Dehydration and change in body mass: not linked after all?


As of September 2017, new Sweat Science columns are being published at 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)


I (and everyone I know) have always taken this for granted: if you weigh yourself before and after a workout, the difference tells you how much fluid you lost to sweat (after correcting for any water that you drank during the workout). If you lose more than about 2% of your bodyweight, dehydration will impair your performance. That’s what the ACSM guidelines on hydration say:

If proper controls are made, BW [bodyweight] changes can provide a sensitive estimate of acute TBW [total body water] changes to access hydration changes during exercise.

But it turns out there’s actually a hot debate currently raging in the literature about this. The latest salvo just appeared online in the British Journal of Sports Medicine, from researchers in South Africa (Pretoria, not Cape Town, though Tim Noakes is indeed listed as a co-author). They studied 18 soldiers doing a 14.6-kilometre march while drinking “ad libitum” (however much they wanted), and took careful measurements of a whole series of physiological parameters. One of those parameters is “total body water” — the sum total of all water stored in the body, typically totalling about 60% of body mass — which they  measured using radioactive tracers. When we talk about hydration and dehydration, that’s what we’re really talking about: is there sufficient TBW to ensure that all the tissues and cellular processes in the body are working optimally.

The surprise: the subjects lost 1.98% of their body mass on average, but their total body water stayed roughly the same (actually, it increased by 0.53% on average). They drank 0.85 litres per hour, but sweated out 1.289 litres per hour. In other words, they were losing fluid — so how did their total body water stay the same or increase?

Some of the possible explanations are explored in this 2007 paper by Ron Maughan. One is “metabolic water”: when your body converts fat or carbohydrate into ATP, it release some water as part of the sequence of chemical reactions (one estimate is that it releases 0.13 g per calorie burned). A more significant possibility, especially for endurance athletes, is that every gram of glycogen you store ties up an estimated 3-4 g of water. A marathoner who carbo-loads and packs in 450 grams of glycogen, for example, could in theory have 1.35 kg of “hidden” water that will gradually be released into the body as carbohydrate stores are released during exercise.

So what this study claims is that these soldiers were sent out on a march and told to drink however much they wanted; they lost 2% of their body mass, but weren’t dehydrated. Their  interpretation: the body’s thirst mechanism is built to maintain the osmolality (the concentration of “stuff,” essentially) in the blood and tissues, which was indeed preserved in this experiment.

Is the debate over? Far, far from it. For one thing, a laboratory experiment at Penn State published last year found exactly the opposite — that the amount of weight you lose during exercises correlates perfectly with the loss in total body water. How to reconcile these diverging views? I’m not sure, but I’m digging into the literature and doing some interviews for an upcoming article.

13 Replies to “Dehydration and change in body mass: not linked after all?”

  1. Interesting. I had thought that the binding of water during the storage of carbohydrate was an established fact. If that were true, it would follow that if you deplete a significant amount of stored carbohydrate, and ingest only water, then unless your weight drops you must necessarily end up more hydrated than when you started. That is, unless your body has a hidden reservoir to store the excess water. What is the answer to this paradox?

  2. Hi Phil: Yes, the binding of water during carbohydrate is established fact. What’s not agreed on is (a) exactly how much water is stored per gram of carb (quite a wide divergence in opinions), and (b) whether the water that is stored should be counted as part of “total body water” before it’s released. The ACSM position statement linked in the post says:

    “The precise fate of water liberated as glycogen is utilized is unknown, but the fact that any water bound to glycogen is part of the starting TBW pool suggests it is of little potential consequence to fluid intake recommendations.”

    I have a hard time understanding why that should be true. From the perspective of hydration and performance, fluid is only relevant if it’s available for essential bodily processes. It seems to me that having fluid bound to glycogen is the equivalent of swallowing a zip-lock baggie full of water. Sure, the water is inside your body, but it’s not playing a role in hydration until the bag is unzipped!

  3. Exactly. And once the baggy is unzipped, who cares what the “precise fate” of the water is? Unless it gets poured into some other, hypothetical, baggy, it is out there, available for metabolic uses. And surely the burden of proof must be on those conjecturing the imaginary second baggy?

    Anyway, the bottom line, alas, is that the bit that you have trouble understanding is the same one that puzzles me! Good luck cracking this riddle, and keep us posted!

  4. Really interesting article, Ian — thanks for posting it. Lots of critiques that, unfortunately, can be applied to both of the studies discussed above. (One of the studies was funded by a sports-drink manufacturer; both are part of a polarized debate with a lengthy history, in which both sides might be more interested in proving themselves correct than in figuring out the truth, at least subconsciously.)

  5. very cool study. I assume that the subtext to all of this is that hydration is somehow important to either performance or health? Which it obviously is to an extent, but I’ve found that some of Noakes’ most convincing arguments have been related to how overblown the importance of hydration/dehydration are among the general public. Has anything related to the debate of the fundamental importance/unimportance of being hydrated come out in your interviews?

  6. Hey Travis — yes, it’s all part of the same debate. This new study (and a few others by Noakes’s group, one of which is currently in review) are essentially their attempts to explain why hydration (in the sense of “make sure you replace all the fluids you lose”) isn’t important. Noakes and his allies acknowledge that yes, if you just drink based on thirst, you’ll lose some weight (and possibly even some total body water) during exercise. But he argues that your body doesn’t really care how much it weighs, it only cares about your plasma osmolality, which determines whether the optimal homeostatic fluid balance is maintained in your cells and tissues.

    So in that sense, hydration IS important in their view — but you’re perfectly fine just listening to your thirst, even if that means that you 2% or significantly more of your body mass, because (as the new studies show) your total body water and more importantly your plasma osmolality are maintained.

  7. Hey Alex: excellent and thought provoking review – i actually commented on the article [see 2nd blog, dated Nov 29 – for my “take” on the local running scene and your insights/conclusions. This also begs the question of “when” to drink [as well as how much] in endurance events.

  8. @JCH
    Hey JCH — Sorry for the long-delayed reply — and thanks for the shout-out in the blog entry. The “when” question is a tough one. The obvious answer is “whenever you’re thirsty,” but in races you generally have to drink only at specific drink stations, so sometimes you have to anticipate your needs to avoid ending up thirsty between aid stations. Definitely a topic that needs further research.

Comments are closed.