RCT of Oxygen vs. Room Air (Delivered by a Concentrator)

Drawing of a nasal cannula from Wikimedia Commons

It's a common assumption amongst both the general population as well as medical professionals that breathlessness equals some problem with oxygen delivery, and therefore, every patient with dyspnea should have at least a little bit of oxygen delivered through a nasal cannula regardless of their oxygen saturation (sometimes endlessly).  Because of the ubiquitous nature of oxygen therapy, why not?  It helps many patients, and it makes sense that the higher concentration of O2 molecules represents the critical component of the therapy.  It's been shown to palliate dyspnea, improve functional status, and prolong life in patients with COPD with hypoxemia. 
Hypoxia isn't always the cause of dyspnea, however, and Abernethy et al. set out to test the hypothesis that room air delivered by a concentrator at 2 liters per minute might rival the efficacy of oxygen delivered at the same rate for patients with advanced disease, refractory dyspnea, and normal blood oxygen concentrations (Pa02 >55 mmHg).  The results of the double-blind, randomized control trial were published in Lancet recently. 
The trial population included patients with a variety of primary diagnoses with COPD (60-63%) the most common, lung cancer (13-15%) second most common, and several others.  The average age was 73-74.  At baseline, patients were also similar with respect to dyspnea measurements, QOL, and partial pressure of arterial oxygen and carbon dioxide.  The majority of patients had an ECOG performance status between 1 and 3.
Of 567 patients referred, 239 were eventually randomized with the most common exclusion criteria including the presence of a recent prescription for O2, acute cardiopulmonary event recently, hypercapnea, hypoxemia, anemia, or delirium.
Patients received seven days of either oxygen or room air delivered by a concentrator.  Why seven days?  Because in a pre-trial survey, it was determined that palliative care practitioners might be influenced by a trial that lasted for at least seven days!  (Great planning.)  The primary outcome measured was the patient's rating of restlessness on a numerical scale (0-10) in the morning after waking up and before going to bed for each of the days.  A clinically significant response was defined as a reduction in breathlessness rating of >1 on the numerical scale, and measurements for the two groups were compared on all days of the study (as well as a few days before).  There were several secondary outcomes. 
The primary outcomes:

• Breathlessness measurements were similar throughout the seven days of the trial when the two groups were compared.  
• Morning dyspnea: 58 (52%) of 112 patients assigned to oxygen and 40 (40%) of 101 patients assigned to room air responded to the interventions (statistically significant difference). 
• Evening dyspnea: response rates were 42% for both interventions
  (oxygen, n=47; room air, n=42).
• Over intervention period, breathlessness scores of both groups improved significantly.  

Secondary outcomes such as other measurements of dyspnea and quality of life did not differ between the groups.  The proportion of patients reporting severe dyspnea or dyspnea that disturbed sleep decrerased in both groups similarly.
Severity of baseline dyspnea did predict response to therapy in both groups for both morning and evening measurements (the patients with worse dyspnea were less likely to respond to either therapy).  Use of oxygen did predict morning response but not evening response.
Breathlessness was not measured with activity, but subjects desire for Oxygen therapy was measured at the end of the trial.  Interestingly, distributions of preferences were similar between groups:

• 18% of all participants didn't want oxygen therapy afterwards.
• 26% said they derived no benefit
• 48% requested oxygen

The population studied was heterogeneous.  The authors draw no conclusions regarding differences based on disease state.  I look forward to the authors' further analysis to help determine whether diagnosis matters.
This study got me thinking a little more about some questions:
When is room air really room air?  
Medical air may have the same concentration of oxygen in it as room air, but the effect of flow/ turbulence produced by the nasal cannula may have a salutary effect on dyspnea regardless of the concentration of oxygen.  In addition to a placebo effect occuring in both groups, there is likely therapeutic benefit of forced air (the authors reference studies indicating the benefit of a simple, hand-held fan, for instance). 
Should we provide air via concentrator rather than concentrated O2 in some cases?
No.  Maybe if oxygen therapy weren't so ubiquitous and medical air were the standard of care already, maybe you could make that case.  Although I don't know for sure, I'm going to guess that a room air concentrator is similar in cost to an oxygen concentrator, so I'm unsure of any cost implications.  Even though the the null hypothesis has been confirmed, the authors suggest that oxygen therapy should still be tried briefly and then "less burdensome" strategies should be considered depending on the effect in the individual patient.
Is that conclusion strong enough?
I'd go a mini-step further.  For hospice patients where I practice (and I'm assuming with most American hospices), even a hint of dyspnea often equals automatic ordering of oxygen therapy.  The symbolism of oxygen arriving at a person's home is rich.  Hospices are rightfully out to win the trust of patients and families, and the efficient delivery of DME (especially something that patients are only accustomed to seeing in ambulances, emergency departments, and hospitals) sends a signal that the hospice means business!  This study doesn't influence any of that symbolism.
Perhaps if a patient with mild to moderate dyspnea does not have oxygen in the home (and especially for those who may have a recent documented normal oxygen saturation), this study helps us put a mental brake on ordering oxygen momentarily. This study may corroborate that there are a number of patients who have experienced nasal cannulas, hate them, don't want them, but many will silently submit to our desire to help by throwing one on them.  For these patients, turn up the fans and give them a little morphine (if they so desire).  This study doesn't prove the efficacy of a nice fan vs. oxygen, but it's probably as close as we'll get, and the results show that O2 isn't the be-all, end-all.  So we shouldn't feel compelled to say "you have to wear it or else."
Other patients may be ambivalent.  Concerns should be explored and perhaps a trial of other therapies offered. Many patients will still ask for oxygen therapy, and this study wouldn't change our response to that request but informs how we teach patients about their symptom regardless.
What about the hospitalized setting and sicker patients?
Indeed, the authors point out that this was a population that still had a fair performance status and little dyspnea at rest.  This was entirely an outpatient population.  The former point is especially important to keep in mind- the results may not apply to a sicker crowd with more severe symptoms. 
What do you think?  Does this study influence your practice or intrinsically change how you view dyspnea?  This is a great one for journal clubs.

Abernethy, A., McDonald, C., Frith, P., Clark, K., Herndon II, J., Marcello, J., Young, I., Bull, J., Wilcock, A., & Booth, S. (2010). Effect of palliative oxygen versus room air in relief of breathlessness in patients with refractory dyspnoea: a double-blind, randomised controlled trial The Lancet, 376 (9743), 784-793 DOI: 10.1016/S0140-6736(10)61115-4