SAN FRANCISCO – In patients on extracorporeal membrane oxygenation (ECMO) support, a 25% drop in regional oxygen saturation index (rSO2) value on a bedside near-infrared spectroscopy (NIRS) system is a potential indicator of acute brain injury (ABI), according to a new pilot study. Patients on ECMO are in dire straits, with about a 50% survival rate, but the odds become much worse in the presence of brain injury – the fatality rate rises to about 90%. In addition, ABIs and other neurological complications are common in ECMO patients, occurring in about half.
Currently, clinicians must rely on clinical signs to pinpoint and address emerging injuries, but a bedside NIRS system could flag problems for earlier intervention. “Changes in brain oxygenation could prompt clinicians to do earlier investigations of what’s going on, and maybe they can intervene and reduce the extent of the injury,”, director of research at Carolinas Medical Center, Charlotte, N.C., said in an interview. Dr. Huynh comoderated the session at the annual clinical congress of the American College of Surgeons, where the research was presented. The study was “very novel,” Dr. Huynh added.
NIRS monitors oxygenation noninvasively. Its rSO2 value balances cerebral oxygen delivery and oxygen consumption, reflecting aerobic metabolism, and therefore has the potential to reveal a patient’s status. A previous study () showed that a drop in rSO2 of more than 25%, or absolute values below 40, could predict ABI in ECMO patients, but the work was done retrospectively and is subject to the usual limitations.
The current work, which was presented by medical student Megan Hunt of Johns Hopkins School of Medicine, Baltimore, sought to determine the potential value of NIRS in a prospective study, as well as to determine the interpretation of NIRS that had the best predictive value.
The researchers enrolled 47 consecutive ECMO patients (39 venoarterial, 8 venovenous). Twenty-seven of 39 (69%) of venoarterial ECMO patients, and 6 of 8 (75%) of venovenous ECMO patients had a desaturation event; 53.2% of patients overall experienced an ABI – a number that was in line with previous studies. The researchers found that ABI risk factors such as atrial fibrillation, a lower pre-ECMO Glasgow Coma Score, and more blood transfusions were linked to greater odds of ABI. “Generally, more critically ill patients [had] acute brain injuries,” Ms. Hunt said during her talk.
A total of 65.6% of patients who had an ABI registered a 25% or more drop in rSO2 from baseline, while 26.7% had no change (P = .03). No other measure showed a statistically significant association. A multivariate regression model of ABI variables found that a 25% decline in rSO2 from baseline predicted ABI (adjusted odds ratio, 7.6; P = .03), which outperformed APACHE II (aOR, 1.2; P = .002). A receiver operating characteristic (ROC) curve showed that a 25% drop from baseline had a sensitivity of 85% and a specificity of 55% for ABI.
The study is limited by its small sample size, the fact that it was done at a single institution, and its lack of a control group for validation, but Ms. Hunt was confident in its value. “Ultimately, I think this data is significant in that it shows that NIRS can serve as a real-time bedside monitoring tool for detecting acute brain injury. A drop of greater than 25% should be worrisome for clinicians using this monitoring tool,” she said, and she called for further research to establish a consistent protocol.
The source of funding was not disclosed. Ms. Hunt and Dr. Huynh had no relevant disclosures.
SOURCE: Hunt MF et al. Clinical Congress 2019. Abstract, doi: 10.1016/j.jamcollsurg.2019.08.666.