Current issues of ACP Journal Club are published in Annals of Internal Medicine


3 hyperbaric oxygen treatments reduced cognitive sequelae of acute carbon monoxide poisoning


ACP J Club. 2003 May-Jun;138:67. doi:10.7326/ACPJC-2003-138-3-067

Source Citation

Weaver LK, Hopkins RO, Chan KJ, et al. Hyperbaric oxygen for acute carbon monoxide poisoning. N Engl J Med. 2002;347:1057-67. [PubMed ID: 12362006] (All 2003 articles were reviewed for relevancy, and abstracts were last revised in 2009.)



In patients with carbon monoxide poisoning, is hyperbaric oxygen (HBO) more effective than normobaric oxygen for reducing cognitive sequelae?


Randomized (allocation concealed*), blinded (patients, outcome assessors, and statisticians),* controlled trial with follow-up at 6 weeks, 6 months, and 1 year.


Emergency departments in Utah, Idaho, and Wyoming, USA.


152 patients (mean age 36 y, 62% men) with documented or obvious exposure to carbon monoxide and any of 11 predetermined symptoms. Patients were excluded if they were < 16 years of age, moribund, or pregnant or if > 24 hours had elapsed since exposure. Follow-up at 6 weeks was 97%.


All patients had 3 hyperbaric chamber sessions at intervals of 6 to 12 hours starting within 24 hours after carbon monoxide exposure. All nonintubated patients received oxygen, 15 L/min via a nonrebreathing face mask and reservoir. Intubated patients were mechanically ventilated with 100% oxygen. 76 patients were allocated to HBO and exposed to 100% oxygen at 3 atmospheres (304 kPa) and then 2 atmospheres absolute (203 kPa) during the first chamber session and 100% oxygen at 2 atmospheres absolute for sessions 2 and 3. Seventy-six patients were allocated to normobaric oxygen and exposed to air at 1 atmosphere absolute (101.3 kPa, or sea-level pressure) for all 3 chamber sessions.

Main outcome measures

Cognitive sequelae at 6 weeks were measured using a battery of neuropsychological tests and were considered present if any subtest score was > 2 standard deviations (SDs) below the mean of demographically corrected standardized scores or if ≥ 2 scores were > 1 SD below the mean.

Main results

Analysis was by intention to treat. The trial was stopped early after 3 of 4 planned interim analyses were completed. At 6 weeks, 6 months, and 12 months, patients in the HBO group had fewer cognitive sequelae than did patients in the normobaric group (Table). The analysis was adjusted for cerebellar dysfunction because it was not equally distributed (4 HBO vs 15 normobaric group patients), with no change in results.


In patients with acute carbon monoxide poisoning, 3 hyperbaric oxygen treatments within 24 hours of exposure reduced cognitive sequelae at 6 weeks, 6 months, and 1 year.

*See Glossary.

Source of funding: Not stated.

For correspondence: Dr. L.K. Weaver, LDS Hospital, Salt Lake City, UT, USA. E-mail

Table. Hyperbaric oxygen vs normobaric oxygen for acute carbon monoxide poisoning†

Outcomes Hyperbaric Normobaric RRR (95% CI) NNT (CI)
Neurologic sequelae at 6 wk 25% 46% 46% (15 to 66) 5 (3 to 18)
Neurologic sequelae at 6 mo 25% 46% 46% (15 to 66) 5 (3 to 18)
Neurologic sequelae at 1 y 18% 33% 44% (2.2 to 68) 7 (4 to 182)

†Abbreviations defined in Glossary; RRR, NNT, and CI calculated from data in article. Analysis includes all patients. Those with missing data at 6 weeks were assumed to have cognitive sequelae; those with cognitive sequelae at 6 weeks and missing data at 6 or 12 months were assumed to have cognitive sequelae at those time points.


Although carbon monoxide, the most frequent cause of poisoning, results in both immediate and delayed cognitive deficits, physicians have been justifiably skeptical about treating it with HBO rather than normobaric oxygen. HBO often incurs the cost and risk associated with transferring a potentially unstable patient to a referral center, and observational studies that initially suggested a benefit were recently contradicted by a well-designed but negative randomized trial by Scheinkestel and colleagues (1).

The study by Weaver and colleagues showed a large neuropsychological benefit in the HBO group. This study avoids 2 of the limitations of the study by Scheinkestel and colleagues: loss to follow-up and cluster rather than patient-level randomization for group exposures. The Scheinkestel trial also differed in other respects. More patients had attempted suicide (69% compared with 31% in the study by Weaver and colleagues), and depression and other toxins may have influenced outcomes. Patients in the Scheinkestel trial control group received continuous face mask oxygen at 15 L/min for 3 days, whereas patients in the study by Weaver and colleagues received oxygen only when needed to maintain oxygen saturation. Both studies concealed allocation and used sham hyperbaric treatments to enhance blinding, and both used persistent or delayed neurologic sequelae as a primary outcome. However, each used a different battery of tests. Similarly, differences in treatment delay, case severity, and HBO protocols might also explain the disparity between the results of these 2 studies.

The study by Weaver and colleagues shifts the balance of evidence to support HBO over normobaric oxygen. Future studies should be designed to determine optimal combinations of the dose and timing of HBO, and planned subgroup analyses should attempt to define which patients could benefit most from hospital transfer. Hopefully, these studies will provide the evidence needed to erase the remaining doubt about the effectiveness of HBO.

Stephen Pitts, MD, MPH
Emory University
Atlanta, Georgia, USA


1. Scheinkestel CD, Bailey M, Myles PS, et al. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: a randomised controlled clinical trial. Med J Aust. 1999;170:203-10. [PubMed ID: 10092916]