10.1161/CIRCULATIONAHA.120.

050645

Comparison Between Ticagrelor and Clopidogrel in Elderly Patients with an

Acute Coronary Syndrome: Insights from the SWEDEHEART Registry

Running Title: Szummer et al.; Ticagrelor vs Clopidogrel Among Elderly MI Patients

Karolina Szummer, MD, PhD1; Maria E. Montez-Rath, PhD2; Joakim Alfredsson, MD, PhD3;
David Erlinge, MD, PhD4; Bertil Lindahl, MD, PhD5; Robin Hofmann, MD, PhD6;
Annica Ravn-Fischer, MD, PhD7; Per Svensson, MD, PhD6; Tomas Jernberg, MD, PhD8

1
Section of Cardiology, Department of Medicine, Huddinge, Karolinska Institutet, 141 86
Stockholm, Sweden; Department of Cardiology, Karolinska University Hospital, Hälsovägen 4,
141 86 Stockholm, Sweden; 2Division of Nephrology, Stanford University School of Medicine,
Palo Alto, CA; 3Department of Cardiology and Department of Medical and Health Sciences,
Linköping University, Linköping, Sweden; 4Department of Cardiology, Clinical Sciences, Lund
University, Lund, Sweden; 5Department of Medical Sciences, Cardiology, Uppsala University,
S-751 85, Uppsala, Sweden; and, Uppsala Clinical Research Center, Uppsala University,
Downloaded from http://ahajournals.org by on September 1, 2020

Uppsala, Sweden; 6Department of Clinical Science and Education, Division of Cardiology,

Karolinska Institutet, Södersjukhuset, Stockholm, Sweden; 7Department of Molecular and
Clinical Medicine, University of Gothenburg and Department of Cardiology, Sahlgrenska
University Hospital, Gothenburg, Sweden; 8Department of Clinical Sciences, Danderyd
University Hospital, Karolinska Institutet, 182 88, Stockholm, Sweden

Address for Correspondence:

Karolina Szummer,
Karolinska University Hospital Huddinge/Karolinska Institutet
141 86 Stockholm, Sweden
Email: Karolina.szummer@ki.se

This work was presented as an abstract at the European Society for Cardiology Congress, August
29 to September 1, 2020.

1
 10.1161/CIRCULATIONAHA.120.050645

Abstract

Background: The comparative efficacy and safety of ticagrelor vs. clopidogrel in older
myocardial infarction (MI) patients has received limited study.
Methods: We performed an observational analysis of all patients ≥80 years (n=14005) who were
discharged alive with aspirin combined with either clopidogrel (60.2%) or ticagrelor (39.8%)
after a MI between 2010 and 2017 registered in the national registry Swedish Web-System for
Enhancement and Development of Evidence-Based Care in Heart Disease Evaluated According
to Recommended Therapies (SWEDEHEART). Inverse probability treatment weighting was
used in Cox regression models to adjust for differences in demographics, in-hospital therapies,
and medications. The primary ischemic outcome (death, MI or stroke), and bleeding were
obtained from national registries at 1 year. A sensitivity analysis in <80-year-old patients was
performed.
Results: In patients ≥80 years, the incidence of the primary ischemic outcome (HR 0.97, 95% CI
0.88-1.06) was similar for ticagrelor- and clopidogrel-treated patients. Ticagrelor was associated
with a 17% and 48% higher risk of death (1.17 (1.03- 1.32)) and bleeding (1.48 (1.25- 1.76)), but
a lower risk of MI (0.80 (0.70- 0.92)) and stroke (0.72 (0.56-0.93)). In <80-year-old patients the
incidence of the primary ischemic outcome was 17% (0.83 (0.77-0.89)) lower with ticagrelor.
Ticagrelor was associated with 15% (0.85 (0.76-0.96)) lower risk of death, 32% higher risk of
bleeding (1.32 (1.18-1.47)), but lower risk of MI (0.82 (0.75-0.91)) and stroke (0.82 (0.69-0.98)).
Conclusions: Ticagrelor use among elderly MI patients was associated with higher risk of
bleeding and death compared with clopidogrel. A randomized study of ticagrelor vs clopidogrel
in the elderly is needed.
Downloaded from http://ahajournals.org by on September 1, 2020

Key Words: Antiplatelet therapy; elderly; myocardial infarction; risk-benefit; prognosis

Non-standard Abbreviations and Acronyms

CABG coronary artery by-pass grafting
CI Confidence interval
HR Hazard ratio
Non-STEMI Non-ST-segment elevation myocardial infarction
IPTW Inverse probability treatment weighting
PLATO trial Platelet Inhibition and Patient Outcomes trial
STEMI ST-segment elevation myocardial infarction
SWEDEHEART registry Swedish Web-System for Enhancement and Development of
Evidence-Based Care in Heart Disease Evaluated According
to Recommended Therapies registry

2
 10.1161/CIRCULATIONAHA.120.050645

Clinical Perspective

What is new?
• In this observational analysis from the nationwide myocardial infarction registry in
SWEDEN (SWEDEHEART), we found that ≥80-year-old patients may have a different
benefit-risk ratio when treated with ticagrelor compared with clopidogrel when discharged
after a myocardial infarction.
• Specifically, we found that patients ≥ 80 years-old had a 17% increased risk of death and
a 48% higher risk of bleeding, when discharged on ticagrelor versus clopidogrel after a
myocardial infarction, whereas there was a 20% lower risk of a new myocardial
infarction and 28% lower risk of stroke.

What are the clinical implications?

• Selecting a combination of aspirin with a P2Y12 inhibitor (either clopidogrel or
ticagrelor) at discharge after a myocardial infarction should be made cautiously among
elderly ≥ 80 years-old, who may have a different risk-benefit than younger individuals.
Downloaded from http://ahajournals.org by on September 1, 2020

• A randomized clinical trial in the elderly is needed to confirm these findings.

3
 10.1161/CIRCULATIONAHA.120.050645

Introduction

Since the publication in 2009 of the Platelet Inhibition and Patient Outcomes (PLATO) trial1, the

current European2, 3 and American guidelines4, 5 on the treatment of myocardial infarction

unanimously recommend a combination of aspirin with ticagrelor instead of clopidogrel, unless

contraindicated. In the PLATO trial, ticagrelor, compared with clopidogrel, used up to 12 months

after an episode of acute coronary syndrome, was shown to be unequivocally superior by

reducing the composite endpoint of vascular death, myocardial infarction or stroke. In addition,

death from any cause was lowered by 1.4% in patients treated with ticagrelor. Because of this

convincing evidence and unambiguous guidelines, there has been a widespread uptake in the use

of ticagrelor after myocardial infarction6.

In the PLATO trial, the reduction in ischemic outcomes came at an increased risk of non-

coronary artery by-pass grafting (CABG) related major bleeding (3.8% with clopidogrel vs.
Downloaded from http://ahajournals.org by on September 1, 2020

4.5% with ticagrelor). One of the most important factors associated with higher bleeding risk is

older age7-10. In PLATO, median patient age was 62 years, which more than 10 years younger

than patients treated in everyday clinical practice11, 12. In real life, elderly patients are common,

have more comorbidities than younger patients, and often have more risk factors for bleeding,

such as reduced kidney function, anemia and inflammation. Hence, the risk of bleeding with a

more potent antiplatelet therapy may be much higher in a non-selected MI population. In fact,

major bleeding events have increased in recent years, occurring in 4.8% of patients within the

first year after discharge, whereas mortality has not declined (7). One possible reason for this

finding could be that the trade-off between reduction of ischemic events and increased risk of

bleeding is not beneficial in elderly patients.

4
 10.1161/CIRCULATIONAHA.120.050645

The aim of this study was to compare clinical outcomes with ticagrelor versus clopidogrel among

≥80-year-old myocardial infarction patients.

Methods

The nationwide cohort from the SWEDEHEART registry

Because of the sensitive nature of the data collected for this study as part of the Swedish Web-

System for Enhancement and Development of Evidence-Based Care in Heart Disease Evaluated

According to Recommended Therapies (SWEDEHEART) registry, data cannot be made

available to other researchers.

All patients were recruited from the SWEDEHEART registry13 (Figure I in the

Supplement). SWEDEHEART is a nationwide registry, that includes all consecutive patients

admitted to a coronary care unit with symptoms suggestive of an acute coronary syndrome.
Downloaded from http://ahajournals.org by on September 1, 2020

Currently, all Swedish hospitals (n=72) that provide care for acute cardiac diseases participate.

Over 100 variables regarding baseline characteristics, medication on admission, in-hospital

therapies, complications, and discharge medication are collected

(http://www.ucr.uu.se/swedeheart/). Monitors evaluate the correctness of data entered in the

registry with the medical records yearly (agreement is around 96%). Patients do not provide

written consent but are informed about their participation and can opt-out.

Our aim was to emulate a pragmatic target trial intended to compare the effect of

ticagrelor and clopidogrel as in the PLATO trial but in an elderly population14. Thus, this study

included all consecutive patients ≥80 years of age and alive at discharge, who had a myocardial

infarction registered for the first time in SWEDEHEART between January 1st, 2010, and

December 31st, 2017 (Figure I in the Supplement). The time period encompassed the

5
 10.1161/CIRCULATIONAHA.120.050645

introduction of ticagrelor in the latter half of 2011. Patients were excluded if they were treated

with clopidogrel or ticagrelor prior to admission for the index myocardial infarction, had a

requirement for direct oral anticoagulant or vitamin K antagonist, a history of prior bleeding,

serious in-hospital bleeding (requiring blood transfusion or operation), or a hemoglobin <100 g/L

(<10g/dL). Patients with a diagnosis rendering them less likely to be prescribed a more potent

antiplatelet agent were also excluded (e.g. dementia, dialysis) (Figure I in the Supplement). Data

on baseline characteristics were enriched with information from the National Patient Registry

with diagnoses of all hospital admissions in Sweden since 1987.

The study protocol was approved by the regional ethics committee in Stockholm,

Sweden. As data in the SWEDEHEART registry is collected as part of health care quality

improvement, informed consent from participants is not required by Swedish law.

Exposure
Downloaded from http://ahajournals.org by on September 1, 2020

Patients were considered exposed to either ticagrelor or clopidogrel if there was a dispensation

recorded in the Swedish drug dispensation registry within two weeks of hospital discharge. If

this data was unavailable (n=553, 3.9%) but the SWEDEHEART registry had an entry with the

discharge medication, the latter information was used.

Outcome definitions

The main outcomes were analyzed both individually and combined, and consisted of death, re-

admission for myocardial infarction, stroke, and bleeding, all occurring within 365 days of the

discharge-date. The endpoints were obtained from ICD codes from the national patient registry

and vital status from the population registry (Table I in the Supplement). A new myocardial

infarction in the first four weeks following discharge was defined by both a new registration in

the SWEDEHEART registry and a new diagnosis in the patient registry. An entry in the

6
 10.1161/CIRCULATIONAHA.120.050645

SWEDEHEART registry was required as the ICD-10 coding does not differentiate between a

new myocardial infarction or a hospitalization for a different reason within 4 weeks of hospital

discharge. After four weeks from the discharge date, an entry in the patient registry was

sufficient. Last date of follow-up was December 31st, 2017.

Statistical analyses

Baseline characteristics are described as medians (interquartile ranges), means (standard

deviation, SD) for continuous variables or as counts and percentages for categorical variables

and compared using standardized differences15.

The main analysis, evaluating the association between treatment assignment and

endpoints, was assessed in weighted Cox regression models. As the treatment was not randomly

assigned and to reduce selection bias, inverse probability treatment weighting (IPTW) was used.

The weights were computed from logistic regression models predicting treatment assignment
Downloaded from http://ahajournals.org by on September 1, 2020

that included all 35 covariates in Table 1. To obtain a similar number in the two treatment arms

after IPTW, the weights were stabilized. Stabilization (multiplying the weight by the probability

of being exposed for those exposed and the probability of being unexposed for those unexposed),

results in a pseudo-population with similar percentage of patients exposed/un-exposed in each

level of the covariates as the overall percentage in the study population. Stabilization does not

affect the point estimates but reduces the variability of the weights. This can also lead to an

exaggerated increase in the weights, but no extreme values were found (10 >weights >0.1). The

final Cox regression model included all 35 covariates from Table 1 in order to obtain a double

robust estimate16. The proportional hazards assumption for each Cox regression model was

examined with Schoenfield residuals and met. Hazard ratios (HRs) are reported with 95%

confidence intervals (CIs).

7
 10.1161/CIRCULATIONAHA.120.050645

The only confounder found to have missing data was creatinine, which was unavailable

in 606 patients (4.3%) (Table II in the Supplement). We assumed that creatinine was missing at

random (i.e., the missingness was related to observed characteristics) and used multiple

imputation by chained equations (mice) to generate 5 complete data sets. The above analysis was

then performed in each data set separately and coefficients were averaged and results

exponentiated to create the final estimate. The imputation models included all the 34 baseline

covariates in the Table 1, as well as the outcome (myocardial infarction, stroke, death, re-

admission for bleeding) and the time to the events. To account for the variability when using

multiple imputation as well as the fact that the probability of treatment was estimated and used to

compute weights, bootstrapping was used to estimate the standard error. A thousand bootstrap

samples with replacement from the original data (with missing values for creatinine) were used.

For each of the bootstrap samples, the above analysis was performed: 1) impute five datasets; 2)
Downloaded from http://ahajournals.org by on September 1, 2020

in each, estimate the probability of treatment, calculate IPTW weights, run weighted Cox

regression model; 3) average the five coefficients to obtain an estimate for the bootstrap sample.

The bootstrap standard error is then computed as the standard deviation of the 1000 estimates17.

Several sensitivity analyses were performed. First, for confirmation that our results in the

observational study were consistent with the results in the PLATO trial that was performed on a

younger population, the treatment effect of ticagrelor versus clopidogrel was assessed in patients

younger than 80 years with the same characteristics as the original cohort. Given that our

complete and imputed analyses provided similar results, for computational ease the sensitivity

analysis was performed on complete cases with the same adjustments as described above.

8
 10.1161/CIRCULATIONAHA.120.050645

Second, since the patients were elderly, the competing risk of death was accounted for by

Fine and Gray models. Third, IPTW adjusted Cox regression analysis was performed on

complete cases (Table II in the Supplement).

All analyses were performed with STATA (Stata/MP 15.1, StataCorp LLC).

Results

Study population – 80 years or older

In total, there were 14,005 patients who were 80 years or older and who survived until discharge

for a myocardial infarction (Figure I in the Supplement), of which 4,428 (31.6%) had a ST-

segment elevation myocardial infarction (STEMI) and the remaining had a non-ST-segment

elevation myocardial infarction (NSTEMI). The majority were prescribed clopidogrel (n=8434,

60.2%) and fewer ticagrelor (n= 5571, 39.8%). Between 2010 to 2017, ticagrelor use increased
Downloaded from http://ahajournals.org by on September 1, 2020

from 0% to 72.5% (Figure 1). Patients prescribed ticagrelor were younger and had fewer co-

morbidities at presentation (Table 1), but after IPTW adjustment, the baseline characteristics

were balanced (Table 1; Figure II and Figure III in the Supplement).

Study outcome in patients 80 years or older

In unadjusted analysis, patients treated with ticagrelor had significantly fewer combined

ischemic events or combined ischemic/bleeding events (Figure IV in Supplements, Table 2). The

event rate for myocardial infarction, stroke and death was lower among those treated with

ticagrelor, whereas the rate of re-admission for bleeding was higher.

In the main analyses, consisting of IPTW Cox regression with imputed missing values, there

was no significant reduction in neither the combined ischemic nor in the combined

ischemic/bleeding endpoints with ticagrelor compared with clopidogrel therapy (Figure 2, Table

2). Ticagrelor-treated patients had 20% and 28% lower hazard of having a new myocardial

9
 10.1161/CIRCULATIONAHA.120.050645

infarction (HR 0.80; 95% CI 0.70- 0.92) or stroke (HR 0.72, 95% CI 0.56-0.93) (Table 2),

respectively, compared with clopidogrel-treated patients within one year. However, ticagrelor-

treated patients had a 17% higher hazard of dying (HR 1.17, 95% CI 1.03- 1.32) and a 48%

higher hazard of being re-admitted with a bleeding event (HR 1.48, 95% CI 1.25- 1.76) than

those treated with clopidogrel.

Competing risk analysis, accounting for the risk of death, showed similar associations to the

endpoints. In complete case analyses, the results from IPTW adjusted Cox regression analysis

(Table III in the Supplement) were consistent with the main analyses.

Comparison of treatment effects among patients younger than 80 years

In the cohort of patients younger than 80 years old, 38.2% (n=22415) of the patients were treated

with clopidogrel and 61.8% (n=36256) with ticagrelor (Table IV and Figure V).

Results from the IPTW analysis on complete cases showed 18% lower hazard of
Downloaded from http://ahajournals.org by on September 1, 2020

myocardial infarction (HR 0.82, 95% CI 0.75-0.91), 18% lower hazard of stroke (HR 0.82 95%

CI 0.60-0.98), and 15% lower hazard of death (HR 0.85, 96% CI 0.76-0.96) (Supplemental Table

IV, Supplemental Table V, Supplemental Figure IV), but 32% higher hazard of bleeding events

(HR 1.32, 95% CI 1.18-1.47). Overall, there was a 17% lower hazard of a combination of

ischemic events (HR 0.83; 95%, 0.77-0.89), but when ischemic and bleeding events were

combined, there was no significant difference between the two treatments (HR 0.95, 95% CI

0.89-1.01).

Discussion

In this study we found that elderly myocardial infarction patients ≥80 years who were treated

with ticagrelor did not have a lower risk for the combined outcome of death, readmission

10
 10.1161/CIRCULATIONAHA.120.050645

because of myocardial infarction or stroke. Although the risk of stroke and myocardial infarction

alone was lower, the risk of bleeding and death was higher compared with patients treated with

clopidogrel. These findings indicate that the current guideline recommendations should be

applied with caution in elderly patients and that we may not expect the same efficacy and safety

as evident in clinical trials. Yet, over the 7 years in the study period, there has been a significant

uptake in ticagrelor use so that nearly three quarters of all elderly patients at the end of 2017

were prescribed ticagrelor and not clopidogrel at discharge after a myocardial infarction. Thus,

there is a need for a randomized study examining the effects of more potent antiplatelet therapy

in the elderly population.

The efficacy and safety of ticagrelor and prasugrel among elderly patients has been the

focus of other studies. This topic is important, as approximately 35% of all myocardial infarction

patients are ≥75 years-old18, and elderly patients are for many reasons at higher risk of bleeding7-
Downloaded from http://ahajournals.org by on September 1, 2020

10, 19
, which could offset the benefits observed in younger patients. Two smaller studies, one

randomized and one non-randomized, have compared potent P2Y12 inhibitors vs clopidogrel in

the elderly. The POPular AGE trial20 from the Netherlands randomized 1003 patients with Non-

ST-elevation acute coronary syndrome patients above the age of 70 years to clopidogrel or a

choice of either ticagrelor/prasugrel for 12 months. In the ticagrelor/prasugrel arm most patients

received ticagrelor (95%), and the risk of bleeding was 29% lower with clopidogrel than with

ticagrelor/prasugrel (HR 0.71, 95% CI 0.54-0.94, p=0.03), while clopidogrel treatment was non-

inferior with regard to the combined clinical endpoint of ischemic and bleeding events. It is not a

surprise that ticagrelor and prasugrel are associated with higher risk of bleeding, as this was

evident in the comparisons of clopidogrel with ticagrelor in the PLATO trial and prasugrel in

TRITON-TIMI-38. The results are expected, as both ticagrelor and prasugrel have a much more

11
 10.1161/CIRCULATIONAHA.120.050645

rapid and consistent onset of action on the platelets’ adenosine diphosphate receptor P2Y12

inhibiting platelet aggregation immediately. In a non-randomized registry study from Germany21,

the effect of ticagrelor versus clopidogrel was assessed among 1087 patients with ST-elevation

myocardial infarction who were ≥75 years of age (54% of whom were≥80 years old). In their

clinical practice, half of the patients were receiving ticagrelor and the other half clopidogrel.

Propensity-score matched analyses showed that compared with clopidogrel ticagrelor was

associated with 31% lower risk of the combined ischemic endpoints (HR 0.69, 95% CI 0.49-

0.97) compared with clopidogrel, without differences in the incidence of bleeding or death.

However, the incidence of out-of-hospital bleeding was exceptionally low (0.4% in the

clopidogrel vs. 1.8% in the ticagrelor group), which probably reflects that not all bleeding events

were captured. This is much lower than the 4.8% one-year risk of out-of-hospital bleeding

reported from other real-world data22.

Downloaded from http://ahajournals.org by on September 1, 2020

The present study is by far the largest study comparing clinical outcomes between

ticagrelor and clopidogrel treated elderly patients. The number of patients collected over several

years from clinical practice allowed for analysis of both elderly and younger patients.

Consequently, it was possible to estimate and confirm the benefit of ticagrelor among younger

patients, in whom there was 15% lower risk of death (Supplemental Table V), 18% lower risk of

myocardial infarction and18% lower risk of stroke, but 32% higher risk of bleeding. These

findings among < 80 year-old patients are hence consistent with and support the results from the

PLATO trial1, in which the median age was 62.0 years. This also suggests that our methodology

to examine outcomes in these populations is adequate.

12
 10.1161/CIRCULATIONAHA.120.050645

We attempted to design a study that resembles a clinical trial as much as possible, by

selecting individuals at low risk of bleeding. Except for age, which is a risk factor for bleeding,

patients were excluded if they had anemia, prior bleeding or new in-hospital bleeding events.

Also, all patients were drug naïve with regard to ticagrelor or clopidogrel on the index hospital

admission. Yet, despite these stringent selection criteria, the risk of bleeding was higher among

ticagrelor-treated patients.

Our results are somewhat in contrast from the PLATO trial, in which there were no signs

of an age-by-treatment interaction23. It is important to note that patients included in clinical trials

are very different from those in real world clinical practice24. But not all elderly are the same and

several other factors influence the risk of bleeding25, and high age is associated with high platelet

reactivity and increased risk of thrombotic complications26, 27. Before we have more evidence

from randomized trials, a more individualized approach may be appropriate in the elderly28.
Downloaded from http://ahajournals.org by on September 1, 2020

The observational, non-randomized design is the main limitation of our study. We could

only adjust for known factors by IPTW in the models, hence residual confounding remains

possible, although it is reassuring that the results obtained among younger individuals were

comparable with those from a randomized clinical trial. For instance, there was a substantial

difference in the number of patients treated invasively at baseline with clopidogrel versus

ticagrelor, but after adjustment with IPTW the two groups were balanced on this covariate. The

study database did not include any falsification endpoint, i.e. an endpoint that is known to be

unrelated to treatment under study, which could have supported that the analyses were

unbiased. In addition, all data were analyzed as intention-to-treat, and early termination of a drug

was not accounted for. It is possible that some patients crossed over from one drug to another,

which could potentially have weakened the observed associations. However, we included a large

number of patients after the publication of the PLATO-trial, which decrease the likelihood of a

13
 10.1161/CIRCULATIONAHA.120.050645

type 2 error. A potential confounder when studies are performed is that clinical practice change,

and other new therapies or interventions are introduced that could further improve outcomes over

the years. This could bias the results towards the treatment which was later introduced. However,

we adjusted for all known therapies in the final models. Another limitation is that the cause of

death is unknown.

In conclusion, this observational study suggests that ticagrelor should be used with

caution among elderly patients 80 years and older, as it may be associated with increased risk of

death and bleeding. There is a need for an adequately powered randomized study examining the

effects of potent anti-platelets in the elderly population.

Sources of Funding

KS was supported by the Stockholm County Council (clinical research appointment), by the
Downloaded from http://ahajournals.org by on September 1, 2020

Swedish Medical Association (Svenska Läkarsällskapet) and the Swedish Heart Association

(Svenska Hjärtförbundet). RH was supported by the Stockholm County Council (clinical

postdoctoral appointment) and by the Swedish Heart-Lung foundation.

SWEDEHEART is supported by the Swedish Heart- and Lung-foundation.

Conflict of Interest Disclosures

KS, MEMR, DE, BL, RH, ARF, PS have no disclosures.

JA has received research grants from AstraZeneca. JA has received lecture fees from

AstraZeneca, Sanofi-Aventis, BMS, Eli-Lilly, MSD and Boeringer-Ingelheim. JA has been on an

advisory board for AstraZeneca. TJ has received institutional grants from MSD and Novartis.

14
 10.1161/CIRCULATIONAHA.120.050645

The employer (Karolinska Institutet) of TJ has received payment from AstraZeneca, Bayer,

Sanofi, Novartis and Merck for TJs lecturing and consulting.

Supplemental Materials

Supplemental Figures I-V

Supplemental Tables I-V

References

1. 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. New England Journal of Medicine. 2009;361:1045-1057.doi:
10.1056/NEJMoa0904327
2. Ibanez B, James S, Agewall S, Antunes MJ, Bucciarelli-Ducci C, Bueno H, Caforio ALP,
Crea F, Goudevenos JA, Halvorsen S, et al. 2017 ESC Guidelines for the management of acute
myocardial infarction in patients presenting with ST-segment elevation: The Task Force for the
management of acute myocardial infarction in patients presenting with ST-segment elevation of
Downloaded from http://ahajournals.org by on September 1, 2020

the European Society of Cardiology (ESC). Eur Heart J. 2018;39:119-177.doi:

10.1093/eurheartj/ehx393
3. Roffi M, Patrono C, Collet JP, Mueller C, Valgimigli M, Andreotti F, Bax JJ, Borger
MA, Brotons C, Chew DP, et al. 2015 ESC Guidelines for the management of acute coronary
syndromes in patients presenting without persistent ST-segment elevation: Task Force for the
Management of Acute Coronary Syndromes in Patients Presenting without Persistent ST-
Segment Elevation of the European Society of Cardiology (ESC). Eur Heart J. 2016;37:267-
315.doi: 10.1093/eurheartj/ehv320
ehv320 [pii]
4. Amsterdam EA, Wenger NK, Brindis RG, Casey DE, Jr., Ganiats TG, Holmes DR, Jr.,
Jaffe AS, Jneid H, Kelly RF, Kontos MC, et al. 2014 AHA/ACC Guideline for the Management
of Patients with Non-ST-Elevation Acute Coronary Syndromes: a report of the American
College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am
Coll Cardiol. 2014;64:e139-e228.doi: 10.1016/j.jacc.2014.09.017
5. O'Gara PT, Kushner FG, Ascheim DD, Casey DE, Jr., Chung MK, de Lemos JA, Ettinger
SM, Fang JC, Fesmire FM, Franklin BA, et al. 2013 ACCF/AHA guideline for the management
of ST-elevation myocardial infarction: a report of the American College of Cardiology
Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol.
2013;61:e78-140.doi: 10.1016/j.jacc.2012.11.019
6. Sahlen A, Varenhorst C, Lagerqvist B, Renlund H, Omerovic E, Erlinge D, Wallentin L,
James SK and Jernberg T. Outcomes in patients treated with ticagrelor or clopidogrel after acute

15
 10.1161/CIRCULATIONAHA.120.050645

myocardial infarction: experiences from SWEDEHEART registry. Eur Heart J. 2016;37:3335-

3342.doi: 10.1093/eurheartj/ehw284
7. Simonsson M, Winell H, Olsson H, Szummer K, Alfredsson J, Hall M, Dondo TB, Gale
CP and Jernberg T. Development and Validation of a Novel Risk Score for In-Hospital Major
Bleeding in Acute Myocardial Infarction:-The SWEDEHEART Score. J Am Heart Assoc.
2019;8:e012157.doi: 10.1161/jaha.119.012157
8. Mathews R, Peterson ED, Chen AY, Wang TY, Chin CT, Fonarow GC, Cannon CP,
Rumsfeld JS, Roe MT and Alexander KP. In-hospital major bleeding during ST-elevation and
non-ST-elevation myocardial infarction care: derivation and validation of a model from the
ACTION Registry(R)-GWTG. Am J Cardiol. 2011;107:1136-1143.doi:
10.1016/j.amjcard.2010.12.009
9. Mehran R, Pocock SJ, Nikolsky E, Clayton T, Dangas GD, Kirtane AJ, Parise H, Fahy
M, Manoukian SV, Feit F, et al. A risk score to predict bleeding in patients with acute coronary
syndromes. J Am Coll Cardiol. 2010;55:2556-2566.doi: 10.1016/j.jacc.2009.09.076
10. Subherwal S, Bach RG, Chen AY, Gage BF, Rao SV, Newby LK, Wang TY, Gibler WB,
Ohman EM, Roe MT, et al. Baseline risk of major bleeding in non-ST-segment-elevation
myocardial infarction: the CRUSADE (Can Rapid risk stratification of Unstable angina patients
Suppress ADverse outcomes with Early implementation of the ACC/AHA Guidelines) Bleeding
Score. Circulation. 2009;119:1873-1882.doi: 10.1161/circulationaha.108.828541
11. Szummer K, Wallentin L, Lindhagen L, Alfredsson J, Erlinge D, Held C, James S,
Kellerth T, Lindahl B, Ravn-Fischer A, et al. Improved outcomes in patients with ST-elevation
myocardial infarction during the last 20 years are related to implementation of evidence-based
treatments: experiences from the SWEDEHEART registry 1995-2014. Eur Heart J.
2017;38:3056-3065.doi: 10.1093/eurheartj/ehx515
Downloaded from http://ahajournals.org by on September 1, 2020

12. Szummer K, Wallentin L, Lindhagen L, Alfredsson J, Erlinge D, Held C, James S,

Kellerth T, Lindahl B, Ravn-Fischer A, et al. Relations between implementation of new
treatments and improved outcomes in patients with non-ST-elevation myocardial infarction
during the last 20 years: experiences from SWEDEHEART registry 1995 to 2014. Eur Heart J.
2018;39:3766-3776.doi: 10.1093/eurheartj/ehy554
13. Jernberg T, Attebring MF, Hambraeus K, Ivert T, James S, Jeppsson A, Lagerqvist B,
Lindahl B, Stenestrand U and Wallentin L. The Swedish Web-system for enhancement and
development of evidence-based care in heart disease evaluated according to recommended
therapies (SWEDEHEART). Heart. 2010;96:1617-1621.doi: 10.1136/hrt.2010.198804
14. Hernan MA and Robins JM. Using Big Data to Emulate a Target Trial When a
Randomized Trial Is Not Available. Am J Epidemiol. 2016;183:758-764.doi:
10.1093/aje/kwv254
15. Austin PC. Using the Standardized Difference to Compare the Prevalence of a Binary
Variable Between Two Groups in Observational Research. Communications in Statistics -
Simulation and Computation. 2009;38:1228-1234.doi: 10.1080/03610910902859574
16. Bang H and Robins JM. Doubly robust estimation in missing data and causal inference
models. Biometrics. 2005;61:962-973.doi: 10.1111/j.1541-0420.2005.00377.x
17. Albee Y. Ling MEM-R, Maya B. Mathur, Kris Kapphahn, Manisha Desai. How to apply
multiple imputation in propensity score matching with partially observed confounders: a
simulation study and practical recommendations. arXiv:190407408. 2019.
18. Hanssen M, Cottin Y, Khalife K, Hammer L, Goldstein P, Puymirat E, Mulak G, Drouet
E, Pace B, Schultz E, et al. French Registry on Acute ST-elevation and non ST-elevation

16
 10.1161/CIRCULATIONAHA.120.050645

Myocardial Infarction 2010. FAST-MI 2010. Heart. 2012;98:699-705.doi: 10.1136/heartjnl-

2012-301700
19. Capodanno D and Angiolillo DJ. Antithrombotic therapy in the elderly. J Am Coll
Cardiol. 2010;56:1683-1692.doi: 10.1016/j.jacc.2010.04.063
20. Gimbel M, Qaderdan K, Willemsen L, Hermanides R, Bergmeijer T, de Vrey E,
Heestermans T, Tjon Joe Gin M, Waalewijn R, Hofma S, et al. Clopidogrel versus ticagrelor or
prasugrel in patients aged 70 years or older with non-ST-elevation acute coronary syndrome
(POPular AGE): the randomised, open-label, non-inferiority trial. Lancet. 2020;395:1374-
1381.doi: 10.1016/s0140-6736(20)30325-1
21. Schmucker J, Fach A, Mata Marin LA, Retzlaff T, Osteresch R, Kollhorst B, Hambrecht
R, Pohlabeln H and Wienbergen H. Efficacy and Safety of Ticagrelor in Comparison to
Clopidogrel in Elderly Patients With ST-Segment-Elevation Myocardial Infarctions. J Am Heart
Assoc. 2019;8:e012530.doi: 10.1161/jaha.119.012530
22. Simonsson M, Wallentin L, Alfredsson J, Erlinge D, Hellstrom Angerud K, Hofmann R,
Kellerth T, Lindhagen L, Ravn-Fischer A, Szummer K, et al. Temporal trends in bleeding events
in acute myocardial infarction: insights from the SWEDEHEART registry. Eur Heart J.
2019.doi: 10.1093/eurheartj/ehz593
23. Husted S, James S, Becker RC, Horrow J, Katus H, Storey RF, Cannon CP, Heras M,
Lopes RD, Morais J, et al. Ticagrelor versus clopidogrel in elderly patients with acute coronary
syndromes: a substudy from the prospective randomized PLATelet inhibition and patient
Outcomes (PLATO) trial. Circ Cardiovasc Qual Outcomes. 2012;5:680-688.doi:
10.1161/circoutcomes.111.964395
24. Wallentin L. Different perspectives on outcomes in patients with non-ST-elevation
myocardial infarction when observed in clinical trials and in real life. Eur Heart J. 2018;39:3821-
Downloaded from http://ahajournals.org by on September 1, 2020

3824.doi: 10.1093/eurheartj/ehy610
25. Urban P, Mehran R, Colleran R, Angiolillo DJ, Byrne RA, Capodanno D, Cuisset T,
Cutlip D, Eerdmans P, Eikelboom J, et al. Defining high bleeding risk in patients undergoing
percutaneous coronary intervention: a consensus document from the Academic Research
Consortium for High Bleeding Risk. Eur Heart J. 2019;40:2632-2653.doi:
10.1093/eurheartj/ehz372
26. Angiolillo DJ, Capodanno D, Danchin N, Simon T, Bergmeijer TO, Ten Berg JM,
Sibbing D and Price MJ. Derivation, Validation, and Prognostic Utility of a Prediction Rule for
Nonresponse to Clopidogrel: The ABCD-GENE Score. JACC Cardiovasc Interv. 2020;13:606-
617.doi: 10.1016/j.jcin.2020.01.226
27. Sibbing D, Aradi D, Alexopoulos D, Ten Berg J, Bhatt DL, Bonello L, Collet JP, Cuisset
T, Franchi F, Gross L, et al. Updated Expert Consensus Statement on Platelet Function and
Genetic Testing for Guiding P2Y(12) Receptor Inhibitor Treatment in Percutaneous
Coronary Intervention. JACC Cardiovasc Interv. 2019;12:1521-1537.doi:
10.1016/j.jcin.2019.03.034
28. Capranzano P and Angiolillo DJ. Tailoring P2Y(12) Inhibiting Therapy in Elderly
Patients With Myocardial Infarction Undergoing Primary Percutaneous Coronary Intervention. J
Am Heart Assoc. 2019;8:e014000.doi: 10.1161/jaha.119.014000

17
 10.1161/CIRCULATIONAHA.120.050645

Table 1. Baseline characteristics before and after inverse probability treatment weighting (IPTW) from one of the five imputed datasets.
See Figure I and Figure II in the Supplement for balance of the covariates for complete (observed data) and all five imputations.

Before IPTW** After IPTW**

Clopidogrel Ticagrelor Standardized Clopidogrel Ticagrelor Standardized
difference*** difference***
8434 (60.2) 5571 (39.8) 8421 (60.0)**** 5607 (40.0)
Demographics
Age mean ± SD; median (IQR) 85.4 (±4.1) 84.0 (±3.5) -37.6 84.9±3.9 85.0±3.9 1.8
85 (82-88) 83 (81-86) 84 (82-87) 85 (82-88)
Male 51.8 45.9 -11.8 49.4 49.2 -0.3
Current smoker 5.4 5.9 2.2 5.6 5.6 -0.1
Prior diseases
Myocardial infarction 29.4 21.5 -18.2 26.2 26.0 -0.6
Heart failure 17.7 10.4 -21.2 14.7 14.5 -0.6
Coronary artery bypass grafting 9.9 7.7 -7.8 9.1 9.0 -0.2
Percutaneous Coronary Intervention 13.2 13.4 0.7 13.3 13.3 -0.0
Diabetes 22.2 22.7 1.3 22.4 22.2 -0.2
Stroke 10.8 7.1 -12.9 9.4 9.5 0.6
Kidney disease 4.7 3.4 -6.4 4.1 4.0 -0.9
Downloaded from http://ahajournals.org by on September 1, 2020

Chronic obstrucktive pulmonary disease 8.4 7.5 -3.3 8.0 7.8 -0.7
Peripheral arterial disease 6.4 5.6 -3.4 6.2 6.2 -0.0
Hypertension 69.2 68.3 -1.9 68.8 69.1 0.6
Cancer within last 3 years 4.1 3.5 -2.9 3.9 3.8 -0.5
Type of myocardial infarction and signs on admission
STEMI (vs NSTEMI) 27.0 38.6 24.7 31.4 30.9 -1.0
Killip>1 3.6 1.9 -10.3 2.9 2.8 -0.3
Medication on admission
Aspirin 53.6 43.6 -20.1 49.7 49.6 -0.3
Beta-blocker 42.7 34.7 -16.4 39.5 39.8 0.6
Calcium channel antagonist 24.3 24.8 1.2 24.5 24.4 -0.0
Digitalis 1.0 0.4 -6.9 0.8 0.7 -0.4
Angiotensin converting enzyme-inhibitors/Angiotensin II
receptor blocker 38.7 38.4 -0.6 38.6 38.3 -0.6
Diuretics 33.5 25.0 -18.8 30.3 30.9 1.3
Statins 26.3 25.1 -2.8 25.8 25.3 -1.2
In-hospital treatments/interventions
Inotropes 1.1 1.9 6.4 1.4 1.4 0.1

18
 10.1161/CIRCULATIONAHA.120.050645

Intravenous diuretics 23.6 19.5 -10.0 21.9 22-0 0.0

In-hospital percutaneous coronary intervention 44.9 78.7 74.0 58.3 57.9 -0.8
In-hospital coronary angiography 55.2 86.1 72.1 67.4 67.0 -0.9
Medication on discharge
Beta-blocker 86.1 87.1 2.9 86.4 86.8 1.1
Calcium channel antagonist 21.3 20.8 -1.1 21.1 21.2 0.2
Digitalis 1.0 0.2 -10.1 0.7 0.5 -2.4
Angiotensin converting enzyme-inhibitors/Angiotensin II
receptor blocker 71.2 80.7 22.4 74.8 74.1 -1.6
Medication for diabetes 15.9 16.5 1.4 16.1 16.0 -0.2
Statins 77.1 90.5 37.1 82.3 81.5 -2.0
Hospital duration
Days in hospital 5.8±4.2 5.2±3.8 -12.8 5.6 ±4.0 5.7±4.4 2.9
5 (3-7) 4 (3-6) 4 (3-7) 4 (3.7)
Laboratory measurement
eGFR 54±26 59±23 21.6 56±25 56±25 -1.1
58 (45-73) 63 (49-77) 60 (46-74) 60 (46-75)
• Reporting: N (%), mean (±standard deviation); median (p25-p75) ** IPTW: inverse probability treatment weighting; *** A standardized difference < 10 was considered
good balance. The standardized difference for eGFR, which had the most missing, varied among the five imputed from -0.9 to -1.2. See also Figure III in the Supplement.
**** The N presented in the Table 1 after IPTW is the effective N after application of weights, but still using the entire sample.
Downloaded from http://ahajournals.org by on September 1, 2020

19
 10.1161/CIRCULATIONAHA.120.050645

Table 2. Incidence rates and inverse probability treatment weights (IPTW) adjusted hazard ratio for 1-year outcomes in ≥80 year-old patients with an
acute myocardial infarction treated with either clopidogrel or ticagrelor from the SWEDEHEART registry between 2010-2017.

Clopidogrel Ticagrelor IPTW adjusted IPTW adjusted

Cause-specific Competing risk
Hazard ratio Hazard ratio
(95% CI)* (95% CI)*
Number Mean±SD/median Incidence rate Number Mean±SD/median Incidence rate
events follow-up time (per 100 person-years) events follow-up time (per 100 person-
years)

Stroke, MI, death 2,230 296±121 /365 32.8 844 297±117 /365 18.7 0.97 (0.88-1.06) ---

Stroke, MI, death, 2,427 290±125 /365 37.4 1,058 288±123 /365 24.2 1.03 (0.94-1.12) ---
readmission for bleeding

MI 1,048 300±119 /365 13.9 360 300±115 /365 7.5 0.80 (0.70-0.92) 0.78 (0.68-0.90)

Stroke 277 318±104 /365 3.44 155 308±108 /365 2.32 0.72 (0.56-0.93) 0.70 (0.54-0.91)
Downloaded from http://ahajournals.org by on September 1, 2020

Death 1344 322 ±99 /365 18.1 511 311±105 /365 10.8 1.17 (1.03-1.32) ---

Readmission for bleeding 388 314±107 /365 4.86 333 300±114 /365 6.90 1.48 (1.25-1.76) 1.45 (1.23-1.72)
* Results obtained after five multiple imputations (see methods for details). Confidence interval (CI) was obtained from 1000 boot-strap sampling with replacement. Inverse-
probability treatment weights (IPTW) was used to account for differences in baseline characteristics.
SD: Standard deviation

20
 10.1161/CIRCULATIONAHA.120.050645

Figure Legends

Figure 1. Temporal use of clopidogrel and ticagrelor in elderly patients

Use of clopidogrel and ticagrelor in patients ≥80 years old between 2010 and 2017.

Figure 2. Events in the first year following discharge

Cumulative incidence for one imputed dataset after inverse probability of treatment weights

(without accounting for the competing risk of death).

A) Death, myocardial infarction or stroke

B) Death, myocardial infarction, stroke, or readmission for bleeding

C) Myocardial infarction

D) Stroke
Downloaded from http://ahajournals.org by on September 1, 2020

E) Death

F) Readmission for bleeding

21
 ≥80 years old
100,0%

90,0%

80,0%

70,0%

60,0%

50,0%
Downloaded from http://ahajournals.org by on September 1, 2020

40,0%

30,0%

20,0%

10,0%

0,0%
2010 2011 2012 2013 2014 2015 2016 2017

Ticagrelor Clopidogrel
 F
C
B

E D
A
Downloaded from http://ahajournals.org by on September 1, 2020