The 30 percent figure exists, and two attempts to reproduce it have failed#
Drinking water does not meaningfully raise your metabolism. What raises it, by a trivial and entirely predictable amount, is warming the water — and the celebrated study behind the claim has now been tested twice by another laboratory and has not held up.
The original is easy to find and hard to forget: in 14 healthy normal-weight adults measured by whole-room indirect calorimetry, drinking 500 ml of water increased metabolic rate by 30 percent, peaking 30 to 40 minutes in, with a total thermogenic response of about 100 kJ — roughly 24 calories. The authors noted that about 40 percent of the effect came from warming the water from 22 to 37 °C, and extrapolated that 2 litres a day would add about 400 kJ1. Even taken at face value, the headline "30 percent" is a 96-calorie-a-day claim, not the metabolic overhaul it gets quoted as. And it does not appear to be taken at face value by the people who tried hardest to reproduce it.
The pillar on what metabolism actually is explains why no beverage can move total expenditure much. This piece is about a case where the mechanism was specific enough to test — and where two research groups reached opposite answers for a reason you can name.
Two labs, three studies, and a specific point of disagreement#
This is not vague inconsistency. The follow-up work was designed to interrogate Boschmann's proposed mechanism directly, and the design is what makes the split informative.
Test one. Researchers gave healthy young volunteers about 518 ml of distilled water at room temperature, 0.9 percent saline, and a 7 percent sucrose solution as a positive control. Distilled water did not raise energy expenditure (P = 0.34). Saline did not either (P = 0.33). Sucrose did, decisively (P < 0.0001) — proving the equipment could see a real thermic effect when one existed. In a subgroup given water at 3 °C, expenditure rose 4.5 percent over 60 minutes (P < 0.01). Their conclusion: the results "cast doubt on water as a thermogenic agent for the management of obesity"2.
The saline arm is the whole point of that design. Boschmann's proposed mechanism ran through osmosensitive sympathetic activation — the body reacting to a sudden drop in blood osmolality. Isotonic saline removes that stimulus. If the mechanism were real, saline and distilled water should have behaved differently. They behaved identically, and both did nothing.
Test two. Nine years later the same group added the control that had been missing from everyone's protocol: sham drinking. Twenty-seven adults (14 men, 13 women, BMI 18.5 to 33.9) had resting expenditure measured for at least 30 minutes before and 130 minutes after either 500 ml of distilled water at 21–22 °C or a sham drink, crossover design. Water produced marginal increases in resting expenditure — under 3 percent above baseline — that were not significantly different from sham drinking. Respiratory quotient fell after the water, which looks like a shift to fat burning, and it fell to a similar extent after the sham drink too. None of the variability tracked body fat, central adiposity or fat-free mass3.
That is the more damaging of the two results, because the RQ drop is the finding people cite as evidence that water makes you burn fat. Going through the motions of drinking produced the same drop. Whatever it is, it is not the water.
Both null studies come from one laboratory in Fribourg, which is a genuine limitation — a replication base concentrated in a single group is thin, however careful that group is. It declares no competing interests, and the 2015 study was funded internally by its own department.
The part everyone agrees on is thermodynamics#
Strip out the contested mechanism and one uncontested cost remains: your body has to bring cold liquid up to 37 °C, and that costs energy no matter what any calorimeter says about sympathetic tone.
The arithmetic is secondary-school physics, and it is ours, not a published result. Water takes about one calorie per gram per degree Celsius. So:
| Water drunk | Temperature rise to 37 °C | Thermodynamic cost |
|---|---|---|
| 500 ml at room temperature (22 °C) | 15 °C | ~7.5 kcal |
| 500 ml straight from the fridge (5 °C) | 32 °C | ~16 kcal |
| 500 ml near-freezing (3 °C) | 34 °C | ~17 kcal |
| 2 litres a day at room temperature | — | ~30 kcal |
Warming half a litre of water from room temperature to body temperature costs about seven and a half calories. That is the floor, the ceiling, and essentially the whole uncontested finding.
Boschmann's own paper is consistent with this: 40 percent of their 24-calorie response — about 9.6 calories — was attributed to warming, which is the same order as the physics. And it is why Brown's group found something only in the 3 °C arm. Cold water is the one version that has a mechanism nobody disputes, which is also the mechanism at work in cold exposure more generally.
Even there the accounting is loose. A 4.5 percent rise over 60 minutes, applied to a typical adult resting rate, works out to roughly 3 calories on our arithmetic — noticeably less than the ~17 calories the warming itself should cost. The two figures do not reconcile cleanly, which is reason to treat both as approximations rather than a completed ledger.
What the most water-friendly review's own tally shows#
The strongest case for water is usually built from a 2016 qualitative review that catalogued randomized trials by outcome. Its counts are worth reading directly, because they undercut the thermogenesis claim from inside the friendliest possible source.
Of 68 short-term energy-expenditure effects it catalogued, 33 were null, 9 showed lower expenditure after drinking water, and only 7 showed higher. Of 115 weight-change effects from longer trials, 83 were null. The place the counts run strongly one way is fat oxidation — 62 of 74 effects showed an increase — and energy intake, where 56 effects showed water lowering intake against 90 null and 30 higher4.
The review's competing-interests statement reads, verbatim: "J.J.D.S. is an occasional consultant for Danone Research." Danone is among the largest bottled-water companies in the world. That does not make the counts wrong — anyone can check them — but it is the disclosure a reader is owed, and it sits on the side of this literature that most consistently reports benefits. The Fribourg nulls carry no such disclosure, and the trial with the clearest practical result, below, was funded by the US National Institutes of Health.
Read on its own terms, the tally says: expenditure effects are mostly nothing and about as often negative as positive; weight effects are mostly nothing; the reliable signal is on intake.
The effect that does show up on a scale, and its odd footnote#
Forty-eight adults aged 55 to 75 with BMI 25 to 40 were assigned to a hypocaloric diet plus 500 ml of water before each meal, or to the hypocaloric diet alone. Over 12 weeks, the water group lost about 2 kg more — a 44 percent greater decline in weight5. That is a substantial result from a small trial, and it is the real basis for the pre-meal-water advice.
The footnote is the interesting part, and it is in the same abstract. The trial also measured whether a water preload reduced how much people ate at a test meal. At baseline it did: 498 ± 25 kcal with the preload against 541 ± 27 without (P = 0.009), a 43-calorie difference. At week 12 it no longer reached significance: 480 ± 25 against 506 ± 25 (P = 0.069).
So the mechanism the trial proposed for its own result had faded by the end of the trial, while the weight difference persisted. Either the acute effect matters mostly early, or something else — displacing caloric drinks, an adherence cue, the ritual of preparing to eat — carried the rest. The authors say the result "may be due in part" to the acute intake reduction, and "in part" is doing real work in that sentence.
What to do with a glass of water#
The practical advice barely changes, but the reason for it does.
- Drink water before meals if it helps you eat less. A 43-calorie reduction is small, and it is on the intake side of the ledger, where reductions actually stick — the general case for that asymmetry is in metabolism-boosting foods.
- The big win is displacement, not thermogenesis. Water's real caloric power is what it replaces: the juice, the latte, the beer. Those are the calories most often missing from a food log, which is the subject of drinks, oils and bites, and swapping one 250-calorie drink for water beats every thermogenic effect in this article combined by an order of magnitude.
- Cold water is a curiosity, not a strategy. Sixteen calories a bottle at the physical maximum, and it helps with hunger management only at the margins.
Drink water because you are thirsty, because it is free, and because every glass of it is a glass of something else you didn't drink. The metabolism part was never the reason.
FAQ#
How many calories does drinking two litres of water a day burn?#
About 30 calories, and only from warming the water to body temperature — our arithmetic from the specific heat of water. The original study reporting a 30 percent rise in metabolic rate put the total at roughly 24 calories per 500 ml, but two later studies from another lab found distilled water produced no significant rise at all: under 3 percent above baseline and no different from sham drinking. Treat 30 calories as a generous ceiling.
Does drinking cold water burn more calories than room-temperature water?#
Slightly, and it is the one part of the claim with an undisputed mechanism. Warming 500 ml from 3 °C rather than 22 °C roughly doubles the thermodynamic cost, from about 7.5 to about 17 calories. Measured directly, water at 3 °C raised energy expenditure 4.5 percent over 60 minutes, while room-temperature water and saline raised it not at all. A daily habit of ice water is worth a few dozen calories at most.
Does drinking water before meals actually help you lose weight?#
There is one reasonable trial supporting it: 48 adults aged 55 to 75 on a hypocaloric diet lost about 2 kg more over 12 weeks when they drank 500 ml before each meal. The acute effect on how much they ate was 43 calories per meal at the start of the trial and no longer statistically significant by week 12 — so the mechanism is less settled than the result. It is a cheap, harmless thing to try, not a proven lever.
Sources#
- Boschmann M, Steiniger J, Hille U, Tank J, Adams F, Sharma AM, Klaus S, Luft FC, Jordan J. Water-induced thermogenesis. J Clin Endocrinol Metab. 2003;88(12):6015-6019.
- Brown CM, Dulloo AG, Montani JP. Water-induced thermogenesis reconsidered: the effects of osmolality and water temperature on energy expenditure after drinking. J Clin Endocrinol Metab. 2006;91(9):3598-3602.
- Charrière N, Miles-Chan JL, Montani JP, Dulloo AG. Water-induced thermogenesis and fat oxidation: a reassessment. Nutr Diabetes. 2015;5(12):e190.
- Stookey JD. Negative, null and beneficial effects of drinking water on energy intake, energy expenditure, fat oxidation and weight change in randomized trials: a qualitative review. Nutrients. 2016;8(1):19.
- Dennis EA, Dengo AL, Comber DL, Flack KD, Savla J, Davy KP, Davy BM. Water consumption increases weight loss during a hypocaloric diet intervention in middle-aged and older adults. Obesity (Silver Spring). 2010;18(2):300-307.
