Comparison of two devices and two breathing patterns for exhaled breath condensate sampling

Abstract

Introduction: Analysis of exhaled breath condensate (EBC) is a noninvasive method to access the epithelial lining fluid of the lungs. Due to standardization problems the method has not entered clinical practice. The aim of the study was to assess the comparability for two commercially available devices in healthy controls. In addition, we assessed different breathing patterns in healthy controls with protein markers to analyze the source of the EBC.

Methods: EBC was collected from ten subjects using the RTube and ECoScreen Turbo in a randomized crossover design, twice with every device--once in tidal breathing and once in hyperventilation. EBC conductivity, pH, surfactant protein A, Clara cell secretory protein and total protein were assessed. Bland-Altman plots were constructed to display the influence of different devices or breathing patterns and the intra-class correlation coefficient (ICC) was calculated. The volatile organic compound profile was measured using the electronic nose Cyranose 320. For the analysis of these data, the linear discriminant analysis, the Mahalanobis distances and the cross-validation values (CVV) were calculated.

Results: Neither the device nor the breathing pattern significantly altered EBC pH or conductivity. ICCs ranged from 0.61 to 0.92 demonstrating moderate to very good agreement. Protein measurements were greatly influenced by breathing pattern, the device used, and the way in which the results were reported. The electronic nose could distinguish between different breathing patterns and devices, resulting in Mahalanobis distances greater than 2 and CVVs ranging from 64% to 87%.

Conclusion: EBC pH and (to a lesser extent) EBC conductivity are stable parameters that are not influenced by either the device or the breathing patterns. Protein measurements remain uncertain due to problems of standardization. We conclude that the influence of the breathing maneuver translates into the necessity to keep the volume of ventilated air constant in further studies.

Figure 1. The figure shows the sample volume of exhaled breath condensate (EBC) after ten minutes collection.

a. After tidal breathing (TB) and hyperventilation (TB) EBC volumes with RTube were higher compared with ECoScreen turbo (p<0.001) Hyperventilation causes higher EBC volumes compared with tidal breathing in both devices (p<0.0001 in RTube, p = 0.056 in ECoScreen). b. Hyperventilation via the ECoScreen turbo caused a 1.78 fold higher movement of ventilated air (p<0.0001).

Figure 2. Bland Altman Plots are shown to display differences in individual measurements of the exhaled breath condensate (EBC) under certain conditions.

Neither the device nor the ventilation pattern changed the EBC pH significantly. a. RTube and ECoScreen showed comparable pH values in EBC in tidal breathing (TB). b. RTube and ECoScreen showed comparable pH values in EBC in hyperventilation (H). c. The breathing manoeuvres did not produce significant differences in EBC pH using the ECoScreen. d. The breathing manoeuvres did not produce significant differences in EBC pH using the RTube.

Figure 3. Bland Altman Plots are shown comparing the conductivity of exhaled breath condensate under different conditions.

Neither the device nor the breathing pattern changed the conductivity of the EBC significantly. a. In tidal breathing (TB) RTube and ECoScreen did not produce statistically different conductivity values. b. RTube and ECoScreen did not produce statistically different conductivity values collecting EBC with a hyperventilation (H) maneuver. c. The breathing manoeuvres did not produce significant differences in EBC conductivity using the ECoScreen. d. The breathing manoeuvres did not produce significant differences in EBC conductivity using the RTube.

Figure 4. Displayed are the overall protein measurements in four different ways.

a. Comparing the ECoScreen and RTube EBC protein concentration after tidal breathing (TB) no statistical significant difference could be shown (p = 0.51). After hyperventilation (H) ECoScreen resulted in higher protein concentrations than the RTube (p<0.001). Comparing the two manoeuvres, hyperventilation yielded higher concentrations than tidal breathing, but this difference was significant only in the ECoScreen (p<0.0001). b. To the volume of ventilated air normalized protein concentrations in EBC collected by the ECoScreen device did not show a difference between tidal breathing and hyperventilation (p = 1). c. Analyzing the total protein amount in EBC, hyperventilation with ECoScreen resulted in higher protein values compared to RTube (p<0.05). Comparing hyperventilation with tidal breathing in the ECoScreen device, hyperventilation resulted in higher absolute protein amounts (p<0.001). d. By normalizing the absolute protein amount in EBC to the volume of ventilated air using the ECoScreen turbo hyperventilation expressed higher overall protein values/ventilated volume (p<0.05).Within the same device hyperventilation yielded higher overall protein concentrations of EBC than tidal breathing, though the difference was statistically significant only in the ECoScreen device (H: 0.033 mg/ml±0.008 mg/ml vs. TB: 0.016 mg/ml±0.003 mg/ml; p<0.05; figure 4a, grey bars).

Figure 5. Specific protein measurements are displayed in four different ways.

a. The breathing manoeuvres tidal breathing (TB) and hyperventilation (H) and also the devices RTube and ECoScreen turbo had no effect on the total concentration of Clara cell protein (CCP) and surfactant protein-A (SP-A), respectively (p = 0.17; p = 0.16). b. Normalizing the CCP and SP-A protein concentrations to ventilated volume revealed lower CCP and SP-A values under hyperventilation conditions (p<0.001; p<0.0001). c. Absolute amount of CCP and SP-A. Hyperventilation leads to significant higher SP-A and CCP levels (p<0.0001 for both). d. Normalizing the absolute amount of SP-A and CCP to the volume of ventilated air resulted in no significant difference of CCP and SP-A levels comparing hyperventilation with tidal breathing.

Figure 6. The three-dimensional plot of the linear discriminant (LD) analysis shows that two breathing patterns and the two devices were clearly separable using the electronic Nose (Cyranose 320).