Thinking outside the Bag (Not Necessarily outside the Lab)

Dear Editor-in-Chief: As avid users of the Douglas bag technique for indirect calorimetry in our own research, we found the commentary “Who Needs a Bag?” deeply affirming (2). We wholeheartedly resonate with the key messages that the method remains a state-of-the-art gold standard and should accordingly form a core element of research training in the “first principles” of metabolic assessment. This commentary was prompted by an equally eloquent original article in the same issue by Hopker et al. (3), who neatly evaluated potential sources of error at various steps in the process of sampling and analyzing expired gases. In doing so, however, the authors have inadvertently highlighted what appears to be an altogether overlooked and therefore widespread error in indirect calorimetric analysis arising from our long-standing reliance on the assumption that inspired gas fractions are both stable and reflective of atmospheric constants. (Hopker et al. quote the commonly cited but rarely referenced values of 20.93% oxygen and 0.03% carbon dioxide.) Although the assumed constant of 0.03% carbon dioxide might well have been valid at the beginning of the 20th century when the principles of indirect calorimetry were first realized (4) and Douglas (1) initially described his now classic method of collecting expired gases, atmospheric carbon dioxide content has since risen exponentially, with mean global values having rounded to 0.04% for over two decades already and due to surpass that absolute figure this year for the first time in human history (6). Moreover, the commonly adopted constant of 20.93% oxygen remains derived from the highly variable range of values reported until recently (5), yet reliable atmospheric oxygen data only became available within the past two decades and already reveal a decrease >0.03% in that time alone, reaching the current nadir of 20.95% (i.e., past values deviated even further from those assumed (7)). Basing contemporary calculations of metabolic substrate oxidation on atmospheric values estimated before the industrial revolution therefore systematically underestimates energy expenditure by approximately 0.6%, yet the outdated constants still widely cited in our field suggest that the extent of recent changes in atmospheric composition is not common knowledge. (Indeed, we had to search beyond the human physiology/metabolism literature to ascertain the above values.) Irrespective of global climate change, however, even well-ventilated laboratories unavoidably become “contaminated” with inhabitants’ expired gases such that ambient air about the participant can readily approach 0.1% carbon dioxide and 20.7% oxygen over the course of testing. Under such conditions, energy expenditure would be gradually overestimated by up to 7%. Unlike random errors, which can largely be managed by replicate samples/analyses to dilute “noise” in measurement, such systematic bias clearly holds profound implications for interpretation of metabolic assessments using indirect calorimetry. We therefore propose that that the composition of the ambient air actually inspired should be measured (i.e., outside the Douglas bag proximal to the participant’s mouthpiece) with these values used relative to each corresponding expired gas sample, as opposed to simply assuming constants for atmospheric air that is neither constant nor inspired. James A. Betts Dylan Thompson Human Physiology Research Group Department for Health University of Bath Bath, United Kingdom