Genotype First or Treat Empirically: Choosing the Second Antiplatelet by CYP2C19 Status
A clopidogrel prodrug that never gets activated protects nobody — a genetic test can catch that in advance, but only if the result comes back before the decision has to be made, and the alternative of treating everyone with the stronger drug carries its own real cost.
C.Y., a 52-year-old man who owns a small landscaping company and is, by his own description, "never not moving," had a brief episode of right facial numbness and word-finding difficulty while loading his truck at 6:45am, resolved within twenty minutes but recognized by him as serious enough to call his wife rather than continue his route. He arrived at the emergency department within an hour of symptom onset, his exam already normal, but a small area of diffusion restriction on MRI confirmed a true minor ischemic stroke rather than a mimic. He has hypertension managed on amlodipine and no history of prior stroke, TIA, or bleeding.
His ABCD2 score comes out to 3 — one point for speech disturbance without weakness, one for a duration between ten and sixty minutes, one for an elevated triage blood pressure, and nothing for age at 52 or for diabetes he doesn't have. That lands in the score's low-to-moderate band, and it is the wrong instrument for him anyway: ABCD2 estimates stroke risk after a transient attack, and his diffusion restriction means he has already had one. Minor ischemic stroke is its own entry criterion in CHANCE and POINT, and it is the one he meets, so the team is planning short-course dual antiplatelet therapy on that basis rather than on the score. The remaining question is which second agent to pair with aspirin. His background is East Asian, a population in which CYP2C19 loss-of-function allele carriage runs substantially higher, roughly 50-60%, than in white populations, where it runs closer to 25-30% — a real, population-level difference in how likely clopidogrel, a prodrug that requires CYP2C19-mediated hepatic activation to work at all, is to actually function as intended for any given patient before it's even started. CHANCE-2 tested this directly: in patients confirmed as CYP2C19 loss-of-function carriers by genotyping, ticagrelor plus aspirin reduced recurrent stroke compared with clopidogrel plus aspirin, without an increase in major bleeding in that carrier population. Point-of-care genotyping is available at this hospital but typically takes several hours to return a result — not instantaneous, but well within the treatment window for a decision that doesn't need to be made in the first few minutes the way a thrombolysis decision does.
In the stroke unit, choosing the second antiplatelet
I'd send the CYP2C19 point-of-care panel now and hold the second agent's specific choice until it results. CHANCE-2 randomized confirmed loss-of-function carriers directly to ticagrelor-plus-aspirin versus clopidogrel-plus-aspirin and found reduced recurrent stroke with ticagrelor in that specific population, without an increase in major bleeding. That's a genotype-matched trial answering a genotype-specific question — about as directly applicable as evidence gets.
The practical problem is turnaround time. Point-of-care genotyping here isn't instantaneous, and if we're starting dual therapy today regardless, why not just start ticagrelor empirically and skip the wait? THALES already established ticagrelor's own evidence base independent of genotype, against aspirin alone — it doesn't need a test result to justify using it.
That's true as far as it goes, but be careful what THALES actually compared. Its control arm was aspirin alone, not clopidogrel — so it tells us adding ticagrelor to aspirin raised severe bleeding from 0.1% to 0.5% and intracranial hemorrhage from 0.1% to 0.4%, but it never weighed ticagrelor against the drug you'd otherwise be giving. For that comparison you have to go back to CHANCE-2, which found no difference in severe bleeding between the two but more total bleeding events with ticagrelor. For a patient who turns out NOT to be a loss-of-function carrier — still the majority even in his higher-prevalence population — clopidogrel would have worked completely normally, and empiric ticagrelor exposes him to that extra bleeding risk for no benefit at all.
Genotyping's real value isn't just catching who needs the stronger drug — it's sparing everyone else from a real risk they didn't need to take. Given his population background alone puts him at meaningfully higher pretest probability of being a carrier, I'd genotype and bridge with clopidogrel until the result comes back, switching to ticagrelor only if confirmed.
Agreed: send point-of-care CYP2C19 genotyping now, start clopidogrel plus aspirin empirically as the bridging regimen, and switch clopidogrel to ticagrelor if the result confirms a loss-of-function allele. Standard 21-day dual-therapy duration planned regardless of which second agent he ends up on.
Not agreed: whether ancestry-informed pretest probability should shorten the empiric-clopidogrel bridge further in specific populations — genuinely raised, not resolved, and left as an institutional practice question rather than settled for this one patient.