Carbon monoxide: not gone, not to be forgotten.

Carbon monoxide is a remarkable poison. Because of the non-specific manner in which it may present, and the consequent multiplicity of possible diagnoses, it has been described as "the silent killer". This issue of the journal contains three publications on the topic, and if that were not enough, the Chief Medical Officer has recently written to all the general practitioners in Britain warning them of the problems associated with carbon monoxide poisoning." Another useful reference document has also appeared from the Health and Safety Executive outlining the diagnosis and management of carbon monoxide toxicity with an emphasis on the industrial aspects.' It is necessary for all those working in accident and emergency departments to be aware of the diagnosis, the management, the pitfalls, and the science of carbon monoxide poisoning. Some may ask why there is such concern about carbon monoxide. The reason is that, apart from being the commonest cause of fatal poisoning in Britain and many other countries, it is now well recognised as a cause of neurotoxicity. There may be clearly evident damage, such as parkinsonian syndromes, cortical blindness, and dementia, or there may be more subtle problems, such as apathy, fatigue, irritability, minor memory disturbances, difficulty with decisions, and personality change.4 5 It is not necessary for the patient to have suffered severe or life threatening poisoning for these effects to occur. Carbon monoxide is also unusual in that delayed neurotoxicity may occur days or even weeks after apparent recovery from the episode of poisoning.' The most immediate problem is that of diagnosis. The clinical assessment of the patient may reveal very few clues. Headache, dizziness, fatigue, and nausea are common symptoms. Tachycardia, tachypnoea, postural hypotension, and signs of central nervous system depression occur in more severe cases. One should always suspect carbon monoxide when people present with non-specific symptoms, particularly in cold weather; unfortunately, this is precisely the time when viral and bacterial infections are more common. When more than one person has suffered typical symptoms, or if a pet animal has also become ill or died, the clinician should actively consider the diagnosis.7 If everyone in the premises has suffered from headache and vomiting, then carbon monoxide poisoning is by far the most likely diagnosis. The carboxyhaemoglobin (COHb) concentration is the most important diagnostic test, and blood should be taken as early as possible after the patient presents, as the blood COHb declines rapidly, especially if the patient is being given oxygen by facemask. The paper by Turner, Esaw, and Clark (p96),8 which investigates the importance of metabolic acidosis, helps to confirm that acidosis is a significant feature of carbon monoxide poisoning and also that initial acidosis is a good predictor of severity and of treatment requirements. It is clear that the initial assessment of the patient should include a full history and clinical examination, a COHb concentration, and also an arterial pH estimation. However, this publication also shows us that better markers of intracellular hypoxia are needed. What is currently being measured reflects whole body hypoxia, but it is mainly cerebral hypoxia that is responsible for the acute presentation and long term sequelae. The update by Turner, Hamilton-Farrell, and Clark (p92) is a useful review of a range of new concepts in carbon monoxide poisoning.9 Ideas about the mechanism of toxicity have developed markedly. The original simplistic picture put forward by Haldane is no longer accepted. He believed that the toxicity of carbon monoxide was secondary to its high affinity for haemoglobin, and that the resulting decrease in the oxygen carrying capacity of the blood caused end organ hypoxia. Other mechanisms are much stronger candidates for causing symptomatology and toxicity. These include combination of carbon monoxide with cytochromes and with myoglobin, inhibition of mitochondrial metabolism, competition with nitrous oxide, promotion of free radical damage and inflammatory mechanisms; their precise contributions remain to be elucidated. However, these mechanisms are not easy to detect or identify in clinical practice, and because of its ease of measurement, blood COHb remains the most important diagnostic test despite its poor predictive value. The authors have posed questions about the roles of glucose and ethanol, but the relationship of these two to acute and chronic human poisoning is not clear. Therapeutic interventions are limited to supportive care and normobaric or hyperbaric Accident and Emergency Department, St Mary's Hospital, Queen Elizabeth the Queen Mother Wing, South WharfRoad, London W2 lNY