How Soon Is Too Soon: Anticoagulation Timing After a Cardioembolic Stroke, By Infarct Size
Two patients, both newly diagnosed with atrial fibrillation after an ischemic stroke, land on opposite ends of a real trial's own confidence range — one whose infarct size argues for real caution even under the newest evidence, one whose imaging clears every bar for starting early.
M.T., a 76-year-old retired librarian, spent four decades cataloguing the county system's rare-book collection and still volunteers Tuesdays sorting donations. She was found by a neighbor slumped in her reading chair, unresponsive on her right side and unable to speak, sometime between 1:00pm when the neighbor saw her watering plants on the porch and 2:40pm when she was discovered. EMS brought her in with an NIHSS of 16 — dense right hemiplegia, global aphasia, and a right gaze preference. She has no prior stroke history and no known atrial fibrillation before today; her only regular medication is a multivitamin.
CT angiography at admission showed a left M1 occlusion, and she was outside the thrombolysis window given the uncertain onset; mechanical thrombectomy achieved complete reperfusion. Telemetry during her admission captured new-onset atrial fibrillation, and follow-up imaging on hospital day 3 confirms a large left MCA territory infarct — well over a third of the vascular territory — with no hemorrhagic transformation on the most recent scan. By the classic teaching heuristic for anticoagulation timing after cardioembolic stroke, an infarct this size would traditionally mean waiting roughly two weeks before starting an anticoagulant, out of concern that reperfused, infarcted tissue this extensive carries a meaningfully elevated hemorrhagic-transformation risk if anticoagulated too early. ELAN, the newest randomized trial on this question, tested earlier initiation directly against that kind of delayed approach — for major infarcts specifically, its early arm started a DOAC at day 6 or 7 rather than waiting the full two weeks — and found no significant difference in its composite outcome, importantly without an increase in symptomatic intracranial hemorrhage in the earlier-treated group. Her own numbers put her exactly in the population that comparison was built around: large enough to have been considered high-risk under the old rule, and now sitting inside a real, if modestly sized, trial's evidence that the old rule may be more cautious than the data requires.
I'd anchor to ELAN's own major-infarct arm rather than the older 1-3-6-12-day heuristic — start her DOAC at day 6 or 7 if the day-5 or -6 scan stays clean, not day 14. ELAN randomized early against later initiation across a range of infarct sizes and found no significant difference in its composite outcome, and critically, no increase in symptomatic intracranial hemorrhage in the earlier-treated group, even among the larger infarcts it enrolled.
You're right about ELAN's headline result, and I'm not arguing against earlier initiation in principle — I just want to flag that the major-infarct subgroup wasn't the trial's best-powered stratum, and the confidence interval around that specific comparison is wide. A trial that's reassuring on average isn't the same as a trial that's precisely powered for her exact infarct size.
I'd want a clean repeat scan close to the actual start date, not just the day-3 scan we already have, before committing to day 6 or 7 — individualized timing off real imaging, not the calendar alone.
Worth weighing the other side of that delay too: her annualized stroke risk from untreated atrial fibrillation isn't zero while everyone waits for reassurance. Every extra day without anticoagulation is a day of real, competing embolic risk, which is part of why ELAN was designed in the first place — the old heuristic wasn't chosen to eliminate risk, it traded one risk for another without ever being tested against the alternative directly.
Agreed: repeat imaging planned for day 5-6 specifically to confirm no interval hemorrhagic transformation, with apixaban starting day 6 or 7 if that scan is clean — following ELAN's early-arm timing for major infarcts rather than the older two-week rule. Aspirin continues as bridging coverage until then.
Not agreed: what to do if the day 5-6 scan shows asymptomatic petechial hemorrhagic transformation rather than a clean result — a real, plausible outcome ELAN's own protocol doesn't specify a clean answer for, left as a decision to make when and if the scan actually shows it rather than pre-committed to now.
J.R., a 64-year-old man, still rides his bicycle most mornings before his commute to the architecture firm where he's worked for over thirty years, and was mid-ride when he noticed his left hand fumbling the brake lever in a way that "just felt wrong." He pulled over, texted his wife rather than calling 911 at first — "in hindsight, not my finest decision," he says — and was eventually brought in by his wife about two hours after symptom onset. His exam on arrival showed mild left hand weakness and clumsiness, NIHSS 2, with no facial droop, no speech involvement, and no gait instability.
MRI confirmed a small, cortically-based embolic-appearing infarct in the right precentral gyrus, consistent with the focal hand weakness, without the diffuse small-vessel pattern that would suggest a lacunar rather than embolic mechanism. Telemetry during his brief observation admission captured paroxysmal atrial fibrillation, previously undiagnosed; he has hypertension on amlodipine and no other known cardiovascular disease. By the classic timing heuristic, a minor infarct like his would already support starting anticoagulation within the first few days; ELAN's own early-treatment arm for minor and moderate infarcts is even more aggressive than that older rule, initiating a DOAC within 48 hours in appropriately selected patients, and this is the trial's best-powered stratum — the subgroup where the confidence around "no increase in symptomatic hemorrhage with earlier initiation" is tightest. His own infarct, small, cortical, without hemorrhagic transformation on either of two scans since admission, sits about as cleanly inside that reassured population as a real patient's imaging ever does.
Start apixaban within 48 hours. This is exactly ELAN's best-powered stratum — minor infarct, no hemorrhagic transformation on two separate scans — and the trial's own early-initiation arm for patients like him went even sooner than the classic day-1-to-3 heuristic already would have. I don't see a reason to wait past 48 hours here.
Agreed, and I'd add the competing-risk argument directly: he's had a real embolic event already, from atrial fibrillation we now know is present, and every additional day without anticoagulation is another day of that same mechanism potentially firing again. Nothing about his imaging argues for caution here.
Compare this directly to the patient we just discussed with the large MCA infarct, where I wanted a fresh scan close to the actual start date before committing — his situation is genuinely different, not just a smaller version of the same caution. Small, cortical, clean on two scans, and sitting in the trial's best-powered, most-reassured subgroup rather than its thinnest.
I'm comfortable with 48 hours here specifically because the imaging and the trial data actually line up cleanly, which is exactly the distinction that made me want more caution in the larger-infarct case.
Agreed: apixaban started within 48 hours of admission, given the clean, small, cortical infarct pattern on two separate scans and his position in ELAN's best-powered early-treatment subgroup. Amlodipine continued unchanged.
Not agreed on anything load-bearing for this patient specifically — the earlier disagreement in this case set belonged entirely to the larger-infarct patient; here, imaging and trial data pointed the same direction for all three voices.