New horizons in the diagnosis and management of dehydration

Abstract

Hydration is a fundamental aspect of clinical practice and yet it is an under-researched topic, particularly in older people, leading to many areas of uncertainty. There are two types of dehydration; hypertonic, which is a water deficit, and isotonic, which is a deficit of both water and salt. Individual clinical signs and bedside tests are poor diagnostic tools, making dehydration difficult to identify. However, the diagnostic value of a holistic clinical approach is not known. The gold-standard clinical test for dehydration is serum osmolality, but this cannot diagnose isotonic dehydration and may delay diagnosis in acute situations. Salivary osmolality point-of-care testing is a promising and rapid new diagnostic test capable of detecting both hypertonic and isotonic dehydration in older people, but further evidence to support its clinical utility is needed. Daily fluid requirements may be less than previously thought in adults, but the evidence specific to older people remains limited. Hydration via the subcutaneous route is safer and easier to initiate than the intravenous route but is limited by infusion speed and volume. Prompting older adults more frequently to drink, offering a wider selection of drinks and using drinking vessels with particular features can result in small increases in oral intake in the short-term. The ongoing clinically-assisted hydration at end of life (CHELsea II) trial will hopefully provide more evidence for the emotive issue of hydration at the end of life.

Key Points

• Hypertonic dehydration is predominantly a water deficit; isotonic dehydration is a deficit of both water and salt.

• Clinical examination is of limited diagnostic value when diagnosing dehydration in older people.

• Recent research suggests daily fluid requirements may not be as high as previously recommended.

• Oral rehydration salts or milk are superior to water alone when treating dehydration orally.

• Subcutaneous fluid replacement is much safer than the intravenous route but is slower

Introduction

Hydration is a fundamental but confusing area of practice for clinicians working with older people. There is little doubt that addressing dehydration can prevent and treat common disorders such as urinary tract infections, acute kidney injury, orthostatic hypotension, constipation and delirium. However, the assessment of dehydration can be rather subjective, non-evidence based and affected by multiple factors. Once identified, reversing dehydration can be challenging as it often involves behavioural change and overcoming barriers such engrained habits and fear of incontinence.

Definitions

Dehydration may, in its simplest form be considered a deficiency of water, but even this simplistic definition is controversial. The lack of formal definition has hindered academic progress, with several professional societies differing in even the basic principles of whether the definition should include factors such as osmolality, intake or output, clinical findings and volume distribution. In 2019, a group of UK academics convened to reach consensus [1]. The group which identifies as multidisciplinary was drawn from anaesthesia/critical care (50%), sports science, cardiology and dietetics and was largely based in southeast England. The group of 12 experts used expert opinion, literature reviews and a modified Delphi process to produce a series of recommendations.

Two types of dehydration were defined by the consensus group:

• Hypertonic dehydration, an uncompensated, predominantly pure water deficit (e.g. inadequate water intake or excessive loss) resulting in a raised extracellular osmolality.

• Isotonic dehydration, which is a deficiency of water with an associated reduction in salt (e.g. caused by diarrhoea) which does not increase the extracellular osmolality [1].

While these definitions may have a role in standardising diagnosis, the narrow geographical membership and lack of diverse expertise will likely lead to a lack of uptake by the wider community. However, the definitions are clear and compared with other existing definitions are easily translated into clinical practice.

Diagnosis

Clinical examination

It is generally accepted that the diagnostic accuracy of clinical examination is poor [2]. This has recently been confirmed in the Dehydration Recognition in our Elders (DRIE) study [3]). In this England-based study, 188 participants, mean age 86 (±8) years, were included from 56 care homes. Each participant underwent thorough clinical evaluation which included symptoms of dry eyes/mouth, thirst and feeling out of sorts and examination of the mouth and tongue, lips, eyes, axillae, palms, skin, capillary refill time, foot vein filling, temperature, heart rate and sitting and standing blood pressure. Serum osmolality >300 mOsm/kg was the reference standard for dehydration, with a pre-defined sensitivity and specificity of 70% for diagnostic accuracy. Thirty-eight participants had serum osmolality >300 mOsm/kg, but none of the 49 signs, symptoms or bed side tests reached the required level of diagnostic accuracy to identify these individuals.

The most recent Cochrane review in this area (Clinical symptoms, signs and tests for identification of impending and current-loss dehydration in older people, 2015 [4]) reaches a similar conclusion to the DRIE study. The review used serum osmolality as the gold standard definition of current water-loss dehydration, with a cut-point of ≥300 mOsm/kg [5]. It should be noted therefore that the results do not apply to isotonic dehydration. The review included 24 studies. Of these, 13 were classed as high risk of bias, 13 as low risk and the risk in the remaining five was unclear. The review evaluated 36 different signs, symptoms and bedside tests, but with variable cut offs and scales used by different studies this resulted in 152 different diagnostic criteria. A pre-specified criteria of sensitivity 60% and specificity 75% was set as an acceptable level of diagnostic accuracy. Of the 152 diagnostic criteria only one met this level of accuracy—bioelectrical impedance analysis (at a resistance level of 50 kHz, ≥450 ohm). However, this was a single study (n = 15, unknown risk of bias [4]), with three other studies (all with a lower risk of bias) using the same criteria not reaching these levels of diagnostic accuracy. Unfortunately, a pooled analysis was not possible.

A limitation of the DRIE study and the Cochrane review, both of which are specific to older people, are that they do not evaluate whether different combinations of symptoms and signs improve diagnostic accuracy, or whether looking at an individual from a holistic point of view is clinically useful. Another criticism is that by using serum osmolality as the gold standard they are unable to evaluate isotonic dehydration, and given that clinical examination may identify isotonic dehydration, this reduces the specificity of their analysis.

Novel tests

Ultrasonography

While serum osmolality may be used to diagnose hypertonic dehydration, it has been proposed that inferior vena cava (IVC) diameter and IVC collapsibility may be used as a test for isotonic dehydration [6–9]. In one recent study, based in a nursing home in Japan, 89 residents were assessed for dehydration (defined as serum osmolality ≥295 mOsm/kg) and underwent assessment of IVC by ultrasonography [9]. It is not clear whether these assessments occurred on the same day. The IVC diameter and the IVC collapsibility index were similar between those with dehydration (n = 15) and those who were euhydrated (n = 74), p 0.22 (although no pre-defined level of clinical significance or statistical power was set). There was also no correlation between IVC diameter and serum osmolality. In addition, the researchers used the blood urea nitrogen/creatine (BUN:Cr) ratio as an alternate marker of dehydration. The BUN:Cr ratio is suggested by some to be a measure of isotonic dehydration, however, this is considered a non-specific outcome which is affected by muscle mass, renal function, catabolic states (such as sepsis or starvation) and protein intake [1]. Nevertheless, in this study there was no correlation between BUN:Cr and IVC diameter or distensibility in nursing home residents.

In contrast, a study based in Germany evaluated IVC diameter and collapsibility index in a series of older adults (>65 years) attending the Emergency Department [6]. Participants were classed as dehydrated (n = 78) or euhydrated (n = 121) based on clinical examination; those with fluid overload or uncertainty of fluid balance were excluded. Of five different measures of IVC diameter, four were significantly lower in the dehydration group (e.g. IVC diameter compression in transverse plane euhydrated median diameter 0.9 (0–1.9) cm, dehydrated 0.6 (0–1.5) cm, P < 0.0001). Of the two measures of IVC collapsibility both were significantly higher in the dehydrated group (e.g. collapsibility index euhydrated median 0.28 (0.02–0.91), dehydrated 0.35 (0.08–1.0), p 0.045). Clearly the use of clinical signs and symptoms as the gold standard reference for dehydration is subjective, open to bias and reduces the quality of the study given that evidence shows is it not accurate. Furthermore, IVC measurements are probably more useful in people with intravascular volume depletion (isotonic dehydration) than hypertonic dehydration, in which case clinical examination as a reference standard is further flawed. However, further work in this area should be encouraged as the diagnosis of isotonic dehydration lacks reliable diagnostic criteria.

Point of care testing

Saliva osmolality [10] has recently been explored as a potentially useful diagnostic test for dehydration in older people. Saliva can be collected by placing a small absorbent pad, linked to a display panel, underneath the tongue until an indicator band appears on the handle (like a lateral flow test). The pad is then inserted into a collection device which compresses the saliva into a tube ready for point of care testing and a rapid saliva osmolality result.

A cross-sectional study set in one Emergency Department in Wales, collected data on 130 consecutive attendees (mean age 78 ± 9 years). Each participant underwent clinical assessment for dehydration, had their serum osmolality and their BUN:Cr measured and also had their saliva analysed for osmolality [10]. A large sample was excluded (242) because they were too unwell, could not consent or had already started treatment. In this study, the device was placed underneath the tongue for 4 minutes and a salivary volume of ≥25 μl was required for analysis. Of the 130 participants, adequate saliva was collected from 98. Twenty-seven were diagnosed with hypertonic dehydration, 25 with isotonic dehydration and 78 were found euhydrated using reference standards. Table 1 demonstrates the level of diagnostic accuracy of saliva osmolality.

Table 1.

Diagnostic accuracy of saliva osmolality [10]

	Hypertonic dehydration	Isotonic dehydration	All dehydration
Saliva osmolality diagnostic cut-off point (mOsm/kg)	95	97	94
ROC Area Under the Curve Analysis (95% CI)	0.76 (0.66 to 0.87), P < 0.001	0.76 (0.62 to 0.89), P 0.001	0.76 (0.66 to 0.86), P < 0.001
Sensitivity	70%a	78%a	76%a
Specificity	68%a	72%a	68%a

Receiver Operating Characteristic (ROC); 95% CI—confidence intervals;

^a95%CI not reported

In contrast to other clinical tests, the saliva osmolality was able to identify both hypertonic and isotonic forms of dehydration with moderate levels of accuracy. It has the advantage of being less invasive and providing a rapid result in comparison to serum osmolality, but results are affected by timing of eating and drinking and reduced levels of saliva. It is also unknown whether embedding these tests into clinical practice would result in clinical improvements and further early phase studies are required to evaluate whether the tests are sensitive to change.

Rehydration

Although treatment principles are simple—replenish the water and salt losses—there are many unanswered questions which make the treatment of dehydration challenging.

Fluid requirements

The UK NHS’s Eatwell Guide recommends a daily intake of six to eight glasses of fluid per day and the European Food Safety Authority recommends 2.5 l of water for males over 14 years and 2 l for females [11]. However, these recommendations have recently been called into doubt following a large international study. This cross-sectional study analysed water turnover (fluid balance, ml/day) data from over 5,604 subjects, from birth to 96 years old, from 23 different countries, available from the International Atomic Energy Agency International Doubly Labelled Water database. Each participant consumed a weight-adjusted dose of deuterated water. The level of deuterated water was measured before and after the isotope drink using either saliva, urine or serum, allowing an estimation of the amount of deuterated water left in the body and therefore the total body water volume [12]. The study team found that water turnover reduced with advancing age and was lowest in females aged >65 years. Total body water also declined with advancing age, presumed secondary to reduced muscle mass. The authors then used extensive data on geography, climate, activity levels and water turnover to produce an equation to estimate an individual’s daily water requirements:

[1,076 × PAL] + [14.34 × body weight (kg)] + [374.9 × sex] + [5.823 × humidity (%)] + [1,070 × athlete status] + [104.6 × HDI] + [0.4726 × altitude (m)]—[0.3529 × age (years)²] + [24.78 × age (years)] + [1.865 × ambient temperature (°C)²]—[19.66 × temperature (°C)]—713.1.

PAL, physical activity level = total energy expenditure/basal energy expenditure, roughly estimated as light work = 1.7, moderate work female (F) 2.2, male (M) 2.7, moderately heavy work F 2.3, M 3.0, heavy work F 2.8, M 3.8; athlete = 1, non-athlete = 0; male = 1, female = 0; HDI, human development index, high (e.g. UK) =0, middle =1, low =2;

For an inactive 60-year-old female weighing 65 kg with average UK home humidity (50%) and temperature (20°C) in Newcastle upon Tyne, the water turnover would be 2169.5 ml/day; assuming 40% of water input were from food, that would equate to a recommended daily fluid intake of 1.3 l (or 5.2 glasses of water). Unfortunately, due to the age-related changes in water turnover beyond the age of 60 years, the authors excluded people over this age to derive the equation. In addition, the equation does not consider that a greater proportion of older people may be dehydrated. However, it does competently call into question the existing NHS advice.

Oral hydration

Orally ingested water is largely absorbed in the small intestine where the absorption of sodium and glucose creates osmotic gradients promoting intracellular water uptake. It could therefore be assumed that plain water is not the most hydrating of fluids to drink. To confirm this, one study compared the effects of 13 different beverages on hydration [13]. Seventy-two young (mean age 25 ± 4 years), healthy and active males consumed four beverages each, under experimental conditions to create a standardised baseline level of hydration, on different days in a randomised order. One litre of each beverage was consumed steadily over 30 minutes. The drinks included were still spring water, sparkling water, cola, diet cola, sports drink, oral rehydration salts (ORS), orange juice, lager, hot black coffee, hot black tea, cold black tea, full-fat milk and skimmed milk. After 1 hour, in comparison with still water, a statistically significant positive fluid balance was seen with full-fat milk, skimmed milk, ORS and orange juice. At 4 hours, this effect persisted for all these drinks except orange juice. A hydration index (urine mass within 2-hours post beverage/urine mass post-still water) was significantly greater for full-fat milk, skimmed milk, ORS and orange juice. An additional interesting finding revealed that hydration was no worse than still-water when consuming caffeinated drinks or alcohol, although perhaps the study duration was too short to adequately capture the pharmacological effects of these drinks [13]. Clearly these results arise from experimental conditions in a highly selective cohort and similar results may not arise in older people with altered physiology, people with co-morbidity or polypharmacy.

Parenteral hydration

Unlike absorption of water from the gut, the absorption of water from the subcutaneous tissues is a passive process and may therefore be an inefficient and slow means of rehydration. However, studies in both younger and healthy older populations have demonstrated moderate absorption rates. Yet healthy populations are not those in need of parental nutrition and similar evidence is required for those in whom it is indicated.

This has recently been confirmed in a small (n = 6) study of older (mean age 81 ± 2 years) hospitalised adults based on a geriatric medicine ward in a single hospital site in Denmark. Those who were unable to consent and those on a fluid restriction or at the end of life were excluded. Participants received 235 ml of radioisotope labelled saline, infused subcutaneously over 1 hour, via butterfly needle sited on the abdomen. Absorption was measured using a gamma detector at the infusion site as well as peripheral blood sampling. At 60 minutes, 53% (95%CI 50–56) had been absorbed into circulation and at 110 minutes 88% (95%CI 86–90) had been absorbed, giving an average absorption rate of 127 ml/hour (95% CI 90–164 ml/hours) [14]. This was a small study, but with a clinically relevant cohort. The volume infused was low and over a short duration of time, so whether absorption rates decline beyond the hour and with larger volumes is unknown.

A recent systematic review and meta-analysis on the benefits and harms of subcutaneous fluid replacement in older people included 29 studies in the review and six in the pooled analyses [15]. None of the included studies were graded as low risk of bias. The pooled analysis compared the rate of adverse events in those randomised to subcutaneous versus intravenous fluid. There was a 38% lower risk of adverse events in the subcutaneous group (risk ratio 0.62 [95%CI 0.53, 0.71], moderate quality of evidence). Intravenous hydration reduced serum osmolality by 5.75 mmol/kg (95%CI 0.13 to 11.4; two trials, very low quality of evidence) more than subcutaneous infusion, although it is not clear at what time point this was measured. Over 24 hours, the intravenous route infused 155 ml (95%CI 60 to 253; three trials; very low quality of evidence) more than the subcutaneous route when administering 1 l of fluid. The clinical significance of this small volume is open to debate. There was low quality evidence that in people with cognitive impairment there was a 58% reduced risk of agitation when using the subcutaneous route compared to the intravenous route (risk ratio 0.42 [95%CI 0.22 to 0.79]; three studies).

There are no studies directly comparing the oral route with parenteral routes when hydrating older people.

Behavioural interventions

A recent systematic review examined the evidence for oral interventions to treat or prevent dehydration in older people in hospital or in care settings [16]. They identified 19 eligible studies in people aged >65 years; 10 of these were rated as low risk of bias and nine as unclear risk of bias. The different interventions were classed into three categories: behavioural (drinking straws, different vessel designs, hydration advice, increased prompting, increased drink rounds and increased drink selection), environmental (high-contrast tableware, increased assistance and improved social atmosphere), multifaceted (advice and increased drink rounds; drink selection, cup size and prompting) and nutritional (premade thickened fluids and thickened fluids plus plain (liquid) water). The review found that behavioural interventions were more consistently associated with increased fluid intake or increased levels of hydration. However, although increased fluid intake was found to be statistically significant in some studies, the clinical significance of the small increase is questionable (for example, the study evaluating a new drinking vessel, increased daily fluid intake by a mean of 66 ml, reported as statistically significant [17]). Findings from environmental, nutritional and multifaceted studies were inconsistent and therefore not recommended by the authors. The review authors highlight the difficulty in implementing complex interventions which typically require high levels of staff and patient motivation, increased resource and that they have low levels of longevity. A 2019 qualitative study exploring the barriers to hydration found that fear of incontinence, physical and cognitive abilities, a lack of understanding and a loss of pleasure all contributed [18].

Future directions

Behavioural mimicry is proposed as a potential solution to increase fluid intake in people with dementia as it does not rely on memory. In one deception study of 42 healthy older adults, who thought they were entering a study about language, an actor posed as a volunteer and either frequently drank from their cup, or just touched it. Participants were noted to spend significantly more time drinking when the actor took a drink from their cup (0.08 drinks/minute versus 0.05, p 0.021) [19]. Further studies will be required to confirm whether this social effect would translate into clinically useful effects in people with dementia.

The clinically-assisted hydration at end of life (CHELsea II) trial is a cluster randomised trial evaluating assisted-hydration in the last few days life, across 80 UK hospices [20]. The trail aims to determine whether intravenous fluids in the last few days of life can prevent delirium; secondary outcomes include upper airway secretions, dyspnoea, nausea and vomiting, and adverse events of assisted hydration. The results are expected in 2025 and given the lack of evidence and the differing opinions between clinicians this trial will hopefully provide some clarity. As this study is limited to hydration in the last few days of life, further research would be helpful to guide clinicians providing care for people with limited life expectancy but which is longer than a few days.

While this New Horizons review has focussed on hydration of older people, what about hydration in the people looking after older people? One current observational study is exploring correlations between self-assessed competence and hydration (measured using urine specific gravity) in 60 NHS doctors [21]. The results are expected in 2023.

Given the size, quality and cohorts included in most hydration research, much research is needed to improve care for older people. This is not an easy task, with a lack of quality basic science studies from which to build a solid evidence base, a lack of accepted definitions, diagnostic criteria and daily fluid requirements and a wide-ranging potential set of clinical outcomes.

Conclusion

As dehydration is associated with significant morbidity and mortality in older people, its recognition and treatment is important. While individual clinical signs and bedside tests are limited in diagnostic value, a holistic clinical evaluation should still be recommended. Serum osmolality has a role in the diagnosis in hypertonic dehydration with further studies required to evaluate point of care testing. Daily fluid requirements may be lower than previously recommended, but this evidence is weaker for older people. While oral hydration is usually the preferred method, the subcutaneous rate offers reasonable rates of fluid absorption and has the advantage over intravenous routes of having much fewer adverse effects. As complex interventions, behavioural and environmental strategies may result in small increases in fluid intake but are limited by motivation and longevity.

Declaration of Conflicts of Interest

None.

Declaration of Sources of Funding

None.