Expired carbon monoxide as a marker of CO poisoning and its application in determining treatment End-points

Carbon monoxide (CO) is a colourless, odourless toxic gas that is able to substitute for oxygen at many levels in the oxygen cascade. CO poisoning is responsible for nearly a quarter of suicide deaths in Australia, and hundreds of individuals sustain non-fatal poisoning every year. Up to two thirds of individuals who survive CO poisoning have long-term neurological or cognitive impairment. Despite years of study by medical researchers, a reliable marker of acute CO poisoning severity that correlates with outcome has not been identified. Oxygen is known to be an antidote to CO poisoning, yet there is significant debate regarding the dose required, and the treatment duration. The end-point of CO excretion from the body is the lungs. Measurement of expired CO has been documented since the 1980's, however there has been limited study of ECO in poisoned patients. In this research ECO was investigated as marker of CO poisoning, and its application in determining treatment end-point. A low cost, portable and non-invasive apparatus was successfully developed for measurement of ECO, oxygen concentration and minute volume. The apparatus was then evaluated in a variety of settings, for adults and children, and to establish baseline ranges for non-smokers, smokers and poisoned individuals, breathing air, NBO and HBO. The technique of measuring ECO was further investigated to determine the relationship between ECO and COHb, and for the diagnosis of CO poisoning. The apparatus was evaluated in the clinical setting to determine pulmonary CO elimination kinetics. A prospective series of CO poisoned patients was enrolled to determine if acute ECO levels correlated with clinical outcomes and to assess whether unrecordable ECO was a suitable marker of treatment end-point. In this research, expired oxygen concentration was also monitored, to ensure that all individuals received the stated dose of oxygen. Baseline levels of ECO were found to be very low in healthy non-smoking volunteers, and in nonsmoking divers treated for decompression illness, consistent with the observation that most CO derives from exogenous sources. Smokers had higher baseline ECO than non-smokers, and smoker ECO levels correlated positively with the number of cigarettes smoked per day, and negatively with the time since last cigarette. Breathing air and NBO, a strong positive linear relationship between the ECO and COHb was observed for non-poisoned smokers, poisoned individuals and pooled data. Expired CO concentration increased in proportion with increasing FIO2 for 0.21 (air) to 1.0 (NBO). While breathing 100% oxygen, increasing ambient pressure from 1 ATA to 2.8ATA did not alter the ECO concentration (ppm) in each breath. However, elimination of CO was greatly enhanced due to the increased density of gas at higher pressures. Each tidal volume at 2.8ATA actually contains 2.8 times as many molecules of CO compared with the same tidal volume at 1ATA ambient pressure. When poisoned subjects breathed NBO and HBO, significant amounts of ECO were detectable when the COHb was unrecordable using the biochemical method. This suggested that ECO more accurately reflected remaining CO in body stores than COHb, however this might have resulted from the limits of the biochemical method for detecting low levels of COHb (< 2%). Concurrent measurement of expired oxygen provided useful confirmation that the intended 100% oxygen dose was delivered to all treated individuals. ECO was a useful non-invasive test to diagnose acute (< 6 hours) CO poisoning, when ECO values were > 40 ppm. For ECO values of 7 ppm to 40 ppm, clinical information would be needed to separate mildly poisoned individuals from smokers. Expired CO and COHb were equally effective in identifying acutely poisoned individuals, from smokers and non-smokers. Critical values of ECO >40 ppm or COHb > 7% were shown to be highly specific for CO poisoning. Expired CO demonstrated single stage exponential elimination kinetics in both NBO and HBO treatment environments. CO elimination in HBO was significantly faster than NBO. There was a seven to ten-fold variation in CO elimination between individuals in either treatment (NBO or HBO). Based on these findings, current empirical regimens may over-treat some individuals and under-treat others. The half-lives determined for ECO elimination were longer than those determined for COHb. This suggests that elimination of CO via the breath may be slower than elimination from Hb. If unrecordable ECO proved useful as a treatment endpoint, this would allow treatment to be tailored to the individual's acute CO load. In the clinical series of 66 acutely poisoned patients, there were a high number of males sustaining CO poisoning from deliberate self-harm. These individuals had longer exposures, greater neurological toxicity, and were more likely to have LOC than accidental exposures. The greater toxic effect and higher CO body load was most likely due to breathing leaded petrol exhaust containing high CO levels to attempt suicide. In keeping with their greater neurological toxicity, there was a positive correlation between ECO, COHb levels, and the severity of poisoning. The ECO measurement breathing oxygen correlated significantly with the severity of neurological impairment in the ED. This provided support for ECO levels as useful guide to acute clinical poisoning severity. However, acute ECO and COHb levels measured in the ED were not predictive of outcome at 3 months. This may have been affected by significant delays in transferring patients for HBO treatment. Just over 28% of patients had poor outcomes at 3 months, using unrecordable ECO as a treatment endpoint. At this point, patients who had abnormal neurological or cognitive function remained abnormal at 3 months. Unfortunately the treatment endpoint using ECO did not prevent cases of DNS, or the need to provide follow-up for CO poisoned patients. The occurrence of DNS after all CO had been removed suggests that DNS may result from mechanisms other than direct CO toxicity. Poor outcomes were associated with delays to study entry, suicide attempts, motor vehicle exhaust as a source of CO and acidosis measured in the ED. Individuals with LOC did not have a significantly worse outcome than those remaining conscious during their CO exposure. HBO and NBO treated patients had similar levels of PNS, however the HBO group had a lower incidence of DNS ‚Äö- an unexpected finding. Because the study was not randomized, it was not possible to conclude this is a definite treatment effect. Compared with NBO, HBO treatment led to faster removal of CO, and shorter treatments. Measurement of ECO constitutes a novel non-invasive method of monitoring of acute CO poisoning. It has potential to compliment existing methods of monitoring acute CO poisoning, and may be useful as a non-invasive test to diagnose CO poisoning. Clinical outcomes in this series compared favourably with other series of similar severity poisoning in the literature. However, further research using a randomized controlled trial is required to determine if unrecordable ECO is a useful guide to treatment endpoint.