Published ahead of print on April 19, 2007, doi:10.1164/rccm.200702-308OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200702-308OC
Effect of Oxygen on Health Quality of Life in Patients with Chronic Obstructive Pulmonary Disease with Transient Exertional Hypoxemia1 Graduate Department of Rehabilitation Sciences, University of Toronto, Toronto, Canada; 2 Department of Respiratory Medicine, West Park Healthcare Centre, Toronto, Canada; 3 Department of Physical Therapy, University of Toronto, Toronto, Canada; 4 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada; and 5 Faculty of Medicine, University of Toronto, Toronto, Canada Correspondence and requests for reprints should be addressed to R. Goldstein, M.B. Ch.B., F.R.C.P., West Park Healthcare Centre, 82 Buttonwood Avenue, Toronto, ON, M6M 2J5 Canada. E-mail: rgoldstein{at}westpark.org
Rationale: Ambulatory oxygen improves acute exercise performance in people with chronic obstructive pulmonary disease (COPD). This improvement may not translate into symptomatic benefit for patients during activities of daily living. Objectives: We undertook a series of individual randomized controlled trials (N-of-1 RCTs) to measure the effect of oxygen in patients with COPD who do not meet criteria for mortality reduction with long-term oxygen therapy. Methods: Twenty-seven patients completed blinded N-of-1 RCTs, each comprising three pairs of 2-week home treatment periods, with oxygen provided during one period of each pair and a placebo mixture during the other.
Measurements and Main Results: Patients completed the Chronic Respiratory Questionnaire (CRQ), the St. George's Respiratory Questionnaire, and a home five-minute-walk test at the end of each period. We defined a positive response as a CRQ dyspnea score greater (less dyspnea) on oxygen than placebo during all three pairs of treatment periods, with a difference Conclusions: This study does not support the general application of long-term ambulatory oxygen therapy for patients with COPD who do not meet criteria for mortality reduction with long-term oxygen therapy. N-of-1 RCTs can identify patients who may benefit.
Key Words: chronic obstructive pulmonary disease • oxygen • placebo • exertion
Chronic obstructive pulmonary disease (COPD) is a leading cause of morbidity and mortality worldwide and results in substantial economic and social burden (1). Many patients with COPD identify dyspnea on exertion as the key adverse impact of their condition (2) and the strongest determinant of their functional status (3, 4). Several guidelines have suggested that COPD management should be individualized and based on assessments of the impact of treatment on dyspnea rather than on physiological measures of pulmonary function or arterial blood gases (5–8). A Cochrane review (9) has summarized the well-recognized mortality benefits of long-term oxygen therapy (LTOT) for individuals with chronic resting hypoxemia. Clinicians sometimes prescribe ambulatory oxygen for patients without resting hypoxemia, who experience hypoxemia only during exercise or activities of daily living (10, 11). Studies of short-term ambulatory oxygen have demonstrated improvements in acute exercise performance among patients with moderate to severe COPD (11). However, these laboratory-based, acute physiological responses may not reflect how a patient responds symptomatically to the longer term use of ambulatory oxygen in their daily lives (12). Justifying the expense and inconvenience of long-term ambulatory oxygen for transient exercise hypoxemia requires an understanding of its effect on the patient's experience in the community, during activities of daily living (13, 14). Four randomized controlled trials (RCTs), evaluating the role of ambulatory oxygen, have reported mixed results (10, 15–18). Given these limited prior RCTs, we undertook a series of individual (N-of-1) RCTs (18) to evaluate the effect of supplementary oxygen on symptoms, quality of life, and exercise tolerance among individuals with COPD who did not meet the criteria for mortality reduction (19), and to explore the variation in response across patients.
See online supplement for further details of methods used.
Patients
Study Design Oxygen and placebo gas mixtures were supplied (VitalAire Healthcare, Inc., Mississauga, ON, Canada) in identical cylinders. Concentrators were modified (AirSep NewLife Elite; AirSep, Buffalo, NY) to provide oxygen or air. Participants were requested to use the gas provided for at least 1 hour per day during activities that made them short of breath. Both the patients and the outcome assessors were blind to the gas mixture provided. Patients completed walking oximetry while breathing room air, followed by titration of oxygen, to establish the flow rate necessary to maintain saturation at 92% or greater during activities, for each of the N-of-1 RCT periods (23).
Outcomes Before commencing the N-of-1 RCT, patients completed an incremental and constant power exercise test (30, 31), the latter to determine the acute effects of supplemental oxygen on endurance time and symptoms. Patients performed the constant power test twice, with the patient blinded and the order randomized as to whether he or she was provided with compressed air (21%) or supplemental oxygen.
Data Analysis An original sample size of 40 was determined to yield a 95% confidence interval (CI) around the proportion of positive N-of-1 RCTs of approximately ±0.15 when that proportion was 0.5. For proportions greater or less than 50%, the CI would be narrower. Recruitment proved more difficult than anticipated, and when we had used the available resources, we performed an interim analysis before making a decision as to whether to apply for more.
All statistical analyses, except for the ICC, were performed using SAS 9.1.3 (SAS Institute, Inc., Cary, NC) and graphs were plotted using SigmaPlot 8.0 (Systat Software, Inc., San Jose, CA). The ICC was calculated using SPSS 12.0.1 for Windows (SPSS, Inc., Chicago, IL). An
See online supplement for further details of results.
Patients
Acute Response to Oxygen In acute studies, immediately before commencing the N-of-1 RCTs, oxygen improved constant power exercise saturation from 88 ± 4 to 94 ± 3% (p < 0.001) and increased endurance time from 4.6 ± 2.0 to 7.0 ± 4.4 min (p = 0.003). It did not influence ratings of dyspnea or leg fatigue, or end-exercise ventilation.
Group N-of-1 RCTs
There were no significant differences between oxygen and placebo in any domain of the CRQ or the SGRQ. Although there was an improvement over time for the CRQ dyspnea (p = 0.05) and the CRQ mastery (p = 0.001), these effects were unrelated to the gas mixture being used, with no main effects of oxygen, nor any interaction between treatment and pair for any of the CRQ domains. Furthermore, the upper boundary of the CI excluded a mean difference greater than the MID for all four CRQ domains (Table 2, Figure 3). There were no significant differences between oxygen and placebo, or improvements with time for any domains of the SGRQ.
The average duration of gas used was 3.5 ± 5.2 hours per day (range, 0.3–21.3) for oxygen and 2.5 ± 3.1 hours per day (range, 0.4–16.2) for placebo. The total oxygen usage consisted of 2.8 ± 5.0 (0.0–20.2) concentrator and 0.7 ± 0.6 (0.0–2.7) cylinder hours per day. The total placebo usage consisted of 1.8 ± 3.0 (0.0–15.4) concentrator and 0.6 ± 0.6 (0.0–3.0) cylinder hours per day. There was a significant correlation between the amount of oxygen used and the amount of placebo used (ICC = 0.85, p < 0.001).
Individual N-of-1 RCTs Table 3 and Figure 4 present the details of one clear responder. For this patient, the CRQ dyspnea showed consistent differences. He also walked further and had less dyspnea at the end of the walk while using oxygen. This patient used oxygen for an average of 17.5, 18.0, and 21.0 hours/day and placebo for 2.2, 3.6, and 6.6 hours/day. This patient requested early termination of all placebo periods at 3 days, whereas oxygen periods averaged 13 days.
In all other individuals, there were no significant statistical or clinical differences between oxygen and placebo for the CRQ, SGRQ, and five-minute home walk test.
Although ambulatory oxygen acutely improved constant power endurance and a standardized home walking test, very few patients experienced benefit from using oxygen at home. Group results showed no apparent effect of oxygen on any of the four domains of the CRQ or the SGRQ (Table 2, Figure 3). Only 2 of 27 patients met our criteria for an oxygen responder. In the patient whose dyspnea response was most convincing, benefit did not extend to any other domains of quality of life (Table 3, Figure 4). The acute improvement in exercise tolerance with supplemental oxygen is consistent with previous studies (11) and the absence of effect of long-term ambulatory oxygen on quality of life is consistent with three previous RCTs, although two of them (16, 17) studied patients with resting hypoxemia. Lilker and colleagues (17) used a 10-week blinded crossover design in nine patients with resting hypoxemia (17). They reported improvements in minute ventilation and arterial oxygen tension, but found no change in dyspnea, subjective assessment of activity, or distance walked per day while using liquid oxygen compared with liquid air. McDonald and coworkers (18) assigned 26 patients with a PaO2 greater than 60 mm Hg to a 12-week blinded crossover design and reported no differences in quality of life while using compressed oxygen compared with compressed air. Lacasse and associates (16) reported on a 1-year blinded, randomized, three-period crossover trial comparing (3 mo of each) the following: (1) standard therapy of home oxygen from a concentrator; (2) standard therapy plus as-needed ambulatory cylinder oxygen; and (3) standard therapy plus ambulatory cylinder compressed air, among patients with COPD and resting hypoxemia. Ambulatory oxygen had no influence on quality of life or exercise tolerance. In contrast, Eaton and colleagues (15), in a crossover study of 41 patients with COPD who received 6 weeks each of oxygen and placebo, reported differences in the CRQ favoring oxygen, with 23 of 41 patients reporting a difference of greater than the MID in their dyspnea scores between oxygen and placebo periods. Their patients had a slightly higher six-minute-walk distance (358 ± 96 vs. 283 ± 105 m) than ours, possibly because they had just completed a pulmonary rehabilitation program, but otherwise appear similar. They failed to identify predictors of benefit; those who showed acute response to oxygen in their six-minute-walk tests did not necessarily also show differences in quality of life between oxygen and placebo. The level of resting oxygenation does not explain differences between the three trials that failed to show group differences with oxygen versus placebo (16–18) and Eaton and colleagues' RCT (15). The room-air PaO2 values among patients in Lilker and coworkers' (17), Lacasse and colleagues' (16), and McDonald and coworkers' (18) studies were 53 ± 8, 53 ± 4, and 69 ± 3.3 mm Hg, respectively. In the study by Eaton and colleagues (15), which reported improvement with oxygen, the resting PaO2 was 69 ± 7.5 mm Hg. Furthermore, the level of exercise hypoxemia does not help clarify the differences between the four trials that showed minimal or no benefit from oxygen (18) and the Eaton and colleagues trial (15). Eaton's study only included patients who desaturated 88% or less and the room-air six-minute-walk saturation nadir was 82 ± 5.4%. Our patients were similar to Eaton's group, with a room-air six-minute-walk saturation nadir of 82 ± 4%. Although exertional desaturation was not an eligibility criterion in McDonald and coworkers' trial, desaturation during the room-air six-minute-walk test was 6.2 ± 5.1%, a decrease from a resting saturation of 94 ± 2.1%. Neither Lilker and colleagues (17) nor Lacasse and colleagues (16) reported the extent of exercise desaturation, but because their patients had resting hypoxemia, desaturation was likely more than in the Eaton trial (15). Adherence to treatment is an important contributor to effectiveness (5) but also fails to explain the differences. Average use of oxygen was as follows: Eaton and colleagues (15), 1.75 hours per day; Lacasse and coworkers (16), 0.5 hours per day; McDonald and colleagues (18), 5.1 ± 3.1 E-sized (capacity of 680 L) cylinders in a 6-week period (likely <0.5 h/d). Our patients used the most, an average of 3.5 ± 5.2 hours per day, concentrator and cylinder combined usage. Although we cannot know with certainty if there were no periods of transient hypoxemia for which the gas mixture was not used, our direction to the patients was to use the gas for any activities limited by dyspnea as this was the time when they were most likely to be hypoxemic. Our results reflect patient behavior that one might expect in response to these instructions. Conceivably, ambulatory oxygen may be more effective among those engaged in a higher level of activity. In studies similar to ours that have examined the influence of ambulatory oxygen among patients who do not meet the usual criteria for LTOT (15, 18), results have been inconsistent. Patients in Eaton and colleagues' study (15) who had recently completed a pulmonary rehabilitation program were at a higher functional status than ours, as reflected by their higher six-minute-walk distance (358 ± 96 vs. 283 ± 105 m) and might have been engaged in more vigorous exercise than our patient population (33). The patients in McDonald's report (18) also had a six-minute-walk distance of 326 ± 97 m, very similar to that in Eaton's study. Cylinder usage is a surrogate for activity levels in the absence of direct measurement of activity levels with activity monitors (34–36). Strengths of the N-of-1 RCT's design include the use of systematic strategies, such as blinding and quantitative measures of patient experience, to minimize bias. The close correlation between the amount of oxygen and placebo used was consistent with the patients being unaware as to which mixture they were using. Another strength of the N-of-1 RCT design is its detailed study of individuals, allowing detection of responders who would remain undetected in group-level analysis (37, 38). Indeed, in the present study, we found no apparent group effects, and yet detected two oxygen responders. Our use of a priori criteria for a positive N-of-1 RCT that emphasized a large and consistent response in CRQ dyspnea strengthens the inferences regarding response. A potential limitation of this trial is the sample size. Although we did not meet our initial sample size, our analysis suggested that it would be extremely unlikely that enrolling more patients would have altered the conclusion. In the group results, there were only very weak trends in favor of oxygen being of benefit, restricted to the dyspnea and fatigue domains of the CRQ, and confidence intervals excluded mean differences above the MID (Table 2, Figure 3). We wished to establish our best estimate of the proportion of responders, and our confidence in that estimate. Our best estimate is 7%, and the range of plausible truth, represented by the 95% CI is 1 to 24%. Although narrowing the CI, by recruiting more patients, would be helpful, the high likelihood of the proportion of responders being less than 25% is highly relevant information in that it suggests serious problems with the indiscriminate use of oxygen to relieve day-to-day dyspnea in patients with COPD without resting hypoxemia. Ambulatory oxygen therapy is routinely prescribed for those who do not meet criteria for mortality reduction, despite the inconsistent evidence regarding its effect (8, 39, 40). Professional guidelines that recommend ambulatory systems have provided laboratory results of exercise testing as supportive evidence (11, 41). Domiciliary oxygen generates substantial costs (42). In 2002, the total U.S. Medicare costs for home oxygen therapy was U.S. $2.2 billion (43). Ambulatory systems represent an additional equipment cost over stationary systems. Even the most cost-effective portable lightweight cylinders and liquid systems require frequent maintenance and delivery (44, 45). In England and Northern Ireland, in 2002–2003, more than £8.6 million (U.S. $19 million) was spent on the ingredient cost of cylinder oxygen alone (46). Oxygen is associated not only with cost but with inconvenience for the patient. These considerations suggest the inadvisability of oxygen use on the basis of laboratory tests. One approach to this issue is to restrict oxygen use to patients who show benefit in an N-of-1 RCT. Adopting such an approach would require including the cost of conducting such trials with the costs associated with ambulatory oxygen. In 1990, the cost of an N-of-1 RCT was estimated at Can $400–500 per trial (47). The many home-based assessments in this trial increased its costs to an estimated Can $1,700.00 (U.S. $1,400.00) per patient, including time spent by the respiratory therapist in patient assessment, equipment delivery and trial administration, assessments by the supervising physician, and time spent by a separate unblended individual preparing, maintaining, and ordering equipment. However, this cost is still considerably less than the cost of providing long-term ambulatory oxygen for patients who do not benefit from it. In summary, this study does not support the general application of ambulatory oxygen therapy for patients with COPD who do not meet criteria for mortality reduction with domiciliary oxygen. Our data suggest that only a small proportion of such patients with mild resting hypoxemia and exercise desaturation receive an important benefit from home oxygen.
The authors thank the patients and staff at West Park Healthcare Centre, especially Tom Dolmage, who carried out hospital-based exercise testing; Rick Collins, for managing the equipment; and John Tagg, the librarian. They thank Christina Lacchetti, Lisa Buckingham, and Diane Heels-Andsell from the data management team at McMaster University. They also thank the respiratory rehabilitation programs, home care companies, and physicians who helped with patient recruitment, especially Krisztina Weinacht. The authors also acknowledge Pierre Emmanuel-Augustin of Vital Healthcare and Maureen Williams of the Ontario Ministry of Health and Long Term Care, Home Oxygen Program.
Supported by the Ontario Ministry of Health and Long Term Care. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200702-308OC on April 19, 2007 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form February 23, 2007; accepted in final form April 17, 2007
Related articles in AJRCCM:
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
0.5 inches for at least two treatment pairs. Oxygen significantly increased the five-minute-walk test (427 vs. 412 steps, p = 0.04). Two of 27 patients met the responder criteria. Among the whole group, neither the CRQ nor the St. George's Respiratory Questionnaire showed any statistical or clinical differences between oxygen and placebo. 
of 0.05 was considered significant. All descriptive data are expressed as means and standard deviations. All estimates are expressed as mean differences and 95% CI unless otherwise stated. 
