|
In 1950 Pace
and colleagues published a report establishing that
hyperbaric oxygen (HBO) significantly accelerates the rate of
carbon monoxide (CO) elimination from haemoglobin.1 Further studies
have shown that, after CO exposure, HBO accelerates the dissociation
of CO from cytochrome a,a3 and induces a more rapid return
to normal of cytochrome redox state,2,3 reduces brain lipid
peroxidation,4 inhibits pathological
endothelial leukocyte adhesion5 and prevents intracranial
hypertension.6
It has therefore been no surprise that clinical observation and
published evidence, including that of randomised trials, have
supported the clinical efficacy of HBO for CO poisoning.7-12 When HBO is
given to individuals with moderate or severe poisoning within
approximately six hours of exposure, most published data strongly
suggest that it not only induces a more rapid recovery, but also
reduces the most-feared complication -- persistent or delayed
neurological sequelae.
Nevertheless, questions remain. As HBO therapy is not immediately
available to all poisoned patients, who should be treated or
transported to a hyperbaric facility? How many treatments should be
given? There have been few randomised trials of HBO therapy, and no
published randomised trial has addressed the question of whether it
is effective in the treatment of severely poisoned patients.
Therefore, a large trial from a reputable institution is welcome.
In this issue of the Journal, Scheinkestel and
colleagues13 have addressed the
question of whether HBO should be used for CO poisoning. Their study
design is among the most rigorous yet published in this area.
Nonetheless, methodological questions remain, and some caution is
required in accepting the authors' conclusions.
In the study, all 191 patients were treated for 100 minutes in a
hyperbaric chamber, but were randomised to receive either
normobaric oxygen (NBO) or HBO at 2.8 atmospheres absolute (ATA) for
60 of the 100 minutes. All patients received continuous high flow
oxygen between treatments, or 100% oxygen if they were intubated.
They received one treatment per day for three days, after which they
underwent neuropsychological assessment. If, after this, they had
clinical or neuropsychological abnormalities, they received an
additional three treatments. Most patients in this study were
poisoned during an attempted suicide, and 73% were categorised as
severe (defined by a carboxyhaemoglobin level >30%,
Mini-mental score  24, confusion, loss of consciousness, focal
neurological deficits, convulsions, pulmonary oedema,
electrocardiogram abnormalities or dysrhythmias, hypotension,
cardiac arrest or acidosis). Nevertheless, inclusion of mildly
poisoned patients and the high average Mini-mental scores suggest
that a significant number of patients had only mild central nervous
system impairment at the time of presentation.
The use of cluster randomisation for patients presenting
simultaneously (to minimise the effect of the study on daily
practice) engenders the risk of bias, as clusters of patients with
similar baseline characteristics are simultaneously assigned to
one or other of the groups. Generalised linear models were used to
adjust for within-cluster correlation, but the authors have not
provided enough information to allow an assessment as to whether the
statistical analysis accounted for all of the resulting bias. A
confirmatory analysis that would not have incurred this potential
bias would have been a reanalysis of the results using only one patient
from each cluster. This might have strengthened the report, as would a
display of the numbers and sizes of clusters and a listing of the
magnitude and consistency of the parameter estimates for the various
models.
The primary results of the study depend heavily on seven
neuropsychological assessments that were performed after
treatment and before discharge. Two or more "abnormal" test results
represented a poor outcome, which then defined persistent
neurological sequelae (PNS). Conclusions based on these tests must
be tempered by the fact that, although baseline Mini-mental state
tests were done, the study did not make baseline neuropsychological
assessments, and hence could not quantify change for these patients.
The only statistically significant difference between NBO and HBO
patients at the end of treatment was in one of seven
neuropsychological scales (Rey auditory verbal learning
test), which favoured patients treated with NBO. Considering
the multitude of statistical tests performed and the lack of a
comprehensive baseline assessment, a single significant test may
not be meaningful.
In view of the weight of evidence in favour of the use of HBO for acute CO
poisoning, how can the conclusions of Scheinkestel and colleagues
that there is no difference in efficacy between HBO and NBO be
explained? Firstly, it is possible that their method of NBO
administration was more effective than in other studies. All their
patients received three or six days of high-flow oxygen, and
intubated patients might have been given 100% oxygen for six
continuous days, a regimen that is considerably more intensive than
common clinical practice. Secondly, their HBO regimen may have
appeared less effective than in other studies, for several reasons.
Depression can confound neuropsychological assessment, and the
high proportion of depressed patients in the study may have minimised
the apparent effect of HBO treatment and contributed to the high
number of patients with a poor outcome. Unfortunately, neither the
inspired oxygen concentration in either group nor the exact
pressure-time profile of the hyperbaric treatments is provided.
However, adding up to six HBO sessions to 100% oxygen for three or six
days would be expected to produce a significant degree of pulmonary
oxygen toxicity. Also, as Scheinkestel and colleagues point out,
repetitive treatments at 2.8 ATA (a pressure higher than many
clinicians use for CO poisoning) might have induced a neurotoxic
effect14 that offset any potential
benefit.
Further, there are significant omissions from the article that
preclude unfettered acceptance of the authors' conclusions. The
surprising observation that there was no significant improvement in
Mini-mental score in either group is weakened by the lack of
information regarding administration of sedative drugs,
especially to intubated patients, which might have confounded the
testing. Importantly, other than mortality, no clinical outcomes or
self-reported assessments of functional ability are reported. The
overall relapse rate at follow-up, defined as new morbidity or
deterioration in any neuropsychological test score, was higher in
the group treated with HBO, but the relapse rate in the various
subgroups of greatest interest (particularly those with short
treatment delays) is not detailed. Moreover, the low follow-up rate
(46%) makes it difficult to draw valid conclusions.
The data regarding comparability of the two groups have two
significant omissions -- the numbers of severely affected patients
with long delays to treatment, and the number of mildly affected
patients. Neither of these subgroups is likely to show a measurable
response to treatment using the chosen endpoints. The possibility
that both types were significantly represented is suggested by
Mini-mental scores that appear disproportionately high for the
degree of severity that is implied, as well as the high geometric mean
of the delays to treatment (>6 hours). It therefore appears
possible that a significant proportion of the patients in this study
were treated at a time after CO exposure that HBO is likely to be
ineffective.15-17 Including a large
number of patients who are unlikely to respond (too mildly affected or
treated too late) in a study will reduce the apparent effectiveness of
the intervention, and might partly explain the surprisingly high
proportion of patients with PNS (74% and 68% of patients,
respectively, in the HBO and NBO groups). What might otherwise have
been the most important conclusion of this investigation -- that even
in the subgroup of severe poisonings treated within four hours, there
was no difference between NBO and HBO -- would have been more
convincing had the authors provided the observational and
statistical details.
What new information can be learned from the work of Scheinkestel and
colleagues? Their results hint that, in the type of patients studied,
prolonged administration of NBO may be more effective than the
shorter regimens that are in general use. With respect to the primary
question addressed by the investigators,13 we feel that there are
still too many unresolved issues in their analysis to discard HBO as a
treatment for acute CO poisoning.
Richard E Moon
Professor of Anesthesiology, and
Associate Professor of Pulmonary
and Critical Care Medicine
Elizabeth DeLong
Associate Professor, Division of Biometry
Duke University Medical Center, Durham, NC, USA
- Pace N, Strajman E, Walker E. Acceleration of carbon monoxide
elimination in man by high pressure oxygen. Science 1950;
111: 652-654.
-
Brown SD, Piantadosi CA. Reversal of carbon monoxide-cytochrome c
oxidase binding by hyperbaric oxygen in vivo. Adv Exp Med
Biol 1989; 248: 747-754.
-
Brown SD, Piantadosi CA. Recovery of energy metabolism in rat brain
after carbon monoxide hypoxia. J Clin Invest 1992; 89:
666-672.
-
Thom S. Antagonism of carbon monoxide-mediated brain lipid
peroxidation by hyperbaric oxygen. Toxicol Appl Pharmacol
1990; 105: 340-344.
-
Thom SR. Functional inhibition of leukocyte
 2
integrins by hyperbaric oxygen in carbon monoxide-mediated brain
injury in rats. Toxicol Appl Pharmacol 1993; 123: 248-256.
-
Jiang J, Tysseborn I. Cerebrospinal fluid pressure changes after
acute carbon monoxide poisoning and therapeutic effects of
normobaric and hyperbaric oxygen in conscious rats. Undersea
Hyperb Med 1997; 24: 245-254.
-
Myers RAM, Snyder SK, Emhoff TA. Subacute sequelae of carbon
monoxide poisoning. Ann Emerg Med 1985; 14: 1163-1167.
-
Norkool DM, Kirkpatrick JN. Treatment of acute carbon monoxide
poisoning with hyperbaric oxygen: a review of 115 cases. Ann Emerg
Med 1985; 14: 1168-1171.
-
Gorman DF, Clayton D, Gilligan JE, Webb RK. A longitudinal study of
100 consecutive admissions for carbon monoxide poisoning to the
Royal Adelaide Hospital. Anaesth Intensive Care 1992; 20:
311-316.
-
Ducasse JL, Celsis P, Marc-Vergnes JP. Non-comatose patients
with acute carbon monoxide poisoning: hyperbaric or normobaric
oxygenation? Undersea Hyperbar Med 1995; 22: 9-15.
-
Thom S, Taber R, Mendiguren I, et al. Delayed neuropsychologic
sequelae after carbon monoxide poisoning: prevention by treatment
with hyperbaric oxygen. Ann Emerg Med 1995; 25: 474-480.
-
Hampson NB. Carbon monoxide poisoning in the United States. In:
Oriani G, Marroni A, Wattel F, editors. Handbook on hyperbaric
medicine. New York: Springer, 1996: 297-304.
-
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: 203-210.
-
Holbach KH, Caroli A, Wassmann H. Cerebral energy metabolism in
patients with brain lesions at normo- and hyperbaric oxygen
pressures. J Neurol 1977; 217: 17-30.
-
Zanetti CL. A review of carbon monoxide poisoning treated at
Edgewater Hospital. In: Kindwall EP, editor. Proceedings of the
Eighth International Congress on Hyperbaric Medicine. 1984; Aug
20-22; Long Beach, CA. San Pedro, CA: Best Publishing, 1987: 258-262.
-
Goulon M, Barois A, Rapin M, et al. Carbon monoxide poisoning and
acute anoxia due to breathing coal gas and hydrocarbons. J
Hyperbaric Med 1986; 1: 23-41.
-
Raphael JC, Elkharrat D, Jars-Guincestre MC, et al. Trial of
normobaric and hyperbaric oxygen for acute carbon monoxide
intoxication. Lancet 1989; 2: 414-419.
©MJA 1998
Make a
comment
|
