OUP user menu

Thirst and hydration status in everyday life

Mindy Millard-Stafford, Deborah Michael Wendland, Namrita K O'Dea, Tracy L Norman
DOI: http://dx.doi.org/10.1111/j.1753-4887.2012.00527.x S147-S151 First published online: 1 November 2012


Water is an essential nutrient for all persons; thus, maintaining a chronic state of optimal hydration is recognized to provide health benefits. Fluid balance is maintained via thirst, a feedback-controlled variable, regulated acutely by central and peripheral mechanisms. However, voluntary drinking is also a behavior influenced by numerous social and psychological cues. Therefore, whether “thirst-guided” drinking maintains optimal hydration status is a multifactorial issue. Thirst perception is typically assessed by subjective ratings using either categorical or visual analog scales; however, which instrument yields greater sensitivity to change in hydration status has not been examined. Ratings of thirst perception do not always yield predictable patterns of voluntary drinking following dehydration; therefore, perceived thirst and ad libitum drinking are not equivalent measures of human thirst. The recommendation “drink to thirst” is frequently given to healthy individuals during daily life. However, factors and conditions (e.g., age, disease) that influence thirst should be recognized and probed further.

  • beverages
  • fluid balance
  • fluid intake
  • water


Water is an essential nutrient for all persons across the lifespan; thus, maintaining a chronic state of optimal hydration is recognized to provide health benefits and to enhance quality of life. Fluid balance is maintained via thirst, a feedback-controlled variable, regulated acutely by central and peripheral mechanisms.1 However, voluntary drinking in humans is also a behavior influenced by social and psychological cues.2,3 Therefore, whether thirst-guided drinking maintains optimal hydration status is a multifactorial question and remains a topic of interest.

Human thirst is an amorphous measure that is somewhat arbitrarily quantified. Lay recommendations such as “drink to thirst” stem from the Institute of Medicine, which stated the following: “The vast majority of healthy people adequately meet their daily hydration needs by letting thirst be their guide.”4 The purpose of the present review is to briefly summarize the thirst mechanism, describe techniques to assess human thirst, and explore recent studies related to thirst-guided drinking in daily life.


The reader is referred to recent reviews on the thirst mechanism.58 Key physiological signals for thirst are plasma hyperosmolality with consequential cellular dehydration, and hypovolemia due to low blood volume and low arterial pressure.9,10 The concept of a thirst “center” per se has been replaced with a complex neural circuitry regulating fluid intake.8 Osmoreceptors housed in the hypothalamus sense fluctuations in osmolality of plasma and cerebrospinal fluid. Located in the circumventricular organs (organum vasculosum of the lamina terminalis [OVLT] and subfornical organ [SFO]) and median preoptic nucleus (MnPO), these receptors have neuronal connections to supraoptic and paraventricular nuclei that synthesize arginine vasopressin (AVP). The posterior pituitary releases AVP with the net result of decreased renal water excretion and thirst stimulation. AVP release is also regulated by peripheral baroreceptors, cardiopulmonary volume receptors, and angiotensin II. Although not totally resolved in humans, SFO mediates the dipsogenic action of angiotensin II in rats.11 Thus, a complex neuroendocrine pathway stimulates “water appetite” to maintain hydration.613

When dehydrated, animals and humans sense thirst and voluntarily drink fluids until satiated. In rats, due to rapid gastric emptying, osmolality is adjusted soon after the onset of drinking, causing inhibition of AVP secretion. Termination of drinking occurs a few minutes after AVP inhibition, but coincides with rapid restoration of plasma sodium.14 In humans, satiation of thirst may also occur rapidly, but usually prior to alterations in plasma sodium or osmolality.15,16 In addition, sensory input from fluid passage through the oro-pharynx and esophagus elicit feedback to satiate thirst.1219

During ad libitum rehydration, subjects report attenuated thirst but continue to drink. Increased regional cerebral blood flow was observed during thirst and continued after drinking commenced.12 Moreover, brain activation correlated with posthydration plasma AVP, suggesting that OVLT and other brain regions monitor hydration status and drinking behavior.12 Most investigations have utilized water for rehydration protocols following thirst.1223 Drinking is also influenced by the need for salt7 such that sodium balance may take priority over fluid restoration.24 Although thirst stimulus occurs within minutes, the onset of sodium appetite can take longer (hours) to develop.25 Ad libitum fluid replacement among individuals similarly dehydrated (3% loss in body mass) may vary due to differential sodium loss even when perceived thirst is similar.26 Thus, it is unclear how sodium concentration and osmotic pressure concomitantly influence human thirst and its satiation with different fluids.

In addition to homeostatic mechanisms, psychological and social cues also influence daily fluid ingestion. Much of daily fluid intake is ingested with meals.328 A gender difference in hydration is possible because girls demonstrate more favorable drinking habits than boys.29 Voluntary fluid restriction could also be related to fear of incontinence (in women and elderly persons), inconvenience of finding or being permitted access to restroom facilities, mobility status,30 lack of safe and/or acceptable beverages, and inability to swallow.31 Thus, drinking to thirst is a multifactorial behavior.


Thirst assessment relies upon an individual's recognition, perception, and explanation of the sensation. Historically, several methods have quantified the subjective measurement of thirst. Both visual analog scales (VAS) and categorical rating instruments (Figure 1) are cited in the literature. Each method may have advantages and limitations. To our knowledge, no formal research has compared the sensitivity or applicability of these various instruments in response to known changes in hydration status to better define accepted methodology to assess thirst.

Figure 1

Representations of instruments used to assess human thirst: 10-cm visual analog scale (top) and 7-point categorical scale (bottom).

An early study measured thirst with a VAS scored as the deviation from perceived thirst at the start of the experiment.32 Subjects placed a mark on a 30-cm horizontal line with the line center anchored by the term “standard.” Deviations to the left and right of standard over time indicated decreased and increased thirst, respectively. Thompson and Campbell also concomitantly used a 7-point categorical scale, with integers of 1 representing extreme thirst, 4 being neutral, and 7 representing no desire to drink even on request.32 Rolls et al.33 later published a VAS that tracked with plasma osmolality. Subjects responded to questions (e.g., “How thirsty do you feel now?”) by placing a mark on a 10-cm horizontal line anchored by phrases “not at all” and “very thirsty” at the extremes. Slight modifications of this instrument have been implemented over the years.1036 Other adaptations, including a vertical response line37 or altered placement of extreme anchors and permitting subjects to extend the given line,38 have also been published.

Another instrument to assess thirst is a categorical Likert rating.39 The Thirst Sensation Scale (TSS) comprises 37 survey items corresponding to various feelings, symptoms, and sensations both related and unrelated to thirst. Although comprehensive, the time-consuming nature of the TSS may have prompted investigators to adopt a 9-point categorical rating scale with integer descriptors ranging from “not thirsty at all” to “very, very thirsty.”2642 Further simplification of a Thirst Likert scale collapsed ratings into three categories: no thirst, mild thirst, or moderate/severe thirst.43

Although there is an increasing reliance upon VAS as a robust psychometric instrument,44 the preferred instrument (categorical versus VAS) in terms of sensitivity and validity to assess thirst is not clear, particularly with regard to the population to be studied (e.g., elderly persons, children). Comparison of VAS and categorical scales (or the combination of the two) to assess another related construct (i.e., “thermal” perception) has not yielded definitive answers to date.45 Thus, a comparison of the various perceived thirst rating scales is clearly needed before a recommendation can be made regarding which to utilize in systematic studies on “thirst.”

Another question remains in assessing thirst. Is perceived thirst synonymous with the volume of fluid ingested ad libitum when a subject is dehydrated? As suggested previously, drinking according to thirst may not correspond to perceived thirst.46 Individuals with cystic fibrosis (CF), who have exceptionally salty sweat, report similar perceived thirst after 3% dehydration compared with non-CF subjects, but drank less fluid (by 40% or approximately 750 mL) during rehydration.26 It was previously reported that CF exhibits blunted thirst drive.47,48 However, Brown et al.26 indicate that thirst sensation is not different in CF, but drinking response is blunted relative to thirst. Thus, depending on the measure used to quantify human thirst (i.e., perception rating or the volume of fluid ingested during ad libitum drinking), the outcome could yield different interpretations. Therefore, perceived thirst and ad libitum drinking are not equivalent measures of human thirst.


Whether individuals can let thirst be their guide has provoked debate. The following guidance was issued from the International Life Sciences Institute North America Conference on Hydration and Health Promotion (November 29–30, 2006, Washington, DC): “Most healthy people adequately meet their daily water needs when they let thirst be their guide. However, this is not true for athletes, military in hot environments, people who are ill, the elderly or infants. The sense of thirst (or the ability to communicate it) of these populations is not an adequate reflection of their water needs.”49

It is frequently reported that sensitivity of the thirst mechanism is diminished in elderly persons.34,48 Presumably, this difference contributes to suboptimal hydration, particularly with reference to institutionalized cohorts.50 In their review, Kenney and Chiu concluded that functionally independent adults (aged =65 years) seem to ingest sufficient fluids to maintain hydration, but age-related differences may influence the thirst mechanism.51 Older adults tend to have higher baseline osmolality52 and a potentially higher osmotic set point for thirst (with little change in sensitivity) but blunted thirst and satiety in response to blood volume challenges.53

Recent evidence that thirst is diminished in apparently healthy older adults and reduces fluid intake is not entirely consistent. Thirst ratings by VAS were not different in older (approximately 60–80 years) versus younger adults (approximately 20–45 years) in studies by Farrell et al.23 and Bossingham et al.35 Despite similar perceived thirst, older men ingested about half the fluid in response to hypertonic saline-induced thirst.23 However, in a free-living study that accounted for all daily fluid intake and output,35 ad libitum water intake was similar in older versus younger adults. Moreover, hydration status based on plasma osmolality was similar between the groups, although urine-specific gravity was different by age and gender (i.e., higher for older and male groups).35 A longitudinal study conducted over a 10-year period reported that daily water consumption increased by =75% in elderly participants.53 Taken together, the drinking response to thirst may not be entirely predictable in healthy older adults, and whether drinking to thirst results in adequate hydration remains unclear.

There are conditions, however, in which drinking to thirst is neither optimal nor advised. The management guidelines for some diseases – such as congestive heart failure,54,55 end-stage renal disease on dialysis,56 and chronic obstructive pulmonary disease57– include fluid restriction and, thus, elevated thirst.58,59 In addition to disease states, medications that elicit dry mouth (e.g., some antidepressants and antihypertensives) may also affect drinking.60 Other nondisease factors noted in the literature that affect thirst include cold exposure41 and hormonal fluctuations.61 Therefore, a variety of conditions or factors could modulate thirst in daily life.


In summary, although thirst is homeostatically controlled, drinking to thirst is also influenced by other psychological and social cues. The best method to assess human thirst in research is unclear and merits further study. More investigation is needed to extend our understanding of whether drinking to thirst is appropriate across all segments of the population in their daily lives and under specific conditions.


The authors thank Dr. Mary Beth Brown and Dr. Ann Grandjean for their input on this manuscript.

Funding.  Partial funding was received from the Cystic Fibrosis Foundation and the Carl Gisolfi Fund from the American College of Sports Medicine Foundation to support the publication by Brown et al.26 for which MM-S is among the authors. The author has received prior research support from The Coca-Cola Company, Atlanta, GA.

Declaration of interest.  The authors have no relevant interests to declare.


View Abstract