Abbreviated Versus Standard Dual Antiplatelet Therapy Times After Percutaneous Coronary Intervention in Patients With High Bleeding Risk With Acute Coronary Syndrome: Insights From the SWEDEHEART Registry

Abstract

Background

Dual antiplatelet therapy (DAPT) reduces ischemic events but increases bleeding risk, especially in patients with high bleeding risk (HBR). This study aimed to compare outcomes of abbreviated versus standard DAPT strategies in patients with HBR with acute coronary syndrome undergoing percutaneous coronary intervention.

Methods and Results

Patients from the SWEDEHEART (Swedish Web‐system for Enhancement and Development of Evidence‐Based Bare in Heart Disease Evaluated According to Recommended Therapies) registry with at least 1 HBR criterion who underwent percutaneous coronary intervention for acute coronary syndrome were identified and included. Patients were divided into 2 groups based on their planned DAPT time at discharge: 12‐month DAPT or an abbreviated DAPT strategy and matched according to their prescribed P2Y12 inhibitor at discharge. The primary outcome assessed was time to net adverse clinical events at 1 year, which encompassed cardiac death, myocardial infarction, ischemic stroke, or clinically significant bleeding. Time to major adverse cardiovascular events and the individual components of net adverse clinical events were considered secondary end points. A total of 4583 patients were included in each group. The most frequently met HBR criteria was age older than 75 years (65.6%) and Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy score ≥25 (44.6%) in the standard DAPT group and oral anticoagulant therapy (79.6%) and age 75 years and older (55.2%) in the abbreviated DAPT group. There was no statistically significant difference in net adverse clinical events (12.9% versus 13.1%; hazard ratio [HR], 0.99 [95% CI, 0.88–1.11], P=0.83), major adverse cardiovascular events (8.6% versus 7.9%; HR, 1.08 [95% CI, 0.94–1.25]), or their components between groups. The results were consistent among all of the investigated subgroups.

Conclusions

In patients with HBR undergoing percutaneous coronary intervention due to acute coronary syndrome, abbreviated DAPT was associated with comparable rates of net adverse clinical events and major adverse cardiovascular events to a DAPT duration of 12 months.

Nonstandard Abbreviations and Acronyms

BARC

Bleeding Academic Research Consortium

DAPT

dual antiplatelet therapy

ESC

European Society of Cardiology

HBR

high bleeding risk

MACE

major adverse cardiac events

MASTER DAPT

Management of High Bleeding Risk Patients Post Bioresorbable Polymer Coated Stent Implantation With an Abbreviated Versus Prolonged DAPT Regimen

NACE

net adverse clinical events

OAC

oral anticoagulant

PRECISE‐DAPT

Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy

PSM

propensity score match

RIKS‐HIA

Register of Information and Knowledge About Swedish Heart Intensive Care Admissions

SCAAR

Swedish Coronary Angiography and Angioplasty Registry

STOPDAPT‐2 ACS

Short and Optimal Duration of Dual Antiplatelet Therapy‐2 Study for Patients With ACS

SWEDEHEART

Swedish Web‐System for Enhancement and Development of Evidence‐Based Care in Heart Disease Evaluated According to Recommended Therapies

TWILIGHT‐ACS

Ticagrelor Alone vs Ticagrelor Plus Aspirin Following Percutaneous Coronary Intervention in Patients With Non‐ST‐Segment Elevation Acute Coronary Syndromes

Clinical Perspective.

What Is New?

• The current study adds real‐world evidence supporting the notion that abbreviating/de‐escalating dual antiplatelet therapy regimens is not associated with an increased ischemic risk or adverse outcome in patients with a high bleeding risk undergoing percutaneous coronary intervention due to acute coronary syndrome.

• This study was able to confirm the results of several recent randomized controlled trials in an all‐comers population.

What Are the Clinical Implications?

• While supporting the viability of abbreviated/de‐escalated dual antiplatelet therapy in patients with high bleeding risk, this study also highlights the complex balance between ischemic and bleeding risks in the heterogenous high bleeding risk population and underlines the importance of a careful assessment of both factors before choosing regimen type.

Current European Society of Cardiology (ESC) guidelines recommend a dual antiplatelet therapy (DAPT) duration of 6 to 12 months with a P2Y12 inhibitor and aspirin following drug‐eluting stent implantation after an acute coronary syndrome (ACS). ¹ , ² For patients with a high bleeding risk (HBR), guidelines recommend an individualized assessment of bleeding versus ischemic risk and some kind of abbreviated (1–6 months) or de‐escalated (eg, switch to a less potent P2Y12 inhibitor) DAPT. ² While the evidence supporting a 12‐month strategy for patients without an HBR is robust, recommendations for patients inherently at HBR are less definitive. Key issues raised are frequent exclusion or under‐representing the highest risk patients with ACS, and the 2023 ESC Task Force emphasizes that an individual evaluation of ischemic/bleeding risk should guide therapy in patients with HBR. The MASTER DAPT (Management of High Bleeding Risk Patients Post Bioresorbable Polymer Coated Stent Implantation With an Abbreviated Versus Prolonged DAPT Regimen) trial is one of the most recent examples of this and showed that abbreviated DAPT was noninferior to standard 12‐month DAPT regarding net adverse clinical events (NACE) as well as major adverse cardiac events (MACE) and resulted in significantly fewer major bleedings and clinically relevant nonmajor bleedings. ³ However, the study population consisted of a large proportion of patients with chronic coronary syndrome (40%), risking bias toward perceived benefit with short APT regimens, and allowed treatment with only 1 stent type (Ultimaster, Terumo), raising concerns on whether results could be generalized to other stents with different risks of stent thrombosis. Finally, randomization occurred 1 month after percutaneous coronary intervention (PCI) and thus excluded patients with early adverse events, complicating the extrapolation of results to real‐world settings. ³

In light of this, the present study aimed to evaluate the outcomes for patients treated with a de‐escalated or abbreviated DAPT strategy compared with a standard 12‐month DAPT duration in a real‐world population of patients with ACS, treated with any modern drug‐eluting stent and including patients from the date of hospital discharge.

METHODS

Data Sources

The nationwide RIKS‐HIA (Register of Information and Knowledge About Swedish Heart Intensive Care Admissions) is a subregistry of the SWEDEHEART (Swedish Web‐System for Enhancement and Development of Evidence‐Based Care in Heart Disease Evaluated According to Recommended Therapies) registry. ⁴ The registry contains extensive information on baseline characteristics (demography, comorbidities, and medications), presentation characteristics, information from examinations, treatments, and complications during hospital stay and discharge medications. Starting January 1, 2013, the registry implemented intention‐to‐treat variables, which included documentation of antiplatelet agents prescribed at discharge along with the intended treatment duration to allow for a comprehensive assessment of the quality of care following ACS. Data on PCI procedure were obtained from SCAAR (Swedish Coronary Angiography and Angioplasty Registry), which is part of the SWEDEHEART registry. Last, baseline data were enriched with data from several national registries. Vital status was obtained from the National Population Register (containing International Classification of Diseases [ICD] data from all admissions in Sweden since 1987), the National Cause of Death Registry, and outcomes data (bleeding and stroke) from the National Patient Registry. The authors had full access to the data, and the corresponding author takes responsibility for the analyses performed. The data set is legally restricted due to Swedish patient privacy and secrecy laws and Uppsala University and Uppsala Clinical Research Center's legal department. Data are available upon reasonable request to the Data Protection Officer at Uppsala County council at landstinget@lul.se.

Study Design and Matching

This study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational studies and was approved by the regional ethical review board in Lund. We included all patients with ACS who were enrolled in the RIKS‐HIA registry between January 1, 2013, and January 15, 2021. Patients were included if they fulfilled ≥1 HBR criteria. HBR criteria were adopted from the MASTER DAPT trial and included age≥75 years, oral anticoagulant therapy (previous or prescribed during hospital stay), anemia (defined as hemoglobin <110 g/L), history of bleeding, history of ischemic stroke, non‐skin malignancy or hematological disorder (International Classification of Diseases, Tenth Revision [ICD‐10], codes: C00–99 and D1–49, excluding C44 and D10–36), or the 4‐item Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy (PRECISE‐DAPT) score ≥25 (a validated scoring system that predicts future bleeding events) (detailed information can be seen in Table S1). ⁵ In addition, patients were only eligible if they were treated with PCI with a drug‐eluting stent and if they underwent complete revascularization of all clinically significant stenosis, defined as ≥50% luminal obstruction or after intracoronary physiology assessment to level baseline risk of de novo ischemic events. Patients were excluded if they had incomplete or missing information on intended treatment duration with P2Y12 inhibitors and acetylsalicylic acid, as well as if they experienced cardiac arrest or cardiogenic shock or died during index hospitalization. A detailed overview of all inclusion and exclusion criteria is shown in Figure 1. Patients who were discharged alive with complete data on intended antiplatelet agent and treatment duration were grouped according to whether the intended treatment duration was a standard DAPT with 12 months with any P2Y12 inhibitor and acetylsalicylic acid, or an abbreviated strategy. Abbreviated DAPT strategies included any combination of shortened or reduced antiplatelet strategy within the first 12 months. Patients were then matched on the choice of P2Y12 inhibitor at discharge using propensity score matching (PSM), after which outcomes were assessed.

Figure 1. Exclusion and inclusion criteria.

Flowchart of data management including data sources, patient inclusion, and crude matching process. Shown are the numbers of patients remaining after each step of the inclusion and exclusion criteria. DAPT indicates dual antiplatelet therapy; HBR, high bleeding risk; OAC, oral anticoagulant; PRECISE‐DAPT, Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy; and SWEDEHEART, Swedish Web‐System for Enhancement and Development of Evidence‐Based Care in Heart Disease Evaluated According to Recommended Therapies.

Outcomes

The primary outcome was NACE at 1 year defined as the first occurrence of: (1) cardiac death, (2) myocardial infarction (MI), (3) ischemic stroke, or (4) significant bleeding event. Secondary outcomes were MACE, defined as first occurrence of: (1) cardiac death, (2) MI, or (3) ischemic stroke, as well as the respective components of the primary outcome. Exact definitions and data sources for the primary end point and all secondary end points are available in Table S1. Landmark analyses starting at 30 days from discharge were conducted in order to facilitate comparisons to the MASTER DAPT trial. Further landmark analyses assessing bleeding and MI events at days 0 to 180 after discharge as well as at days 181 to 365 were also performed. This was performed to elucidate outcomes for the abbreviated group as most patients in that group were treated ≤6 months.

Sensitivity analyses using all‐cause death as well as NACE and MACE based on all‐cause death as the outcome were also performed. Censorship dates and death status were ascertained by the National Board of Health and Welfare by deterministic linkage to the National Population Register, and dates and rates of cardiac death, stroke, and bleeding outcomes were obtained through similar linkage to the National Patient Registry. Bleeding and stroke events were obtained from the National Patient Registry and defined using ICD codes (Table S1). The use of health care registries in identifying bleeding and stroke as outcomes have been previously assessed and validated. ⁶ , ⁷ The sensitivity of fatal bleeding or bleeding requiring hospitalization and fatal bleeding or nonfatal major bleeding was 99.5% and 84.5%, respectively, and a positive predictive value corresponding to 95.5% and 93.1%. ⁶ The sensitivity of first stroke or any stroke was 85.9% and 82.7%, respectively, and a positive predictive value corresponding to 94.0% and 82.7%. ⁷ The ICD codes used to define bleeding were chosen to, as far as possible, reflect bleeding events corresponding to Bleeding Academic Research Consortium (BARC) type 2, 3, or 5 events and included hemorrhagic stroke. ⁸ Information on MI was obtained from the SWEDEHEART registry and defined as a subsequent entry to a coronary care unit with a discharge diagnosis of MI according to the fourth universal definition of MI. Follow‐up was calculated from the discharge date, and all outcomes were ascertained up to June 29, 2021, with complete follow‐up available for all patients.

Statistical Analysis

Baseline data are presented as means with SDs for continuous variables and counts with percentages for categorical variables. Differences between groups were calculated using independent t or χ² tests. The event rates for both primary and secondary outcomes were calculated using the Kaplan–Meier estimator. The association between DAPT strategy and hazards of primary and secondary outcomes were assessed using Cox proportional hazards regression and presented as hazard ratios (HRs) with 95% CIs. PSM was used to balance differences related to the choice of P2Y12 inhibitor, and thus DAPT intensity, at discharge. The propensity of receiving each given treatment regimen was estimated using the intended choice of P2Y12 inhibitor as the independent variable, and patients were then matched using the nearest neighbor method with a caliper set to 0.1. All survival analyses were conducted on the PSM population. Multivariable Cox regression models adjusting for baseline differences remaining after PSM were also conducted.

Subgroup analyses with interaction testing were performed for each bleeding risk–defining criteria as well as indication (unstable angina pectoris, ST‐segment–elevation MI, and non–ST‐segment–elevation MI), number of diseased vessels (1‐vessel disease and >1‐vessel disease), vascular access (radial or femoral artery access), history of coronary artery bypass graft surgery, and history of PCI in addition to clopidogrel and prasugrel/ticagrelor. Sensitivity analyses substituting cardiac death with all‐cause death were also performed. Finally, we analyzed the complete study population before PSM using multivariable Cox regression models (Table S2).

All analyses were performed on complete case data as the proportions of missing values were generally low after application of all inclusion and exclusion criteria. A 2‐sided P value ≤0.05 was considered statistically significant in all analyses. All statistical analyses were performed in Stata SE version 17 (StataCorp).

RESULTS

Patient Characteristics

A total of 57 910 patients with uncomplicated ACS undergoing complete revascularization with PCI were identified from the SWEDEHEART registry (Figure 1). After exclusion of patients without HBR, 22 419 were eligible (5901 with an abbreviated DAPT strategy and 16 409 with a standard DAPT strategy). After matching, a total of 9166 patients were included (4583 discharged with a planned standard DAPT strategy and 4583 discharged with a planned abbreviated strategy). The final study population with their characteristics and clinical presentation are presented in Table 1. Patient characteristics before matching are presented in Table S3. The different abbreviated DAPT strategies are presented in Table 1. Adherence to a P2Y12 inhibitor at discharge was 92.9% in the standard DAPT group and 92.4% in the abbreviated DAPT group (Table S4). Most patients not adhering to the planned P2Y12 inhibitor was due to de‐escalation from ticagrelor/prasugrel to clopidogrel in both groups. Overall adherence to the planned treatment duration was estimated at 86.1% in the standard group, whereas 77.2% adhered to the planned duration with the P2Y12 inhibitor in the abbreviated DAPT group. The median time of the last retrieved prescription was 291 days (range, 248–325 days) in the standard DAPT group and 175 days (range, 84–295 days) in the abbreviated DAPT group. Overall, the average age of the study population was 74.8±9.6 years, with 34.6% of patients being women. A total of 60.4% of patients were 75 years or older, 46.6% were taking oral anticoagulants (OACs) at discharge, 9.5% had a history of clinically relevant bleeding, 13.8% had a history of ischemic stroke, 8.4% had anemia (hemoglobin <110 g/L), 2.5% had an ongoing malignancy or hematologic disorder, and 46.4% had a PRECISE‐DAPT score of ≥25. The most frequently met HBR criteria in the standard DAPT population was age 75 years and older and PRECISE‐DAPT score ≥25. In the abbreviated DAPT group, the most frequently met criterion was OAC therapy at discharge. The proportion of patients with each HBR criteria is displayed in Table 2. Patients received a total of 16 950 stents. Lesion‐level data on lesion location, lesion complexity, and stent types are presented in Table 1.

Table 1.

Baseline Characteristics of the Study Population

	Standard DAPT	Abbreviated DAPT	P value
Total number of patients	4583	4583
Age, mean±SD, y	75.2±9.9	74.4±9.3	<0.001
Women	1648 (36.0%)	1529 (33.4%)	0.009
BMI, mean±SD	26.9±4.6	27.4±4.8	<0.001
PRECISE‐DAPT score, mean ±SD	28.1±14.9	25.0±14.6	<0.001
Smoking status			0.44
Never	2020 (44.1%)	1964 (42.9%)
Ex‐smoker (>1 mo)	1836 (40.1%)	1912 (41.7%)
Current smoker	555 (12.1%)	535 (11.7%)
Unknown	172 (3.8%)	172 (3.8%)
Medical history
Diabetes	1188 (25.9%)	1269 (27.7%)	0.056
Hypertension	2225 (48.5%)	2469 (53.9%)	<0.001
MI	1289 (28.1%)	1352 (29.5%)	0.15
Hyperlipidemia	2223 (49.3%)	2320 (51.3%)	0.054
Coronary artery disease	1510 (32.9%)	1606 (35.0%)	0.034
PCI	976 (21.3%)	1010 (22.0%)	0.39
CABG	265 (5.8%)	341 (7.4%)	0.001
Chronic heart failure	592 (12.9%)	923 (20.1%)	<0.001
Peripheral artery disease	295 (6.4%)	315 (6.9%)	0.40
Chronic obstructive pulmonary disease	393 (8.6%)	447 (9.8%)	0.051
Renal failure	338 (7.4%)	308 (6.7%)	0.22
Previous medications
ACEI	1047 (22.9%)	1134 (24.9%)	0.028
ARB	1122 (24.6%)	1269 (27.9%)	<0.001
β‐Blockers	1971 (43.2%)	2540 (55.8%)	<0.001
Statins	1771 (38.8%)	1827 (40.1%)	0.20
Clinical presentation	1089 (23.8%)	1241 (27.2%)	<0.001
Unstable angina	838 (18.3%)	833 (18.2%)	0.60
NSTEMI	2335 (51.0%)	2296 (50.1%)
STEMI	1410 (30.8%)	1454 (31.7%)
P2Y12 inhibitor treatment at presentation			<0.001
None	3949 (86.2%)	4238 (92.5%)
Clopidogrel	530 (11.6%)	239 (5.2%)
Prasugrel	2 (<1%)	2 (<1%)
Ticagrelor	71 (1.5%)	70 (1.5%)
Other/unknown	30 (0.7%)	31 (0.7%)
OAC treatment at presentation			<0.001
None	4249 (92.7%)	2388 (52.1%)
Warfarin	164 (3.6%)	1071 (23.4%)
Dabigatran, rivaroxaban, and apixaban	152 (3.3%)	1090 (23.8%)
Other/unknown	18 (0.4%)	34 (0.7%)
eGFR, mean±SD	67.8±21.8	68.8±20.5	0.038
Hemoglobin (g/L), mean±SD	133.6±17.7	136.5±17.4	<0.001
Angiographic findings			0.23
One‐vessel disease not LM	2920 (63.7%)	2996 (65.4%)
Two‐vessel disease not LM	1075 (23.5%)	1032 (22.5%)
Three‐vessel disease not LM	345 (7.5%)	345 (7.5%)
LM+1‐ to 3‐vessel disease	243 (5.3%)	210 (4.6%)
Lesion location*	8590 (100.0%)	8360 (100.0%)	†
Proximal left anterior descending artery	1837 (21.4%)	1912 (22.9%)	0.02
Middle left anterior descending artery	1658 (19.3%)	1679 (20.1%)	0.20
Distal left anterior descending artery	269 (3.1%)	272 (3.3%)	0.65
Left anterior descending branches	352 (4.1%)	321 (3.8%)	0.39
Left main stem	319 (3.7%)	314 (3.8%)	0.88
Proximal circumflex artery	696 (8.1%)	595 (7.1%)	0.02
Distal circumflex artery	305 (3.6%)	272 (3.3%)	0.29
Circumflex branches	687 (8.0%)	709 (8.5%)	0.25
Proximal right coronary artery	821 (9.6%)	718 (8.6%)	0.22
Middle right coronary artery	853 (9.9%)	784 (9.4%)	0.22
Distal right coronary artery	481 (5.6%)	426 (5.1%)	0.15
Right coronary artery branches	196 (2.3%)	218 (2.6%)	0.17
Lesion complexity according to ACC/AHA*	8590 (100.0%)	8360 (100.0%)
A	658 (7.7%)	741 (8.9%)	<0.001†
B1	2688 (31.3%)	2621 (31.4%)
B2	2520 (29.3%)	2483 (29.7%)
C	1384 (16.1%)	1238 (14.8%)
B1 bifurcation	457 (5.3%)	395 (4.7%)
B2 bifurcation	636 (7.4%)	563 (6.7%)
C bifurcation	240 (2.8%)	315 (3.8%)
Unclassified	7 (0.08%)	4 (0.05%)
No. of stents implanted			0.002
1	2580 (56.3%)	2732 (59.6%)
2	1234 (26.9%)	1179 (25.7%)
≥3	769 (16.8%)	672 (14.7%)
Stent type*	8590 (100.0%)	8360 (100.0%)	<0.001
Boston Scientific SYNERGY	1914 (22.3%)	2149 (25.7%)
Medtronic RESOLUTE ONYX	1545 (18.0%)	1829 (21.9%)
Boston Scientific PROMUS PREMIER	1556 (18.1%)	1029 (12.3%)
Biotronic ORSIRO	790 (9.2%)	622 (7.4%)
Medtronic RESOLUTE INTEGRITY	772 (9.0%)	527 (6.3%)
Boston Scientific PROMUS ELITE	284 (3.3%)	475 (5.7%)
Other	1729 (20.1%)	1729 (20.7%)
ECG rhythm at discharge			<0.001
Sinus rhythm	4232 (92.3%)	3127 (68.2%)
Atrial fibrillation/flutter	225 (4.9%)	1240 (27.1%)
Unknown/other	126 (2.7%)	216 (4.7%)
OACs at discharge			<0.001
None	4011 (87.5%)	965 (21.1%)
Warfarin	325 (7.1%)	1504 (32.8%)
Dabigatran, rivaroxaban, and apixaban	230 (5.0%)	2084 (45.5%)
Unknown/other	17 (0.4%)	30 (0.7%)
P2Y12 inhibitor at discharge			0.99
Clopidogrel	2927 (63.9%)	2927 (63.9%)
Prasugrel	24 (0.5%)	24 (0.5%)
Ticagrelor	1632 (35.6%)	1632 (35.6%)
Abbreviated antiplatelet strategy
Short DAPT with <6 mo P2Y12‐inhibitor and 12 mo ASA		1126 (24.6%)
Short DAPT <6 mo P2Y12 inhibitor and 6 mo ASA		488 (10.6%)
Short DAPT with <6 mo ASA and 12 mo P2Y12 inhibitor		892 (19.5%)
Short SAPT with <6 mo P2Y12 inhibitor		1041 (22.7%)
SAPT with 12 mo P2Y12 inhibitor		1036 (22.6%)
Standard DAPT with 12 mo ASA and P2Y12	4583 (100%)	0 (0%)
ASA duration, mo
1	0 (0.0%)	947 (37.8%)
3	0 (0.0%)	328 (13.1%)
6	0 (0.0%)	105 (4.2%)
12	4583 (100%)	1126 (44.9%)
P2Y12 duration, mo
1	0 (0.0%)	319 (7.0%)
3	0 (0.0%)	550 (12.0%)
6	0 (0.0%)	1786 (39.0%)
12	4583 (100%)	1928 (42.1%)
Clopidogrel duration, mo
1	0 (0.0%)	227 (7.8%)
3	0 (0.0%)	329 (11.2%)
6	0 (0.0%)	936 (32.0%)
12	2927 (100%)	1435 (49.0%)
Prasugrel/ticagrelor duration, mo
1	0 (0.0%)	92 (5.6%)
3	0 (0.0%)	221 (13.4%)
6	0 (0.0%)	850 (51.3%)
12	1656 (100.0%)	493 (29.8%)

Baseline demographic and medical characteristics for the study population stratified on standard abbreviated vs standard DAPT strategy. Differences between groups were calculated using independent t and χ² tests for nominal and ordinal data. ACC/AHA indicates American College of Cardiology/American Heart Association; ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; ASA, acetylsalicylic acid; BMI, body mass index; CABG, coronary artery bypass graft; DAPT, dual antiplatelet therapy; eGFR, estimated glomerular filtration rate; LM, left main (coronary artery); MI, myocardial infarction; NSTEMI, non–ST‐segment–elevation myocardial infarction; OAC, oral anticoagulant; PCI, percutaneous coronary intervention; PRECISE‐DAPT, Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy; SAPT, single antiplatelet therapy; and STEMI, ST‐segment–elevation myocardial infarction.

^*Lesion‐level data. The denominator is the total number of treated lesions (16 950) by each group.

^†No statistical differences in patient‐level data.

Table 2.

Distribution of HBR Features Between Study Groups

	Standard DAPT, n (%)	Abbreviated DAPT, n (%)	P value	Missing, n (%)
Age≥75 y	3009 (65.6%)	2528 (55.2%)	<0.001	0
OAC	621 (13.6%)	3648 (79.6%)	<0.001	0
Previous bleeding*	504 (11.0%)	369 (8.1%)	<0.001	0
Previous stroke*	685 (14.9%)	578 (12.6%)	<0.001	0
Hemoglobin <110 g/L	392 (8.6%)	299 (6.5%)	<0.001	943 (10.3%)
Malignancy/hematologic disorder	135 (3.0%)	96 (2.1%)	<0.001	0
PRECISE‐DAPT score ≥25	2044 (44.6%)	1763 (38.5%)	<0.001	954 (10.4%)

Distribution of HBR features between study groups. Differences between groups were calculated with χ² test. P<0.05 is regarded as significant. DAPT indicates dual antiplatelet therapy; HBR, high bleeding risk; OAC, oral anticoagulant; and PRECISE‐DAPT, Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy.

^*Past 12 months: International Classification of Diseases, Tenth Revision (ICD‐10), codes: C00–99 and D1–49, excluding C44 and D10–36 during index hospitalization.

Results of Primary and Secondary Analyses

The Kaplan–Meier events and event rate of NACE was 584 (12.9%) for patients with an abbreviated DAPT time and 594 (13.1%) for patients with a standard DAPT time (HR, 0.99 [95% CI, 0.88–1.11], P=0.83). Event rates of MACE were 8.6% and 7.9% for the abbreviated DAPT time and standard DAPT time group, respectively (HR, 1.08 [95% CI, 0.94–1.25], P=0.28) (Figure 2). There was no statistically significant difference in incidence of any of the end points making up the composite end points: clinically significant bleeding (HR, 0.88 [95% CI, 0.74–1.04], P=0.13), ischemic stroke (HR, 1.18 [95% CI, 0.89–1.58], P=0.26), MI (HR, 1.04 [95% CI, 0.85–1.28], P=0.68), or cardiac death (HR, 1.04 [95% CI, 0.81–1.35], P=0.73) (Figure 2). There was no statistically significant difference in NACE and MACE based on all‐cause mortality or in all‐cause mortality between groups (Table 3). Sensitivity analyses adjusting for baseline differences after PSM showed consistent results with outcomes of crude univariable regression models (Table S5).

Figure 2. Results of primary and secondary outcomes.

Kaplan–Meier failure‐function, results of Cox proportional hazards regressions, and the cumulative incidence of both primary composite end points. (A) NACE consisting of first event of cardiac death, myocardial infarction, ischemic stroke, or hospitalization due to bleeding according to planned dual antiplatelet therapy time and (B) MACE consisting of first event of cardiac death, myocardial infarction, or ischemic stroke, as well as the same data for all of the individual subgroups of the composite primary end point NACE. DAPT indicates dual antiplatelet therapy; HR, hazard ratio; MACE, major adverse cardiac events; and NACE, net adverse clinical events.

Table 3.

Results of Sensitivity and Landmark Analyses

	Standard DAPT, % (n)	Abbreviated DAPT, % (n)	HR (95% CI)	P value
Standard follow‐up (0–365 d)
Using all‐cause death
NACE	15.6 (715)	15.7 (720)	1.01 (0.91–1.12)	0.83
MACE	10.8 (495)	12.0 (549)	1.12 (0.99–1.26)	0.07
All‐cause death	6.3 (289)	6.8 (311)	1.08 (0.92–1.27)	0.35
Thirty‐day landmark analyses (30–365 d)
Using cardiac death
NACE	10.0 (437)	9.6 (419)	0.96 (0.84–1.10)	0.60
MACE	6.0 (266)	6.6 (289)	1.10 (0.93–1.30)	0.28
Bleeding	5.0 (216)	4.0 (176)	0.82 (0.67–1.00)	0.04
MI	3.2 (138)	3.5 (153)	1.12 (0.89–1.41)	0.35
Stroke	1.5 (67)	1.7 (76)	1.14 (0.82–1.58)	0.43
Cardiac death	1.9 (85)	2.0 (87)	1.03 (0.76–1.38)	0.87
Using all‐cause death
NACE	12.4 (547)	12.3 (541)	0.99 (0.88–1.12)	0.93
MACE	8.7 (389)	9.8 (438)	1.14 (0.99–1.30)	0.07
All‐cause death	5.3 (238)	5.7 (258)		0.35

Event rates of Kaplan–Meier failure function for primary and secondary end points together with results of Cox proportional hazards regression analyses shown as HRs and 95% CI. DAPT indicates dual antiplatelet therapy; HR, hazard ratio; MACE, major adverse cardiac events; MI, myocardial infarction; and NACE, net adverse clinical events.

Landmark Analyses

Thirty‐day landmark analysis after discharge showed no statistically significant difference in NACE between groups with a Kaplan–Meier event rate of 9.6% in the abbreviated group and 10.0% in the standard group (HR, 0.96 [95% CI, 0.84–1.10], P=0.60) (Table 3). There was no statistically significant difference between groups regarding NACE within the first month of discharge (HR, 1.05 [95% CI, 0.85–1.31], P=0.65). There was no statistically significant difference between groups in the 30‐day MACE landmark analyses (6.6% and 6.0%, respectively) (HR, 1.10 [95% CI, 0.93–1.30], P=0.28) or any other secondary end point, except for bleeding, which was lower in the abbreviated DAPT group (5.0% versus 4.0%) (HR, 0.82 [95% CI, 0.67–1.00], P=0.04) (Table 3). Last, landmark analyses analyzing the cumulative incidence and HR of bleeding and MI at 0 to 180 days and 181 to 365 days showed no statistically significant difference between study groups during days 0 to 180 (HR, 0.99 [95% CI, 0.81–1.22], P=0.92; and HR, 1.07 [95% CI, 0.83–1.39], P=0.59). The landmark analysis for days 181 to 365 showed statistically significantly fewer bleeding complications in the abbreviated DAPT group, whereas no statistically significant difference in risk of MI was observed (HR, 0.69 [95% CI, 0.50–0.92], P=0.013; and HR, 1.00 [95% CI, 0.72–1.38], P=0.98) (Figure 3).

Figure 3. Landmark analyses.

Kaplan–Meier failure‐function as well as results of Cox proportional hazards regressions shown as HRs and corresponding 95% CIs for both bleeding events and myocardial infarction for the landmark analyses between days 0 and 180 and 181 and 365, respectively. DAPT indicates dual antiplatelet therapy; and HR, hazard ratio.

Subgroup Analyses

Subgroup analyses assessing the primary end point are presented in Figure 4. Results showed no effect modifiers of the primary outcome as indicated by interaction tests with the exception of clopidogrel versus prasugrel/ticagrelor. Abbreviated DAPT was associated with a lower risk of NACE in the subgroup of patients discharged with clopidogrel and a higher risk of NACE in the subgroup discharged with more potent P2Y12 inhibitors (P value of interaction <0.001) (Figure S1). Subgroup analyses of each HBR group were investigated for all secondary end points and are presented in Tables S6 through S10. These analyses were in line with the primary analyses showing no statistically significant difference between abbreviated versus standard DAPT in any subgroup and any outcome measure, except for patients with cancer, who showed a statistically significant decrease in the rate of MACE (Table S6).

Figure 4. Subgroup analyses.

Forest plot depicting cumulative incidence and results of Cox proportional hazard regression for each respective subgroup shown. Results of interaction tests assessing each subgroup for significant interaction with the effects of abbreviated DAPT on the primary end point of net adverse clinical events. CABG indicates coronary artery bypass graft; DAPT, dual antiplatelet therapy; HR, hazard ratio; NSTEMI, non–ST‐segment–elevation myocardial infarction; OAC, oral anticoagulant; PCI, percutaneous coronary intervention; PRECISE‐DAPT, Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy; and STEMI, ST‐segment–elevation myocardial infarction.

Unmatched Population

Multivariable Cox regressions adjusted for baseline differences in the unmatched population were performed for all end points. Results adhered to the finding of the PSM cohort, showing no statistically significant difference between the abbreviated versus standard DAPT strategy with regard to NACE (adjusted HR, 1.0 [95% CI, 0.87–1.16], P=0.98) (Table S2). In addition, there was no statistically significant difference in MACE, cardiac death, bleeding, stroke, and MI. There was a higher risk of all‐cause death with the abbreviated DAPT strategy (adjusted HR, 1.27 [95% CI, 1.0–1.55], P=0.02) (Table S2).

Post Hoc Analyses

To address the difference in OAC treatment among the groups, we conducted a post hoc analysis by assessing a cohort matched for the intended OAC therapy at discharge, in addition to choice of P2Y12 inhibitor. A total of 1763 patients were included in each arm after matching, and their demographic characteristics are detailed in (Table S11). There were no statistically significant differences observed in the administration of P2Y12 inhibitors or OAC treatment within this cohort. Furthermore, the Kaplan–Meier event rates for NACE within this subgroup closely mirrored those of the primary analysis, with rates at 12.0% versus 12.6% for the standard and respective abbreviated DAPT groups, respectively. No statistically significant differences were observed in NACE or any other outcome measures as detailed in Table S12.

DISCUSSION

In this nationwide population‐based study of patients with HBR who had ACS treated with PCI, there was no statistically significant difference in the incidence of NACE and MACE at 1 year between patients with HBR with an intended abbreviated DAPT strategy compared with those with a planned 12‐month DAPT regimen. These findings align with the results of recent randomized controlled trials indicating that abbreviated DAPT strategies are not associated with an adverse outcome and further underlines the importance of considering an abbreviated or de‐escalated DAPT strategy as a treatment option for patients with HBR who have ACS.

Outcome studies based on treatment duration using observational data are inherently complicated by various systematic errors such as selection bias arising from the absence of information on treatment intention. In addition, time‐dependent bias is frequently encountered, with patients with shorter survival having a shorter treatment duration and fewer prescriptions. ⁹ However, previous research has shown that intention‐to‐treat variables in registry data are able to provide reliable effect estimates. ⁹ , ¹⁰ , ¹¹ , ¹² The incorporation of intention‐to‐treat variables, specifically related to planned APT and its duration within the SWEDEHEART registry, allowed us to evaluate outcomes of a real‐world DAPT abbreviation strategy. The application of our inclusion and exclusion criteria resulting in at least 1 HBR criterion for each patient reduced the heterogeneity of the study population. On top of this, patients were matched on their planned P2Y12 inhibitor to ensure comparability between patients receiving an abbreviated or standard DAPT in terms of P2Y12 agents. This matching process aimed to isolate the effects of DAPT duration from the impact of varying potency of antiplatelet agents, which is a crucial predictor of bleeding complications in general but particularly so for patients with HBR. However, to address confounding arising from imbalances after matching, we also conducted a comprehensive set of analyses. First, we employed multivariable‐adjusted Cox regression models on the PSM population, accounting for imbalances between the treatment groups. These analyses consistently showed no statistically significant differences in outcomes between the 2 treatment strategies but with the expected trends towards decreased bleeding and increased ischemic events in the abbreviated treatment arm. Second, we performed multivariable‐adjusted Cox regression models on the unmatched population. The results of these analyses also aligned with the main findings. Third, we assessed outcomes across a wide set of subgroups with consistent results. One noteworthy exception was the fact that the abbreviated DAPT strategy was associated with statistically significantly worse outcomes in patients discharged with more potent P2Y12 inhibitors (ticagrelor and prasugrel) and statistically significantly better outcomes in those receiving clopidogrel. This may well be due to the fact that a higher proportion of patients in the abbreviated DAPT group were discharged with both OACs and prasugrel/ticagrelor (28.9% compared with 2.1% in in the standard DAPT group), a combination of therapies not recommended by guidelines. Fourth, we assessed the composite end points based on cardiac death, which we believe to be superior (as opposed to all‐cause death) when attempting to evaluate APT in a heterogenic population as it is more sensitive to changes in key events such as in‐stent thrombosis. However, we also assessed all‐cause death and the composite outcomes NACE and MACE based on all‐cause death instead of cardiac death to rule out the possibility of conflicting results and ensure consistency as well as enhance comparability with other trials. The results of these analyses consistently demonstrated no statistically significant differences between treatment strategies with any mortality definition but somewhat larger differences between treatment groups.

The assessment of patients meeting the HBR eligibility criteria adopted from the MASTER DAPT trial resulted in a study population with comparable age and sex distribution, prevalence of comorbidities, and bleeding risk (mean PRECISE‐DAPT score of 26.5 versus 26.8). In the MASTER DAPT trial, patients were randomized 1 month after PCI. In the substudy of MASTER DAPT, which compared abbreviated DAPT with standard DAPT in patients with a recent ACS (within 12 months), event rates were comparable to the event rates observed in this study (8.9% and 7.6%) for NACE and MACE, respectively, versus 9.6% and 6.6%, respectively, in this study. ¹³ Moreover, the rate of cardiac death in our study was 1.9% compared with 2.0% in the MASTER DAPT trial, further emphasizing the validity of our study population and the transferability of the trial results to a real‐world setting. However, it is important to note that the rates of bleeding complications in the ACS substudy of the MASTER DAPT trial were reported as 6.2%, whereas in our study, they were observed to be 5.7% in the primary analysis after discharge and 4.0% in the 30‐day landmark analysis. These discrepancies can partially be explained by differences in outcome definitions but also underscore the challenges associated with studying bleeding complications using registry data. This should not affect our study results as no systematic differences in reporting of bleeding outcomes should exist between the 2 treatment strategies in our study. We observed a statistically nonsignificant trend toward fewer bleeding complications observed with abbreviated DAPT. This is in contrast to what is often seen in randomized controlled de‐escalation trials such as the 1‐month DAPT followed by clopidogrel monotherapy in ACS trial (STOPDAPT‐2 ACS [Short and Optimal Duration of Dual Antiplatelet Therapy‐2 Study for Patients With ACS]), the MASTER DAPT trial, and the TWILIGHT‐ACS (Ticagrelor Alone vs Ticagrelor Plus Aspirin Following Percutaneous Coronary Intervention in Patients With Non‐ST‐Segment Elevation Acute Coronary Syndromes) trial, in which the abbreviated/de‐escalation arms showed statistically significantly lower rates of bleeding events. ¹⁴ , ¹⁵ We believe that the lack of effect on bleeding rates in the abbreviated DAPT group can be explained by a chronically elevated bleeding risk due to OAC treatment that continues beyond DAPT discontinuation, which balances the elevated bleeding risk of standard DAPT duration and mitigates the expected bleeding events. Similarly, the lack of difference in MI risk could be explained by the fact that only patients with complete revascularization were included but also by the skewed distribution of OAC treatment, which theoretically counteracted stent thrombosis in the abbreviated DAPT group. The ESC guidelines recommend 6 months of DAPT for patients with HBR with ACS undergoing PCI, but, at the same time, recommend an individualized approach based on ischemic versus bleeding risk. More than half of patients (57.9%) in the abbreviated DAPT group in our study received ≤6 months of DAPT and the remaining 42.1% of patients received 12 months of P2Y12 inhibitors, among whom half received reduced aspirin duration or no aspirin at all. The diversity in abbreviation strategy reflects real‐world clinical practice, wherein APT is individualized after careful assessment of ischemic and bleeding risks, with consideration given to OAC treatment. To address the differences in OAC treatment and increase comparability between groups, numerous additional analyses were performed, such as multivariable Cox regression models adjusted for OAC treatment, subgroup analyses of both patients with and without OAC treatment, as well as a separate cohort of patients matched for OAC treatment. These analyses aligned with the primary findings of the study both in terms of event rates and relative risk between strategies, across all end points. Finally, results from the landmark analyses at 30 and 180 days showed that abbreviated DAPT was associated with fewer bleeding complications and thus aligns with previous evidence.

Altogether, the results of the current study supports the notion that abbreviated regimens can be successful in decreasing bleeding risk without dramatically increasing ischemic risk. However, the evidence from this study and previous research is not irrefutable in any way. Important subgroups such as patients with ST‐segment–elevation MI in trials such as TWILIGHT‐ACS, or patients taking OACs in the TWILIGHT‐ACS and STOPDAPT2‐ACS trials, continue to be excluded from clinical trials. As approximately one‐third of patients with ACS present with ST‐segment–elevation MI and a sizeable proportion are discharged with an OAC, the transferability of findings from such randomized clinical trials to real‐world scenarios is complicated and more studies are needed to give clear evidence of safe DAPT de‐escalation in a wide ACS population.

Limitations

This study has several limitations. Despite our efforts to address measured confounding, it is essential to acknowlege that the results may be constrained by unmeasured confounding. The rationale of study design was to include patients with at least 1 HBR criterion and balance their bleeding risk related to antiplatelet therapy (clopidogrel versus ticagrelor/prasugrel) to isolate the effects of differing DAPT duration from the differing DAPT potency. The use of PSM was therefore not used to emulate the perfectly balanced treatment/control populations of randomized clinical trials. The imbalances between the study groups, particularly regarding OAC treatment, are critical to keep in mind when interpreting the study results. PSM has been shown to carry the risk of introducing hidden bias in data sets when used to perfectly balance confounders. ¹⁶ We believe that our descriptive review with comparisons between matched, unmatched, multivariable‐adjusted, and stratified analyses provides a more nuanced insight with less risk of hidden bias. Further limitations include the utilization of ICD codes for outcome identification. Although all analyses were conducted based on intention‐to‐treat data, which minimize systematic errors, it is important to note that the actual adherence to the intended regimen, while generally high, may lead to misclassification of exposure, potentially impacting the results. Last, it must be emphasized that the short treatment arm of this study comprised several differing de‐escalation/abbreviated DAPT regimens and that the outcomes of each one of these might well differ from the group.

CONCLUSION

In this population‐based study reflecting real‐world conditions, patients with HBR with ACS undergoing PCI showed comparable outcomes for patients discharged with an abbreviated or de‐escalated DAPT strategy and those receiving a standard DAPT strategy in terms of NACE and MACE. These findings contribute to the growing body of evidence supporting the viability of abbreviated DAPT after PCI in patients with HBR who have ACS. However, the decision to adopt an abbreviated regimen should be made after a thorough evaluation of the balance between prothrombotic and bleeding risk factors.

Sources of Funding

This work was supported by the Swedish Heart and Lung Foundation (M.A.M. and D.E.), Swedish Scientific Research Council (D.E.), Skane University Hospital funds (D.E.), Thorsten Westerström's foundation (M.A.M.), ALF (M.A.M. and D.E.), The Crafoord Foundation (Crafordska stiftelsen) (M.A.M.), the Swedish Society of Medicine, and the Anna‐Lisa and Sven‐Eric Lundgren Foundation for medical research (M.A.M.). The sponsors were not involved in the study design, collection of data, analysis of data, interpretation of data, writing of the article, approval of the article, or in the decision to submit the article for publication. The decision to submit the article was solely the authors.

Disclosures

None.

Supporting information

Tables S1–S12

Figure S1