Detection of hydrogen peroxide and ethanol in exhaled breath using biosensors:Considerations in standardizing breath collection

Breath analysis;exhaled breath;sampling conditions;breath collection standardization;biosensor;uncertainty analysis

Chen, Shih-Fang

关键词: Breath analysis;    exhaled breath;    sampling conditions;    breath collection standardization;    biosensor;    uncertainty analysis;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/31994/Chen_Shih-Fang.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Breath monitoring is a non-invasive, safe, and repeatable approach to determining the health status of humans and other mammals. Breath samples could be detected in two ways  directly sensing exhaled breath (EB) or chilling the EB to obtaining the exhaled breath condensate (EBC). Each has its advantages and disadvantages but they are both affected by different sampling conditions. Additionally, volatile organic compounds (VOCs) and nonvolatile organic compounds (non-VOCs) in the breath matrix are retained differently under varied sampling conditions. The dearth of information on how sampling conditions affect the intrinsic properties of biomarkers in breath and the lack of standardization information hinder the use of breath monitoring in clinical use.The study aims to develop predictive models to standardize the varied sampling conditions of breath temperatures, flow rates, condensing temperatures, and sensing durations in EB and EBC sensing. Ethanol (VOC) and H2O2 (non-VOC) were chosen as model biomarkers, which were potential biomarkers of liver function and respiratory diseases, respectively. A breath output simulator was developed to simulate the conditions of exhaled breath. Screen printed carbon electrodes (SPCEs) were used solely or immobilized with alcohol oxidase as biosensors for detecting the chosen biomarkers amperometrically. Akaike's information criterion, Bayesian information criterion, and cross validation were adopted in predictive model selections, and uncertainty analyses were surveyed to further clarify the margin of doubt for the measurement of each sampling factors. Final predictive models were developed for ethanol in EB and EBC, and H2O2 in EBC for specific sensing time (5 min) and full sensing duration (3-10 min). Results showed that the EBC model for ethanol in 5 min measurement performed a better regression result (R2 = 0.9471) than the EB model for ethanol (R2 = 0.8261) and the EBC model for H2O2 (R2 = 0.8261). Furthermore, in 5 min predictive models, both ethanol and H2O2 concentrations in EBC samples were affected by condensing temperature, but only H2O2 detection was affected by breath temperature and breath rate. The results indicated that sampling conditions were more critical and were more constrained for non-VOC sensing than VOC sensing. Uncertainty analyses showed that the 5 min EBC predictive models had 18.53 – 26.55% percentage uncertainty and the 5 min EB predictive model had up to 44.21% percentage uncertainty. The major source of the uncertainties was due to the sensing system which included the SPCEs and used enzyme.