22
The causal anchor for treating the initial sweep gas setting as a neuroprotective decision rather than a default. It removes the confounding objection that defeats the registry literature, because the animals were identical and sweep gas was the only variable.
Control (sweep at 100% of blood flow) and slow correction (25% of blood flow; about 10 mmHg over 10 minutes)
DOI 10.1007/s12975-025-01376-8
Rapid carbon dioxide correction (sweep gas at 200% of ECMO blood flow; PaCO2 60 to about 30 mmHg within 5 minutes)
Six animals; anaesthetised porcine model; abstract only; the CA-3 result contradicts the other three regions; no human dose-response established.
Rapid correction impaired cerebral autoregulation at every phase (pressure reactivity index 0.50 versus 0.11 at 10 to 20 minutes) and produced more ischaemic neuronal injury in caudate (43.1 vs 10.6%), putamen (66.6 vs 23.9%) and temporal cortex (34.9 vs 8.9%). NUMERICAL FLAG: hippocampal CA-3 is reported as 4.7% versus 18.0%, the opposite direction, and is not corrected here.
Six pigs; 15 minutes untreated ventricular fibrillation, then extracorporeal support, return of circulation attempted at 20 minutes