Turn down the thermostat to battle the obesity epidemic


As of September 2017, new Sweat Science columns are being published at www.outsideonline.com/sweatscience. Check out my bestselling new book on the science of endurance, ENDURE: Mind, Body, and the Curiously Elastic Limits of Human Performance, published in February 2018 with a foreword by Malcolm Gladwell.

- Alex Hutchinson (@sweatscience)


Here’s one I hadn’t heard before:

Increases in winter indoor temperatures in the United Kingdom, United States and other developed countries may be contributing to rises in obesity in those populations, according to [University College London] research published today.

The claim comes from a new paper in Obesity Reviews (press release here), and on the surface, at least, seems plausible.

Reduced exposure to cold may have two effects on the ability to maintain a healthy weight: minimizing the need for energy expenditure to stay warm and reducing the body’s capacity to produce heat.

In particular, the authors note, exposure to cold is thought to stimulate the production of the famous but elusive “brown fat” that burns calories to generate heat. The paper itself goes into great depth about the proposed mechanisms and the expected effect size. Two graphs that I found interesting; first, indoor temperature trends in the U.S. and U.K.:

indoortempI love that U.K. bedroom temperature data! I lived in an unbelievably leaky apartment in Montreal for a couple of years — it was the former “coach house” behind a grander building, so it had no insulation. We were paying for our own heat, and had no money, so we kept the thermostat in the living room down about 13-15 C, and didn’t bother heating the bedrooms, kitchen or bathroom. We kept a “guest blanket” on the couch for anyone foolish enough to visit. And both my roommate and I were pretty skinny throughout those two years…


This is the crucial data — 24-hour energy expenditure as a function of temperature. Could it make a difference? Maybe. But before anyone gets too upset about this, I should note that even the authors of the paper aren’t proposing that this is the cause of obesity — they’re just suggesting it could be one of the many contributing factors.

17 Replies to “Turn down the thermostat to battle the obesity epidemic”

  1. I haven’t seen the original paper, but based on the graphs you’ve provided this seems like a pretty small effect. The US bedrooms appear to have risen about 1 degree and the living rooms look unchanged to me. The slopes of the lines in the second graph seem to average about -0.1 MJ/day/degree. So, for a one degree increase in ambient temperature the reduction in daily energy expenditure would be about 24 kcal. Not much. Do the authors of the paper suggest that living in warmer climates has the same effect? If so, shouldn’t the obesity rate be less the colder the climate? This shouldn’t be too hard to examine. It might show a pretty good fit in the US (I recall seeing a chart with higher obesity rates in the south).

  2. Interesting points, Brian. My initial summary of the paper was a little cursory, so I owe it to the authors to expand on a couple of points. That being said, I just went back and re-read the paper, and can’t find any reference to the point you raise about how this effect would play out in warmer climates. Certainly, you’d think if cold were key the epidemiological data as a function of latitude would be pretty clear.

    Regarding your estimate of 24 kcal per day, I’d say that would actually be pretty significant — IF those calories weren’t offset by compensatory mechanisms (i.e. the cold makes you hungrier so you eat 24 extra kcal). That number of calories per day would add up to a difference of 25 pounds after a decade! One of the papers cited in the review makes a more modest claim: “exposure to mild cold for 10% of the time over 10 years could be equivalent to an 8 kg difference in body weight, if other factors such as energy intake and external insulation remained equal.”

    But the calculations derived from the graph I included are only half the story: that’s how much energy you burn through “thermogenesis” — your body’s mechanisms for generating heat in the cold (shivering is the most obvious, though it’s not necessary). The other half is that cold seems to trigger the accumulation of brown fat, which also has the ability to generate heat and burn calories. Another study cited in the review used PET scans to measure brown fat levels in different temperature conditions, and concluded that brown fat triggered by mild cold could burn 4.4 kg of fat per year.

    Another point that I found interesting is that there’s evidence that cold-induced thermogenesis (the first mechanism I mentioned above) triggers the same pathways as diet-induced thermogenesis, the calories you burn after eating a meal. Some animal studies suggest that if you’re adapted to mild cold, you end up burning more calories after each meal because your thermogenesis response is amped up.

    The real hole in all these arguments (in addition to the question of warmer climates that you already mentioned) is whether adaptive responses kick in. For the most part, our bodies are pretty finely tuned to stay at the same weight, so if you sit in a cold room, your appetite should go up to compensate; if you sit in a warm room, you should be less hungry. So the paper concludes that it’s “biologically plausible” that indoor heating trends are contributing to obesity, but for now it remains just a hypothesis.

  3. Clothing and blankets are a confounding variable that would have to be controlled for before any conclusion can be drawn.

  4. ‘Confouning variable’ is not really the right term in my previous post, since clothing and blankets will probably mask any relationship between the dependent and independent variable. It sounded good, though.

  5. “And both my roommate and I were pretty skinny throughout those two years…”
    And you weren’t before and after that? =D

  6. RH: good point, of course. In the second graph posted above, all but one of the studies shown used “standard clothing” — so the participants weren’t allowed to throw on an extra sweater as the room got colder. The one study that did allow participants to adjust their clothing (Warwick and Busby) is — not coincidentally — the line with the flattest slope in the graph.

    barnee: Hmmm, speaking of confounding variables… 🙂

  7. The data don’t actually support 24 kcal per day since only the bedrooms have warmed up. 8 kcal per day is closer to what the data support. Yes, I know this still means 380g of fat in one year, but only if you believe that energy consumed and energy expended are independent of one another.

    A nice little thought experiment is to imagine drinking one extra liter of cold water every day. Let’s say the water starts at 4 C and your body heats it up to 38 C before it passes out of your system. This requires 34 kcal of energy (by definition). Done every day this should lead to weight loss of 3.5 pounds per year. Does it? Or alternatively, does it prevent a weight gain of 3.5 pounds? Not very likely in my opinion, again because I don’t believe that energy expended and energy consumed are independent of one another.

  8. AG – I think we’d expect it to go the other way around, no? Fat keeps people warm, so they’d be inclined to turn the thermostat down. On the other hand, being sedentary makes people colder, so that could make them turn the heat up. More likely, I think, is that rising prosperity and better technology and insulation have allowed more people to keep themselves warmer.

    Brian – Okay, but given your conclusion, is there any amount of calories burned (by being cold, or exercise, or whatever) that you think would lead to a change in body mass? It seems like your argument is independent of whether it’s 8 kcal, 24 kcal, or 500 kcal per day.

  9. I am inclined agree with Brian, but for slightly different reasons.

    I don’t think intake and expenditure are as closely linked as Brian supposes, but the correlation between room temperature and energy expenditure is, under normal (as opposed to laboratory) circumstances, influenced by so many other variables (for one: not being home all day) that a very slight theoretical effect may be easily offset, or not come into being at all, because of all the other variables. It is easily lost in the ‘noise’.

    The example with the cold water on the other hand, seems like a neat, straightforward mini-liposuction. Having a belly full of water may even lessen calorie intake.

    The only downside I can think of is that, if you drink normally besides the cold water and you retain al that fluid, you may get heavier because you carry a kilo of water to lose 380g.

  10. “On the other hand, being sedentary makes people colder, so that could make them turn the heat up. ” I think this is significant and I find it hard to dismiss.

  11. The amount of energy our bodies must expend to thermoregulate depends on our environmental temperatures, among other things, and that may contribute in some small way contribute to our body weight. When you have such an idea that is so directly derived from first principles, confirmation is as easy as simply finding confirmatory correlations, but the exercise is hardly necessary unless you’ve got the data to develop predictive models to determine whether the reality matches the first principle predictions. If not, then you have some potentially interesting science to do in finding the sources of variation. I hope when you say the paper went “… into great depth about the proposed mechanisms and the expected effect size…,” that the rest of the paper summarized an effort to do just that. Just the plots you show above are otherwise interesting only as hypothesis development or funny trivia about how cold British bedrooms are kept! 🙂

Comments are closed.