Hyperbaric Oxygen for Carbon Monoxide Poisoning

Abstract & Commentary

Synopsis: In this carefully done, double-blind, placebo-controlled clinical trial, hyperbaric oxygen administered in 3 treatment sessions over 24 hours reduced the incidence of impaired cognitive function following acute carbon monoxide poisoning.

Source: Weaver LK, et al. Hyperbaric oxygen for acute carbon monoxide poisoning. N Engl J Med. 2002; 347(14):1057-1067.

In an important, large-scale study addressing a controversial issue, Weaver and colleagues at the University of Utah compared hyperbaric oxygen (HBO) therapy to normobaric oxygen therapy in patients with acute carbon monoxide (CO) poisoning, using a randomized, double-blind, controlled (sham HBO treatment) design. Beginning in late 1992, Weaver et al evaluated 98% of all patients with known CO poisoning seen in the hospital emergency departments of hospitals in Utah, Idaho, and Wyoming for possible inclusion in the study. Patients with documented CO poisoning were enrolled if they had loss of consciousness or other central-nervous-system manifestation, nausea, vomiting, cardiac ischemia, or metabolic acidosis, and could be enrolled and begun on HBO treatment within 24 hours of CO exposure.

Patients received either HBO therapy (sessions with 100% oxygen at pressures of 3, 2, and 2 atmospheres absolute over 24 hours) or sham HBO (sessions with air at sea-level pressure as compared to ambient Salt Lake City barometric pressure of 0.85 atmospheres absolute) in a monoplace HBO chamber. The respiratory therapists operating the chamber knew whether HBO or sham treatment was being administered, but this was concealed from both the patients and Weaver et al throughout the study and through 12 months of follow-up. Psychologists unaware of the treatment group administered neuropsychological tests, immediately after chamber sessions 1 and 3, and at 2 and 6 weeks and 6 and 12 months following original exposure. The primary study outcome was the incidence and severity of cognitive sequelae at 6 weeks.

A total of 460 patients were screened, of whom 332 were potentially eligible for the study and 152 were enrolled (76 patients in each group). According to the predetermined regimen for interim analysis of the results, the study was stopped after 150 of the planned 200 patients had been enrolled, because HBO was shown to be efficacious (P < 0.01). The patients (29% females) were 35 ± 10 yrs old in the HBO group and of similar demographics in the placebo group. Mean blood levels of carboxyhemoglobin (COHb) in the 2 groups were the same, 25% initially and approximately 5% at the time of the first treatment in the chamber. Eight patients in each group were initially intubated. None of the patients died.

Cognitive sequelae at 6 weeks were present in 25% (19 of 76) of the HBO patients and in 46% (35 of 76) of the control patients (adjusted odds ratio, 0.45; with 95% confidence interval, 0.22-0.92; P = 0.03). Although more of the control patients had cerebellar dysfunction on initial examination (15% vs 4%), the significant increase in cognitive sequelae persisted even after adjustment for this and for other stratification variables. Patients randomized to the HBO group continued to have significantly less cognitive disturbances at 12 months. Adverse effects of the HBO treatment were minor and consisted mainly of anxiety and ear problems.

Comment by David J. Pierson, MD

This is a landmark study that will be quoted widely and will influence practice for a long time. The New England Journal of Medicine made a big deal of its publication: an 11-page lead article (longer than most of the Journal’s original research reports), an accompanying editorial1 and a clinical perspective.2 With a very well-designed and carefully described study, Weaver et al have substantially improved the quality of clinical evidence now available for managing patients with acute CO poisoning. They have shown that patients with this condition who demonstrate initial neurological or cardiac abnormalities, or who present with nausea and vomiting or metabolic acidosis, benefit from HBO therapy with respect to the incidence of long-term cognitive sequelae.

A recent review of HBO in these pages by Takezawa3 concluded that this form of therapy had been shown to be effective in diabetic foot ulcer, and that the rationale and clinical imperative for treatment in decompression sickness were sufficiently compelling to justify its use in that condition. However, based on the literature published to date, including the results of 5 randomized controlled trials, HBO could not be recommended as an effective treatment for CO poisoning. Although Takezawa’s review drew criticism from several well-known authorities on HBO in the form of letters and phone calls to the Editor, his conclusions could be justified on the basis of the evidence published prior to the appearance of the present paper.

This statement is supported by a Cochrane Review by Juurlink et al4 earlier this year. Juurlink and colleagues point out that the favorable results of a number of nonrandomized trials of HBO in CO poisoning have led to its widespread use for this indication. However, their review of 6 randomized controlled trials failed to show an overall benefit of HBO with respect to long-term neurologic sequelae: symptoms possibly related to CO poisoning were present in 34% (81 of 237) of patients who received HBO (albeit with varying regimens) as compared with 37% (81 of 218) of patients treated with normobaric oxygen (OR for benefit from HBO, 0.82, 95% CI, 0.41-1.66).

One reason the literature on this has been suboptimal is that this kind of research is complex and difficult to do well. The Utah group has a track record of successful performance of careful, large-scale clinical outcome studies, and they appear to have "gotten it right" with this one. Their report has been a long time coming. It has been a decade since the first patients were enrolled. The study was first discussed at a national meeting 7 years ago, and the upcoming positive results have been touted for several years. It is a relief to see it, finally, in print.

Clinicians should be cautious in applying the results of this study to their practice. Weaver et al used 3 HBO sessions, whereas a study published in 19955 found that 74% of practitioners who provided HBO treatment used only a single HBO session for CO poisoning. Whether benefit similar to that shown in the study by Weaver et al would occur with only one HBO session is unknown and should probably not be assumed.

Dr. Pierson is Professor of Medicine University of Washington Medical Director Respiratory Care Harborview Medical Center, Seattle, WA.


1. Thom SR. Hyperbaric-oxygen therapy for acute carbon monoxide poisoning. N Engl J Med. 2002; 347(14):1105-1106.

2. Piantadosi CA. Carbon monoxide poisoning. N Engl J Med. 2002;347(14):1054-1055.

3. Takezawa J. Hyperbaric oxygen therapy. Critical Care Alert. 2000;8(8):88-91.

4. Juurlink DN, Stanbrook MB, McGuigan MA. Hyperbaric oxygen for carbon monoxide poisoning (Cochrane Review). In: The Cochrane Library, Issue 3, 2002. Oxford: Update Software.

5. Hampson NB, et al. Selection criteria utilized for hyperbaric oxygen treatment for carbon monoxide poisoning. J Emerg Med. 1995;13:227-231.