Home > Uncategorized > The more you exercise, the less diet matters

The more you exercise, the less diet matters

June 13th, 2011

This week’s Jockology column in the Globe and Mail takes a closer look at some new results (which I blogged about back in April) from the long-running National Runners’ Health Study:

At a public debate in May on the relative importance of exercise and diet in battling obesity, Yoni Freedhoff began his opening arguments with some basic physics.

“There’s no debate about whether the laws of thermodynamics exist,” said Dr. Freedhoff, the medical director of the Bariatric Medical Institute in Ottawa. Energy can’t be created or destroyed, so weight loss ultimately depends on burning more calories than you consume. But which side of that equation should you focus on? [READ THE FULL ARTICLE]

The basic finding of the new study is that the more you exercise, the weaker the link between diet and weight. I exchanged a few e-mails with Yoni Freedhoff (of Weighty Matters fame) about this idea, and his initial reaction was that the findings could be interpreted as simply the result of calories burned while running. After all, running 8 km per day (as the “top” group in the analysis does) burns quite a few calories. I tend to think that there’s more going on here (as I explain in the article), but I’d certainly be interested in hearing what others think. Am I making too big a deal about something that’s completely obvious?

  1. Phil Koop
    June 13th, 2011 at 14:02 | #1

    If, on the one hand, Gary Taubes claims that if you exercise more you’ll eat even more than you burn, and on the other, Yoni Freedman suggests that marginal exercise exactly matches marginal calorie deficit, then how can the issue be “completely obvious”? Unless you look at the data, of course. So keep up the good work!

  2. Seth Leon
    June 13th, 2011 at 18:08 | #2

    Just looking at the graph and not having read the report or article it looks to me like there might be something more than a ‘weaker the link between diet and weight’ as exercise goes up.

    With increasing exercise the relationship between meat intake and waist size change decreases and actually goes negative with the highest exercise group. The relationship with fruit intake looks pretty flat although there appears to be a slight trend in the opposite direction of meat intake.

    Looks like there might be some interaction here. The high exercize group is 8k which would be low for marathon training. I wonder if the reduction in waist size with meat consumption would continue for higher volume runners? I guess I should read the paper.

  3. June 13th, 2011 at 21:28 | #3

    Alex,

    For Doctors/Scientists to say you can eat what you like as long as you burn it through exercise is not a great message. I am a very active individual and can burn most of what I consume but that “indestructable’ ideology may end up hurting me in the long run as I eat more treats and fatty foods to satisfy my never ending cravings.
    Therefore, someone that eats a healthier diet, but does not excercise, may in fact be in better overall condition than I am.

  4. alex
    June 14th, 2011 at 00:07 | #4

    Forwarding a comment from HVR, posted on another thread but related to this discussion:

    “Hi, Just read your article in the Globe. Very interesting and helpful. Curious about one thing though: “It’s well established that the more meat you eat, the heavier you tend to be; similarly, the more fruit you eat, the lighter you tend to be.” – Do you know what study this draws on? I would like to read the whole thing. Thanks for the info. Can’t wait to read the book.
    H”

    The new study by Paul Williams that I describe in the article references six different papers for the claim about the relationship between meat and fruit consumption and prospective weight gain/loss. Here are links to two that have the full text freely available:
    http://jn.nutrition.org/content/132/6/1335.long
    http://www.ncbi.nlm.nih.gov/pubmed/9184500

    Looking through the references, the relationship seems to be much clearer for meat than fruit (something that Williams acknowledges). But I should emphasize again that the point of the study wasn’t to single out meat and/or fruit as uniquely bad and/or good. They’re simply convenient markers that stand in for broader dietary patterns.

    It may be that (to pick a ridiculous example) meat has no effect on health, but people who eat lots of meat also tend to eat lots of barbecue sauce, and barbecue sauce is what causes weight gain. But this study doesn’t really care what the mechanism is: it’s just trying to show that, whatever the mechanism, the links between dietary patterns and weight are weaker at higher levels of exercise.

  5. alex
    June 14th, 2011 at 00:17 | #5

    @Phil Koop
    Thanks, Phil! In glancing through the responses to the article posted on the Globe site, I see many people insisting that the results are trivial and obvious — but many of these claims are mutually contradictory, which reassures me that it’s not at all obvious. :)

    I should also clarify that Dr. Freedhoff was just giving me his off-the-cuff reaction in response to an informal e-mail from me, without having read the study. While I do think there’s more going on than just calorie burn, his response did prompt me to think more carefully about the data — and consequently to wonder about some of the points that Seth brings up.

  6. alex
    June 14th, 2011 at 00:36 | #6

    @Seth Leon

    Yes, I was wondering if anyone would bring up the anomalies in the data!

    “With increasing exercise the relationship between meat intake and waist size change decreases and actually goes negative with the highest exercise group.”

    And the same (or opposite, actually) thing happens with the fruit: the relationship changes sign at the highest levels of exercise. The title of the paper is “Exercise Attenuates the Association of Body Weight with Diet in 106,737 Runners.” If running was really just “attenuating” the relationship, we’d expect both relationships to tend toward zero at the highest levels of exercise. Instead, they seem to overshoot zero, leading us to the somewhat odd conclusion that, for people running more than 8K a day, the best weight loss strategy is to cut out fruit and gorge on meat. (Gary Taubes would agree with this, of course, but that doesn’t explain why the opposite appears to hold true for those exercising less.)

    Of course, it’s possible to come up with explanations for this offset. Perhaps some disproportionate fraction of the most active runners are the type of people who’ve always been skinny no matter what they ate, so they’ve had no incentive to adopt “healthy” eating patterns. Their high meat and low fruit consumption could pull the data past zero. But this is exactly the kind of tap-dancing you want to avoid in interpreting epidemiological data, because it’s possible to justify pretty much anything if you already think you know the answer!

    “I wonder if the reduction in waist size with meat consumption would continue for higher volume runners?”

    I did ask Williams if he saw any plateau in the relationship at high levels of exercise. He said no, but the >8K/day group was the highest stratification in his data set.

    Ultimately, my take is that the sign change is likely an artifact of the fact that meat and fruit are just markers of more complex dietary patterns (i.e. this study ISN’T intended to show that fruit and meat CAUSE weight gain and loss). However, this anomaly does make me less confident in the conclusions.

  7. alex
    June 14th, 2011 at 00:44 | #7

    @Larry Bradley
    “For Doctors/Scientists to say you can eat what you like as long as you burn it through exercise is not a great message.”

    I don’t think any doctors or scientists are saying that. If you read my Globe article, the last line (and the main message) is: “Instead, the results suggest that, in the great “forks v. feet” debate, trying to fix one without considering the other simply doesn’t make sense.”

    Still, I understand your point. The question of “message” is a deep and longstanding debate in public health. I tend to side with those who argue that the best and most honest thing we can do is present all the information we have, and give people the tools to make their own decisions. The counterargument is that many people don’t have the necessary background and context to be able to interpret information properly — so, for example, they’ll read my article in the Globe and think the message is that diet doesn’t matter if you exercise. But frankly, if that’s what they conclude, it’s clear that they didn’t read the whole article, so I don’t feel too guilty!

  8. Jill
    June 14th, 2011 at 19:43 | #8

    Hi Alex,
    I must say, the increase in weight with the serving of fruit for the 8km/day + runners still leaves me confused….understanding it’s not a causal study, but still….

    I have Williams’ article “asymmetric weight gain and loss from increasing and decreasing exercise” (medicine & science in sprots). I’m not sure if this is the same cohort,but wondering if you are familiar with this article. One conclusion is that for those runners (high mileage runners, I believe) who decrease their mileage and gain weight can’t lose that said weight gain if their running is increased to match their former mileage (clear?). Can you explain why this is? My understanding is that for women to lose weight running, they needed to run about 48km/wk before they noticed a difference. I don’t believe they kept food diaries, so not sure how calorie intake would influence this as well. Are these the kinds of things you write about in your new book? thanks!

  9. RH
    June 16th, 2011 at 08:20 | #9

    For the heck of it, let’s argue that a low ‘meat elasticity of the waist’ at higher mileage is counter intuitive.

    The assumption is, that one daily serving of meat represents one unit of energy and one unit of energy translates to one standard portion of mass.

    Obviously, if two people of equal size with a waist, say, 80 cm, both decide to eat one extra portion of meat(=energy=mass), this would add the same mass to both of them, no matter how many km’s they run in their daily lives. So both should see the same increase in waist size with one extra serving.

    Actually, the abstract sais that the runners are skinnier than the sedentary group, so, appartenly the group that runs >8 km has a smaller waist line than the <2 km group. Obviously, one standard volume of mass adds less cm's to the circumference of someone with a big mass and thus a big waist line than to that of a thin person. So, actually the group of sedentary 8km athletes.

    That is, unless there is more at work here, like a change in metabolism.

  10. RH
    June 16th, 2011 at 08:22 | #10

    (sorry)

    For the heck of it, let’s argue that a low ‘meat elasticity of the waist’ at higher mileage is counter intuitive.
    The assumption is, that one daily serving of meat represents one unit of energy and one unit of energy translates to one standard portion of mass.
    Obviously, if two people of equal size with a waist, say, 80 cm, both decide to eat one extra portion of meat(=energy=mass), this would add the same mass to both of them, no matter how many km’s they run in their daily lives. So both should see the same increase in waist size with one extra serving.
    Actually, the abstract sais that the runners are skinnier than the sedentary group, so, appartenly the group that runs >8 km has a smaller waist line than the <2 km group. Obviously, one standard volume of mass adds less cm's to the circumference of someone with a big mass and thus a big waist line than to that of a thin person. So, actually the group of sedentary people should have a lower waist increase per serving 8km athletes.
    That is, unless there is more at work here, like a change in metabolism.

  11. Evan Plaice
    June 16th, 2011 at 18:11 | #11

    The explanation for ‘why’ runners running distances greater than 8k may be very simple.

    When you run longer distances, if you burn through all of your available energy your body switches to metabolizing it’s reserves. Without a sufficient store of available fat your body kicks into catabolism (Ie, it starts to metabolize muscle).

    The distance runners on a higher meat diet are capable of restoring more muscle mass due to the larger quantities of meats (and essential amino acids) that they’re taking in.

    Those on the higher fruit diet are not only burning less energy (due to decreased muscle mass from catabolism) but they are storing more fat (due to the higher levels of fructose intake).

    There is no single relationship between running longer distances and waist size, there are actually two.

    1. The more muscle mass you have, the higher your energy requirements are. Running greater distances (>8km) requires a higher protein intake to repair/maintain muscle mass after losses incurred through catabolism (due to lack of available energy during extended workouts). With a lack of available fat energy stores, increased protein intake becomes even more important.

    2. Even with a low GI source of sugar (fructose) such as fruit, fructose is a high energy food that will be stored as fat if it isn’t immediately used. Having an increased amount of muscle mass will ensure that your body has a high enough daily caloric expenditure to consume the increased amount of energy (and thus stop the excess from being stored as fat). Without sufficient muscle mass, a high sugar diet is not advised unless gaining fat mass is the desired end result.

    I could be wrong but from everything I know and have read, this assessment seems to make the most sense.

    BTW, using ‘waist line’ to measure fat mass is pretty useless. Different people store fat differently on their bodies. Measuring ‘waist lines’ is just for vanity. Although it works for making macro measurements, it’s pretty useless when it comes to measuring small variations in body fat.

  12. alex
    June 19th, 2011 at 23:38 | #12

    @RH
    I like that line of reasoning! The conclusion, I guess, is that there IS something more at work (like a change in metabolism). And I wasn’t just wasting everyone’s time by posting this study! :)

  13. alex
    June 20th, 2011 at 00:09 | #13

    @Jill
    Thanks for the comments, Jill. I agree that the weight results for people running 8km/day are confusing. I don’t have an explanation. And, as unsatisfying as it is to say this, I think it would be a mistake to get too focused on finding a real, physical explanation for an anomaly like this. Given the size of the data set and the various statistical manipulations that were used to produce these numbers, there are lots of places where subtle distortions could have crept in. In judging the “hierarchy of evidence,” we can be more confident about the trend (diet has less influence in heavy exercisers) than the specific numbers themselves.

    As for Williams’ “asymmetric weight gain and loss” paper, I do indeed cover that (briefly) in Cardio and Weights! Again, Williams’ epidemiological data shows us what’s happening, but doesn’t tell us why it happens. But the assumption is that is has to do with the way the body attempts to keep a stable weight. In general, the body has a bunch of homeostatic mechanisms that keep your weight roughly the same. If you miss a few meals, your metabolism slows down in order to prevent weight loss; if you stuff yourself, your metabolism ramps up to burn off the extra calories (if everything is working properly — this is one of the responses that malfunctions in some obese people). The problem is that the body is more vigilant about preventing weight loss than weight gain. In evolutionary terms, wasting away was always a bad thing, whereas putting on a few extra pounds to provide some extra energy stores was sometimes helpful. That’s basically what’s going on with the asymmetric weight gain/loss.

    In terms of women needing to run more than 48K/week before they noticed a difference — I don’t think that’s quite right! In fact, the study says: “Above 32 km/wk in men and 16 km/wk in women, the effects of training and detraining are comparable, such that weight gains and losses associated with changes in exercise levels are probably reversible.” In other words, there was no asymmetry at all for women running more than 16km/week!

  14. alex
    June 20th, 2011 at 00:16 | #14

    @Evan Plaice
    Thanks for the thoughts, Evan — interesting suggestions. I’m not sure I buy the argument about protein intake and catabolism, though. The runners in this study were maxing out at not much more than 8K/day. Most people can store enough carbohydrate to take them ~30K, so it seems unlikely that many of these people were catabolizing muscle on a regular basis!

    “BTW, using ‘waist line’ to measure fat mass is pretty useless. Different people store fat differently on their bodies. Measuring ‘waist lines’ is just for vanity.”

    For what it’s worth, the study outcomes were BMI, waist circumference, chest circumference and hip circumference, so they did indeed look at fat storage in various places.

    Having said that, I don’t agree that “measuring waist lines is just for vanity.” In fact, waist circumference is probably the single most important measurement related to fat storage that you can take, since it tells you about visceral fat, which is a much greater health concern that other areas of fat storage. Here’s an excerpt from an article I wrote a few years ago:

    [...] Still, BMI remains a fairly blunt tool with which to judge your body shape.

    “It’s most useful in population studies,” says Travis Saunders, a graduate researcher at the Children’s Hospital of Eastern Ontario Research Institute in Ottawa who blogs about obesity research at ObesityPanacea.com. “But if you try to apply it to individuals, it doesn’t work.”

    That’s because where you store fat is as important as how much you have. Fat in the abdominal region, particularly the visceral fat that accumulates between internal organs rather than fat stored just beneath the skin, is particularly problematic. In contrast, Mr. Saunders says, fat on the hips, buttocks and lower body appears to be less of a concern.

    For that reason, many doctors now measure waist circumference as a proxy for visceral fat. Ideally, men should be less than 102 centimetres (40 inches) and women should be less than 88 centimetres (35 inches), Mr. Saunders says. [...]

  15. RH
    June 21st, 2011 at 06:57 | #15

    @alex
    Thanks and sorry for the typo’s. I should read my contributions before I post them.

    Speaking of elasticities, perhaps the proper way to think about it is sort of an economic equilibrium equasion.

    To maintain weight, your food intake has to equal your energy expenditure. Your expenditure is the energy you spend on running plus your resting metabolism. Your energy expenditure on running is (for some proper unit and disregarding speed, drag etc.) your daily mileage times your mass. Your resting metabolism is said to scale with your mass as a 3/4 power. That would make the equilibrium equation:

    Food = mileage * mass + mass^3/4

    If we model people as cylinders of height 1, the circumference (waist line) scales with the mass (for some proper unit) as a square root. Expressing mass in circumference would make the equation

    Food = mileage * circumference^2 + circumference^3/2

    Obviously, the higher the mileage, the steeper the line that links food intake on the y-axis to circumference on the x-axis. So, for a given food intake, and a high mileage, you’d be stuck at a lower equilibrium weight/mass/circumference.

    Put differently, If you are sedentary and you increase your daily food intake by 1, you’d have to rely solely on your resting metabolism to bring your intake and expenditure into equilibrium, which means increasing mass. If you run daily, an increase in mass burns much more food, since your resting metabolism increases and you have to carry the extra weight across many kilometres.

    Sorry, but it’s completely obvious again, though I can’t explain the negative elasticity at >8 km.

  1. June 13th, 2011 at 13:38 | #1
  2. June 15th, 2011 at 23:47 | #2