0021-972X/07/$15.

00/0 The Journal of Clinical Endocrinology & Metabolism 92(8):3334 –3337

Printed in U.S.A. Copyright © 2007 by The Endocrine Society
doi: 10.1210/jc.2006-1438

BRIEF REPORT
Water Drinking Induces Thermogenesis through
Osmosensitive Mechanisms
Michael Boschmann, Jochen Steiniger, Gabriele Franke, Andreas L. Birkenfeld, Friedrich C. Luft, and
Jens Jordan

Downloaded from https://academic.oup.com/jcem/article/92/8/3334/2598681 by guest on 05 February 2021

Franz-Volhard Clinical Research Center and Helios-Klinikum-Berlin, Charité Campus Buch, Universitary Medicine Berlin,
D-13125 Berlin, Germany

Context: Recently, we showed that drinking 500 ml water induces and 50 ml water had no effect. Heart rate and blood pressure did not
thermogenesis in normal-weight men and women. change in these young, healthy subjects.

Objective: We now repeated these studies in a randomized, con- Conclusions: Our data exclude volume-related effects or gastric
trolled, crossover trial in overweight or obese otherwise healthy sub- distension as the mediator of the thermogenic response to water
jects (eight men and eight women), comparing also the effects of 500 drinking. Instead, we hypothesize the existence of a portal osmo-
ml isoosmotic saline or 50 ml water. receptor, most likely an ion channel. (J Clin Endocrinol Metab 92:
3334 –3337, 2007)
Results: Only 500 ml water increased energy expenditure by 24%
over the course of 60 min after ingestion, whereas isoosmotic saline

E XTRACELLULAR HYPEROSMOLARITY causes cell

shrinkage, whereas extracellular hypoosmolarity
causes cell swelling. Osmotically induced cell volume
the same thermogenic response in obese as in normal-weight
subjects and whether it is related to gastrointestinal disten-
tion, hypoosmotic stimulation, or both.
changes modulate important cellular functions, both in vitro
and in vivo, which may influence the organism directly or Subjects and Methods
indirectly through homeostatic reflex pathways (1, 2). Subtle We studied overweight or obese otherwise healthy subjects on no
changes in osmolarity occur in daily life even in healthy medications [eight men, age 29 ⫾ 2 (range 20 – 42) yr, body mass
people. For example, water drinking changes osmolarity in index (BMI) 33 ⫾ 1.3 (range 29 –39) kg/m2; eight women, age 33 ⫾ 2
the portal circulation (2) and also acutely activates the sym- (range 27–39) yr, BMI 30.3 ⫾ 0.4 (range 29 –32) kg/m2]. The institutional
review board approved all studies. Written informed consent was ob-
pathetic nervous system (3, 4). Indeed, water drinking raises tained before study entry.
both plasma norepinephrine concentrations and muscle All subjects included in our study had a sedentary lifestyle. Never-
sympathetic nerve activity (3–5). The sympathetic activation theless, we asked our subjects to abstain from strenuous physical ex-
drives a profound pressor response in autonomic failure ercise 48 h before testing. Furthermore, subjects were asked to abstain
patients (4, 6). The response is abolished with pharmacolog- from smoking and consuming alcoholic drinks and caffeine-containing
beverages or snacks 48 h before testing. They were also advised to take
ical ganglionic blockade (4). Moreover, metabolic rate in- a light dinner no later than 13 h before testing. Subjects did not eat 12.5 h
creases almost 30% in healthy, normal-weight subjects (7). and did not drink 1.5 h before testing. They appeared in our laboratory
The response is attenuated with ␤-adrenoreceptor blockade at 0830 h. Testing was begun at 0900 h after they had emptied the
(7). Only one third of the increase in metabolic rate was bladder. Oxygen consumption (VO2) and carbon dioxide production
(VCO2) were measured by using a respiratory chamber, considering the
explained by the energy demand to warm the water from 22 methodological improvements introduced by Brown et al. (11), to assess
C to body temperature. The metabolic response was atten- changes in energy expenditure and respiratory quotient (RQ ⫽ VCO2/
uated with systemic ␤-adrenoreceptor blockade (7). The in- VO2). The respiratory chamber had a volume of 5800 liters and was
crease in metabolic rate was not reproduced in a recent study calibrated with acetone burning. In earlier tests, we confirmed that the
using a different methodology (8). Cardiovascular sympa- respiratory chamber was able to follow rapid changes in metabolic rate.
When using 20-min integration periods, the time course of the post-
thetic responses were elicited by water but not saline, thereby prandial thermogenic response after ingesting a standardized test meal
implicating osmosensitive mechanisms (5, 9, 10). The objec- was virtually identical over 360 min with the respiratory chamber and
tive of our study was to test whether water drinking elicits a canopy system, respectively (n ⫽ 6). We measured heart rate and
brachial blood pressure every 5 min (Dinamap). Subjects remained
seated throughout the experiment and were continuously monitored
First Published Online May 22, 2007 through a window in the metabolic chamber. After a run-in period of 15
Abbreviations: BMI, Body mass index; RQ, respiratory quotient; TRP, min, resting energy expenditure was determined for 30 min. Thereafter,
transient receptor potential. the subjects ingested 50 ml (control) or 500 ml tap water (pH 7.5; [Na⫹]
JCEM is published monthly by The Endocrine Society (http://www. 1.5 mmol/liter; [Ca2⫹] 3.1 mmol/liter) or 500 ml isoosmotic saline in less
endo-society.org), the foremost professional society serving the en- than 5 min. Testing was conducted on separate days in a randomized,
docrine community. crossover fashion such that each subject was tested with all fluids.

3334
Boschmann et al. • Water and Thermogenesis J Clin Endocrinol Metab, August 2007, 92(8):3334 –3337 3335

Temperature of the fluids was 22 C. After completion of drinking,

measurements were continued for another 90 min.
Energy expenditure was calculated according to Ferrannini (12). All
data are given as means ⫾ sem. ANOVA with repeated measures was
used followed by post tests. A value for P ⬍ 0.05 was considered
statistically significant.

Results
Resting energy expenditure was 5.66 ⫾ 0.25 kJ/min before
500 ml water, 5.62 ⫾ 0.25 kJ/min before saline, and 5.68 ⫾
0.24 kJ/min before 50 ml water drinking (not significant).
Within 10 min after drinking 500 ml water, energy expen-
diture started to increase, reaching a maximum of 24 ⫾ 3%
above baseline after 60 min (Fig. 1, top). The response to 500

Downloaded from https://academic.oup.com/jcem/article/92/8/3334/2598681 by guest on 05 February 2021

ml water was significantly greater than the response to 50 ml
water or to 500 ml saline (P ⬍ 0.0001 by ANOVA for both
comparisons). We calculated the individual thermogenic re-
sponse between 0 and 90 min after fluid ingestion (Fig. 1,
bottom). On average, subjects used an additional 95 ⫾ 10 kJ
with 500 ml water, 28 ⫾ 4.7 kJ with 50 ml water, and 43 ⫾
4.6 kJ with 500 ml saline (P ⬍ 0.0001 by ANOVA). Resting RQ
was 0.81 ⫾ 0.01 before drinking 500 ml water, 50 ml water,
or 500 ml saline, and it did not change significantly over 60
min after drinking the test fluids. None of the interventions
elicited blood pressure or heart rate changes (Fig. 2).

Discussion
Drinking water induces a much greater thermogenic re-
sponse than drinking the same volume of isoosmotic saline.
We do not have a good explanation why water drinking
induced a thermogenic response in some but not in all studies
(7, 8). In the present and in our previous study (7), we applied
a metabolic chamber. Brown et al. (8) obtained measurements
using a canopy system. It is difficult to explain the discrepant
results with these methodological differences. Despite these
issues, we suggest that our observation provides clues re-
FIG. 1. Top, Relative change in energy expenditure (EE) over time in
garding the underlying mechanism of water-induced ther- obese subjects (eight men, eight women) after drinking 50 ml water
mogenesis. We previously showed a gender difference in the (50 water), 500 ml water (500 water), or 500 ml isoosmotic saline (500
substrates used for fueling water-induced thermogenesis (7). NaCl). At 0 min, subjects started to drink the fluids in less than 5 min.
Water drinking increased lipid oxidation in men and carbo- Testing was conducted in a randomized and crossover fashion on
separate days. *, P ⬍ 0.05; **, P ⬍ 0.01, for the comparison between
hydrate oxidation in women. In obese subjects, the response 500 ml water and 500 ml saline; #, P ⬍ 0.001 for the comparison
to water appears to be more uniform, i.e. RQ did not change between 500 and 50 ml water. Bottom, Individual thermogenic re-
significantly in men and women, indicating no change in the sponses to 50 ml water, 500 ml water, or 500 ml normal saline. The
ratio between carbohydrate and lipid oxidation rate. We response was calculated between 0 and 90 min after drinking. The
dotted line indicates the energy that is required to heat 500 ml water
speculate that obese subjects may be less able to switch be-
or saline from room temperature to body temperature. The P values
tween carbohydrate and lipid oxidation; the phenomenon is are given for the analysis with Bonferoni’s post test (PANOVA ⬍
sometimes referred to as metabolic inflexibility. 0.0001).
Similarly to previous studies (3, 4), water drinking did not
elicit a pressor response in younger subjects without auto- afferent stimulus and the afferent pathway causing activa-
nomic dysfunction. In contrast, water raises blood pressure tion of efferent sympathetic neurons is unknown. Water tem-
in autonomic failure patients (4). Thus, intact autonomic perature, distention of gastrointestinal organs, or changes in
cardiovascular regulation appears to attenuate the water- osmolarity could be involved. Water-drinking-induced car-
induced pressor response through baroreflex mechanisms. diovascular and metabolic responses are not solely explained
Similarly, low doses of the sympathetic stimulant yohimbine by a thermal stimulus. In autonomic failure patients, drink-
raise blood pressure in autonomic failure patients (4). The ing colder or warmer water elicited an identical pressor re-
same doses change blood pressure little or not at all in sponse (4). In healthy normal-weight subjects, approximately
healthy subjects. 60 –70% of the water-induced thermogenesis could not be
The mechanism causing sympathetic activation with wa- attributed to heating of the ingested water (7). Indeed, drink-
ter drinking is not fully understood. Studies in tetraplegic ing 37 C warm water elicited a substantial thermogenic re-
patients suggest a spinal mechanism (13). The nature of the sponse (7). Gastric distention increases sympathetic nerve
3336 J Clin Endocrinol Metab, August 2007, 92(8):3334 –3337 Boschmann et al. • Water and Thermogenesis

humans caused a greater increase in sweat production, a

sympathetic response, than infusion of isoosmotic solu-
tions (2).
We suggest that water drinking induces a local decrease in
osmolarity in the gastrointestinal tract, portal vein, and liver.
A local decrease in extracellular osmolarity may influence
thermogenesis through local changes in organ function, al-
tered activity of osmosensitive neural pathways, or both
mechanisms combined. Indeed, osmotic cell swelling acti-
vates anabolic processes including glycogen and protein syn-
thesis in the presence of suitable substrates (16). Further-
more, the decrease in osmolarity may stimulate afferent
neurons projecting to the spinal chord. Osmoreceptive af-

Downloaded from https://academic.oup.com/jcem/article/92/8/3334/2598681 by guest on 05 February 2021

ferent nerve fibers in the portal vein and in the liver have
been identified in animals (17). Physiological studies pro-
vided additional evidence for existence of peripheral osmo-
receptors (2). Although our study suggests that water-induced
sympathetic activation may be secondary to stimulation of os-
mosensitive afferent neurons, the transduction mechanism is
unknown. Serotonin, glutamine, and substance P release might
be involved (18). Hypoosmolarity causes cell swelling and,
therefore, membrane stretch. We hypothesize that osmosensory
transduction may rely on stretch-sensitive ion channels. Mem-
bers of the transient receptor potential (TRP) family, such as
TRPV1 (19) or TRPV4 (20), are prime suspects.

Acknowledgments

Received July 5, 2006. Accepted May 11, 2007.

Address all correspondence and requests for reprints to: Jens Jordan,
M.D., Franz Volhard Clinical Research Center, Charité Campus Buch,
Universitary Medicine Berlin, Wiltbergstrasse 50, Hs. 129, D-13125 Ber-
lin, Germany. E-mail: [email protected].
This work was supported by the Helmholtz Society and an unre-
FIG. 2. Changes in heart rate (HR) and systolic blood pressure (SBP) stricted grant from Forum Trinkwasser e.V.
over time in obese subjects (eight men, eight women) before and after Author Disclosure: M.B., J.S., G.F., A.L.B., F.C.L., and J.J. have nothing
drinking 50 ml water (50 water), 500 ml water (500 water), or 500 ml to declare.
isoosmotic saline (500 NaCl).
References
traffic in human subjects (14). We did not measure gastric
1. Lang F, Busch GL, Ritter M, Volkl H, Waldegger S, Gulbins E, Haussinger
distention directly. However, the differential response to D 1998 Functional significance of cell volume regulatory mechanisms. Physiol
drinking water or saline in our study suggests that gastric Rev 78:247–306
distention is not the crucial mechanism for water-drinking- 2. Haberich FJ 1968 Osmoreception in the portal circulation. Fed Proc 27:1137–
1141
induced sympathetic activation. The idea is supported by the 3. Scott EM, Greenwood JP, Gilbey SG, Stoker JB, Mary DA 2001 Water in-
observation that water drinking elicits more pronounced gestion increases sympathetic vasoconstrictor discharge in normal human
cardiovascular responses than drinking the same volume of subjects. Clin Sci Colch 100:335–342
4. Jordan J, Shannon JR, Black BK, Ali Y, Farley M, Costa F, Diedrich A,
saline (5, 9, 10). Similarly, in dogs, gastric water infusion Robertson RM, Biaggioni I, Robertson D 2000 The pressor response to water
elicited a much greater pressor response than isoosmotic drinking in humans: a sympathetic reflex? Circulation 101:504 –509
5. Raj SR, Biaggioni I, Black BK, Rali A, Jordan J, Taneja I, Harris PA,
saline in the same amounts or the inflation of a gastric balloon Robertson D 2006 Sodium paradoxically reduces the gastropressor response
(2). Furthermore, human magnetic resonance imaging stud- in patients with orthostatic hypotension. Hypertension 48:329 –334
ies demonstrated that after 40 min, only 25% of the ingested 6. Jordan J, Shannon JR, Grogan E, Biaggioni I, Robertson D 1999 A potent
pressor response elicited by drinking water. Lancet 353:723
water remains in the stomach (15). 7. Boschmann M, Steiniger J, Hille U, Tank J, Adams F, Sharma AM, Klaus S,
The main difference between water and isoosmotic saline Luft FC, Jordan J 2003 Water-induced thermogenesis. J Clin Endocrinol Metab
is the osmolarity of the solution. Water is hypoosmotic in 88:6015– 6019
8. Brown CM, Dulloo AG, Montani JP 2006 Water-induced thermogenesis re-
contrast to isoosmotic saline. Thus, our data suggest that the considered: the effects of osmolality and water temperature on energy expen-
water-induced changes may be explained by stimulation of diture after drinking. J Clin Endocrinol Metab 91:3598 –3602
9. Lipp A, Tank J, Franke G, Arnold G, Luft FC, Jordan J 2005 Osmosensitive
osmosensitive structures. Indeed, the time course of the mechanisms contribute to the water drinking-induced pressor response in
changes in sympathetic activity (3), blood pressure (4, 6, 13), humans. Neurology 65:905–907
and metabolic rate (7) parallel the time course of altered 10. Brown CM, Barberini L, Dulloo AG, Montani JP 2005 Cardiovascular re-
sponses to water drinking: does osmolality play a role? Am J Physiol Regul
plasma osmolarity after water drinking (7). Moreover, in- Integr Comp Physiol 289:R1687–R1692
fusion of hypoosmotic solutions through a gastric tube in 11. Brown D, Cole TJ, Dauncey MJ, Marrs RW, Murgatroyd PR 1984 Analysis
Boschmann et al. • Water and Thermogenesis J Clin Endocrinol Metab, August 2007, 92(8):3334 –3337 3337

of gaseous exchange in open-circuit indirect calorimetry. Med Biol Eng Com- 16. Haussinger D 1996 The role of cellular hydration in the regulation of cell
put 22:333–338 function. Biochem J 313(Pt 3):697–710
12. Ferrannini E 1988 The theoretical bases of indirect calorimetry: a review. 17. Adachi A 1984 Thermosensitive and osmoreceptive afferent fibers in the he-
Metabolism 37:287–301 patic branch of the vagus nerve. J Auton Nerv Syst 10:269 –273
13. Tank J, Schroeder C, Stoffels M, Diedrich A, Sharma AM, Luft FC, Jordan 18. Wu XY, Zhu JX, Gao J, Owyang C, Li Y 2005 Neurochemical phenotype of
J 2003 Pressor effect of water drinking in tetraplegic patients may be a spinal vagal afferent neurons activated to express C-FOS in response to luminal
reflex. Hypertension 41:1234 –1239 stimulation in the rat. Neuroscience 130:757–767
14. Rossi P, Andriesse GI, Oey PL, Wieneke GH, Roelofs JM, Akkermans LM 19. Sharif NR, Witty MF, Seguela P, Bourque CW 2006 An N-terminal variant
1998 Stomach distension increases efferent muscle sympathetic nerve activity of Trpv1 channel is required for osmosensory transduction. Nat Neurosci
and blood pressure in healthy humans. J Neurol Sci 161:148 –155 9:93–98
15. Ploutz-Snyder L, Foley J, Ploutz-Snyder R, Kanaley J, Sagendorf K, Meyer 20. Alessandri-Haber N, Dina OA, Yeh JJ, Parada CA, Reichling DB, Levine JD
R 1999 Gastric gas and fluid emptying assessed by magnetic resonance im- 2004 Transient receptor potential vanilloid 4 is essential in chemotherapy-
aging. Eur J Appl Physiol Occup Physiol 79:212–220 induced neuropathic pain in the rat. J Neurosci 24:4444 – 4452

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the
endocrine community.

Downloaded from https://academic.oup.com/jcem/article/92/8/3334/2598681 by guest on 05 February 2021