Warmer bypass meant fewer strokes in emergency aortic surgery. Then the authors adjusted for confounders, and the four-band story fell apart. One number held.
The Research
Chan JTN, Fazmin IT, Vokshi I, de Silva R, Farid S. Impact of hypothermia level on neurological and renal outcomes in emergency aortic surgery: a single-centre experience. Perfusion, 2026. doi:10.1177/02676591261500968. · PubMed 42911832
Warmer bypass meant fewer strokes in emergency aortic surgery. Then the authors adjusted for confounders, and the four-band story fell apart. One number held.
Reviewed 175 consecutive patients having emergency surgery for acute aortic syndrome at one center between 2018 and 2023. Median age 64, 37 percent women, 89 percent having a hemiarch. Split them into four temperature bands: deep at 20 degrees or below, low-moderate to 24, high-moderate to 28, and mild above 28.
In hemiarch patients, stroke ran 27 percent in deep, 26 percent in low-moderate, 8.8 percent in high-moderate, and 11 percent in mild, p of 0.049 before any adjustment. Bypass time fell steadily with warmth, from 270 minutes deep to 164 minutes mild. Overall 30-day mortality was 18.9 percent and overall stroke 19.4 percent. After adjustment for confounders, the four-band comparison did not hold. A sensitivity analysis at a single 24.0 degree threshold did, with stroke odds 0.23, interval 0.08 to 0.64, p of 0.005.
The Hallway
Once you allow for the fact that the sickest patients probably got the coldest runs, the four groups stop looking different.
Palm TreeHowever, multivariate regression showed that the four-category comparison did not reach significance for the assessed outcomes once adjusted for confounding variables.
A real question with a hard outcome, stroke, in a population where the temperature decision gets made at two in the morning. The bypass-time gradient is coherent, because colder cases ran a hundred minutes longer, and that entanglement is the actual clinical problem rather than a statistical nuisance. The authors also did the thing that kills tidy stories: they adjusted, and they reported that the tidiest version died.
Retrospective, one center, 175 patients split four ways, which leaves bands small enough to swing on a few events. The gradient is not even monotonic, since high-moderate at 8.8 percent beat mild at 11 percent, so warmer was not uniformly better. Temperature was not randomized and the sickest patients most likely went coldest. And a single threshold found by sensitivity analysis inside the same dataset it is reported from is a hypothesis, not a validated cut-point.
The InFlow-mation
Critical Conclusion
The four-band story did not survive adjustment. A single binary at 24 degrees did, with stroke odds of 0.23, which is a large effect to take from a small retrospective sample.
Take It Back to the Pump
Keep a warm hemiarch protocol, because the direction agrees with the EACTS guidance already sitting in your room. Do not build a cut-point rule on one center sensitivity analysis. If someone quotes 24 degrees, agree on the direction and ask what the next study would have to enroll.
Pump off.
On the source
Peer reviewed, single-center retrospective review covering 2018 to 2023. Not randomized, 175 patients, four temperature bands. The 24.0 degree threshold is a sensitivity analysis within the same cohort, so treat it as hypothesis generating. Emergency acute aortic syndrome only.