P2Y12 Inhibitor Selection After NSTEMI: How Much Weight Does a Healed Ulcer Still Carry?
A single patient, less than an hour past a stented NSTEMI, whose chart already carries one closed chapter of GI bleeding. The disagreement isn't about whether he needs a second antiplatelet on top of aspirin — it's about which one is safe to add given his bleeding history.
P.N., a 66-year-old man, spent thirty-four years walking a postal route before retiring two years ago, and now fills three mornings a week stocking shelves at the food pantry two blocks from his house. He has treated hypertension and a decade-long cholesterol history on lisinopril and atorvastatin, and quit a half-pack-a-day habit eight years ago after a bad bout of bronchitis sent him to urgent care. Until this morning he had no cardiac history at all. That changed a few hours ago at the pantry, when he set down a case of canned vegetables and felt a chest pressure that didn't ease with rest; it returned twice more before he let a coworker call 911 that afternoon. Troponin was rising by the time he reached the emergency department, the ECG showed new anterior ST depressions, and catheterization forty-five minutes ago found a 90% mid-LAD stenosis — now open, with a single drug-eluting stent.
His chart already carries one other hospitalization, nine months back: a bleeding duodenal ulcer, found on endoscopy after several days of dark stools, treated with eradication therapy for the H. pylori it grew on biopsy and two units of blood, confirmed cleared on a follow-up scope three months later. He's taken no NSAIDs since, and his hemoglobin has held steady in the low 12s at every visit since — not still bleeding, just permanently a point or two under where it used to sit. The stent he now carries is a surface his own platelets will try to cover with clot until the vessel wall grows over it, and a P2Y12 inhibitor alongside aspirin exists to prevent that — but the three available drugs don't do the job identically, and the difference matters more in a patient like him. Clopidogrel needs two sequential liver conversions through CYP2C19 before it activates, and roughly three in ten white patients carry at least one reduced-function copy of that gene — enough that a real minority look adequately treated on paper while staying under-protected in practice. Ticagrelor and prasugrel skip that dependency and inhibit platelets more reliably, at the measured cost of more bleeding. His own chart clears two, not one, of the criteria a bleeding-risk framework built for exactly this decision uses to flag a patient before any drug is chosen: a spontaneous bleed requiring transfusion within the past year, and a hemoglobin that never fully recovered. Whether that should push the team toward the gentler drug, the more effective one, or a third option that tries to find out which he actually needs first, is what they're now working out at his bedside.
At the bedside, less than an hour after the stent
Start ticagrelor today, not clopidogrel, and don't wait on a genotype result to make that call. He's forty-five minutes past a drug-eluting stent to his mid-LAD — this is exactly the window where under-protection costs the most, and clopidogrel's problem isn't a small-print footnote: something like three in ten white patients carry a reduced-function CYP2C19 allele, and we won't know which kind he is for days. PLATO put ticagrelor against clopidogrel across eighteen thousand ACS patients, and it didn't just reduce ischemic events — it reduced all-cause mortality, without a significant difference in overall major bleeding.
I'd read this differently if his ulcer were still active, or if this were a second bleed rather than one closed episode nine months ago. It isn't, and that distinction is doing real work in my reasoning, not being waved past.
I'd read his chart differently. He clears two of the criteria the bleeding-risk consortium built specifically to flag a patient like him before the drug gets chosen — a transfusion-requiring bleed inside the past year, and a hemoglobin that's never come back to where it was. That's the framework doing exactly the job it exists to do. "No significant difference in overall major bleeding" in PLATO still sits on top of a real increase in non-CABG bleeding underneath it, and his own denominator isn't the trial's average enrollee. Start clopidogrel, send the genotype today if this center can turn it around inside twenty-four hours, and escalate to ticagrelor only if the result actually says he needs it.
I'm not disputing that a poor metabolizer left on clopidogrel for even a day or two carries real risk too. I'm disputing that the risk only runs in one direction, which is how the ticagrelor-first argument tends to get made.
You're both right about different halves of the same problem, and the actual constraint is turnaround time, not a real disagreement about the pharmacology. If genotyping came back in an hour, I'd side with starting clopidogrel and escalating only the roughly one in three patients who turn out to need it. It doesn't come back in an hour here — realistically two to three days — and stent-thrombosis risk is highest in exactly that window. So: load and start ticagrelor today, send the genotype in parallel, and use pantoprazole rather than omeprazole for the PPI, since if the result does bring clopidogrel back into play later, omeprazole's own interaction with CYP2C19 would undercut it before it started.
What we haven't agreed on is what happens once that genotype is back. A normal-metabolizer result would mean clopidogrel was always going to be adequate for him — is that a reason to switch him off ticagrelor once he's past the highest-risk early window, or a reason to finish the twelve months on the drug that already has a demonstrated mortality benefit in a trial population that was never genotyped in the first place? The genotype-guided de-escalation data that exist were built in STEMI patients going straight to primary PCI — not his presentation — so applying that evidence here is an extrapolation of the principle, not a direct match to it.
Agreed within minutes: aspirin 81 mg daily, ticagrelor 180 mg load then 90 mg twice daily, pantoprazole 40 mg daily — and a CYP2C19 genotype sent today rather than waited on before starting DAPT.
Not agreed, and the reason the plan carries a hard branch point rather than a single expectation:
Once he's past roughly the first month and the genotype confirms normal CYP2C19 activation, switch to clopidogrel and accept the bleeding-safety trade for the remainder of the year.
PLATO's mortality benefit was shown in a population that was never stratified by genotype. Finishing the studied regimen — not improvising a switch strategy no trial has actually tested in an NSTEMI population — is the more defensible read of the evidence as it stands.
Gastroenterology and cardiology left the room with different priors about what that genotype result should mean once it's back. Nobody set an expectation for which reading is more likely correct — only that the result, whichever it says, gets discussed together before any change is made, not enacted unilaterally by whichever service happens to see it first.