Ernstige dehydratie behoeft volumesuppletie met intraveneuze vochttoediening, indien orale toediening niet wordt verdragen. In het geval van hypovolemische shock wordt gestart met 10 ml/kg Ringerlactaat (RL) of NaCl 0.9% in 10 minuten tot zo nodig 40-60ml/kg tot herstel van de weefselperfusie. De huidige WHO-richtlijn adviseert ernstige dehydratie – zonder tekenen van shock – ook met snelle intraveneuze volumesuppletie (minimaal 100ml/kg in 3-6 uur) te herstellen. Er is niet goed onderbouwd of dit effectief en veilig is. De huidige WHO-richtlijn is gebaseerd op expert opinion, nadien zijn enkele studies gepubliceerd.

Introduction

Severe dehydration requires volume suppletion through intravenous fluid administration. In case of hypovolemic shock, initial treatment consists of 10 mL/kg Ringer Lactate (RL) or NaCl 0.9% in 10 minutes to 40-60mL/kg until recovery of tissue perfusion. The current WHO-guideline advises to also treat severe dehydration – without signs of shock – with rapid intravenous volume suppletion (a minimum of 100 mL/kg within 3-6 hours). Currently, it is unknown wheter this is effective and safe.

Sub question 1: Fast (≥20 ml/kg in the first hour) versus slow (<20 ml/kg in the first hour) volume suppletion

Sub question 2: volume suppletion in 24 h versus 48-72 h

None of the studies fulfilled the selection criteria for sub question 2 on volume suppletion in 24 hours versus 48-72 hours.

Sub question 1: Fast (≥20 ml/kg in the first hour) vs. slow (<20 ml/kg in the first hour) volume suppletion

Three studies (1 SR and 2 RCTs) were included in the analysis of the literature for sub question 1 (Iro, 2018; Houston, 2019 and Freedman, 2013). Results of Azarfar (2014), which were part of the SR by Iro (2018) are not included in this analysis because the intervention did not match the research question discussed here.

Sub question 2: volume suppletion in 24 h versus 48-72 h

None of the studies fulfilled the selection criteria for sub question 2 on volume suppletion in 24 h versus 48-72 hours.

Description of studies

Iro (2018) performed a systematic review and meta-analysis to compare use of different rates of intravenous fluid therapy in children with acute gastroenteritis and moderate or severe dehydration. They searched several databases for RCTs, including Ovid Medline (1946 – May 11, 2014), Embase (1974 – May 2017), Global Health (1973 – May 2017), Global Health Library, Science Citation Index Expanded, and Conference proceedings Citation Index-Science (Web of Science, 1945 – May 2017). Three RCTs were included with a total of 468 patients (Freedman, 2011; Nager, 2010; Azarfar, 2014). One trial was performed in the USA, one in Canada, and one in Iran. All trials were conducted at an emergency department. Freedman (2011) compared 60mL/kg of normal saline (NS) over 1 h (rapid, n=114) with 20 mL/kg over 1 h (slow, n=112). Nager (2010) compared administration of 50 mL/kg of NS over 1 h (n=46) with 3 h (n=46). Azarfar (2014) compared 20-30 mL/kg of a crystalloid solution (CS) over either 2 h (n=75) or 24 h (n=75). The mean age in the selected studies ranged from 1.4 to 2.7 years. Length of follow-up was 24 hours, three days or seven days. The Cochrane Collaboration’s tool for assessing risk of bias was used to assess quality of the studies. The study by Freedman (2011) was rated as having a low risk of bias, the other two studies were of lower methodological quality (moderate or high risk of bias). Results of the study by Azarfar (2014) are not further discussed in this literature summary, because volume suppletion was too slow, i.e., less than 20 ml/kg in the first hour, for the research question discussed here.

Houston (2019) performed an open-label randomized control trial in children admitted to the hospital with severe dehydration secondary to gastroenteritis in Uganda/Kenya. During the period of study, there were no reported epidemics of cholera in either location. The RCT was of high quality (low risk of bias). Patients were randomized to rapid or slow rehydration. In the rapid group (n=61), patients younger than one year old received 30 ml/kg in the first hour plus 70 ml/kg over the five hours after the first hour. Children who were one year old or older received 30 ml/kg in the first 30 minutes plus 70 ml/kg over the 2.5 hours after the first half hour. In addition, they received boluses for treatment of shock (n=2). In the slow rehydration group (n=61), patients received 100 ml/kg over 8 h regardless of age, without fluid boluses. Children with severe malnutrition were excluded. At baseline, the groups were well matched in terms of number of patients, gender, and age. The primary outcome measure was frequency of fluid-related serious adverse events (including mortality, cardiovascular collapse, suspected raised intracranial pressure, and pulmonary oedema) within 48 h. Secondary outcome measures included time to correct signs of severe dehydration, time to tolerate oral fluids/feeds and time to discharge, hypo- or hypernatremia (defined as sodium levels < 135 mmol or > 145 mmol/L) at 8 h and readmission to hospital (within 7 days post-discharge).

Freedman (2013) refers to the same prospective RCT as described by Freedman (2011) (as part of the SR by Iro, 2018). However, in Freedman (2013) different outcome measures are reported than in Freedman, 2011. The primary outcome measure in Freedman, 2013, was the development of hyponatremia at four hours. Secondary outcome measures were change in sodium relative to baseline value, the magnitude of decrease in sodium levels among those who experienced a decrease, risk of hypernatremia, correlations between urine parameters and hyponatremia, and fluid overload. Similar to Freedman (2011) children (> 90 days old) diagnosed with dehydration secondary to gastroenteritis were assessed, in which oral rehydration therapy (ORT) failed and intravenous therapy (IVT) was prescribed. Patients were randomized to receive 60 mL/kg (large) or 20 mL/kg (standard) 0.9% saline bolus followed by maintenance 0.9% saline for 3 h. Biochemical tests were performed at baseline and 4 h. No further follow-up data was recorded for these outcome measures.

Results

Mortality (critical outcome)

In only one of the five included RCTs, cases of mortality occurred (Houston, 2019). Overall mortality in this study was 4 (3.3%) out of 122 children. Houston (2019) reported that two (3.3%) out of 61 patients in the plan C arm (rapid rehydration group) had a cardiovascular collapse that resulted in death. The same number of deaths (2 (3.3%) out of 61 patients) occurred in the slow rehydration arm and was also the result of cardiovascular collapse. The number of reported deaths is too low in both arms to assess whether this difference is clinically relevant.

PICU admission (critical outcome)

None of the included RCTs reported on PICU admission. This is likely due to the fact that most included studies were executed at emergency departments (all studies except for Houston, 2019). In addition, Houston, 2019, was executed in non-Western countries (Uganda/Kenya). It is therefore questionable whether a PICU was available.

Severe adverse events (critical outcome)

Iro (2018) defined serious adverse events (SAEs) as the proportion of participants with a pre-defined serious adverse event other than death. Where this was not clear they used the definition from the International Conference on Harmonisation (ICH) Harmonised Tripartite Guideline: “any adverse event is life threatening, or requires hospitalization or prolongation of hospital stay, results in persistent or significant disability/incapacity”. Pre-specified SAEs were new onset seizures, pulmonary oedema, cerebral oedema and cardiac failure. In the two relevant RCTs (Freedman, 2011; Nager, 2010; N=318), no reports of any of these pre-specified SAEs were found.

Houston (2019) reported frequency of fluid-related serious adverse events (including mortality, cardiovascular collapse, suspected raised intracranial pressure, pulmonary

oedema) within 48 h. In the plan C (rapid) arm, 3 patients (5%) had a serious adverse event (two cardiovascular collapse (resulting in death) and one status epilepticus (resolved) within 48 h of randomization, compared with 2 (3%) (suspected pulmonary oedema and cardiovascular collapse (both resulting in death) in the slow arm (risk ratio 0.67, (95% CI 0.12, 3.85); risk difference − 1.6%, (95% CI − 8.7%, 5.4%); p = 0.65). There was one additional SAE (convulsions) in the slow arm that occurred at 101 hours from randomization, which resolved without other complications. Only one of the SAEs was judged to be probably related to the study fluid (in the rapid arm), with the other five unlikely to be related. Note that these results partly overlap with the results presented for the outcome measure “mortality”, as adverse events resulting in death are also included here.

Freedman (2013) did not report any sodium-related serious adverse events.

Taken together, the number of severe adverse events does not differ between fast and slow rehydration interventions. The total sample size of the studies is too low to draw definite conclusions of rare events such as severe adverse events.

Hospital admission (important outcome)

In the SR by Iro (2018), two studies reported on hospital admission (Freedman, 2011; Nager, 2010). Houston (2019) assessed patients who were already hospitalized. Therefore, this study is not included in the analysis of this outcome measure.

Freedman (2011) reported that 33 (29%) out of 114 children in the rapid intravenous rehydration group were admitted to hospital compared with 19 (17%) out of 112 children in the slow group (risk ratio: 1.71 (95% CI 1,03 to 2,82)). Thus, more children were admitted to hospital after rapid treatment. This difference persisted following exclusion of children admitted to hospital because of their metabolic acidosis.

Nager (2010) reported that 1 out of 46 (2%) patients in the rapid rehydration group were admitted to hospital due to treatment failure compared with 3 out of 46 (6.5%) in the standard (slow) group (risk ratio: 0.33 (95% CI 0,04 to 3,09)). Thus, two more patients were hospitalized after slow rehydration. Note that patients requiring admission to the hospital were considered a study failure and were excluded from analysis because all patients were required to be treated as outpatients based on the study protocol.

Taken together, the studies of Freedman (2011) and Nager (2010) do not show convincing evidence in favor of rapid or slow rehydration for the outcome measure hospital admission (RR 1.09, 95% CI: 0.26 to 4.58). The results of both studies show clinically relevant differences, however in opposite direction, which resulted in a very wide confidence interval. Therefore, these results are very imprecise, which is likely due to the low sample size and number of events in these studies.

Length of stay (important outcome)

In the SR by Iro (2018), only one study reported on length of stay. Freedman (2011) reported that the median time to discharge was higher in the rapid versus slow rehydration group (6.3 h versus 5.0 h, p=0.03, in both arms (n=114 versus 112)), yet this did not achieve statistical significance after correction for multiple testing. The interquartile range was not reported and therefore clinical relevance of this result cannot be assessed properly. The number of patients that had to stay at the emergency department longer than six hours was 40 (35%) out of 114 in the rapid group, and 37 (33%) out of 112 in the slow group (risk ratio 1.06 (95% CI: 0.74 to 1.53, p=0.78)). This difference is not clinically relevant and imprecise due to the wide confidence interval and low number of patients.  

Nager did not analyze the differences in length of stay between the two arms, because all patients in the rapid arm were discharged after two hours from the onset of the study intervention, and all patients in the slow arm after four hours. They reported 45 (98%) patients discharged after 2 hours in de rapid group and 43 (94%) patients discharged after 4 hours in the standard group.

Houston (2019) reported no difference between the rapid and slow arm (both n=61) in time to discharge (p=0.8). Means or standard deviations were not reported, thus clinical relevance cannot be assessed.

Overall, no convincing evidence was found that length of stay differed after rapid or slow rehydration.

Electrolyte disturbance: sodium (important outcome)

Electrolyte disturbance (sodium) was defined as a serum sodium outside the normal range of 135-145 mmol/L.

In the SR by Iro (2018), only Freedman (2011) reported on electrolyte disturbance. However, because Freedman 2013 reflects the same data and provides a more extensive report of this outcome measure, results from Freedman 2013 are reported here. Nager excluded patients prior to randomization when their sodium (Na) level was lower than 130 mmol/L or higher than 150 mmol/L and this study is therefore not included in the results.

At baseline, Houston (2019) reported median sodium levels of 139 (IQR 132 to 149) mmol/L in the rapid arm and 143 (IQR 137 to 156) mmol/L in the slow arm shifting to 142 (138, 154) mmol/L in the rapid arm and 143 (138, 156) mmol/L in the slow arm at 24 hours. On the whole sodium levels trended to achieve more normal range at 8hrs and 24hrs of presentation.

At baseline, 31 (58%) of 53 children in the rapid arm were hypo- or hypernatremic compared with 31 (62%) out of 50 children in the slow arm. Mean changes of sodium of the complete study group were reported. Hyponatremic patients showed a mean increase of 9 mmol/L in the first 8 hrs and a total increase of 12 mmol/L after 24 hrs. Hypernatremic patients showed a mean decrease of 15 mmol/L in the first 8 hrs and a total decrease of 10 mmol/L after 24hrs. Patients with normal sodium levels at baseline showed a increase of 1 mmol/l in the first 8 hrs and a total increase of 9 mmol/L at 24 hrs.

Taken together, sodium levels or number of patients with hypo- or hypernatremia did not show clinically relevant differences between the two arms, but changes in the mean sodium levels in the complete study group of the hypo- and hypernatremic subgroups were higher than clinically preferred (maximum 8-10 mmol/L per 24 hrs).

At baseline, Freedman (2013) reported sodium levels of 136 ± 4.2 mEq/L in the group that received large-volume intravenous rehydration (n=114), compared with 137 ± 3.8 mEq/L in the standard group (n=110). Four hours after treatment, the mean (± SD) change in sodium levels showed an increase of 1.6 ± 2.4 mEq/L in the large-volume group versus and increase of 0.9 ± 2.2 mEq/L in the standard group. The increase in sodium levels over four hours was larger for the large-volume group compared with the standard group and statistically significant. After four hours, 23 (21%) of 112 patients in the large-volume group were still hyponatremic (sodium level < 136 mEq/L) and 21 (20%) out of 105 patients in the standard group. Three (3%) patients were still hypernatremic (sodium level > 145 mEq/L) in the large-volume group and 1 (1%) out of 105 patients in the standard group. None of these differences were clinically relevant.

Freedman (2013) also performed a subgroup analysis in the children who were hyponatremic at baseline (n=84, the number of patients per arm was not reported). In this subgroup, the mean (± SD) change in sodium levels was 2.9 ± 1.9 mEq/L in the large-volume group, and 2.4 ± 2.2 mEq/L in the standard volume group. This group difference was not statistically significant (p= 0.47), nor clinically relevant.

Taken together, the included studies (Houston, 2019; Freedman, 2013) do not seem to show clinically relevant differences in electrolyte disturbance (sodium) after rapid or slow volume suppletion.

Level of evidence of the literature

The level of evidence regarding the outcome measure mortality could not be graded due to the low number of deaths in the included studies.

The level of evidence regarding the outcome measure PICU admission could not be graded due to the absence of PICU admissions in the included studies.

The level of evidence regarding the outcome measure severe adverse events could not be graded due to the low number of severe adverse events in the included studies.

The level of evidence regarding the outcome measure hospital admission was downgraded by 3 levels because of study limitations (risk of bias, 1 level); the low number of included patients and a very wide confidence interval including clinically relevant differences in both directions (imprecision, 2 levels).

The level of evidence regarding the outcome measure length of stay was downgraded by 4 levels because of study limitations (risk of bias, 1 level); conflicting results (inconsistency); applicability (bias due to indirectness, 1 level); the low number of included patients and wide confidence intervals (imprecision, 2 levels).

The level of evidence regarding the outcome measure electrolyte disturbance was downgraded by 2 levels because of applicability (bias due to indirectness, 1 level); the low number of included patients (imprecision).

A systematic review of the literature was performed to answer the following questions:

What are the beneficial and/or harmful effects of fast volume suppletion (≥20mL/kg in the first hour) compared with slow volume suppletion (<20 mL/kg in the first hour) in children with moderate to severe dehydration?

What are the beneficial and/or harmful effects of fast volume suppletion (in 24 hours) compared with slow volume suppletion (in 48-72 hours) in children with moderate to severe dehydration?

P: Moderately or severely dehydrated children (0-12 years).

I1: Fast volume suppletion (≥20 ml/kg in the first hour).

C1: Slow volume suppletion (<20 ml/kg in the first hour).

I2: Volume suppletion in 24 hours.

C2: Volume suppletion in 48-72 hours.

O: Mortality, PICU admission, severe adverse events (cerebral oedema, pulmonal oedema, seizures, cardiac failure), hospital admission, length of stay, electrolyte disturbance.

Relevant outcome measures

The guideline development group considered the following outcome measures as critical outcome measures for decision making: mortality, admission to the pediatric intensive care unit (PICU), and severe adverse events (e.g. cerebral oedema, pulmonal oedema, seizures, cardiac failure, and life-threatening electrolyte disbalance). Hospital admission, length of stay, and electrolyte disturbance were considered important (but not critical) outcome measures for decision making.

A priori, the working group did not further define the outcome measures listed above but used the definitions used in the studies.

For hard outcome measures of mortality (dehydration-related deaths) and serious adverse events the working group did not define thresholds for clinical decision making a priori. For all other outcome measures, the default thresholds proposed by the international GRADE working group were used: a 25% difference in risk ratio (RR) for dichotomous outcomes, and 0.5 standard deviations (SD) for continuous outcomes.

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms until 18 November 2020. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 336 unique hits. Studies were selected based on the following criteria: systematic reviews, randomized controlled trials and observational studies on the optimal amount and speed of intravenous volume suppletion in severe dehydration in children. Studies were initially selected based on title and abstract screening. After reading the full text, 1 systematic review (Iro, 2018; discussing two RCTs that matched the PICO (Nager, 2010 and Freedman, 2011) and 2 RCTs were included for sub question 1 (Houston, 2019 and Freedman, 2013; see the table with reasons for exclusion under the tab Methods), and 36 studies were excluded. None of the studies fulfilled the selection criteria for sub question 2.

Results

One SR (discussing two relevant RCTs) and two RCTs were included in the analysis of the literature. Important study characteristics and results are summarized in the evidence tables. The assessment of the risk of bias is summarized in the risk of bias tables.