Reply: Fixed breathing protocols in multiple-breath-washout testing: truly an option in children?

Reply to A-C. Kentgens and co-workers:

We agree with the challenges discussed by A-C. Kentgens and co-workers for the adoption of fixed breathing protocols outside the adult range assessed in our study, but we want to clarify that we were not recommending imposing 1 L breathing outside the age range studied [1]. The main intention of our research letter was to illustrate the impact of different breathing protocols on variability of multiple breath washout (MBW) outcomes by choosing a cohort where both protocols were readily feasible. By expressing the tidal volume (V_T) dependence that we observed in adults in terms of mL·kg⁻¹, we then aimed to make observations about whether the pronounced variability shown with tidal breathing may affect existing paediatric data. Consistent with a concurrent publication by Handley et al. [2], we found that controlled versus tidal breathing had a negligible impact on functional residual capacity and lung clearance index (LCI), but large effects on acinar and conductive airway ventilation heterogeneity (S_acin and S_cond) in healthy adults. If the latter two indices are used in interventional studies, we also contended that interpretation of treatment effects is potentially complicated by confounding V_T effects, if breathing before or after treatment changes in any way.

A-C. Kentgens and co-workers bring up an interesting point: whether a test with a fixed, or a larger than natural, V_T can be of physiological relevance. This really depends on the physiological information one is after; a case in point is diffusing capacity, which is sampled at a volume largely exceeding natural lung inflation. Particularly with respect to ventilation distribution, absolute lung inflation is less relevant than the relative expansion of lung regions [3], and it is highly unlikely that ventilation distribution per se would be affected by adopting a fixed or a larger than natural tidal volume. While the rate of overall lung volume expansion changes as a subject approaches total lung capacity, the relative expansions are likely similar, as exemplified by linear portions of the onion-skin model for interregional differences [4]. This would still hold in lung disease with different expansion rates in different parts of the lungs, unless their relative expansion rates are modified near total lung capacity.

While insights about the impact of breathing patterns on MBW variability are important to consider when designing future testing protocols, adult studies can also inform ways that re-analysis of existing MBW data can improve data quality. For instance, LCI variability which was not shown to be sensitive to breathing pattern in recent adult data [1, 2], can be reduced by computing LCI based on mean expired concentration (LCI_meanexp) instead of end-expiratory concentration (LCI_endexp) as was done in initial MBW studies (e.g. [5]) and many adult MBW studies since. Because LCI is intrinsically based on a single value of gas concentration that is to represent the efficiency of gas mixing of the lungs at each subsequent breath, any choice of concentration will be biased to some extent (e.g. end-tidal concentration is biased towards the low ventilation compartment). In addition, the actual value of end-tidal concentration measured at the mouth is governed by complex dynamic intra-breath aspects of ventilation heterogeneity, such as convective flow asynchrony and diffusion-convection dependent inhomogeneities [4]. We have previously discussed the respective behaviour of LCI_meanexp and LCI_endexp in asthma [6]. Here, we share a re-analysis of N₂ MBW data from an adult cystic fibrosis study [7] to show how LCI variability increases when LCI_endexp is considered instead of LCI_meanexp (figure 1). In those 22 patients (mean±sd forced expiratory volume in 1 s 81±16% predicted) for whom MBW tests were long enough to allow concurrent computation of LCI_endexp (average LCI_endexp 10.5) in addition to LCI_meanexp (average LCI_meanexp 7.8), there was a dramatic increase in inter-subject coefficient of variation (CoV) just by considering end-expiratory (CoV(LCI_endexp) 32%) instead of mean expired concentration (CoV(LCI_meanexp) 19%) on the same curves. In the 25 normal subjects of that same study, the increased variability effect on LCI was still present, albeit to a lesser extent (LCI_endexp 6.9 with CoV 8.6% versus LCI_meanexp 6.2 with CoV 6.7%).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Re-analysed multiple breath N₂ concentration curves obtained from 45 cystic fibrosis patients [7], determining lung clearance index (LCI) using either a) mean expired or b) end-expired N₂ concentration. Grey areas delimit the 1/40 pre-test N₂ thresholds. Open circles: cystic fibrosis patients for whom both mean expired and end-tidal N₂ concentration reached the 1/40th pre-test N₂ concentration threshold (n=22). Their mean±sd forced expiratory volume in 1 s (FEV₁) was 81±16% predicted and LCI_meanexp was 7.8±1.5; corresponding LCI_endexp was 10.5±3.3. Crosses: cystic fibrosis patients for whom mean expired but not end-tidal N₂ concentration reached the 1/40 pre-test N₂ concentration threshold (n=23). Their mean±sd FEV₁ was 56±17% predicted and LCI_meanexp was 11.0±2.0.

In essence, our intention here [1] and in previous work [6] is to share our experience with MBW testing in adult patients, to illustrate sources of increased experimental variability, for instance the breathing protocol chosen or the method of LCI computation used. Considering that LCI_meanexp instead of LCI_endexp also obviously reduces test duration, it offers an additional approach to consider when exploring methods to shorten MBW test whilst preserving the distal part of the washout curve [8]. Depending on disease severity, sensitivity of LCI to detect disease change would need to be explored, but this can be readily done by re-analysis of existing data.

Shareable PDF