Optimised oxygenation improves functional capacity during daily activities in patients with COPD on long-term oxygen therapy: a randomised crossover trial

Abstract

Background

Minimising hypoxaemia during submaximal walking tests has a positive effect on exercise capacity and dyspnoea in patients with chronic obstructive pulmonary disease (COPD) on long-term oxygen therapy (LTOT). However, the impact of optimising oxygenation during everyday tasks remains unexplored. Therefore, we investigated the effects of maintaining a target saturation on activities of daily living (ADL) using automated oxygen titration compared with conventional fixed oxygen flow.

Methods

In a double-blinded, randomised crossover trial, patients with COPD on LTOT performed two GlittreADL tests to assess the functional capacity of everyday activities using (1) their fixed oxygen dose and (2) an adjusted flow from 0 to 8 L/min targeting a peripheral oxygen saturation (SpO₂) of 90–94%. A closed-loop device automatically titrated the oxygen based on information from a Bluetooth wrist pulse oximeter.

Results

31 patients (mean±SD age: 72.8±5.9 years, forced expiratory volume in 1 s of % predicted: 36.7±12.7) were included. The patients reduced the time to perform the ADL test by median (IQR) 38 (12–73) s, p<0.001, using automated titration compared with the fixed oxygen flow. The oxygen flow in the automated arm more than tripled to 5.4 (4.1–6.8) versus 1.6 (1.1–2.1) L/min (fixed) during the test, p<0.001, while the time spent within SpO₂-target was increased from 19% to 49%, p=0.002. Correspondingly, the patients experienced less dyspnoea (BorgCR10); 5 (3–7) versus 6 (4–8), p<0.001, in favour of the automated oxygen titration.

Conclusions

Improving oxygenation and extending the time spent within target saturation reduced dyspnoea and improved functional capacity in ADL in patients with COPD on LTOT.

Trial registration number

NCT05553847.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Patients with chronic obstructive pulmonary disease on long-term oxygen therapy often desaturate during physical activity, demonstrated in standardised walking tests, with intermittent hypoxaemia of varying severity depending on the intensity of the activity. Whether these episodes of hypoxaemia impact the patients’ daily living is unclear.

WHAT THIS STUDY ADDS

• The present study shows that optimising oxygenation to the patient’s immediate need could have a significant impact on their ability to participate in activities of daily living and experience less dyspnoea in doing so.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The study contributes to our understanding of the potential benefits that technological developments could offer for future home oxygen treatment. It underscores the need for additional research and development to establish a novel approach to prescribing home oxygen, emphasising optimised and individually targeted oxygen therapy for every activity throughout the day.

Introduction

Two landmark studies conducted more than 40 years ago established the benefit of long-term oxygen therapy (LTOT) in terms of increased survival in patients with chronic obstructive pulmonary disease (COPD) who suffer from hypoxaemic chronic respiratory failure.^{1 2} The aim of oxygen therapy is to provide patients with enough oxygen to bring the partial pressure of oxygen (PaO₂) above 8 kPa at rest.³ However, during physical activity and exercises, the oxygen need increases, and the patients often experience periods of hypoxaemia despite the use of LTOT.³ In a previous study by our group, the participants with COPD on LTOT spent 65% of the time during a walking test with a peripheral oxygen saturation (SpO₂) <85%, when receiving their prescribed fixed oxygen dose.⁴

The consequences of intermittent hypoxaemia seen in patients on LTOT on both patient-reported outcomes, functional capacity and mortality are poorly understood.^{3 5} Nevertheless, it is known that patients on LTOT have higher mortality, are very limited physically and are less physically active than patients without the need for LTOT.⁶ Moreover, they have difficulties participating in activities of daily living (ADL) without desaturation and hypoxaemia.⁷ Hypoxaemia triggers dyspnoea, which stands as the most debilitating symptom in the everyday life of patients with COPD,^{8 9} and the ability to engage in everyday tasks is essential to maintain an independent lifestyle.⁹ Yet, the influence of optimal oxygenation on ADL has not been investigated.

Closed-loop systems for oxygen delivery adjust the oxygen flow automatically based on input of oxygen saturation from a pulse oximeter attached to the patient’s finger. These closed-loop devices have been shown to increase the time spent within a target saturation both during hospital admissions and in walking tests.410,18 The increased time within a target saturation during walking tests translates into meaningful improvements for patients with COPD on LTOT in terms of higher exercise capacity and alleviated dyspnoea during walking.^{4 17 18} The movements during walking are uniform and repetitive, whereas during ADL, the patient engages in diverse movements and may pause for rest. Daily activities such as dressing, showering and cooking involve the use of both arms and legs, which amplifies ventilatory demands and dyspnoea compared with walking only.⁹ This variability in the activities requires different levels of oxygen consumption, which could lead to significant fluctuations in the saturation.⁷ Closed-loop devices may be able to maintain a target saturation despite variability in oxygen needs. An aimable saturation is recommended to be above 90%; however, various guidelines seldom specify an upper saturation limit.¹⁹ The typical target for patients with stable COPD hovers around 90–94%.^{2 4 17 18 20} Targeting a specific saturation during everyday tasks is relevant due to the physiological effect of preventing hypoxaemia, which could alleviate dyspnoea during activities.

We hypothesised that by using a closed-loop system with automated oxygen flow fluctuating according to the actual oxygen demand, we could optimise oxygenation, thereby enabling patients to efficiently accomplish daily tasks with reduced dyspnoea.

The purpose of this study was to examine the effect of targeting an SpO₂ of 90–94% on ADL compared with the response of the usual fixed oxygen flow in a standardised ADL-test in patients with COPD on LTOT.

Methods

Trial design and setting

This two-centre, double-blinded, randomised crossover study enrolled patients with COPD and chronic respiratory failure with resting hypoxaemia (PaO₂≤7.3 kPa). Participants were recruited from the Departments of Pulmonology at Copenhagen University Hospital, Hvidovre and Copenhagen University Hospital, Bispebjerg-Frederiksberg, Denmark in the period from November 2022 to November 2023.

Patients and recruitment

Inclusion criteria involved patients with clinically stable COPD who received LTOT according to the international criteria for home oxygen treatment,³ who were able to walk independently with or without walking aid and cognitively able to participate. Exclusion criteria were exacerbation in COPD treated with either antibiotics or prednisolone within the preceding 3 weeks or comorbidities known to impact physical functioning.

According to clinical routine, patients on LTOT received a home care visit from a nurse specialised in oxygen treatment. In connection to the nurse visit, the patients’ medical records were screened, and eligible patients were invited to participate.

Procedure

The patient’s ability to perform ADL was assessed using the GlittreADL test developed by Skumlien et al.²¹ On a single test day at the hospital, the patients conducted two ADL tests using (A) their usual fixed oxygen dose and (B) an automatically and continuously adjusted oxygen flow ranging from 0 to 8 L/min, set to achieve SpO₂ in the target of 90–94%.

ADL assessment

GlittreADL is characterised as a semi-laboratory test validated in patients with COPD.^{22 23} It measures functional capacity (the maximum ability to perform activities) and reflects functional performance in daily physical activities known to be challenging for patients with COPD.^{22 24} A cut-off point of 3½ min in the GlittreADL test has been proven effective in identifying patients with abnormal functional capacity, in terms of worse dyspnoea, poorer health status and decreased quality of life.²⁵

Prior to the first test, baseline data was collected and the patients performed a practice round for familiarisation.²³ In the GlittreADL test, the patients rose from a seated position and walked 10 m to a bookshelf, where they moved three 1 kg cartons from the top shelf at shoulder height to the middle shelf at hip height, down to the floor and then stepwise back up to the top shelf. After turning, they walked back, sat down and directly began the next lap. This was repeated for five laps. Rest was allowed during testing, but activity had to be resumed as soon as possible. Women and men carried a 2.5 and 5.0 kg backpack, respectively, throughout the test.²¹ To standardise the test, all patients used a rollator (which also carried the oxygen cylinder and the closed-loop device), and the test was conducted without the usual two steps between the chair and the bookshelf (due to the use of rollator). Immediately after the test, the patients were asked to rate their level of dyspnoea using the Borg Dyspnoea Scale CR10.²⁶

Oxygen equipment

The oxygen supply came from the hospital’s 3 L oxygen cylinder at 200 bar (equivalent to 600 L of oxygen when released). The cylinder was attached to the closed-loop device, O2matic HOT (O2matic, Herlev, Denmark) and the equipment was placed on the rollator, figure 1. The patients wore a Nonin Wrist Pulse Oximeter (Nonin Medical, USA) which sent information on heart rate and oxygen saturation through Bluetooth to the closed-loop device. In both arms, the closed-loop device delivered the oxygen and collected data every second on oxygen flow, SpO₂ and heart rate. Data was transmitted to a cloud-based platform accessible only to the investigators. All patients wore the Optiflow nasal cannula (Fisher & Paykel), which features wider prongs designed to allow higher flow rates while maintaining patient comfort, for oxygen supplementation.

Figure 1. The GlittreADL test setup was identical in both arms. The patients carried a backpack, a pulse oximeter on the wrist and used a rollator. A high-flow nasal cannula was connected to the closed-loop device, which was attached to an oxygen cylinder. The patients rose from a chair, walked 10 m to a shelf, moved three 1 kg cartons, turned, walked back and sat down. This was repeated five times in each arm.

In the control arm, the flow was kept fixed according to each patient’s medical prescription, if it was sufficient to maintain an SpO₂ >90% at rest.

In the active arm, the closed-loop device was set to aim at keeping an SpO₂ target range of 90–94%. The oxygen flow was automatically adjusted up to 8 L/min based on the SpO₂. The adjustments were done every second based on average SpO₂ for the last 15 s.

Outcomes

The primary outcome was the difference between arms in time taken to complete the GlittreADL test. Secondary outcomes included the difference in Borg dyspnoea score immediately after ending the GlittreADL test. Average oxygen flow during tests, differences in time spent within acceptable SpO₂-interval (SpO₂ 90–94%), time spent with moderate hypoxaemia (SpO₂ 85–89%) and with severe hypoxaemia (SpO₂<85%) were also assessed.

Demographics including body mass index, smoking status, spirometry and comorbidities were collected, as well as the modified Medical Research Council dyspnoea scale (MRC, range 0–4) and the COPD assessment test (CAT, range of 0–40).^{27 28}

Randomisation and blinding

The patients were randomised after the familiarisation of the test to either AB or BA. In the AB arm, the patients received the usual fixed oxygen during the GlittreADL test first, followed by the target saturation titration, and vice versa in the BA arm. The crossover design was chosen to evaluate the response of the individual patient to different oxygen saturations. The randomisation list was computer-generated and compiled for each patient in REDCap electronic data capture tools (REDCap Consortium, Nashville, USA) hosted at Capital Region of Denmark.

An independent person randomised and prepared the oxygen setup. The coding of the device to either fixed flow or automated titration was done in the cloud solution. The closed-loop device and the pulse oximeter were covered with black opaque tape to prevent visual observation. Both the assessor conducting the tests as well as the patient were blinded to the intervention. The minimum interval between test arms was 20 min to avoid carryover effect.³

Statistical analysis

The sample size was determined based on the primary outcome, time to complete the GlittreADL test. The minimal important difference (MID) was 23 s and SD expected to be 0.74 s.²⁵ Based on an alpha of 0.05 and a power of 80%, a power analysis revealed that a sample of 32 patients was needed to detect a statistical difference between arms.

MID in the BorgCR10 dyspnoea scale is one point difference in score.²⁹

Continuous variables were examined for normality and analysed with either paired t-test (in case of normality) or Wilcoxon signed-rank test (in case of non-normality). Data presented as median with IQR or mean±SD. The Spearman test was used in an additional analysis for correlation. Test for carryover effect was performed by comparing the first and the second test using the Wilcoxon signed-rank test. Potential confounding variables were examined using an independent t-test or a χ² test, as appropriate. IBM SPSS Statistics for Windows, V.28.0 was used for all statistical analyses. GraphPad Prism V.10.1.2 for Windows was used for figures.

Results

In the study period, 145 patients on LTOT were listed for nurse visits at Hvidovre Hospital. Out of these, 79 patients had a diagnosis of COPD and severe resting hypoxaemia. After screening of medical records, 39 patients were eligible for the study, and 20 of them accepted participation. Additionally, patients were referred and included from Hvidovre Hospitals Pulmonary Rehabilitation Unit and from Bispebjerg-Frederiksberg Hospital. Subsequently, 32 patients were included, as illustrated in figure 2. One patient was unable to complete the GlittreADL test due to severe dyspnoea attack, significant desaturation and the need for medical assistance, and was therefore excluded after randomisation. 31 patients (17 men, 14 women) with a mean±SD age of 72.8±5.9 years completed the tests. The patients had a resting mean SpO₂ 92.7±2.0 with a home oxygen dose of 1.6±1.0 L/min, and a high symptom burden, with mean CAT score of 17.0±6.3 and median (IQR) mMRC score of 3.0 (IQR 3–3). Additional patient characteristics are presented in table 1.

Figure 2. Flow diagram. AHH, Copenhagen University Hospital, Hvidovre; BFH, Copenhagen University Hospital, Bispebjerg-Frederiksbjerg; COPD, chronic obstructive pulmonary disease; LTOT, long-term oxygen therapy.

Table 1. Characteristics of the study patients with COPD on LTOT.

Variables, n=31
Gender, male/female, no.	17/14
Age, years	72.8±5.9
Body mass index, Kg/m2	27.6±7.1
LTOT dose, fixed dose, L/min	1.6±1.0
SpO2 at rest with LTOT, %	92.7±2.0
Borg CR10 dyspnoea at rest, 0–10	1.1±1.1
CAT score, 0–40	17.0±6.3
mMRC, 0–4, median score (IQR)	3.0 (3–3)
mMRC, 0/1/2/3/4, no.	0/1/5/20/5
GOLD classification A/B/E, no.	0/19/12
FEV1, litre	0.9±0.4
FEV1, % of predicted	36.7±12.7
FVC, litre	1.9±0.7
FVC, % of predicted	61±18
FEV1/FVC, ratio	0.48±14
Hospital admissions the last year, no.	1.48±1.5
Smoking status, no. (%)
Tobacco user	1 (3.2)
Former tobacco user	30 (96.8)
Pack years	52.9 (33.7)
Rollator dependent, no. (%)	18 (58.1)
Comorbidities, no. (%)
None	3 (9.7)
Heart failure	10 (32.3)
Ischaemic heart disease	1 (3.2)
Diabetes	4 (12.9)
Osteoarthritis (hip or knee)	9 (29)
Osteoporosis	14 (45.2)

Data are presented as mean±SD or number with percentage (%).

CAT, COPD assessment test; COPD, chronic obstructive pulmonary disease; FEV₁, forced expiratory volume in 1 s; FVC, forced vital capacity; GOLD, Global Initiative for Chronic Obstructive Lung Disease; LTOT, long-term oxygen therapy; mMRC, modified Medical Research Council Dyspnoea Scale; SpO₂, peripheral oxygen saturation.

Nine (29%) patients were unable to wear the backpack during the GlittreADL due to ankylosing spondylitis, severe osteoarthritis, clavicle/shoulder fracture or very severe dyspnoea.

The primary outcome, time to complete the GlittreADL, was significantly reduced by a median of 38 (12-73) sec, p<0.001, when using automated oxygen compared with the fixed dose, table 2. The corresponding difference in Borg dyspnoea score was 1 (0–2), p<0.001 in favour of automated oxygen titration.

Table 2. Differences in outcomes between arms.

Performance in the GlittreADL test, n=31
Variables	Fixed oxygen dose	Automated oxygen titration	Difference	P value
Time to complete GlittreADL, seconds	427 (365–569)	409 (300–502)	38 (12–73)	<0.001
Dyspnoea, Borg CR10, score from 0 to 10	6 (4–8)	5 (3–7)	1 (0–2)	<0.001
Time within target SpO2 during ADL test, %	18.7 (10.8–29.9)	49.2 (20.3–72.4)	21.4 (-3.2–52.6)	<0.001
Median SpO2 during test %	86.6 (84.5–88.4)	91.0 (86.9–92.0)	3.4 (1.4–5.6)	<0.001
Oxygen flow, L/min	1.6±1.0	5.2±1.9	3.4±1.7	<0.001
Heart rate, bpm	97.6±14.9	93.7±13.5	2.1±8.3	0.06

Primary and secondary outcome measures during the GlittreADL test, when using usual fixed oxygen dose versus automated oxygen titration based on the peripheral oxygen saturation (SpO₂). Target SpO₂: 90–94%. Data presented as median with (IQR) or mean±SD. Borg Dyspnoea Scale CR10 at the end of the test. SpO₂, oxygen flow and heart rate are presented as mean values throughout the test (n=28 in these three values).

ADL, activities of daily living; bpm, beats per minute; L/min, litre per minutes.

27 (87%) patients exceeded the MID of 23 s or 1 point score, respectively, in either time to complete the GlittreADL or the Borg dyspnoea score. 19 (61%) improved by 23 s or more, and 23 (74%) reported alleviated dyspnoea with at least 1 point. 15 patients (48%) demonstrated improvements above the MID in both outcome measures. The fastest patient completed the GlittreADL in 3 min and 25 s using automated oxygen titration.

A significantly larger proportion of time spent within target saturation was seen in the automated oxygen arm, p=0.002, figure 3. The mean SpO₂ was 92% at the beginning of both test arms. In more than 33% of the time in the ADL test, the patients experienced severe hypoxaemia with an average SpO₂<85%, when using the fixed oxygen dose. This time was significantly reduced to 17%, when adjusting the oxygen flow automatically, p=0.007. No unintended effects were noted.

Figure 3. Percentage of time spent within saturation intervals. Boxplot showing median, IQR, minimum and maximum time spent in the different oxygen saturation (SpO₂) intervals. X-axis: Four predefined oxygen saturation intervals. Y-axis: Percentage of the time taken to complete the GlittreADL test with Automated Oxygen titration (AutOx) and Fixed Oxygen dose (FixedOx). *p<0.05, **p<0.01. ADL, activities of daily living; ns, not significant.

In the analysis testing for carryover effect, no statistical difference was found between test 1 and test 2, p=0.3. There were no significant demographic differences between the AB and BA arms. As nine patients were unable to wear a backpack, we performed a sensitivity analysis including only those who used the backpack. This showed no change in the primary outcome, and the difference remained significant (p<0.001).

A post hoc analysis showed a correlation between degree of desaturation with fixed oxygen and the time difference between arms in completing the ADL test, Spearman ρ=0.69, p<0.0001, figure 4.

Figure 4. Time spent with hypoxaemia relative to the effect of optimised oxygenation. Each data point represents an individual patient, illustrating the percentage of time spent with hypoxaemia (SpO₂ <90%) in the fixed oxygen arm relative to the change in performance of the GlittreADL test between arms. The minimal important difference (MID) in GlittreADL is 23 s illustrated with the dotted line. ADL, activities of daily living.

Discussion

The present study showed that the capacity to perform ADL improved in patients with COPD on LTOT when optimising oxygenation. Correspondingly, the patients reported less dyspnoea. The use of automated oxygen titration during the ADL test increased time spent within a target saturation of 90–94% compared with the use of a fixed oxygen flow. However, it required more than three times as high an oxygen flow, which on average was 5.4 L/min in the adjusted flow arm compared with 1.6 L/min in the fixed flow arm.

GlittreADL test performance

The GlittreADL test involves various functional tasks, such as walking, sitting down, rising from a chair, crouching and moving objects. In our study, the patients took an average of 7.3 min to complete the test. We removed the steps from the original test to ensure that all patients began under the same conditions, and in order to use a rollator. It can be speculated that patients would spend less time on the test without the need to climb stairs. However, only one patient managed to complete the test in less than the cut-off point of 3½ min, which distinguishes between normal and abnormal functional capacity.³⁰ Patients who are more severely affected by dyspnoea and COPD require longer time to complete the test, resulting in lower functional capacity in daily life.^{31 32} This emphasises that our study participants were severely limited in daily activities due to their condition.

The observed improvement of 38 s, while statistically significant, is also clinically relevant as it exceeds the MID of 23 s, which was originally established following a pulmonary rehabilitation programme.²⁵ Both Gulart et al and Mendes et al used the GlittreADL test before and after an 8 week exercise training programme, reporting mean improvements of 43.8 and 31.2 s, respectively.^{33 34} Similarly, Skumlien et al observed a 53.4 s improvement after 4 weeks of rehabilitation.²¹ In contrast, Calik-Kutukcu et al found no significant change in GlittreADL time following arm strength training, though they reported an improvement in dyspnoea.³⁵ We observed both an improvement in dyspnoea and a substantial reduction in test completion time. 87% of the participating patients, in the present study, had a clinically relevant improvement in either the time to complete the ADL task or on dyspnoea. Notably, this improvement occurred not over weeks of exercise training but within the same day, with the only variable being the oxygen dose. However, it required three times as high an oxygen flow, and still the patients were within the SpO₂-target for only 47% of the time during the test.

Correlation between desaturation and improvement

The patients were encouraged to perform the test as quickly as possible, although they were still allowed to pause and ‘catch their breath’. In our previous study,⁴ the patients experienced significant desaturation during a fixed-pace walking test, which led to a large correction in hypoxaemia and consequently marked improvements in dyspnoea (a 50% reduction) and performance (a 98% increase). In contrast, the present study employed a self-paced test. This design allowed patients to adjust their effort, and the possibility to rest probably gave them some degree of recovery. The improvements in dyspnoea and performance following the reduction in hypoxaemia were less pronounced in the present study.

Our findings are further supported by the observed correlation between the time spent in hypoxaemia during the test while using fixed oxygen and the improvement in time to complete the test when oxygen flow was optimised (figure 4). In other words, the more the patients desaturated, the more they profited from the adjusted oxygen flow. Similar findings were reported by O’Donnell et al, who found that the greater the intensity of dyspnoea during exercise while breathing room air, the greater the reduction on dyspnoea intensity when exercising with oxygen.³⁶ Overall, these results strengthen the hypothesis that reducing hypoxaemia leads to an increase in ADL performance.

Correcting hypoxaemia

Patients requiring LTOT experience desaturation both during exercise and daily activities. Even patients with milder COPD may approach their maximal ventilation (up to 80% of peak) during daily tasks.³⁷ Research has demonstrated that a high concentration of oxygen can reduce ventilatory demands,^{36 38} improve muscle function³⁹ and leg fatigue.³⁶ In short, enough available oxygen increases the patients’ capacity to move. Lessening the ventilatory burden could be a physiological explanation that may underpin the improvements observed in the present study, with the correction on hypoxaemia being the key factor.

Currently, the primary rationale for providing patients on LTOT with portable oxygen devices for ambulatory oxygen is the possibility to prolong the time spent with LTOT.^{3 5} Lacasse et al showed in 2005 that patients were not more physically active when receiving ambulatory oxygen versus no ambulatory oxygen.⁴⁰ However, the saturation levels of the participating patients were not assessed. It could be argued that if the patients desaturated while moving outdoors, they were still very limited by hypoxaemia, thus lacking capacity to move. Garrod et al concluded in a study that oxygen during training improved dyspnoea but did not further enhance overall exercise tolerance.⁴¹ They noted, however, that the patients, despite 4 L/min, still desaturated during exercise, suggesting that persistent hypoxaemia may still have limited the patients’ performance.⁴¹ Automated oxygen administration offers the opportunity to continuously adjust oxygen flow to the patient immediate needs and thereby reduce episodes of desaturation.

Clinical implication

A standardised ADL test provides an objective and reliable assessment of patients’ ability to perform ADL. This minimises the risk of patients’ own interpretation of the effects and gives us insight into the implications of correcting hypoxaemia with measurable, comparable quantitative outcomes. Our findings offer initial evidence that patients can perform daily activities more easily when their oxygen delivery is adjusted to meet their immediate needs. Additionally, the current study preceded another investigation in which the setup was implemented in patients’ homes to assess the feasibility of automated oxygen therapy in daily life.⁴² The setup in the home-based study was similar to the present study, using a wrist pulse oximeter for continuous monitoring and oxygen flow adjustments of up to 8 L/min. The only difference was that the oxygen source was a stationary concentrator instead of an oxygen cylinder, making it a part of patients’ normal life with LTOT. The improvement in the ADL seen in the present study could potentially impact patients’ daily lives by facilitating activities such as showering, cooking, being mobile and engaging socially.

Strengths and limitations

A strength of our study was the crossover design with both tests conducted on the same day, which ensured that the patients were in similar health status and that the results reflected the individual response to the optimisation of the oxygen level. A further strength was the blinding of the patients with the closed-loop device used for both fixed flow and variable flow. Also, the use of the Optiflow nasal cannula, typically used in an inpatient hospital setting, minimised the risk of the patients noticing an increased flow. None of the patients reported any burning sensation in the nose as experienced in our earlier study.⁴ However, we acknowledge that we did not formally assess blinding by directly surveying the patients.

We chose to modify the GlittreADL test to ensure inclusion of patients in need of a walking aid, which was 58% of the patients (table 1). We did not find any studies describing the use of the GlittreADL without steps, leading us to conclude that all studies have involved patients capable of walking without a rollator and climbing stairs without assistance. This modification of the otherwise validated test represents a clear limitation of our study. Nevertheless, since the patients served as their own controls, we maintain confidence in the results. Also, some patients were unable to carry the backpack during the test, mainly due to severe comorbidity. However, in this case, a study showed validity and comparable SpO₂ responses between GlittreADL with and without backpack.³³ Our study involved a relatively small sample size of 31 patients, which could limit the generalisability of the findings. Lastly, the closed-loop device and Bluetooth pulse oximeter may have technical limitations or inaccuracies that could affect the reliability of the measurements.

In conclusion, the functional capacity in ADL in patients with COPD on LTOT improved when improving oxygenation and extending the time spent within target saturation. Concurrently, the patients reported less dyspnoea with enhanced oxygenation. Future studies should focus on the possibilities and benefits of improving oxygenation and maintaining a target saturation when moving outdoors and doing activities at home.

Data availability statement

Data are available upon reasonable request.