Impact of different fixed flow sampling protocols on flow‐independent exhaled nitric oxide parameter estimates using the Bayesian dynamic two‐compartment model

Abstract

Exhaled nitric oxide (FeNO) is an established respiratory biomarker with clinical applications in the diagnosis and management of asthma. Because FeNO depends strongly on the flow (exhalation) rate, early protocols specified that measurements should be taken when subjects exhaled at a fixed rate of 50 ml/s. Subsequently, multiple flow (or "extended") protocols were introduced which measure FeNO across a range of fixed flow rates, allowing estimation of parameters including C aw NO and C A NO which partition the physiological sources of NO into proximal airway wall tissue and distal alveolar regions (respectively). A recently developed dynamic model of FeNO uses flow-concentration data from the entire exhalation maneuver rather than plateau means, permitting estimation of C aw NO and C A NO from a wide variety of protocols. In this paper, we use a simulation study to compare C aw NO and C A NO estimation from a variety of fixed flow protocols, including: single maneuvers (30, 50,100, or 300 ml/s) and three established multiple maneuver protocols. We quantify the improved precision with multiple maneuvers and the importance of low flow maneuvers in estimating C aw NO. We conclude by applying the dynamic model to FeNO data from 100 participants of the Southern California Children's Health Study, establishing the feasibility of using the dynamic method to reanalyze archived online FeNO data and extract new information on C aw NO and C A NO in situations where these estimates would have been impossible to obtain using traditional steady-state two compartment model estimation methods. K E Y W O R D SBayesian inference, exhaled breath, FeNO, sampling protocol