Evidence overview

After primary screening of 206 articles and abstracts identified from the literature search, 20 articles were retrieved. Most of these studies had in fact assessed the effectiveness of oral fluids in the treatment of gastroenteritis and dehydration rather than in the prevention of dehydration. Only one prevention study was identified. In that study, continued breastfeeding and use of ORS solution at home were evaluated as potential strategies for preventing dehydration.

In a case–control study from Bangladesh,⁸² children aged between 1 and 35 months were selected for study inclusion if they had watery diarrhoea for 6 days or less at first presentation and had been breastfeeding up to the time of onset of diarrhoea. All were assessed for dehydration and were classified as ‘cases’ (with moderate to severe dehydration if there was a definite decrease in skin elasticity and presence of one or more of following signs: sunken eyes, failure to urinate for 6 hours, sunken anterior fontanelle, rapid and weak pulse) or as ‘controls’ (with no dehydration or mild dehydration if they did not fulfil those clinical criteria). Home ORT use was defined as giving either pre-packaged ORS solution or home-made salt and sugar solution. Information on socioeconomic and demographic characteristics, medical history and fluid/feeding interventions at home was collected using a field-tested structured questionnaire administered by an interviewer. Cases and controls were recruited from the same reporting area. Observer bias was reduced by blinding the interviewers to the hypothesis being tested and by blinding both mothers and interviewer to the case and control allocation. After analysing the association of each factor of interest with dehydration and identifying various confounding variables, logistic regression analysis was conducted to identify factors independently associated with dehydration. [EL = 2+]

There were 285 cases and 728 controls. After controlling for confounding factors (lack of maternal education, history of vomiting, high stool frequency, young age and infection with Vibrio cholerae), the risk of dehydration was five times higher in infants whose mothers stopped breastfeeding compared with infants whose mothers continued to breastfeed following the onset of diarrhoea (OR 5.23; 95% CI 1.37 to 9.99; P = 0.016). Similarly, the risk of dehydration was 1.5 times higher in infants who did not receive any ORT at home compared with those who received plentiful ORT (total volume ≥ 250 ml) (OR 1.57; 95% CI 1.08 to 2.29; P = 0.019). Infants receiving smaller amounts of ORT (≤ 250 ml) before admission had an 18% higher risk of dehydration compared with those receiving plentiful ORT, but the risk was not statistically significant (OR 1.18; 95% CI 0.84 to 1.66; P = 0.343).

Evidence summary

Evidence from a case–control study [EL = 2+] indicated that cessation of breastfeeding in children with gastroenteritis was associated with an increased risk of dehydration. This study also suggested that oral fluid supplementation begun at home and given in good quantity was associated with a reduced risk of dehydration.

GDG translation from evidence to recommendation

Evidence, though limited, suggests that continued breastfeeds and provision of oral fluid supplementation to children with gastroenteritis reduces the risk of dehydration. The lack of available evidence was not surprising, given the ethical difficulties with undertaking an RCT comparing the administration and withholding of oral fluid supplementation. Given that oral fluids are effective in the management of the dehydrated child, as discussed in Section 5.2, the GDG considered that it was reasonable to assume that liberal fluid supplementation is effective in the prevention of dehydration. While it was recognised that some children may prefer other oral fluids, ORS solution has advantages (Section 5.3) and so should be used if possible for children at increased risk of dehydration (Section 4.1).

5.2.1. ORT versus IVT for children with all degrees and types of dehydration

5.2.2. ORT versus IVT for children with severe dehydration

5.2.3. ORT versus IVT for children with hypernatraemic dehydration

Evidence overview

No published study was identified that examined the effectiveness of fluids other than ORS solution in the treatment of dehydration. However, one study⁸⁷ characterised the composition of a wide range of readily available fluids and commercially produced drinks. The analysis included tap water and about 90 commercial ‘clear’ fluids including soups, juices, fruit-flavoured drinks and carbonated drinks. Sodium concentrations ranged from 0.1 to 251 mmol/l, potassium concentration from 0.0 to 65 mmol/l, and osmolalities ranged from 246 to 2000 mOsm/l. Among the fruit juices tested (apple, grape, lemon, orange), none had a sodium concentration greater than 10 mmol/l or a potassium concentration less than 24 mmol/l. The osmolality of the soups tested ranged from 293 to 543 mOsm/l. Soups prepared from crystals had slightly higher osmolalities than those prepared from liquid concentrates.

Evidence summary

No study was found to evaluate the effectiveness of juices, tap water or other commercial clear fluids in the treatment of dehydration. Evidence from one cross-sectional study showed great variation in the concentration of sodium and potassium, and in the osmolarity of readily available commercial clear fluids such as juices, soups and carbonated drinks.

GDG translation from evidence to recommendation

Although there were no clinical trials on the effectiveness of fluids other than ORS solution in the treatment of dehydration, the GDG considered that the composition of such fluids was generally inappropriate. In dehydration due to gastroenteritis, both water and electrolyte replacement is essential, and non-ORS solution fluids do not usually contain appropriate constituents. ORS solution was considered the appropriate fluid for oral rehydration.

ORS solution has been manufactured using a range of constituents in differing concentrations. Various organic solutes have been included such as glucose, starch and amino acids. Sodium chloride has been used in varying concentrations. Other non-essential constituents, including potassium, bicarbonate and acetate are often included. Much research has been carried out to evaluate the effectiveness and safety of these various solutions. Two key areas of research have focused on the optimal sodium/osmolar concentration in ORS solution and on the relative efficacies of glucose versus rice starch as the organic constituent in ORS solution.

The composition of the original WHO ORS solution (glucose 111, sodium 90, potassium 20, chloride 80 and bicarbonate 30, all in mmol/l) was selected to allow for use of a single solution that would effectively treat dehydration secondary to diarrhoea caused by various infectious agents and resulting in varying degrees of electrolyte loss.¹⁹ However, in developed countries, viral gastroenteritis is common and is associated with less severe salt losses, and so there was concern that the sodium content of the original WHO ORS solution might be excessive.⁸⁸ From the 1970s, efforts focused on improving the efficacy of ORS solution by altering its composition. It was found that solutions with higher concentrations of co-transporters (such as sugars) and higher osmolarity decreased rather than increased intestinal sodium and water absorption. Additionally, hypernatraemia was reported with their use. The current formulation WHO ORS solution adopted in 2002 (glucose 75, sodium 75, potassium 20, chloride 65 and citrate 10, all in mmol/l) preserves the 1 : 1 molar ratio of sodium to glucose that is critical for efficient co-transport of sodium. It has a reduced osmolar load (245 mOsm/l) compared with the original formulation (311 mOsm/l). It also has a longer pre-mixed shelf life owing to its citrate content.

The evidence searches on this question were limited to include only those studies that compared the effectiveness of high-osmolarity/low-sodium ORS solution with low-osmolarity/low-sodium ORS solution or the glucose-based ORS solution with the rice-based ORS solution. Evidence on other types of ORS solution using different carbohydrate substrates or organic substitutes such as cereals or amino acids was not reviewed in this section since these products are not available in the UK and are not currently recommended by the WHO.

Evidence overview

Three systematic reviews have been included – two^89,90 provided evidence relating to effectiveness of low-sodium/low-osmolarity ORS solution versus high-sodium/high-osmolarity ORS solution, while the third review⁹¹ compared rice-based ORS solution with the glucose-based ORS solution.

One systematic review⁸⁹ compared the effectiveness of the previously recommended WHO ORS solution (osmolarity 311 mmol/l with 90 mmol/l of sodium) with reduced osmolarity ORS solution (osmolarity 250 mmol/l or less with reduced sodium) in the treatment of children with acute diarrhoea. Only RCTs with adequate randomisation were considered for inclusion while quasi-randomised trials were excluded. Participants included children with acute diarrhoea (with history of less than 5 days). The primary outcome of interest was the need for ‘unscheduled’ IV infusion during the course of treatment, while the secondary outcomes were stool output, vomiting and asymptomatic hyponatraemia (serum sodium < 130 mmol/l) during follow-up. Results from the various studies were pooled using the fixed effect model. [EL = 1++]

Fourteen RCTs were included in this review and they were conducted in Egypt (two), Bangladesh (three), Mexico (one), Columbia (one), India (three), Panama (one) and the USA (one). All the studies recruited children younger than 5 years suffering from acute non-cholera diarrhoea, with the exception of three trials that did include children with cholera. In five trials, children with severe dehydration were included while five other trials included malnourished children. Nine trials reported allocation concealment that was adequate and six were double-blinded. Loss to follow-up was less than 10% of the randomised participants in all trials. The protocol of this review had initially defined reduced osmolarity as < 250 mOsm/l but during the course of reviewing this limit was increased to 270 mOsm/l or less since some trials had used this higher limit of definition. Osmolarity of the control group was also increased from 311 to 331 mmol/l to include two additional trials. Nevertheless, in both the trials the concentration of sodium and glucose was similar to the WHO-recommended ORS solution. Since stool output was measured in various ways using different units in the RCTs, their results were pooled and expressed as standardised mean difference.

Out of 14 trials, 11 reported the need for unscheduled IVT. Three trials reported that none of the children required an IV infusion and hence odds ratios could not be calculated. Results from the meta-analysis of the other eight trials (n = 1996) showed a significant reduction in the need for additional IVT for children receiving the low-osmolarity ORS solution compared with children treated with the WHO-recommended high-osmolarity ORS solution (OR 0.59; 95% CI 0.45 to 0.79). Sensitivity analysis conducted with studies where allocation concealment was clearly described as adequate suggested little difference in the result for the primary outcome. Eleven trials (n = 1776) measured stool output during the rehydration period and the pooled results showed a significant reduction in stool output with the low-osmolarity ORS solution (SMD −0.23; 95% CI −0.33 to −0.14). Hyponatraemia and vomiting during rehydration were reported in six trials each. Children treated with the reduced osmolarity ORS solution showed a lower tendency for vomiting (OR 0.71; 95% CI 0.55 to 0.92) compared with the WHO ORS solution group, but no statistically significant difference was observed for the presence of hyponatraemia (OR 1.44; 95% CI 0.93 to 2.24). There was no evidence of statistical heterogeneity for any of the results.

In the second systematic review from the USA,⁹⁰ the effectiveness of ORT was evaluated in comparison with IVT among well-nourished children with gastroenteritis in developed countries, and this was followed by a comparison between high-sodium ORS solution and low-sodium ORS solution. Trials were included if they were published in English, conducted in populations of well-nourished children during the late 1970s through to the early 1990s and included more than ten patients. A total of 13 trials were included in this review and all were conducted in the USA or Canada – six RCTs of ORT with IVT arms and seven RCTs without IVT arms (i.e. comparing oral solutions with differing sodium content). The age of the study population ranged from 3 months to 3 years but one study included children aged 1 month to 14 years. Clinical efficacy was defined as the success of ORT in rehydrating children with gastroenteritis within 12–24 hours of starting treatment, while failure was defined as the need to use IVT for rehydration. High sodium content was defined as a sodium level of 90 mmol/l, medium as 50–75 mmol/l and low as 26–45 mmol/l. Safety was measured by the relative incidence of hypernatraemia (serum sodium level > 146 mmol/l) and hyponatraemia (serum sodium level < 132 mmol/l) induced by the treatment. [EL = 1+]

Altogether, eight trials (one RCT with an IVT arm and seven trials without an IVT arm) had compared ORT solutions of differing sodium content and their results are reported in this section. The high-sodium formula had the lowest failure rate among the three groups at 1.9% (95% CI 0% to 5.4%), while the low-sodium group had a failure rate of 3.6% (95% CI 0% to 7.3%) and the medium-sodium group a failure rate of 5% (95% CI 1.9% to 8.1%). However, there were no statistically significant differences in the failure rates of the three groups treated with high-, medium- and low-sodium ORS solution. Only one trial with an IVT arm gave information on the cases of hypo- and hypernatraemia. It reported three cases of hyponatraemia that corrected to normal within 24 hours of treatment. Another trial with no IVT arm reported one case of hyponatraemia in the high-sodium group and six cases each in the medium- and low-sodium groups. Estimates of effect could not be calculated for incidences of hyponatraemia and hypernatraemia because the total numbers of individuals in each group were not available. Moreover, there were no statistically significant differences between the high- and low-sodium ORS solution for other outcomes (weight gain, volume, frequency and duration of diarrhoea, or length of hospital stay).

A Cochrane review⁹¹ compared rice-based ORS solution (50–80 g/l of rice powder) to the glucose-based WHO ORS solution (20 g/l of glucose) for the treatment of diarrhoea. Trials were included only if the rice-based ORS solution was made by replacing glucose in the standard WHO ORS solution with 50–80 g/l of rice powder and all the other electrolyte concentrations remained the same. Participants included both children and adults with signs of dehydration due to acute diarrhoea. The outcomes reported were stool output during the first 24 hours, total stool output (from admission to cessation of diarrhoea) and duration of diarrhoea. For studies in children, data from girls were excluded owing to difficulty in measuring the stool volumes.

Of 22 hospital-based RCTs included for this review, 12 trials included children – five with cholera and seven with non-cholera diarrhoea. Four of these trials were conducted in Bangladesh, two in India, and one each in Indonesia, Pakistan, Mexico, Chile, Peru and Egypt. Two trials included children younger than 6 months whereas the others included children above 4–6 months. Allocation concealment was adequate in 15 of the 22 studies while the method of randomisation and concealment was not reported in the remaining studies. None of the trials reported whether patients with severe dehydration were randomised or whether outcome measurement started before or after initial IVT. In six trials, 1–15% of randomised patients were excluded from the final analysis, but these patients should have continued to be monitored and their data included using intention-to-treat analysis. Outcomes were analysed separately for children with cholera and non-cholera diarrhoea. [EL = 1++]

Twelve studies (n = 2854) reported the duration of diarrhoea for children younger than 5 years suffering from non-cholera diarrhoea. Results from the meta-analysis suggest that children receiving rice-based ORS solution had shorter duration of diarrhoea compared with those receiving the glucose-based ORS solution, but the difference was not statistically significant (WMD −1.26 days; 95% CI −4.4 to 1.9 days). There were also no statistically significant differences between the two types of ORS solution for the outcomes of stool output in the first 24 hours of the intervention (15 trials; WMD −4.3 g/kg; 95% CI −9.4 to 0.8 g/kg) or total stool output (nine trials; WMD −28.2 g/kg; 95% CI −52.4 to 3.9 g/kg). For children with cholera, a significant reduction in the 24 hour stool output was seen with rice-based ORS solution (four trials; WMD −67.4 g/kg; 95% CI −94.3 to −40.5 g/kg). Only one trial (n = 48) reported total stool output and duration of diarrhoea in children with cholera and that showed a significant decrease in both outcomes with the use of rice-based ORS solution.

Evidence summary

There is evidence from one high-quality systematic review [EL = 1++] indicating a significant reduction in the need for unscheduled IV fluids for the treatment of dehydration in children with diarrhoea when using low-osmolarity ORS solution compared with the previously recommended high-osmolarity WHO ORS solution. Moreover, the results suggest that low-osmolarity ORS solution leads to a greater reduction in stool output and vomiting. No difference was seen in the incidence of hyponatraemia. The other systematic review [EL = 1+] reported no statistically significant difference between ORS solutions with different sodium contents in terms of treatment failure (need for IVT) in well-nourished children with gastroenteritis and dehydration. It also failed to show any consistent trend in favour of either high- or low-sodium ORS solution for rehydration.

Evidence from a well-conducted systematic review [EL = 1++] found no statistically significant differences in stool output or duration of diarrhoea when children with non-cholera diarrhoea were treated with rice-based ORS solution compared with the traditional glucose-based ORS solution. However, in children with cholera, rice-based ORS solution was associated with a reduction in stool output and duration of diarrhoea.

GDG translation from evidence to recommendation

ORS solution of reduced osmolarity (<270 mOsm/l) is preferable to solutions with a high osmolarity (>311 mOsm/l). Reduced osmolarity ORS solution was associated with a lower incidence of failure to rehydrate and also a greater reduction in stool output.

The original (pre-2002) WHO ORS solution had an intermediate osmolarity of 311 mOsm/l and a sodium concentration of 90 mmol/l. In 2002, the WHO recommended a new ORS solution formulation of lower osmolarity and sodium concentration (245 mOsm/l and 75 mmol/l, respectively). This is closer in composition to the ORS solution products commonly used in the UK (sodium concentrations 50–60 mmol/l, osmolarities 240–250 mOsm/l; see Table 5.1).

Table 5.1

Compositions of WHO ORS solution and of ORS solution products available in the UK.

There are no studies that compare the new (2002) WHO ORS solution with products currently used in the UK. It is therefore unknown whether there are any clinically important differences between these solutions, particularly in terms of rehydration failure, cessation of vomiting, duration or volume of stool losses, or incidence of symptomatic hyponatraemia. Hyponatraemia was important to consider as it has been suggested that low-sodium ORS solution might be associated with hyponatraemia. While some studies suggested this might be so, the effect was not statistically significant and there were no reports of clinically significant hyponatraemia with adverse effects such as convulsions. There were no studies with hyponatraemia as a primary outcome. The GDG therefore concluded that ORS solution products currently available in the UK were appropriate for use in ORT.

Cereal-based ORS solution may have a beneficial effect in reducing diarrhoeal losses compared with glucose-based ORS solution. However, the available evidence applied only to cholera and was of low quality. Rice-based ORS solution is currently not available in the UK. The GDG did not consider that there was evidence to support its use.

No studies were identified that compared the clinical effectiveness of various oral fluid regimens in the treatment of dehydrated children with gastroenteritis. Likewise, there were no studies that specifically addressed the optimal volume of fluid to be used for rehydration, the optimal route of administration (bottle, beaker, cup, spoon, syringe or nasogastric), the frequency of administration, the time interval over which rehydration should be attempted, or the indicators for reintroduction of oral fluids after IVT.

Evidence overview

Since no study was identified directly answering the question, evidence was included from three RCTs that had recruited children with dehydration for the primary purpose of comparing different ORS solution products. The study populations in all three trials included some children presenting with severe dehydration and/or shock, and all these children were started on ORT following initial rehydration with IVT.

In the first multicentre trial,⁹² the efficacy of reduced osmolarity ORS solution was compared with that of the pre-2002 WHO ORS solution. Children presenting with severe dehydration were initially rehydrated with IVT for 2 hours and then randomised to the two groups as soon as they were able to accept fluids orally. The second trial⁹³ compared oral glucose electrolyte solution with oral sucrose solution in equimolar concentrations using the WHO-recommended electrolyte formula. In children with severe dehydration (fluid deficit ≥ 10% body weight), 70% of their estimated fluid deficit was replaced within the first 2 hours by IVT and further rehydration was achieved by ORT. The third trial⁹⁴ evaluated the safety and efficacy of glycine-based ORS solution compared with ORS solution containing no glycine. Children with severe dehydration in both groups were initially given IVT until blood pressure and pulse returned to normal, and then rehydration was completed within 4 hours by giving either of the two ORS solutions.

Evidence summary

There were no studies that provided direct evidence on the effectiveness of various oral fluid regimens in terms of the route of administration, frequency of administration or volume of fluid to be used. However, the procedures used in studies suggest that children with gastroenteritis and severe dehydration can be successfully rehydrated with ORT after an initial rehydration with IVT. After an initial rehydration with IVT, ORT was usually introduced within 2–4 hours of starting rehydration.

GDG translation from evidence to recommendation

There was no evidence to support recommendations on how much fluid should be given, and over what time frame, when treating a dehydrated child. The WHO recommends rehydration over a period of 4 hours.²⁰ The GDG agreed that clinical experience showed this to be generally possible. It was considered important to achieve rehydration as quickly as possible, but more rapid rehydration might be associated with an increased risk of vomiting.

The traditional approach to oral rehydration has been to offer ORS solution in small quantities at frequent intervals. That seemed appropriate advice and the GDG agreed that it would improve tolerance.

Given that precise determination of dehydration severity is not possible, the exact volume of fluid required for rehydration cannot be calculated accurately at the outset. At the mildest end of the spectrum, dehydration may be clinically undetectable. It is likely that clinical signs of dehydration first become apparent in patients with about 3–5% weight loss. Children who are at the most severe end of the spectrum may have lost 10% or more of their body weight. The GDG therefore considered that a reasonable approach in a child presenting with clinical manifestations of dehydration was to assume 5% dehydration at the outset. Based on that assumption, rehydration should be attempted by giving 50 ml/kg over the initial 4 hour rehydration period. In some cases, this may be somewhat more than is required, but that will be of no clinical consequence. In other more severely dehydrated children, 50 ml/kg may be insufficient. It would therefore be important to regularly reassess the child’s state of hydration and, when necessary, to increase the final volume of replacement fluid administered (see worked example in Tables 5.2 and 5.3). Children with red flag symptoms or signs (see Table 4.6) would require frequent reassessment during rehydration, with adjustment of the deficit replacement depending on that assessment.

Table 5.2

Worked example of an oral rehydration strategy in a 12-month-old child weighing 10 kg.

Table 5.3

Maintenance fluid volume requirements based on body weight.

The use of a nasogastric tube to deliver ORS solution is common but not universal practice. It may allow oral rehydration of children who will not drink ORS solution. If children vomit ORS solution persistently, continuous infusion through a nasogastric tube may improve tolerance but there are no studies on this method of administrating ORT. Placement of a nasogastric tube may be somewhat unpleasant or distressing for children. There are possible complications associated with nasogastric feeding.⁹⁶ These concerns needed to be balanced against the alternative of IVT, which might also be associated with distress and complications such as phlebitis or cellulitis. The GDG considered that nasogastric fluid administration was a reasonable and perhaps preferable alternative to IVT for some children. However, each case needed to be assessed on its own merits. Moreover, studies comparing nasogastric tube therapy with IVT should be undertaken.

Introduction

Although most children with dehydration can be successfully managed using ORT, occasionally IVT may be indicated. In children with hypovolaemic shock, immediate IVT might be required. For various reasons, ORT might be unsuccessful and so IVT might be necessary. The GDG considered the indications for starting IVT, the rate at which IV rehydration should be performed, the choice of IV fluid, and the option of changing from IVT to ORT to complete the rehydration process.

A systematic literature search identified 381 publications and 26 articles were retrieved for reviewing. Secondary screening of these studies led to four studies being finally included as evidence; however, these studies were of poor quality or gave indirect evidence to inform the questions. The majority of questions included in this section yielded no evidence considered to be of sufficient quality on which to base recommendations. Thus the recommendations developed in this section are based mostly on a combination of non-RCT studies, observational data, anecdotal reports and expert consensus opinion.

Evidence overview

No studies were identified which gave evidence on the indications for starting IVT in children with dehydration

GDG translation from evidence to recommendation

In the absence of any published evidence to inform this question, the GDG based its recommendation on consensus decision. The group fully accepted established practice with regard to the initial management of patients with shock.⁹⁷ Consequently, all children with hypovolaemic shock due to dehydration require IVT. Patients with severe dehydration may be at risk of becoming shocked. As discussed in Chapter 4, the clinical features associated with severe dehydration may partially overlap those associated with shock. The GDG considered that if the clinician was uncertain as to whether the child was actually in a state of shock, the child should receive IVT.

The GDG agreed that. although ORT was recommended as the first-line treatment for dehydration. it was occasionally unsuccessful. In cases where, despite appropriate treatment with ORS solution, the child’s state of hydration does not improve or where it shows signs of worsening, IVT would be required. For example, a child might fail to tolerate the necessary quantities of oral (or perhaps nasogastric tube administered) ORS solution.

The decision to use IVT should only be taken, however, following a thorough assessment of the child’s condition and careful consideration as to whether ORT is truly failing. It is also important that the child’s state of hydration be monitored carefully and regularly.

A range of IVT regimens and fluids have been employed in the treatment of dehydration and shock. The fluids have included Ringer’s lactate and saline in various concentrations.

Evidence overview

No study was identified which specifically addressed the question of immediate IVT regimen in children with hypovolaemic shock. For the second question, one good-quality RCT was found that compared 0.9% saline plus 2.5% dextrose (NS) with 0.45% saline plus 2.5% dextrose (N/2) for the treatment of dehydration.

In the absence of any direct evidence to answer the first question, information was collected from various studies which had described fluid regimens in the treatment of dehydration in children with severe dehydration and/or shock. The various regimens are summarised in Table 5.4.

Table 5.4

Fluid regimens used for rehydration in children in different studies.

For the optimal composition of fluid for IVT, one study was identified. It was a prospective randomised study¹⁰² conducted in Australia to determine whether the use of 0.9% saline rather than 0.45% saline reduced the risk of hyponatraemia. Children with gastroenteritis aged between 6 months and 14 years were eligible for enrolment in the study only after a decision to treat with IVT was made by their treating physician, independent of the study. IVT was administered for dehydration if, while being treated in the emergency department, children were vomiting or had an inadequate intake of oral fluids. None of the children were severely dehydrated. After enrolment, participants were randomly allocated to receive either 0.9% saline plus 2.5% dextrose (NS) or 0.45% saline plus 2.5% dextrose (N/2). The rate of infusion was decided by the treating physician. The options used were a ‘rapid replacement protocol’ (RRP) consisting of 10 ml/kg per hour for 4 hours or a slow replacement protocol in which children received their fluid deficit based on estimated percentage dehydration over a 24 hour period (in addition to their maintenance fluids). The primary outcome examined was the incidence of hyponatraemia defined as plasma sodium < 135 mmol/l. The authors presented the results separately for those with hyponatraemia and those with normal plasma sodium levels measured prior to starting IVT.

Altogether, 102 children were enrolled in the study. Of these, 36% (37/102) were hyponatraemic before starting IVT. The median duration of illness prior to presentation was longer in the hyponatraemic group than in those with normal plasma sodium, but apart from this there were no statistically significant differences in their baseline clinical characteristics or biochemical test results. In total, 51 children were randomly assigned to each treatment group. In those with initial hyponatraemia given 0.45% saline (n = 16), there was no change in mean plasma sodium after 4 hours, but in those with an initially normal plasma sodium (n = 35) there was a significant decrease in the mean sodium concentration after 4 hours (135 ± 1.8 mmol/l versus 137 ± 1.7 mmol/l; P < 0.001). Hyponatraemic children given 0.9% saline (n = 21) had a significant increase in mean sodium concentration (134 ± 2.1 mmol/l versus 132 ± 2.4 mmol/l; P < 0.001), but in those with an initially normal plasma sodium (n = 30) there was no statistically significant change. [EL = 1+]

Evidence summary

No study was identified which gave direct evidence on the immediate IVT regimen in children with severe dehydration and/or hypovolaemic shock. However, processes followed in various trials suggest that these children were initially rehydrated with Ringer’s lactate solution given at a rate of 20–40 ml/kg per hour over a period of 1–2 hours or until there was improvement in blood pressure and pulse volume.

Evidence from another RCT [EL = 1+] suggested that rehydration with 0.9% saline IVT led to a significant increase in the mean plasma sodium levels in children with hyponatraemic dehydration while the use of 0.45% saline did not correct this abnormality. Moreover, the use of 0.45% saline was associated with a significant decrease in the plasma sodium concentration in those with normal plasma sodium concentrations prior to IVT while the use of 0.9% saline was not.

No study was identified that provided direct evidence on the volume of deficit or the duration over which IV rehydration should be performed. For the second part of the question, three papers were identified.

Evidence overview

All three studies considered the role of ‘rapid rehydration therapy’ or delivering fluid deficit replacement over a short period of time in children with severe dehydration (without shock) who were unable to tolerate oral fluids.

The first was a prospective study with historical controls conducted in Australia¹⁰³ to evaluate the effectiveness of rapid rehydration with IV fluid or with ORT (administered through nasogastric tube) in the treatment of moderately dehydrated children. Inclusion criteria were age 6 months to 16 years, duration of illness less than 48 hours, presence of vomiting and diarrhoea with mild to moderate dehydration, normal respiratory rate and level of consciousness, and a CRT of less than 2 seconds. All the participants were initially given a trial of oral fluids using Gastrolyte-R® or apple juice diluted to 25% (2.5 g carbohydrate, 1.25 mg sodium, 20 mg potassium) if the former was refused. Parents were educated by nurses on the importance of initial oral rehydration. Moderately dehydrated children who were unable to tolerate 100 ml of oral fluid over 1 hour (50 ml for children younger than 2 years) were given rapid rehydration. The options for administration were intravenously using N/2 saline + 2.5% dextrose over 2 hours at 20 ml/kg per hour or by nasogastric tube with Gastrolyte-R at the same rate. Following rapid rehydration, children were given another trial of 100 ml of oral fluid (50 ml for children younger than 2 years) over 1 hour. Children who tolerated and satisfied the discharge criteria were discharged while those not tolerating orally were admitted to the hospital to continue rehydration. The historical control group was made of children admitted 2 years earlier in the same hospital with a similar diagnosis, and their hospital records were checked for data collection. These children were given a non-standard regimen of initial oral fluid trial, failing which they were rehydrated intravenously over a period of 24 hours. There was no specific education on oral fluid therapy geared towards parents and volume of fluid intake was estimated from parents’ reports. Outcomes reported were admission to hospital, discharge in 8 hours or less after presentation to the emergency department and re-presentation requiring admission within 48 hours of discharge from the emergency department. The outcomes were measured for moderately dehydrated patients as well as for mildly dehydrated patients. Patients having rehydration via the IV route or the nasogastric tube route were analysed together. [EL = 2−]

In this study, 145 patients were recruited in the intervention group (rapid rehydration therapy) and 170 in the control group (IV rehydration over 24 hours). The two groups were similar with regard to age and sex but the intervention group had recruited significantly more moderately dehydrated children. On comparing moderately dehydrated patients only between the two groups, a statistically significant reduction was observed in the hospital admission rates in the intervention group compared with the control group (55.8% versus 96.3%; P < 0.001). Moreover, significantly more patients in the intervention group were discharged at 8 hours or less after presentation to the emergency department (44.2% versus 3.7%; P < 0.001). No statistically significant difference was seen for rates of re-presentation requiring admission within 48 hours of discharge from the emergency department. For mildly dehydrated patients in the two groups, no statistically significant difference was seen for the above outcomes. In the intervention group, electrolytes were analysed for 78 children and 17 were found to be hyponatraemic on initial assessment. Two of these patients presented with serum sodium levels < 130 mmol/l (128 and 125 mmol/l). However, they did not suffer from any complications or clinical sequelae and their serum sodium levels returned to normal levels by 12 hours.

The second study was a prospective cohort study from the USA¹⁰⁴ that evaluated the efficacy of rapid IV rehydration in children with mild to moderate dehydration due to gastroenteritis in the hospital outpatient department. The criteria for inclusion were age at least 6 months, clinical diagnosis of acute gastroenteritis with exclusion of other causes, vomiting for less than 48 hours in duration with at least five episodes in the 24 hours preceding presentation, presence of normal serum sodium levels (130–149 mEq/l) and metabolic acidosis (serum bicarbonate < 18 mEq/l) at the time of presentation. Each patient received an infusion of 20–30 ml/kg isotonic crystalloid solution over 1–2 hours, followed by a trial of oral rehydration. Children who subsequently vomited were admitted for continued IV rehydration therapy, while those tolerating oral fluids were discharged with home-care instructions. To identify variables that might identify children who would not tolerate oral fluids after outpatient rapid IV rehydration, regression analysis was conducted with data from the two groups of children – those successfully tolerating oral fluids and those requiring admission for continued IVT. [EL = 2−]

This study enrolled a convenience sample of 58 children with age ranging from 6 months to 13 years (median age 22 months). One-third of the children were clinically assessed to have moderate dehydration (deficit of 6–10 % body weight) while the rest had mild dehydration. After rapid outpatient IV rehydration, 16 patients (28%) did not tolerate oral fluids while the rest 42 (72%) tolerated orally and were discharged home. The baseline characteristics of these two groups were not described. Of the discharged patients, 14% (6/42) were re-admitted owing to recurrent vomiting and dehydration. A significantly higher proportion of children who did not tolerate orally after rapid IV rehydration had metabolic acidosis (69% versus 2%; P < 0.001) and were moderately dehydrated (56% versus 24%; P < 0.01) compared with the patients discharged home. There were no differences between the two groups regarding the age and severity of diarrhoea or vomiting.

In another non-comparative study, from Canada,¹⁰⁵ children aged between 1 and 6 years with mild or moderate dehydration secondary to gastroenteritis were recruited. Children were included if they had diarrhoea and/or vomiting for less than 5 days with mild to moderate dehydration, had normal nutritional status and were unable to retain small amounts of clear fluid or refused to take them. Children who had taken medication, those having an underlying disease and those with electrolyte abnormalities were excluded. A trial of rehydration was initially attempted with small amounts of clear fluids (the authors did not specify how they defined ‘clear fluid’), and if the fluid was refused or vomited, the child was considered for the study. IVT was administered by giving 3.3% dextrose and 0.3% saline at a rate of 10 ml/kg per hour for 3 hours (total 30 ml/kg). During IVT, patients did not receive any oral fluid. Discharge was allowed if there were no clinical signs of dehydration, no persistent vomiting, normal central nervous system examination and if the parents felt the child had improved. [EL = 3]

Seventeen children (mean age 2.6 ± 1.7 years) met the study inclusion criteria. All had had vomiting for an average 2.1 ± 1.2 days prior to presentation at the emergency department, and 10 of them had had diarrhoea for the preceding 1.9 ± 1.9 days. Seven children had at least 6% dehydration and 7/12 (58.3%) had mild metabolic acidosis with a base deficit of 5 or more. All patients improved after IVT and only 6/17 had vomited after therapy. One patient continued vomiting till 48 hours after IVT and required another course of IVT, following which there was no vomiting. None of the patients required hospital admission after discharge from the emergency department.

Evidence summary

There was a lack of good-quality evidence available for the clinical effectiveness of rapid IV rehydration in children with gastroenteritis and moderate or severe dehydration. The first study with a historical control group [EL = 2−] suggested that rapid rehydration by ORT or IVT in moderately dehydrated children led to a significant reduction in the hospital admission rate and an increase in discharge from the emergency department within 8 hours of presentation. No statistically significant difference was seen for these outcomes in the group of mildly dehydrated children. Results from the other two studies (a poorly conducted cohort study and a non-comparative study) showed rapid IVT to be successful in achieving rehydration in most of the dehydrated children. However, the study population in these two studies was not homogeneous and included children with mild dehydration. In the cohort study, more than 70% of the children given rapid rehydration were able to tolerate orally and the majority of children not tolerating orally had metabolic acidosis and/or moderate dehydration.

GDG translation from evidence to recommendation

There was no definitive evidence on the optimum IV fluid regimen for the management of hypovolaemic shock in the dehydrated child with gastroenteritis. However, there was widespread consensus that whatever the cause of shock, a bolus of IV fluid should immediately be given.^97,106

Currently, the main focus of discussion regarding fluid administration in patients with shock is on the optimal choice of fluid. This includes discussion of crystalloid versus colloid fluids, especially in relation to septicaemic or critically ill patients. The use of albumin solutions in particular is controversial.^107–109 The GDG considered that for children with shock due to dehydration from gastroenteritis, 0.9% sodium chloride solution is an appropriate and readily available fluid for bolus administration.

In hypovolaemic shock associated with gastroenteritis, rapid recovery would be expected in most children following the administration of a 20 ml/kg bolus of 0.9% sodium chloride solution.^97,106 If an immediate response does not occur, a further 20 ml/kg bolus should be given. In the absence of prompt response to bolus fluid administration, it is important to consider whether factors other than hypovolaemia, for example septicaemia, might be responsible. In this case, expert advice should be sought from a paediatric intensive care specialist.

With regard to the treatment of dehydration (as opposed to shock) with IVT, the GDG recognised that there was a lack of evidence from clinical trials to inform practice in this area. For example, the WHO recommends the use of Ringer’s lactate solution for IV rehydration. There has been much discussion on the optimal choice of IV maintenance fluids for children, particularly since the National Patient Safety Agency (NPSA) issued a specific alert on this matter in 2007.¹¹⁰ The NPSA alert reinforced an existing trend away from the use of hypotonic solutions such as 0.18% sodium chloride. This reflected concerns about the risk of hyponatraemia.¹¹¹

The NPSA has recommended the use of isotonic solutions such as 0.9% sodium chloride with or without 5% glucose for children with a range of conditions associated with a risk of hyponatraemia, including those with diarrhoea. There has been some controversy regarding the NPSA recommendations. The use of 0.9% sodium chloride solution provides more salt to the child than is to be found in an average diet. It appears that most children do not experience adverse effects with this increased salt load, but some may do.¹¹² Concern has been expressed at a possible risk of hypernatraemia and hyperchloraemic acidosis. However, because gastroenteritis is associated with increased salt losses, the use of 0.9% sodium chloride solution in these children has not been challenged. Furthermore, the NPSA recommendations emphasise the importance of careful monitoring of plasma electrolytes during IVT.

The GDG agreed with the NPSA recommendations on this matter. The GDG also considered that it was important for clinicians to be aware that metabolic acidosis might not be due to persisting shock. Such misinterpretation could lead to the administration of excessive fluid.¹¹³

As discussed earlier in relation to ORT, there is usually no precise way of determining the exact fluid deficit and hence the required replacement volume. If a child is considered to be in a state of hypovolaemic shock, their intravascular volume has been significantly depleted owing to fluid loss. To reach this stage, it is generally accepted that they are at least 10% dehydrated, i.e. they will have lost at least 100 ml/kg of total body fluid. The GDG agreed that once shock was corrected with fluid bolus administration, it was safest to assume that the child might be 10% dehydrated and they should therefore receive 100 ml/kg as deficit replacement.

For those who are dehydrated but without shock, in keeping with the recommendation for ORT, a volume of 50 ml/kg IV was considered by the GDG to be an appropriate initial deficit replacement volume. Regular reassessment would be necessary to determine whether this was sufficient to reverse the signs of clinical dehydration. Further replacement would be given subsequently if this was necessary.

The GDG was conscious of a lack of high-quality evidence regarding the optimal time period over which IV fluids should be administered to children with dehydration. With ORT it is widely accepted that rehydration can and should be completed quickly – typically over a 4 hour period. Traditionally, IV rehydration has been performed over a 24 hour time period, and this approach is still in common use. This practice probably evolved from experience with the use of subcutaneous fluid replacement in children in the 1940s.¹¹⁴ The NPSA patient safety alert on IV infusions for children¹¹⁰ stated that fluid deficits should be replaced over a minimum of 24 hours. That alert was published with the intention of minimising the risk of hyponatraemia when administering IV infusions to children. The GDG was conscious of the fact that opinions regarding the optimal rate of IV fluid replacement in dehydration due to gastroenteritis are varied, and more rapid rehydration regimens have been widely advocated. In 1996, the American Academy of Pediatrics recommended a rapid rehydration model, for example giving 20 ml/kg or more over the first hour depending on the individual child’s clinical condition.⁵ For many years, the WHO has advocated rapid IV rehydration for severe dehydration, recommending administration of 100 ml/kg over the first 6 hours, and even more rapid rehydration in older children.⁷⁶ The GDG was aware of an increasing trend towards the use of rapid IV rehydration regimens in children with gastroenteritis, as illustrated by the studies listed in Table 5.4. It has been argued that rapid administration of IV fluids may improve gastrointestinal perfusion so that oral feeding can be reinstituted earlier, and that improved renal perfusion may assist in correcting acidosis and electrolyte imbalances.¹¹⁴ The GDG was aware of a study on rapid IV rehydration currently in progress at The Hospital for Sick Children, Toronto, Canada (see www.clinicaltrials.gov/ct2/show/NCT00392145). In this trial, children attending the emergency department with dehydration secondary to gastroenteritis and requiring IV rehydration are being randomised to a 20 ml/kg 0.9% normal saline bolus over 1 hour or to ‘rapid intravenous rehydration’ consisting of a 60 ml/kg 0.9% normal saline bolus over 1 hour. The primary outcome measure in this study is the clinical rehydration status after 2 hours, and secondary outcomes include duration of hospitalisation and ability to tolerate oral rehydration.

The GDG concluded that it was not currently possible to make a clinical recommendation on the optimal rate of IV fluid administration in children with dehydration due to gastroenteritis. The GDG considered that this was a crucially important matter, and a priority area for research. A research recommendation was therefore made regarding rapid IV rehydration. Importantly, the GDG did recommend early implementation of ORT to complete rehydration as soon as fluids are tolerated.

The GDG considered that it is essential that plasma electrolyte concentrations are measured at baseline when commencing IVT, and regularly thereafter. The NPSA patient safety alert on administering IV infusions to children emphasises the importance of monitoring the plasma sodium concentration regularly in order to avoid dangerous hyponatraemia or hypernatraemia.¹¹⁰ The GDG did not consider that it could make precise recommendations on the frequency with which blood testing should be undertaken, and this would depend both on the results of prior tests and on the particular risk factors in individual cases. Moreover, it did not consider that it should provide specific advice on the actions needed in the event of electrolyte disturbances being present. Consideration should be given to including potassium supplementation in the IV fluid solution following measurement of the plasma potassium concentration. The fluid solutions recommended in this guideline are all available preconstituted with potassium, at either 10 mmol/l or 20 mmol/l.

The GDG recognised the specific risks associated with hypernatraemic dehydration in gastroenteritis and the importance of safe management. In such cases, expert advice on fluid management should be sought. No high-quality evidence was found on the relative safety of different fluid regimes. However, the GDG recognised that there is common agreement that in those with clinically significant hypernatraemia (plasma sodium concentration > 160 mmol/l), fluid deficit replacement should be delivered slowly (typically over 48 hours) using an isotonic solution (0.9% sodium chloride solution). The NPSA advice on IV infusions is that plasma sodium should be reduced at a maximum rate of 0.5 mmol/l per hour, or more slowly if it has prevailed for more than 5 days. Frequent monitoring of the plasma concentration is therefore essential in such patients.

Evidence overview

In the absence of any direct evidence to answer this question, information was again collected from various studies which had described fluid regimens in which ORS solution was introduced after a period of initial IVT in children with severe dehydration and/or shock. The different fluid regimens have been tabulated in Table 5.4. In these studies, rapid IV fluid was given initially (for example, 20–40 ml/kg per hour) and ORT usually introduced after about 1–2 hours to complete the rehydration process.

Evidence summary

No direct evidence was identified on the appropriate time for switching IVT to ORT. Processes followed in various studies indicate that children with severe dehydration and/or hypovolaemic shock were initially rehydrated with IVT over a period of 1–2 hours or until there was improvement in blood pressure and pulse volume, and ORT were usually introduced after this period to complete the rehydration process.

GDG translation from evidence to recommendation

In current practice the GDG believed that once IVT is begun, children often remain in hospital for lengthy periods, for example 24 hours or more. Although formal research trials are not available, nevertheless clinical studies have reported success with regimens in which children with shock or severe dehydration due to gastroenteritis received IVT for 1–2 hours, with subsequent rehydration given as ORT. Hence there was some evidence (Table 5.4) to support early introduction of ORT in those requiring initial IV rehydration. The GDG agreed that if a child is able to tolerate orally, IVT should be stopped as soon as possible and further rehydration completed with ORT alone. This was desirable in that it would reduce the risk of prolonged IV fluid administration. The GDG also anticipated that early introduction of ORT could reduce the need for hospital admission and facilitate early hospital discharge.

Introduction

Following rehydration, some children may be at risk of recurrence of dehydration. In those cases, it might be possible to prevent this by giving supplementary fluids. The GDG wished to consider the circumstances in which this should be advised and an appropriate strategy for effective supplementation.

Evidence overview

A detailed literature search failed to identify any relevant good-quality studies to answer these questions.

GDG translation from evidence to recommendation

In the absence of clinical studies on the recurrence of dehydration, the GDG based its recommendations on consensus. Once a child is rehydrated, whether by ORT or IVT, it is important to ensure that they receive adequate fluids for normal maintenance and if necessary to compensate for significant ongoing fluid loss from diarrhoea. The frequency with which dehydration recurs following successful rehydration is not clear, but recurrence of dehydration certainly does happen. Intuitively, it seemed probable that some children would be at increased risk of recurrence, for example very young infants, those with continuing severe diarrhoea and those with persistent vomiting in whom oral fluids might not be tolerated.

There were therefore many variables to consider. The GDG considered that each individual child should be assessed for risk and managed appropriately. No prescriptive recommendation could be made regarding the exact volume or frequency of supplementary oral fluid that might be required. However, the GDG agreed that in general the aim should be to offer the infant or child plenty of their normal fluids. Supplementary ORS solution could be used if practical and appropriate – particularly for those at increased risk of dehydration (Section 4.1). The WHO previously advised giving 10 ml/kg of ORS solution after each diarrhoeal stool and more recently giving 50–100 ml if younger than 2 years and 100–200 ml if above 2 years.²⁰ The post-rehydration diarrhoeal losses in the WHO recommendations’ intended population (including children with cholera) are probably greater than is often the case in children in the UK. The GDG considered that specific fluid supplementation regimens are not necessary for most children. However, they should be considered for children at increased risk of dehydration, and a pragmatic approach is to give an additional 5 ml/kg of ORS solution for each large watery stool passed. It is important to bear in mind that if, for some reason, a child remains on IVT following rehydration they too might be at risk of dehydration recurrence, and they too might require fluid supplementation.

The GDG also agreed that if dehydration recurs, fluid management should be started again with the ORT.