Summary: In a post hoc analysis of the TALOS-AMI trial, switching from ticagrelor to clopidogrel one month after heart attack was associated with benefit consistent across diabetes status, but within the diabetic subgroup the individual bleeding and ischemic differences did not reach statistical significance.
PICO Summary
| Element | Detail |
|---|---|
| Population | 859 patients with diabetes (31.9% of the trial) who tolerated 1 month of ticagrelor-based DAPT after MI and PCI; post hoc analysis of the TALOS-AMI RCT, South Korea. |
| Intervention | Unguided de-escalation to clopidogrel-based DAPT for the remaining 11 months. |
| Comparison | Continued ticagrelor-based DAPT. |
| Outcome | De-escalation was associated with a lower primary composite endpoint (mainly fewer BARC type 2 bleeds) consistently regardless of diabetes status (interaction p=0.467). Within the diabetes cohort, the differences in bleeding (HR 0.48; 95% CI 0.22–1.05; p=0.066) and ischemic (HR 0.57; 95% CI 0.25–1.29; p=0.176) endpoints were not statistically significant. Significant treatment-by-diabetes interactions were seen for BARC 3 or 5 bleeding (p=0.042) and target vessel revascularisation (p=0.014). |
Antiplatelet De-Escalation in Diabetes
Post hoc RCT · MI + diabetes · 11 months
Within the diabetic subgroup, both bleeding (HR 0.48) and ischemic (HR 0.57) hazard ratios favoured de-escalation but crossed 1.0, so neither reached significance. The benefit was consistent across diabetes status rather than independently proven in diabetes.
Expert Commentary
This is a clinically relevant analysis of an attractive idea, that ischemic risk is front-loaded after a heart attack while bleeding risk persists, so stepping down to clopidogrel after the first month might improve the net balance, and it usefully focuses on diabetic patients who are often deemed too high-risk to de-escalate. The reassuring finding is that the de-escalation benefit seen overall did not differ by diabetes status, with no statistical interaction, so diabetic patients did not appear to fare worse from stepping down. The honest qualifier, which the post should foreground, is that within the diabetic subgroup the individual bleeding and ischemic differences did not reach significance, at p of 0.066 and 0.176, so this is consistency rather than independent proof of benefit in diabetes, and it is a post hoc subgroup of a trial not powered for it. CYP2C19 status, which strongly affects clopidogrel response, was not addressed. Can I use this with my patients? As supportive context for shared decisions. For a stabilised diabetic patient a month after MI, especially with bleeding risk or ticagrelor intolerance, de-escalation is a reasonable option, while I would avoid it in very high ischemic-risk anatomy and treat this evidence as hypothesis-refining rather than definitive.
References
Kim SH, Lee KY, Byeon J, et al. De-escalation dual antiplatelet strategy in stabilized myocardial infarction patients with diabetes mellitus. JACC Cardiovasc Interv. 2025;18(22):2713–2724. doi:10.1016/j.jcin.2025.09.026