++
Survivors of CO poisoning are faced with potential impairments to cardiac and neurological functions. CO poisoning can cause acute cardiac compromise, and survivors exhibit an increased risk for cardiovascular related death in the subsequent 10 years.88,175 Some patients exhibit acute abnormalities wherein they have impaired consciousness and or focal neurological findings from the time of initial presentation and never recover.3,47,50,69 Other patients seemingly recover from acute poisoning but then manifest neurological or neuropsychiatric abnormalities from 2 days to about 5 weeks after poisoning.42,55,73,92,131,139,143,152,166,183,186,215,235 Events occurring after a clear or “lucid” interval are termed “subacute or delayed” neurological sequelae. These terms have gained popularity and may have some clinical utility, but animal studies suggest neuropathology is more of a continuum. That is, various pathways of injury occur in close proximity, and in some cases they occur concomitantly.68,97,140,153,208,209,213
++
Since 1960, HBO has been used with increasing frequency for severe CO poisoning because clinical recovery appeared to improve beyond that expected with ambient pressure oxygen therapy. Support for HBO use comes from this experience.73,74,93,115,138,152,155,171 The clinical efficacy of HBO for acute CO poisoning has been assessed in six prospective, randomized trials published in peer-reviewed journals. Only one clinical trial satisfies all items deemed to be necessary for the highest quality of randomized controlled trials.235 This double-blinded, placebo-controlled clinical trial involved 152 patients, who received treatment with either three sessions of HBO therapy or normobaric O2 with sham pressurization to maintain blinding. The group treated with HBO had a lower incidence of cognitive sequelae than the group treated with NBO after adjustment for pre-treatment cerebellar dysfunction and stratification (odds ratio 0.45, 95% confidence interval 0.22–0.92, p= 0.03).
++
The only other blinded, prospective, randomized trial was published in 1999, and it involved 191 patients of different severity treated with either daily HBO (2.8 ATA for 60 minutes) with intervening high flow oxygen for 3 or 6 days versus high flow normobaric oxygen for 3 or 6 days.178 Additional HBO treatments (up to six daily) were performed in patients without adequate neurological recovery. The primary outcome measure for this trial was testing performed at completion of treatment (3–6 days) and not from long term follow-up. This study had a high rate of adverse neurological outcomes in all patients, regardless of treatment assignment. Neurological sequelae were reported in 74% of HBO treated patients and in 68% of controls. No other clinical trial has described this magnitude of neurological dysfunction. The high incidence may be related to the assessment tool which could not discern true neurological impairments from poor test taking related to depression.179 Suicide attempts with CO represented 69% of cases in this trial. Moreover, 54% of subjects were lost to follow-up. Outcomes at one month were not reported, but were stated to show no difference. Multiple statistical comparisons were reported without apparent planning or the requisite statistical correction. Both treatment arms received continuous supplemental mask O2 for 3 days between their hyperbaric treatments (both true HBO and “sham”), resulting in greater overall O2 doses than conventional therapy. Multiple flaws in the design and execution of this study are discussed in the literature, so it is impossible to draw meaningful conclusions from the data.80,148 Despite these issues, conclusions from this trial have been accepted in systematic reviews.29,238
++
Other randomized trials of HBO for CO poisoning were not blinded. The first prospective clinical trial involving HBO therapy did not demonstrate therapeutic benefits.166 This study was criticized because the authors used a low oxygen partial pressure (2 ATA) versus the more usual protocols with 2.5 to 3 ATA, an unvalidated questionnaire to assess neurological function, and because nearly half of the patients received hyperbaric treatments more than 6 hours after they were identified.27 This 1989 study protocol was used again in a more recent trial with no modifications despite the criticisms voiced over a decade earlier. Not surprisingly, the outcome was virtually the same. If study results are expressed on an intention-to-treat basis, patients with transient loss of consciousness had an incidence of neurological sequelae based on the self-assessment questionnaire of 48%, and those treated with HBO had an incidence of 51%. Patients with initial coma who were treated once with HBO had an incidence of sequelae of 47%; patients who were treated with two HBO treatments had an incidence of sequelae of 60% (not significantly different from any of the other groups).6
++
HBO was found effective in several other prospective investigations. In a trial involving mildly to moderately poisoned patients, 23% of patients (7/30) treated with ambient pressure oxygen developed neurologic sequelae, whereas no patients (0/30; p<0.05) who were treated with HBO (2.8 ATA) developed sequelae.215 In another prospective, randomized trial, 26 patients were hospitalized within 2 hours of discovery and were equally divided between two treatment groups: ambient pressure oxygen or 2.5 ATA O2.55 Three weeks later, patients treated with HBO had significantly fewer abnormalities on electroencephalogram, and single-photon emission computed tomography (SPECT) scans showed that cerebral vessels had nearly normal reactivity to carbon dioxide, in contrast to diminished reactivity in patients treated with ambient pressure oxygen.
++
In conclusion, efficacy of HBO for acute CO poisoning is supported in animal trials, and studies provide a mechanistic basis for treatment. In this era of evidence-based medicine a great deal of emphasis has been placed on systematic reviews, which have stressed the need for new studies because of the range in quality of published clinical trials.29,238 Practice recommendations are published based on existing studies with the best design that most closely address the actual practical handling of patients.83 HBO should at least be considered in all cases of serious acute CO poisoning and normobaric 100% oxygen continued until the time of HBO administration. CO poisoned patients who receive three HBO treatments within 24 hours following presentation manifest approximately one half the rate of cognitive sequelae at 6 weeks, 6 months, and 12 months following treatment as those treated with only normobaric oxygen. Risk factors for long term cognitive impairment in patients not treated with hyperbaric oxygen include age >36 years, exposure >24 hours, loss of consciousness, and COHb >25%. Recommendations for children can pose special challenges, but in clinical series there appear to be no marked differences in clinical manifestations compared to those reported in adults.62,82,108,162 Pregnancy poses another special situation in that CO readily crosses the placenta and may cause fetal distress and fetal death. HBO has been administered safely to pregnant women, but there are no prospective studies of efficacy, so standard adult guidelines for treatment are often utilized.
++
Another study has shown that HBO is only beneficial in reducing neurological sequelae among patients who do not possess the apolipoprotein ε4 allele.89 Because genotype is typically unknown, this report does not alter existing treatment guidelines, but it may become important for future research. Although the basic mechanisms are unknown, it is well established that the apolipoprotein genotype can have profound effects on risk for a variety of neuropathological events.1,64,141,176 Whether apolipoprotein ε4 modifies the primary pathophysiological insults of CO or the mechanisms of HBO are currently unknown. As of yet, no objective method is available for staging the severity of CO poisoning, although preliminary reports suggest plasma markers may be used in the future.209 Psychometric screening tests have not proved reliable because abnormalities during the initial screening do not correlate with development of delayed sequelae.215