.0 (18.0, 50.0) | |
| Mean stent diameter, mm | <0.001 | ||
| Mean (SD) | 3.08 (0.49) | 3.01 (0.45) | |
| Median (Q1, Q3) | 3.0 (2.8, 3.5) | 3.0 (2.8, 3.3) | |
| Total contrast, mL | <0.001 | ||
| Mean (SD) | 193.10 (88.44) | 177.83 (83.45) | |
| Median (Q1, Q3) | 173.0 (130.0, 234.0) | 160.0 (120.0, 219.0) | |
| Arterial access — no. (%) | <0.001 | ||
| Femoral | 6572 (54.3) | 598 (20.8) | |
| Radial | 5224 (43.2) | 2158 (75.0) | |
| Other | 303 (2.5) | 120 (4.2) | |
| Severity of coronary disease — no (%) | 0.036 | ||
| Normal | 44 (0.4) | 3 (0.1) | |
| Single vessel | 5935 (48.8) | 1470 (50.2) | |
| Multi‐vessel | 6189 (50.9) | 1456 (49.7) | |
| Treated vessel — no. (%) | 0.166 | ||
| RCA | 3791 (31.2) | 916 (31.3) | |
| LAD | 5415 (44.5) | 1288 (44.0) | |
| LCx | 2792 (22.9) | 668 (22.8) | |
| LM | 169 (1.4) | 57 (1.9) | |
| Type of lesion — no. (%) | 0.002 | ||
| De novo | 11787 (96.9) | 2874 (98.1) | |
| Other restenosis | 38 (0.3) | 6 (0.2) | |
| In‐stent restenosis | 334 (2.7) | 49 (1.7) | |
| Lesion classification — no. (%) | <0.001 | ||
| A | 1186 (9.8) | 282 (9.6) | |
| B1 | 3808 (31.3) | 957 (32.7) | |
| B2 | 4085 (33.6) | 837 (28.6) | |
| C | 2430 (20.0) | 571 (19.5) | |
| B1 Bifurcation | 218 (1.8) | 122 (4.2) | |
| B2 Bifurcation | 277 (2.3) | 113 (3.9) | |
| C Bifurcation | 139 (1.1) | 42 (1.4) | |
| Other | 15 (0.1) | 5 (0.2) | |
| Culprit lesion occlusion — no. (%) | 0.367 | ||
| No | 8276 (68.0) | 1951 (66.7) | |
| Yes <3 mo | 3736 (30.7) | 938 (32.0) | |
| Yes >3 mo 21 | 156 (1.3) | 40 (1.3) | |
| Methods used — no. (%) | <0.001 | ||
| POBA | 405 (3.3) | 66 (2.3) | |
| Direct stent | 3944 (32.4) | 832 (28.4) | |
| POBA and stent | 7661 (63.0) | 1972 (67.3) | |
| Atherectomy | 14 (0.1) | 1 (0.0) | |
| Rotablator | 9 (0.1) | 2 (0.1) | |
| Wire‐attempt | 53 (0.4) | 12 (0.4) | |
| DEB | 14 (0.1) | 7 (0.2) | |
| DEB and stent | 6 (0.0) | 1 (0.0) | |
| Other | 61 (0.5) | 36 (1.3) | |
| Type of stent — no. (%) | <0.001 | ||
| BMS | 9324 (76.6) | 309 (10.5) | |
| DES | 2844 (23.4) | 2620 (89.5) | |
| Thrombus aspiration — no. (%) | 0.090 | ||
| Yes | 1054 (8.7) | 283 (9.7) | |
| Missing | 11 (0.1) | 0 | |
| FFR/iFR — no (%) | <0.001 | ||
| Yes | 224 (1.8) | 150 (5.1) | |
| Missing | 2 (0.0) | 1 (0.0) | |
| IVUS — no. (%) | 0.060 | ||
| Yes | 428 (3.5) | 129 (4.4) | |
| Missing | 2 (0.0) | 1 (0.0) | |
| OCT — no. (%) | <0.001 | ||
| Yes | 5 (0.0) | 27 (0.9) | |
| Missing | 1 (0.0) | 1 (0.0) | |
| Complete revascularisation — no. (%) | <0.001 | ||
| Yes | 7263 (59.7) | 1882 (64.3) | |
| Missing | 8 (0.1) | 7 (0.2) | |
| Intra‐aortic balloon pump — no. (%) | <0.001 | ||
| Yes | 360 (2.9) | 20 (0.7) | |
| Missing | 12 (0.1) | 0 (0) | |
| UH/LMWH — no. (%) | <0.001 | ||
| Yes | 6225 (51.2) | 2378 (81.2) | |
| Missing | 2 (0.0) | 0 (0.0) | |
| Bivalirudin — no. (%) | <0.001 | ||
| Yes | 7500 (63.7) | 1511 (51.6) | |
| Missing | 1 (0.0) | 0 (0.0) | |
| GP2b/3a inhibitor — no. (%) | <0.001 | ||
| Yes | 3366 (27.7) | 119 (4.1) |
BMS indicates bare metal stent; CTO, chronic total occlusion; DEB, drug‐eluting balloon; DES, drug‐eluting stent; FFR, fractional flow reserve; GP2b/3a, glycoprotein 2b/3a receptor‐inhibitor; iFR, instantaneous wave–free ratio; IVUS, intravascular ultrasound; LAD, left anterior descending coronary artery; LCx, left circumflex artery; LM, left main; LWMH, low molecular weight heparin; OCT, optical coherence tomography; POBA, plain old balloon angioplasty; RCA, right coronary artery; and UH, unfractionated heparin.
For sensitivity analysis we used propensity score models based on the inverse probability of treatment weighting. 22 Significant predictors of treatment with ticagrelor for each patient were identified by fitting a logistic regression model with (1) a binary dependent variable representing treatment ticagrelor or clopidogrel and (2) candidate variables consisting of the patient‐related predictors of the type of therapy used. All variables in Tables 1 and 2 and hospital, were entered into the logistic model.
Post‐Estimation Diagnostics
Goodness‐of‐fit (calibration) for the models was assessed with the Hosmer–Lemeshow test. Multi‐collinearity between the variables in the model was assessed by calculation of the variance inflation factor. All statistical analyses were performed using Stata software (version 16.0, StataCorp, College Station, Texas, USA). Instrumental variable 2‐stage least squares regression models were completed using the IVREG2 module. 23 All tests were 2‐tailed, and a P<0.05 was considered statistically significant. Because of multiple analyses, P<0.05 was expected to occur by chance in 1 out of 20 analyses. The validity of instrumental variables was examined by calculation of the standardized difference of variables that reflects known patient’s characteristics and procedural details in treated and untreated patients stratified on the calendar year during the study period. We used logistic regression to evaluate the predictive power of instruments for treatment with P2Y12 antagonists as well as for primary and secondary outcomes. All tests were 2‐tailed, and P<0.05 was considered statistically significant. All statistical analyses were performed using Stata software (version 16.0, StataCorp, College Station, Texas, USA). All tests were 2‐tailed and a P<0.05 was considered statistically significant.
Results
Baseline Characteristics and Treatments
A total of 15 097 patients were included in the study of which 12 168 (80.6%) were treated with clopidogrel and 2929 (19.4%) were treated with ticagrelor. Baseline patients’ characteristics are presented in Table 1 and procedure‐related details are presented in Table 2. Women represented 27% of all patients and STEMI was reported in 44%; 30% of all patients were >75 and 16% had diabetes mellitus. We found no difference between the 2 groups on age, sex, creatine clearance, body mass index, diabetes mellitus, smoking habits, treated vessel, culprit vessels occlusion, and the use of thrombus aspiration during PCI. Patients treated with a ticagrelor were more likely to have had previous myocardial infarction, prior PCI, and prior coronary artery by‐pass surgery. Patients treated with clopidogrel were more likely to have hyperlipidemia, non‐ST segment elevation (NSTE)‐ACS, cardiogenic shock, and to be treated with intra‐aortic balloon pump but were less likely to have a radial access and to be treated with PCI during off‐hours. Patients treated with ticagrelor were more often treated with bivalirudin but less often with GP2b/3a receptor antagonist. They were more likely to be completely revascularized and to have thrombus aspiration before stent placement but were less likely to receive direct stenting. Drug‐eluting stents, fractional flow reserve/instantaneous wave–free ratio, intravascular ultrasound and optical coherence tomography were more often used in patients treated with ticagrelor; 4515 (30%) patients were treated with PCI after the change from clopidogrel to ticagrelor as the default P2Y12 receptor antagonist in February 2012 (Figure 1). Of these, 1587 (35%) were treated with clopidogrel.
Figure 1.
Percentage breakdown per calendar month for use of the 2 P2Y12 antagonists—clopidogrel and ticagrelor—over the study period between January 2005 and January 2015 in Västra Götaland County. In February 2012, the regional executive board decided to replace clopidogrel with ticagrelor as the default P2Y12 receptor antagonist for patients with ACS. Figure 1 depicts the percentage breakdown for all patients treated during the study period with either clopidogrel (n=12 168; blue) or ticagrelor (n=2929; red) per calendar month during the study period. Percentages for each column per month thus add up to the entire study population (100%) for each treatment group in all fig captions.
Clinical Outcome
Primary End Point
The primary outcome was assessed at 30 days after the index hospitalization. Information about vital status and stent thrombosis was available for all included patients. There were 555 events at 30 days of which 53 (9.5%) were stent thromboses. Treatment with ticagrelor was not associated with a lower risk for the primary end point (adjusted odds ratio [aOR], 1.20; 95% CI, 0.87–1.61; P=0.250, Table 3).
Table 3.
Clinical Outcomes
| Clinical Outcome | Clopidogrel (n=12 168) | Ticagrelor (n=2929) | Adjusted Odds Ratio | 95% CI | P Value | Missing n (%) |
|---|---|---|---|---|---|---|
| Primary end point | ||||||
| Death or definite stent thrombosis at 30 d — n (%) | 439 (3.6) | 116 (4.0) | 1.20 | 0.87–1.61 | 0.250 | 0 |
| Secondary end points | ||||||
| Death, myocardial infarction, or stroke at 30 d — n (%) | 697 (6.5) | 205 (8.1) | 1.25 | 0.95–1.63 | 0.104 | 2128 (14.1) ‡ |
| Death, myocardial infarction, or stroke at 1 y — n (%) | 1441 (13.4) | 340 (13.4) | 1.15 | 0.93–1.45 | 0.192 | 2128 (14.1) ‡ |
| Death at 30 d — n (%) | 439 (3.6) | 116 (4.0) | 1.18 | 0.88–1.64 | 0.287 | 0 |
| Definite stent thrombosis at 30 d — n (%) | 45 (0.4) | 8 (0.3) | 1.30 | 0.54–3.10 | 0.556 | 0 |
| Death at 1 y — no. (%) | 729 (6.1) | 115 (6.3) | 1.28 | 0.86–1.93 | 0.222 | 0 |
| Definite stent thrombosis at 1 y — no. (%) | 76 (0.7) | 10 (0.4) | 1.18 | 0.54–2.56 | 0.682 | 0 |
| In‐hospital bleeding — no. (%)* | 489 (4.2) | 163 (6.6) | 2.88 | 1.53–5.44 | 0.001 | 979 (6.5) |
| Neurologic complication — no. (%) † | 19 (0.2) | 5 (0.2) | 0.95 | 0.44–2.02 | 0.891 | 778 (5.2) |
Major bleeding (Bleeding Academic Research Consortium type 3), minor bleeding (Bleeding Academic Research Consortium type 2).
Stroke or transient ischemic attack.
Data missing for myocardial infarction and stroke.
Secondary End Points
Between 2000 and 2015, the overall mortality at 30 days (aOR, 1.01; 95% CI, 0.97–1.06, Ptrend=0.514), (Figure 2) and at 1‐year (aOR, 1.02; 95% CI, 0.99–1.06, Ptrend=0.174), (Figure 3) did not change. A total of 844 (5.6%) were dead at 1‐year after the index hospitalization. Estimated risk of death at 30 days (aOR, 1.18; 95% CI, 0.88–1.64; P=0.287) and 1‐year (aOR, 1.28; 95% CI, 0.86–1.93; P=0.222) was not different between the 2 groups (Table 3). There was no difference in definite stent thrombosis at 30 days (aOR, 1.30; 95% CI, 0.54–3.10; P=0.556), (Table 3)]. Data about in‐hospital bleeding and in‐hospital neurologic complications were missing in 1757 (11.6%) patients. The risk of in‐hospital bleeding was higher with ticagrelor (aOR, 2.88; 95% CI, 1.53–5.44; P=0.001), Table 3). There was no difference in neurologic complications during hospitalization (aOR, 0.95; 95% CI, 0.44–2.02; P=0.891), The mean follow‐up time was 1633±1048 (range 0–3692) in the whole cohort. Patients treated with clopidogrel had longer mean follow‐up time (1918±963, range 0–3692) than patients treated with ticagrelor (454±280, range 0–1179). There were 2618 (17.3%) deaths during the total follow‐up time. We found no significant difference in long‐term mortality between the 2 groups (adjusted hazard ratio [aHR], 1.07; 95% CI, 0.89–1.29; P=0.437), (Figure 4). We found no significant difference in long‐term major adverse cardiovascular events (death, myocardial infarction, or stroke) between the 2 groups (aHR, 1.08; 95% CI, 0.91–1.28; P=0.400). We found no interaction between the P2Y12 antagonist and age, sex, diabetes mellitus, STEMI/non‒ST‐segment‒elevation myocardial infarction, the use of GP2b/3a inhibitor and access site for any of primary or secondary end points (all P interaction>0.05).
Figure 2.
Cumulative incidence of mortality at 30‐days in Västra Götaland County between 2005 and 2015.
Figure 3.
Cumulative incidence of mortality at 1 year in Västra Götaland County between 2005 and 2015.
Figure 4.
Long‐term mortality in patients treated with ticagrelor and clopidogrel in Västra Götaland County between 2005 and 2015.
Sensitivity Analyses
The results from the sensitivity analyses based on logistic regression adjusted with inverse probability of treatment weighted propensity score are presented in Table S1. Generally, the results were in agreement with the primary model with the estimated risk of in‐hospital bleeding (with ticagrelor) being higher than in the primary model.
Data Analysis and Post‐Estimation Diagnostics
Post‐estimation analysis for the logistic regression models by Hosmer–Lemeshow test showed adequate goodness‐of‐fit for the models (P>0.05). Squared covariate terms had no explanatory power in any of the models (link test, P>0.05). The average variance inflation factor was <5.0 for all models, indicating a lack of multicollinearity between the variables.
Discussion
Among 15 097 patients with ACS undergoing PCI between January 2005 and January 2015 in Sweden, treatment with ticagrelor was not associated with a lower risk of all‐cause mortality and stent thrombosis. However, ticagrelor‐treated patients had a higher risk of in‐hospital bleeding. These findings are different from the PLATO trial and raise important matters for discussion.
Our results are supported by several observational studies. In a study performed within the British Cardiovascular Intervention Society PCI registry, the authors investigated the impact of P2Y12 receptor antagonists on all‐cause mortality at 1 year in patients with STEMI treated with primary PCI. 9 In agreement with our findings, this study showed no difference in mortality between ticagrelor and clopidogrel. However, the authors reported improved outcome in patients treated with prasugrel compared both with clopidogrel and ticagrelor. Moreover, in a population of Dutch patients with ACS (contemporaneous with our population), treatment with ticagrelor increased risk of the composite net clinical end point of all‐cause death, myocardial infarction, stroke, or major bleeding at 1 year. 8 While ticagrelor was not associated with benefits on ischemic end points, the risk of major bleeding was almost 3 times higher in the patients treated with ticagrelor. Finally, a recently published Canadian registry analysis of comparable size and patient case–mix as in our study, did not show that ticagrelor was superior to clopidogrel regarding major adverse cardiovascular events in patients with ACS undergoing PCI. 11
One previous observational study from SWEDEHEART reported that ticagrelor was associated with lower all‐cause mortality but with higher risk of bleeding. 5 However, while the previous study evaluated only survival after hospital discharge (1434 cases of in‐hospital deaths were excluded from the analysis), the present study includes all deaths from hospital admission until the end of the follow‐up time. We argue that our approach is more appropriate for the assessment of the external validity of the PLATO trial for all‐cause mortality.
Our results are also supported by the evidence from several randomized controlled trials. First, the cluster‐randomized ARTEMIS (Affordability and Real‐World Antiplatelet Treatment Effectiveness After Myocardial Infarction Study) investigators assessed the impact of medication co‐payment vouchers on the P2Y12 receptor antagonists compliance and major adverse cardiovascular events. 24 In >11 000 patients with myocardial infarction from 301 different hospitals in the United States, patients receiving vouchers had higher ticagrelor persistence, but this did not translate into clinical benefit. Also, in the TOPIC trial (Timing of Platelet Inhibition after ACS), de‐escalation of P2Y12 receptor antagonist (from ticagrelor to clopidogrel) 1 month after PCI was studied in a contemporary ACS population. 25 The risk for the composite net clinical end point (cardiovascular death, unplanned hospitalization, urgent revascularization, stroke, bleeding) at 1‐year was lower in patients who switched from ticagrelor to clopidogrel. The difference was driven by an increased incidence of bleeding with no variance in ischemic end points. 8 , 9
Why our results differ from the PLATO trial may be explained by the following features. While in PLATO more than one third of patients were treated conservatively, all patients in our study were treated with PCI. Compared with our study, patients in PLATO were younger and more frequently men which confirms well‐known discrepancies between randomized clinical trials and unselected patient populations. 26 Consequently, based on the established predictors for bleeding, 27 , 28 our unselected population may have been more prone to bleeding when exposed to a potent antiplatelet agent. Indeed, the increased risk of bleeding may constitute mediating pathway 29 for the observed absence of mortality benefit with ticagrelor in an unselected population. 8 , 9 Another simple but potent mediator for attenuation of ticagrelors’ efficacy outside randomized clinical trial may be decreased compliance because of the higher rate of adverse events (dyspnea, bleeding), administration of the drug twice per day and higher costs. Indeed, in the PLATO trial, ticagrelor had higher discontinuation rate than clopidogrel with 23% of patients being non‐compliant. It seems reasonable to argue that non‐compliance with ticagrelor is even higher outside the settings of well‐controlled clinical trial and with a different (more non‐compliant prone) patient population.
Like many medical novelties after introduction to the market, ticagrelor has imposed a considerable economic burden to healthcare resources. In our region, this additional cost reaches an extra 2.000.000 Euros annually compared with clopidogrel. Yet, the totality of evidence at present time does not unequivocally support the cost‐effectiveness of ticagrelor.
Several limitations need to be addressed. Swedish practice and outcomes do not necessarily reflect cardiovascular care in other areas in the world. We acknowledge that observational design of our study produces a risk for residual confounding. However, our statistical models were based on instrumental variable analysis which is one of the best methods to eliminate both known and unknown confounders in observational studies. In addition, most known confounders related to patient’s characteristics and procedural details favored ticagrelor. It is plausible that possible unmeasured confounders also favored ticagrelor because these patients were treated during the recent era. The events in the study were not adjudicated by an adjudication committee. However, regular external monitoring and data validation are performed in SWEDEHEART to ascertain high data accuracy. 13 Our data did not allow differentiation between cardiac and non–cardiac death, but mortality is a robust end point. Possibility for administrative errors about citizens’ vital status in Sweden is low and nearly 100 percent of deaths are registered within the first month. 30 Finally, we could not account for the duration of antiplatelet treatment and possible cross–over between the groups as this information is not reported to the registry. The analysis was, therefore, based on the intention to treat principle. The recommended duration for dual antiplatelet therapy in patients with ACS has been 12 months during the study period.
In conclusion, in this observational study, treatment with ticagrelor was not superior to clopidogrel in patients with ACS treated with PCI and was associated with an increased risk of bleeding. Benefits of ticagrelor, as demonstrated in the PLATO trial, may not be externally valid when applied to unselected patients with ACS undergoing PCI.
Sources of Funding
None.
Disclosures
OA and EO report speaker fees from Astra Zeneca. The remaining authors have no disclosures to report.
Supporting information
(J Am Heart Assoc. 2020;9:e015990 DOI: 10.1161/JAHA.119.015990.)
For Sources of Funding and Disclosures, see page 10.
References
- 1. Neumann FJ, Sousa‐Uva M, Ahlsson A, Alfonso F, Banning AP, Benedetto U, Byrne RA, Collet JP, Falk V, Head SJ, et al. 2018 ESC/EACTS Guidelines on myocardial revascularization. Eur Heart J. 2019;40:87–165. [DOI] [PubMed] [Google Scholar]
- 2. Levine GN, Bates ER, Bittl JA, Brindis RG, Fihn SD, Fleisher LA, Granger CB, Lange RA, Mack MJ, Mauri L, et al. 2016 ACC/AHA guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease: a report of the American College of Cardiology/American Heart Association Task Force on clinical practice guidelines: an update of the 2011 ACCF/AHA/SCAI guideline for percutaneous coronary intervention, 2011 ACCF/AHA guideline for coronary artery bypass graft surgery, 2012 ACC/AHA/ACP/AATS/PCNA/SCAI/STS Guideline for the diagnosis and management of patients with stable ischemic heart disease, 2013 ACCF/AHA guideline for the management of ST‐elevation myocardial infarction, 2014 AHA/ACC guideline for the management of patients with non‐ST‐elevation acute coronary syndromes, and 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery. Circulation. 2016;134:e123–e155. [DOI] [PubMed] [Google Scholar]
- 3. Wallentin L, Becker RC, Budaj A, Cannon CP, Emanuelsson H, Held C, Horrow J, Husted S, James S, Katus H, et al. Ticagrelor versus clopidogrel in patients with acute coronary syndromes. N Engl J Med. 2009;361:1045–1057. [DOI] [PubMed] [Google Scholar]
- 4. Angeras O, Hasvold P, Thuresson M, Deleskog A, ÖBraun O. Treatment pattern of contemporary dual antiplatelet therapies after acute coronary syndrome: a Swedish nationwide population‐based cohort study. Scand Cardiovasc J. 2016;50:99–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Sahlen A, Varenhorst C, Lagerqvist B, Renlund H, Omerovic E, Erlinge D, Wallentin L, James SK, Jernberg T. Outcomes in patients treated with ticagrelor or clopidogrel after acute myocardial infarction: experiences from SWEDEHEART registry. Eur Heart J. 2016;37:3335–3342. [DOI] [PubMed] [Google Scholar]
- 6. Peyracchia M, Saglietto A, Biolè C, Raposeiras‐Roubin S, Abu‐Assi E, Kinnaird T, Ariza‐Solé A, Liebetrau C, Manzano‐Fernández S, Boccuzzi G, et al. Efficacy and Safety of clopidogrel, prasugrel and ticagrelor in ACS patients treated with PCI: a propensity score analysis of the RENAMI and BleeMACS registries. Am J Cardiovasc Drugs. 2020;20:259–269. [DOI] [PubMed] [Google Scholar]
- 7. Ahn JH, Ahn Y, Jeong MH, Kim JH, Hong YJ, Sim DS, Kim MC, Hwang JY, Yoon JH, Seong IW, et al. Ticagrelor versus clopidogrel in acute myocardial infarction patients with multivessel disease; From Korea Acute Myocardial Infarction Registry‐National Institute of Health. J Cardiol. 2020;75:478–484. [DOI] [PubMed] [Google Scholar]
- 8. Zocca P, van der Heijden LC, Kok MM, Lowik MM, Hartmann M, Stoel MG, Louwerenburg JW, de Man F, Linssen GCM, Knottnerus IL, et al. Clopidogrel or ticagrelor in acute coronary syndrome patients treated with newer‐generation drug‐eluting stents: CHANGE DAPT. EuroIntervention. 2017;13:1168–1176. [DOI] [PubMed] [Google Scholar]
- 9. Olier I, Sirker A, Hildick‐Smith DJR, Kinnaird T, Ludman P, de Belder MA, Baumbach A, Byrne J, Rashid M, Curzen N, et al. Association of different antiplatelet therapies with mortality after primary percutaneous coronary intervention. Heart (British Cardiac Society). 2018;8:1683–1690. [DOI] [PubMed] [Google Scholar]
- 10. Alexopoulos D, Xanthopoulou I, Deftereos S, Hamilos M, Sitafidis G, Kanakakis I, Pentara I, Vavouranakis M, Davlouros P, Hahalis G, et al. Contemporary antiplatelet treatment in acute coronary syndrome patients undergoing percutaneous coronary intervention: 1‐year outcomes from the GReek AntiPlatElet (GRAPE) Registry. J Thromb Haemost. 2016;14:1146–1154. [DOI] [PubMed] [Google Scholar]
- 11. Turgeon RD, Koshman SL, Youngson E, Har B, Wilton SB, James MT, Graham MM. Association of ticagrelor vs clopidogrel with major adverse coronary events in patients with acute coronary syndrome undergoing percutaneous coronary intervention. JAMA Intern Med. 2020;180:420–428. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Stenestrand U, Wallentin L. Early statin treatment following acute myocardial infarction and 1‐year survival. JAMA. 2001;285:430–436. [DOI] [PubMed] [Google Scholar]
- 13. Jernberg T, Attebring MF, Hambraeus K, Ivert T, James S, Jeppsson A, Lagerqvist B, Lindahl B, Stenestrand U, Wallentin L. The Swedish Web‐system for enhancement and development of evidence‐based care in heart disease evaluated according to recommended therapies (SWEDEHEART). Heart (British Cardiac Society). 2010;96:1617–1621. [DOI] [PubMed] [Google Scholar]
- 14. van Buuren S. Multiple imputation of discrete and continuous data by fully conditional specification. Stat Methods Med Res. 2007;16:219–242. [DOI] [PubMed] [Google Scholar]
- 15. Angeras O, Albertsson P, Karason K, Ramunddal T, Matejka G, James S, Lagerqvist B, Rosengren A, Omerovic E. Evidence for obesity paradox in patients with acute coronary syndromes: a report from the Swedish Coronary Angiography and Angioplasty Registry. Eur Heart J. 2013;34:345–353. [DOI] [PubMed] [Google Scholar]
- 16. White IR, Royston P. Imputing missing covariate values for the Cox model. Stat Med. 2009;28:1982–1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Rubin DB. Inference and missing data. Biometrika. 1976;63:581–590. [Google Scholar]
- 18. Harris KM, Remler DK. Who is the marginal patient? Understanding instrumental variables estimates of treatment effects. Health Serv Res. 1998;33:1337–1360. [PMC free article] [PubMed] [Google Scholar]
- 19. Mcclellan M, Mcneil BJ, Newhouse JP. Does more intensive treatment of acute myocardial‐infarction in the elderly reduce mortality ‐ analysis using instrumental variables. J Am Med Assoc. 1994;272:859–866. [PubMed] [Google Scholar]
- 20. Rassen JA, Schneeweiss S, Glynn RJ, Mittleman MA, Brookhart MA. Instrumental variable analysis for estimation of treatment effects with dichotomous outcomes. Am J Epidemiol. 2009;169:273–284. [DOI] [PubMed] [Google Scholar]
- 21. Converse RL Jr, Jacobsen TN, Toto RD, Jost CM, Cosentino F, Fouad‐Tarazi F, Victor RG. Sympathetic overactivity in patients with chronic renal failure. N Engl J Med. 1992;327:1912–1918. [DOI] [PubMed] [Google Scholar]
- 22. Heckman JJIH, Petra T. Matching as an econometric evaluation estimator. Rev Econ Stud. 1998;65:261–294. [Google Scholar]
- 23. Christopher FBSME, Stillman S. IVREG2: Stata module to extended instrumental variables/2SLS, GMM and AC/HAC. LIML and k‐class regression. In: Boston College Department of Economics; 2002.
- 24. Wang TY, Kaltenbach LA, Cannon CP, Fonarow GC, Choudhry NK, Henry TD, Cohen DJ, Bhandary D, Khan ND, Anstrom KJ, et al. Effect of medication co‐payment vouchers on P2Y12 inhibitor use and major adverse cardiovascular events among patients with myocardial infarction: the ARTEMIS randomized clinical trial. JAMA. 2019;321:44–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Cuisset T, Deharo P, Quilici J, Johnson TW, Deffarges S, Bassez C, Bonnet G, Fourcade L, Mouret JP, Lambert M, et al. Benefit of switching dual antiplatelet therapy after acute coronary syndrome: the TOPIC (timing of platelet inhibition after acute coronary syndrome) randomized study. Eur Heart J. 2017;38:3070–3078. [DOI] [PubMed] [Google Scholar]
- 26. Steg PG, Lopez‐Sendon J, Lopez de Sa E, Goodman SG, Gore JM, Anderson FA Jr, Himbert D, Allegrone J, Van de Wer F. External validity of clinical trials in acute myocardial infarction. Arch Intern Med. 2007;167:68–73. [DOI] [PubMed] [Google Scholar]
- 27. Moscucci M, Fox KA, Cannon CP, Klein W, Lopez‐Sendon J, Montalescot G, White K, Goldberg RJ. Predictors of major bleeding in acute coronary syndromes: the Global Registry of Acute Coronary Events (GRACE). Eur Heart J. 2003;24:1815–1823. [DOI] [PubMed] [Google Scholar]
- 28. Kinnaird TD, Stabile E, Mintz GS, Lee CW, Canos DA, Gevorkian N, Pinnow EE, Kent KM, Pichard AD, Satler LF, et al. Incidence, predictors, and prognostic implications of bleeding and blood transfusion following percutaneous coronary interventions. Am J Cardiol. 2003;92:930–935. [DOI] [PubMed] [Google Scholar]
- 29. Eikelboom JW, Mehta SR, Anand SS, Xie C, Fox KA, Yusuf S. Adverse impact of bleeding on prognosis in patients with acute coronary syndromes. Circulation. 2006;114:774–782. [DOI] [PubMed] [Google Scholar]
- 30. Ludvigsson JF, Almqvist C, Bonamy AK, Ljung R, Michaelsson K, Neovius M, Stephansson O, Ye W. Registers of the Swedish total population and their use in medical research. Eur J Epidemiol. 2016;31:125–136. [DOI] [PubMed] [Google Scholar]
- 31. Stukel TA, Fisher ES, Wennberg DE, Alter DA, Gottlieb DJ, Vermeulen MJ. Analysis of observational studies in the presence of treatment selection bias: effects of invasive cardiac management on AMI survival using propensity score and instrumental variable methods. JAMA. 2007;297:278–285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Brookhart MA, Rassen JA, Schneeweiss S. Instrumental variable methods in comparative safety and effectiveness research. Pharmacoepidemiol Drug Saf. 2010;19:537–554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Earle CC, Tsai JS, Gelber RD, Weinstein MC, Neumann PJ, Weeks JC. Effectiveness of chemotherapy for advanced lung cancer in the elderly: instrumental variable and propensity analysis. J Clin Oncol. 2001;19:1064–1070. [DOI] [PubMed] [Google Scholar]
- 34. Bound J, Jaeger DA, Baker RM. Problems with instrumental variables estimation when the correlation between the instruments and the endogeneous explanatory variable is weak. J Am Stat Assoc. 1995;90:443–450. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.