‘Narrow therapeutic window’ for oxygen in brain injury
MedWire News: Hyperoxia, as well as hypoxia, early after hospital admission is associated with poor outcomes and mortality in patients with traumatic brain injury, report researchers in the Archives of Surgery.
In an invited critique, H Gill Cryer (University of California at Los Angeles Medical Center, USA) suggests that doctors may tend to give patients high oxygen supplementation to prevent hypoxia, with its known detrimental effect on outcomes.
This seemed evident in the current study, which included 1547 patients with severe brain injury, defined as a head Abbreviated Injury Score of 3 or more. Megan Brenner (University of Maryland School of Medicine, Baltimore, USA) and team report that 43% of the patients were hyperoxic during the first 24 hours of hospitalization, with average partial pressure of oxygen in the blood (PaO2) values exceeding 200 mmHg. By contrast, just 7% were hypoxic, with average levels below 100 mmHg.
"Unfortunately, this hyperoxia appears to be detrimental to outcome compared with normoxia and is just as bad as hypoxia," says Cryer. "If true, these are important findings that could change the way we approach patients with traumatic brain injury."
The mortality rates were 31% and 38% in the hyperoxic and hypoxic groups, respectively, compared with 25% among patients with PaO2 values of 100-200 mmHg. These associations persisted after accounting for confounders, with hyperoxia increasing mortality risk 1.5-fold and hypoxia raising it 2.2-fold, relative to normoxia.
Also, hyperoxic patients had a significant 1.52-fold increase in risk for poor short-term outcomes, defined as a discharge Glasgow Coma Scale score of 3-8, relative to patients with normal PaO2 levels, while hypoxic patients had a 1.66-fold increased risk.
There were no significant differences in most outcomes when comparing hyperoxic and hypoxic patients. Hyperoxic patients had longer stays in the intensive care unit and in hospital, but this was explained by a larger proportion of hypoxic patients who died doing so within 7 days of admission, at 90%, compared with 76% of hyperoxic and 74% of normoxic patients.
Cryer notes that several questions remain to be answered, including whether hyperoxia is associated with specifically with death from brain injury.
But he concludes: "I believe this article raises the important possibility that overshooting normal PaO2 values in an attempt to prevent hypoxia in patients with traumatic brain injury may be detrimental."
MedWire (www.medwire-news.md) is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2012
By Eleanor McDermid