Declaring conflict of interest

Letters to the Editor

Emergency Medicine (2000) 12, 354–359

Carbon monoxide poisoning

Dr Doherty is to be congratulated on his review of thecomplex topic of carbon monoxide (CO) poisoning inhis paper ‘History, pathophysiology, clinicalpresentation and role of hyperbaric oxygen in acutecarbon monoxide poisoning’.1 He summarized that themainstays of treatment include immediate rescue fromthe source, application of 100% oxygen (NBO) as soonas possible, and supportive care, including appropriateairway management. This emphasizes the critical roleof prehospital providers and emergency physicians inthe diagnosis and management of CO poisoning. Wewish to comment on his conclusions regardingpathophysiology and clinical outcomes.

CO pathophysiologyFocal neurological injury has been detected in CT andpost-mortem studies of CO victims. Doherty did notexplain why a systemic toxin (CO) should selectivelytarget specific areas of the brain.1 Recent work hasidentified another toxic mechanism of CO; as a secondmessenger.2 It acts via haem oxygenase enzymesystems in the brain, which show high concentrationsin the olfactory bulb, hippocampus, islands of Callejae,piriform cortex, pontine nucleus, habenula, tenia tecta,cerebellum and basal ganglia including the globuspallidus. Permanent structural injury from CO hasbeen demonstrated in similar areas, and correlates withclinical neurological syndromes of memory loss,urinary incontinence, Parkinsonian gait disorders andpersonality changes.3,4

Clinical outcomesWe do not agree with Doherty that ‘…multiple studieshave failed to show any firm scientific benefit in theuse of hyperbaric oxygen…’ (compared with 100%oxygen) for CO poisoning. There are three componentsto the neuropsychiatric effects of CO: acute reversibleinjury, persistent early morbidity, and delayedneurological syndrome.4 These are summarized fromavailable trial data, some not referenced in Doherty’sreview.

Acute reversible toxicity

This is due to the cellular effects of CO, including lipidperoxidation, as outlined in Doherty’s paper. This is

corrected by rescue from the CO environment, andadministration of 100% oxygen or hyperbaric oxygen(HBO). Hyperbaric oxygen treatment considerablyreduces the elimination half-life of CO compared withoxygen at 1.0 atmospheres absolute (ATA).5,6,

Standard HBO treatment for CO poisoning inAustralia and New Zealand is 2.8 ATA, the pressurerequired to maintain cellular respiration withouthaemoglobin. Two 2.8 ATA HBO treatments 12 h apartallows definitive treatment of CO poisoning, anddischarge of the patient who is free of persistentsequelae or psychiatric morbidity.4 Data from therandomized trial by Ducasse and Celcis support thisearlier recovery with HBO.7 Resolution of clinicalabnormalities occurred more rapidly with HBO thanNBO in their series; five out of the 13 NBO-treatedpatients still had clinical abnormalities at 12 h versusnone of the HBO-treated group. A 3-day admission isrequired using the Scheinkestel NBO regimen.8

Further analysis of our data has demonstrated COelimination half-lives of 37.23 min (95% CI31.19–43.26 min, range 14.49–83.0 min) in HBO 2.8 ATA(25 patients), compared with 144.6 min (95% CI86.9–202.2 min, range 52.19–321.8 min) for 100%oxygen 1.0 ATA (13 patients).6 Our values are of thesame magnitude as those described by Pace et al.(15 volunteers).5 We also found a sixfold variation inmeasured elimination half-lives within treatmentgroups, NBO and HBO. Empirical regimens toeliminate CO may be inaccurate; undertreating somepatients and overtreating others. Our ongoing researchis attempting to titrate the oxygen dose to match theindividual’s CO load.

Persistent early morbidity — Irreversibleneurological injury

At some point the neurological injury becomesirreversible, probably due to a combination of tissuehypoxia, hypotension, and continued effects outlinedunder the heading ‘acute reversible toxicity’. Studies ofpersistent early morbidity (PEM) as an outcomemeasure have demonstrated the greatest discrepancies.Definitions of PEM and follow up vary considerablyamong trials.

We present a summary of the outcome data forPEM from five major prospective trials (key:

P, prospective; R, randomized; C, controlled; S, single-blinded; D, double-blinded; U, unblinded):

• Raphael J-C et al. (1989) P, R, C, U(629 patients – 89.5% follow up 1 month)9

Evaluation at 1 month (group A)100% oxygen PEM = 36/89 (40.4%) HBO PEM = 41/88 (46.5%), P = 0.61

• Gorman DF et al. (1992) P, non-R, C, U(100 patients – 76% follow up 1 month)10

100% oxygen PEM = 2/8 (25%), 1 HBO treatmentPEM = 9/20 (45%), multiple HBO treatmentsPEM = 9/50 (18%), P = 0.039 multiple HBO versusother (favours multiple HBO)

• Weaver LK et al. (1995) P, R, C, D(50 patients – 98% follow up 6 weeks)11

Group A 4/25 (16%) PEM, group B 4/25 (16%)P = NSTreatment arms not known

• Mathieu D et al. (1996) P, R, C, S(575 patients – 100% follow up at 12 months)12

Difference noted at 3 months (no difference at 1, 6 or12 months)100% oxygen PEM = 42/234 (15%) HBO PEM = 26/273 (9.5%), P = 0.016 (favours HBO)

• Scheinkestel C et al. (1999) P, R, C, D(191 patients – 46% follow up at 1 month)8

(i) Severely poisoned100% oxygen PEM = 44/67 (65%),HBO PEM = 61/72 (85%), P = 0.03 (favours NBO)(ii) Not severely poisoned (by Kehat and Schupakfrom Scheinkestel’s data)13

100% oxygen PEM = 15/20 (75%)HBO PEM = 16/32 (50%), P = 0.04 (favours HBO)

Comparing studies that have used PEM as ameasure of clinical outcome is difficult due to variabletreatment regimens, definition of PEM, variabledocumentation of severity of CO exposure, rescue to100% oxygen delay, actual oxygen concentrationduring treatment, treatment duration, and patients’premorbid status.

The interpretation one can make from the above fivestudies is that neither treatment is superior. Thissuggests that a certain amount of neurological damageis determined at the time of the poisoning, or duringthe prehospital period. For this group, HBO provides afaster treatment regimen, with earlier recovery, if it isavailable.

Delayed neurological syndrome

Delayed neurological syndrome (DNS) occurs afterthere has been an apparent recovery during acutetreatment. As outlined in Doherty’s paper, thesyndrome may follow coma-inducing exposure, andmay be worse in the elderly.9 Available evidencesuggests a treatment advantage for HBO in preventingthis syndrome in seven prospective trials summarizedbelow:• Myers RAM et al. (1985) P, non-R, C, U

(213 patients – follow up not stated)14

100% oxygen DNS = 10/82 (12.2%), HBO DNS = 0/131, P < 0.0001All relapses successfully treated with HBOFavours HBO

• Gorman DF et al. (1992) P, non-R, C, U(100 patients 76% follow up at 1 month)10

100% oxygen DNS = 2/8 (25%), 1 HBO treatmentDNS = 2/20 (10%), multiple HBO treatments PEM= 0/50 (0%), P = 0.014 multiple HBO versus otherNo difference single HBO versus 100% oxygenFavours multiple HBO

• Ducasse JL et al. (1995) P, R, C, U(26 patients – 69% follow up at 1 month)7

EEG abnormalities at 3 weeks after exposure100% oxygen 4/10 (40%) HBO 0/8 (0%)Favours HBO

• Thom S et al. (1995) P, R, C, U(65 patients – 100% follow up at 3 months)15

100% oxygen DNS = 7/32 (22%), HBO DNS = 0/33(0%), P = 0.014Favours HBO

• Weaver et al. (1995) P, C, R, D(50 patients – 98% follow up at 6 weeks)11

Group A 4/25 (16%) DNS, group B 0/25 (0%) P = 0.054Treatment arms not known

• Smart DR et al. (1995) P, C, non-R, S(66 patients – 95% follow up 3 months)6

100% oxygen DNS = 4/15 (26.7%), HBO DNS = 1/43 (2.3%) P = 0.013Favours HBOOur study of 66 patients allocated to HBO and NBO

by a prospective protocol was reported in abstract in1995. We measured elimination of expired CO; zero inthe breath defining the treatment end-point. Wetailored treatment to each individual. There were twodeaths, and 63 patients were followed up to 3 months;58 had full testing. One out of 43 HBO-treated patientshad suffered DNS by 3 months compared with four out

355


of 15 treated with 100% oxygen. All DNS was treatedwith HBO, and fully recovered.• Scheinkestel et al. (1999) P, C, R, D

(191 patients – 46% follow up 1 month)8

100% oxygen DNS = 0/87 (0%) HBO DNS = 5/104 (4.8%), (P = 0.03)Favours NBOScheinkestel’s is the only study that found an

increase in DNS in the HBO-treated group. Theirtreatment end-point was at 1 month, yet five patientswho ‘relapsed’ after hyperbaric oxygen treatment didso at a median of 40 days (IQR 29–80 days). Based onthese figures, at least three of the five patients withDNS presented after the study protocol had finished,introducing a possible bias (i.e. that their detection wasnot part of the study). The HBO-treated group were amean of 3 years older than the NBO-treated group(P = 0.13), which may have been relevant.

When analysing outcomes for DNS, five of theabove studies favour HBO, one favours NBO, and one isyet to be concluded. HBO has data supporting itsefficacy in treatment of DNS, no such data exist forNBO.

The pregnant patient

Dr Doherty did not mention HBO as a treatment optionfor CO-poisoned pregnant patients. Two series from1991 (44 and 40 patients) showed use of HBO tosuccessfully treat pregnant patients, albeit with smallsubgroup numbers.16,17 In the second multicentreprospective study, three out of five in the severelypoisoned group receiving high flow oxygen hadadverse fetal outcomes compared with none in the HBOgroup.17 In the absence of RCT for pregnant patients,HBO is favoured as a treatment option.

SummaryWe agree with Doherty’s assertion that there arelimitations in all studies of CO to date. Despitelimitations, a strong case remains, supporting HBO intreatment of CO poisoning.• Hyperbaric oxygen removes CO from the body

faster than 100% oxygen at 1 ATA, and mayshorten hospital stay of poisoned patients who arefree of PEM, and psychiatric comorbidity.

• HBO guarantees that a poisoned individual willreceive 100% oxygen in a dose appropriate for thepatient. The authors are aware of abbreviated 100%oxygen regimens since Scheinkestel’s publication —these are unproven.

• There is short-term recovery advantage, andpotentially reduced hospital stay; however, lessconvincing long-term outcome advantage for HBOover NBO in preventing Persistent Early Morbidity.

• HBO has case series and basic science supportingits safety and efficacy in treating CO-poisonedpregnant patients.

• There is a treatment advantage for HBO inpreventing DNS in a majority of studies.We acknowledge the need for more precise outcome

definitions and further refinement of oxygen dose,titrating treatments to individual patient CO load, aswell as further RCT of treatment. The RCT by Weaverand colleagues is nearing completion, and should shedfurther light on this complex field.11

References1. Doherty S. History, pathophysiology, clinical presentation and

role of hyperbaric oxygen in acute carbon monoxide poisoning.Emerg. Med. 2000; 12: 55–61.

2. Gorman D. Carbon monoxide: From toxic poison to brainmessenger. S.P.U.M.S. J. 1995; 25: 77–81.

3. Min SK. A brain syndrome associated with delayedneuropsychiatric sequelae following acute carbon monoxidepoisoning. Acta Psychiatr. Scand. 1986; 73: 80–6.

4. Mark PD. Carbon monoxide poisoning: A review. S.P.U.M.S. J.1992; 22: 127–35.

5. Pace N, Strajman E, Walker EL. Acceleration of carbonmonoxide elimination in man by high pressure oxygen. Science1950; 111: 652–4.

6. Smart DR, Oxer HF, Mark PD, Banham NDG. Carbon monoxideoff gassing: A new modality for assessing the treatment endpoint in CO poisoning. Emerg. Med. 1995; 7: 240.

7. Ducasse JL, Celsis P, Marc Vergnes JP. Non-comatose patientswith acute carbon monoxide poisoning: Hyperbaric ornormobaric oxygenation? Undersea Hyperb. Med. 1995; 22:9–15.

8. Scheinkestel CD, Bailey M, Myles PS et al. Hyperbaric ornormobaric oxygen for acute carbon monoxide poisoning: Arandomised controlled clinical trial. Med. J. Aust. 1999; 170:203–10.

9. Raphael JC, Elkharrat D, Jars-Guincestre MC et al. Trial ofnormobaric and hyperbaric oxygen for acute carbon monoxideintoxication. Lancet 1989; 2: 414–19.

10. Gorman DF, Clayton D, Gilligan JE, Webb RK. A longitudinalstudy of 100 consecutive admissions for carbon monoxidepoisoning to the Royal Adelaide Hospital. Anaes. Intens. Care1992; 20: 311–16.

11. Weaver LK, Hopkins RO, Larson-Lohr V, Howe S, Haberstock D.Double Blind, controlled, prospective, randomized clinical trialin patients with acute carbon monoxide poisoning: Outcome ofpatients treated with normobaric oxygen or hyperbaric oxygen— An interim report. Undersea Hyperb. Med. 1995; 22: S14.

12. Mathieu D, Wattel F, Mathieu-Nolf M et al. Carbon monoxidepoisoning randomized, prospective study comparing the effectsof HBO versus 12 h NBO in non-comatose CO poisoned


356

patients. Proceedings of the Twelfth International Congress onHyperbaric Medicine, 1996, Milano, Italy. Flagstaff, AZ: BestPublishing Company, 1996.

13. Kehat I, Shupak A. Hyperbaric oxygen vs. normobaric oxygen incarbon monoxide poisoning. Undersea Hyperb. Med. 2000; 27:47 (Letter).

14. Myers RAM, Snyder SK, Emhoff TA. Subacute sequelae ofcarbon monoxide poisoning. Ann. Emerg. Med. 1985; 14:1163–7.

15. Thom SR, Taber RL, Mendiguren II, Clark JM, Hardy KR, FisherAB. Delayed neuropsychologic sequelae after carbon monoxidepoisoning: Prevention by treatment with hyperbaric oxygen.Ann. Emerg. Med. 1995; 25: 474–80.

16. Elkharrat D, Raphael J, Korach J et al. Acute carbon monoxidepoisoning and hyperbaric oxygen in pregnancy. Intens. CareMed. 1991; 17: 289–92.

17. Koren G, Sharav T, Pastuszak A et al. A multicentre,prospective study of fetal outcome following accidental carbonmonoxide poisoning in pregnancy. Reprod. Toxicol. 1991; 5:397–403.

David R Smart

Royal Hobart Hospital,

Tas., Australia

Paul D Mark

Fremantle Hospital,

WA, Australia

Reply

I thank Drs Smart and Mark for their interest and letteron carbon monoxide (CO) poisoning. I welcome theirnoting that CO has been shown to act as a secondmessenger, a mechanism of toxicity not included in myreview.1

With regard to the statement ‘…multiple studieshave failed to show any firm scientific benefit in theuse of hyperbaric oxygen…’, this is still supported bythe lack of adequate scientific studies, as outlined inmy review. The two additional studies cited by Smartand Mark were by Ducasse et al.2 and Smart et al.3

(published as an abstract). The paper by Ducasse et al.2

was deliberately excluded from my review, which onlyconsidered prospective trials of more than 30 patients.This study enrolled 26 patients, all with a GCS of 15,except one with a GCS of 14. There was no pre-treatment neuropsychiatric assessment. Follow-upassessment involved EEG recordings in 18 patients andcerebral blood flow in 10 patients. Smart and Mark arecorrect in asserting that this study found earlierrecovery with hyperbaric oxygen (HBO), yet the lownumbers make this a low-power study. In addition, the

initial EEG were performed within the first 24 h and itis not clear whether this was prior to or after treatmentin some, none or all of the patients. Of the 13 patientstreated with HBO, four initially had class II EEGchanges. Five of these patients were lost to follow upand it is not clear if this included any or all of the fourwith class II changes. Ducasse et al.2 conclude bystating that when a hyperbaric facility is notimmediately available, as is the case in the vastmajority of Australasian hospitals, that normobaricoxygen (NBO) should be administered and that they‘…propose to treat with HBO all the patients who didnot fully recover clinically two hours after breathingpure O2’.

With regard to delayed neurological sequelae (DNS),the bias and flaws in the papers by Myers et al.,4

Gorman et al.,5 Thom et al.6 and Scheinkestel et al.7

were all elucidated in the review article. Smart andMark have re-cited the results without listing the flaws,except with the paper by Scheinkestel et al.,7 a paperwhich is contradictory to their views on HBO. All thesestudies have flaws and do not stand up to rigorousscientific scrutiny, hence my assertion that there is no‘firm scientific basis’ to support HBO. The article byWeaver et al.8 is on-going and still blinded and it isdifficult to evaluate the study by Smart et al.3 in itsabstracted form.

Raphael et al.9, in a study including 629 patients,concluded that ‘HBO had no advantage over NBO inthe treatment of patients without loss ofconsciousness’ and that in patients with loss ofconsciousness, two HBO sessions had no benefit overone.

In his editorial, Emerson10 noted ‘there are nodouble-blind controlled studies of the use ofhaemodialysis for severe methanol and ethylene glycolpoisoning, but haemodialysis is still performed’. This istrue, but haemodialysis is more readily available inmost intensive care units, including rural centres.Hyperbaric centres are few in number, with mostcapital cities only possessing one. To treat all cases ofCO poisoning with HBO would necessitate a largenumber of transfers, including medical retrievals ofobtunded patients and aeromedical evacuation of ruralpatients. This would be at a substantial cost. Clearevidence of benefit is required to warrant such anapproach.

A point well made by Smart and Mark, and also byEmerson10 is that if an NBO approach is taken on thebasis of Scheinketsel et al.’s7 study, then this mandates3 days of 100% oxygen therapy.

357


I did mention HBO as a treatment option forpregnant women, mentioning why the foetus is moresusceptible and quoted from Jain,11 that up to fivetimes the duration of HBO has been recommended forpregnant patients. The debate about HBO therapy hasbeen ongoing for many years. Even hyperbaricspecialists have questioned whether HBO is a ‘cure insearch of a disease’.12 My review of HBO presented theevidence and discussed the flaws in this evidence.Smart and Mark may well prove to be correct inasserting ‘evidence suggests a treatment advantage forHBO in preventing DNS’. This does not alter the factthat the studies, at present, are scientifically flawed andare not conclusive. If there is benefit from HBO, it isstill not known whether this benefit extends to allpatients or only to the most seriously poisoned. AsSmart and Mark suggest, it is hoped that the study byWeaver et al.8 may shed further light on this area.

References1. Doherty S. History, pathophysiology, clinical presentation and

role of hyperbaric oxygen in acute carbon monoxide poisoning.Emerg. Med. 2000 12: 55–61.

2. Ducasse JL, Celsis P, Marc Vergnes JP. Non-comatose patientswith acute carbon monoxide poisoning: Hyperbaric ornormobaric oxygenation? Undersea Hyperb. Med. 1995; 22:9–15.

3. Smart DR, Oxer HF, Mark PD, Banham NDG. Carbon monoxideoff gassing: A new modality for assessing the treatment endpoint in CO poisoning. Emerg. Med. 1995; 7: 240.

4. Myers RAM, Snyder S, Emhoff TA. Subacute sequelae ofcarbon monoxide poisoning. Ann. Emerg. Med. 1985; 14:1163–7.

5. Gorman DF, Clayton D, Gilligan JE, Webb RK. A longitudinalstudy of 100 consecutive admissions for carbon monoxidepoisoning to the Royal Adelaide Hospital. Anaes. Intens. Care1992; 20: 311–16.

6. Thom Sr, Taber RL, Mendiguren II, Clark JM, Hardy KR,Fisher AB. Delayed neuropsychological sequelae after carbonmonoxide poisoning: Prevention by treatment with hyperbaricoxygen. Ann. Emerg. Med. 1995; 25: 474–80.

7. Scheinkestel CD, Bailey M, Myles PS et al. Hyperbaric ornormobaric oxygen for acute carbon monoxide poisoning: Arandomized controlled clinical trial. Med. J. Aust. 1999; 170:203–10.

8. Weaver LK, Hopkins RO, Larson-Lohr V, Howe S, Haberstock D.Double-blind, controlled, prospective, randomized clinical trialin patients with acute carbon monoxide poisoning: Outcome ofpatients treated with normobaric or hyperbaric oxygen — Aninterim report. Undersea Hyperb. Med. 1995; 22: S14.

9. Raphael J-C, Elkharrat D, Jars-Guincestre M-C et al. Trial ofnormobaric and hyperbaric oxygen for acute carbon monoxideintoxication. Lancet 1989; 2: 414–19.

10. Emerson G. Hyperbaric medicine and carbon monoxidepoisoning: Oxygen under pressure. Emerg. Med. 2000; 12: 9–10.

11. Jain KK. Textbook of Hyperbaric Medicine. Boston: Hogrefe andHuber, 1990.

12. Torda T. Hyperbaric oxygen — Its role in poisoning.Proceedings, Australasian Clinical Toxicology Conference,September 1995, Sydney, Australia, 1995; A23–6.

Steven Doherty MB BS, FACEM

Emergency Physician, Tamworth Base Hospital

Tamworth, NSW, Australia

Editor’s Note

This correspondence is now closed. Readers who wishto continue exploring this controversy are directed to asimilar airing of the argument concerning the use ofhyperbaric oxygen in acute carbon monoxide poisoningin the ‘Letters’ section of 8 July British Medical Journal(2000; 321: 109–111) where Scheinkestel CD et al.respond to comments made by Weaver LK in hisprevious Editorial (BMJ 1999; 319: 1083–4). TigheSQM also concludes that 100% oxygen is the bestoption and Weaver LK continues to disagree with themboth!

Assoc. Prof. Anthony FT Brown,

Editor, Emergency Medicine

Declaring conflict of interest

I was disappointed by Dr Brendon Smith’s commentsin his Letter to the Editor (Emerg. Med. 2000; 12:157–8).

An otherwise important issue (conflict of interest)was clouded by an initial statement that was intendedto cast doubt on the integrity of the organizers of ascientific meeting. To quote from Dr Smith’s firstparagraph: ‘The recent Australasian College forEmergency Medicine and Australasian Society forEmergency Medicine Scientific Conference inAuckland, New Zealand, included presentations highlyfavourable to the sponsors’ products but denieddelegates the opportunity to consider potential bias’.

I would like to address the two inferences in thisstatement. My response is based on my absoluteknowledge of the organization of the AucklandConference.

The ACEM/ASEM Annual Scientific Meeting(ASM) relies on company sponsorship to subsidizecosts and to keep the conference affordable to Fellows,


358

trainees and allied clinical personnel. The scientificcontent of the meeting is the responsibility of theScientific Committee and is the basis on which theconference will be considered educational. These twoaspects of the ASM are organizationally separate,although company commercial interest is obviously thebasis of sponsorship. No speaker at the AucklandASM was paid by an individual company, despiteinclusion in one of the two sponsored symposia. Nospeaker selection was influenced by an individualcompany. Sponsorship monies were paid into theconference budget, from which expenses for thekeynote speakers were provided according to theguidelines drafted by ACEM. All abstracts were

accepted on merit, and companies were not involvedwith the selection of speakers or content. Clearly, theScientific Committee cannot influence the views heldby the speakers with regard to particular products.

Dr Smith also states that delegates were denied theopportunity to consider potential bias. Surely the paneldiscussion time would have provided the opportunityto challenge any content or bias of the speakers?

I trust that this clarifies the facts and reassuresFellows of the College that the ASM is conducted withfull awareness of the potential for conflict of interest.

Peter Freeman

ASM ’99 Convenor

359


Documents

Declaring conflict of interest