Exposure to environmental contaminants from air pollution is being increasingly identified as a significant cause for a large portion of the global death burden and associated with severe adverse health outcomes, including pulmonary, cardiovascular or immunologic and neurologic diseases. Among these air pollutants, carbon monoxide (CO) finds its place as one of the main “suspects”. Especially in enclosed spaces CO poses a great health risk – whether its low level chronic exposures leading to permanent neurologic disorders or acute exposures that lead to death. In both cases, the job of a toxicologist is to accurately detect CO by using the appropriate analytical method. The job of an epidemiologist is to associate the CO levels to a health outcome. Errors in the exposure measurement can lead to real associations not being detected. Therefore, we reviewed measurement methods used for CO measurement and identified key sources of error, which could contribute to exposure misclassification. A novel approach to more accurately determine the total amount of CO in blood (TBCO) via gas-chromatography mass spectrometry (GC-MS) was developed and validated. The results obtained from this method were then used to “feed” CO exposure assessments. Exposure assessments in public health represent the tool that from exposure estimates lead to determination of risk estimates. Using the previously identified key sources of error together with the results obtained from the improved measurement method, we developed a simulation model to quantify the impact that changes in the error due to measurement improvement could have on the relative risk (RR) of CO exposure estimates in a study using exhaled breath CO (exCO) and personal CO monitors (pCO) to measure CO. Choice of biomarker, analytical method, number of samples/measurements, time of sampling, inter-individual variabilities and use of averages for population data were some of the identified potential sources of error and can contribute from 5% up to 80% to the total measurement error. The novel CO measurement method leads to improvements in CO concentrations ranging from 20%–80% (compared to currently used methods). When applying this to the simulation model, the RR was increased up to 18.3% for exCO and up to 30% for pCO. An improved measurement method for TBCO was validated and has shown promising results in improving accuracy of CO determinations in blood as well as showing better stability to a variety of storage conditions in clinical and forensic settings. Application of this method to exposure assessment data lead to significant reduction of measurement error, resulting in CO exposure estimates closer to the true CO burden on the population. Improvement of the CO measurement method not only directly affects the results obtained in clinical and forensic cases by allowing otherwise undetected cases to be known, but it can also indirectly influence the data that is used by public health authorities to create guidelines, to generate prevention, treatment and training plans, ultimately improving global health. The model used here for CO exposure can easily be expanded and applied to other types of exposures.
Read full abstract