Shedding light into the black box of infant multiple‐breath washout

Abstract

Background: Multiple-breath inert gas washout (MBW) is a sensitive technique to assess lung volumes and ventilation inhomogeneity in infancy. Poor agreement amongst commercially available setups and a lack of transparency in the underlying algorithms for the computation of infant MBW outcomes currently limit the widespread application of MBW as a surveillance tool in early lung disease. Methods:We determined all computational steps in signal processing and the calculation of MBW outcomes in the current infant WBreath/Exhalyzer D setup (Exhalyzer D device, Eco Medics AG; WBreath software version 3.28.0, ndd Medizintechnik AG; Switzerland). We developed a revised WBreath version based on current consensus guidelines and compared outcomes between the current (3.28.0) and revised (3.52.3) WBreath version. We analyzed 60 visits from 40 infants with cystic fibrosis (CF) and 20 healthy controls at 6 weeks and 1 year of age.Results: Investigation into the algorithms in WBreath 3.28.0 revealed discrepancies from current consensus guidelines, which resulted in a potential overestimation of functional residual capacity (FRC) and underestimation of lung clearance index (LCI).We developed a revised WBreath version (3.52.3), which overall resulted in 6.7% lower FRC (mean (SD) −1.78 (0.99) mL/kg) and 14.1% higher LCI (1.11 (0.57) TO) than WBreath version 3.28.0. Conclusion:Comprehensive investigation into the signal processing and algorithms used for analysis of MBW measurements improves the transparency and robustness of infant MBW data. The revised software version calculates outcomes according to consensus guidelines. Future work is needed to validate and compare outcomes between infant MBW setups.

Comparison of functional residual capacity between updated infant SF6 multiple‐breath washout setups

“…Multiple‐breath inert gas washout (MBW) is a sensitive technique to assess lung volumes and ventilation inhomogeneity from infancy, but technical and methodological issues limit its widespread application 1 . There are two setups currently being used for the collection of infant MBW measurements, the WBreath (ndd Medizintechnik AG) and the more recent Spiroware (Eco Medics AG) setup.…”

“…We recently identified, characterized, and corrected a significant sensor‐crosstalk error in the Exhalyzer D device that led to an overestimation of tracer gas concentration 4 . Likewise, we identified and corrected discrepancies from current ATS/ERS consensus guidelines in WBreath analysis software that led to an overestimation of FRC 1 . As a result, the respective manufacturers released new software versions which incorporated updated analysis algorithms.…”

“…In WBreath, outcomes are based on changes in the main‐stream molar mass of expired air caused by the wash‐in and wash‐out of the tracer gas sulfur hexafluoride (SF 6 ). However, previous studies have shown that the outcomes are heavily dependent on the software version, system settings, and analysis protocol 1 . The Spiroware setup utilizes additional information from simultaneously measured and synchronized main‐stream CO 2 and side‐stream molar mass and O 2 signals to calculate SF 6 concentrations 2 .…”

ERS International Congress 2021: highlights from the Paediatric Assembly

“…O estreich et al . [ 32 ] also examined the evolution of LCI in healthy and CF cohorts by measuring SF 6 MBW at 6 weeks and 1 year of age. Despite similar lung function at 6 weeks of age, LCI significantly deteriorated over the first year of life in infants with CF [ 33 ].…”

A computerized tool for the systematic visual quality assessment of infant multiple-breath washout measurements

“…Currently, there are two software packages available for the analysis of infant SF 6 -MBW measurements: The historic WBreath (ndd Medizintechnik AG, Zurich, Switzerland) and the more recent Spiroware software (Eco Medics AG, Duernten, Switzerland), both applying indirect measurement principles for the inert tracer gas. Due to fundamental differences in the signals used and algorithms applied, the data of the two systems cannot be used interchangeably ( 14 , 15 ). Moreover, both software packages have recently received relevant updates affecting the main outcomes, so that previously collected data needs to be reanalyzed ( 14 , 16 ).…”