Assessing the calculation of conductive and acinar ventilatory heterogeneity indices Scond and Sacin from multiple-breath washout data

• Published in: Journal of applied physiology (1985) 2023-04, Vol.134 (4), p.879-886
• Main Authors: Prisk, G Kim, Rutting, Sandra, Bozier, Jack, Thamrin, Cindy, Robinson, Paul D, Thompson, Bruce R
• Format: Article
• Language: English
• Subjects: Algorithms; Error correction; Heterogeneity; Lungs; Mechanical ventilation; Sensors; Simulation
• ISSN: 8750-7587; 1522-1601
• DOI: 10.1152/japplphysiol.00423.2022

Sensor errors resulting in elevated values of N2 concentration [N2] in commercial multiple-breath washout (MBW) devices have been shown to prolong the washout and result in erroneously high functional residual capacity (FRC) and lung clearance index (LCI) values. The errors also affect the indices of conductive and acinar ventilatory heterogeneity (Scond and Sacin) although the mechanism by which this change in values occurs remains unclear. Exploring these effects also provides a timely opportunity to examine the appropriateness of the algorithm used to calculate these indices. Using a two-compartment model with differing specific ventilation (SV) such that the lower SV unit empties late, noise-free MBW were simulated both corrected and uncorrected for the recent sensor error. Scond was calculated using regression of normalized phase III slope (SnIII) against lung turnover (TO) from a TO range starting at 1.5 and ending at an upper turnover (TOupper) between 4 and 8 (default 6) over a range of simulated values. The principal effect of the sensor error was that as the MBW proceeded the phase III slope of successive breaths was normalized by an increasingly overestimated [N2], resulting in SnIII values that fell precipitously at high TO, greatly reducing Scond. Reanalysis of previously published data and of simulated data showed a large proportional bias in Scond, whereas Sacin was only minimally affected. In adult subject data, reducing TOupper below 5.5 was associated with a large drop of up to ∼60% in Scond calculated from data corrected for sensor error. Raising the upper TO limit elevated Scond by ∼20% but with a large concomitant increase in variability. In contrast to Scond, Sacin was relatively unaffected by changes in TOupper with changes of