Intrabreath diffusing capacity of the lung in healthy individuals at rest and during exercise

Application of Machine Learning in Pulmonary Function Assessment Where Are We Now and Where Are We Going?

Analysis of pulmonary function tests (PFTs) is an area where machine learning (ML) may benefit clinicians, researchers, and the patients. PFT measures spirometry, lung volumes, and carbon monoxide diffusion capacity of the lung (DLCO). The results are usually interpreted by the clinicians using discrete numeric data according to published guidelines. PFT interpretations by clinicians, however, are known to have inter-rater variability and the inaccuracy can impact patient care. This variability may be caused by unfamiliarity of the guidelines, lack of training, inadequate understanding of lung physiology, or simply mental lapses. A rules-based automated interpretation system can recapitulate expert’s pattern recognition capability and decrease errors. ML can also be used to analyze continuous data or the graphics, including the flow-volume loop, the DLCO and the nitrogen washout curves. These analyses can discover novel physiological biomarkers. In the era of wearables and telehealth, particularly with the COVID-19 pandemic restricting PFTs to be done in the clinical laboratories, ML can also be used to combine mobile spirometry results with an individual’s clinical profile to deliver precision medicine. There are, however, hurdles in the development and commercialization of the ML-assisted PFT interpretation programs, including the need for high quality representative data, the existence of different formats for data acquisition and sharing in PFT software by different vendors, and the need for collaboration amongst clinicians, biomedical engineers, and information technologists. Hurdles notwithstanding, the new developments would represent significant advances that could be the future of PFT, the oldest test still in use in clinical medicine.

Alveolar-capillary reserve during exercise in patients with chronic obstructive pulmonary disease

Background

Factors limiting exercise in patients with COPD are complex. With evidence for accelerated pulmonary vascular aging, destruction of alveolar-capillary bed, and hypoxic pulmonary vasoconstriction, the ability to functionally expand surface area during exercise may become a primary limitation.

Purpose

To quantify measures of alveolar-capillary recruitment during exercise and the relationship to exercise capacity in a cohort of COPD patients.

Methods

Thirty-two subjects gave consent (53% male, with mean ± standard deviation age 66±9 years, smoking 35±29 pack-years, and Global Initiative for Chronic Obstructive Lung Disease (GOLD) classification of 0-4: 2.3±0.8), filled out the St George's Respiratory Questionnaire (SGRQ) to measure quality of life, had a complete blood count drawn, and underwent spirometry. The intrabreath (IB) technique for lung diffusing capacity for carbon monoxide (IBDLCO) and pulmonary blood flow (IBQc, at rest) was also performed. Subsequently, they completed a cycle ergometry test to exhaustion with measures of oxygen saturation and expired gases.

Results

Baseline average measures were 44±21 for SGRQ score and 58±11 for FEV₁/FVC. Peak oxygen consumption (VO₂) was 11.4±3.1 mL/kg/min (49% predicted). The mean resting IBDLCO was 9.7±5.4 mL/min/mmHg and IBQc was 4.7±0.9 L/min. At the first workload, heart rate (HR) increased to 92±11 bpm, VO₂ was 8.3±1.4 mL/kg/min, and IBDLCO and IBQc increased by 46% and 43%, respectively, compared to resting values (p,0.01). The IBDLCO/Qc ratio averaged 2.0±1.1 at rest and remained constant during exercise with marked variation across subjects (range: 0.8-4.8). Ventilatory efficiency plateaued at 37±5 during exercise, partial pressure of mix expired CO₂/partial pressure of end tidal CO₂ ratio ranged from 0.63 to 0.67, while a noninvasive index of pulmonary capacitance, O₂ pulse × PetCO₂ (GxCap) rose to 138%. The exercise IBDLCO/Qc ratio was related to O₂ pulse (VO₂/HR, r=0.58, p<0.01), and subjects with the highest exercise IBDLCO/Qc ratio or the greatest rise from rest had the highest peak VO₂ values (r=0.65 and 0.51, respectively, p<0.05). Of the noninvasive gas exchange measures of pulmonary vascular function, GxCap was most closely associated with DLCO, DLCO/Qc, and VO₂ peak.

Conclusion

COPD patients who can expand gas exchange surface area as assessed with DLCO during exercise relative to pulmonary blood flow have a more preserved exercise capacity.

Influence of resting lung diffusion on exercise capacity in patients with COPD

Background

Lung diffusing capacity for carbon monoxide (DLCO) gives an overall assessment of functional lung surface area for gas exchange and can be assessed using various methods. DLCO is an important factor in exercise intolerance in patients with chronic obstructive pulmonary disease (COPD). We investigated if the intra-breath (IBDLCO) method may give a more sensitive measure of available gas exchange surface area than the more typical single breath (SBDLCO) method and if COPD subjects with the largest resting DLCO relative to pulmonary blood flow (Qc) would have a more preserved exercise capacity.

Methods

Informed consent, hemoglobin, spirometry, SBDLCO, IBDLCO, and Qc during IBDLCO were performed in moderate to severe COPD patients, followed by progressive cycle ergometry to exhaustion with measures of oxygen saturation (SaO₂) and expired gases.

Results

Thirty two subjects (47% female, age 66 ± 9 yrs., BMI 30.4 ± 6.3 kg/m², smoking hx 35 ± 29 pkyrs, 2.3 ± 0.8 on the 0-4 GOLD classification scale) participated. The majority used multiple inhaled medications and 20% were on oral steroids. Averages were: FEV₁/FVC 58 ± 10%Pred, peak VO₂ 11.4 ± 3.1 ml/kg/min, and IBDLCO 72% of the SBDLCO (r = 0.88, SB vs IB methods). Using univariate regression, both the SB and IBDLCO (% predicted but not absolute) were predictive of VO₂peak in ml/kg/min; SBDLCO/Qc (r = 0.63, p < 0.001) was the best predictor of VO₂peak; maximal expiratory flows over the mid to lower lung volumes were the most significantly predictive spirometric measure (r = 0.49, p < 0.01). However, in multivariate models only BMI added additional predictive value to the SBDLCO/Qc for predicting aerobic capacity (r = 0.73). Adjusting for current smoking status and gender did not significantly change the primary results.

Conclusion

In patients with moderate to severe COPD, preservation of lung gas exchange surface area as assessed using the resting SBDLCO/Qc appears to be a better predictor of exercise capacity than more classic measures of lung mechanics.

Underdiagnosis and overdiagnosis of asthma in the morbidly obese

BACKGROUND

The prevalence of obesity and asthma has increased concurrently over the last decades, suggesting a link between obesity and asthma. However, asthma might not be adequately diagnosed in this population.

AIM

To investigate whether not only overdiagnosis but also underdiagnosis of asthma is present in an obese population.

METHODS

Morbidly obese subjects with or without physician-diagnosed asthma were recruited from a pre-operative screening programme for bariatric surgery, and were characterized using an extensive diagnostic algorithm.

RESULTS

473 subjects were screened; 220 met inclusion criteria, and 86 agreed to participate. Among the 32 participating subjects who had a physician diagnosis of asthma, reversible airway obstruction and/or bronchial hyperresponsiveness could only be detected in 19 patients (59%, 95% CI [0.41-0.76]), whereas in 13 patients (41%, 95% CI [0.24-0.50]) the diagnosis of asthma could not be confirmed (overdiagnosis). In contrast, in the remaining 54 patients, 17 (31%, 95% CI [0.20-0.46]) were newly diagnosed with asthma (underdiagnosis).

CONCLUSION

Besides overdiagnosis, there is also substantial underdiagnosis of asthma in the morbidly obese. Symptoms could be incorrectly ascribed to either obesity or asthma, and therefore also in the morbidly obese the diagnosis of asthma should also be based on pulmonary function testing.

Pulmonary gas exchange and acid-base balance during exercise

As the first step in the oxygen-transport chain, the lung has a critical task: optimizing the exchange of respiratory gases to maintain delivery of oxygen and the elimination of carbon dioxide. In healthy subjects, gas exchange, as evaluated by the alveolar-to-arterial PO2 difference (A-aDO2), worsens with incremental exercise, and typically reaches an A-aDO2 of approximately 25 mmHg at peak exercise. While there is great individual variability, A-aDO2 is generally largest at peak exercise in subjects with the highest peak oxygen consumption. Inert gas data has shown that the increase in A-aDO2 is explained by decreased ventilation-perfusion matching, and the development of a diffusion limitation for oxygen. Gas exchange data does not indicate the presence of right-to-left intrapulmonary shunt developing with exercise, despite recent data suggesting that large-diameter arteriovenous shunt vessels may be recruited with exercise. At the same time, multisystem mechanisms regulate systemic acid-base balance in integrative processes that involve gas exchange between tissues and the environment and simultaneous net changes in the concentrations of strong and weak ions within, and transfer between, extracellular and intracellular fluids. The physicochemical approach to acid-base balance is used to understand the contributions from independent acid-base variables to measured acid-base disturbances within contracting skeletal muscle, erythrocytes and noncontracting tissues. In muscle, the magnitude of the disturbance is proportional to the concentrations of dissociated weak acids, the rate at which acid equivalents (strong acid) accumulate and the rate at which strong base cations are added to or removed from muscle.

Reduced exercise tolerance and pulmonary capillary recruitment with remote secondhand smoke exposure

Rationale

Flight attendants who worked on commercial aircraft before the smoking ban in flights (pre-ban FAs) were exposed to high levels of secondhand smoke (SHS). We previously showed never-smoking pre-ban FAs to have reduced diffusing capacity (Dco) at rest.

Methods

To determine whether pre-ban FAs increase their Dco and pulmonary blood flow () during exercise, we administered a symptom-limited supine-posture progressively increasing cycle exercise test to determine the maximum work (watts) and oxygen uptake () achieved by FAs. After 30 min rest, we then measured Dco and at 20, 40, 60, and 80 percent of maximum observed work.

Results

The FAs with abnormal resting Dco achieved a lower level of maximum predicted work and compared to those with normal resting Dco (mean±SEM; 88.7±2.9 vs. 102.5±3.1%predicted ; p = 0.001). Exercise limitation was associated with the FAs' FEV₁ (r = 0.33; p = 0.003). The Dco increased less with exercise in those with abnormal resting Dco (mean±SEM: 1.36±0.16 vs. 1.90±0.16 ml/min/mmHg per 20% increase in predicted watts; p = 0.020), and amongst all FAs, the increase with exercise seemed to be incrementally lower in those with lower resting Dco. Exercise-induced increase in was not different in the two groups. However, the FAs with abnormal resting Dco had less augmentation of their Dco with increase in during exercise (mean±SEM: 0.93±0.06 vs. 1.47±0.09 ml/min/mmHg per L/min; p<0.0001). The Dco during exercise was inversely associated with years of exposure to SHS in those FAs with ≥10 years of pre-ban experience (r = −0.32; p = 0.032).

Conclusions

This cohort of never-smoking FAs with SHS exposure showed exercise limitation based on their resting Dco. Those with lower resting Dco had reduced pulmonary capillary recruitment. Exposure to SHS in the aircraft cabin seemed to be a predictor for lower Dco during exercise.

High-altitude exposure of three weeks duration increases lung diffusing capacity in humans

Background: high-altitude adaptation leads to progressive increase in arterial PaO2. In addition to increased ventilation, better arterial oxygenation may reflect improvements in lung gas exchange. Previous investigations reveal alterations at the alveolar-capillary barrier indicative of decreased resistance to gas exchange with prolonged hypoxia adaptation, but how quickly this occurs is unknown. Carbon monoxide lung diffusing capacity and its major determinants, hemoglobin, alveolar volume, pulmonary capillary blood volume, and alveolar-capillary membrane diffusion, have never been examined with early high-altitude adaptation. Methods and Results: lung diffusion was measured in 33 healthy lowlanders at sea level (Milan, Italy) and at Mount Everest South Base Camp (5,400 m) after a 9-day trek and 2-wk residence at 5,400 m. Measurements were adjusted for hemoglobin and inspired oxygen. Subjects with mountain sickness were excluded. After 2 wk at 5,400 m, hemoglobin oxygen saturation increased from 77.2 ± 6.0 to 85.3 ± 3.6%. Compared with sea level, there were increases in hemoglobin, lung diffusing capacity, membrane diffusion, and alveolar volume from 14.2 ± 1.2 to 17.2 ± 1.8 g/dl ( P < 0.01), from 23.6 ± 4.4 to 25.1 ± 5.3 ml·min−1·mmHg−1 ( P < 0.0303), 63 ± 34 to 102 ± 65 ml·min−1·mmHg−1 ( P < 0.01), and 5.6 ± 1.0 to 6.3 ± 1.1 liters ( P < 0.01), respectively. Pulmonary capillary blood volume was unchanged. Membrane diffusion normalized for alveolar volume was 10.9 ± 5.2 at sea level rising to 16.0 ± 9.2 ml·min−1·mmHg−1·l−1 ( P < 0.01) at 5,400 m. Conclusions: at high altitude, lung diffusing capacity improves with acclimatization due to increases of hemoglobin, alveolar volume, and membrane diffusion. Reduction in alveolar-capillary barrier resistance is possibly mediated by an increase of sympathetic tone and can develop in 3 wk.

[Non-invasive study of pulmonary vascular recruitment during exercise]

The in vivo study of the pulmonary microcirculation, and its recruitment, is currently not common, although it may be of interest. The intrabreath analysis (IB) of the carbon monoxide (CO) and acetylene (C(2)H(2)) diffusion is used to study the transfer of CO (TLCO) and the pulmonary capillary blood flow (Qs), particularly during exercise. The evolution of the Qs during different stages of exercise has never been reported in healthy subjects. The authors measured the Qs and TL at rest and then during and after short bouts of exercise in 12 healthy subjects. The Qs increased from 5.6 L/min at rest to 13.8 L/min during exercise while the TLCO increased from 11 to 16.7 mmol/kPa/min. A linear relationship was found between the Qs and the TLCO, with Qs values close to those obtained with other techniques. The Qs returned to rest values more rapidly than the TLCO (probably because of the membrane factor). Pulmonary vascular recruitment can be easily studied in healthy subjects. This parameter may be important in the study in pulmonary vascular diseases.

Assessment of the alveolar volume when sampling exhaled gas at different expired volumes in the single breath diffusion test

BACKGROUND

Alveolar volume measured according to the American Thoracic Society-European Respiratory Society (ATS-ERS) guidelines during the single breath diffusion test can be underestimated when there is maldistribution of ventilation. Therefore, the alveolar volume calculated by taking into account the ATS-ERS guidelines was compared to the alveolar volume measured from sequentiallly collected samples of the expired volume in two groups of individuals: COPD patients and healthy individuals. The aim of this study was to investigate the effects of the maldistribution of ventilation on the real estimate of alveolar volume and to evaluate some indicators suggestive of the presence of maldistribution of ventilation.

METHODS

Thirty healthy individuals and fifty patients with moderate-severe COPD were studied. The alveolar volume was measured either according to the ATS-ERS guidelines or considering the whole expired volume subdivided into five quintiles. An index reflecting the non-uniformity of the distribution of ventilation was then derived (DeltaVA/VE).

RESULTS

Significant differences were found when comparing the two measurements and the alveolar volume by quintiles appeared to have increased progressively towards residual volume in healthy individuals and much more in COPD patients. Therefore, DeltaVA/VE resulted in an abnormal increase in COPD.

CONCLUSION

The results of our study suggest that the alveolar volume during the single breath diffusion test should be measured through the collection of a sample of expired volume which could be more representative of the overall gas composition, especially in the presence of uneven distribution of ventilation. Further studies aimed at clarifying the final effects of this way of calculating the alveolar volume on the measure of DLCO are needed. DeltaVA/VE is an index that can help assess the severity of inhomogeneity in COPD patients.

Comparison of Total-Breath and Single-Breath Diffusing Capacity in Healthy Volunteers and COPD Patients

BACKGROUND

The measurement of single-breath diffusing capacity (Dlco(SB)) assumes that diffusing capacity per liter of alveolar volume (Dlco/VA) determined in a 750-mL gas sample represents the diffusing capacity (Dlco) of the entire lung. Fast-responding gas analyzers provide the opportunity to verify this assumption because of the possibility to measure CO and CH(4) fractions continuously throughout the entire expiration. Continuous gas sampling provides more information per measurement, but this information cannot be expressed in the traditional parameters. Our goals were to find new parameters to express the extra information of the continuous gas sampling, and to compare these new parameters with the traditional parameters.

METHODS

We compared a new method to determine Dlco with the traditional method in 62 healthy volunteers and 26 COPD patients. Traditionally, Dlco(SB) is determined by multiplying Dlco/VA with alveolar volume, both calculated from gas concentrations in a 750-mL gas sample. The new method calculates total-breath Dlco (Dlco(TB)) by integration of Dlco/VA against exhaled volume.

RESULTS

In healthy volunteers, Dlco/VA shows a slight upward slope during exhalation, while in COPD patients Dlco/VA shows a horizontal line. Total-breath total lung capacity (TLC) is larger than single-breath TLC both in healthy volunteers and in COPD patients, leading to a Dlco(TB) that is significantly larger than Dlco(SB) in both groups (p < 0.001).

CONCLUSION

The assumption that a 750-mL gas sample represents the entire lung seems to be correct for Dlco/VA but not for the CH(4) fraction in case of ventilation inhomogeneity.

Stroke volume response during exercise measured by acetylene uptake and MRI

The intra-breath technique to measure acetylene absorption offers the possibility to determine augmentation of the pulmonary blood flow per heart beat (Q(C)) as an estimate of the stroke volume response during exercise. However, this method has not been compared with a validated test until now. Therefore, the aim of this study was to compare Q(C) with stroke volume (SV(MRI)) determined by magnetic resonance imaging (MRI) at rest and during exercise in healthy subjects and patients. For this purpose, ten healthy subjects and ten patients with idiopathic pulmonary arterial hypertension (iPAH) with expected impaired stoke volume response during exercise were measured by both methods. Exercise-induced changes in Q(C) and SV(MRI) were correlated in healthy controls (r = 0.75, p < 0.05). Compared to healthy controls, Q(C) increased less during exercise in iPAH patients (11 +/- 17 ml versus 33 +/- 12 ml, p < 0.05). A similar difference in stroke volume response to exercise between the two groups was measured by MRI (-0.6 +/- 8 ml versus 23 +/- 12 ml, p < 0.05, respectively). Hence, intra-breath and MRI measurements showed similar differences in exercise-induced changes in stroke volume between controls and patients. From these results it can be concluded that the intra-breath measurement of acetylene absorption might be of value as a non-invasive tool to estimate stroke volume augmentation during exercise and can detect differences in stroke volume responses between iPAH patients and healthy subjects.

Intrabreath analysis of carbon monoxide uptake during exercise in patients at risk for lung injury

The single exhalation analysis of carbon monoxide, acetylene, and methane allows the determination of intrabreath (regional) DL, pulmonary capillary blood flow and ventilation inhomogeneities during rest and exercise. We reasoned that this technique might be more sensitive in detecting regional pulmonary capillary abnormalities than resting single breath DL (DL(sb)). We selected a group of breast cancer patients in high-dose chemotherapy (HDCT) protocols who were at risk for pulmonary injury. We grouped the patients into pre-HDCT and post-HDCT, and used resting DL(sb) to further categorize the latter into those with and without pulmonary injury. We found that exercise DL increases were blunted in post-HDCT patients with low resting DL(sb). More importantly, even in post-HDCT patients with normal resting DL(sb), exercise DL response was reduced in the slowest emptying lung units along with evidence for ventilation inhomogeneities (increased methane slope). We conclude that exercise assessments of DL at low lung volumes and gas mixing properties may be sensitive indicators of lung injury.

An open-circuit method for determining lung diffusing capacity during exercise: comparison to rebreathe

To avoid limitations associated with the use of single-breath and rebreathe methods for assessing the lung diffusing capacity for carbon monoxide (DlCO) during exercise, we developed an open-circuit technique. This method does not require rebreathing or alterations in breathing pattern and can be performed with little cognition on the part of the patient. To determine how this technique compared with the traditional rebreathe (DlCO,RB) method, we performed both the open-circuit (DlCO,OC) and the DlCO,RBmethods at rest and during exercise (25, 50, and 75% of peak work) in 11 healthy subjects [mean age = 34 yr (SD 11)]. Both DlCO,OCand DlCO,RBincreased linearly with cardiac output and external work. There was a good correlation between DlCO,OCand DlCO,RBfor rest and exercise (mean of individual r2= 0.88, overall r2= 0.69, slope = 0.97). DlCO,OCand DlCO,RBwere similar at rest and during exercise [e.g., rest = 27.2 (SD 5.8) vs. 29.3 (SD 5.2), and 75% peak work = 44.0 (SD 7.0) vs. 41.2 ml·min−1·mmHg−1(SD 6.7) for DlCO,OCvs. DlCO,RB]. The coefficient of variation for repeat measurements of DlCO,OCwas 7.9% at rest and averaged 3.9% during exercise. These data suggest that the DlCO,OCmethod is a reproducible, well-tolerated alternative for determining DlCO, particularly during exercise. The method is linearly associated with cardiac output, suggesting increased alveolar-capillary recruitment, and values were similar to the traditional rebreathe method.