Critical review · May 21, 2026
STOP-IMH: ticagrelor without aspirin right after primary PCI in STEMI
Yosofi B, Woelders ECI, Peeters DAM, von Birgelen C, van Rees JB, Heestermans AACM, et al. Ticagrelor monotherapy versus dual antiplatelet therapy after primary percutaneous coronary intervention: one-year outcomes of the pilot STOP-IMH randomised trial. EuroIntervention. 2026;22(14):e773-e783.
doi:10.4244/EIJ-D-26-00421 · Open-label multicentre randomised pilot trial (the Netherlands; NCT05986968), n = 200
Summary
STOP-IMH randomised 200 STEMI patients to ticagrelor monotherapy or ticagrelor plus aspirin, directly after primary PCI. Follow-up was 13 months. A cardiac magnetic resonance (CMR) substudy assessed intramyocardial haemorrhage (IMH).
| Outcome | Ticagrelor (n = 98) | DAPT (n = 101) | HR (95% CI) |
|---|---|---|---|
| MACCE | 4.1% | 4.0% | 1.04 (0.26 to 4.17) |
| Clinically relevant bleeding | 4.1% | 10.9% | 0.37 (0.12 to 1.17) |
| Non-access-site bleeding | 2.0% | 9.9% | 0.20 (0.04 to 0.92) |
Strengths
- Protocol published in advance, DSMB oversight and blinded event adjudication.
- CMR substudy built to test the mechanism.
- The authors state the study is a pilot and exploratory.
- The bleeding reduction points the same way as TICO and TARGET-FIRST.
Methodological issues
- Uninformative MACCE interval. HR 1.04 (0.26 to 4.17) with about 4 events per arm. The minimum detectable effect is around HR > 7. A 4-fold increase in ischaemic risk could go undetected. Concluding “no difference in ischaemic events” reads absence of evidence as evidence of absence.
- Fragile bleeding finding. HR 0.20 (0.04 to 0.92) has a fragility index of 1. It is the only significant result among many secondary endpoints.
- No Kaplan-Meier curves or Schoenfeld test. Cox HRs are reported without curves. Bleeding hazard clusters in months 1 to 3 and ischaemic hazard spreads across the year. Proportional hazards remain an untested assumption.
- CMR signal favouring DAPT. In 51 patients, 7 of 8 structural readouts favour DAPT: infarct size (15.4% vs 9.9% of LV), IMH presence (42% vs 32%), microvascular obstruction presence (44% vs 35%), LVEF (47% vs 53%, p = 0.06), LV mass index (p = 0.04), LV end-systolic volume index and LV thrombus (22% vs 13%). The exception is microvascular obstruction size (3.6% vs 4.4% of LV), with heavily overlapping IQRs. By binomial test, 7 of 8 in the same direction gives p ≈ 3.5%.
- Incomplete explanation by the authors. The authors attribute the signal to chance and to universal aspirin loading. VerifyNow shows 58% of the monotherapy arm still had full aspirin inhibition at days 2 to 4, when IMH expands. That would explain similar arms, not worse numbers in the monotherapy arm. Another reading: the design cannot test the hypothesis it was built for.
- Highly selected population. Killip I in 98.5%, final TIMI 3 in 91.5%, no shock, no severe CKD, no oral anticoagulation, no recent MI. Five high-volume Dutch centres with 97% radial access.
Opinion
As a pilot, STOP-IMH does its job: it shows the design is feasible and raises the question of whether immediate aspirin omission holds up mechanistically. It does not inform practice.
I disagree with the clinical translation in the abstract, discussion and “Impact on Daily Practice” section. Higher-risk patients, where the bleeding-ischaemia trade-off matters most, were not in the trial. The CMR signal deserves a more cautious read. To me, it points against monotherapy.
NEO-MINDSET STEMI and STOP-DAPT 3 signalled trouble with early aspirin withdrawal, and STOP-IMH points the same way. COMPARE STEMI ONE should answer the question. Until then, the reading the methods section gives holds: a hypothesis-generating study.