Hyperbaric oxygen therapy in carbon monoxide poisoning: effects on neurological sequelae

Journal Club

CJEM 2000;2(1):22-24

Clinical question
Is hyperbaric oxygen superior to normobaric oxygen for preventing neurologic sequelae in carbon monoxide (CO) poisoned patients?

Article chosen
Scheinkestel CD, Bailey M, Myles PS, Jones K, Cooper DJ, Millar IL, et al. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: a randomized controlled clinical trial. Med J Aust 1999;170:203-10.

The search
National Library of Medicine, PubMed, MEDLINE MeSH headings: carbon monoxide AND poisoning (5 years, human, all studies) Yield: 57 citations

The evidence

Design: A double-blind randomized controlled trial with 1-month follow-up.

Population: Over a 28-month period, 230 patients with CO poisoning were referred to a hyperbaric unit. Of these, 191 who received treatment within 24 hours of exposure were enrolled. Pregnant women, children and burn victims were excluded. Eligible patients were stratified based on the need for mechanical ventilation and severity of poisoning. Severe poisoning was defined as: mini-mental status score < 24, carboxyhemoglobin level over 30%, confusion, focal neurological deficits, loss of consciousness, ECG abnormalities, arrythmias, pulmonary edema, metabolic acidosis, hypotension, convulsions, cardiac arrest or need for mechanical ventilation.

Intervention: All patients underwent 100-minute treatments in a hyperbaric chamber on 3 successive days. Hyperbaric oxygen (HBO) patients received 60 minutes of HBO therapy at 2.8 atmospheres plus 40 minutes of 100% oxygen. Normobaric oxygen (NBO) patients received 100 minutes of 100% oxygen at 1.0 atmosphere. NBO patients underwent "sham" treatment, whereby the chamber was flushed regularly to simulate pressurization. Treatment was extended to 6 days for patients who were clinically abnormal or who had poor neuropsychological parameters after 3 treatments. All patients received continuous high-flow oxygen between chamber exposures.

Outcomes measured: Patients underwent neuropsychological testing at baseline, after treatment (at 3 or 6 days), and at 1 month. Persistent neurologic sequelae (PNS) was defined as failure to improve to pre-treatment levels after oxygen therapy. Delayed neurologic sequelae (DNS) was defined as post-treatment improvement to normal, followed by a deterioration over days to weeks.

Results: 73% of the patients had severe CO poisoning, and mean time to treatment was 7.1 hours. Prognostic baseline variables (proportion treated within 4 hours of exposure, proportion requiring mechanical ventilation, CO level, and number of severe poisonings and accidental poisonings) were balanced between groups, suggesting adequate randomization.

After 3 days of treatment, 28% of HBO patients and 15% of NBO patients had abnormal neuropsychological tests (p = 0.01). For patients with severe poisoning, these figures were 35% and 13% respectively (p = 0.001). The incidence of persistent neurological sequelae (PNS) was 71% at discharge and 62% at 1-month follow-up, with no significant difference between groups. Five patients (2.6%) had delayed neurologic sequelae (DNS) at a median of 40 days after treatment. All 5 belonged to the HBO group. At follow-up, NBO recipients scored significantly better on 1 of 7 neuropsychological tests (the Rye auditory verbal learning test) and the same on the others.

Conclusion
HBO therapy did not improve neuropsychological outcomes in this heterogeneous group of CO poisoned patients.

Comments
Carbon monoxide poisoning is common^1,2 and hyperbaric facilities are not. HBO therapy is costly and relatively inaccessible in Canada and other countries with low population densities and long transport times. This well designed randomized clinical trial suggests that, while high-flow oxygen may be important, HBO does not improve neuropsychological outcomes after CO poisoning.

The incidence of PNS in this study (71% at discharge and 62% at follow-up) is high compared to previous studies.3 This probably reflects the high proportion of severe poisonings (73%) and suicide attempts (69%), as it is known that suicidal patients are often depressed and tend to score poorly on neuropsychological testing.

To put their results in context, the authors discuss 6 prior randomized studies of HBO therapy in CO poisoned patients. Two of these (combined n = 91) showed benefit while 4 (combined n = 1395) showed none. The current study is the first randomized trial to use standardized treatment and follow-up protocols as well as "sham treatment" for controls. Other studies have chosen not to include "sham" treatments because of the risk of chamber-related complications.^3-5

Advocates of hyperbaric oxygen therapy suggest that treatment is most effective when applied within 6 hours of exposure. In this study, mean time from exposure to treatment was 7.1 hours, and it is conceivable that the delay may have limited HBO's potential efficacy. To address this concern, the authors performed a retrospective subgroup analysis of patients who presented from 0-3 hours, 3-6 hours, 6-12 hours and >12 hours. This analysis failed to show a time-dependent HBO treatment effect.

It is interesting to note that - based on number of treatments, treatment duration, and the administration of high-flow oxygen between chamber exposures - patients in the Scheinkestel study received a higher total oxygen dose than those in previous studies. In this study, HBO patients received oxygen therapy equivalent to 35.7 carboxyhemoglobin dissociation half-lives and NBO patients received the equivalent of 28.5 dissociation half-lives. In previous studies, patients received between 7 and 18 carboxyhemoglobin dissociation half-lives. Since total oxygen dose may be an important determinant of outcome,^3-8 it is possible that the NBO group in this study did relatively better because of the higher administered oxygen dosage.

An important drawback to the study is the fact that the investigators achieved only 46% follow-up at 1 month. While this rate is comparable to previous studies, it raises the possibility that patients lost to follow-up might have done significantly better or worse than those captured. If so, the true outcomes could differ from the reported outcomes.

Readers should also remember that these findings may not apply to pregnant women, children and burn victims, and that the methodological problems described above raise minor concerns about the study conclusions.

Clinical bottom line
This study is compatible with the bulk of previous literature. It suggests that most patients can be managed with NBO and that HBO does not improve neuropsychological outcomes after CO poisoning - especially in severely poisoned patients like the ones studied. Emergency physicians who manage CO poisoned patients without a hyperbaric facility will take comfort from these findings; however, it is still possible that some subgroups do benefit from HBO, and it may be prudent for physicians to collaborate with local hyperbaric facilities to establish protocols for dealing with specific patient groups.

References

1. Weaver LK. Carbon monoxide poisoning. Crit Care Clin 1999;15:297-317.
2. Varon J, Marik PE, Fromm RE Jr, Gueler A. Carbon monoxide poisoning: a review for clinicians. J Emerg Med 1999;17:87-93.
3. Tibbles PM, Perrotta PL. Treatment of CO poisoning: a critical review of human outcome studies comparing NBO with HBO. Ann Emerg Med 1994;24:269-76.
4. Hopkins RO, Weaver LK. CO controversies: neuropsychologic testing, mechanism of toxicity and hyperbaric oxygen. Ann Emerg Med 1995;25:272-3.
5. Seger D, Welch L. CO controversies: neuropsychologic testing, mechanism of toxicity and hyperbaric oxygen. Ann Emerg Med 1994;24:242-8.
6. Thom SR, Taber RL, Mendiguren II, Clark JM, Hardy KR, Fisher AB. Delayed neuropsychologic sequelae (NPS) after CO poisoning: prevention by treatment with HBO. Ann Emerg Med 1995;25:474-80.
7. Olson KR, Seger D. Hyperbaric oxygen for carbon monoxide poisoning: Does it really work? Ann Emerg Med 1995;25:535-7.
8. Hampson NB, Dunford RG, Kramer CC, Norkool DM. Selection criteria utilized for HBO treatment of CO poisoning. J Emerg Med 1995;13:227.
   

Date appraised: Nov. 7, 1999