General Anesthesia or Sedation for the Thrombectomy Suite
A retired choir director's large-vessel stroke is thrombectomy-eligible, but she can't hold still and won't stop trying to sit up. The anesthesia team has to pick an approach the interventionalist can actually work through.
B.C., a 68-year-old woman, directed her church choir for over twenty years before retiring last spring, and was found by her daughter unable to speak and unable to move her right side, last known well two and a half hours earlier. Imaging confirmed a left M1 occlusion with a moderate, favorable penumbra — a genuine thrombectomy candidate — but she is agitated on the table, repeatedly trying to sit up despite redirection, poorly cooperative with the positioning the procedure requires. The choice in front of the team is not simply comfort versus discomfort; a patient who can't hold still during catheter navigation through an M1 vessel risks a longer procedure, a higher complication rate, or an incomplete recanalization, any of which would cost her more than either anesthesia approach itself.
Older observational data linking general anesthesia to worse thrombectomy outcomes shaped a generation of practice toward conscious sedation as the default, but three more recent randomized trials — SIESTA, AnStroke, and GOLIATH — each tested general anesthesia against conscious sedation directly, under a protocol that held mean arterial pressure tightly controlled throughout, and none found general anesthesia worse than sedation on its primary endpoint, with GOLIATH and SIESTA both trending favorably for the general anesthesia arm on secondary measures. None was built as a formal non-inferiority trial, and all three were single-center, so what they establish is the absence of the harm the observational data had predicted rather than a positive demonstration of equivalence. The consistent read across all three is that the earlier observational signal against GA reflected periprocedural hypotension in practice at the time, not general anesthesia as a mechanism — a population-level lesson that applies directly to her, since her agitation itself, not a theoretical anesthesia risk, is the more immediate threat to a clean, timely recanalization.
Her daughter, still in the room despite the team's efforts to move family to the waiting area, keeps repeating that her mother was singing in the car twenty minutes before this happened — a detail that carries no clinical weight but that the team has learned not to rush past, since it's often the only version of the patient a family has to hold onto while a procedure this time-pressured unfolds around someone who can no longer speak for herself.
Stillness versus speed to groin puncture
Every minute we spend intubating is a minute the penumbra isn't being reperfused. Her gag reflex is intact and she doesn't have an aspiration event yet. I'd rather go with conscious sedation and manage her agitation with medication titrated at the table than commit to a general anesthesia induction before we've even tried.
I understand the time pressure, and I'm not dismissing it — but three separate randomized trials, SIESTA, AnStroke, and GOLIATH, tested exactly this question under protocols that kept her blood pressure controlled throughout, and none of the three found general anesthesia worse on its primary endpoint, with two favoring it on secondary measures — they weren't formal non-inferiority designs, but the harm the old observational data predicted simply didn't appear. The older data against GA reflected hypotension in practice, not the anesthesia itself. Given how agitated she is right now, I think GA with a tight pressure protocol is actually the faster path to a clean procedure, not the slower one.
We don't have to fully commit before we know how she responds. Attempt a controlled sedation load with a low threshold to convert to general anesthesia the moment positioning or catheter navigation is genuinely compromised — that gives the interventionalist the fast path if it works, and gives us the trial-supported fallback immediately if it doesn't, rather than deciding the whole case on a prediction of how cooperative she'll be.
Sedation was attempted first but converted to general anesthesia within eight minutes when repositioning attempts failed to control her movement. A tight blood-pressure protocol was maintained throughout, and complete recanalization was achieved with no periprocedural hypotension.
The interventionalist's time-pressure concern was real but ultimately cost less than feared — the conversion added roughly ten minutes to door-to-puncture time, well within the procedural window, and the team agreed the trial-supported protocol made general anesthesia the safer choice once conversion became necessary rather than a compromise accepted reluctantly.