Does wearing a mask during workouts affect exercise? Research review

Published 1 July 2021
Man wearing a mask lifing weights in a gym

The conclusion by Canadian researchers that the wearing of face masks has no effect on exercise may be overstating the definitiveness of their study’s findings says Southern Cross University exercise physiologist Dr Mike Climstein.


It’s January last year and we have booked our bi-yearly Bali surf trip. Impossibles surf break at the Bukit peninsula is calling me, but thanks to a new virus called COVID-19 I can’t answer the call and miss out on the trip (along with 25% of the money spent on it). That’s OK, I’m sure by September 2020 the coronavirus will be a memory and we’ll return to Bali for the much-needed break. So confident am I, in fact, that I tack on an extra week, because we truly deserve it by then. How unrealistic was I. Now it’s February 2021 and there’s no Bali in slight, or this year’s pencilled-in Hawaii surf trip. Any overseas travel for that matter.

March 2021 and although University staff are working from home, my unreliable home internet means I have special permission to continue working from the campus. The University’s air conditioning offers welcome respite from the hot South East Queensland summer, so in the scheme of things I’m doing OK, but with the gym closed, I’m going to have to get creative about my training regime… Let’s see; in the training room and my research lab I have access to a number of treadmills, cycle ergometers, a dual axis Smith machine, 200 kilograms of weights and a power rack. OK, this will work. With the COVID safety plan only one person is allowed in the lab at a time, equipment must be wiped down prior to and immediately after use, and a surgical facemask and gloves must be worn at all times. I’m ready to train, or am I? This will be very interesting.

The treadmill is first up, and it’s always high intensity interval training (the sooner the cardio is over, the quicker I can get to the ‘real’ workout, haha). Whoa, it’s only the first interval, a mere 14.0 kilometres per hour for one minute (followed by one minute walking at 6.0 kilometres per hour) and I am not liking this at all with a surgical mask on… but rules are rules. OK, let’s try the next interval, a big increase to 14.5 kilometres per hour. Same, this is not comfortable. I am not sure if I am getting enough ventilation, it seems a lot harder than it should be. I decide to give the cycle ergometer a go. Still high intensity interval training – and still the same sensation. Am I not getting enough air to breath, or is it all in my mind and simply different from what I’m used to?

This leads us to the topic of this Research Review. Canadian kinesiologist Dr Keely Shaw and her colleagues must have had similar experiences to mine, because they decided to investigate the effects of wearing a cloth face mask or disposable surgical face mask while completing high intensity exercise.

Key points from the paper

  • Researchers investigated the effects of wearing a cloth face mask or disposable surgical face mask while completing high intensity exercise
  • There were no significant differences in any of the outcome measures of time to exhaustion, peak heart rate, RPE, oxygen saturation or peak power
  • The authors concluded that exercise performance measures were not affected by wearing either a disposable surgical face mask, nor non-disposable cloth face mask compared to no face mask
  • Closer analysis of the study findings, when also considered in conjunction with possible limitations of the study, may conclude that the study’s key finding of ‘no effect’ is an oversimplification.

The authors concluded that exercise performance measures were not affected by wearing either a disposable surgical face mask, nor non-disposable cloth face mask compared to no face mask. These results are interesting, as recent research by Chandrasekaran and colleagues proposed that wearing a face mask would result in rebreathing of carbon dioxide (usually exhaled from the body), which would compromise exercise performance and lead to lower oxygen saturation levels. Chandrasekaran also proposed that the wearing of a face mask, regardless of type, would provide resistance to breathing, thereby making breathing more difficult and decreasing exercise performance.


This is a very timely study with COVID impacting exercise throughout Australia and the world. With a number of States or Territories requiring face masks be worn on an intermittent basis, it is beneficial to know that aerobic exercise appears not to be affected significantly by the wearing of a face mask.

Based upon the conclusions of Shaw and colleagues in Canada, the difficulty I was experiencing while wearing a face mask would appear at first glance to be imagined (ouch!).


I want to start this section by stating that we have published research encouraging the use of masks to limit the spread of COVID-19 at a time when the World Health Organisation and others were still claiming that masks should not be worn. I am definitely not anti-mask and I recommend their use. The comments below are purely related to the statistical interpretation of the results in the paper. While I feel that the authors should have toned down the message they relayed, I am not advocating against the use of masks in gyms or elsewhere.

“I would be very cautious concluding there is no effect on exercise from wearing a mask.”

Man wearing a mask while exercising with battle ropes in the gym

I like to take a conservative approach in the reporting of scientific results. I do not agree with the approach the authors have taken in reporting their findings. There are so many things I am concerned with about the way this data is interpreted, I will pick a few key examples only so as not to overly focus on technicalities.

If I was a reviewer or editor responsible for vetting this paper, I would have requested a number of small changes in the language used. The experiment, which the authors conducted on 14 people (one of whom was excluded due to equipment issues), was originally designed to show that there was a difference in exercise related to mask use. These statistical tests specifically look for a difference to be detected, and if a difference is not detected, that does not tell us that there was no difference – only that in the experiment no difference was detected. This could be for many reasons, such as random noise in data or the sample size not being large enough to detect differences that exist. In these cases, we may, for example, repeat the experiment with a larger sample size, to see if we can detect a difference with more data.

When we wish to run an experiment to show there is no difference from an activity, such as wearing a mask, one option is to make an assumption on what size of effect we would expect to detect a difference. The authors did exactly this: they made an assumption about how much peak power output would be reduced by ‘mild hypoxia” and they assumed that it would be a 5% reduction. They then concluded that if mild hypoxia was present, with their sample size it should be enough to detect it 95% of the time. So, if they ran their experiment 20 times they should detect mild hypoxia if it was present 19 times out of 20. So far, this is acceptable. However, their finding should then read ‘mild hypoxia not present when performing a progressive exercise test on a Monark cycle’, as opposed to ‘Wearing a face mask has no effect on vigorous exercise’. From the conservative standpoint that I take with regards scientific results, the title of the paper, although more exciting, over-reports the results of the study.

Further concerns about their study design revolve around the number of significant figures reported for result values in their data. I will focus on one value they recorded, as an example, namely blood oxygen saturation (%). This value is reported to the nearest 1%, with approximate mean values of 96% for no mask or a surgical mask and 95% for a cloth mask. If the authors are recording data accurately to more significant figures, it should be reported more precisely. It looks like the values are approximately the same because the data is rounded in their results table, but looking at their graph it appears at every single mean power output value from 0 to 100%, that the arterial oxygen saturation is lower when wearing a mask. Visual inspection would therefore imply that there may be a difference when wearing a mask. I can go into this in more detail, however in concise terms for the purposes of this Research Review I would be very cautious concluding there is no effect on exercise from wearing a mask.

The effect is similar for RPE; the mean RPE is clearly higher for the majority of peak power output % values (from their graph) when wearing a mask than without. This may not be statistically significant; however, I believe that the statistical analysis and assumptions the authors have selected are not enough to rule out this effect – which is what they have done in their interpretation of results. This could be a purely psychological effect, but it at least establishes that I am perhaps not alone in perceiving a greater rate of exertion.

Another caution is that the study authors used Bonferroni analyses, which are susceptible to false negatives, namely not detecting differences when they exist (exactly what is being claimed).

Therefore, in my interpretation, the study authors’ conclusion of no differences from wearing a mask does not agree with the literature they quote in their own manuscript and is not well supported by their own data analysis. Regardless, I would like to be clear that people should still wear a mask during exercise when mandated by state regulations: doing so might save the life of an elderly or unfit fellow patron working out in the gym, someone simply walking past while an exerciser goes for a jog, or anyone who later interacts with an infected person.

This was a very timely study with good methodology, however, it would have been appropriate for the investigators to also use the Rating of Perceived Dyspnea (RPD) scale, as this is specific to shortness of breath (SOB, one of my favourite medical abbreviations) and breathlessness. Most exercise enthusiasts and personal trainers will be familiar with Borg’s rating of perceived exertion scale. This scale was developed by Borg and is widely used with any patients who have respiratory disorders like chronic obstructive pulmonary disease and asthma. The scale is from 0 (nothing at all) to 10 (maximal) and pertains to the participants rating of the difficulty of their breathing.

Mike Climstein, Southern Cross University

Joe Walsh, Sport Science Institute

This article has been republished with permission from the Australian Fitness Network. Click here to view the original article. 

Media contact: Sharlene King, media officer at Southern Cross University +61 429 661 349 or