Long-Term Oxygen Therapy in COPD Patients Who Do Not Meet the Actual Recommendations

ABSTRACT

Chronic respiratory failure due to chronic obstructive pulmonary disease (COPD) is an increasing problem worldwide. Many patients with severe COPD develop hypoxemic respiratory failure during the natural progression of disease. Long-term oxygen therapy (LTOT) is a well-established supportive treatment for COPD and has been shown to improve survival in patients who develop chronic hypoxemic respiratory failure. The degree of hypoxemia is severe when partial pressure of oxygen in arterial blood (PaO₂) is ≤55 mmHg and moderate if PaO₂ is between 56 and 69 mmHg. Although current guidelines consider LTOT only in patients with severe resting hypoxemia, many COPD patients with moderate to severe disease experience moderate hypoxemia at rest or during special circumstances, such as while sleeping or exercising. The efficacy of LTOT in these patients who do not meet the actual recommendations is still a matter of debate, and extensive research is still ongoing to understand the possible benefits of LTOT for survival and/or functional outcomes such as the sensation of dyspnea, exacerbation frequency, hospitalizations, exercise capacity, and quality of life. Despite its frequent use, the administration of “palliative” oxygen does not seem to improve dyspnea except for delivery with high-flow humidified oxygen. This narrative review will focus on current evidence for the effects of LTOT in the presence of moderate hypoxemia at rest, during sleep, or during exercise in COPD.

KEYWORDS:

• Chronic obstructive pulmonary disease
• long-term oxygen
• moderate hypoxemia
• exercise
• nocturnal
• desaturation

Introduction

With recent advances both in the treatment and management of chronic obstructive pulmonary disease (COPD), mortality in patients with chronic respiratory failure has decreased over time, and end-stage disease with chronic respiratory failure has become a more evident problem (1,2). End-stage COPD patients can present with gas exchange abnormalities either with hypoxemia only or with hypoxemia accompanied by hypercapnia. Long-term oxygen therapy (LTOT) has long been the standard of care for hypoxemic COPD patients. However, domiciliary non-invasive ventilation support added to LTOT may be beneficial in hypercapnic respiratory failure when daytime arterial carbon dioxide level is ≥ 45–50 mmHg in the stable period.

Approximately 7% of patients with moderate to severe COPD develop resting hypoxemia within 5 years (3,4). The main risk factors for the development of hypoxemia in COPD are still not clear; however, male sex, living at a high altitude, increased body mass index, lower forced expiratory volume in one second (FEV₁), increased functional residual capacity (%), presence of heart failure, pulmonary artery enlargement on computerized tomography, previous history of severe COPD exacerbation, low baseline resting oxygen saturation, and increased resting heart rate have been reported as possible risk factors (3–6). Although FEV₁ is an important parameter for the assessment of disease severity, there is a weak correlation between FEV₁ and the degree of hypoxemia. (7). This finding is also supported by the observation that lung function does not predict survival in established hypoxemic patients (8). Emphysematous COPD has been reported as a risk factor for exercise-induced hypoxemia but not for resting hypoxemia (9,10). In addition, recent studies suggest that race and some genetic variations (polymorphisms in chromosomes 14 and 15) may be predictive of resting hypoxemia (11).

The degree of hypoxemia is directly related to mortality in COPD, and correcting hypoxemia with LTOT increases survival (12). Most national and international guidelines recommend the use LTOT in the presence of resting hypoxemia, which is defined as a resting partial pressure of oxygen in arterial blood (PaO₂) ≤55 mmHg (or oxygen saturation [SaO₂] ≤ 88%) or resting PaO₂ between 56 and 59 mmHg (or SaO₂ ≤ 88%) with evidence of pulmonary hypertension, cor pulmonale, or polycythemia with a hematocrit value greater than 55% (Table 1) (13–18). The current recommendation is based on the results of two historical studies published in the early 1980s: the Nocturnal Oxygen Therapy Trial (NOTT) in 1980 and the Medical Research Council (MRC) study in 1981 (19,20). Although both studies had relatively few patients when compared to current randomized controlled trials (RCTs; MRC study included 87 patients, NOTT study included 203 patients) and did not represent the ideal group for LTOT (e.g., 25–52% of patients were still^{Figure 1} smokers and all patients are aged < 70 years in the MRC study, smoking status was not declared in NOTT), the results showed that LTOT increased survival in hypoxemic patients, and this effect on mortality was directly correlated with the duration of LTOT use.

Table 1. Definitions.

Resting hypoxemia	Resting PaO2 ≤55 mmHg or SaO2 ≤88%
	OR
	Resting PaO2 between 56-59 mmHg (or SaO2 ≤88%) with evidence of either pulmonary hypertension, cor pulmonale or polycythemia (hematocrit >55%)
Moderate hypoxemia	Resting PaO2 ranging between 56 and 69 mmHg
Nocturnal desaturation	Spending ≥30% sleep time with SaO2 < 90% without evidence of associated sleep apnea
	OR
	A fall in SaO2 below 90% for 5 minutes or more with a nadir SaO2 below 85% during sleep
Exercise-induced desaturation	SaO2 ≤ 88% during exercise
	OR
	A fall in SaO2 ≥ 4% during exercise

Abbreviations:

PaO₂: partial pressure of oxygen in arterial blood, SaO₂: oxygen saturation

However, approximately 20% of COPD patients were prescribed LTOT outside published guidelines, and there is still debate whether the use of oxygen therapy is beneficial during the earlier phase of the disease before the development of resting hypoxemia, especially for symptom relief (21). Patients with moderate to severe COPD often have moderate hypoxemia (PaO₂: 56–69 mmHg) at rest, and patients who do not have hypoxemia at rest become hypoxemic during exercise or sleep (22). Although these patients are not qualified for oxygen therapy with current indications, research is still ongoing whether oxygen therapy may have potential positive effects in these patients. In this review, we touch upon the answers of some clinical questions of utmost importance regarding the effects of LTOT in moderate hypoxemia at rest and during sleep or exercise in COPD.

We searched Medline, EmBase, and the Cochrane database, using the keywords “COPD and oxygen therapy,” “COPD and moderate hypoxemia,” “COPD and nocturnal desaturation,” “COPD and exercise induced desaturation” for all reports of adults published up to January 2017 (Figure 1). Our search was not limited to publications in English. We selected relevant reports and comprehensive reviews. We also searched the reference lists of the identified publications, selecting relevant articles with an emphasis on oxygen therapy research.

Figure 1. Flow chart for articles included in the review (RCT, Randomized controlled trial).

The rationale of LTOT use in COPD

Consequences of chronic hypoxemia

The main mechanisms of hypoxemia in COPD are ventilation perfusion mismatch and emphysematous destruction of the alveolocapillary membrane. Pulmonary vasoconstriction occurs in response to alveolar hypoxia in an effort to match ventilation and perfusion (12). However, ongoing pulmonary vasoconstriction causes pulmonary remodeling and pulmonary hypertension, resulting in right heart failure. The degree of pulmonary hypertension is inversely correlated with the severity of resting hypoxemia. Pulmonary hypertension in COPD is also related to an increased hospitalization frequency and decreased survival (23–25). From a cardiac perspective, hypoxemia causes peripheral vasodilation, which induces compensatory tachycardia with a subsequent increase in cardiac output to improve oxygen delivery (24). Persistent hypoxemia may induce secondary erythrocytosis, causing polycythemia and increasing the serum viscosity. Muscle dysfunction, neurocognitive impairment, and impaired quality of life (QoL) (13,25,26). All these consequences contribute to increased mortality in COPD.

Physiological and clinical effects of oxygen therapy

If hypoxemia persists, tissue hypoxia and organ dysfunction will be inevitable. Thus, the correction of hypoxemia is important for survival. Oxygen therapy may reverse pulmonary vasoconstriction and stabilize or reduce pulmonary artery pressure (24,27,28). Improvement of pulmonary hemodynamics is enhanced if oxygen is used continuously, and this effect seems to be sustainable over the years (19,29). Oxygen administration improves breathlessness, especially during exercise in hypoxemic subjects. Moreover, exercise tolerance in the acute setting is improved with oxygen, including in patients with mild to moderate hypoxemia (30). Selinger et al. showed that positive changes induced by oxygen are reversible by the interruption of therapy (31). LTOT may improve the sleep profile and quality by preventing nocturnal hypoxemia (24). In addition, LTOT may be helpful for improving respiratory and general health-related QoL (32).

Oxygen therapy in COPD patients with moderate hypoxemia

PaO₂ has been shown to exhibit a decreasing trend that is directly related to mortality in COPD (33,34). Therefore, theoretically, oxygen supplementation to achieve higher PaO₂ levels in the presence of moderate hypoxemia may have a positive impact on disease outcomes. There are a limited number of studies assessing the effect of LTOT in moderately hypoxemic COPD patients (Table 2). Gorecka et al. randomized COPD patients with moderate hypoxemia into control (n = 67) and LTOT (n = 68) groups and followed them for at least three years or until death (35). There were no significant differences in survival rates between patients treated with LTOT and controls. Longer oxygen use (>15 hours per day) did not improve survival. Ringbaek et al. studied 170 moderately hypoxemic COPD patients before and after 10 months of initiation of home oxygen therapy. Comparing the two periods, there were no differences in admission rates or the number of patients with at least one hospitalization (36). They concluded that home oxygen therapy does not reduce hospitalization in patients with COPD without severe hypoxemia. In a trial performed in hypercapnic COPD patients with a mean PaO₂ of 67 mmHg at rest, LTOT did not provide a survival benefit. However, it halted the decline in endurance time and reduced the exertional dyspnea score significantly after 1 year compared to the controls (37).

Table 2. Randomized Controlled Trials For Long-Term Oxygen Therapy In COPD Patients wit Moderate Hypoxemia.

				Resting
	Study population	Comparison (n)	FEV1 (L and/or % predicted)	PaO2 (mmHg)	SaO2	Mean O2 use (hours/day)	Study duration	Impact on outcomes
Gorecka^35 1997	Patients with PaO2 between 56-65 mmHg	LTOT (68): Oxygen was prescribed with a flow rate adjusted to maintain a PaO2 >65 mmHg and >17 hours/day	0.85 L (29.7%)	60	NA	13.5	3 years	Primary outcome: LTOT did not prolong survival
		Control (67)	0.81 L (29.8%)	61	NA	-		Secondary outcome: There was no correlation between the duration of oxygen use and survival in LTOT group
Haidl^37 2004	COPD patients with reversible hypercapnia (PaCO2 > 45 mmHg at rest on 2 different days or increase in PaCO2 after cycle testing > 45 mmHg) with PaO2 at rest > 55 mmHg and a mean nocturnal oxygen saturation > 90%, and before discharge, PaCO2 at rest should have returned to ≤45 mm Hg)	LTOT (14): Oxygen was prescribed at 2 L/min for ≥15 hours/day	38.8%	67	NA	10.4	1 year (the planned study duration was 3 years, but due to the high drop-out rate, assessment after one year was chosen by the authors)	Primary outcome: LTOT did not decrease mortality in COPD patients with reversible hypercapnia
		Control (14)	42.7%	66	NA	-		Secondary outcomes: LTOT stopped the decline in endurance time and reduced the exertional dyspnea score significantly after 1 year compared to controls
The Long Term Oxygen Treatment Trial Group^38 2016	Patients with moderate resting hypoxemia (SpO2 89-93%) or resting SpO2 >93% and exercise-induced moderate desaturation (SpO2 ≥ 80% for ≥ 5 min and <90% for ≥10 seconds during 6 MWT)	LTOT (386): Oxygen (2 L/min) was prescribed 24 hours in patients with moderate hypoxemia (220) and during exercise and sleep in patients with exercise-induced desaturation (148)	47%	NA	93%	15.1 hours in the 24-hour group and 11.3 hours in the sleep–exercise group	1-6 years (median follow-up: 18.4 months)	Primary outcome: LTOT for patients with resting or exercise-induced moderate desaturation did not affect the time to death or first hospitalization for any cause.
		Control (370)	46%	NA	94%	1.8 hours (median 0 hours)*		Secondary outcomes: There was no difference between groups for the rates of all hospitalizations and COPD exacerbations. There was no difference between the groups in the change from baseline in measures of quality of life (assessed with quality of well being scale, SF 36, SGRQ, Pittsburgh sleep quality index), depression and anxiety (assessed with HAD scale), lung function, distance for 6MWT and other measures of functional status

 

 

Abbreviations:

6MWT: 6-minute walk test, COPD: chronic obstructive pulmonary disease, FEV₁: forced expiratory volume in one second (% predicted), L/min: liters per minute, LTOT: long-term oxygen therapy, NA: data not available, O₂: oxygen, PaCO₂: partial pressure of carbon dioxide in arterial blood, PaO₂: partial pressure of oxygen in arterial blood, SaO₂: oxygen saturation, SpO₂: oxygen saturation with pulse oximetry

* Use of supplemental oxygen limited to periods of severe resting desaturation (SpO₂ ≤88%) or severe exercise-induced desaturation (SpO₂ <80% for ≥1 minute); the oxygen requirement was reassessed after 30 days

A recent multicenter study provided important insight concerning this issue (38). This is the largest study to date examining the efficacy of LTOT in resting or exercise-induced moderate hypoxemic patients (n = 738). In the LTOT Trial Research Group study, patients were randomized to either supplemental oxygen or no supplementation. In the supplemental oxygen group, patients with moderate resting desaturation (Oxygen saturation with pulse oxymeter [SpO₂] 89–93%) were prescribed 24-hour oxygen, and those with desaturation only during exercise (during the 6-minute walk test [6MWT], SpO₂ ≥ 80% for more than 5 minutes and < 90% equal to or more than 10 seconds) were prescribed oxygen during exercise and sleep. The median follow-up time was 18 months. LTOT had no effect on time to death or time to first hospitalization. In addition, the exacerbation rate, QoL, and functional measures (depression, anxiety, and functional status) were similar between the study groups. The study also had some limitations. There is the possibility of the exclusion of patients with a greater number of symptoms because they might have declined to participate. Second, neither nocturnal desaturation nor effects of oxygen on exercise performance were assessed in the study. Finally, the estimated use was 13 hours/day in the supplemental oxygen group, and it is uncertain whether oxygen use for longer periods of time in these patients might have provided different results. This is especially important because we know that based on NOTT results, the duration of oxygen use is predictive of survival in hypoxemic COPD patients

Although theoretically initiating LTOT in the early phase of COPD, before the onset of irreversible changes, seems to be reasonable, the data reported to date have shown no benefit. Regarding current evidence, oxygen therapy is not indicated in COPD patients with moderate hypoxemia.

Oxygen therapy in COPD patients with nocturnal desaturation

Chronic obstructive pulmonary disease patients experience oxygen desaturation of clinical importance during sleep due to worsening of ventilation/perfusion mismatch and nocturnal alveolar hypoventilation (39–41). Most of these patients have near normal oxygen saturation during wakefulness, and daytime PaO₂ values may not be helpful for the discrimination of nocturnal desaturators (41–43). The prevalence of nocturnal desaturation varies widely according to the definitions used for nocturnal desaturation and the selected patient populations in the studies. Fletcher et al. defined nocturnal desaturation as a fall in SaO₂ below 90% for 5 minutes or more, reaching a nadir saturation of at least 85%, and reported that approximately one-third of patients with a PaO₂ greater than 60 mmHg, have nocturnal desaturation (44). Other studies used the definition of spending more than 30% of at least one of two nights with an oxygen saturation of less than 90% and showed that 50–70% of patients with mild to moderate hypoxemia have nocturnal desaturation (41,45). Although nocturnal oximetry is the simplest way to show nocturnal desaturations, polysomnography is the best tool for the evaluation of sleep disorder breathing problems in COPD patients, especially for the differential diagnosis of obstructive sleep apnea (46).

The impact of nocturnal desaturation on pulmonary hemodynamics is still not well understood. Levi-Valensi et al. reported that COPD patients who desaturated at night have a higher mean pulmonary artery pressure when compared to patients who did not desaturate (19 vs 17 mmHg; p < 0.05). In addition patients who developed pulmonary hypertension had higher number of desaturation dips, longer dip durations, and lower mean nocturnal arterial oxygen saturation compared with patients who desaturated but did not develop pulmonary hypertension and patients who did not desaturate (42). In contrast to this finding, Chaouat and colleagues showed that the mean pulmonary artery pressure was not correlated with the degree and duration of nocturnal desaturation in COPD patients with mild daytime hypoxemia (45).

The prognostic significance of nocturnal desaturation for mortality in COPD is not clear. Fletcher et al. studied non-hypoxemic COPD patients (daytime PaO₂ greater than 60 mmHg) with and without nocturnal desaturation and reported that survival was significantly improved in patients without nocturnal desaturation (48). They defined nocturnal desaturation based on two different definitions: episodic desaturation associated mainly with REM sleep (definition 1) and desaturation > 30% of the time spent in bed below an oxygen saturation of 90% (definition 2). When stratified for supplemental oxygen use, survival was only improved in subjects referencing definition 1. A double-blind, randomized trial evaluated the long-term effect of oxygen supplementation in COPD subjects with nocturnal desaturation (Table 3) (49). Patients treated with supplemental oxygen during sleep for 36 months showed a significant downward trend in pulmonary artery pressure (−3.7 mmHg) compared to patients treated with room air (+3.9 mmHg). In another RCT, nocturnal oxygen supplementation did not affect the evolution of pulmonary hypertension and mortality in COPD patients with nocturnal desaturation (50). Nocturnal oxygen supplementation did not prevent initiation of conventional LTOT as well.

Table 3. Randomized Controlled Trials For Nocturnal Oxygen Therapy In COPD Patients With Nocturnal Desaturation.

	Study population	Definition for nocturnal desaturation	PSG	Comparison (n: recruited/completed the study)	FEV1 (L and/or % predicted)	Resting PaO2 (mmHg)	Nocturnal SaO2	T Sa <90%	Mean O2 use (hours)	Study duration	Impact on outcomes
Fletcher^49 1992	Patients with resting daytime PaO2 ≥60 mm and nocturnal desaturation	i. Baseline awake and non-rapid eye movement (NREM) ear oximeter saturation ≥ 90%,	Yes	Nocturnal oxygen therapy 3 L/min (19/7)	1.42L	74	91%	29%	7.6	3 years	Primary outcome: Nocturnal desaturating patients who received oxygen during sleep for 3 years showed a significant downward trend in pulmonary artery pressure compared to desaturators treated with compressed air
		ii. A fall below this value for 5 min or more, usually coinciding with but not limited to REM sleep, iii. A nadir SaO2 value reaching 85% or lower		Compressed air (19/9)	1.42L	77	91%	24%	9.4		Secondary outcomes: There was no statistically significant difference between groups for mortality. and nonvascular parameters (hemoglobin, red blood cell mass).
Chaouat^50 1999	Patients with daytime PaO2 between 56-69 mmHg and nocturnal desaturation	Spending ≥30% of the recording time with transcutaneous arterial O2 sat <90%	Yes	Nocturnal oxygen therapy 2 L/min (41/24)	1.1L (41%)	63	88%	63%	8.9	2 years	Primary outcome: Nocturnal oxygen therapy did not alter the evalution of pulmonary hemodynamics
				Control (35/22)	0.98L (35%)	63	89%	65%	-		Secondary outcomes: Nocturnal oxygen therapy did not prevent initiation of conventional LTOT (number of patients who required LTOT during follow-up). Nocturnal oxygen therapy did not improve survival

Abbreviations:

COPD: chronic obstructive pulmonary disease, FEV₁: forced expiratory volume in one second (% predicted), L/min: liters per minute, LTOT: long-term oxygen therapy, O₂: oxygen, PaO2: partial pressure of oxygen in arterial blood, PSG: polysomnography, SaO₂: oxygen saturation, SpO₂: oxygen saturation with pulse oximetry, T Sa <90%: recording time spent with an oxygen saturation <90%

Nocturnal desaturation is a known factor in sleep fragmentation and the decrease in QoL in patients with COPD. However, Lewis et al. reported that patients with nocturnal desaturation did not have worsened sleep quality, QoL, and daytime somnolence when compared to patients without desaturation (41). Studies assessing the effects of the correction of desaturation with nocturnal oxygen therapy on sleep and QoL also had heterogeneous results. Nocturnal oxygen supplementation in COPD patients with isolated nocturnal hypoxemia did not improve QoL within 6 weeks (51). A recent study showed that nocturnal oxygen therapy may be beneficial for improving QoL only in emphysematous but not in chronic bronchitis type COPD (52). The authors explained this finding as a consequence of various responses to the hypoxemic ventilation stimulus in these two phenotypes.

Due to the paucity of data and the relatively small sample size of the studies mentioned above, the role of oxygen therapy in clinically relevant outcomes remains unclear. An ongoing study, the International Nocturnal Oxygen Trial (INOX), will address this question and facilitate our understanding of whether supplemental oxygen during sleep is effective (53). This study is a three-year, multicenter, placebo-controlled, randomized trial of nocturnal oxygen therapy added to usual care, and the primary outcome is a composite of all-cause mortality or the requirement for LTOT.

Oxygen therapy in COPD patients with exercise-induced desaturation

Approximately 20% of COPD patients without resting hypoxemia develop desaturation during exercise (10,54,55). Exercise-induced desaturation (EID) is defined as a fall in oxygen saturation to 88% or lower during exercise (Table 1). EID is considered an important risk factor for pulmonary hypertension, rapid decline in lung function, and recurrent exacerbations (54–57). It has been reported that patients with low oxygen saturation at rest, emphysematous phenotype, and low diffusion capacity are more likely to desaturate during exercise (10,58). EID diagnosis is usually confirmed by the 6MWT, and it has been shown that desaturation during the 6MWT can predict a future need for LTOT (59). In addition, patients who desaturate during the 6MWT have been reported to have an approximately twofold increased risk of death compared with those who did not desaturate (55,60).

Oxygen supplementation during exercise results in significant improvements in exercise tolerance both in healthy people and in COPD patients. Oxygen administration causes less breathlessness and improves exercise capacity by decreasing minute ventilation and dynamic hyperinflation, improving pulmonary hemodynamics by alleviating hypoxic vasoconstriction and increasing cardiac output and oxygen delivery in the acute setting (24). It has been shown that supplemental oxygen protects against desaturation during exercise in COPD patients without severe resting hypoxemia (61). Moreover, oxygen supplementation improves exercise duration, exercise endurance, and symptom perception in non-hypoxemic patients (30). The likely underlying mechanism is a reduction of the hypoxemia-related increase in ventilatory demand, dynamic hyperinflation, and increased exertional breathlessness (25,30).

However, evidence for the long-term use of oxygen to prevent EID during exercise in non-hypoxemic COPD patients outside the laboratory setting is limited. Moore et al. showed that ambulatory oxygen had no benefit in terms of exercise capacity (62). Similar to this finding, the LTOT trial subgroup analysis showed that patients with an exercise desaturation profile (43% of patients) experienced no benefit from ambulatory oxygen supplementation with respect not only to mortality and hospitalizations but also to nadir oxygen saturation during exercise, distance walked in the 6MWT, functional status, and other measures of QoL (38). In a meta-analysis, it was shown that the clinical utility and effectiveness of ambulatory oxygen in improving mortality and exercise capacity were not evident in patients who were not hypoxemic at rest (63).

We can conclude that oxygen supplementation during exercise may prevent desaturation and improve exercise capacity and symptoms in the acute setting in patients who are not eligible for LTOT; however, long-term domiciliary oxygen use for EID has no effect on important outcomes such as mortality, functional capacity, and QoL.

Oxygen therapy during pulmonary rehabilitation and exercise training in COPD patients

It is well established that pulmonary rehabilitation (PR) and physical training enhance functionality and QoL in patients with COPD. It also has positive impacts such as improvements in exercise capacity and breathlessness. However, as most of these patients have limited exercise capacity because of desaturation episodes during training, supplemental oxygen during training would be beneficial. Voduc et al. evaluated the effect of supplemental oxygen during PR and showed that exercise duration increased with oxygen (64). Neunhauserer et al. showed that supplemental oxygen in non-hypoxemic COPD doubled the effect of endurance training but had no effect on strength gain (65). Several RCTs assessed the effects of oxygen on long-term PR and training (66–74; Table 4). Emtner et al. studied non-hypoxemic COPD patients undergoing high-intensity endurance training with a cycle ergometer for 7 weeks while breathing either supplemental oxygen or room air and showed that the oxygen training group had an improved maximal work rate, increased endurance, and reduced exercise breathing frequency than the air group (70). Dyer et al. analyzed the effects of supplemental oxygen during the PR program (72). They compared the results of PR undertaken either with or without ambulatory oxygen in patients with a demonstrable benefit from oxygen at baseline (patients who walked an additional 10% or more with ambulatory oxygen in an endurance shuttle walk test). Participants who were randomized to the oxygen group were asked to use oxygen at their prescribed flow rate at home during all activities that induced dyspnea. It was shown that supplemental oxygen significantly improved the endurance walking distance.

Table 4. Randomized Controlled Trials For Long-Term Oxygen Therapy For Pulmonary Rehabilitation And Exercise Training In COPD Patients.

	Study population	Comparison (n recruited/ Completed the study)	FEV1 (L and/or % predicted)	Resting		Training type	Training duration	Impact on outcomes
				PaO2 (mmHg)	SaO2
Rooyackers^66 1997	Patients with SaO2 <90% at the peak incremental cycle exercise	Oxygen 4 L/min (12/12)	0.9 L (29%)	77	NA	In-patient pulmonary rehabilitation program (general exercise training included dynamic and isometric strength training and specific training of daily life activities like breathing retraining, physical therapy)	80 minutes5x/week10 weeks	Primary outcome: Although supplemental oxygen had acute beneficial effects on exercise performance, oxygen-supplemented exercise training did not add to the effects of training while breathing room air.
		Room air (12/12)	1.2 L (38%)	79	NA			Secondary outcomes: Pulmonary rehabilitation improved exercise performance and quality of life (assessed with CRQ) in both groups.
Fichter ^67 1999	Patients with stable disease	Oxygen 35% (5/5)	43.2%	74	NA	Cycle ergometer training at a constant workload of 80% of their highest achieved work-rate	45 minutes5x/week4 weeks	Primary outcome: Change in maximally achieved power in the incremental exercise test after training was 15.9% in the supplemental oxygen group and 36.3% in room air group
		Room air (5/5)
Garrod^68 2000	Patients with limited exercise tolerance due to dyspnea and having a fall in arterial saturation of at least 4% from baseline to 90% or below upon exercise testing	Oxygen 4 L/min (13/11)	0.77 L (28.5%)	64	92%	Upper and lower limb training	60 minutes3x/week6 weeks	Primary outcome: There was no l benefit of oxygen on exercise tolerance or health status (no change in SWT, CRQ, HAD or London Chest Activity of Daily Living scales)
		Compressed air (12/11)	0.84 L (35.0%)	64	93%			Secondary outcome: The supplemental oxygen group had a greater reduction in dyspnea upon exertion, as measured with the Borg scale compared with the air group
Wadell^69 2001	Patients with PaO2 >60 mmHg at rest and hypoxemia during exercise (SaO2 ≤92% in the 6MWD)	Oxygen 5 L/min (11/10)	39.3%	71	95%	Interval walking on a treadmill (intensity set according to Borg ratings)	30 minutes3x/week8 weeks	Primary outcome: Supplemental oxygen did not improve the training effect compared with training with air
		Compressed air (11/10)	51.6%	70	95%			Secondary outcomes: Training significantly improved distance in 6MWT in both groups. Oxygen group showed no decrease in Borg ratings for dyspnea, higher increase in lactate during exercise after training and increase in time for SaO2<90% during the test
Emtner^70 2003	Patients who do not experience appreciable desaturation during exercise (SpO2 ≥88% during a constant work rate test)	Oxygen 3 L/min (15/14)	1.0 L (35%)	72	94%	High-intensity endurance training with a cycle ergometer	45 minutes3x/week7 weeks	Primary outcome: Supplemental oxygen group achieved higher training intensity and endurance capacity and the breathing pattern improved significantly
		Compressed air (15/15)	1.0 L (38%)	74	93%			Secondary outcome: Quality of life (assessed with CRQ) increased significantly in both groups
Scorsone^71 2010	Patients with FEV₁ <60% predicted	Oxygen 40% (10/10)	47%	74	NA	Leg-cycle training at 80% of the initial peak work rate for at least 20 minutes	40 minutes3x/week8 weeks	Primary outcome: Supplemental oxygen and heliox did not provide an advantage for improving exercise tolerance and symptoms over air
		Humidified air (10/10)	50%	76	NA			Secondary outcomes: Physical training resulted in a significant improvement in peak oxygen consumption, power output and exercise endurance time in all groups
		Heliox 60-40 (10/10)	49%	77	NA
Dyer^72 2012	Patients desaturated by >4% and to <90% upon exertion, and walked 10% or more further with ambulatory oxygen on ESWT	Oxygen 2-6 L/min to keep SpO2 >90% (28/24)	0.96 L (39%)	NA	94%	Strengthening and endurance exercise tailored to suit individual needs with aerobic exercise (walking) set at 85% the estimated peak oxygen uptake together with an education program	2x/week6-7 weeks	Primary outcome: Supplemental oxygen significantly improved the endurance walking distance during pulmonary rehabilitation program
		Room air (27/23)	1.12 L (44%)	NA	94%	Home exercise program: Participants randomized to the oxygen group were asked to use oxygen at their prescribed flow rate at home during all activities that induced dyspnea		Secondary outcome: There was no statistically significant change in quality of life (assessed with CRQ, HAD scale, SIFT questionnaire)
Ringbaek^73 2013	Patients desaturated by >4% and <90% during ESWT	Oxygen 2 L/min (22/12)	32.8 %	NA	94%	Supervised walking and cycling and unsupervised home training (patients were instructed to continue the unsupervised training at home at least 30 min every day during the whole study period)	1st phase: 30 minutes 2x/week for 7 weeks;2nd phase: 30 minutes 1x/week for 12 weeks;3rd phase: Only unsupervised home training at least 30 minutes for 12 weeks	Primary outcome: Supplemental oxygen improved oxygen saturation during ESWT but there was no difference between the groups at 33-weeks of evaluation with regard to the change in ESWT
		Control (23/17)	10.9%	NA	94%	Participants randomized to the oxygen group were asked to use oxygen during supervised and unsupervised training		Secondary outcomes: Oxygen therapy did not provide an additional benefit in terms of quality of life (assessed with SGRQ), risk of acute exacerbations and hospitalizations
Spielmanns^74 2014	Patients who were normoxemic (SaO2 >90%) at rest and during exercise without an oxygen supply	Oxygen 4 L/min (42/19)	1.2 L (44%)	NA	> 90%	Endurance training program with progressively increasing loads involving large muscle groups	30 min3x/week24 weeks	Primary outcome: Supplemental oxygen during the training program did not result additional change in 6MWT distance
		Compressed air (43/17)	1.5 L (43%)					Secondary outcomes: Endurance training resulted in significant improvements in quality of life (assessed with SF-36) and exercise capacity in subjects with moderate-to-severe COPD

 

Abbreviations:

6MWT: 6-minute walk test, COPD: chronic obstructive pulmonary disease, CRQ: chronic respiratory disease questionnaire, CRQ: chronic respiratory questionnaire, ESWT: endurance shuttle walk test, FEV₁: forced expiratory volume in one second (% predicted), HAD: hospital anxiety and depression scale, L/min: liters per minute, LTOT: long-term oxygen therapy, NA: data not available, O₂: oxygen, PaO₂: partial pressure of oxygen in arterial blood, SaO₂: oxygen saturation, SF 36: short form 36, SGRQ: St George's respiratory questionnaire, SIFT: surrey information on fuctional tool, SpO₂: oxygen saturation with pulse oximeter

In contrast to these findings, some studies have reported negative results. Exercise training has been shown to result in significant improvements in exercise capacity and QoL in subjects with moderate-to-severe COPD; however, oxygen supplementation did not accentuate the benefits of training (66,68,69). Spielmanns et al. used oxygen during a 24-week training program with progressively increasing loads involving the large muscle groups in a randomized controlled study, and showed that training caused statistically significant improvements in QoL, the maximal tolerated load during cycling, peak oxygen uptake, and the 6MWT. However this benefit was independent of oxygen supplementation (74). Ringbaek et al. randomized normoxemic COPD patients to control or ambulatory oxygen of 2 liters/minute during exercise and reported that there was no difference between the groups in the endurance shuttle walk test, change in St George's Respiratory Questionnaire scores, or number of patients with acute exacerbation, hospital admission or dropout (73). In a meta-analysis, it was reported that oxygen supplementation during training may augment the benefits of exercise; however, the available data remain inconsistent due to small sample sizes and methodological variations among the studies (75).

There is insufficient evidence for the beneficial effects of oxygen supplementation during PR and exercise training. A recent trial protocol was published to understand the role of oxygen in PR (SuppORT trial). This randomized controlled study will recruit COPD patients who demonstrate oxygen desaturation lower than 90% during the 6MWT and address whether supplemental oxygen during exercise training is more effective than medical air in improving exercise capacity and QoL (76). The results will help us to further understand the role of oxygen supplementation during exercise training.

Oxygen therapy for the relief of dyspnea in COPD

Relief of dyspnea during exercise

Many COPD patients experience breathlessness during small tasks even if they are not hypoxemic. Oxygen relieves dyspnea and improves recovery from breathlessness (25). Oxygen supplementation may be helpful for patients to carry on their daily routine activities and therefore may improve QoL. However, the results of RCTs (Table 5) have shown that the use of ambulatory oxygen therapy for daily activities have limited effects for the relief of dyspnea, exercise performance, and QoL (62,77–79). Eaton et al. conducted a 12-week double-blind randomized cross-over trial in severe COPD patients with exertional desaturation (78). Approximately 70% of the patients had a reduced Borg dyspnea score and improved health-related QoL with supplemental oxygen. In contrast, McDonald et al. reported that supplemental oxygen increased the walking distance in the acute setting but did not provide any long-term benefit in terms of dyspnea and QoL (77). Similar results were also reported by Nonoyama et al., who showed that only a small proportion of patients with mild hypoxemia experienced a benefit from home oxygen after long-term use (79). Moore et al. evaluated a 12-week, parallel, double-blinded, randomized trial of cylinder air versus cylinder oxygen, provided at 6 liters/minute for use during any activity provoking breathlessness, and no benefits were observed in terms of dyspnea or QoL (62).

Table 5. Randomized Controlled Trials For Long-Term Oxygen Therapy For Exertional Dyspnea in COPD Patients.

	Study population	Comparison	FEV1 (L and/or % predicted)	Resting		Study Duration	Impact on Outcomes
				PaO2 (mmHg)	SaO2
McDonald^77 1995	Patients describing exertional dyspnea sufficient to interfere with daily activities	Oxygen (4 L/min)	0.9 L	69	94%	6 weeks of oxygen and 6 weeks of air	Primary outcome: Suplemental oxygen increased 6MWT distance and steps achieved by small, statistically significant increments acutely at baseline and after 6 and 12 weeks without corresponding falls in Borg dyspnea scores in the acute studies however there was no difference in 6MWT distance, the number of steps achieved and Borg dyspnea scores between oxygen and air group in domiciliary studies
	Study with crossover design: 33 participants-26 completed	Compressed air					Secondary outcomes: Quality of life indices improved by statistically significant amounts after home oxygen therapy. Quality of life (assessed with CRQ) did not change after 6 weeks use of domiciliary oxygen supplementation during exertion compared to domiciliary supplemental air
Eaton^78 2002	Patients who experience exertional dyspnea impacting daily activities and having exertional desaturation ≤88% on air	Oxygen 4 L/min	25.9%	69	94%	6 weeks of oxygen and 6 weeks of air	Primary outcome: Improvements were seen in all domains of the disease-specific CRQ, both domains of the HAD and several domains of the SF-36 for oxygen compared with cylinder air
	Study with crossover design: 50 participants-41 completed	Compressed air					Secondary outcome: Baseline characteristics and acute response to oxygen therapy did not predict short-term (12 weeks) response
Nonoyama^79 2007	Patients who experience dyspnea limiting daily activities, and with desaturation ≤88% for 2 continuous minutes during room-air 6MWT	Oxygen 2 L/min (Range 1-3 L/min)	NA	NA	NA	Three pairs of 2-week home treatment periods, with oxygen provided during one period of each pair and a placebo mixture for the other period	Primary outcome: No apparent effect of ambulatory oxygen was observed on any of the four domains of the CRQ or the SGRQ
	N-of-1 RCT design: 38 participants-27 completed	Placebo mixture*					Secondary outcome: Ambulatory oxygen improved constant power exercise saturation and increased endurance, 5-min home walking test and dyspnea when compared to placebo in acute studies
Moore^62 2011	Patients with moderate to severe exertional dyspnea (Medical Research Council Dyspnea Scale grade >3)	Oxygen 6 L/min	1.15 L (47%)	70	NA	12 weeks	Primary outcome: Domiciliary ambulatory oxygen did not cause any improvement in dyspnea assessed with dyspnea domain of the CRQ
	Study with parallel groups: 143 participants-139 completed	Compressed air	1.17 L (47%)	73	NA		Secondary outcomes: There was no difference between groups for baseline/transition dyspnea index, quality of life (CRQ, HAD and assessment of quality of life utility index) or function (6MWT)

Abbreviations:

6MWT: 6-minute walk test, COPD: chronic obstructive pulmonary disease, CRQ: chronic respiratory disease questionnaire, HAD: hospital anxiety and depression scale, NA: data not available FEV₁: forced expiratory volume in one second (% predicted), L/min: liters per minute, LTOT: long-term oxygen therapy, O₂: oxygen, PaO2: partial pressure of oxygen in arterial blood, SaO₂: oxygen saturation, SF 36: short form 36, SGRQ: St George's respiratory questionnaire

*24% oxygen diluted with ambient air provided an FiO₂ of approximately 21.2% inhaled at 2 L/min.

In a recent Cochrane review, the efficacy of oxygen therapy for breathlessness and health-related QoL was assessed in COPD patients who do not meet the criteria for LTOT (80). They included 44 studies encompassing 1195 participants in the review, among which 33 studies (901 participants) were included. The results showed that breathlessness during exercise or daily activities was reduced by oxygen compared with air (low-quality evidence) and that short-burst oxygen given before exercise was ineffective (low-quality evidence). Oxygen reduced breathlessness during exercise tests (moderate-quality evidence), but the effect on breathlessness in daily life was limited (low-quality evidence). There was no effect on health-related QoL (low-quality evidence). The authors concluded that oxygen can relieve breathlessness when given during exercise to mildly hypoxemic and non-hypoxemic people with COPD who would not otherwise qualify for home oxygen therapy. Most evidence is attributable to acute effects during exercise tests, and no evidence indicates that oxygen is beneficial for reducing breathlessness in daily life and improving QoL.

Relief of dyspnea for palliative purposes

End-stage COPD patients suffer from a high burden of symptoms. Dyspnea is the most debilitating symptom in advanced COPD and profoundly affects QoL. Patients who have intractable dyspnea usually have a fear of death by suffocation (81), and it was shown that COPD patients usually suffer from poor symptom control including dyspnea in the last week of their life (82). Supplemental oxygen is commonly used as palliative treatment in end-stage COPD patients for the relief of distress experienced by patients and their families (83). Palliative oxygen therapy is defined as the use of oxygen to relieve the sensation of refractory persistent breathlessness in advanced disease or life-limiting illness irrespective of the underlying pathology, in which all reversible causes have been or are being treated optimally (14). Despite the widespread use, data supporting this approach are contradictory, and there is limited evidence to support the routine use of supplemental oxygen to reduce dyspnea in non-hypoxemic patients with advanced COPD (84–87). In a RCT conducted in patients with life-limiting illnesses who were ineligible for LTOT, supplemental oxygen provided no additional symptomatic benefit compared to room air (86). There were no statistically significant differences between the two groups in terms of breathlessness, frequency of side effects, or change in QoL. Campbell et al. showed that when compared to air, oxygen therapy did not decrease respiratory distress in patients who were near death because of respiratory failure, as measured by the Respiratory Distress Observation Scale (87). Hypoxemic patients with corrected saturation after oxygen did not experience a significant difference in symptoms between air and palliative oxygen. A meta-analysis reported that oxygen therapy reduced dyspnea compared with medical air in COPD patients who do not qualify for home oxygen (88). Dyspnea was reduced by continuous oxygen during exertion but not short-burst oxygen therapy. Concerning the limited options in these patients and the relatively high expectations of patients and their relatives, we think that decisions regarding palliative oxygen should be considered on an individual basis. If palliative oxygen therapy is considered, a therapeutic trial is recommended after 3 days to assess the net clinical benefit (89).

One other option for the palliative oxygen support is high-flow nasal oxygen (HFNO) administered through nasal prongs. HFNO has some advantages, such as the maintenance of mucociliary function by the delivery of warmed and humidified gas (90). It also has the advantage of more reliable delivery of the desired FiO₂ when the patient have high peak inspiratory flow rates due to respiratory distress (91,92). Compared to standard oxygen delivery systems, HFNO improves ventilatory efficiency by washing out carbon dioxide from the anatomic dead space and causes a positive end expiratory pressure, when the mouth is closed (93). It has recently been shown that HFNO reduces the work of breathing in patients with both hypoxemic and hypercapnic COPD. (94,95). HFNO may reduce dyspnea successfully with greater comfort and tolerance when compared with standard conventional oxygen masks (96,97). The effect of HFNO for palliative oxygen therapy is promising and merits further research.

Conclusion

The evolution of oxygen therapy in the medical field has been a long journey (98). Despite relatively good evidence for the benefits of LTOT in hypoxemic COPD patients in improving survival, similar benefits have not been demonstrated in the majority of COPD patients with moderate or intermittent hypoxemia, either nocturnal or exercise-induced (Table 6). The role of oxygen in symptom relief, such as dyspnea, has also been controversial both during exercise and as a palliative tool. New studies are currently in the pipeline (the INOX and SuppORT trials), and thus, it is likely that in the next few years, the “oxygen journey” will continue.

Table 6. Summary of evidence for long-term oxygen use in in specific circumstances in COPD.

Moderate hypoxemia	Current evidence does not support the use of LTOT in COPD patients with moderate desaturation
Nocturnal desaturation	The benefit of LTOT for clinically relevant outcomes remains unclear due to the paucity of data
Exercise-induced desaturation	Oxygen administration causes less breathlessness and improves exercise capacity during exercise however the effectiveness of long term use of ambulatory oxygen on exercise capacity and mortality is not evident
Pulmonary rehabilitation and exercise training	There is insufficient evidence to confirm the benefits of supplemental oxygen during exercise training but oxygen administration may reduce breathlessness and improve exercise tolerance
Relief of dyspnea for palliation	Current evidence does not support the use of LTOT for the relief of dyspnea in non-hypoxemic COPD patients

Abbreviations:

COPD: chronic obstructive pulmonary disease, LTOT: long-term oxygen therapy.