Ch 28 (s28.13); Ch 22 (s22.3); Ch 27
Refines Chapter 22's null finding rather than contradicting it: the peri-cannulation carbon dioxide signal is strongest for haemorrhage and weakest for brain injury as a composite. Supports recording the pre-cannulation carbon dioxide so the ELSO 50 percent ceiling has a denominator.
Patients with versus without acute brain injury, and with versus without intracranial haemorrhage
DOI 10.1097/MAT.0000000000001699
Peri-cannulation change in arterial carbon dioxide tension, calculated as the mean of all values in the 12 hours before minus the mean of all values in the 12 hours after cannulation (median 11 values per patient, IQR 8 to 14)
Single centre, retrospective, 129 patients. The association is with intracranial haemorrhage, not with acute brain injury as a composite — an important distinction that explains why a large ELSO extracorporeal cardiopulmonary resuscitation study of 3,125 patients found no significant association with brain injury overall.
Pre-cannulation (39 versus 42 mmHg, p=0.38) and post-cannulation (37 versus 36 mmHg, p=0.82) values did not differ between those with and without acute brain injury. However higher pre-cannulation carbon dioxide (odds ratio 2.10, 95 percent CI 1.10 to 4.00, p=0.02) and a larger peri-cannulation fall (odds ratio 2.69, 1.18 to 6.13, p=0.02) were associated with intracranial haemorrhage specifically.
129 adult venoarterial ECMO patients at a tertiary centre with standardised neuromonitoring (median age 60, 63 percent male). Acute brain injury in 43 (33 percent), defined to include intracranial haemorrhage, ischaemic stroke, hypoxic-ischaemic brain injury, cerebral oedema, seizure and brain death.