Novel volumetric capnography indices measure ventilation inhomogeneity in cystic fibrosis

Sotirios Fouzas, Anne-Christianne Kentgens, Olga Lagiou, Bettina Sarah Frauchiger, Florian Wyler, Ilias Theodorakopoulos, Sophie Yammine, Philipp Latzin
ERJ Open Research 2022 8: 00440-2021; DOI: 10.1183/23120541.00440-2021

Abstract

Background Volumetric capnography (VCap) is a simpler alternative to multiple-breath washout (MBW) to detect ventilation inhomogeneity in patients with cystic fibrosis (CF). However, its diagnostic performance is influenced by breathing dynamics. We introduce two novel VCap indices, the capnographic inhomogeneity indices (CIIs), that may overcome this limitation and explore their diagnostic characteristics in a cohort of CF patients.

Methods We analysed 320 N2-MBW trials from 50 CF patients and 65 controls (age 4–18 years) and calculated classical VCap indices, such as slope III (SIII) and the capnographic index (KPIv). We introduced novel CIIs based on a theoretical lung model and assessed their diagnostic performance compared to classical VCap indices and the lung clearance index (LCI).

Results Both CIIs were significantly higher in CF patients compared with controls (mean±sd CII1 5.9±1.4% versus 5.1±1.0%, p=0.002; CII2 7.7±1.8% versus 6.8±1.4%, p=0.002) and presented strong correlation with LCI (CII1 r2=0.47 and CII2 r2=0.44 in CF patients). Classical VCap indices showed inferior discriminative ability (SIII 2.3±1.0%/L versus 1.9±0.7%/L, p=0.013; KPIv 3.9±1.3% versus 3.5±1.2%, p=0.071), while the correlation with LCI was weak (SIII r2=0.03; KPIv r2=0.08 in CF patients). CIIs showed lower intra-subject inter-trial variability, calculated as coefficient of variation for three and relative difference for two trials, than classical VCap indices, but higher than LCI (CII1 11.1±8.2% and CII2 11.0±8.0% versus SIII 16.3±13.5%; KPIv 15.9±12.8%; LCI 5.9%±4.2%).

Conclusion CIIs detect ventilation inhomogeneity better than classical VCap indices and correlate well with LCI. However, further studies on their diagnostic performance and clinical utility are required.

Abstract

Novel volumetric capnography indices are promising markers of ventilation inhomogeneity in patients with cystic fibrosis https://bit.ly/32ngbS6

Introduction

Volumetric capnography (VCap) is a simple and noninvasive lung function technique that describes the dynamics of carbon dioxide (CO2) exhalation breath-by-breath [1]. The volume-based capnogram is the plot of CO2 concentration against the exhaled air volume and consists of three phases: phase I represents the washout of the uppermost conductive airways that contain atmospheric – CO2-free – air; phase II is characterised by a steep CO2 rise and reflects the mixing between atmospheric air and CO2-rich gas from the alveolar compartment; and phase III, the so-called “alveolar plateau”, represents the expiration of alveolar gas. Volumetric capnography allows for the assessment of dead space ventilation and ventilation–perfusion abnormalities [2–7], while the slope of phase III (SIII) is considered an index of ventilation inhomogeneity in obstructive lung disorders, such as cystic fibrosis (CF) [7–11].

Since VCap does not require a complex measurement setup or extensive signal processing, it may be considered a simpler alternative to techniques aiming to detect ventilation inhomogeneity, such as multiple-breath washout (MBW) [2, 7, 10]. However, classical VCap indices are considerably dependent on expiratory volume (VE) and respiration dynamics [12–14]. This is particularly important in children, where variable breathing patterns may significantly affect the diagnostic performance of the method [6, 7, 9, 15].

In this paper, we introduce novel VCap indices called capnographic inhomogeneity indices (CIIs) that may overcome the above limitations. Moreover, they may serve as promising ventilation inhomogeneity indices in clinical settings where advanced MBW setups are not available, or as faster and more attainable markers in obstructive lung disorders that require continuous follow-up. We assess the feasibility and repeatability of these novel CIIs in clinical practice, and we explore their diagnostic performance in a cohort of CF patients.

Methods

Study design and population

In this pilot study, novel VCap indices were validated using existing nitrogen (N2) MBW data from 4- to 18-year-old CF patients and healthy controls. All measurements were obtained at the University Children's Hospital of Bern, Switzerland, between 2013 and 2019 (see Sample size estimation) [16–21]. All subjects were free from pulmonary exacerbations [16–21]. Only subjects with at least two good-quality MBW trials were included. Local ethics committees approved all studies, and participants or caregivers provided informed written consent.

Lung function

MBW trials were performed with the children in sitting position while breathing normally, according to international guidelines [22]. N2-MBW measurements were obtained as previously described [20], using the Exhalyzer D device (Ecomedics, Duernten, Switzerland) that incorporates an ultrasonic flowmeter and a main-stream CO2 sensor (Capnostat, infrared single beam, dual-wavelength technology; rise time <60 ms) with Spiroware version 3.1.6 or 3.2.1. All data were reloaded and reanalysed using the updated software provided by the manufacturer of Spiroware 3.3.1 (Ecomedics, Duernten, Switzerland). MBW quality control was performed by experienced operators, based on established criteria [21, 22].

The following MBW parameters were investigated: functional residual capacity (FRC), lung clearance index (LCI), tidal volume, respiratory rate and minute ventilation [22]. Additional information on age, sex, weight and height were obtained from patient files.

Volumetric capnography

Volumetric capnograms were obtained from N2-MBW trials after CO2 signals were corrected for setup-dependent signal alignment and sensor-specific delays (Spiroware software). Breath-by-breath volumetric capnograms were obtained by plotting the CO2 fraction against the corresponding VE. The expired CO2 volume (VECO2) was calculated by integrating the CO2 signal over VE. The slope of phase II (SII) and the SIII were obtained by fitting a linear regression line (least squares method) over the capnogram, between 10 and 60% of the end-tidal CO2 value and 65–95% of VE, respectively [6]. The capnographic index (KPIv) was calculated as the ratio SIII/SII. The slopes were normalised by the corresponding mixed expired CO2 fraction (FECO2), which equals the VECO2 divided by VE [12]. The airway dead space (VD) was calculated using the equal-area method proposed by Fowler [23].

To avoid inclusion of irregular capnograms, MBW breaths were automatically excluded from analysis if: 1) end-tidal CO2 (ETCO2) was <3.5%; 2) their corresponding VE deviated more than two SDs from trial mean; 3) VD was <10% or >30% of VE; 4) intersection of SII and SIII was not calculable, fell outside the capnogram range (i.e.intersection at VE<0, VE>VE at the end of expiration, CO2>CO2max) or below the CO2 curve; 5) SIII was steeper than SII; and 6) coefficient of determination (r2) for SIII fitting was <0.7. For all acceptable breaths of each MBW trial, we calculated mean ETCO2, SII, SIII, normalised SII (nSII), normalised SIII (nSIII), KPIv and VD values. Calculation of VCap indices and quality control were performed using a custom Python script (Python Software Foundation, Wilmington, DE, USA; www.python.org/).

Capnographic inhomogeneity indices

Modelling ventilation inhomogeneity

To understand the calculation of novel CIIs, we consider a lung model (figure 1) in which we define VA, the volume of the alveolar compartment that contributes to expiration, VD the volume of the dead space compartment and VE the expired volume of air. Let us assume that the alveolar compartment comprises three equal sub-compartments connected in series, which empty sequentially without air mixing among them (figure 1). At end-inspiration, the dead space compartment is filled with atmospheric air (with a CO2 concentration of zero). All CO2 is contained in VA, at a mixed alveolar concentration of FACO2, which is the mean CO2 concentration of alveolar sub-compartments (i.e. FA1CO2, FA2CO2 and FA3CO2) (figure 1). As expiration commences, dead space air emerges first, followed by gas from the three alveolar sub-compartments. At end-expiration, the concentration of CO2 in the expired air (FECO2) results from the mixing between the CO2-free dead space air and the CO2-rich alveolar air; an amount of CO2 remains in the dead space compartment at a concentration FDCO2 (figure 1).

FIGURE 1
FIGURE 1

The three-compartment model of our study and the respective volumetric capnograms. a) Without inhomogeneity (ideal lung); b) with sequential inhomogeneity. See text for complete description. VE: expiratory volume; VA: alveolar compartment volume; VD: dead space compartment volume; VECO2: expired CO2 volume; FECO2: mixed expired CO2 fraction; SIII: slope of phase III; FACO2: CO2 fraction in alveolar compartment; FDCO2: CO2 fraction in dead space compartment; FA1–3: alveolar sub-compartments.

In an ideal lung, FA1CO2, FA2CO2 and FA3CO2 are equal (figure 1a). Thus, time-dependent inhomogeneities do not exist, the rate of CO2 exhalation is constant and the phase III of the capnogram is horizontal (figure 1a). In this case, FDCO2 at end-expiration equals FACO2 (figure 1a). When time-dependent inhomogeneities exist, due to stratified ventilation inhomogeneity distal to the airway–alveolar interface and delayed emptying of respiratory units with low ventilation–perfusion (V′/Q′) ratios [24, 25], it applies that FA1CO2<FA2CO2<FA3CO2 (figure 1b). In this case, the rate of exhaled CO2 increases as expiration commences, and the SIII rises (figure 1b). At end-expiration, FDCO2 equals FA3CO22 (figure 1b). The magnitude of sequential inhomogeneity in this model is reflected by the difference among FA1CO2, FA2CO2 and FDCO2 (figure 1b). However, since neither FACO2 nor FDCO2 can be measured in vivo, these differences cannot be computed.

Estimating FDCO2

Let us consider the volumetric capnogram of figure 2a and analyse it according to the concept of Aitken and Clark-Kennedy [26], as follows: a) The volume of CO2 that remains in the dead space compartment at end-expiration (VDCO2) can be obtained by extending the line of phase III to the right (line be) using linear regression until the distance cf becomes equal to VD, and calculating the area of the trapezoid befc; the FDCO2 is the area befc divided by VD. b) The total volume of CO2 leaving the alveolar compartment (VACO2) is the area of the trapezoid aefd; FACO2 can be obtained by dividing this area by VE (figure 2a).

FIGURE 2
FIGURE 2

VCap analysis with extension by VD. a) As originally proposed by Aitken and Clark-Kennedy [26]; b) as applied in our study. See text for detailed explanations. VE: expiratory volume; VD: dead space compartment volume; SIII: slope of phase III; VECO2: expired CO2 volume; VDCO2: CO2 volume in dead space compartment.

The above concept was applied to volumetric capnograms of our study (figure 2b). After the VD was calculated by Fowler's method [23], each capnogram was extended to the “right” (linear extension according to SIII) by VD. The VECO2 was computed as the integral of the CO2 signal over VE, and the VDCO2 as the integral of the extended part of the capnogram over VD (figure 1b); the VACO2 was computed as the sum of VECO2 and VDCO2. Then, by dividing VECO2 by VE, VDCO2 by VD, and VACO2 by VE, the FECO2, FDCO2 and FACO2, respectively, were calculated. In addition, the concentration of CO2 in the exhaled air coming exclusively from the alveolar compartment (FexCO2) was calculated by dividing VECO2 by the difference VE − VD.

Capnographic inhomogeneity indices

Two novel CIIs were calculated:

A. CII1, which is the relative difference between FDCO2 and FACO2 or (FDCO2− FACO2)/FACO2 and represents a raw estimate of sequential inhomogeneity.

B. CII2, which is the relative difference between FDCO2 and FexCO2 or (FDCO2− FexCO2)/FexCO2. Since CII2 reflects the difference of CO2 concentration between the alveolar sub-compartments A1+A2 and A3 (figure 1b), it should theoretically represent a more precise estimate of sequential ventilation inhomogeneity.

Statistics

Sample size estimation

Since CIIs were introduced for the first time in clinical practice, no data were available for a priori sample size estimation. Based on the hypothesis that CIIs (as markers of VI) would be correlated with the LCI, we estimated that MBW measurements from at least 112 subjects would be required to reveal a significant correlation (p<0.05), with a Pearson's r ≥0.3 and at least 90% power. By reviewing the database of our laboratory (different independent studies [16–21]), we found that such a sample with at least two acceptable MBW trials could be obtained by including the children measured between January 2013 and December 2019 (n=115; 50 CF and 65 healthy controls). We preferred to include all 115 children (instead of 112) for reasons of consistency. Post hoc effect size calculation was also performed.

Statistical analyses

VCap indices were assessed both per subject (mean of at least two acceptable trials) and per MBW trial (analysis per trial) to allow better physiological appraisal. Per subject data are presented throughout the manuscript, while per trial analyses are presented in the online supplementary material. Between-group comparisons were performed using t-test. The percentage of breaths acceptable for VCap analysis was calculated as percentage of all washout breaths. Linear and nonlinear regression was applied to assess the relationship between SIII and VE, and SIII and CIIs. Pearson correlation analysis was used to assess the relationship between LCI and VCap outcomes. The intra-trial (i.e. between breaths of the same trial) and inter-trial (i.e. between different trials of the same subject) variability was calculated using the coefficient of variation (CV). In the presence of only two acceptable MBW trials, the inter-trial variability was calculated as relative difference between the two trials. Receiver operation characteristics analysis was used to estimate the overall diagnostic ability of LCI and VCap indices by means of area under the curve (AUC). Optimal cut-off values for each index were determined using Youden Index analysis. All analyses were performed in Stata (StataCorp, College Station, TX, USA).

Results

All 320 MBW trials from 50 patients with CF (137 trials) and 65 HCs (183 trials) were analysed. The clinical characteristics and lung function parameters of the two study groups are presented in table 1. The LCI was significantly higher in CF patients compared to healthy children (mean±sd LCI 7.8±1.7 versus 6.2±0.4, p<0.001).

View this table:
TABLE 1

Characteristics and lung function of the study groups

VCap indices

Calculation of VCap indices was feasible in all trials of all study participants, independent of disease status. The percentage of breaths within each trial that were acceptable for VCap analysis was higher in CF patients compared with healthy children (72.4±17.8 (42.7–98.3)% versus 63.5±19.3 (19.7–96.7)%). Further information regarding non-accepted breaths (percentages, stratification per exclusion criteria) is presented in supplementary table S2. Slope III and KPIv were higher in CF patients compared with healthy children (SIII 2.3±1.0%/L versus 1.9±0.7%/L, p=0.013; KPIv 3.9±1.3% versus 3.5±1.2%, p=0.07), but only SIII was significantly higher. The CIIs were also significantly higher in CF patients compared with healthy controls (CII1 5.9±1.4% versus 5.1±1.0%, p=0.002; CII2 7.7±1.8% versus 6.8±1.4%, p=0.002). Post hoc effect size analysis revealed that these differences yield a study power of 71% and 68%, respectively, for the given level of statistical significance, or 96% and 89.5%, respectively, for a p-value <0.05. VCap parameters and indices are presented in tables 2 and 3, and figure 3. AUC values and the corresponding optimal cut-off values of LCI and VCap indices are presented in supplementary table S5). Above the cut-off value of LCI were classified 76% (n=38) of patients with CF and 10.8% (n=7) of controls, above the cut-off value of SIII 34% (n=17) of patients with CF and 10.8% (n=7) of controls, above the cut-off value of KPIv 76% (n=38) of patients with CF and 47.7% (n=31) of controls, above the cut-off value of CII1 44% (n=22) of patients with CF and 20% (n=13) of controls, and above the cut-off value of CII2 48% (n=24) of patients with CF and 18.5% (n=12) of controls.

View this table:
TABLE 2

VCap parameters and indices

View this table:
TABLE 3

Novel volumetric capnography (VCap) parameters and capnographic inhomogeneity indices

FIGURE 3
FIGURE 3

Boxplots showing the difference between patients with cystic fibrosis and healthy controls in mean values of a) CII1 and b) CII2, c) LCI, d) SIII and e) KPIv. The individual black dots represent mean values per subject. CII: capnographic inhomogeneity index; LCI: lung clearance index; SIII: slope of phase III; KPIv: capnographic index.

Correlations of VCap indices

There was a strong inverse curvilinear relationship between SIII and VE and a weak linear relationship between CII1 and SIII and CII2 and SIII (figure 4, supplementary figure S3). The correlations between classical VCap indices and CIIs are shown in supplementary table S6. The correlation SIII–LCI and KPIv–LCI was weak (SIII–LCI r2=0.03; KPIv–LCI r2=0.08 in CF patients), while the correlation between CIIs and LCI was stronger (CII1–LCI r2=0.47 and CII2–LCI r2=0.44 in CF patients). Overall (all participants), the Pearson's correlation coefficient between CII1 and LCI was 0.572 and between CII2 and LCI 0.557; both values yield a study power of 100% (p<0.001). More correlations between VCap parameters and LCI are presented in table 4 and supplementary table S7. Both CII1 and CII2 were significantly correlated with age (r2=0.162 and r2=0.118, respectively), while age was also significantly correlated with the LCI (r2=0.09). There was no relationship between CIIs and sex in our cohort (data not shown).

FIGURE 4
FIGURE 4

Relationship between a) SIII and VE, b) CII1 and SIII and c) CII2 and SIII in CF (green dots) and healthy control trials (black open dots). CF: cystic fibrosis; SIII: slope of phase III; VE: expiratory volume; CII: capnographic inhomogeneity index.

View this table:
TABLE 4

Correlations between VCap parameters and the lung clearance index

Intra- and inter-trial variability

The intra-trial variability CVs of SIII and KPIv were higher compared with CII1 and CII2 (CV of SIII 37.5±19.2%, KPIv 35.3±15.5% versus CII1 31.1±8.9% and CII2 31.7±9.4%, in all trials). For all these VCap indices, the intra-trial variability was lower in CF patients compared with controls (supplementary table S8). Similarly, the inter-trial (intra-subject) variability of SIII and KPIv was higher compared with CII1 and CII2 in all trials (SIII 16.3±13.5%, KPIv 15.9±12.8% vs CII1 11.1±8.2% and CII2 11.0±8.0%). However, the inter-trial variability was comparable between CF patients and controls (supplementary table S9). The LCI showed lowest inter-trial variability (LCI 5.9±4.2%) (supplementary table S9, figure S2).

Discussion

In this pilot study, we introduced novel capnographic indices of ventilation inhomogeneity and we assessed their diagnostic performance in comparison with classical VCap parameters (i.e. SIII and KPIv) and the LCI. We found that SIII and the novel CIIs were higher in CF patients than in controls. However, the novel capnographic inhomogeneity indices showed better correlation with the LCI and lower intra-trial and inter-trial variability than SIII and KPIv, although their overall diagnostic performance was inferior to the LCI.

Performance of VCap indices

As expected, SIII and KPIv were increased in CF patients. These findings are in line with previous studies, suggesting that classical VCap indices may be useful ventilation inhomogeneity markers in adults and children with CF [8–11, 27]. In CF, ventilation inhomogeneity results in delayed CO2 mixing within the conductive airways and, eventually, to non-homogeneous CO2 exhalation (i.e. steeper phase III) [1, 3, 8] (figure 1); the KPIv (i.e. the SIII to SII ratio) increases respectively [8]. Thus, increased SIII and KPIv are consistent findings in CF [9–11, 27], albeit their discriminative ability is moderate [11] and, in any case, inferior to that of the LCI [10]. Of note, the correlation between SIII and LCI or KPIv and LCI was rather weak. The latter contrasts the findings of Fuchs et al. [10] who showed a stronger correlation between those indices and LCI. The calculation of SII and SIII on an “averaged”, user-defined capnogram in their study [10], as opposed to breath-by-breath calculation of VCap indices using well-defined criteria in ours, might explain these differences.

Conversely, CII1 and CII2 presented better discriminative characteristics than the classical VCap indices. In addition, both CIIs correlated significantly with the LCI – a robust index of ventilation inhomogeneity in CF patients [28, 29], and did so better than the classical VCap parameters (table 4). Thus, CIIs may be considered valid measures of ventilation inhomogeneity, which may also relate to the severity of the disease.

Theoretical advantages of CIIs over classical VCap indices

The SIII reflects the rate of CO2 exhalation beyond mid-expiration and is an unstandardised index that depends on the dynamics of expiration, especially on VE [12–14]. However, normalisation of SIII by VE [11] is not justified, because as shown in previous studies [12–14] and confirmed in ours (figure 4a, supplementary figure S3A), the SIII–VE relationship is not linear. SIII normalisation by the FECO2 (i.e. the normalised SIII) may allow for intra- or inter-subject adjustment for different CO2 concentrations but does not eliminate the dependency from VE. Therefore, in subjects with variable respiratory patterns and/or changing lung volumes, as it typically is in children, the utility of SIII is limited [14]. In our study, this disadvantage is also reflected by the unexpected negative correlation between SIII and LCI (table 4). Higher LCI values are typically seen in older CF patients due to the progression of the disease [21]; but since older children also have larger lung volumes, their SIII is lower due to the strong SIII–VE relationship (figure 4a, supplementary figure S3A).

The theoretical background for calculation of CIIs is different: in lung disease, variable gas mixing within the respiratory units (serial inhomogeneities) and/or regional V′/Q′ variations (parallel inhomogeneities) result in sequential variations of CO2 concentration distal to the airway–alveolar interface, which are exacerbated further by the delayed emptying of respiratory units with altered mechanical properties and low V′/Q′ [23, 24]. Overall, these phenomena result in time-dependent inhomogeneities of CO2 concentration that can be detected at the airway opening [1, 12]. The CIIs, which according to the proposed model (figure 1) are calculated as differences of CO2 concentrations, reflect these time-dependent inhomogeneities. The CO2 concentrations (i.e. VCO2 − VE fractions) also include an inherent normalisation for VE, and their use may thus overcome the limitations of classical VCap indices. The significant positive correlation between CIIs and LCI and the weak correlation between CIIs and SIII further support this hypothesis.

Feasibility and repeatability of CIIs

CIIs calculation was feasible in all MBW trials, independent of disease status. The percentage of acceptable breaths for VCap analysis was ∼75% in CF and 65% in control trials. We found lower variability in CIIs compared with SIII and KPIv, but higher compared with LCI. The higher percentage of acceptable breaths together with a lower intra-trial variability found in CF patients may be because they were familiar with the MBW procedure and, thus, able to maintain more stable breath patterns during MBW measurement.

Strengths and limitations

This study presents an innovative method for ventilation inhomogeneity assessment that in principle requires only flow and CO2 signals, thus being potentially more attainable in clinical settings. In our large cohort of both healthy children and children with CF, with a wide age range and diversity in disease severity, these novel CIIs showed a good association with the more complicated LCI, thus confirming their clinical potential. The use of N2-MBW files facilitated direct comparison of MBW and VCap outcomes derived from the same files, limiting the effect of possible influencing factors (e.g. related to breathing pattern or testing circumstances). All MBW trials were analysed with the newest Spiroware version (i.e. 3.3.1), so that our results are not affected by the recently revealed sensor crosstalk error in the Exhalyzer D device [30]. Overall, a posteriori effect size calculations yielded study power in the range of 68–100% for the differences between CIIs in CF and controls and the Pearson's correlation coefficients between CIIs and LCI.

Inevitably, our study also has some limitations. First, VCap analysis was performed retrospectively using previously collected N2-MBW data from one centre. This implies that the baseline quality-control criteria (e.g. those related to breathing pattern and variability) were specific to MBW [21, 22] and not to capnography. This may have affected the intra- and inter-trial variability of CIIs (which was larger than the inter-trial variability of the LCI) and influenced their discriminative ability. Prospective data acquisition for capnographic analysis, including instructions and/or incentives to reduce breathing variability and VCap-specific quality control, may improve the quality of capnograms, decrease the variability of CIIs and increase their diagnostic performance. Second, the effect of 100% O2 concentration (N2-MBW technique) on VCap parameters is unknown. However, the effect, if any, would not be different between CF and controls.

Finally, CIIs calculation was based on the extension of capnograms “to the right”. Arguably, the evolution of events that determine phase III cannot be predicted [25]. Yet, if we assume that these events remain stable during the end of expiration, a forward extension of capnograms may be justified [26]. In fact, this assumption is the basis of the universally accepted Fowler's method to calculate VD [23], where the capnogram is extended “to the left”.

Clinical implications

Volumetric capnography is a simple noninvasive technique that does not require complex respiratory manoeuvres, exogenous gases, extensive signal processing or operator's expertise [2, 5, 7]. Current evidence suggests that classical VCap indices, such as SIII and KPIv, may be sensitive markers of early lung changes in patients with CF [11]; however, their diagnostic performance is limited due to the strong dependence on breathing dynamics [12–14]. Our results indicate that novel CIIs have lower variability compared to classical VCap indices, thus suggesting that they may be less influenced by the breathing pattern; however, further studies are needed to clarify this important issue. Future research should also focus on defining the methodological requirements and the proper quality-control criteria to obtain high-quality capnograms that might reduce variability and increase the performance of these novel indices. The novel CIIs should be assessed using simpler (i.e. non-MBW) setups, ideally on multiple occasions to allow for repeatability assessment, and in large cohorts with a range of obstructive lung disorders (e.g. asthma). Finally, further multicenter research is required to assess the external validity and the potential clinical value of this method.

Conclusions

In conclusion, the herein introduced CIIs performed better than the classical VCap parameters in detecting ventilation inhomogeneity in CF patients. The CIIs also correlated well with the LCI and had lower variability compared to classical VCap indices. Their calculation was feasible in all study participants, independently of disease status. Thus, although their overall diagnostic performance was inferior to the LCI, they may be considered as promising and simpler markers of ventilation inhomogeneity. Further research is required to define the exact methodological requirements, improve the diagnostic performance and assess the true clinical value of CIIs, especially at the bedside.

Supplementary material

Supplementary Material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material 00440-2021.SUPPLEMENT

Footnotes

  • Provenance: Submitted article, peer reviewed.

  • This article has supplementary material available from openres.ersjournals.com

  • Conflicts of interest: S. Fouzas has nothing to disclose.

  • Conflicts of interest: A-C. Kentgens repots no other conflicts of interest.

  • Conflicts of interest: O. Lagiou has nothing to disclose.

  • Conflicts of interest: B.S. Frauchiger has nothing to disclose.

  • Conflicts of interest: F. Wyler has nothing to disclose.

  • Conflicts of interest: I. Theodorakopoulos has nothing to disclose.

  • Conflicts of interest: S. Yammine repots no other conflicts of interest.

  • Conflicts of interest: P. Latzin reports grants from Vertex and Vifor paid to his institution, personal fees and honoraria paid to his institution from Vertex, Vifor and OM Pharma, fees for participation on a data safety monitoring or advisory board from Santhera (DMC), and personal fees paid to his institution from Polyphor, Vertex, OM pharma and Vifor, all outside the submitted work.

  • Support statement: The work was supported by Swiss National Science Foundation Grants (Yammine 179905 and Latzin 182719). A-C. Kentgens is a recipient of the Swiss Government Excellence Scholarship from The Swiss Confederation. Funding information for this article has been deposited with the Crossref Funder Registry.

  • Received July 1, 2021.
  • Accepted December 14, 2021.
http://creativecommons.org/licenses/by-nc/4.0/

This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org

References

  1. ↵
    1. Fletcher R,
    2. Jonson B,
    3. Cumming G, et al.
    The concept of deadspace with special reference to the single breath test for carbon dioxide. Br J Anaesth 1981; 53: 77–88. doi:10.1093/bja/53.1.77
    OpenUrl
  2. ↵
    1. Suarez-Sipmann F,
    2. Bohm SH,
    3. Tusman G
    . Volumetric capnography: the time has come. Curr Opin Crit Care 2014; 20: 333–339. doi:10.1097/MCC.0000000000000095
    OpenUrlCrossRefPubMed
  3. ↵
    1. Fletcher R
    . Airway deadspace, end-tidal CO2, and Christian Bohr. Acta Anaesthesiol Scand 1984; 28: 408–411. doi:10.1111/j.1399-6576.1984.tb02088.x
    OpenUrlPubMed
    1. Tusman G,
    2. Suarez-Sipmann F,
    3. Bohm SH, et al.
    Capnography reflects ventilation/perfusion distribution in a model of acute lung injury. Acta Anaesthesiol Scand 2011; 55: 597–606. doi:10.1111/j.1399-6576.2011.02404.x
    OpenUrlCrossRefPubMed
  5. ↵
    1. Kremeier P,
    2. Böhm SH,
    3. Tusman G
    . Clinical use of volumetric capnography in mechanically ventilated patients. J Clin Monit Comput 2020; 34: 7–16. doi:10.1007/s10877-019-00325-9
    OpenUrl
  6. ↵
    1. Fouzas S,
    2. Hacki C,
    3. Latzin P, et al.
    Volumetric capnography in infants with bronchopulmonary dysplasia. J Pediatr 2014; 164: 283–8.e1-3. doi:10.1016/j.jpeds.2013.09.034
    OpenUrl
  7. ↵
    1. Verscheure S,
    2. Massion PB,
    3. Verschuren F, et al.
    Volumetric capnography: lessons from the past and current clinical applications. Crit Care 2016; 20: 184. doi:10.1186/s13054-016-1377-3
    OpenUrlPubMed
  8. ↵
    1. Krauss B,
    2. Deykin A,
    3. Lam A, et al.
    Capnogram shape in obstructive lung disease. Anesth Analg 2005; 100: 884–888. doi:10.1213/01.ANE.0000146520.90393.91
    OpenUrlCrossRefPubMed
  9. ↵
    1. Strömberg NO,
    2. Gustafsson PM
    . Ventilation inhomogeneity assessed by nitrogen washout and ventilation-perfusion mismatch by capnography in stable and induced airway obstruction. Pediatr Pulmonol 2000; 29: 94–102. doi:10.1002/(SICI)1099-0496(200002)29:2<94::AID-PPUL3>3.0.CO;2-B
    OpenUrlCrossRefPubMed
  10. ↵
    1. Fuchs SI,
    2. Junge S,
    3. Ellemunter H, et al.
    Calculation of the capnographic index based on expiratory molar mass-volume-curves: a suitable tool to screen for cystic fibrosis lung disease. J Cyst Fibros 2013; 12: 277–283. doi:10.1016/j.jcf.2012.08.011
    OpenUrl
  11. ↵
    1. Almeida-Junior A,
    2. Marson FAL,
    3. Almeida CCB, et al.
    Volumetric capnography versus spirometry for the evaluation of pulmonary function in cystic fibrosis and allergic asthma. J Pediatr (Rio J) 2020; 96: 255–264. doi:10.1016/j.jped.2018.10.008
    OpenUrl
  12. ↵
    1. Neufeld GR,
    2. Schwardt JD,
    3. Gobran SR, et al.
    Modelling steady state pulmonary elimination of He, SF6 and CO2: effect of morphometry. Respir Physiol 1992; 88: 257–275. doi:10.1016/0034-5687(92)90001-D
    OpenUrlCrossRefPubMed
    1. Schwardt JD,
    2. Gobran SR,
    3. Neufeld GR, et al.
    Sensitivity of CO2 washout to changes in acinar structure in a single-path model of lung airways. Ann Biomed Eng 1991; 19: 679–697. doi:10.1007/BF02368076
    OpenUrlCrossRefPubMed
  14. ↵
    1. Ream RS,
    2. Schreiner MS,
    3. Neff JD, et al.
    Volumetric capnography in children. Influence of growth on the alveolar plateau slope. Anesthesiology 1995; 82: 64–73. doi:10.1097/00000542-199501000-00010
    OpenUrlCrossRefPubMed
  15. ↵
    1. Schmalisch G
    . Current methodological and technical limitations of time and volumetric capnography in newborns. Biomed Eng Online 2016; 15: 104. doi:10.1186/s12938-016-0228-4
    OpenUrlCrossRefPubMed
  16. ↵
    1. Korten I,
    2. Kieninger E,
    3. Yammine S, et al.
    The Swiss Cystic Fibrosis Infant Lung Development (SCILD) cohort. Swiss Med Wkly 2018; 148: w14618.
    OpenUrlCrossRef
    1. Fuchs O,
    2. Latzin P,
    3. Kuehni CE, et al.
    Cohort profile: the Bern infant lung development cohort. Int J Epidemiol 2012; 41: 366–376. doi:10.1093/ije/dyq239
    OpenUrlCrossRefPubMed
    1. Yammine S,
    2. Nyilas S,
    3. Casaulta C, et al.
    Function and ventilation of large and small airways in children and adolescents with inflammatory bowel disease. Inflamm Bowel Dis 2016; 22: 1915–1922. doi:10.1097/MIB.0000000000000779
    OpenUrl
    1. Nyilas S,
    2. Bauman G,
    3. Pusterla O, et al.
    Ventilation and perfusion assessed by functional MRI in CF patients: reproducibility in comparison to lung function. J Cyst Fibros 2019; 18: 543–550. doi:10.1016/j.jcf.2018.10.003
    OpenUrl
  20. ↵
    1. Frauchiger BS,
    2. Binggeli S,
    3. Yammine S, et al.
    Longitudinal course of clinical lung clearance index in children with cystic fibrosis. Eur Respir J 2020; 24: 2002686.
    OpenUrl
  21. ↵
    1. Frauchiger BS,
    2. Carlens J,
    3. Herger A, et al.
    Multiple breath washout quality control in the clinical setting. Pediatr Pulmonol 2021; 56: 105–112. doi:10.1002/ppul.25119
    OpenUrlPubMed
  22. ↵
    1. Robinson PD,
    2. Latzin P,
    3. Verbanck S, et al.
    Consensus statement for inert gas washout measurement using multiple- and single- breath tests. Eur Respir J 2013; 41: 507–522. doi:10.1183/09031936.00069712
    OpenUrlAbstract/FREE Full Text
  23. ↵
    1. Fowler WS
    . Lung function studies. II: the respiratory deadspace. Am J Physiol 1948; 154: 405–416. doi:10.1152/ajplegacy.1948.154.3.405
    OpenUrlCrossRef
  24. ↵
    1. Fishman AP,
    2. Farhi LE,
    3. Tenney SM,
    4. Geiger SR
    1. Anthonisen NR,
    2. Fleetham JA
    . Ventilation: total, alveolar and dead space. In: Fishman AP, Farhi LE, Tenney SM, Geiger SR, eds. Handbook of Physiology: Section 3: The Respiratory System. Volume IV: Gas Exchange. Bethesda, American Physiological Society, 1987; pp. 113–129.
  25. ↵
    1. Fishman AP,
    2. Farhi LE,
    3. Tenney SM,
    4. Geiger SR
    1. Piiper J,
    2. Scheid P
    . Diffusion and convection in intrapulmonary gas mixing. In: Fishman AP, Farhi LE, Tenney SM, Geiger SR, eds. Handbook of Physiology: Section 3: The Respiratory System. Volume IV: Gas Exchange. Bethesda, American Physiological Society, 1987; pp. 51–69.
  26. ↵
    1. Aitken RS,
    2. Clark-Kennedy AE
    . On the fluctuation in the composition of the alveolar air during the respiratory cycle in muscular exercise. J Physiol 1928; 65: 389–411. doi:10.1113/jphysiol.1928.sp002485
    OpenUrlPubMed
  27. ↵
    1. Veronez L,
    2. Moreira MM,
    3. Soares ST, et al.
    Volumetric capnography for the evaluation of pulmonary disease in adult patients with cystic fibrosis and noncystic fibrosis bronchiectasis. Lung 2010; 188: 263–268. doi:10.1007/s00408-009-9213-z
    OpenUrlPubMed
  28. ↵
    1. Gustafsson PM,
    2. Aurora P,
    3. Lindblad A
    . Evaluation of ventilation maldistribution as an early indicator of lung disease in children with cystic fibrosis. Eur Respir J 2003; 22: 972–979. doi:10.1183/09031936.03.00049502
    OpenUrlAbstract/FREE Full Text
  29. ↵
    1. Aurora P,
    2. Gustafsson P,
    3. Bush A, et al.
    Multiple breath inert gas washout as a measure of ventilation distribution in children with cystic fibrosis. Thorax 2004; 59: 1068–1073. doi:10.1136/thx.2004.022590
    OpenUrlAbstract/FREE Full Text
  30. ↵
    1. Wyler F,
    2. Oestreich MH,
    3. Frauchiger BS, et al.
    Correction of sensor crosstalk error in Exhalyzer D multiple-breath washout device significantly impacts outcomes in children with cystic fibrosis. J Appl Physiol (1985) 2021; 131: 1148–1156. doi:10.1152/japplphysiol.00338.2021
    OpenUrlPubMed
Back to top
Email
Print
Citation Tools
Novel volumetric capnography indices measure ventilation inhomogeneity in cystic fibrosis
Sotirios Fouzas, Anne-Christianne Kentgens, Olga Lagiou, Bettina Sarah Frauchiger, Florian Wyler, Ilias Theodorakopoulos, Sophie Yammine, Philipp Latzin
ERJ Open Research Jan 2022, 8 (1) 00440-2021; DOI: 10.1183/23120541.00440-2021
del.icio.us logo Technorati logo Twitter logo Connotea logo Facebook logo Mendeley logo
Full Text (PDF)

Jump To

Subjects