Rebuttal from Mehul S. Patel, Nicholas Hart and Michael I. Polkey

Professor McConnell's thesis seems to be that a single meta-analysis (Illi et al. 2012) in healthy individuals concludes that respiratory muscle training (RMT) improves exercise tolerance (McConnell, 2012, this issue), coupled with eloquent mechanistic arguments that it ought indeed to do so. Overall, we retain our view that the literature does not sufficiently support the argument that RMT improves exercise tolerance (Patel et al. 2012, this issue).

The thorough work of Illi and colleagues is insightful with regard to the quality of the literature; a point raised by us. First, from an initial screening of 236 potentially relevant studies, only 28 were randomised controlled studies and the median quality score was 2 out of 7, illustrating the low standard of work in this area. Subsequently, only 8 studies provided sufficient data to be included in a meta-analysis. However, we are indebted to Professor McConnell for calling to our attention what we believe to be the most revealing figure from Illi and co-workers’ meta-analysis (see Fig. 1). This is a textbook example of publication bias, in which small negative studies remain unpublished, so fatally undermining their conclusions.

Figure 1. Funnel plot revealing the absence of negative smaller studies.

CLT = constant load test; TT = time trial; IIT = intermittent incremental test; IT = conventional (non-intermittent) incremental test; SMD = standardised mean difference; SE = standard error. Reproduced from Illi et al. (2012) with permission from Springer International AG 2012. All rights reserved.

Illi and colleagues’ work also sidesteps the question of whether respiratory muscle training is better than general physical training. We have previously noted, in the case of chronic obstructive pulmonary disease (COPD), that time spent on respiratory muscle training may entail an ‘opportunity cost’ for physical training (Polkey et al. 2011). Illi et al.'s sub-analysis suggested that less fit individuals benefited more from RMT, unsurprising given the recognised benefits of whole body training (Bernard et al. 1999; Mickleborough et al. 2008). Moreover, considered mechanistically, the peripheral muscles seem to represent the better training target given that preferential perfusion of peripheral rather than respiratory muscles during exercise may limit global benefits with RMT (Vogiatzis et al. 2010).

Space does not permit an in-depth discussion of the detailed physiology of the metaboreflex, and in any case, like Professor McConnell, we applaud the physiology that elucidated it. Nevertheless, we caution that interventions used in physiological scenarios, for example by using assisted ventilation (Harms et al. 2000; Kyroussis et al. 2000; Romer et al. 2006), may be of an order of magnitude significantly different to the benefits conferred by RMT. For this reason, we do not believe that mechanistic studies can be a substitute for straightforward experimentation when answering the question posed.

Call for comments

Readers are invited to give their views on this and the accompanying CrossTalk articles in this issue by submitting a brief comment. Comments must not exceed 250 words, with a maximum of six references from peer reviewed publications only. To submit a comment, use the online form available in the centre panel on the HighWire site. If other responses have already been submitted, a ‘view comments’ link will be visible.

All comments will be moderated, and those deemed to add significantly to the discussion will be published online-only as footnotes to the articles. Comments may be posted up to 6 weeks after publication of the article, at which point the discussion will close and authors will be invited to submit a ‘final word’.

Questions about this call should be directed to Jerry Dempsey at jdempsey@wisc.edu.

To submit a comment, go to: http://jp.physoc.org/letters/submit/jphysiol;590/15/3399