92 Turner, Hamilton-Farrell, Clark

oxygen. Anecdotal reports of alternative ap- presented in this issue of the journal. One cer- proaches have not been followed up.'0 There is tainly wants to avoid the error of telling a little evidence from clinical studies that patient with undiagnosed carbon monoxide J Accid Emerg Med: first published as 10.1136/emj.16.2.92 on 1 March 1999. Downloaded from normobaric oxygen is effective, and even the poisoning to "go home and stay warm", with evidence concerning hyperbaric oxygen is dif- the result that they suffer even worse poison- ficult to interpret despite animal studies and ing. cellular research which provide strong evi- dence in its favour. One important factor is 1 Chief Medical Officer and Chief Nursing Officer. Carbon monoxide: the forgotten kilter. London: Department of that it reduces free radical concentrations, and Health, September 1998. this means that the benefit from hyperbaric 2 Chief Medical Officer. CO poisoning: pulse oximetry. oxygen may be far reaching. It provides an CMO's update 20. London: Department of Health, November 1998: 3. explanation why hyperbaric oxygen prevents 3 Carbon monoxide: health hazards and precautionary meas- the delayed manifestations of carbon monox- ures. Guidance note EH43. 2nd Ed. London: Health and Safety Executive, 1998. ide poisoning. 4 Smith JS, Brandon S. Morbidity from acute carbon monox- The presentation by Hamilton-Farrell ofthe ide poisoning at three year follow up. BMJ 1973;i:318-21. 5 Chale SN. Carbon monoxide poisoning. In: Vicciello P, ed. British Hyperbaric Association's carbon mon- Handbook ofmedical toxicology. Boston: Little, Brown 1993: oxide database (p98) provides an impressive 639-47. 6 Choi IS, Kim SK, Lee SS, et al. Evaluation of outcome of panorama of the use of hyperbaric oxygen in delayed neurologic sequelae after carbon monoxide poison- Britain." The indications for referral are ing by technetium 99m hexamethylpropylene amine oxime brain singe photon emission computed tomography. Eur clearly set out, and the reasons for delay to Neurol 1995;35: 137-42. referral and delay to arrival in a hyperbaric 7 Kirkpatrick JN. Occult carbon monoxide poisoning. West J7 facility are listed. Many of these can be Med 1987;146:52-6. 8 Turner M, Esaw M, Clark RJ. Carbon monoxide poisoning reduced by making a critical examination at a treated with hyperbaric oxygen: metabolic acidosis as a local level and addressing the deficiencies predictor of treatment requirements. J Accid Emerg Med 1999;16:96-8. identified. The rising number of referrals may 9 Turner M, Hamilton-Farrell MR, Clark RJ. Carbon monox- be a result of increased awareness, since the ide poisoning: an update. J Accid Emerg Med 1999;16: 92-6. number of cases of fatal poisoning is falling. It 10 Howard RJ, Lake DR, Pall H, et al. Allopurinol/N- has been suggested that this is due to the acetylcysteine for carbon monoxide poisoning. Lancet 1987;ii:628-9. introduction of catalytic converters in motor 11 Hamilton-Farrell MR. British Hyperbaric Association vehicles. 2 carbon dioxide database, 1993-96. Y Accid Emerg Med 1999;16:98-103. Those working in any branch of emergency 12 Kendell RE. Catalytic converters and prevention of suicide. medicine would do well to study the articles Lancet 1998;352:1525.

Carbon monoxide poisoning: an update http://emj.bmj.com/ Mark Turner, Martin R Hamilton-Farrell, Robin J Clark

Carbon monoxide is the major cause of death Death may be due to cardiac toxicity,

from poisoning in the UK. In 1996 there were neurotoxicity, systemic acidosis, or respiratory on September 29, 2021 by guest. Protected copyright. 877 deaths from carbon monoxide poisoning arrest. (CMP) in England and Wales; the majority were suicides due to car exhaust fumes.' Pathophysiology The commonly held beliefthat CMP is mainly due to binding of haemoglobin was shown to Department of Causes be false over 20 years ago.6 Goldbaum et al Cardiology, Derriford The most common causes of CMP are demonstrated that exchange transfusion, using Hospital, Plymouth, deliberate attempted suicide using car exhaust a red cell concentrate with a carboxyhaemo- Devon PL6 8HD fumes, domestic heater malfunction, open globin (COHb) concentration of 80%, led to M Turner heating fires, burning buildings, and indefinite survival in one group of dogs, Whipps Cross accidental exposure to car exhaust. Rarer despite equilibrium COHb concentrations of Hospital, London causes include recreational boating,2 indoor 60%. The potential to endure a diminished M R Hamilton-Farrell burning of charcoal briquettes,' obstruction oxygen carrying capacity was demonstrated by snow of vehicle exhaust systems,4 and the when other dogs were venesected isovolaemi- Royal Naval Hospital hepatic metabolism of methylene chloride cally, reducing mean packed cell Haslar, Gosport volume by R J Clark from paint stripper fumes.5 The British 68%, and survived indefinitely. In contrast, Hyperbaric Association (BHA) CMP data- dogs in a third group inhaled 13% carbon Correspondence to: base suggests that, for cases managed with monoxide for 15 minutes, producing a mean Dr Turner, Specialist hyperbaric oxygen, approximately half are COHb of 65%, and all died. Goldbaum et al Registrar in Cardiology. deliberate and half accidental.5 This contrasts deduced that the slow rate of carbon monox- Accepted 10 September with the mortality figures, with 85% of CMP ide binding by haemoglobin led to blood 1998 deaths being due to deliberate poisoning. entering the systemic capillary beds with a Carbon monoxide poisoning 93

Clinical manifestations Box 1: Clinical manifestations of car- Severe CMP is characterised by neurological bon monoxide poisoning and cardiovascular manifestations (box 1). J Accid Emerg Med: first published as 10.1136/emj.16.2.92 on 1 March 1999. Downloaded from Neurological Unconsciousness is the most common neuro- * Loss of consciousness logical manifestation although memory loss, * Neurological and cognitive abnormality drowsiness, poor cognitive performance, and * Gait disturbance disturbance of balance or gait (including Cardiovascular cerebellar features) may be seen. Less severe * Hypotension cases may have non-specific symptoms includ- * Cardiac ischaemia ing headache, nausea, vomiting, and lethargy. * Arrhythmia In such cases a careful history should be taken Biochemical to identify potential exposure to carbon * Carboxyhaemoglobinaemia monoxide. It is possible that many mild cases * Metabolic acidosis are never diagnosed and there is evidence that much late morbidity from CMP goes unrecognised.'4 The commonly quoted cherry red discolora- tion mucous may be significant burden of dissolved carbon monox- of skin and membranes ide available for diffusion. Conversely, in the seen at autopsy but is extremely rare in patients exchange transfusion experiment described presenting with acute CMP.'5 The absence of above, minimal dissolved carbon monoxide the cherry red colour does nothing to exclude had been infused and the carbon monoxide the diagnosis.'6 was so tightly bound to haemoglobin that there was little tissue burden. The paper con- Clinical assessment cluded that the toxicity of carbon monoxide is A careful history, specifically seeking all of the predominantly a direct action on cells, rather possible manifestations listed above should be than mainly the denial of oxygen carriage by taken. The type of heating in the patient's haemoglobin. home should be established. A venous COHb The mechanism for tissue damage in CMP concentration should be performed if there is is the poisoning of intracellular oxygen any suspicion of the diagnosis. carrying haem proteins, such as cytochrome a, Even in obvious and severe cases, clinical and myoglobin. This leads to cellular history and examination are most important. dysfunction and produces the clinical mani- Specific features of the history should be festations most rapidly in tissues with high sought: cause, duration of exposure, loss of energy requirements (brain and heart). Fur- consciousness (even if transient), the ther injury is caused by radical oxygen species, possibility of other poisons (cyanide or drugs), including nitric oxide, produced locally by and chest pain. A history of possible chronic neutrophils.7 exposure should be sought in accidental cases, Hyperglycaemia and hypoglycaemia have an since this may have preceded the acute http://emj.bmj.com/ adverse effect in rats poisoned with carbon presentation. Neurological examination monoxide.8 Investigation of the brain energy should include assessment of cognitive func- metabolites ATP and phosphocreatine failed to tions and in particular memory; patients define clearly a mechanism for this finding, should all be asked to stand and walk, if they although lactate concentrations were lower in are able, because balance and gait disturbance the normoglycaemic animals than in those with may be the only abnormality in the physical hyperglycaemia.9 The role of plasma glucose in examination. on September 29, 2021 by guest. Protected copyright. humans has not been investigated, however it Most patients with severe poisoning will present with a reduced level of consciousness, would seem logical to maintain normoglycae- so bedside blood glucose analysis should be mia in patients. performed immediately to exclude hypoglycae- Ethanol has been shown to modify the mia, which may complicate CMP or may be severity of poisoning in mice.'0 Survival time the sole cause of unconsciousness. The COHb was improved in mice pretreated with concentration should be measured in either ethanol. The authors propose that the ethanol venous or arterial blood (no significant differ- reduced blood flow and hence the rate of ence in value is obtained'7) in all patients with delivery of carbon monoxide to the tissues, undiagnosed unconsciousness: greater than resulting in a lower total body carbon monox- 10% is diagnostic of CMP in all but the heavi- ide burden. est of tobacco smokers. In both the acute and chronic phases of Pulse oximetry may give falsely high CMP, cerebral perfusion has been shown to be readings, since COHb absorbs light at an abnormal. In the acute phase, large focal almost identical frequency to that of defects may occur" whereas diffuse abnormal- oxyhaemoglobin: thus the measured value ity is present in patients with late sequelae."' It is the sum of oxyhaemoglobin and remains unclear whether abnormal perfusion is carboxyhaemoglobin.'8 Arterial blood gases a cause of brain injury or the result of other should therefore be measured using a pathological mechanisms. Assessment of cer- co-oximeter, to assess oxyhaemoglobin and ebral perfusion by isotope imaging has been COHb, to confirm adequacy of spontaneous suggested as a possible indicator of severity of or assisted ventilation, and to measure pH. poisoning."3 Measurement of lactate may be helpful, if 94 Turner, Hamilton-Farrell, Clark

available. Venous blood should also be drawn HYPERBARIC OXYGEN TREATMENT for assays of other toxins and glucose. Currently accepted indications for hyperbaric

Electrocardiography (ECG) should be per- oxygen (Guy's Hospital Poisons Unit) are: loss J Accid Emerg Med: first published as 10.1136/emj.16.2.92 on 1 March 1999. Downloaded from formed in all cases because ischaemic of consciousness (even iftransient), neurologi- changes, ST depression, or even ST elevation cal or cognitive abnormality, ECG evidence of (myocardial infarction pattern) can occur cardiac ischaemia (ST abnormality), COHb despite normal coronary arteries.'9 ECG >20%, and pregnancy (as the fetus cannot be monitoring is important, because the presence assessed and is more susceptible to poisoning of arrhythmias may influence decisions about than an adult). Although blood COHb the wisdom of interhospital transfer, although measured at the time of admission helps to haemodynamically significant arrhythmia is confirm the diagnosis, it does not always rare in our experience. If ST elevation is correlate well with severity of poisoning.2' present in acute CMP, thrombolysis is prob- Both the type of exposure and the timing of ably not appropriate since the cause is measurement, relative to the removal from generally myocardial toxicity rather than exposure, contribute to this. A prolonged thrombotic occlusion of an epicardial coron- exposure to relatively low concentrations of ary artery. Reversible biventricular dysfunc- carbon monoxide may produce a large tissue tion, leading to haemodynamic instability, is load of carbon monoxide with low COHb, also described after CMP20 due to the toxic whereas a short exposure to a high concentra- effect of carbon monoxide on cardiac myo- tion can produce a low tissue load (less severe) cytes. and a high COHb. Despite this it is considered by some an indication for hyperbaric oxygen. Interestingly, of patients in the BHA database, Treatment, indications for hyperbaric 298 (51.8%) had recorded COHb concentra- oxygen, and assessment of severity tions exceeding 25%; of these cases all but Carbon monoxide binds more avidly than oxy- seven satisfied the criteria for hyperbaric oxy- gen to haemoglobin and to tissue haem gen other than the COHb concentration proteins, but these remain reversible processes. alone. This included patients in whom neuro- The premise underlying treatment is that logical abnormalities were not reported at the removal of carbon monoxide will minimise fur- time of referral for hyperbaric oxygen; abnor- ther damage and reverse cellular metabolic malities were either not recognised initially or dysfunction. The half life of COHb is variable developed later, in transit. Therefore, despite but typically 270 minutes when breathing air, the lack of specific evidence that arbitrary compared with 90 minutes when breathing concentrations of COHb alone are an indica- 100% oxygen2" and 25 minutes while breathing tion for hyperbaric oxygen, in practice high oxygen at a pressure of 3 bar. Cardiac output concentrations of COHb may be an indication (hence circulation time) is an additional deter- that referral is necessary but is rarely the only minant of clearance rate. pretext for treatment. Recovery depends on prompt removal from Hyperbaric oxygen has been shown to

the toxic exposure and institution of effective reduce the free radical concentration where http://emj.bmj.com/ resuscitation for life threatening cases. 100% normobaric oxygen did not.24 The para- dox that higher partial pressures of oxygen NORMOBARIC OXYGEN TREATMENT reduce oxygen free radical production is Although there is controversy surrounding explained by the observation that neutrophil hyperbaric oxygen, there is little argument that adherence to the vascular endothelium medi- 100% normobaric oxygen should be adminis- ated by P2 integrins is reduced by hyperbaric

tered. There are no studies suggesting a good oxygen. The mechanism for this appears to be on September 29, 2021 by guest. Protected copyright. outcome without any treatment, even in mild an effect on the membrane guanylate cyclase cases, therefore it should be considered man- of the neutrophil.25 These data suggest that datory to provide oxygen. As soon as the diag- any benefit from hyperbaric oxygen may not nosis is suspected, high concentration oxygen be solely an effect on the half life of carbon should be delivered via a tight fitting facemask monoxide in the blood and tissues, but also a and non-rebreathing reservoir bag, aiming to local action in the tissues preventing further achieve as high a concentration as possible. In injury. order to provide 100% oxygen, an air filled In the absence of a controlled clinical trial cushion rimmed facemask, or an endotracheal directly comparing hyperbaric oxygen with tube with the cuff inflated, is necessary. If normobaric oxygen this area remains hyperbaric oxygen is not administered, six controversial, as discussed in an article hours of treatment with 100% oxygen at ambi- published in 1994.26 Since then, further ent pressure, as has been compared with important data have emerged suggesting that hyperbaric oxygen in a randomised trial,22 is hyperbaric oxygen reduces the incidence of recommended as a minimum. If any symptoms late sequelae27: new neurological or cognitive persist, the oxygen should be continued and deficits presenting between three and 21 hyperbaric oxygen should be considered. Be- days after poisoning that contribute to the fore hospital discharge, all patients should be unrecognised chronic morbidity associated carefully re-examined to ensure that no neuro- with CMP.'4 It had been thought that the out- logical or cognitive abnormality (which would come, in patients who never lose conscious- indicate the need for hyperbaric oxygen) has ness at the time of original poisoning, is not developed during normobaric oxygen treat- changed by hyperbaric oxygen22; however, ment. this study compared hyperbaric oxygen and Carbon monoxide poisoning 95

concentrations." A normal tomogram has Box 2: Indications for hyperbaric been correlated with a good outcome,32 oxygen although this does not exclude clinical J Accid Emerg Med: first published as 10.1136/emj.16.2.92 on 1 March 1999. Downloaded from * Loss of consciousness problems. In the same series, abnormalities on * Neurological abnormality the computed tomography also correlated * Cardiac arrhythmia/ischaemia with a poorer outcome despite hyperbaric * COHb >20% oxygen. It remains unclear if early computed * Pregnancy tomography has any role in risk stratification. Magnetic resonance imaging (MRI) has also been performed in a small number of normobaric oxygen and used a self adminis- patients and has a greater sensitivity for tered questionnaire as its outcome measure, white matter abnormalities." Again insuffi- which may not have been a sensitive enough cient data are available to judge if there is a tool to identify subtle cognitive problems. place for MRI in the acute situation. There is Recent data suggest that patients who never certainly a place for brain imaging if patients lose consciousness derive early benefit from fail to make a complete recovery after hyperbaric oxygen, since they recover more hyperbaric oxygen, to exclude coexistent quickly.28 Until more data become available, pathology.'4 we recommend using the guidelines from the National Poisons Unit at Guy's Hospital (box 2). Prevention Unconsciousness is the most common Regular servicing of gas appliances, adequate reason for referral for hyperbaric oxygen; 65% ventilation, and cleaning of chimneys should of the patients in the BHA database (374 of prevent domestic exposures. Carbon monox- 575 cases) had a definite history of loss of ide sensors are commercially available, but consciousness.5 Patients who never lose con- may not detect the low levels that could cause sciousness and have no other indication for clinical manifestations if repeated or pro- hyperbaric oxygen, despite a thorough clinical longed exposure occurs. They do, however, assessment, may be treated with normobaric appear to alert victims to heavy exposures and oxygen.22 General practitioners should be con- encourage early presentation.'5 Catalytic con- tacted and advised to assess regularly patients verters reduce the carbon monoxide content in the first three weeks after poisoning, as late of exhaust gases and may help to reduce the sequelae may develop in this period. Late overall severity of exposures to car exhaust sequelae have been shown to respond to fumes. hyperbaric oxygen29 and therefore patients should be referred for this treatment in this circumstance. Summary Follow up of 50 accidentally poisoned CMP is an important toxicological patients presenting to Whipps Cross Hospital emergency. The majority of cases present to

has been undertaken from 1993 to 1996, three accident and emergency departments, a few to http://emj.bmj.com/ months after initial presentation (personal general practitioners. In some cases the communication, M Hamilton-Farrell). In this diagnosis is obvious, but in other cases group all patients with a history of carbon manifestations may be non-specific when a monoxide exposure for longer than 24 hours high index of suspicion, a careful history, and had residual symptoms, whereas 75% of those venous or arterial COHb are necessary to with shorter exposures were asymptomatic. establish the diagnosis. A thorough clinical assessment, including arterial blood gases and More data will be collected to investigate this on September 29, 2021 by guest. Protected copyright. observation. ECG, are necessary to estimate severity of Metabolic acidosis may provide a closer poisoning. The role of imaging investigations reflection of tissue poisoning and has been is unclear in the acute phase. Computed tom- shown to be associated with more severe ography or MRI of the brain should be poisoning as judged by treatment requirements considered if patients do not make a complete and was certainly a better indicator of severity neurological recovery after hyperbaric oxygen than COHb.'0 More specific markers of to exclude other pathology. cellular dysfunction, such as lactate, should be For clinical advice and access to hyperbaric oxygen treatment investigated and may prove helpful, but clinical contact: the Royal Navy duty diving medical officer (hyperbaric evaluation remains the most valuable tool for medicine clinician), Institute of Naval Medicine, Gosport, tel: 0831 151523 (24 hour service) or the Hyperbaric Medical initial assessment of severity. Centre (DDRC), Plymouth 24 hour carbon monoxide national telephone advice service, tel: 01752 209999/261910 (location Brain imaging of nearest chamber and contact information). The place of imaging investigations has not 1 Office for National Statistics. Mortality statistics cause: been established in the acute phase. Brain England and Wales. London: HMSO, 1996. Series DH2 No 23. computed tomography is not indicated before 2 Silvers SM, Hampson NB. Carbon monoxide poisoning hyperbaric oxygen unless there is doubt about among recreational boaters. JAMA 1995;274: 1614-6. 3 Hampson NB, Kramer CC, Dunford RG, et al. Carbon the cause of unconsciousness. monoxide poisoning from indoor burning of charcoal The most common abnormalities noted on briquets. JAMA 1994-271:52-3. 4 Rao R, Touger M, Gennis P, et al. Epidemic of accidental computed tomography of the brain are low carbon monoxide poisonings caused by snow-obstructed density lesions in the globus pallidus and deep exhaust systems. Ann Emerg Med 1997;29:290-2. 5 Hamilton-Farrell MR. British Hyperbaric Association white matter changes, although these changes carbon monoxide database, 1993-96. Jf Accid Emerg Med do not correlate with outcome or GOHb 1999;16:98-103. 96 Turner, Esaw, Clark

6 Goldbaum LR, Orellano T, Dergal E. Mechanism of the 21 Levassuer L, Galliot-Guilley M, Richter F, et al. Effects of toxic action of carbon monoxide. Ann Clin Lab Sci 1976;6: mode of inhalation of carbon monoxide and of normobaric 372-6. oxygen administration on carbon monoxide elimination 7 Thom SR. Carbon monoxide mediated brain lipid peroxida- from the blood. Hum Exp Toxicol 1996;15:898-903. J Accid Emerg Med: first published as 10.1136/emj.16.2.92 on 1 March 1999. Downloaded from tion in the rat.JAppl Physiol 1990;68:997-1003. 22 Raphael JC, Elkharrat D, Jars-Guincestre MC, et al. Trial of 8 Penney DG, Helfman CC, Dunbar JC Jr, et al. Acute severe normobaric and hyperbaric oxygen for acute carbon mon- carbon monoxide exposure in the rat: effects of hypergly- oxide intoxication. Lancet 1989;ii:414-19. caemia and hypoglycaemia on mortality, recovery and neu- 23 Norkool DM, Kirkpatrick JN. Treatment of acute carbon rologic deficit. Can J Physiol Pharmacol 199 1;69: 1168-77. monoxide poisoning with hyperbaric oxygen: a review of 9 Matsuoka M, Igisu H, Tanaka I, et al. Effects of hypo- and 115 cases. Ann Emerg Med. 1985;14:1168-71. hyperglycaemia on brain energy metabolites in mice 24 Thom SR. Antagonism of carbon monoxide-mediated brain exposed to carbon monoxide. Toxicol Lett 1994;73:135-43. lipid peroxidation by hyperbaric oxygen. Toxicol Appl Phar- 10 Tomita M, Okuyama T, Shimosato K, et al. Effect of ethanol macol 1990;105:340-4. on fatal carbon monoxide poisoning in awake mice. Toxicol 25 Thom SR. Functional inhibition of luekocyte f2 integrins by Lett 1990;50:151-7. hyperbaric oxygen in carbon monoxide-mediated brain 11 Denays R, Makhoul E, Dachy B, et al. Electroencephalo- injury in rats. Toxicol Appl Pharmacol 1993;123:248-56. graphic mapping and '9mTc-HMPAO single-photon emis- 26 Tibbles PM, Perrotta PL. Treatment of carbon monoxide sion computed tomography in carbon monoxide poisoning. poisoning: a critical review of human outcome studies Ann Emerg Med 1994;24:947-52. comparing normobaric oxygen with hyperbaric oxygen. 12 Choi IS, Kim SK, Lee SS, et al. Evaluation of outcome of Ann Emerg Med 1994;24:269-76. delayed neurological sequelae after carbon monoxide 27 Thom SR, Taber RL, Mendiguren II, et al. Delayed poisoning by technetium-99m hexamethylpropylene amine neuropsychologic sequelae after carbon monoxide oxime brain single photon emission computed tomography. poisoning: prevention by treatment with hyperbaric oxy- Eur Neurol 1995;35: 137-42. gen. Ann Emerg Med 1995;25:474-80. 13 Turner M, Kemp PM. Isotope brain scanning with 28 Ducasse JL, Celsis P, Marc-Vergnes JP. Non-comatose Tc-HMPAO: a predictor of outcome in carbon monoxide patients with acute carbon monoxide poisoning: hyperbaric poisoning? _rAccid Emerg Med 1997;14: 139-41. or normobaric oxygen? Undersea Hyperb Med 1995;22:9-15. 14 Smith JS, Brandon S. Morbidity from acute carbon monox- 29 Myers RAM, Snyder SK, Emhoff TA. Sub-acute sequelae of ide poisoning at three-year follow-up. BMJ7 1973;i:318-21. carbon monoxide poisoning. Ann Emerg Med 1985;14: 15 Gorman DF, Clayton D, Gilligan JE, et al. A longitudinal 1163-7. study of 100 consecutive admissions for carbon monoxide 30 Turner M, Esaw M, Clark RJ. Carbon monoxide poisoning poisoning to the Royal Adelaide Hospital. Anaesth Intensive treated with hyperbaric oxygen: metabolic acidosis as a Care 1992;20:311-16. predictor of treatment requirements. J Accid Emerg Med 16 Turner M. The dangers of carbon monoxide poisoning. N 1999;16:96-8. EnglJ Med 1995;332:894. 31 Tom T, Abedon S, Clark RI, et al. Neuroimaging character- 17 Touger M, Gallagher EJ, Tyrell J. Relationship between istics in carbon monoxide toxicity. J Neuroimaging 1996;6: venous and arterial carboxy-haemoglobin levels in patients 161-6. with suspected carbon monoxide poisoning. Ann Emerg 32 Pracyk JB, Stolp BW, Fife CE, et al. Brain computerised Med 1995;25:481-3. tomography after hyperbaric oxygen therapy for carbon 18 Buckley RG, Aks SE, Eschom JL, et al. The pulse oximetry monoxide poisoning. Undersea Hyperb Med 1995;22:1-7. gap in carbon monoxide intoxication. Ann Emerg Med 33 Horowitz AL, Kaplan R, Sarpel G. Carbon monoxide 1994;24:252-5. toxicity: MR imaging in the brain. Radiology 1987;162: 19 Marius-Nunez AL. Myocardial infarction with normal cor- 787-8. onary arteries after acute exposure to carbon monoxide. 34 Turner M, Clark RJ. Glioblastoma presenting as carbon Chest 1990;97:491-4. monoxide poisoning. JAccid Emerg Med 1999;16:155-6. 20 McMeekin JD, Finegan BA. Reversible myocardial dysfunc- 35 Krenzelok EP, Roth R, Full R. Carbon monoxide .... the tion following carbon monoxide poisoning. Can J Cardiol silent killer with an audible solution. Am J Emerg Med 1987;3:118-21. 1996;14:484-6.

Carbon monoxide poisoning treated with hyperbaric oxygen: metabolic acidosis as a http://emj.bmj.com/ predictor of treatment requirements

Mark Turner, Michelle Esaw, Robin J Clark on September 29, 2021 by guest. Protected copyright.

Abstract tion showed only a trend to being higher A retrospective case note analysis was in the multiple treatment group. made of patients who received hyperbaric Although metabolic acidosis is well oxygen for carbon monoxide poisoning recognised, its relationship to treatment and were admitted to the Royal Naval requirements has not been shown previ- Hospital Haslar between 1991 and 1995. ously. Initial COHb does not always Males predominated (38 v 10) as did cases correlate well with severity of poisoning Department of of deliberate self poisoning (31 v 17). The which relates to the mechanism of toxic- Cardiology, Derriford most common presenting feature was ity ofcarbon monoxide: binding ofcarbon Hospital, Plymouth, monoxide to the intracellular oxygen Devon PL6 8HD unconsciousness, which is an indication M Turner for hyperbaric oxygen and therefore re- carrying proteins (for example cyto- flects referral patterns. chromes) rather than solely to haemo- Royal Naval Hospital If patients had not recovered com- globin. These findings are consistent with Haslar, Gosport pletely after one hyperbaric exposure fur- this mechanism and suggests that initial M Esaw acidosis is a better predictor of treatment R J Clark ther treatments were given. The initial hydrogen ion concentration of those requirements and severity than initial Correspondence to: requiring more than one treatment was COHb. Dr Turner, Specialist (7Accid Emerg Med 1999;16:96-98) Registrar in Cardiology. significantly higher than those who recov- ered after the first treatment. The initial Keywords: carbon monoxide poisoning; hyperbaric Accepted 4 June 1998 carboxyhaemoglobin (COHb) concentra- oxygen treatment