Dual Antiplatelet Therapy: A Bifurcated Stent and a Bleed Three Months Old
A single patient, thirty days from a two-stent bifurcation PCI. The disagreement isn't about whether she still needs dual antiplatelet therapy — it's about which half of the regimen she can no longer safely continue, and for how much longer.
D.F., a 68-year-old woman, spends her Tuesday and Thursday mornings shelving returns at the public library two towns over from the house she's lived in alone since her husband died two years ago — a twenty-five-minute drive she still makes herself, in her own car, without anyone reminding her to go. She has carried hypertension for twelve years, managed on lisinopril and amlodipine, alongside a mild, stable decline in kidney function her primary care physician has watched rather than treated. Three months ago that quiet baseline was interrupted: she woke to painless rectal bleeding, was found on colonoscopy to have a bleeding diverticulum, and needed two units of packed red cells before it settled on its own, without any endoscopic or surgical intervention. Her hemoglobin has sat at 11.2 g/dL ever since — mild anemia by any standard definition, and squarely inside the band that counts as a minor bleeding-risk criterion in a woman — and the diverticulum itself was never treated. It is still there, the same tissue that bled once already.
Three days of exertional chest pressure that finally didn't resolve with rest brought her back in, and a troponin rise confirmed an NSTEMI. Catheterization found two-vessel disease, with the culprit lesion sitting at a calcified bifurcation of the LAD and its first diagonal branch — anatomy that could not be covered with a single stent. She left the lab with two drug-eluting stents and a two-stent bifurcation result, the kind of PCI that carries a materially higher early risk of stent thrombosis than a single, straightforward vessel would. That result now sits next to the other fact in her chart: by formal ARC-HBR criteria, she already qualifies as high bleeding risk outright. The transfusion-requiring bleed inside the past six months is a major criterion on its own; the persisting mild anemia and the moderate chronic kidney disease each independently add a minor criterion on top of it. Two findings, both real, both hers, pulling the antiplatelet decision in opposite directions at once.
Thirty days out, at the first follow-up
I put two stents across that bifurcation because a single stent would have left her with a real risk of restenosis or acute closure at the side branch, and that decision doesn't stop mattering the day she walks out of the lab. Two-stent bifurcation results carry a demonstrably higher early stent thrombosis rate than single-vessel PCI, and stent thrombosis, when it happens, kills close to a third of the patients it happens to. The trials behind the short-DAPT, high-bleeding-risk pathway — MASTER DAPT above all — did enroll patients like her: roughly a quarter of that trial met complex-PCI criteria, bifurcation with two stents among them, and its prespecified complex-PCI analysis found no interaction. What I'd press on is that the investigators themselves called that analysis underpowered and hypothesis-generating. A trial powered for its overall population doesn't become powered for her stent geometry just because her stent geometry was represented in it.
If this had been a single, uncomplicated stent in a straightforward vessel, I would be arguing the opposite side of this conversation. The concern here is specific to what that bifurcation actually required, not a general resistance to shortening DAPT in bleeding-risk patients.
She isn't borderline high bleeding risk — she's high bleeding risk twice over. A bleed requiring transfusion inside six months is a major ARC-HBR criterion by itself, full stop, and her hemoglobin and kidney function each add a minor criterion on top of a call that was already made. The diverticulum that bled hasn't been treated; it's still sitting in her colon, and nothing about dual antiplatelet therapy makes it less likely to bleed again. A second transfusion-requiring bleed in a 68-year-old living alone is not a trivial event to weigh against a stent thrombosis risk that, however real, is still a minority outcome even after complex PCI.
I'm not asking to stop antiplatelet therapy entirely — nobody is arguing that. The question is only which single agent she keeps carrying past thirty days, and the data on early aspirin discontinuation after DAPT, with the P2Y12 inhibitor continued alone, doesn't show a meaningful rise in ischemic events even in higher-risk PCI cohorts, once she's past the highest-risk early window.
Both of you are actually arguing about the same thirty-to-ninety-day window from opposite ends of it, and the pharmacology doesn't require picking one of you to be wrong. Keep her on aspirin through this first month — it covers exactly the early, bifurcation-specific stent thrombosis risk that concerns the interventional side, and one month is not the same exposure as three. But most of what dual therapy adds beyond a single P2Y12 agent is aspirin's own contribution: irreversible COX-1 inhibition across the entire circulating life of every platelet, plus a direct, non-platelet mucosal effect on a colon that has already bled once. Clopidogrel alone still blocks P2Y12-mediated platelet aggregation for that same platelet lifespan; it doesn't carry aspirin's separate GI toxicity. Dropping aspirin first, rather than dropping either drug at random or defaulting to a fixed guideline number, is the choice that actually tracks where each drug's risk and benefit live.
One more thing worth deciding now rather than later: if a proton pump inhibitor goes on the discharge list for gastric protection, it should be pantoprazole, not omeprazole. Omeprazole is a strong CYP2C19 inhibitor, and clopidogrel needs CYP2C19 to convert to its active metabolite — pairing the two would blunt exactly the drug we may be about to ask to do the antithrombotic work on its own. Pantoprazole doesn't share that interaction.
Agreed: aspirin and clopidogrel both continue to the thirty-day mark; pantoprazole, not omeprazole, if gastric protection is added; hemoglobin, creatinine, and a direct bleeding history are checked at that visit alongside a routine ischemic-symptom review.
Not agreed, and the reason the one-month visit carries a hard branch point rather than a single expectation:
Aspirin stops; clopidogrel continues alone, indefinitely.
Aspirin continues to at least ninety days, given the bifurcation result, before the same question is revisited.
The interventional cardiologist and the hematologist left the visit with different defaults for what a clean thirty-day check-in should mean; the branch point exists so that disagreement resolves on a fixed date, rather than by whichever specialist she happens to see next.