Application of end-expired breath sampling to estimate carboxyhemoglobin levels in community air pollution exposure assessments

Measurement of carbon monoxide (CO) in end-expired air after breath-holding permits the estimation of blood carboxyhemoglobin (COHb) levels. Some literature suggests that the precision of the method decreases at low COHb levels. As part of a community exposure and health study, the end-expired breath method was applied to estimate COHb levels in 28 men with ischemic heart disease. Paired samples of blood and breath were collected at the beginning and end of the 24-h CO monitoring periods. The aggregate regression of all subjects' COHb on breath CO displayed high variability. However, the variability was substantially reduced for any particular subject, promoting the use of individualized blood-breath standard curves to improve the precision of COHb estimates made from breath CO. The ultimate accuracy of the blood-breath relationship could not be resolved by our data. Two major sources of error are identified. The observed person-to-person variability may be caused by physiologic factors or differences in ability to deliver an end-expired breath sample representative of alveolar air. This variation may also be due to instrumentation factors, specifically the accuracy of the IL282 CO-Oximeter at 0-3% levels. Further research into the sources of variability in the end-expired breath method is recommended. Epidemiologists using similar end-expired breath measurements to predict COHb levels should be cognizant of the magnitude and probable direction of the error in COHb estimates. This non-invasive method should continue to allow evaluation of the success of personal monitoring efforts and pharmacokinetic modeling of CO uptake in community exposure research.