Supplemental oxygen and dypsnoea in interstitial lung disease: absence of evidence is not evidence of absence

To the Editor:

It is with great interest that we read the recent systematic review by Bell et al. [1] concerning the effects of oxygen therapy on dyspnoea and exercise capacity in patients with interstitial lung disease (ILD). The authors report that, while supplemental oxygen increases exercise capacity, it does not improve dyspnoea. Overall, this is a well-executed systematic review that accurately reflects the current literature; however, we believe the conclusion regarding the lack of benefit of supplemental oxygen on dyspnoea in ILD is misleading. This opinion is not based on the quality of the systematic review, but rather on the quality of the existing literature that was evaluated. We base our opinion on two important lines of evidence. First, previous studies that have evaluated the effect of supplemental oxygen on exertional dyspnoea in ILD patients only report peak or end-exercise dyspnoea ratings, thereby ignoring important clinically and physiologically relevant changes occurring at submaximal exercise intensities. Secondly, the exercise testing modalities employed are variable, such as self-paced walk tests [2–5] or incremental cycle exercise tests [6], which are often insensitive to changes in dyspnoea. An additional concern relates to the measures in place to reduce experimental bias, as the authors have addressed. Indeed, the only study included in their review that showed a beneficial effect of oxygen on dyspnoea did not have an appropriate control condition [5], making it impossible to rule out the placebo effect.

The measurement of dyspnoea is founded on the principle of psychophysics, whereby a stimulus is linked to a sensation [7]. It has been established that humans can reliably detect, quantify and discriminate qualitatively distinct sensations of dyspnoea provoked by different respiratory stimuli, either applied experimentally or as a result of disease [8]. Accordingly, the only way to properly evaluate the effects of any therapeutic intervention on exertional dyspnoea is to standardise the stimulus intensity. Specifically, this requires the assessment of dyspnoea at the same absolute exercise intensity and measurement time. It is extremely challenging to standardise exercise intensity during self-paced walk tests and thus, cycle exercise tests are more appropriate by design. Additionally, peak or end-exercise dyspnoea ratings are often insensitive to change following an intervention. This may be attributed to exercise tolerance improving as a result of the intervention and/or the fact that symptom-limited exercise cessation typically occurs at a similar level of dyspnoea for a given individual [9]. In other words, peak exercise is not the same stimulus if exercise time is increased (i.e. for incremental and constant work-rate cycle exercise tests) or if walking pace and/or distance changes (i.e. during self-paced walking tests).

In addition to standardising exercise intensity and measurement times, the magnitude of inspiratory oxygen fraction (F_IO2) delivered should be the same between individuals and should be of sufficient magnitude to reverse or, at the very least, cause a meaningful improvement in arterial oxygen saturation. With the exception of one study [6], all previous studies that examined dyspnoea in ILD used nasal cannulae or venturi masks to deliver the oxygen at various flow rates [2–5, 10]. These gas delivery methods have well-established limitations [11, 12] and the effective F_IO2 is likely to vary between individuals based on breathing pattern. These systems also have modest effects on improving arterial oxygen saturation compared with control conditions [2–5], which may explain, at least in part, previous negative studies of oxygen on dyspnoea in ILD. By contrast, administering oxygen through a reservoir bag connected to a two-way non-rebreathing valve, while less practical, allows for the precise control of F_IO2 and is more effective at preventing arterial oxygen desaturation. This design is critical for testing the potential benefit of increased F_IO2 on dyspnoea. If a benefit of supplemental oxygen during exercise is proven, there will be a need for additional research to identify the best means of translating these findings into clinical practice within a standard pulmonary rehabilitation setting. Thus, in order to make an unbiased evaluation of supplemental oxygen on dyspnoea, dyspnoea ratings should be measured at the same absolute submaximal work rate and exercise time (i.e. iso-time) while administering identical levels of effective F_IO2 compared with an appropriate placebo condition.

We recently completed a study that addresses the aforementioned limitations in order to evaluate the effect of supplemental oxygen on dyspnoea during exercise in patients with ILD [13]. During this single-blind, randomised, placebo-controlled crossover study, 20 ILD patients performed two symptom-limited constant work-rate cycle exercise tests at 75% of peak work rate while breathing room air or supplemental oxygen (60% oxygen) in randomised order. Gas was delivered via a two-way non-rebreathing valve connected to a Douglas bag, such that the effective F_IO2 was constant in all patients and was sufficient to reverse arterial oxygen desaturation, even at peak exercise (mean±sd arterial oxygen saturation measured by pulse oximetry was 98±1% with hyperoxia versus 90±4% with normoxia; p<0.001). Gas cylinders and oxygen saturation monitors were obstructed from patient view, and gas delivery was identical for both experimental conditions. Briefly, dyspnoea was significantly reduced at iso-time with supplemental oxygen versus room air (mean±sd 2.5±2.1 versus 4.4±3.1 Borg units; p<0.01), but was not different at peak exercise (5.4±3.6 versus 5.6±3.0 Borg units; p=0.69) despite a significant improvement in exercise endurance time by an average of 10.3 min (mean±sd 21.9±12.9 versus 11.6±10.0 min; p<0.001). Figure 1 shows mean data as well as the dyspnoea response from an individual patient to illustrate the fundamental importance of measuring and reporting submaximal dyspnoea ratings. The data clearly show a substantial and clinically relevant reduction in dyspnoea at iso-time with hyperoxia in both examples, with no improvement at peak exercise. Relying exclusively on peak dyspnoea, as has been done in previous studies, results in the erroneous conclusion that supplemental oxygen is ineffective at reducing dyspnoea. The results of our study clearly demonstrate a reduction in dyspnoea with supplemental oxygen by an amount that well exceeds the minimal clinically important difference (derived from chronic obstructive pulmonary disease studies) of 1 Borg unit (scale of 0–10) [14]. The positive results from this study formed the impetus for a large multicentre clinical trial evaluating the role of hyperoxia in a long-term exercise-training programme in patients with idiopathic pulmonary fibrosis [15].

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

a) Dyspnoea response at rest, iso-time, and peak exercise in 20 patients with fibrotic interstitial lung disease during constant work-rate cycle exercise testing while breathing 60% oxygen or room air. There was a statistically significant improvement in dyspnoea at iso-time but no change at peak exercise. Data are presented as mean±sem. *: p<0.05. b) Dyspnoea ratings every 2 min during constant load cycling in an individual patient while breathing 60% oxygen or room air. Dyspnoea ratings were reduced at all submaximal exercise times but the rating was increased at peak exercise. The Borg dyspnoea scale ranges from 0 to 10. Data from [13].

The comprehensive systematic review by Bell et al. [1] did not identify a clinically significant benefit of supplemental oxygen on dyspnoea in patients with ILD; however, this review is subject to the methodological limitations of the available literature. We believe that the primary conclusion to be drawn from the existing literature is that there is currently insufficient evidence to suggest either a benefit or a lack of benefit from supplemental oxygen during exercise in patients with ILD. Methodological factors, such as the mode of gas delivery, F_IO2, the timing of dyspnoea measurements, and the exercise testing protocols, have been inadequately considered in most previous studies and probably contribute to the difference in conclusions between our study and the available literature. Before clinicians dismiss supplemental oxygen as a dyspnoea-relieving therapy in ILD, they should recognise that all studies reviewed by Bell et al. [1] were inadequately designed to evaluate the impact of supplemental oxygen on dyspnoea.

Disclosures