Among clinical trials investigating long-term weight loss, exercise as the only intervention is virtually useless. When total daily energy expenditure among industrialized societies is compared to contemporary hunter-gatherer tribesmen, they are similar despite the hunter-gatherers performing markedly more physical activity. The risk of dying is reduced dramatically as time spent being active each day increases towards 1 hour, after which it somewhat flat-lines (no additional benefit).
Why? Shouldn’t being more active increase energy expenditure? According to the additive energy expenditure model, it certainly should (figure 1). However, as Pontzer et al show in their latest publication, the relationship between physical activity and energy expenditure is far more complex. They provide support for the constrained total energy expenditure model in which energy allocation among physiological tasks responds dynamically to long-term shifts in physical activity. As such, daily energy expenditure is maintained within a relatively narrow range.
Five populations hold the answer
In Pontzer’s novel study, total energy expenditure (via double-labelled water), resting metabolic rate (via respirometry), and physical activity (via accelerometers) were measured among 332 adult men and women from five populations across Africa and North America (Ghana, South Africa, Seychelles, Jamaica, and the USA).
Anthropometric measurements explained 52% of the variation in energy expenditure, with fat-free mass being the strongest predictor of how many calories an individual will burn throughout the day. Moreover, the relationship between fat-free mass and daily energy expenditure remained significant after further adjustment for differences in lifestyle (amount of manual labor, study site, and amount of vigorous physical activity).
After controlling for the above variables, it was shown that physical activity accounted for only 7-13% of the variation in total energy expenditure. In accordance with the constrained total energy expenditure model, total daily energy expenditure plateaued around the sixth decile of physical activity levels (figure 2).
To be specific, the point at which increased physical activity had no further impact on daily energy expenditure was ~220 counts per minute (CPM) per day. If we multiply this by the number of minutes in a day (1440), then we see that this is a total activity level of 316,800 counts. Walking 3.5 miles per hour corresponds to about 3866 counts per minute, meaning that the threshold level of physical activity is roughly similar to walking at a 3.5 mph pace for 82 minutes, which is about 4.78 miles.
For everyone who walks more than this, what then explains the similar total energy expenditure? This is a question that the current study cannot answer. We know it isn’t changes in resting metabolic rate, as that was not correlated with physical activity levels. It could be changes in the thermic effect of food, but that doesn’t seem likely. Accordingly, Pontzer et al speculate that it is owed to changes in other forms of physical activity – called activity energy expenditure 1 (AEE1) and AEE2 (figure 3).
Although physical activity requires immediate energy to supply its demands (AEE1), compensatory changes in energy expended on other activities (AEE2) apparently negated the additive effect of additional physical activity on total energy expenditure among individuals about 220 CPM/day. What activities this comprises is not readily apparent, but could be something as simple as not fidgeting as much as normal, or sitting more often than usual.
More fascinating, in my opinion, is the finding that physical activity energy expenditure among persons with a CPM/day of zero is still a whopping 600 kcal – well above the estimated daily cost of standing and fidgeting. It could be that this corresponds to the energy requirements of miscellaneous bodily processes, all of which are reduced when physical activity increases. This reduction, in turn, could therefore improve health by reducing the energy expenditure on inflammation and detrimental immune system activity.
This explanation would certainly line up nicely with data showing the risk of dying is reduced dramatically as time spent being active each day increases towards 1 hour, after which it somewhat flat-lines (no additional benefit). It would also explain why exercise sucks for weight loss and why industrial and hunter-gatherer populations have similar daily energy expenditure despite the latter being far more active and far healthier.
Exercise for health, not weight loss
The constrained model of energy expenditure is an appealing concept. Yet, it is based on cross-sectional data and longitudinal studies in which these analyses are repeated as physical activity was increased over several months are needed. Further work investigating what comprises AEE2 would also provide insight.
It appears that physical activity accounts for only 7-13% of the variation in total energy expenditure after controlling for anthropometric differences between individuals. This is an average, and a quick glance at the tiny grey dots in figure 2 that each represent a single individual show a great level of variation among individuals. So really it all comes back down to you and how you respond to being more active.