Skip to main content
NCBI home page
Respiratory Care logoLink to Respiratory Care
editorial
. 2024 Jan;69(1):1–14. doi: 10.4187/respcare.10810

Automated O2 Titration Alone or With High-Flow Nasal Cannula During Walking Exercise in Chronic Lung Diseases

Felix-Antoine Vézina 1, Pierre-Alexandre Bouchard 2, Émilie Breton-Gagnon 3, Geneviève Dion 4, Damien Viglino 5, Pascalin Roy 6, Lara Bilodeau 7, Steeve Provencher 8, Marie-Hélène Denault 9, Didier Saey 10, François Lellouche 11, François Maltais 12,✉
PMCID: PMC10753602  PMID: 37491073

Abstract

BACKGROUND:

Exercise-induced O2 desaturation contributes to dyspnea and exercise intolerance in various respiratory diseases. This study assessed whether automated O2 titration was superior to fixed-flow O2 to improve exertional dyspnea and walking exercise endurance. We also aimed at evaluating possible additive effects of high-flow nasal cannula coupled with automated O2 titration on these outcomes.

METHODS:

Subjects with chronic respiratory diseases and exercise-induced desaturation performed a 3-min constant-speed shuttle test (CSST) and an endurance shuttle walking test (ESWT) with either (1) fixed-flow O2, (2) automated O2 titration targeting an SpO2 of 94% (± 2%), and (3) automated O2 titration + high-flow nasal cannula according to a randomized sequence. The main outcome was Borg dyspnea score at the end of the 3-min CSST. Secondary outcomes included endurance time and dyspnea during ESWT and oxygenation status during exercise.

RESULTS:

Ten subjects with COPD, 10 with interstitial lung disease, 5 with pulmonary hypertension, and 3 with cystic fibrosis completed the study. Compared to fixed-flow O2, automated O2 titration did not reduce dyspnea at the end of the 3-min CSST. Endurance time during the ESWT was prolonged with automated O2 titration (mean difference 298 [95% CI 205–391] s, P < .001), and dyspnea at isotime was reduced. No further improvement was noted when high-flow nasal cannula was added to automated O2 titration. Compared to fixed-flow O2, O2 flows were higher with automated O2 titration, resulting in better oxygenation.

CONCLUSIONS:

Automated O2 titration was superior to fixed-flow O2 to alleviate dyspnea and improve exercise endurance during the ESWT in subjects with a variety of chronic respiratory diseases. Adding high-flow nasal cannula to automated O2 titration provided no further benefits.

Keywords: oxygen supplementation, exercise, desaturation, automated oxygen titration, high-flow nasal oxygen, dyspnea, COPD, interstitial lung disease, pulmonary hypertension, cystic fibrosis

Introduction

Exertional dyspnea is a shared feature and often the most troublesome symptom in a variety of chronic respiratory diseases. Patients with COPD1, interstitial lung disease (ILD),2 pulmonary hypertension,3 and cystic fibrosis (CF)4 all complain of dyspnea as a limiting symptom in their daily life. The pathophysiology of exertional dyspnea in chronic respiratory diseases is complex, with each specific disease involving specific mechanisms. However, in all of them, exercise-induced O2 desaturation may play a role.

When tested in laboratory settings, supplemental oxygen reduces exertional dyspnea and improves exercise tolerance in subjects with COPD and ILD.5–8 Despite this strong supportive physiological rationale, ambulatory oxygen has negligible and unconvincing effects on functional status and quality of life when used in daily life.9,10 The fact that there is little to no translation of the laboratory findings to the clinical settings is multifactorial, including the inconvenience of transporting cumbersome oxygen systems and perceived stigma of going out with portable oxygen that lead to little or no use of ambulatory oxygen.9,11 Importantly, however, currently available ambulatory oxygen systems are often unable to maintain oxygenation as the oxygen needs increase in parallel with the progression of exercise,12 thus failing to reduce ventilatory drive and, therefore, dyspnea.13–15

Automated O2 titration has been developed with the objective of maintaining O2 saturation at the desired level by providing continuous adjustment of O2 flows in the context of fluctuating oxygen needs.16 This strategy produced better maintenance of O2 saturation within the target zone in a variety of clinical situations,17,18 including exercise where this novel O2 delivery system may reduce exertional dyspnea and enhance exercise tolerance in subjects with COPD.15,19–21 Whether these encouraging findings would similarly apply to other types of chronic respiratory diseases also associated with exercise-induced O2 desaturation is uncertain.

High-flow nasal cannula is another technique that is being considered to enhance exercise tolerance in subjects with chronic respiratory diseases.22–25 Proposed mechanisms of action that are complementary to those of supplemental oxygen include (1) reduced breathing frequency, (2) reduced dead-space ventilation, and (3) generation of a small PEEP-like effect counterbalancing intrinsic PEEP and reducing work of breathing.26,27 Interestingly, automated O2 titration can be coupled with high-flow nasal cannula to potentiate their respective efficacy to alleviate exertional dyspnea and improve exercise tolerance.

The objectives of this physiological study were to evaluate the efficacy of automated O2 titration in comparison to fixed-flow O2 on exertional dyspnea and exercise tolerance in subjects with various chronic respiratory diseases leading to oxygen desaturation during exercise. We also aimed at evaluating if the addition of high-flow nasal cannula to automated O2 titration was associated with further benefits compared to automated O2 titration alone on the same outcomes. The 3-min constant-speed shuttle test (3-min CSST), a walking exercise test that was specifically designed to quantify the effects of interventions on exertional dyspnea,28,29 was used to document the comparative efficacy of study interventions on the primary outcome of dyspnea, whereas the endurance shuttle walking test (ESWT) was used to measure their impact on walking endurance time.30,31

QUICK LOOK.

Current knowledge

Fixed-flow oxygen supplementation has been proposed to enhance exercise tolerance and to reduce exertional dyspnea in subjects with chronic respiratory diseases. Whether automated O2 titration alone or with high-flow nasal cannula during exercise provides further benefits on these outcomes compared to fixed-flow oxygen is uncertain.

What this paper contributes to our knowledge

In this study, we found that automated O2 titration did improve exercise endurance and dyspnea during the endurance shuttle walking test in comparison to fixed-flow O2 in subjects with a variety of chronic respiratory diseases featuring exercise-induced O2 desaturation. The use of high-flow nasal cannula in conjunction with automated O2 titration did not provide further benefits.

Methods

This was a 3-treatment arm, crossover, and randomized trial (ClinicalTrials.gov NCT05267418). After the evaluation (visit 1) and preparation for the treatment phase of the study (visit 2), participants were requested to complete 3 treatment visits (visits 3, 4, and 5) during which they performed one 3-min CSST and one ESWT (Fig. 1). Experimental visits were identical with the exception of the study treatment to which participants were exposed: (1) fixed-flow O2, (2) automated O2 titration (FreeO2, OxyNov, Lebourgneuf, Québec City, Canada), and (3) automated O2 titration with high-flow nasal cannula (Airvo 2, Fisher & Paykel Healthcare, Auckland, New Zealand). Study treatments were administered according to one of the 3 sequences to which study participants were randomly assigned (Fig. 2).

Fig. 1.

Fig. 1.

Open in a new tab

Schematic representation of study visits.

Fig. 2.

Fig. 2.

Open in a new tab

Flow chart.

The main study objective was to compare the effects of automated O2 titration versus fixed-flow O2 on Borg dyspnea score at the end of the 3-min CSST (primary outcome). Exploratory outcomes included the endurance time during the ESWT, dyspnea at isotime during the ESWT, mean and nadir O2 SpO2 during the 3-min CSST and ESWT, proportion of the time spent with SpO2 < 90% and 85% during the 3-min CSST and ESWT, and changes in capillary PCO2 from rest to end of the ESWT. We also aimed at evaluating whether providing high-flow nasal cannula in addition to automated O2 titration would provide further benefits on the same outcomes. The study was approved by the intuitional ethics committee (CER 21726), and all study participants provided written consent.

Patients followed at the ambulatory pulmonary clinics at the Institut Universitaire de Cardiologie et de Pneumologie de Québec with a diagnosis of COPD, ILD, pulmonary hypertension, or CF were considered for study participation. Study participants had prior documentation of exercise-induced oxygen desaturation as defined by a ≥ 5% fall in SpO2 with a nadir SpO2 < 88% during a 6-min walk test. Participants were in stable condition, with no exacerbation of their primary disease during the 8-week period preceding study participation. COPD was ascertained from current or past smoking exposure of at least 10 pack-years, compatible symptoms, and fixed air-flow obstruction (FEV1 < 70% predicted and FEV1/FVC < 0.7). Subjects with ILD had computed tomography imaging compatible with usual interstitial pneumonia, non-specific interstitial pneumonitis, or hypersensitivity pneumonitis along with a restrictive ventilatory pattern (total lung capacity < 80% predicted). Individuals with class 1 or 4 pulmonary hypertension were allowed into the trial based on a mean pulmonary artery pressure ≥ 25 mm Hg. CF was diagnosed in the presence of a positive sweat chloride > 60 mmol/L or in the presence of 2 CF-causing mutations. Apart from their baseline respiratory condition, study participants had no other disease that would limit or contraindicate the study-related exercise testing procedures. Patients with symptomatic cardiovascular disease, sarcoidosis, or those with connective tissue–related ILD/pulmonary hypertension were excluded.

After determination of study eligibility, subjects were asked to complete 5 visits as illustrated in Fig. 1. During visit 1, pulmonary function was assessed followed by an incremental shuttle walking test (ISWT) to characterize exercise capacity and document the walking speed for the ESWT (see related supplementary materials at http://www.rcjournal.com). At visit 2, two or three 3-min CSST were done to determine the shuttle speed that would produce a Borg dyspnea score of at least 4 points (moderate) while being sustainable for the entire 3-min test duration.29,32 An ESWT was then performed at a walking speed aiming for a 4–8-min test duration.33 Visits 3, 4, and 5 included, always in this order, a 3-min CSST that was performed at the speed determined at visit 2, followed, after 1 h of rest, by an ESWT. These visits were identical except for the treatment that was administered, fixed-flow O2, automated O2 titration alone, or with high-flow nasal cannula. Considering an estimated study sample size of 30 subjects (see below), we limited the number of possible treatment sequences to 3: (1) fixed-flow O2, automated O2 titration alone, and then automated O2 titration + high-flow nasal cannula; (2) automated O2 titration alone, automated O2 titration + high-flow nasal cannula, and then fixed-flow O2; and (3) automated O2 titration + high-flow nasal cannula, fixed-flow O2 alone, and then automated O2 titration. Subjects were randomized to one of these 3 sequences using a randomization table produced by the study biostatistician. Capillary blood was collected at the end of each ESWT to measure capillary PCO2. Besides oxygen, there were no other therapeutic interventions, and study participants continued their standard pharmacologic treatment regimen.

Study participants and research staff supervising the study procedures could not be fully blinded from experimental conditions, but several strategies to minimize bias were deployed. Supplemental oxygen was always delivered by the FreeO2 system, which was used either in the fixed-flow or automatic titration mode depending on the experimental condition so that study participants and study assessor (the person who recorded Borg dyspnea scores and endurance time) could not determine which type of oxygen system was being used. It was not possible to blind for the use of high-flow nasal cannula, but study participants were not aware of the expected benefits of this therapy and of the study objectives. The research assistant whose role was to ensure that the oxygen and high-flow equipment was working properly could not be blinded to the experimental condition.

Fixed-flow O2 was administered with the FreeO2 system in the fixed mode using nasal prong at a flow of 2 L/min or, for oxygen-dependent subjects, at the basal O2 flow + 1 L/min, as commonly done in clinical practice.19 Automated O2 titration was done with the FreeO2 system in the automated mode that continuously adjusts O2 flow using a closed-loop algorithm based on a target SpO2,15 which was set at 94 ± 2% for the purpose of this study. With this system, maximum O2 flow is 20 L/min. The FreeO2 system was coupled with a finger sensor linked to an oximeter (Nonin Medical, Plymouth, Minnesota), providing continuous monitoring and recording of SpO2. For the third experimental condition, FreeO2 was used in conjunction with the Airvo 2 system set at 40 L/min of a mixture of air provided from ambient air via a turbine and compressed O2 humidified and warmed at 37°C using the nasal prong specifically designed for use with this system. FreeO2 was used in the closed-loop mode in this condition with SpO2 target identical to those of the automated O2 titration arm. FreeO2 and Airvo 2 systems were attached on a mobile stand that was displaced by the research staff to follow the participant during the walk, thus ensuring no interference with the walking tests.

The 3-min CSST, with its fixed 3-min duration and externally imposed walking speed, was specifically designed to assess the effects of interventions on dyspnea.28,32 At visit 2, and following a published algorithm for shuttle speed selection,29 subjects were instructed to walk between 2 cones positioned as for the ISWT at 2 different speeds in order to determine, among the 5 possible walking speeds (2.5, 3.25, 4, 5, and 6 km/h), the highest shuttle speed that could be sustained for the entire 3 min and that would produce a score of at least 4 points on the Borg modified 0–10 scale34 at end exercise. In doing so, our objective was to induce a sufficiently high level of dyspnea to be amenable to therapy. If subjects were not able to carry through the 3 min at 2.5 km/h or could not reach a Borg dyspnea score of at least 4 points at the quickest speed, they were excluded. The selected speed was used throughout the remainder of the study.

The ESWT was performed on the same course as the ISWT. The initial walking speed was set at a speed corresponding to 80% of oxygen consumption peak, as estimated from the ISWT35 and then adjusted at a lower or faster pace if test duration was below or above the expected 4–8-min duration, respectively.33 The outcome measures were endurance time, expressed in seconds, excluding the 1.5-min warm-up period; and Borg dyspnea score at isotime that was defined as the longest exercise time that was reached during the 3 experimental conditions. During the warm-up period, FreeO2 was set in the automated mode, aiming at a target SpO2 of 92%, with the goal of having all subjects starting the endurance test at the same level of oxygenation. Depending on treatment allocation, subjects were switched to either fixed-flow O2 or automated O2 titration at the end of the warm-up period.

Statistical Analyses

Results are reported as mean ± SD, unless otherwise specified. The level of significance was set at an α = 0.05 for all analyses. Comparisons of values observed with the 3 experimental conditions were made using a linear mixed model: One fixed experimental factor was linked to 3 experimental conditions (fixed-flow O2, automated O2 titration, and automated O2 titration + high-flow nasal cannula) and where the subjects were analyzed as a random blocking factor. Two levels of comparisons were made: (1) fixed-flow O2 versus automated O2 titration and (2) automated O2 titration versus automated O2 titration + high-flow nasal cannula. Post hoc comparisons were performed using the Tukey comparison technique when the overall P value of the linear mixed model was < .05. Borg dyspnea scores at isotime were compared between experimental conditions using their mean changes from one condition to the other,36 but recognizing that dyspnea ratings are whole numbers (except for the first rating, 0.5), we also report their median value and interquartile ranges (IQRs), which were compared with the Friedman test. In addition, we tested whether the underlying pulmonary disease or the magnitude of O2 desaturation at baseline influenced the main clinical outcomes (Borg dyspnea score at the end of the 3-min CSST, endurance time, and dyspnea at isotime during the ESWT) in the statistical linear mixed model using an interaction term between pulmonary disease and conditions. The univariate normality assumption was verified with the Shapiro-Wilk tests on the error distribution from the statistical model after a Cholesky factorization. The Brown and Forsythe variation of Levene test statistic was used to verify the homogeneity of variances. Some variables were analyses using a non-parametric mixed statistical model on longitudinal data proposed by Brunner37 as the normality and variance assumptions were not fulfilled. The values were transformed by their ranks, and the statistical model proposed previously was applied with corrections for P values on the fixed factor. The results were considered significant with P values ≤ .05. The data were analyzed using the statistical package program SAS v9.4 (SAS Institute, Cary, North Carolina). The sample size was estimated based on a Borg dyspnea score difference between fixed-flow O2 and automated O2 titration of 1.0 at the end of the 3-min CSST and a SD of 1.0 with fixed-flow O2.38 We estimated that 30 completed subjects would provide a statistical power of 85% with an α of 0.05.

Results

The first study subject was seen on October 1, 2019, and the last study visit was completed on March 5, 2021. From the 72 patients who were assessed for eligibility, 30 were initially enrolled in the study, including 11 with COPD, 11 with ILD, 5 with pulmonary hypertension, and 3 with CF (Fig. 2). After completing the baseline evaluation, one subject with COPD and one with ILD declined further study participation and were, therefore, excluded from further analysis. Due to the COVID-19 pandemic shutdown and related sanitary measures, recruitment was stopped at 28 subjects, short of the expected sample size of 30. The characteristics of study participants are provided in Table 1. They had a mean age of 66 ± 12 y, including 16 male and 12 females. Two participants were on long-term oxygen therapy at a flow of 2 L/min. Subjects with COPD and CF had, on average, severe air-flow limitation and gas trapping, whereas subjects with ILD had a moderate to severe restrictive ventilatory pattern with reduced diffusion capacity. All subgroups had, on average, impaired diffusion capacity. Subjects with pulmonary hypertension had a mean pulmonary artery pressure of 38 ± 10 mm Hg. The 3 subjects with CF had severe air-flow limitation. The mean nadir SpO2 during the ISWT at baseline was < 88%, with a fall > 5% during exercise in all study participants. The study procedures were well tolerated by study participants with no adverse effects, apart from dyspnea related to the walking tests.

Table 1.

Subject Characteristics

graphic file with name DE-RESC230151T001.jpg

Open in a new tab

The 3-min CSST and the ESWT were performed at a mean walking speed of 4.82 ± 0.93 km/h and 4.70 ± 0.87 km/h, respectively. The primary outcome of Borg dyspnea score at the end of the 3-min CSST was not reduced using automated O2 titration compared to fixed-flow O2 (mean difference 0.19 [95% CI −0.58 to 0.97] point, P = .82) (Table 2). No further reduction was seen with the addition of high-flow nasal cannula to automated O2 titration. In contrast, there was a 56% increase in the endurance time during the ESWT with automated O2 titration compared to fixed-flow O2 (mean difference 298 [95% CI 205–391] s, P < .001) (Fig. 3). This was accompanied by a reduction in Borg dyspnea score at isotime (mean difference 1.05 [95% CI 0.33–1.78] point, P = .02), which occurred at 499 ± 36 s (Fig. 3). Median values and IQR for Borg dyspnea score at isotime amounted to 5.0 (IQR 4.0–6.5) versus 4.0 (IQR 3.0–5.0) for fixed-flow O2 and automated O2 titration alone, respectively (P = .01). Further, 46% (13/28) of subjects reported a ≥ 1-point reduction in Borg dyspnea score with automated O2 titration compared to fixed-flow O2, a threshold that is considered clinically important.39 No further benefits were seen for either parameter when high-flow nasal cannula was added to automated O2 titration (Table 2). As expected from the prolonged exercise duration with automated O2 titration, Borg dyspnea scores were similar at the end of ESWT for the 3 experimental conditions.

Table 2.

Effects of Fixed-Flow O2, Automated O2 Titration Alone, or With High-Flow Nasal Cannula on Main Study Outcomes for All Study Participants

graphic file with name DE-RESC230151T002.jpg

Open in a new tab

Fig. 3.

Fig. 3.

Open in a new tab

Changes in endurance time (panel A) and in isotime dyspnea (panel B) during the endurance shuttle walking test (ESWT) in all study participants and according to the specific disease subtype. Mean values and 95% CI are presented. The dash lines represent what is considered as the clinically important difference for endurance time during the ESWT (65 s) and Borg dyspnea score (1 point). ILD = interstitial lung disease; CF = cystic fibrosis.

Mean SpO2 and O2 flows values during the 3-min CSST and ESWT under fixed-flow O2 and automated O2 titration in subjects with COPD or ILD are provided in Figure 4. Because of the similarity in clinical response between automated O2 titration alone or with high-flow nasal cannula and for clarity purpose, data from this last experimental condition are not provided. As can be seen, the SpO2 curves during fixed-flow O2 and automated O2 titration were similar during the 3-min CSST. In contrast, a longer exercise duration during the ESWT allowed for a separation between the SpO2 and O2 flow signals between fixed-flow O2 and automated O2 titration.

Fig. 4.

Fig. 4.

Open in a new tab

Mean SpO2 (panel A, B, E, F) and O2 flows (panel C, D, G, H) values during the 3-min constant-speed shuttle test (CSST) and endurance shuttle walking test (ESWT) under fixed-flow O2 (blue lines) and automated O2 titration (orange line) in subjects with COPD (panel A to D) or interstitial lung disease (ILD) (panel E to H). The last data points on panels B, D, F, and H represent the mean end-exercise values for all subjects with COPD (panel B and D) or ILD (panel F and H) obtained with fixed-flow O2 or automated O2 titration. For these data points, standard error of the mean is presented (instead of SD) for clarity purposes. The SpO2 curves during fixed-flow O2 and automated O2 titration were similar during the 3-min CSST, both in COPD and ILD, in such a way that was not differentiation between the 2 O2 delivery modes during the short exercise. In contrast, a longer exercise duration during the ESWT allowed for a differentiation in SpO2 and O2 flow signals between fixed-flow O2 and automated O2 titration.

The oxygenation data for the 3-min CSST and the ESWT are presented in Tables 3 and 4, respectively. Mean O2 flows were slightly higher with automated O2 titration alone or with high-flow nasal cannula compared to fixed-flow O2 during the 3-min CSST. However, besides a small improvement in the minimum SpO2 during the 3-min CSST, higher O2 flows with automated O2 titration did not translate into better oxygenation status compared to fixed-flow O2 (Table 3). Conversely, mean O2 flows during the ESWT were significantly increased with automated O2 titration alone (6.9 ± 2.9 L/min) or with high-flow nasal cannula (7.6 ± 3.8 L/min) compared to fixed-flow O2 (2.3 ± 0.6 L/min, P < .001). Consequently, mean SpO2 was significantly higher with automated O2 titration alone compared to fixed-flow O2 (91 ± 2% vs 87 ± 3%, P = .001). There was no further improvement in oxygenation when high-flow nasal cannula was added to automated O2 titration. The proportion of time spent in the 92–96% target zone was 3-fold higher when automated O2 titration was used compared to fixed-flow O2 (Table 3, P < .001). In a reciprocal way, the time spent with SpO2 < 90% or < 85% was markedly reduced with the use of automated O2 titration. There was no further improvement when high-flow nasal cannula was added to automated O2 titration. Mean capillary PCO2 was 40 ± 7 mm Hg, 42 ± 8 mm Hg, and 41 ± 7 mm Hg at the end of the ESWT during fixed-flow O2, automated O2 titration, or high-flow nasal cannula, respectively.

Table 3.

Oxygenation Data During 3-Min Constant-Speed Shuttle Test for All Study Participants

graphic file with name DE-RESC230151T003.jpg

Open in a new tab

Table 4.

Oxygenation Data During Endurance Shuttle Walking Test for All Study Participants

graphic file with name DE-RESC230151T004.jpg

Open in a new tab

We explored whether the efficacy of automated O2 delivery was modulated by the underlying respiratory diseases. Considering that the study was not designed nor powered to make within- or between-disease category comparisons, data interpretation is limited to qualitative and numerical statements. Results for the main study outcomes and for the oxygenation data during the 3-min CSST and ESWT for each pulmonary disease category are presented in the data supplement (Tables S1–S12, see related supplementary materials at http://www.rcjournal.com). In general, it was found that the magnitude of improvement in endurance time during the ESWT with automated O2 titration was numerically larger (172%) in subjects with ILD as compared with the other diseases (118–157%) (Fig. 3). Consistent with this, the proportion of time spent with SpO2 < 85% was large in subjects with ILD (47%) despite the use of fixed-flow O2, giving room for improvement with automated O2 titration. The corresponding figure in subjects with COPD, pulmonary hypertension, and CF was 0, 39, and 22%, respectively.

According to the linear mixed model, the type of respiratory disease was not statistically associated with the changes in Borg dyspnea score at the end of the 3-min CSST, endurance time, and dyspnea at isotime during the ESWT with automated O2 titration. The SpO2 nadir during the baseline ESWT was found to significantly correlate with changes in dyspnea at isotime with automated O2 titration (P = .03) but not with those of endurance time during ESWT (P = .53) with automated O2 titration.

Discussion

The main study findings were that automated O2 titration did not reduce exertional dyspnea during the 3-min CSST although it did improve exercise endurance and dyspnea during the ESWT in comparison to fixed-flow O2 in subjects with a variety of chronic respiratory diseases featuring exercise-induced O2 desaturation. These benefits were associated with higher O2 flows and better preserved SpO2 during ESWT when automated O2 titration was used. The use of high-flow nasal cannula in conjunction with automated O2 titration did not provide further benefits beyond those obtained with automated O2 titration alone. Subjects experiencing more profound O2 desaturation derived the most benefit from the use of automated O2 delivery.

In the present study, dyspnea was selected as the primary outcome because it is the most prominent symptom in chronic respiratory diseases. We recently developed and validated the 3-min CSST with the specific purpose of quantifying dyspnea on exertion and evaluating response to therapy.28,32 During this test, participants walk at a fixed and externally imposed cadence for 3 min at a speed sufficient to produce exertional dyspnea that is amenable to therapy. The short exercise duration means that a dyspnea score can be obtained in most participants in a clinical trial, avoiding the need for data interpolation when dyspnea scores are missing at certain time points. This exercise test is responsive to therapy as shown by the reduction in dyspnea with bronchodilation in subjects with COPD.29 Despite these features, we were unable to demonstrate further reduction in dyspnea at the end of the 3-min CSST with automated O2 titration in comparison to fixed-flow O2. As illustrated in Figure 4, the SpO2 curves during fixed-flow O2 and automated O2 titration were similar during the 3-min CSST in such a way that differentiation between the 2 O2 administration modes during the short exercise period did not occur. One important difference between bronchodilation and oxygen is that the physiological benefits of the former, with improved expiratory flows and reduced hyperinflation, are already present at the start of the test,40 whereas those of oxygen require some time to occur as exercise intensity and O2 desaturation progress. The response time of the FreeO2 device was likely too long to adapt to a dynamic situation such as a 3-min exercise where the requirements for oxygen are rapidly changing. A further issue is the fact that O2 desaturation was relatively mild during the 3-min CSST, making it less likely for automated O2 titration to show its benefits. One last potential explanation is that automated O2 titration was not compared to room air breathing but to a potentially effective dose of oxygen, therefore reducing the magnitude of any potential effect.

Conversely, the ESWT appeared more appropriate to feature potential benefits of automated O2 titration on dyspnea and exercise tolerance. With a mean duration of 9 min, the exercise period was sufficiently long for the SpO2 curves to show distinct patterns, with clear separation between fixed-flow O2 and automated O2 titration. Consistent with previous reports in COPD,19–21 there was a clear oxygenation benefit of automated O2 titration during the ESWT that translated into prolonged exercise duration and reduced dyspnea in subjects with various chronic respiratory diseases. The ability to improve oxygenation with automated O2 titration was likely critical to enhance exercise performance considering that better oxygenation is associated with reduced ventilatory drive, less dynamic hyperinflation (in COPD), and dyspnea alleviation, collectively leading to prolonged exercise duration.5,41,42 The 298 s mean gain in exercise endurance with automated O2 titration was considerable when viewed in the context of other therapies where improvements ranging from 125–300 s are reported for bronchodilation43 and pulmonary rehabilitation.36 Benefit of automated O2 titration on exercise endurance and dyspnea at isotime also surpassed what is considered clinically important for these outcomes.38,44

One original feature of the present study is the inclusion of subjects with diseases other than COPD. As others, we observed profound O2 desaturation in subjects with ILD that was numerically greater than in COPD.12 Despite the use of O2 at 2.0 L/min, the mean nadir SpO2 was 78%, and O2 flows of up to 15 L/min were provided during automated O2 titration to maintain SpO2 in the target zone during the ESWT in subjects with ILD (Table S10). With fixed-flow O2, most of the walking period was spent with SpO2 < 90% and half of the time with SpO2 < 85%. We studied a small number of subjects who may not be representative of all subjects with ILD, but still these observations are not unprecedented. In 134 subjects with ILD not selected on the basis of O2 desaturation during exercise, Du Plessis et al12 reported an SpO2 nadir < 88% in approximately half of them, and similar results were reported from Japanese investigators.45 This was measured during the 6-min walking test, which is somewhat less intense than the ESWT. Considering the inability of fixed-flow O2 to correct O2 desaturation in a satisfactory manner, this questions the current practice with ambulatory oxygen in chronic respiratory diseases where the capacity of ambulatory O2 equipment is often surpassed by the requirement for high O2 flows to correct the problem.7,46 Indeed, current ambulatory oxygen has shown little to no benefit in subjects with chronic lung diseases.9,10 In several subjects exhibiting profound O2 desaturation during exercise, low O2 flows are unlikely to sufficiently correct oxygenation to make a large difference in the practice of physical activities during daily life. Consequently, the time spent below the target oxygenation was considerable when fixed-flow O2 was used. The FreeO2 device is not currently designed for use in daily activities, but technical development could make it possible to miniaturize the system and to use it in conjunction with a portable O2 concentrator, thus optimizing the odds of having a true clinical impact with ambulatory oxygen.

Another unique aspect of our trial compared to previous work in the field19–21 is the inclusion of a third study arm in which automated O2 titration was coupled with high-flow nasal cannula. We expected to see further benefits in terms of dyspnea and exercise endurance when high-flow nasal cannula was coupled with automated O2 titration, particularly in COPD where the small external PEEP effect may help counterbalance the adverse consequences of dynamic hyperinflation and dyspnea, but also in ILD.24 In a recent crossover trial, it was found that high-flow nasal cannula was more effective than fixed-flow O2 to improve 6-min walk distance in subjects with COPD despite no change in dyspnea or SpO2 compared to fixed-flow O2 delivered with a Venturi mask.25 Similar results were documented using a constant work-rate cycling exercise test.22 In these studies, flows of 55–60 L/min were provided with the high-flow nasal cannula compared to 40 L/min in our study. Somewhat lower flows were used in the present study to obtain the best compromise between subjects' tolerance to high flows and the ability to maintain SpO2 in the desired target zone considering that O2 flows > 20 L/min cannot be delivered with the automated O2 titration system we used. It may well be that by delivering only 40 L/min of high-flow nasal cannula did not allow to optimize the effects of this therapy.

Direct comparisons of our results with previous works19–21 is difficult because of the inclusion of subjects with diseases other than COPD and considering differences in methodological choices across studies. For example, in one study, fixed-flow O2 was adjusted to ensure a resting PaO2 ≥ 60 mm Hg.20 In another trial, O2 flows were individually titrated to secure PO2 ≥ 90% during a 6-min walking test.21 Although more personalized, this latter method is also more cumbersome to implement in daily practice. As previously done,19 we used a pragmatic approach of increasing resting O2 flows by 1 L/min. Another methodological decision regard the SpO2 level to be targeted with automated O2 titration, which varied from SpO2 90–94% across studies.19–21 Here, an SpO2 value of 94% was targeted, considering the dose-dependent effects of oxygen on exercise capacity.41 Despite this, PCO2 remained in the normocapnic range at the end of exercise, which is reassuring considering the need for relatively high O2 flows to maintain oxygenation in the preset target zone. These methodological differences across studies have to be considered with the perspective that mean O2 flows with fixed-flow O2 were not terribly different across studies, ranging from 2–3 L/min. Also, the mean improvement in walking endurance time that we saw with automated O2 titration in subjects with COPD (267 s) was in the reported range (125–325 s).19–21

Potential limitations of our study should be considered. We decided to study the effects of automated O2 delivery in various pulmonary diseases. We took the view of clinicians who would want to know if automated O2 delivery is appropriate in subjects with exercise-induced O2 desaturation, irrespective of the underlying disease. The study was not designed, however, to make within- or between-disease comparisons of the efficacy of automated O2 titration. In this context, the possibility of a larger effect of automated O2 titration in specific disease subtype cannot be excluded, which should be assessed in a dedicated trial. Nonetheless, exploratory analyses suggested that automated O2 delivery was the most effective way to deliver oxygen supplementation in all studied diseases, and this was particularly true for subjects who experienced severe O2 desaturation during exercise whatever the underlying disease. We conducted statistical analyses on key secondary outcomes such as endurance time and dyspnea during the ESWT despite the fact that the primary outcome failed to reach statistical significance. As such, the present results should not be viewed as definitive, and further studies would be useful to confirm the capacity of automated O2 titration to improve walking endurance in patients with various pulmonary diseases that are characterized by O2 desaturation during exercise. The automated O2 titration device that we used did not allow immediate correction of exercise-induced desaturation. As seen from inspection of Figure 4, differentiation of the SpO2 curves between automated O2 titration and fixed O2 was obvious at minute 3, but the time response of the device to bring SpO2 in the target zone was about 7–10 min. We speculate that an algorithm with a faster response time specifically designed for use during exercise would result in even larger effect size. Considering the nature of the study interventions, it was not possible to blind subjects and research staff to them. Despite strategies used to minimize potential biases, we cannot rule that issues with study blinding might have somewhat influenced our results. Finally, since documenting the physiological mechanisms of improvement was beyond the scope of the present study, the mechanisms responsible for increased endurance and decreased dyspnea during the ESWT remain speculative.

Conclusions

The present study provides evidence that automated O2 titration allowed for better maintenance of oxygenation and improved endurance capacity and dyspnea during the ESWT in subjects with a variety of chronic respiratory diseases in comparison to fixed O2 flow. No further benefits were seen when high-flow nasal cannula was coupled to automated O2 titration.

Acknowledgments

The authors acknowledge the assistance of Cynthia Brouillard, Mickaël Martin, Sophie Tanguay, and Andrée-Anne Therrien in conducting the experimental visits and of Serge Simard for the statistical analyses.

Footnotes

Drs Dion, Bilodeau, Lellouche, and Maltais are shareholders of OxyNov, the maker of FreeO2. Dr Lellouche is co-founder of OxyNov. The remaining authors have disclosed no conflicts of interest.

The study was funded by Fondation de l'Institut Universitaire de Cardiologie et de Pneumologie de Québec, fonds général et fonds Alphonse L'Espérance.

Supplementary material related to this paper is available at http://www.rcjournal.com.

See the Related Editorial on Page 149

REFERENCES

  • 1. Müllerová H, Lu C, Li H, Tabberer M. Prevalence and burden of breathlessness in patients with chronic obstructive pulmonary disease managed in primary care. PLoS One 2014;9(1):e85540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Bonini M, Fiorenzano G. Exertional dyspnea in interstitial lung diseases: the clinical utility of cardiopulmonary exercise testing. Eur Respir Rev 2017;26(143):160099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Dumitrescu D, Sitbon O, Weatherald J, Howard LS. Exertional dyspnea in pulmonary arterial hypertension. Eur Respir Rev 2017;26(145):170039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Stenekes SJ, Hughes A, Gregoire MC, Frager G, Robinson WM, McGrath PJ. Frequency and self-management of pain, dyspnea, and cough in cystic fibrosis. J Pain Symptom Manage 2009;38(6):837–848. [DOI] [PubMed] [Google Scholar]
  • 5. O'Donnell DE, D'Arsigny C, Webb KA. Effects of hyperoxia on ventilatory limitation during exercise in advanced chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;163(4):892–898. [DOI] [PubMed] [Google Scholar]
  • 6. Bradley JM, Lasserson T, Elborn S, Macmahon J, O'Neill B. A systematic review of randomized controlled trials examining the short-term benefit of ambulatory oxygen in COPD. Chest 2007;131(1):278–285. [DOI] [PubMed] [Google Scholar]
  • 7. Schaeffer MR, Ryerson CJ, Ramsook AH, Molgat-Seon Y, Wilkie SS, Dhillon SS, et al. Effects of hyperoxia on dyspnea and exercise endurance in fibrotic interstitial lung disease. Eur Respir J 2017;49(5):1602494. [DOI] [PubMed] [Google Scholar]
  • 8. Ulrich S, Schneider SR, Bloch KE. Effect of hypoxia and hyperoxia on exercise performance in healthy individuals and in patients with pulmonary hypertension: a systematic review. J Appl Physiol (1985) 2017;123(6):1657–1670. [DOI] [PubMed] [Google Scholar]
  • 9. Lacasse Y, Lecours R, Pelletier C, Begin R, Maltais F. Randomized trial of ambulatory oxygen in oxygen-dependent COPD. Eur Respir J 2005;25(6):1032–1038. [DOI] [PubMed] [Google Scholar]
  • 10. Visca D, Mori L, Tsipouri V, Fleming S, Firouzi A, Bonini M, et al. Effect of ambulatory oxygen on quality of life for patients with fibrotic lung disease (AMBOX): a prospective, open-label, mixed-method, crossover randomized controlled trial. Lancet Respir Med 2018;6(10):759–770. [DOI] [PubMed] [Google Scholar]
  • 11. Casaburi R, Porszasz J, Hecht A, Tiep B, Albert RK, Anthonisen NR, et al. ; COPD Clinical Research Network. Influence of lightweight ambulatory oxygen on oxygen use and activity patterns of COPD patients receiving long-term oxygen therapy. COPD 2012;9(1):3–11. [DOI] [PubMed] [Google Scholar]
  • 12. Du Plessis JP, Fernandes S, Jamal R, Camp P, Johannson K, Schaeffer M, et al. Exertional hypoxemia is more severe in fibrotic interstitial lung disease than in COPD. Respirology 2018;23(4):392–398. [DOI] [PubMed] [Google Scholar]
  • 13. Visca D, Montgomery A, de Lauretis A, Sestini P, Soteriou H, Maher TM, et al. Ambulatory oxygen in interstitial lung disease. Eur Respir J 2011;38(4):987–990. [DOI] [PubMed] [Google Scholar]
  • 14. Martí S, Pajares V, Morante F, Ramón M-A, Lara J, Ferrer J, Güell M-R. Are oxygen-conserving devices effective for correcting exercise hypoxemia? Respir Care 2013;58(10):1606–1613. [DOI] [PubMed] [Google Scholar]
  • 15. Lellouche F, L'Her E, Bouchard PA, Brouillard C, Maltais F. Automatic oxygen titration during walking in subjects with COPD: a randomized crossover controlled study. Respir Care 2016;61(11):1456–1464. [DOI] [PubMed] [Google Scholar]
  • 16. Lellouche F, L'Her E. Automated oxygen flow titration to maintain constant oxygenation. Respir Care 2012;57(8):1254–1262. [DOI] [PubMed] [Google Scholar]
  • 17. L'Her E, Dias P, Gouillou M, Riou A, Souquiere L, Paleiron N, et al. Automatic versus manual oxygen administration in the emergency department. Eur Respir J 2017;50(1):1602552. [DOI] [PubMed] [Google Scholar]
  • 18. L'Her E, Jaber S, Verzilli D, Jacob C, Huiban B, Futier E, et al. Automated closed-loop versus standard manual oxygen administration after major abdominal or thoracic surgery: an international multi-center randomized controlled study. Eur Respir J 2021;57(1):2000182. [DOI] [PubMed] [Google Scholar]
  • 19. Vivodtzev I, L'Her E, Vottero G, Yankoff C, Tamisier R, Maltais F, et al. Automated O2 titration improves exercise capacity in patients with hypercapnic chronic obstructive pulmonary disease. Thorax 2019;74(3):298–301. [DOI] [PubMed] [Google Scholar]
  • 20. Kofod LM, Westerdahl E, Kristensen MT, Brocki BC, Ringbæk T, Hansen EF. Effect of automated oxygen titration during walking on dyspnea and endurance in chronic hypoxemic patients with COPD: a randomized crossover trial. JCM 2021;10(21):4820. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Schneeberger T, Jarosch I, Leitl D, Gloeckl R, Hitzl W, Dennis CJ, et al. Automatic oxygen titration versus constant oxygen flow rates during walking in COPD: a randomized controlled, double-blind, crossover trial. Thorax 2023;78(4):326–334. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Cirio S, Piran M, Vitacca M, Piaggi G, Ceriana P, Prazzoli M, et al. Effects of heated and humidified high-flow gases during high-intensity constant-load exercise on severe COPD patients with ventilatory limitation. Respir Med 2016;118:128–132. [DOI] [PubMed] [Google Scholar]
  • 23. Bitos K, Furian M, Mayer L, Schneider SR, Buenzli S, Mademilov MZ, et al. Effect of high-flow oxygen on exercise performance in COPD patients. Randomized trial. Front Med (Lausanne) 2020;7:595450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Badenes-Bonet D, Cejudo P, Rodo-Pin A, Martin-Ontiyuelo C, Chalela R, Rodriguez-Portal JA, et al. Impact of high-flow oxygen therapy during exercise in idiopathic pulmonary fibrosis: a pilot crossover clinical trial. BMC Pulm Med 2021;21(1):355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Carlucci A, Rossi V, Cirio S, Piran M, Bettinelli G, Fusar Poli B, et al. Portable high-flow nasal oxygen during walking in patients with severe chronic obstructive pulmonary disease: a randomized controlled trial. Respiration 2021;100(12):1158–1164. [DOI] [PubMed] [Google Scholar]
  • 26. Dysart K, Miller TL, Wolfson MR, Shaffer TH. Research in high-flow therapy: mechanisms of action. Respir Med 2009;103(10):1400–1405. [DOI] [PubMed] [Google Scholar]
  • 27. Spoletini G, Alotaibi M, Blasi F, Hill NS. Heated humidified high-flow nasal oxygen in adults: mechanisms of action and clinical implications. Chest 2015;148(1):253–261. [DOI] [PubMed] [Google Scholar]
  • 28. Sava F, Perrault H, Brouillard C, Darauay C, Hamilton A, Bourbeau J, Maltais F. Detecting improvements in dyspnea in COPD using a three-minute constant-rate shuttle walking protocol. COPD 2012;9(4):395–400. [DOI] [PubMed] [Google Scholar]
  • 29. Maltais F, Aumann JL, Kirsten AM, Nadreau E, Macesic H, Jin X, et al. Dual bronchodilation with tiotropium/olodaterol further reduces activity-related breathlessness versus tiotropium alone in COPD. Eur Respir J 2019;53(3):1802049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Maltais F, Mahler DA, Pepin V, Nadreau E, Crater GD, Morris AN, et al. Effect of fluticasone propionate/salmeterol plus tiotropium versus tiotropium on walking endurance in COPD. Eur Respir J 2013;42(2):539–541. [DOI] [PubMed] [Google Scholar]
  • 31. Maltais F, O'Donnell D, Galdiz Iturri JB, Kirsten AM, Singh D, Hamilton A, et al. Effect of 12 weeks of once-daily tiotropium/olodaterol on exercise endurance during constant work-rate cycling and endurance shuttle walking in chronic obstructive pulmonary disease. Ther Adv Respir Dis 2018;12:1753465818755091. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Perrault H, Baril J, Henophy S, Rycroft A, Bourbeau J, Maltais F. Paced-walk and step tests to assess exertional dyspnea in COPD. COPD 2009;6(5):330–339. [DOI] [PubMed] [Google Scholar]
  • 33. Degani-Costa LH, O'Donnell DE, Webb K, Aranda LC, Carlstron JP, Cesar TDS, et al. A simplified approach to select exercise endurance intensity for interventional studies in COPD. COPD 2018;15(2):139–147. [DOI] [PubMed] [Google Scholar]
  • 34. Borg G. Psychophysical bases of perceived exertion. Med Sci Sports Exer 1982;14:377–381. [PubMed] [Google Scholar]
  • 35. Revill SM, Morgan MD, Singh SJ, Williams J, Hardman AE. The endurance shuttle walk: a new field test for the assessment of endurance capacity in chronic obstructive pulmonary disease. Thorax 1999;54(3):213–222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Puente-Maestu L, Palange P, Casaburi R, Laveneziana P, Maltais F, Neder JA, et al. Use of exercise testing in the evaluation of interventional efficacy: an official ERS statement. Eur Respir J 2016;47(2):429–460. [DOI] [PubMed] [Google Scholar]
  • 37. Brunner E, Domhof S, Langer F. Non-parametric analysis of longitudinal data in factorial experiments. New York: Wiley; 2002. [Database] [Google Scholar]
  • 38. Jones PW, Beeh KM, Chapman KR, Decramer M, Mahler DA, Wedzicha JA. Minimal clinically important differences in pharmacological trials. Am J Respir Crit Care Med 2014;189(3):250–255. [DOI] [PubMed] [Google Scholar]
  • 39. Ries AL. Minimally clinically important difference for the UCSD shortness of breath questionnaire, Borg scale, and visual analog scale. COPD 2005;2(1):105–110. [DOI] [PubMed] [Google Scholar]
  • 40. O'Donnell DE, Fluge T, Gerken F, Hamilton A, Webb K, Aguilaniu B, et al. Effects of tiotropium on lung hyperinflation, dyspnea, and exercise tolerance in COPD. Eur Respir J 2004;23(6):832–840. [DOI] [PubMed] [Google Scholar]
  • 41. Somfay A, Porszasz J, Lee SM, Casaburi R. Dose-response effect of oxygen on hyperinflation and exercise endurance in nonhypoxaemic COPD patients. Eur Respir J 2001;18(1):77–84. [DOI] [PubMed] [Google Scholar]
  • 42. Johannson KA, Pendharkar SR, Mathison K, Fell CD, Guenette JA, Kalluri M, et al. Supplemental oxygen in interstitial lung disease: an art in need of science. Ann Am Thorac Soc 2017;14(9):1373–1377. [DOI] [PubMed] [Google Scholar]
  • 43. Bédard ME, Brouillard C, Pepin V, Provencher S, Milot J, Lacasse Y, et al. Tiotropium improves walking endurance in COPD. Eur Respir J 2012;39(2):265–271. [DOI] [PubMed] [Google Scholar]
  • 44. Pepin V, Laviolette L, Brouillard C, Sewell L, Singh SJ, Revill SM, et al. Significance of changes in endurance shuttle walking performance. Thorax 2011;66(2):115–120. [DOI] [PubMed] [Google Scholar]
  • 45. Nishiyama O, Miyajima H, Fukai Y, Yamazaki R, Satoh R, Yamagata T, et al. Effect of ambulatory oxygen on exertional dyspnea in IPF patients without resting hypoxemia. Respir Med 2013;107(8):1241–1246. [DOI] [PubMed] [Google Scholar]
  • 46. Jacobs SS, Krishnan JA, Lederer DJ, Ghazipura M, Hossain T, Tan AM, et al. Home oxygen therapy for adults with chronic lung disease. An official American Thoracic Society clinical practice guideline. Am J Respir Crit Care Med 2020;202(10):e121–e141. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Respiratory Care are provided here courtesy of Mary Ann Liebert, Inc.

RESOURCES