Highlights
– TWILIGHT posthoc analysis (n = 7,119) categorized patients by number of predefined high‑risk enrollment criteria and evaluated bleeding (BARC 2/3/5) and ischemic (death, MI, stroke) outcomes through 1 year after randomization.
– Ischemic events increased stepwise as the number of high‑risk criteria rose; bleeding events did not.
– Ticagrelor monotherapy (after 3 months of DAPT) consistently reduced BARC 2/3/5 bleeding compared with ticagrelor plus aspirin across all risk‑burden strata without a significant increase in ischemic events (no significant interaction by risk group).
Study background and disease burden
Patients undergoing percutaneous coronary intervention (PCI) often receive dual antiplatelet therapy (DAPT) to reduce stent‑ and procedure‑related ischemic events, but prolonged aspirin in addition to a potent P2Y12 inhibitor increases bleeding risk. Major bleeding is independently associated with morbidity, mortality, and prolonged hospitalization; conversely, ischemic events (myocardial infarction, stroke, cardiovascular death) remain the primary driver of long‑term adverse outcomes after PCI. Identifying strategies that reduce bleeding without compromising ischemic protection is therefore a major unmet need in contemporary interventional cardiology.
The randomized TWILIGHT strategy (randomization at 3 months after PCI among patients free of ischemic and bleeding events) tested whether stopping aspirin and continuing ticagrelor alone could achieve this balance. The present posthoc analysis (Steg et al., Am Heart J. 2025) interrogates how the predefined high‑risk enrollment criteria—both clinical and angiographic—map to outcomes and whether the relative benefit of ticagrelor monotherapy is preserved across risk burden strata.
Study design
This is a posthoc analysis of the TWILIGHT randomized trial cohort (ClinicalTrials.gov NCT02270242) including 7,119 patients who were event‑free at 3 months after PCI and were randomized to ticagrelor monotherapy versus ticagrelor plus aspirin. The analysis classified participants by the number of high‑risk criteria at enrollment (≤3, 4, 5, or ≥6 criteria). High‑risk features included clinical items (for example, troponin‑positive acute coronary syndrome [ACS], atherosclerotic vascular disease) and angiographic/procedural items (for example, multivessel disease [MVD], left main or proximal LAD lesion, total implanted stent length > 30 mm).
Primary outcome: Bleeding Academic Research Consortium (BARC) type 2, 3, or 5 bleeding through 1 year after randomization.
Key secondary outcome: composite of death, myocardial infarction (MI), or stroke through 1 year.
Statistical approach: event rates and hazard ratios (HR) with 95% confidence intervals (CI) were estimated across risk strata and treatment arms; interaction testing assessed whether the effect of ticagrelor monotherapy versus ticagrelor plus aspirin differed by number of high‑risk criteria.
Key findings
Population and risk feature prevalence
– Total analyzed: 7,119 patients. Distribution by number of high‑risk features: ≤3 criteria, 21.5%; 4 criteria, 32.7%; 5 criteria, 27.4%; ≥6 criteria, 18.4%.
– Troponin‑positive ACS was the most common clinical criterion (64.9%).
– Multivessel disease was the most common angiographic criterion (66.5%).
– The most frequent combination (intersection) of criteria included troponin‑positive ACS, atherosclerotic vascular disease, MVD, left main or proximal LAD lesion, and total stent length >30 mm.
Relationship between number of risk features and outcomes
– Ischemic risk (composite of death/MI/stroke) increased stepwise with the number of high‑risk enrollment criteria. This suggests the predefined criteria successfully enriched for ischemic risk.
– Bleeding risk (BARC 2/3/5) did not show the same stepwise increase with higher numbers of high‑risk criteria, indicating that the chosen high‑risk features were more predictive of thrombotic events than of bleeding in this cohort.
Treatment comparisons by risk strata (1‑year outcomes after randomization)
Bleeding (BARC 2/3/5): ticagrelor monotherapy vs ticagrelor + aspirin
– ≤3 RF: 3.5% vs 5.8% — HR 0.59 (95% CI 0.38–0.93).
– 4 RF: 3.7% vs 6.4% — HR 0.57 (95% CI 0.37–0.86).
– 5 RF: 3.8% vs 8.6% — HR 0.44 (95% CI 0.29–0.66).
– ≥6 RF: 5.3% vs 7.9% — HR 0.65 (95% CI 0.44–0.96).
– Interaction p‑value = 0.56, indicating no statistically significant heterogeneity of the bleeding benefit across risk groups.
Ischemic composite (death/MI/stroke): ticagrelor monotherapy vs ticagrelor + aspirin
– ≤3 RF: 1.6% vs 2.1% — HR 0.75 (95% CI 0.38–1.50).
– 4 RF: 3.5% vs 2.2% — HR 1.58 (95% CI 0.91–2.75).
– 5 RF: 4.1% vs 5.0% — HR 0.80 (95% CI 0.51–1.24).
– ≥6 RF: 6.7% vs 6.9% — HR 0.98 (95% CI 0.67–1.43).
– Interaction p‑value = 0.22, indicating no statistically significant heterogeneity in ischemic outcomes by risk group.
Interpretation of effect sizes and clinical significance
– Ticagrelor monotherapy reduced clinically relevant bleeding by approximately 35–56% (HR range ~0.44–0.65) across the risk strata, with consistent point estimates favoring monotherapy.
– There was no consistent or statistically significant signal of increased ischemic events with aspirin withdrawal across risk groups; point estimates varied but confidence intervals overlapped and interaction tests were non‑significant.
– The magnitude of absolute bleeding reduction was greater in groups with higher absolute bleeding rates; however, in this cohort bleeding rates were relatively low overall, likely reflecting selection of patients who were event‑free at 3 months.
Safety and secondary observations
– There was no clear trade‑off between less bleeding and more ischemia across the predefined high‑risk burden groups within this selected population.
– The results reinforce that the predefined high‑risk criteria enriched for ischemic risk more than for bleeding risk in this trial population.
Expert commentary and implications for practice
Clinical meaning
– The analysis provides useful granularity for clinicians considering aspirin discontinuation after an initial 3‑month period of DAPT in high‑risk patients who remain event‑free. The consistent bleeding reduction with preserved ischemic safety across risk‑burden strata supports ticagrelor monotherapy as a viable de‑escalation strategy in this selected population.
Limitations and caveats
– Posthoc nature: subgroup analyses by number of risk criteria are exploratory and hypothesis‑generating; the trial was not prospectively powered for all subgroup comparisons reported here.
– Selection bias: only patients event‑free for 3 months after PCI were randomized; the results do not apply to patients with early post‑PCI complications or early ischemic/bleeding events.
– Generalizability: all randomized patients received ticagrelor as the P2Y12 agent; findings may not directly apply to clopidogrel or prasugrel monotherapy strategies, or to populations with different bleeding/ischemic risk profiles.
– Unmeasured confounders: although the predefined criteria were comprehensive, other variables (e.g., frailty, concomitant anticoagulation, gastrointestinal risk factors) could modify bleeding and ischemic risks.
Context with current evidence
– The concept of early aspirin withdrawal to reduce bleeding while continuing a potent P2Y12 inhibitor reflects a growing body of work evaluating antiplatelet de‑escalation strategies after PCI. This analysis adds nuance by showing that the predefined high‑risk features enroll a population with a graded ischemic risk but that ticagrelor monotherapy’s bleeding benefits and ischemic neutrality are maintained across that gradient—within the limits of selection at 3 months.
Practical recommendation for clinicians
– For patients similar to the TWILIGHT trial population (high‑risk by enrollment criteria and event‑free at 3 months after PCI), stopping aspirin and continuing ticagrelor may be considered to reduce bleeding risk without materially increasing ischemic risk. Decisions should be individualized, accounting for patient comorbidities, bleeding predisposition, concomitant therapies (e.g., oral anticoagulation), and patient preference.
Conclusion
This posthoc analysis of the TWILIGHT trial demonstrates that the trial’s predefined high‑risk enrollment criteria correlated more strongly with ischemic than bleeding risk and that ticagrelor monotherapy initiated at 3 months after PCI consistently reduced clinically relevant bleeding versus continued aspirin plus ticagrelor, without a statistically significant increase in ischemic events across strata of increasing risk feature burden. The findings support ticagrelor monotherapy as a valid de‑escalation option for appropriately selected, event‑free patients after 3 months—but clinicians must recognize the selection limits of the cohort and apply results only to similar patients.
References
– Steg PG, Nicolas J, Baber U, Sartori S, Zhang Z, Feng Y, Angiolillo DJ, Briguori C, Cohen DJ, Collier T, Dangas G, Dudek D, Escaned J, Gibson CM, Han YL, Huber K, Kastrati A, Kaul U, Marx SO, Kornowski R, Kunadian V, Vogel B, Oliva A, Mehta SR, Moliterno D, Sardella G, Krucoff M, Shlofmitz RA, Sharma S, Pocock S, Mehran R. Characterizing high‑risk enrollment criteria and impact on clinical outcomes in a large randomized clinical trial: Insights from the TWILIGHT trial. Am Heart J. 2025 Aug;286:97-107. doi: 10.1016/j.ahj.2025.01.016 IF: 3.5 Q2 . Epub 2025 Jan 30. PMID: 39889917 IF: 3.5 Q2 .- TWILIGHT ClinicalTrials.gov registration: NCT02270242.
