Correction of measurement error in a commercial multiple-breath washout device

Rationale Nitrogen multiple-breath washout (N 2 MBW) is an established technique to assess functional residual capacity (FRC) and ventilation inhomogeneity in the lung. Accurate measurement of gas concentrations is essential for the appropriate calculation of clinical outcomes. Objectives We investigated the accuracy of oxygen and carbon dioxide measurements used for the indirect calculation of nitrogen concentration in a commercial MBW device (Exhalyzer D, Eco Medics AG, Duernten, Switzerland) and its impact on FRC and lung clearance index (LCI). Methods High precision calibration gas mixtures and mass spectrometry were used to evaluate sensor output. We assessed the impact of corrected signal processing on FRC and LCI in a dataset of healthy children and children with cystic fibrosis using custom analysis software. Results We found inadequate correction for the cross sensitivity of the oxygen and carbon dioxide sensors in the Exhalyzer D device. This results in an overestimation of expired nitrogen concentration, and consequently FRC and LCI outcomes. Breath-by-breath correction of this error reduced mean (SD) FRC by 8.9 (2.2)% and LCI by 11.9 (4.0)%. It also resulted in almost complete disappearance of the tissue nitrogen signal at the end of measurements. Conclusions Inadequate correction for cross sensitivity between the oxygen and carbon dioxide gas sensors of the Exhalyzer D device leads to an overestimation of FRC and LCI. Correction of this error is possible and could be applied by re-analysing the measurements breath-by-breath in an updated software version. Grants, Gifts, Equipment, Drugs Eco Medics AG (Duernten, Switzerland) provided a research version of their commercial software Spiroware 3.2.1 including insight on signal processing algorithms and helped with the acquisition of mass spectrometry measurements. This project was funded by the Swiss National Science Foundation, Grant Nr. 182719 (P. Latzin) and 168173 (K. Ramsey)