Balanced crystalloid solutions versus 0.9% saline for treating acute diarrhoea and severe dehydration in children

Abstract

Background

Although acute diarrhoea is a self‐limiting disease, dehydration may occur in some children. Dehydration is the consequence of an increased loss of water and electrolytes (sodium, chloride, potassium, and bicarbonate) in liquid stools. When these losses are high and not replaced adequately, severe dehydration appears. Severe dehydration is corrected with intravenous solutions. The most frequently used solution for this purpose is 0.9% saline. Balanced solutions (e.g. Ringer's lactate) are alternatives to 0.9% saline and have been associated with fewer days of hospitalization and better biochemical outcomes. Available guidelines provide conflicting recommendations. It is unclear whether 0.9% saline or balanced intravenous fluids are most effective for rehydrating children with severe dehydration due to diarrhoea.

Objectives

To evaluate the benefits and harms of balanced solutions for the rapid rehydration of children with severe dehydration due to acute diarrhoea, in terms of time in hospital and mortality compared to 0.9% saline.

Search methods

We used standard, extensive Cochrane search methods. The latest search date was 4 May 2022.

Selection criteria

We included randomized controlled trials in children with severe dehydration due to acute diarrhoea comparing balanced solutions, such as Ringer's lactate or Plasma‐Lyte with 0.9% saline solution, for rapid rehydration.

Data collection and analysis

We used standard Cochrane methods. Our primary outcomes were 1. time in hospital and 2. mortality. Our secondary outcomes were 3. need for additional fluids, 4. total amount of fluids received, 5. time to resolution of metabolic acidosis, 6. change in and the final values of biochemical measures (pH, bicarbonate, sodium, chloride, potassium, and creatinine), 7. incidence of acute kidney injury, and 8. adverse events. We used GRADE to assess the certainty of the evidence.

Main results

Characteristics of the included studies

We included five studies with 465 children. Data for meta‐analysis were available from 441 children. Four studies were conducted in low‐ and middle‐income countries and one study in two high‐income countries. Four studies evaluated Ringer's lactate, and one study evaluated Plasma‐Lyte. Two studies reported the time in hospital, and only one study reported mortality as an outcome. Four studies reported final pH and five studies reported bicarbonate levels. Adverse events reported were hyponatremia and hypokalaemia in two studies each.

Risk of bias

All studies had at least one domain at high or unclear risk of bias. The risk of bias assessment informed the GRADE assessments.

Primary outcomes

Compared to 0.9% saline, the balanced solutions likely result in a slight reduction of the time in hospital (mean difference (MD) −0.35 days, 95% confidence interval (CI) −0.60 to −0.10; 2 studies; moderate‐certainty evidence). However, the evidence is very uncertain about the effect of the balanced solutions on mortality during hospitalization in severely dehydrated children (risk ratio (RR) 0.33, 95% CI 0.02 to 7.39; 1 study, 22 children; very low‐certainty evidence).

Secondary outcomes

Balanced solutions probably produce a higher increase in blood pH (MD 0.06, 95% CI 0.03 to 0.09; 4 studies, 366 children; low‐certainty evidence) and bicarbonate levels (MD 2.44 mEq/L, 95% CI 0.92 to 3.97; 443 children, four studies; low‐certainty evidence). Furthermore, balanced solutions likely reduces the risk of hypokalaemia after the intravenous correction (RR 0.54, 95% CI 0.31 to 0.96; 2 studies, 147 children; moderate‐certainty evidence).

Nonetheless, the evidence suggests that balanced solutions may result in no difference in the need for additional intravenous fluids after the initial correction; in the amount of fluids administered; or in the mean change of sodium, chloride, potassium, and creatinine levels.

Authors' conclusions

The evidence is very uncertain about the effect of balanced solutions on mortality during hospitalization in severely dehydrated children. However, balanced solutions likely result in a slight reduction of the time in the hospital compared to 0.9% saline. Also, balanced solutions likely reduce the risk of hypokalaemia after intravenous correction.

Furthermore, the evidence suggests that balanced solutions compared to 0.9% saline probably produce no changes in the need for additional intravenous fluids or in other biochemical measures such as sodium, chloride, potassium, and creatinine levels. Last, there may be no difference between balanced solutions and 0.9% saline in the incidence of hyponatraemia.

Plain language summary

Balanced crystalloid solutions versus 0.9% saline for severely dehydrated children with acute diarrhoea

What is dehydration and how it is treated?

Children with acute diarrhoea or gastroenteritis who get severely dehydrated require intravenous (into a vein) rehydration with fluids (called intravenous correction). The most common intravenous solutions used for this purpose are the so‐called crystalloids, which are solutions of mineral salts (e.g. electrolytes such as sodium, potassium, or chloride). The most commonly used crystalloid for rehydrating children is 0.9% saline solution. It is unclear whether this solution is the best intervention for these children because, when compared to other fluids in other diseases and conditions, 0.9% saline use has been associated with generating or worsening established metabolic acidosis and increasing the length of hospital stay.

What is metabolic acidosis?

Metabolic acidosis is defined as a reduction in low serum pH caused by different diseases, including dehydration. The pH is a measure of how acidic/basic solutions (such as water or the body fluids) are. Another measure of the degree of the acidosis is serum bicarbonate levels. Bicarbonate is a body buffer that helps to compensate the pH when metabolic acidosis occurs. The lower the level of serum bicarbonate, the more severe the acidosis. Metabolic acidosis is a common complication of dehydration, which may cause vomiting and hamper food intake in children recovering from dehydration, which may increase the hospital length of stay. When metabolic acidosis is severe (very low pH) and not treated, it affects metabolic body functions. Another fear of using 0.9% saline is the potential increase in the risk of hypokalaemia (low potassium levels in the blood), commonly affecting dehydrated children. Hypokalaemia may hamper children's capacity to receive oral fluids and feeding, amongst other complications, due to muscular weakness and a decrease in gastrointestinal motility.

What is not known about the treatment of severe dehydration in children with diarrhoea?

The alternatives to 0.9% saline solution are the so‐called balanced solutions, which are defined as intravenous fluids having an electrolyte composition close to that of human plasma (a component of blood). In comparison to the 0.9% saline solution, which only contains sodium and chloride, balanced solutions have a composition of electrolytes (sodium, potassium, and chloride) similar to the composition of human plasma, including additional cations (calcium, potassium, or magnesium), and anions such as lactate, acetate, or gluconate. A composition more similar to human plasma is expected to be more beneficial for rehydrating than the traditional 0.9% saline solution. This Cochrane Review aimed to determine whether rehydrating dehydrated children with balanced solutions results in better outcomes when compared to 0.9% saline.

What did we want to find out?

We wanted to know whether there were any differences between rehydrating a child with severe dehydration due to diarrhoea with 0.9% saline solution and doing so with balanced solutions.

What did we do?

We searched medical databases and identified five studies that evaluated 465 children. These studies randomly compared balanced solutions (Ringer's lactate or Plasma‐Lyte) with 0.9% saline solutions for severely dehydrated children with acute diarrhoea. Studies were conducted in India, Pakistan, the USA, and Canada.

What did we find?

In severely dehydrated children with diarrhoea, rehydration with balanced solutions likely results in a slight reduction in the time children are in hospital, while we are very uncertain about their effect on deaths during hospitalization when compared to rehydration with 0.9% saline.

Balanced solutions may produce a higher increase in blood pH and bicarbonate levels after correction, which may indicate a faster improvement of metabolic acidosis. However, balanced solutions produce no changes in the need for additional intravenous fluids after the initial correction; in the volume of fluids given; and in the average change of electrolytes and creatinine (a waste product that comes from muscles) levels.

Also, in terms of side effects, balanced solutions likely reduce the risk of hypokalaemia after intravenous correction (that is, fewer children with low values of serum potassium) after the intravenous correction, and they probably make no difference the incidence of hyponatraemia (low blood sodium levels), when compared to 0.9% saline.

Our results are mostly applicable to Ringer's lactate as most of the evidence came from studies comparing 0.9% saline to this solution. The evidence on Plasma‐Lyte (another balanced solution) is scarce and warrants more studies.

What are the limitations of the evidence?

The evidence comparing 0.9% saline solution and balanced solutions is scarce. The available studies evaluated very low numbers of children and is possible that people involved in the studies were aware of which treatment the children received, which gives us little confidence in the results.

How up to date is this evidence?

This review summarized the evidence up to 4 May 2022.

Summary of findings

Summary of findings 1. Balanced crystalloid solutions versus 0.9% saline for treating acute diarrhoea and severe dehydration in children.

Balanced crystalloid solutions versus 0.9% saline for treating acute diarrhoea and severe dehydration in children
Patient or population: children with severe dehydration due to diarrhoea Setting: emergency departments in India, Pakistan, the USA, Canada Intervention: balanced solutions (Ringer's lactate or Plasma‐Lyte) Comparison: 0.9% saline
Outcome	Anticipated absolute effects (95% CI)		Relative effect (95% CI)	№ of children (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with 0.9% saline	Risk difference with balanced solutions
Time in hospital (days)	The mean time in hospital was 0 days (reference value)	MD 0.35 days lower (0.60 lower to 0.10 lower)	—	90 (2 RCTs)	⊕⊕⊕⊖ Moderatea	—
Mortality (during hospitalization)	91 per 1000	61 fewer per 1000 (89 fewer to 581 more)	RR 0.33 (0.02 to 7.39)	22 (1 RCT)	⊕⊖⊖⊖ Very lowb	—
Need for additional fluids	545 per 1000	180 fewer per 1000 (404 fewer to 175 more)	RR 0.67 (0.26 to 1.72)	22 (1 RCT)	⊕⊕⊖⊖ Lowc	—
Total amount of fluids received	The mean total amount of fluids received was 0 mL/kg (reference value)	MD 2.61 mL/kg lower (7.36 lower to 2.13 higher)	—	138 (2 RCTs)	⊕⊕⊖⊖ Lowd	—
Final pH after correction (corrected pH)	The mean final pH after correction was 0 (reference value)	MD 0.06 higher (0.03 higher to 0.09 higher)	—	366 (4 RCTs)	⊕⊕⊖⊖ Lowe	—
Final bicarbonate level after correction	The mean final bicarbonate after correction was 0 mEq/L (reference value)	MD 2.44 mEq/L higher (0.92 higher to 3.97 higher)	—	443 (5 RCTs)	⊕⊕⊖⊖ Lowf	—
Adverse events: hyponatraemia	301 per 1000	124 more per 1000 (12 fewer to 322 more)	RR 1.41 (0.96 to 2.07)	147 (2 RCTs)	⊕⊕⊖⊖ Lowg	—
Adverse events: hypokalaemia	342 per 1000	158 fewer per 1000 (236 fewer to 14 fewer)	RR 0.54 (0.31 to 0.96)	145 (2 RCTs)	⊕⊕⊕⊖ Moderateh	—
CI: confidence interval; MD: mean difference; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate of effect. Low certainty: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate of effect. Very low certainty: we are very uncertain about the estimate of effect.

^aDowngraded one level for study limitations. Two trials, both at high risk of other bias. One trial was stopped prematurely after an interim analysis (Kartha 2017), and the second showed a significant baseline imbalance between groups that may have affected the primary outcomes (Mahajan 2012).
^bDowngraded one level for study limitations and two levels for extremely serious imprecision. Only one study was at high risk of selection bias and other bias (Mahajan 2012). Allocation concealment was not reported and there was significant imbalance between baseline groups (mean baseline pH was 0.08 lower in 0.9% saline group), which could have had an impact on the outcomes. Moreover, the study included only 22 children, and, as a result, the CIs of the effect estimate suggested both benefit and harm. Further, as there was only one single death in the study, we downgraded by two levels for extremely serious imprecision.
^cDowngraded one level for study limitations and one level for imprecision. Only one study was at high risk of selection bias and other bias (Mahajan 2012). Allocation concealment was not reported and there was significant imbalance between baseline groups (mean baseline pH was 0.08 lower in 0.9% saline group, which may have affected the outcomes). Moreover, the study included only 22 children, thus, the CIs of the effect estimate suggested both a benefit and harm.
^dDowngraded one level for study limitations and one level for imprecision. Data for the meta‐analysis were from two trials. One trial was stopped prematurely after an interim analysis (Kartha 2017), and the second was at unclear risk of bias in blinding of participants and personnel, blinding of outcome assessors, and selective reporting outcome criteria (Naseem 2020). Furthermore, the pooled estimate was based on an analysis of 138 children with CIs that suggested both benefit and harm.
^eDowngraded one level for study limitations and one level for inconsistency. Three studies, all at high risk of bias due to unclear blinding of participants and personnel and selective reporting outcome, and high risk of other bias (Kartha 2017; Mahajan 2012; Naseem 2020). There was high heterogeneity (I² = 73%).
^fDowngraded one level for study limitations and one level for inconsistency. Four studies, all at high risk of bias due to unclear blinding of participants and personnel, and selective reporting outcome, and high risk of other bias (Kartha 2017; Mahajan 2012; Naseem 2020; Rasheed 2020). There was high heterogeneity (I² = 83%).
^gDowngraded one level for study limitations and one level for imprecision. Two trials were at high risk of other bias (Allen 2016; Naseem 2020), and a third, at unclear risk of bias in the blinding criteria. Allen 2016 was stopped earlier, had baseline imbalances between groups in several key variables (mean age, vomiting episodes, diarrhoea episodes, weight, and bicarbonate levels), and was entirely funded by the manufacturer of the intervention. Naseem 2020 had unclear blinding of participants, personnel, and outcome assessors, and it was unclear whether it had a selective reporting outcome as it lacked a pre‐established protocol. Moreover, the effect estimate came from pooled data from 147 children, with a CI range that suggested both a benefit and harm.
^hDowngraded one level for imprecision: the effect estimate came from pooled data from 147 children, with a CI range that suggested both a benefit and harm.

Background

Description of the condition

The World Health Organization (WHO) defines acute diarrhoea as the passage of unusually loose or watery stools, usually at least three times in 24 hours, which lasts less than 14 days (WHO 2005). The American Academy of Pediatrics defines acute gastroenteritis as a diarrhoeal disease of rapid onset, with or without additional symptoms and signs such as nausea, vomiting, fever, or abdominal pain (American Academy of Pediatrics 1996). Both terms are commonly used to describe the gastrointestinal infection caused by specific micro‐organisms such as rotavirus, norovirus, Campylobacter, Escherichia coli, Salmonella, Shigella, and others (Guerrant 1990). We use the term 'acute diarrhoea' in this review to cover both acute diarrhoea and acute gastroenteritis.

Acute diarrhoea accounts for over 0.5 million deaths annually in children under five years old (Liu 2015). Although most cases of acute diarrhoea are self‐limiting, dehydration, the most common and dangerous complication, may occur in some children. During diarrhoea, there is an increased loss of water and electrolytes (sodium, chloride, potassium, and bicarbonate) in the liquid stools. Water and electrolytes are also lost through vomit, sweat, and urine. Dehydration occurs when these losses are not replaced adequately and a deficit of water and electrolytes develops (WHO 2005).

Serum bicarbonate is the only laboratory measurement that appears to be of some value for identifying dehydration in children with acute diarrhoea (Steiner 2004). However, it is not practical to obtain blood samples from every child with acute diarrhoea, and the most commonly used approach to assess the level of dehydration is clinical assessment. Dehydration is classified according to its degree, which reflects the magnitude of fluid loss. The gold standard for determining the degree of dehydration has been the child's weight loss (Pruvost 2013). There are three commonly used scales for the classification of dehydration in children: the Clinical Dehydration Scale (CDS) (Friedman 2004), Gorelick Scale (Gorelick 1997), and WHO scale (WHO 2005). The CDS classifies children to have 'no dehydration', 'some dehydration', or 'moderate dehydration' (Friedman 2004; Jauregui 2014). The Gorelick Scale classifies children to have 'no dehydration' or 'moderate to severe dehydration' (Gorelick 1997). The WHO's categories are: 'no dehydration', 'some dehydration', and 'severe dehydration' (WHO 2005). Although scales vary in their particular classifications, their value is to determine fluid management (Table 2).

1. Classification of the hydration status according to the most common dehydration scales.

Dehydration	WHO	CDS	Gorelick Scale	Gorelick Scale
No dehydration	No signs of dehydration (< 5% fluid deficit)  Well, alert Normal eyes Drinks normally, not thirsty Skin fold goes back quickly	Score 0 (< 3% fluid deficit)  Normal general appearance Normal eyes Moist mucous membranes Tears	Alert Normal capillary refill Tears present Moist mucous membranes	Alert Normal capillary refill Tears present Moist mucous membranes Normal eyes Normal breathing Normal pulses Instant recoil of skin Normal heart rate Normal urine output
Mild dehydration	Some dehydration (5–10% fluid deficit) Restless, irritable, sunken eyes Thirsty, drinks eagerly Skin fold goes back slowly	Score 1–4 (3–6% fluid deficit) Thirsty, restless, or lethargic but irritable when touched Slightly sunken eyes Sticky Decreased tears	Alert Normal capillary refill Tears present Moist mucous membranes
Moderate dehydration		Score 5–8 (> 6% fluid deficit)  Drowsy, limp, cold, sweaty; comatose or not Very sunken Dry Absent tears	Restless, lethargic, unconscious Prolonged capillary refill Tears absent Dry or very dry mucous membranes Note: deficit varies according to the number of the above signs present in the childa	Restless, lethargic, unconscious Prolonged capillary refill Tears absent Dry or very dry mucous membranes Sunken eyes Deep and rapid breathing Thready, weak, or impalpable pulses Recoil of skin slowly Tachycardia Reduced urine output Note: deficit varies according to the number of the above signs present in the childb
Severe dehydration	Severe dehydration (> 10% fluid deficit)  Lethargic or unconscious Sunken eyes Drinks poorly or not able to drink Skin fold goes back very slowly.	—	—	—

Source: Florez 2020a.
CDS: Clinical Dehydration Scale; WHO: World Health Organization.
^a ≥ 2 clinical signs, moderate dehydration, ≥ 5%; ≥ 3 clinical signs, severe dehydration, ≥ 10%.
^b ≥ 3 clinical signs, moderate dehydration, ≥ 5%; ≥ 7 clinical signs, severe dehydration, ≥ 10%.

Depending on which dehydration scale is used, it is estimated that children with 'no dehydration' have had no significant losses (depending on the dehydration scale used this means less than 3% or 5% of their bodyweight), children with 'some dehydration' have lost between 3% and 6% of their bodyweight, and those with 'severe dehydration' have lost more than 7% or more than 10% of their bodyweight (see Table 2 for more details on the dehydration classification according to the most commonly used dehydration scales). Since it is challenging, if not impossible, to know the precise weight of a child before the start of the diarrhoeal episode, the assessment of dehydration has traditionally relied on clinical evaluation of signs and symptoms (Falszewska 2018).

Appropriate classification of dehydration is crucial as it determines appropriate management. Children with 'no dehydration' may be managed at home, with increased fluid intake to prevent dehydration. Those with 'some dehydration' are managed by healthcare services and receive oral rehydration therapy, and children with severe dehydration require intravenous rehydration (WHO 2005). Rehydration of severely dehydrated children is achieved with rapid or slow rehydration therapy. Most of the management guidelines recommend a rapid rehydration therapy to treat children with severe dehydration due to diarrhoea (WHO 2005). In rapid therapy the fluid deficit is replaced over three to six hours, while in slow therapy, replacement occurs over more than six to eight hours (typically using one or two 10 mL/kg to 20 mL/kg boluses of 0.9% saline before the full replacement).

Although some micro‐organisms, such as rotavirus, have been associated with an increased risk of severe dehydration in high‐income countries (HIC) (Albano 2007), reports from low‐ and middle‐income countries (LMIC) have shown that micro‐organism identification does not predict the occurrence of severe dehydration (Andrews 2017). Severe dehydration may cause a significant loss of intravascular volume (hypovolaemia). When this loss is substantial, hypovolaemic shock, characterized by tachycardia and either prolonged capillary refill time or hypotension, may occur. Severe dehydration can also cause electrolyte disturbances and metabolic acidosis. Electrolyte disturbances may appear due to significant water, sodium, bicarbonate, and potassium losses. Large losses of bicarbonate in stools, hyperlactatemia caused by hypovolaemia, and long fasting periods causing hyperketonaemia explain metabolic acidosis in these children (Hirschhorn 1980; Levy 2013).

Last, deaths associated with severe dehydration may occur when a hypovolaemic shock is not appropriately managed and when it is associated with severe malnutrition and the coexistence of sepsis (Singh 2019). The mortality rate of children with severe dehydration may be as high as 13% (Akech 2018). Children with severe dehydration should receive intravenous rehydration therapy to restore volume and perfusion to vital organs, resolve metabolic acidosis, and replace the fluid deficit.

Description of the intervention

Crystalloids are aqueous solutions that contain ions at different concentrations. Crystalloids are widely recommended and used for volume replacement, fluid maintenance, and resuscitation in medicine in all care settings in adults and children (Holliday 2007; Long 2016; Myburgh 2013). Despite their ubiquitous use, the optimal concentration of electrolytes is currently unknown. Due to its low cost and widespread availability, 0.9% saline (also called sodium chloride) is the most widely used crystalloid (Santi 2015). The 0.9% saline solution contains 154 mmol/L each of sodium and chloride ions, with no additional electrolytes (Holliday 2007).

Balanced solutions (also called buffered solutions) are crystalloids that are an alternative to 0.9% saline but differ in several aspects. These solutions have lower concentrations of sodium and chloride; contain additional cations such as calcium, potassium, or magnesium; and contain anions such as lactate, acetate, or gluconate, which are metabolized to bicarbonate and may exert an additional buffering effect (Antequera Martin 2019; Santi 2015). These concentrations are more similar to the ones found in human plasma than 0.9% saline. The most common balanced solutions are Ringer's lactate (or Hartmann's solution) and Plasma‐Lyte 148 (or Plasma‐Lyte A).

How the intervention might work

There are two major approaches to the understanding of acid–base disturbances: the standard base‐excess approach (Henderson 1913; Siggaard‐Andersen 1977), and the Stewart approach (Stewart 1978). Under the standard base‐excess approach, metabolic acidosis is mainly explained by an increase in total body acids or by bicarbonate losses, while, under the Stewart approach, acidosis is a consequence of alterations in the water dissociation of body fluids (Morgan 2009). Regardless of which acid–base approach is used, 0.9% saline solution has been shown to produce metabolic acidosis both in vitro and in vivo (Chowdhury 2012; Kellum 2006). Following the standard base‐excess approach, the infusion of 0.9% saline will produce an increase in the amount of body chloride (hyperchloraemia), which will result in a normal anion gap acidosis. The Stewart approach uses the calculation of the strong ion difference (SID) (i.e. the difference between strong cations and anion concentrations). For solutions where this difference is zero, such as 0.9% saline (because [Na⁺] – [Cl^‐] = 154 mmol/L – 154 mmol/L = 0 mmol/L), the body's SID will be reduced, hence metabolic acidosis occurs (Stewart 1978).

In vitro and animal studies have shown several adverse effects of elevated extracellular chloride concentration on physiological variables. Hyperchloraemia has been associated with an increase in renal vascular resistance, and decreases in glomerular filtration rate (Quilley 1993; Wilcox 1987) and renin activity (Kotchen 1983). Hyperchloraemic acidosis has been shown to increase inflammatory molecules in rat models (Kellum 2006). Data demonstrated that washing red blood cells with 0.9% saline is associated with a near doubling of haemolysis during the first 24 hours after washing compared with Plasma‐Lyte A (Refaai 2018). One clinical study showed that an infusion of 0.9% saline induced hyperchloraemic acidosis at a higher rate than balanced solutions (Chowdhury 2012). Based on these characteristics, the use of balanced solutions is expected to produce a more physiological volume replacement and a faster improvement of metabolic acidosis than 0.9% saline in children with dehydration and hypovolaemia (Santi 2015).

Why it is important to do this review

The WHO recommends the use of Ringer's lactate to rehydrate children with severe dehydration due to acute diarrhoea, and suggests the use of 0.9% saline when Ringer's lactate is not available (WHO 2005). The American Academy of Pediatrics recommends the use of either Ringer's lactate or 0.9% saline (American Academy of Pediatrics 1996). In contrast, the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN)/European Society for Pediatric Infectious Diseases (ESPID) guidelines recommend the use of 0.9% saline, and only recommend Ringer's lactate in severe cases of shock (Guarino 2014). The UK's National Institute for Health and Care Excellence (NICE) guidelines only recommend the use of 0.9% saline (NICE 2009). Additional solutions, such as Dhaka Solution, with different electrolyte compositions have been used in different contexts, and are recommended by some guidelines as alternatives to Ringer's lactate, if this is not available, before considering 0.9% saline (WHO 2005), although there is no evidence to support their use. Table 3 displays the electrolyte composition of some solutions that have been used for intravenous rehydration of children with acute diarrhoea. Moreover, a survey amongst paediatric emergency physicians in Canada showed that 0.9% saline is the preferred solution for intravenous rehydration (Freedman 2011). Thus, there are several alternative treatments and heterogeneity in the recommendations of the most important guidelines. This heterogeneity may be caused by the time of publication of the guidelines, but mostly reflects the lack of appropriate evidence.

2. Composition of some intravenous solutions used for rehydration in children with diarrhoea.

Solution	Sodium	Potassium	Chloride	Calcium	Magnesium	Acetate	Gluconate	Lactate	Glucose
	(mEq/L)					(mmol/L)
Common/available balanced solutions
Ringer's lactate	140	4	109	2.7–5.0	0	0	0	27–28	0
Plasma‐Lyte 148/A	140	5	98	0	1.5	27	23–28	0	0
Less common/available balanced solutions
Dhaka solution	133	13	98	0	0	48	0	48	140
Ringer's lactate with dextrose 5%	140	4	109	5	0	0	0	28	278
Unbalanced solutions
Saline solution	154	0	154	0	0	0	0	0	0

Sources: Baxter Healthcare Corporation; Hughes 2021; Kliegman 2020; WHO 2013.

One recent Cochrane Review found that critically ill children and adults treated with 0.9% saline for resuscitation had higher chloride levels, lower bicarbonate levels, and lower pH levels than children treated with balanced solutions, but failed to demonstrate differences in mortality (Antequera Martin 2019). The review's conclusions are not applicable to children with severe dehydration due to acute diarrhoea because authors aimed to determine the effect of the solutions on resuscitation of critically ill people regardless of their diagnoses. Although the Cochrane Review's authors considered subgroup analyses by age or disease, they could not perform them. Additionally, the review authors did not consider studies in children with severe dehydration unless they were considered to be critically ill, which potentially resulted in missing specific evidence for children with acute diarrhoea and dehydration.

Another Cochrane Review showed that balanced solutions reduce the incidence of hyperchloraemia and metabolic acidosis in elective surgery in adults, in comparison to treatment with 0.9% saline (Bampoe 2017). Similar to the Antequera Martin 2019 review, the authors did not find any difference in mortality between the groups. In addition, Bampoe 2017 addressed a population very different to children with severe dehydration due to acute diarrhoea.

In summary, some evidence from different conditions and patients (mostly from adults) has shown that the use of 0.9% saline may be associated with worse biochemical outcomes. Nevertheless, many guidelines still recommend this solution for replacing volume in children with severe dehydration due to acute diarrhoea, and to date there is no systematic review that has summarized the evidence. This review aims to synthesize all the available evidence on the efficacy and safety of balanced solutions in comparison to 0.9% saline to rehydrate children with severe dehydration due to acute diarrhoea. This evidence will be useful to inform future guidelines and updates on the management of children with acute diarrhoea and dehydration.

Objectives

To evaluate the benefits and harms of balanced solutions for the rapid rehydration of children with severe dehydration due to acute diarrhoea, in terms of time in hospital and mortality compared to 0.9% saline.

Methods

Criteria for considering studies for this review

Types of studies

We included randomized controlled trials (RCTs), irrespective of publication status and language.

Types of participants

We included children aged one month to 18 years old, with:

• clinically defined acute infectious diarrhoea or acute gastroenteritis of less than 14 days' duration (defined as passing loose or watery stools at least three times in a 24‐hour period (WHO 2005)); and

• severe dehydration (defined by the dehydration scale recommended by the WHO (WHO 2005); or classified using the CDS or Gorelick Scales, as 'moderate to severe dehydration'), managed with a rapid intravenous rehydration therapy.

Types of interventions

Interventions

The interventions of interest were intravenous balanced crystalloid solutions, defined as any solution that contained lower sodium and chloride concentrations than 0.9% saline, maintained a difference between sodium and chloride levels of at least 20 mmol/L, and that contained bicarbonate or its precursors, such as acetate, lactate, or gluconate. Examples of balanced solutions included, but were not limited to: Ringer's lactate (also called 'Hartmann's solution') and multiple electrolytes' solution (Plasma‐Lyte 148 or Plasma‐Lyte A).

Control

The control was 0.9% saline solution (also called 0.9% sodium chloride or 0.9% NaCl). We considered trials that gave oral rehydration salt (ORS) solution only when ORS was given to complement intravenous infusion in both trial groups.

Types of outcome measures

Primary outcomes

• Time in hospital (or hospital length of stay)

• Mortality (overall during hospitalization)

Secondary outcomes

• Need for additional fluids

• Total amount of fluids received

• Time to resolution of metabolic acidosis (plasma pH greater than 7.35 or plasma bicarbonate greater than 19.9 mmol/L)

• Change in and final pH level after correction

• Change in and final bicarbonate level after correction

• Change in sodium level after correction

• Change in chloride level after correction

• Change in potassium level after correction

• Change in creatinine level after correction

• Acute kidney injury (AKI) at any time (according to authors' definition)

Adverse effects

• Number of children with one or more adverse effects. Expected adverse effects were any electrolytes disturbances as measured by study authors.

Search methods for identification of studies

We included RCTs regardless of language or publication status (published, unpublished, in press, and in progress).

Electronic searches

We searched in the following databases from their inception:

• Cochrane Central Register of Controlled Trials (CENTRAL), Issue 4, 2022;

• MEDLINE (Ovid) from 1946;

• Embase (Ovid) from 1974; and

• LILACS (Latin American and Caribbean Health Sciences Literature) from 1982.

Search strategies are provided in Appendix 1. We conducted searches on 18 June 2020, with an update on 4 May 2022.

Searching other resources

We searched ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp/en/), both accessed on 4 May 2022. On the same date, we searched for abstracts from the most relevant meetings and conferences, such as those from the American Academy of Pediatrics; World Federation Pediatric Intensive and Critical Care Societies; North American and European Society of Pediatric Gastroenterology, Hepatology and Nutrition; and the International Pediatric Association. We did not limit these searches by date of publication. We contacted authors of included trials and experts in the field to ask for any ongoing, missed, or unreported studies. Finally, we checked reference lists of all studies identified by the above methods.

Data collection and analysis

Selection of studies

Two review authors (GPG and JMS) independently screened all relevant titles and abstracts to determine their eligibility. We retrieved the full‐text references of studies that at least one review author considered eligible. Subsequently, two review authors (GPG and JMS) independently assessed the eligibility of full‐text reports of potentially eligible studies to determine their inclusion. Review authors resolved disagreements through discussion or by consulting a third review author (IDF). We contacted the authors of one study to clarify its eligibility. We use the PRISMA flow diagram to describe the selection process (Moher 2009). We performed the screening of titles and abstract and full‐text selection of studies using Covidence.

Data extraction and management

Two review authors (GPS and JMS) independently extracted prespecified characteristics of each trial using a standardized, piloted data extraction form. We resolved any disagreements by discussion or by consulting a third review author (IDF). We extracted the following information:

• characteristics of the study (first author, design, year of publication, sample size per group, and country);

• characteristics of participants (age in months, days of disease, aetiology, electrolyte concentrations, pH, and degree of dehydration);

• details of interventions (type of solution, composition, administered volume per treatment group); and

• outcome results (for dichotomous outcomes the number of events and number of children per group, and for continuous outcomes, the mean and standard deviation (SD) for each treatment group). If authors did not provide these, we proceeded as explained in the Dealing with missing data.

Assessment of risk of bias in included studies

Two review authors (GPG and JMS) independently assessed the risk of bias of the included studies using the Cochrane RoB 1 tool, which includes the following domains: sequence generation; allocation concealment; blinding (masking) of participants, personnel, and outcome assessors; incomplete outcome data; selective outcome reporting; and other sources of bias (Higgins 2008). We judged each domain as low, unclear, or high risk of bias. We did not exclude trials on the basis of risk of bias, but conducted sensitivity analyses to explore the potential effects of high risk of bias in the meta‐analyses (see Sensitivity analysis). We considered industry funding as a potential source of bias (other bias) (Lundh 2017).

Measures of treatment effect

We performed all analyses according to the standards specified in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2020). We calculated risk ratios (RR) for dichotomous outcomes and mean differences (MD) for continuous outcomes, both with their corresponding 95% confidence intervals (CIs).

Unit of analysis issues

We did not consider cross‐over or cluster‐RCTs due to the nature of the condition and the interventions under study. In cases of outcomes with repeated measurements, we planned to use a single time point (e.g. hyperkalaemia at six hours). In cases of studies with more than two groups, we considered only the data from the two groups of interest in the analyses and when more than two groups were eligible, we planned to include each balanced pair‐wise comparison separately, and divide the sample size of the 0.9% saline group amongst the comparisons. However, we did not identify any multiple‐group studies.

Dealing with missing data

We attempted to contact trial authors to request any missing outcome data. If the trial authors did not respond within four to eight weeks, we considered alternatives. First, we imputed some data. Most of our outcomes were continuous, and it is common that authors provide central tendency and dispersion measures that are different from the ones we require (i.e. mean and SD). If, for these outcomes, authors provided medians with ranges or interquartile ranges, we applied the approach of Wan and colleagues to calculate the best estimation of mean and SD (Wan 2014). In cases in which authors provided the mean and no dispersion measure, we tried to impute the SD by borrowing the SD from one or more other studies, ideally from those with a similar sample size (Furukawa 2005). In those cases where we performed a transformation or imputation, we conducted sensitivity analyses to assess the robustness of the results (see Sensitivity analysis).

Assessment of heterogeneity

We assessed heterogeneity by considering clinical and methodological characteristics of studies, visual overlap of CIs in the forest plots, and statistical tests (Chi² and I² statistics) (Higgins 2003). We considered heterogeneity meaningful when the Chi² test had a P value of 0.10. Furthermore, we interpreted the I² value following Cochrane Handbook for Systematic Reviews of Interventions guidance, as follows: 0% to 40%: might not be important; 30% to 60%: may represent moderate heterogeneity; 50% to 90%: may represent substantial heterogeneity; and 75% to 100%: considerable heterogeneity (Deeks 2020). We performed subgroup or sensitivity analyses to explore the potential reasons behind substantial or considerable heterogeneity.

Assessment of reporting biases

We minimized reporting bias by including both published and unpublished studies (see Searching other resources). We visually assessed the likelihood of reporting bias per outcome, by creating a funnel plot and assessing its symmetry when there were 10 or more studies (Egger 1997).

Data synthesis

We performed statistical analyses according to the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2020). We used a random‐effects model for statistical combination because we considered the intervention effects across studies were not identical. In this scenario, a random‐effects is preferable to a fixed‐effect model (Deeks 2020). We used Review Manager Web for data syntheses and analyses (RevMan Web 2023). When meta‐analysis was not appropriate, we provided a narrative description of the results.

Subgroup analysis and investigation of heterogeneity

We planned to perform subgroup analyses for the primary outcomes based on:

• type of balanced crystalloid solution (e.g. Ringer's lactate and Plasma‐Lyte 148/Plasma‐Lyte A versus other balanced solutions);

• severity of dehydration (studies in children with 'severe dehydration' versus studies that included children with 'moderate‐to‐severe dehydration').

Sensitivity analysis

We planned to conduct sensitivity analyses for the primary outcomes to assess the robustness of the results. First, we planned to conduct analyses excluding studies that considered mixed populations (i.e. eligible and ineligible populations). We planned to exclude studies with high or unclear risk of bias in the domains of sequence generation and allocation concealment, and re‐run the analyses; however, given the scarcity of the data, we could to perform these analyses. We chose these domains because they address the bias arising from the randomization process (selection bias), which we considered key to define a rigorous RCT. Biases in this process will lead to unbalanced distribution of participants between the groups, a core element of the trials. Last, we planned to conduct analyses that excluded studies that required estimations and data extraction from figures, and studies in which we had to perform data transformation (i.e. estimating the best mean and SD) or imputation of SD. Nonetheless, we could not perform these analyses due to the scarcity of the studies in most of the outcomes.

Summary of findings and assessment of the certainty of the evidence

We presented the certainty of the evidence using the GRADE approach (Guyatt 2011a). Two review authors (IF and JS) independently assessed the certainty of the evidence for each outcome. The assessment was performed for each outcome and started by assigning an evidence level according to the designs of the included studies. Evidence from RCTs is considered to be of high certainty. The body of evidence summarized was assessed against five criteria: study limitations (risk of bias), inconsistency, indirectness, imprecision, and publication bias (Guyatt 2011b). Since we only included RCTs, we started our assessment with high certainty, and assessed each criterion to consider whether we downgraded. We presented a summary of the evidence in the summary of findings table, which provides key information about the best estimate of the magnitude of the effect in relative terms, and absolute differences for each relevant comparison (Guyatt 2011b). In the summary of findings table, we provided information for the following outcomes: time in hospital, mortality, time to resolution of metabolic acidosis, final pH after correction, final bicarbonate after correction, and adverse effects.

Results

Description of studies

See Characteristics of included studies and Characteristics of excluded studies tables.

Results of the search

Our searches from electronic databases, trial registries, and handsearching identified 738 potentially relevant records. We screened 644 titles and abstracts after eliminating duplicates, and after removing 624 non‐relevant references, we identified 20 full‐text articles (Figure 1). We retrieved and reviewed these full texts against the inclusion criteria. We excluded 11 articles with reasons given in the Characteristics of excluded studies table, and we included five RCTs (nine articles). There are no studies awaiting classification or ongoing.

1.

Study flow diagram.

Included studies

All trials were published between 2012 and 2020. Four studies were conducted in LMICs (India and Pakistan) (Kartha 2017; Mahajan 2012; Naseem 2020; Rasheed 2020), and one recruited children from two HICs (the USA and Canada) (Allen 2016). In total, the five studies evaluated 465 children.

Studies included children aged one month to 18 years and all used the WHO definition of acute diarrhoea (WHO 2005). The mean age of the included children ranged from 15.5 months (Kartha 2017) to 65.5 months (Mahajan 2012). Three studies included children with severe dehydration using the WHO scale (Kartha 2017; Mahajan 2012; Naseem 2020), one study included children with moderate‐to‐severe dehydration using the Gorelick Scale (Allen 2016), and one study did not explicitly use a dehydration scale, but study authors created their own definition of severe dehydration (Rasheed 2020). None of the studies reported aetiology of the diarrhoea. The mean number of days with diarrhoea before the recruitment ranged from 1.5 days to 2.5 days.

Four studies evaluated Ringer's lactate (Kartha 2017; Mahajan 2012; Naseem 2020; Rasheed 2020), and one study evaluated Plasma‐Lyte (Allen 2016), both compared to 0.9% saline solution. In addition to the study fluids, two studies reported the administration of replacement fluids for ongoing gastrointestinal losses and maintenance, using oral rehydration solution or intravenous 0.45% saline in 5% dextrose (Kartha 2017; Mahajan 2012). Three studies supplemented children in both groups with zinc as per WHO guidance (Kartha 2017; Mahajan 2012; Naseem 2020). One study reported that other medications such as paracetamol, non‐steroidal anti‐inflammatory drugs, and narcotics were allowed orally or by intravenous infusion (or both) as cointerventions (Allen 2016). Only one study was a multicentric and set in eight paediatric emergency departments in North America (Allen 2016).

See Characteristics of included studies table.

Excluded studies

We excluded 12 studies (Akech 2014; Allen 2014; Freedman 2013; Golshekan 2016; Houston 2019; Jucá 2005; Levy 2013; Mahalanabis 1972; Neville 2006; Rahman 1988; Sendarrubias 2018; Shaikh 2022). The exclusions were due to different populations (Akech 2018), different interventions (Golshekan 2016; Houston 2019; Jucá 2005; Levy 2013), different comparators (Freedman 2013; Mahalanabis 1972; Rahman 1988; Sendarrubias 2018), and one study in which we detected potential data fabrication (Shaikh 2022).

See Characteristics of excluded studies table.

Studies awaiting classification

We identified no studies awaiting classification.

Ongoing studies

We identified no ongoing studies.

Risk of bias in included studies

See Characteristics of included studies table and Figure 2 for the risk of bias of the included studies.

2.

Risk of bias summary: review authors' judgements about each risk of bias criterion for each included study.

Allocation

Four studies had an adequate description of the randomization process and were considered at low risk of bias for this domain (Kartha 2017; Mahajan 2012; Naseem 2020; Rasheed 2020), and one study did not detail the randomization procedure and, therefore, was at unclear risk of bias (Allen 2016).

Three studies adequately described allocation concealment, and were at low risk of bias (Allen 2016; Kartha 2017; Naseem 2020). Two studies did not provide information on the allocation concealment procedure and were at unclear risk of bias (Mahajan 2012; Rasheed 2020).

Blinding

Three studies clearly reported the blinding of participants and personnel and were at low risk of bias (Allen 2016; Kartha 2017; Mahajan 2012). Two studies did not provide any information on blinding and were at unclear risk of bias (Naseem 2020; Rasheed 2020).

Incomplete outcome data

All included trials had appropriate follow‐up and analysed more than 90% of participants (Allen 2016; Kartha 2017; Mahajan 2012; Naseem 2020; Rasheed 2020). Thus, we considered all trials at low risk of attrition bias.

Selective reporting

Two studies were at low risk of selective reporting bias as protocol were available or planned outcomes were the same reported (or both) (Kartha 2017; Mahajan 2012). Two studies did not provide registration numbers or access to the protocol to assess the prespecified outcomes and were at unclear risk of reporting bias (Naseem 2020; Rasheed 2020). The remaining study was at high risk of bias because the study authors reported "length of stay" as secondary outcome in the registered protocol but did not report this in the final publication (Allen 2016).

Other potential sources of bias

One study was at low risk of other bias (Naseem 2020). One study was at unclear risk of other bias because it did not provide information to verify the baseline balance between groups in the participants' characteristics or provide a funding statement (Rasheed 2020). Three studies were at high risk of other bias due to different reasons. One study had substantial imbalance between intervention and control groups in several variables (mean age, vomiting episodes, diarrhoea episodes, weight, and bicarbonate levels) (Allen 2016). One study was stopped early: (quote) "after an interim analysis by the independent data monitoring committee contended that it was futile to continue the study further" (Kartha 2017). One study had unbalanced groups at baseline (i.e. baseline mean pH value was lower in the 0.9% saline group) (Mahajan 2012).

Effects of interventions

See: Table 1

Primary outcomes

1. Time in hospital

Two studies evaluating 90 children reported results on time in hospital or length of stay (Kartha 2017; Mahajan 2012). Balanced solutions likely results in a slight reduction of the time in hospital compared to 0.9% saline (MD −0.35 days, 95% CI −0.60 to −0.10; moderate‐certainty evidence; Analysis 1.1).

1.1. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 1: Time in hospital

Sensitivity and subgroup analyses

We planned to conduct sensitivity analyses excluding studies with high or unclear risk of bias in the sequence generation or allocation concealment (selection bias) domains. Only one study was classified as such (Allen 2016). Its removal from the meta‐analysis did not yield any difference in comparison to the primary analysis (MD −0.33 days, 95% CI −0.59 to −0.07). We could not perform the planned subgroup analyses based on the type of balanced solution or the severity of the dehydration due to the low number of studies.

2. Mortality (during hospitalization)

One study including 22 children reported results on mortality (Mahajan 2012). There was one death in the 0.9% saline group and none in the balanced solution group. The evidence is very uncertain about the effect of balanced solutions on mortality compared to 0.9% saline (RR 0.33, 95% CI 0.02 to 7.39; very low‐certainty evidence; Analysis 1.2).

1.2. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 2: Mortality (during hospitalization)

Sensitivity and subgroup analyses

We could not perform the planned sensitivity analysis excluding studies of high or unclear risk of selection bias. There was only one study for this analysis, and we judged it an unclear risk of selection bias. Similarly, we could not perform the planned subgroup analyses based on the type of balanced solution or the severity of the dehydration due to the low number of studies.

Secondary outcomes

3. Need for additional fluids

One study including 22 children reported results on the need for additional fluids (Mahajan 2012). Four children required additional fluids after the initial fluid therapy with balanced solutions compared to six children in the 0.9% saline group. The evidence suggests that balanced solutions probably produce no changes in the need for additional fluids compared to 0.9% saline (RR 0.67, 95% CI 0.26 to 1.72; low‐certainty evidence; Analysis 1.3).

1.3. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 3: Need for additional fluids

4. Total amount of fluids received

Two studies including 138 children reported results on the total amounts of fluids received (Kartha 2017; Naseem 2020). The evidence suggests that balanced solutions do not reduce the total amount of fluids compared to 0.9% saline (MD −2.61 mL/kg, 95% CI −7.36 to 2.13; low‐certainty evidence; Analysis 1.4).

1.4. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 4: Total amount of fluids received

5. Time to resolution of metabolic acidosis (plasma pH greater than 7.35 or plasma bicarbonate greater than 19.9 mEq/L)

No studies reported the time for metabolic acidosis resolution.

6. Change in and final pH level after correction

Four studies including 366 children reported results as mean final pH after correction (Kartha 2017; Mahajan 2012; Naseem 2020; Rasheed 2020). Meta‐analysis showed that balanced solutions may result in a higher pH after correction compared to 0.9% saline (MD 0.06, 95% CI 0.03 to 0.09; low‐certainty evidence; Analysis 1.5).

1.5. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 5: Final mean pH after correction

One study evaluating 70 children reported the change in the pH level (difference between mean final pH and mean pH at baseline per group) after correction (Naseem 2020). The study found that balanced solutions may result in a larger increase in mean pH after correction compared to 0.9% saline (MD 0.05, 95% CI 0.02 to 0.08).

7. Change in and final bicarbonate level after correction

Five studies including 443 children reported results as mean bicarbonate value after correction (Allen 2016; Kartha 2017; Mahajan 2012; Naseem 2020; Rasheed 2020). Meta‐analysis showed that balanced solutions may result in a larger bicarbonate level compared to 0.9% saline (MD 2.44 mEq/L, 95% CI 0.92 to 3.97; low‐certainty evidence; Analysis 1.6).

1.6. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 6: Final bicarbonate level after correction

One study including 70 children reported results as the change in bicarbonate level (mean final bicarbonate minus mean bicarbonate at baseline per group) after correction (Naseem 2020). The study found that balanced solutions may result in a larger increase of mean bicarbonate level after correction compared to 0.9% saline (MD 2.22 mEq/L, 95% CI 1.09 to 3.35).

8. Change in sodium level after correction

One study including 70 children reported results as change in sodium level (difference between mean final sodium level after correction and mean sodium level at baseline, per group) (Naseem 2020). The study found that balanced solutions may result in no difference in mean change in sodium level after correction compared to 0.9% saline (MD −0.70 mEq/L, 95% CI −2.90 to 1.50; Analysis 1.7).

1.7. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 7: Change in sodium level after correction

9. Change in chloride level after correction

No studies reported change in chloride level after correction (difference between mean final chloride level after correction and mean chloride level at baseline, per group)

10. Change in potassium level after correction

One study including 70 children reported results as change in potassium level (difference between mean final potassium level after correction and mean potassium level at baseline, per group) (Naseem 2020). The study found that balanced solutions may result in no difference in mean change in the potassium level after correction compared to 0.9% saline (MD 0 mEq/L, 95% CI −0.21 to 0.21; Analysis 1.8).

1.8. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 8: Change in potassium level after correction

11. Change in creatinine level after correction

One study including 70 children reported results as change in creatinine level (difference between mean final creatinine level after correction and mean creatinine level at baseline, per group) (Naseem 2020). The study found that balanced solutions may result in no difference in mean change in the creatinine level after correction compared to 0.9% saline (MD 0.10 mg/dL, 95% CI −0.09 to 0.29; Analysis 1.9).

1.9. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 9: Change in creatinine level after correction

12. Acute kidney injury at any time

One study including 68 children reported results on the incidence of AKI (Kartha 2017). The study reported that seven children in the balanced solutions group had AKI in comparison to six children in the 0.9% saline group. The evidence suggests that balanced solutions may result in no difference in AKI compared to 0.9% saline (RR 1.17, 95% CI 0.44 to 3.11; Analysis 1.10).

1.10. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 10: Acute kidney injury at any time

13. Adverse effects

13.1 Hyponatraemia

Two studies including 147 children reported the number of cases of hyponatraemia after correction (Allen 2016; Naseem 2020). Meta‐analysis showed that balanced solutions may result in no difference in hyponatraemia compared to 0.9% saline (RR 1.41, 95% CI 0.96 to 2.07; low‐certainty evidence; Analysis 1.11).

1.11. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 11: Adverse events (hyponatraemia)

13.2 Hypokalaemia

Two studies including 147 children reported the number of cases of hypokalaemia after correction (Allen 2016; Naseem 2020). Meta‐analysis showed that balanced solutions likely reduce the risk of hypokalaemia in comparison to 0.9% saline (RR 0.54, 95% CI 0.31 to 0.96; moderate‐certainty evidence; Analysis 1.12).

1.12. Analysis.

Comparison 1: Balanced crystalloid solutions versus 0.9% saline, Outcome 12: Adverse events (hypokalaemia)

Discussion

Summary of main results

In this systematic review, we included five studies that evaluated 465 children. The analyses from the primary outcomes showed that, in children with severe dehydration due to acute diarrhoea, balanced solutions likely results in a slight reduction of the time in the hospital compared to 0.9% saline, but we are uncertain whether there is an effect on mortality. The analyses from the secondary outcomes showed that balanced solutions may produce a higher increase in blood pH and bicarbonate levels after the intravenous correction.

However, the evidence suggests that balanced solutions result in no difference in the need for additional fluids; the total amount of fluids received; or in the change in sodium, potassium, chloride, and creatinine levels.

Last, in terms of adverse effects, in comparison to 0.9% saline, balanced solutions likely reduce the risk of hypokalaemia but may result in no difference in the incidence of hyponatraemia after intravenous correction.

Overall completeness and applicability of evidence

The included studies were conducted primarily in LMIC. Three studies recruited children in India, one in Pakistan, and one in North America (the USA and Canada). Three studies included children with severe dehydration, and one included children with moderate‐to‐severe dehydration (one study did not explicitly use a dehydration scale). Therefore, these findings may be more applicable to LMIC than to HIC contexts. Furthermore, severe dehydration cases due to acute diarrhoea are much more common in LMIC, and these findings will definitely be of interest to clinicians and decision‐makers in these settings.

We have no information about the aetiology of the diarrhoeal episodes of the children included in the studies except for very few isolated cases reported in one study. Therefore, we cannot link these results to dehydration cases caused by specific micro‐organisms. Additionally, all studies excluded children with severe malnutrition; thus, this evidence may not apply to this population, which has had different recommendations for intravenous rehydration regimens (i.e. intravenous rehydration only in cases of hypovolaemic shock) (Ashworth 2004).

We evaluated 0.9% saline and balanced crystalloid solutions, as they are the most common solutions in clinical practice in children with severe dehydration. Nevertheless, some solutions may have been studied in this clinical scenario and did not meet our criteria to be considered in the 0.9% saline or the balanced solution groups. As a result, we did not consider trials using 0.45% saline or colloids, or comparing saline to dextrose in saline solutions. Likewise, out of five included trials, four evaluated Ringer's lactate solution and one evaluated Plasma‐Lyte. We found no evidence for other balanced crystalloids, such as Sterofundin/Ringerfundin, which has been used for resuscitation in paediatric sepsis (Trepatchayakorn 2021). Although balanced solutions may share the benefits as their compositions are similar, our results may only be applicable to the use of Ringer's lactate and Plasma‐Lyte.

Only one study was funded a pharmaceutical industry (Allen 2016). Specifically, this study was funded by the manufacturer of Plasma‐Lyte. Moreover, two of this study's authors were employees of the manufacturer. The other studies had no direct funding or were funded by a not‐for‐profit organization (hospital or university).

We consider our results may be useful to inform practice guidelines. As previously described, some guidelines either suggest the use of 0.9% saline in preference to Ringer's lactate (Guarino 2014; NICE 2009), or do not recommend one over the other (American Academy of Pediatrics 1996). Therefore, our results may inform revisions of the recommendations. In contrast, WHO guidelines have explicitly recommended Ringer's lactate over 0.9% saline for rehydrating severely dehydrated children with acute diarrhoea and only recommend 0.9% saline when the Ringer's lactate is unavailable (WHO 2005). Our results may be used to support the WHO recommendation on Ringer's lactate over 0.9% saline. However, additional factors, such as costs and availability of solutions need to be considered to support future recommendations

Certainty of the evidence

The certainty of the evidence ranged between very low and moderate. The results for the primary outcomes were judged as moderate certainty (time in hospital) and low certainty (mortality) (Table 1). The reasons for downgrading the certainty were mainly due to study limitations (risk of bias of individual studies) and imprecision, and in one case, due to inconsistency. The study limitations for time in hospital were due to other biases and unclear allocation concealment.

All the trials, except Naseem 2020 were judged at high or unclear risk of other bias due to early stopping of the study, manufacturer funding, or unbalanced groups at baseline. We downgraded due to imprecision in the outcomes: mortality, need for additional fluids, the total amount of fluids received, and adverse effects (hyponatraemia). We downgraded for inconsistency for the final bicarbonate after correction outcome.

Potential biases in the review process

Even though we applied a comprehensive search strategy without language restrictions, we may have missed relevant studies from other databases. Moreover, we could not graphically explore the risk of publication bias due to the low number of studies. In addition, some data were missing. For instance, in Allen 2016, we could not obtain the SD for the total amount of fluids received per group, from the published paper, and we received no response after contacting the authors by email. Therefore, we excluded this study from this outcome analysis. In addition, we excluded one study due to potential fabricated data (Shaikh 2022), because it had identical results in many outcomes to another study already included (Mahajan 2012).

One study presented outcome data as medians and ranges (Kartha 2017). We transformed these data to the best mean and SD, according to the approach by Wan 2014. This approach could have had an impact on the results. Moreover, we planned sensitivity analyses based on the outcomes that had some data transformation. However, there was not enough information for the primary outcomes.

Some outcomes from the trials did not contribute to the meta‐analyses because they were reported in a different form than that defined by our protocol. For instance, some outcomes were presented as final values, and we were interested in evaluating the change between final and baseline values (e.g. change in sodium, potassium, chloride, and creatinine). As a result, we did not use the information for these outcomes.

Agreements and disagreements with other studies or reviews

To our knowledge, this is the first systematic review that summarizes the evidence comparing 0.9% saline with balanced solutions in children with severe dehydration due to acute diarrhoea. Nonetheless, the effectiveness and safety of balanced crystalloid solutions compared to 0.9% saline have been extensively studied in different clinical scenarios, and several systematic reviews have summarized the evidence.

One Cochrane Review found that critically ill children and adults treated with balanced solutions had lower chloride levels, and higher pH and bicarbonate levels (Antequera Martin 2019). Another Cochrane Review focusing on perioperative care found that compared to 0.9% saline, balanced solutions reduced the incidence of hyperchloraemia and metabolic acidosis in elective surgery in adults (Bampoe 2017).

Non‐Cochrane reviews have also evaluated balanced solutions against 0.9% saline for critically ill people. One recent comparing balanced and saline solutions in critically ill people found that balanced solutions were similar to the saline solution for mortality, major kidney adverse events, length of intensive care unit stay, and level of creatinine (Zhu 2021). Similarly, another recent review focusing on critically ill children found that a balanced solution improved serum bicarbonate and blood pH values compared with the unbalanced fluid, but found no differences in AKI or mortality (Lehr 2022).

In summary, previous Cochrane and non‐Cochrane reviews have found some benefits in biochemical outcomes, such as improvement in pH and bicarbonate, and lower chloride levels, which are similar to our findings. However, only one review has focused on children, which found that balanced solutions improve serum bicarbonate and blood pH values compared with the unbalanced fluids, but found no differences in AKI or mortality in critically ill children (Lehr 2022). Our review found that 0.9% saline was associated with significantly more hypokalaemia, which may be relevant for clinicians who need to choose between interventions to rehydrate dehydrated children due to acute diarrhoea.

Authors' conclusions

Implications for practice.

Guidelines for the management of children with acute diarrhoea differ in their recommendations about which intravenous solution to use for rehydration. As a result, in clinical practice, both 0.9% saline and balanced solutions (mostly Ringer's lactate) are used for this purpose. We found that balanced solutions probably reduce slightly time in hospital, but the evidence was uncertain about their effect on mortality. Moreover, the evidence suggests that balanced solutions are likely superior to 0.9% saline for reducing the time in hospital and in achieving higher pH and bicarbonate levels after correction. Moreover, balanced solutions were associated with fewer episodes of hypokalaemia.

In summary, the differences between balanced crystalloid solutions, although mostly at biochemical levels, might be considered to inform clinical practice. We believe the difference between interventions for rehydrating children with severe dehydration due to acute diarrhoea favours the use of balanced solutions over 0.9% saline, and updates of guideline may need to consider preferring balanced solutions over 0.9% saline in most contexts. Nonetheless, other key factors, such as availability, feasibility, and costs, need to be considered by guidelines committees before recommending one solution over the other. A detailed analysis of benefits and harms, feasibility, and costs, needs to be established to formulate the best recommendations for each context.

Implications for research.

Several areas may need further research to identify the best intravenous solution for these children. First, we need more research focused on patient‐important outcomes. Apart from mortality, we measured the hospital stay and the need for additional fluids, which may reflect the speed of children's recovery from severe dehydration. However, other clinical outcomes may need to be measured in trials to understand the real differences between the interventions. Furthermore, although there is some work on the development of core outcomes set (COS) to be measured in trials in children with diarrhoea (Karas 2015), this work has not expanded to trials specifically focused on oral or intravenous rehydration. For instance, we found no trials in children with severe dehydration measuring quality of life of the children and their parents.

Second, we need larger and more powerful trials. Since the cases of severe dehydration due to acute diarrhoea are expected to be less frequent for several reasons, including a global increase in ORS coverage, sanitation improvement, and rotavirus immunization, amongst others (Black 2019), those trials will be more challenging to design. The most common scenario will be children with moderate or moderate‐to‐severe dehydration requiring intravenous fluids due to the impossibility of achieving oral rehydration. In this scenario, we may also need larger, high‐quality trials comparing 0.9% saline to balanced solutions.

Third, although we do not anticipate substantial differences amongst balanced solutions in the key outcomes, we may need more studies evaluating other balanced solutions. Almost all the synthesized evidence came from studies evaluating Ringer's lactate, with only one study evaluating Plasma‐Lyte. More trials evaluating Plasma‐Lyte, and other solutions, such as Sterofundin, are needed before concluding they may be useful in these children.

Fourth, further research should focus on aspects such as costs and applicability. For instance, most diarrhoeal morbidity occurs in low‐ to middle‐income settings and 0.9% saline is an inexpensive and available solution. Studies such as cost‐effective analyses in these settings of these interventions should be encouraged before considering recommending one intervention over the other.

Last, more research is needed on the use of glucose during intravenous rehydration. Some studies have suggested that adding glucose to the fluids may facilitate oral intake recovery and reduce vomiting (Rahman 1988; Reid 2009). However, the trials examining this effect have only compared 0.9% saline solution against the same solution plus dextrose at different concentrations. Some solutions, such as Dhaka solution and Ringer's lactate with dextrose, have not been appropriately studied and could be evaluated in trials. Nonetheless, hyperglycaemia and osmotic diuresis may hamper the fluid replacement and need to be considered.

History

Protocol first published: Issue 6, 2020

Appendices

Appendix 1. Detailed search strategies

Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Daily and Versions(R)

<1946 to 3 May 2022>

1. exp Diarrhea/

2. diarrh$.mp.

3. exp Gastroenteritis/

4. gastroenteritis.mp.

5. gastrointestinal infection$.mp.

6. enteritis.mp.

7. exp Rotavirus Infection/

8. rotavir$.mp.

9. dysenter$.mp.

10. or/1‐9

11. dehydration.tw. or exp DEHYDRATION/

12. rehydration.tw. or Fluid Therapy/

13. hydration.tw.

14. hypovolemia.tw. or hypovolaemia.tw. or exp Shock/ or exp HYPOVOLEMIA/

15. (intraven* or IV or parenteral).tw.

16. or/11‐15

17. 10 and 16

18. exp isotonic solutions/

19. exp hypotonic solutions/

20. exp REHYDRATION SOLUTIONS/

21. rehydration solution$.tw.

22. exp electrolytes/

23. balanced fluid$.mp.

24. unbalanced fluid$.mp.

25. buffer$ fluid$.mp.

26. exp sodium chloride/

27. (nacl adj "0.9").mp.

28. ("0.9" adj3 (saline or sodium chloride or solution or Nacl)).tw.

29. saline solution.tw.

30. (normal adj2 saline).tw.

31. (physiological adj2 (saline or solution$)).tw.

32. (isotonic adj2 (saline or solution$)).tw.

33. (Ringer$ solution or Ringer$ lactate).tw.

34. Hartman$ solution.tw.

35. hartman$.tw.

36. ringer$ acetate.mp.

37. (plasma?lyte or sterofundin or inosteril or iso?lyte).mp.

38. (poly?electrolyte or multi?electrolyte).mp.

39. (solution$ adj3 (physiological or buffer$ or non?buffer$ or balanced or unbalanced or isotonic or hypotonic or sodium chloride or NaCL or bicarbonat$ or electrolyte or acetated)).tw.

40. or/18‐39

41. 17 and 40

42. ((randomised controlled trial or controlled clinical trial).pt. or randomized.ab. or placebo.ab. or clinical trials as topic.sh. or randomly.ab. or trial.ti.) not (animals not (humans and animals)).sh.

43. (Infan$ or newborn$ or new‐born$ or perinat$ or neonat$ or baby or baby$ or babies or toddler$ or minors or minors$ or boy or boys or boyhood or girl$ or kid or kids or child or child$ or children$ or schoolchild$ or schoolchild).mp. or schoolchild.tw. or schoolchild$.tw. or adolescen$.mp. or juvenil$.mp. or youth$.mp. or teen$.mp. or under$age$.mp. or pubescen$.mp. or exp Pediatrics/ or pediatric$.mp. or paediatric$.mp. or peadiatric$.mp. or school.tw. or school$.tw. or prematur$.mp. or preterm$.mp.

44. 41 and 42 and 43

Embase

<1974 to 2022 Week 17>

1 exp Diarrhea/

2 diarrh$.mp.

3 exp Gastroenteritis/

4 gastroenteritis.mp.

5 gastrointestinal infection$.mp.

6 enteritis.mp.

7 exp Rotavirus Infection/

8 rotavir$.mp.

9 dysenter$.mp.

10 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9

11 dehydration.tw. or exp DEHYDRATION/

12 rehydration.tw. or Fluid Therapy/

13 hydration.tw.

14 (hypovolemia or hypovolaemia).tw. or exp Shock/ or exp HYPOVOLEMIA/

15 (intraven* or IV or parenteral).tw.

16 11 or 12 or 13 or 14 or 15

17 10 and 16

18 exp isotonic solutions/

19 exp hypotonic solutions/

20 exp REHYDRATION SOLUTIONS/

21 rehydration solution$.tw.

22 exp electrolytes/

23 balanced fluid$.mp.

24 unbalanced fluid$.mp.

25 buffer$ fluid$.mp.

26 exp sodium chloride/

27 (nacl adj "0.9").mp.

28 ("0.9" adj3 (saline or sodium chloride or solution or Nacl)).tw.

29 saline solution.tw.

30 (normal adj2 saline).tw.

31 (physiological adj2 (saline or solution$)).tw.

32 (isotonic adj2 (saline or solution$)).tw.

33 (Ringer$ solution or Ringer$ lactate).tw.

34 Hartman$ solution.tw.

35 hartman$.tw.

36 ringer$ acetate.mp.

37 (plasma?lyte or sterofundin or inosteril or iso?lyte).mp.

38 (poly?electrolyte or multi?electrolyte).mp.

39 (solution$ adj3 (physiological or buffer$ or non?buffer$ or balanced or unbalanced or isotonic or hypotonic or sodium chloride or NaCL or bicarbonat$ or electrolyte or acetated)).tw.

40 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39

41 17 and 40

42 (Infan$ or newborn$ or new‐born$ or perinat$ or neonat$ or baby or baby$ or babies or toddler$ or minors or minors$ or boy or boys or boyhood or girl$ or kid or kids or child or child$ or children$ or schoolchild$ or schoolchild).mp. or schoolchild.tw. or schoolchild$.tw. or adolescen$.mp. or juvenil$.mp. or youth$.mp. or teen$.mp. or under$age$.mp. or pubescen$.mp. or exp Pediatrics/ or pediatric$.mp. or paediatric$.mp. or peadiatric$.mp. or school.tw. or school$.tw. or prematur$.mp. or preterm$.mp.

43 41 and 42

44 crossover procedure/ or double blind procedure/ or single blind procedure/

45 (random* or factorial* or placebo* or assign* or allocat* or crossover*).tw.

46 randomized controlled trial.mp. or randomized controlled trial/

47 ((blind* or mask*) and (single or double or triple or treble)).tw.

48 44 or 45 or 46 or 47

49 43 and 48

CENTRAL

Issue 4, 2022

#73 MeSH descriptor: [Diarrhea] explode all trees

#74 diarrh*

#75 MeSH descriptor: [Gastroenteritis] explode all trees

#76 gastroenteritis

#77 gastrointestinal infections

#78 enteritis

#79 MeSH descriptor: [Rotavirus Infections] explode all trees

#80 rotavir*

#81 dysenter*

#82 #73 or #74 or #75 or #76 or #77 or #78 or #79 or #80 or #81

#83 dehydration

#84 rehydration

#85 MeSH descriptor: [Fluid Therapy] explode all trees

#86 hydration

#87 hypovolemia or hypovolaemia or Shock

#88 (intraven* or IV or parenteral)

#89 #83 or #84 or #85 or #86 or #87 or #88

#90 #82 and #89

#91 isotonic solution*

#92 hypotonic solution*

#93 rehydration solution*

#94 electrolytes

#95 balanced fluid*

#96 unbalanced fluid*

#97 buffered fluid*

#98 sodium chloride

#99 nacl and "0.9"

#100 "0.9" and (saline or sodium chloride or solution or Nacl)

#101 saline solution

#102 normal saline

#103 "physiological saline" or "physiological solution"

#104 "isotonic saline" or "isotonic solution"

#105 "Ringer's solution"

#106 "Ringer's lactate"

#107 "Hartman's solution"

#108 Hartman

#109 "Ringer's acetate"

#110 plasmalyte or sterofundin or inosteril or isolyte

#111 polyelectrolyte or multielectrolyte

#112 solution* and (physiological or buffer* or nonbuffer* or balanced or unbalanced or isotonic or hypotonic or sodium chloride or NaCL or bicarbonate or electrolyte or acetated)

#113 #91 or #92 or #93 or #94 or #95 or #96 or #97 or #98 or #99 or #100 or #101 or #102 or #103 or #104 or #105 or #106 or #107 or #108 or #109 or #110 or #111 or #112

#114 #90 and #113

#115 Infan$ or newborn$ or new‐born$ or perinat$ or neonat$ or baby or baby$ or babies or toddler$ or minors or minors$ or boy or boys or boyhood or girl$ or kid or kids or child or child$ or children$ or schoolchild$ or schoolchild

#116 #114 and #115

LILACS

diarrhea or gastroenteritis or rotavirus or dysentery [Words] and and isotonic or ipotonic or rehydration or fluid$ or sodium or electrolytes [Words] and randomized OR randomised OR controlled trial OR clinical trial OR random OR randomly [Words]

ClinicalTrials.gov

fluid therapy | Diarrhea | Child

WHO ICTRP

(diarrhea or gastroenteritis ) and (fluid therapy or rehydration)

Data and analyses

Comparison 1. Balanced crystalloid solutions versus 0.9% saline.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Time in hospital	2	90	Mean Difference (IV, Random, 95% CI)	‐0.35 [‐0.60, ‐0.10]
1.2 Mortality (during hospitalization)	1	22	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.02, 7.39]
1.3 Need for additional fluids	1	22	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.26, 1.72]
1.4 Total amount of fluids received	2	138	Mean Difference (IV, Random, 95% CI)	‐2.61 [‐7.36, 2.13]
1.5 Final mean pH after correction	4	366	Mean Difference (IV, Random, 95% CI)	0.06 [0.03, 0.09]
1.6 Final bicarbonate level after correction	5	443	Mean Difference (IV, Random, 95% CI)	2.44 [0.92, 3.97]
1.7 Change in sodium level after correction	1	70	Mean Difference (IV, Random, 95% CI)	‐0.70 [‐2.90, 1.50]
1.8 Change in potassium level after correction	1	70	Mean Difference (IV, Random, 95% CI)	0.00 [‐0.21, 0.21]
1.9 Change in creatinine level after correction	1	70	Mean Difference (IV, Random, 95% CI)	0.10 [‐0.09, 0.29]
1.10 Acute kidney injury at any time	1	68	Risk Ratio (IV, Fixed, 95% CI)	1.17 [0.44, 3.11]
1.11 Adverse events (hyponatraemia)	2	147	Risk Ratio (M‐H, Random, 95% CI)	1.41 [0.96, 2.07]
1.12 Adverse events (hypokalaemia)	2	147	Risk Ratio (IV, Random, 95% CI)	0.54 [0.31, 0.96]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Allen 2016.

Study characteristics
Methods	Multiple‐site RCT Setting: 8 paediatric EDs in the USA and Canada Trial registration number: NCT01234883
Participants	Number: 100 Mean age: 0.9% saline: 34.2 months; Plasma‐Lyte A: 45.9 months Inclusion criteria: children aged ≥ 6 months to < 11 years with moderate‐to‐severe dehydration due to acute gastroenteritis, defined as ≥ 3 episodes of diarrhoea or non‐bilious vomiting within the previous 24 h and a Gorelick dehydration score ≥ 4 Exclusion criteria: acute gastroenteritis that did not require intravenous therapy per clinicians' judgement, chronic health conditions such as renal failure affecting the ability to tolerate fluids or those that result in electrolyte abnormalities, or the use of prohibited medications (e.g. antacids/antidiarrhoeals within 24 hours or systemic corticosteroids within 72 h prior of presentation).
Interventions	Intervention: Plasma‐Lyte PLA (multiple electrolyte injection, Type 1) Control: 0.9% saline solution Cointerventions: during the prestudy period, all children received routine care, and oral rehydration therapy and intravenous therapy per clinician judgement plus ondansetron and analgesics
Outcomes	Primary outcome: change in venous serum bicarbonate, as measured by total carbon dioxide Secondary outcomes: assessments of the Gorelick score; BARF score for nausea/vomiting; pain (FLACC scale for ages 6 months to 3 years, Faces scale for ages 3–11 years); volume and duration of intravenous therapy; time to clinical rehydration; length of stay in the ED; and safety assessments including physical examinations, laboratory assessments, and any reported or observed adverse events
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Eligible patients were randomly assigned in a 1:1 ratio". No further information on the randomization procedure provided.
Allocation concealment (selection bias)	Low risk	Quote: "Concealed treatment allocation was via an Interactive Voice Recognition System/Interactive Web‐based System".
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Stated as "triple‐blinded" without further information. But the registration (ClinicalTrials.gov) stated: "Masking: Quadruple (Participant, Care Provider, Investigator, Outcomes Assessor)".
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Stated as "triple‐blinded" without more information. But the registration (ClinicalTrials.gov) stated: "Masking: Quadruple (Participant, Care Provider, Investigator, Outcomes Assessor)".
Incomplete outcome data (attrition bias) All outcomes	Low risk	Stated as "triple‐blinded" without more information. But the registration (ClinicalTrials.gov) stated: "Masking: Quadruple (Participant, Care Provider, Investigator, Outcomes Assessor)".
Selective reporting (reporting bias)	High risk	The trial was registered (clinicaltrials.gov/ct2/show/NCT01234883). The protocol included length of stay in the ED outcome, but it was not reported in the full paper.
Other bias	High risk	High risk due to substantial imbalance between groups in several variables (mean age, vomiting episodes, diarrhoea episodes, weight, and bicarbonate levels).

Kartha 2017.

Study characteristics
Methods	Single‐site RCT Trial registration number: Clinical Trials Registry, India (www.ctri.nic.in) (CTRI/2015/01/005481) Setting: tertiary care referral institute in South India, between April 2015 and July 2016
Participants	Number: 68 children Median age: 0.9% saline: 17 (IQR 11.3–29.3) months; balanced solution (Ringer's lactate): 14 (IQR 9.8–30.5) months Inclusion criteria: children aged 1 month to 12 years with severe dehydration owing to acute diarrhoea presenting to the paediatric ED. Acute diarrhoea defined as 3 episodes of loose, watery, or semi‐solid stools within 24 hours for 7 days. Severe dehydration assessed using WHO scale of dehydration; children were considered severely dehydrated if they had any 2 of: lethargy/unconsciousness, sunken eyeballs, drinks poorly/unable to drink, and very slow skin pinch (> 2 seconds) with or without hypotension (mean blood pressure < 5th percentile for age, sex, and height). Exclusion criteria: children with severe acute malnutrition as per WHO classification, duration of diarrhoea > 7 days or blood in stools, or had an underlying serious systemic illness; children who had received Ringer's lactate or 0.9% saline in the 24 hours preceding enrolment into study.
Interventions	Intervention: Ringer's lactate Control: 0.9% saline Cointerventions: all children received oral zinc supplements 10–20 mg/day for 14 days. Along with study fluid, 0.45% normal saline 5% dextrose (with 2 mEq/kg potassium if urine output was present) was administered as maintenance fluid and for ongoing losses (10 mL/kg for each episode of loose stool/vomitus) until the child was accepting oral rehydrating solutions well. WHO oral rehydrating solutions and oral feeds were started as soon as the child was accepting orally.
Outcomes	Primary outcome: proportion of children who showed the disappearance of signs of severe dehydration and an improvement in pH 7.35 at end of 6 hours Secondary outcomes: changes in the serum electrolytes (sodium, potassium, chloride); renal parameters (blood urea, serum creatinine); blood gas parameters (bicarbonate, base‐deficit, anion gap, lactate); volume of intravenous fluid required to correct severe dehydration; time taken to start oral feeding; adverse effects; length of hospital stay; and cost‐effectiveness analysis
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The randomization sequence was generated by the computer by variable block size design with an allocation ratio of 1:1 by a person not involved in the study".
Allocation concealment (selection bias)	Low risk	Quote: "Sequentially numbered boxes were then packed with these bottle‐sets (1 bottle set consists of 10 bottles each of 500 mL 1/4 5000 mL of the respective study fluids) according to the computer‐generated random sequence, after replacing the previous labels with new labels bearing the study name and box number".
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Sequentially numbered boxes were then packed with these bottle‐sets (1 bottle set consists of 10 bottles each of 500 mL 1/4 5000 mL of the respective study fluids) according to the computer‐generated random sequence, after replacing the previous labels with new labels bearing the study name and box number".
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The participants, the treating team of doctors, the nurses administering the study fluid, outcome assessor, and the investigators who collected the data were blinded to the treatment assignments".
Incomplete outcome data (attrition bias) All outcomes	Low risk	No losses to follow‐up.
Selective reporting (reporting bias)	Low risk	Registered protocol was accessed, and we found no evidence of selective reporting.
Other bias	High risk	Quote: "The trial had to be stopped prematurely after an interim analysis by the independent data monitoring committee contended that it was futile to continue the study further. The interim point was the completion of 1 year of study with the enrollment of a minimum of 60 participants; this was to give the study a power of 80%". Furthermore, the results for the primary outcome (time in hospital) were reported as medians and IQRs, and we estimated the best means and standard deviations for the analyses.

Mahajan 2012.

Study characteristics
Methods	Single‐site RCT Setting: paediatric emergency facilities at a tertiary‐care referral teaching hospital Trial registration number: CTRI/2009/091/001084
Participants	Number: 22 children Mean age: 0.9% saline: 58 (SD 24) months; balanced solution (Ringer's lactate): 73 (SD 28) months Inclusion criteria: children aged 1 month to 18 years, with acute diarrhoea and severe dehydration. Acute diarrhoea defined as ≥3 liquid stools in previous 24 hours. Exclusion criteria: children with persistent diarrhoea (> 14 days), clinical signs of severe malnutrition (WHO criteria), known systemic disease (cardiac, endocrine, neurological, chronic renal failure), lethal malformations, and hypoglycaemia (dextrostix value < 40 mg/dL)
Interventions	Intervention: Ringer's lactate Control: 0.9% saline solution Cointerventions: all children received replacement fluids for ongoing losses (watery stools or vomit) and maintenance fluids using either reduced‐osmolarity WHO oral rehydration solution or intravenous 0.45% saline in 5% dextrose and 2 mEq/L potassium chloride as replacement fluids depending upon child's ability to drink. Volume of replacement fluids calculated as 10 mL/kg of bodyweight per stool or vomit. The replacement fluids were charted every 2 hours after assessment of ongoing losses. Children also received age‐appropriate maintenance fluids throughout the study period. All children received oral zinc supplements 10–20 mg/day.
Outcomes	Primary outcome: change in pH from baseline Secondary outcomes: changes in serum bicarbonate levels, base deficit, and serum electrolytes (sodium, potassium, and chloride) from baseline; all‐cause mortality during hospital stay; duration of hospital stay; volume of fluids required for rehydration; and local complications at cannula site
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Random allocation sequence was computer generated (www.randomizer.org) by an independent pediatrician, not involved in patient management".
Allocation concealment (selection bias)	Unclear risk	Not reported.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The participants, treating physicians and assessors managing the patients were, thereby, blinded to the intervention".
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The participants, treating physicians and assessors managing the patients were, thereby, blinded to the intervention".
Incomplete outcome data (attrition bias) All outcomes	Low risk	No losses to follow‐up.
Selective reporting (reporting bias)	Low risk	Prospectively registered trial protocol available. Planned outcomes were the same as reported.
Other bias	High risk	Imbalanced groups: mean baseline pH was lower in 0.9% saline group (0.08 units fewer). Furthermore, the results for the primary outcome (time in hospital) were reported as medians and IQRs, and we estimated the best means and SDs for the analyses.

Naseem 2020.

Study characteristics
Methods	RCT Setting: Department of Pediatrics, Maulana Azad Medical College and associated Lok Nayak Hospital, New Delhi during the period May 2016 to April 2017. Trial registration number: no registration reported
Participants	Number: 70 children Mean age: 0.9% saline: 56.4 (SD 34.8) months; balanced solution (Ringer's lactate): 51.6 (SD 34.8) months Inclusion criteria: children aged 1–12 years with acute diarrhoea and severe dehydration were enroled after taking informed consent from their parents. Acute diarrhoea defined as ≥ 3 loose stools in previous 24 hours and duration of diarrhoea < 14 days. Severe dehydration defined as per WHO guidelines with ≥ 2 of: lethargic or unconscious, drinks poorly or unable to drink, skin pinch goes back very slowly (> 2 second) and sunken eyes. Exclusion criteria: children with dysentery, severe acute malnutrition (WHO criteria), severe anaemia (significant palmar pallor), meningitis, seizures, known surgical problems (e.g. ileostomy), known systemic disease, and hypoglycaemia (blood glucose < 54 mg/dL)
Interventions	Intervention: Ringer's lactate Control: 0.9% saline solution Cointerventions: age‐appropriate maintenance fluids and oral elemental zinc supplementation 20 mg/day
Outcomes	Sodium (mEq/L), potassium (mEq/L), blood urea (mg/dL), creatinine (mg/dL), pH, bicarbonate (mEq/L), and base‐deficit (mmol/L) levels
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Allocation sequence was computer generated (www.randomization.com)".
Allocation concealment (selection bias)	Low risk	Quote: "Allocation concealment was done through serially numbered opaque sealed envelopes (SNOSE)".
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information on blinding on participants or personnel.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information on outcome assessors.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Very few losses to follow‐up (2/70).
Selective reporting (reporting bias)	Unclear risk	No previous trial registration and no access to prespecified protocol.
Other bias	Low risk	No additional source of biases.

Rasheed 2020.

Study characteristics
Methods	Single‐site RCT Setting: Department of Pediatrics, Pak Emirates Military Hospital, Rawalpindi, Pakistan Trial registration number: not reported as registered
Participants	Number: 206 children Mean age: 3.34 (SD 1.19) years Inclusion criteria: children aged 1–5 years with watery diarrhoea and severe dehydration presenting within first 96 hours of diarrhoea. Watery diarrhoea defined as passage of ≥ 3 stools/24 hours that were softer than normal consistency as per history from the mother. Severe dehydration defined as drowsiness, sunken eye, and decreased skin turgor (pinched skin goes back > 1 second) on clinical examination. Exclusion criteria: children with bloody diarrhoea or severe malnutrition (weight for height below −3 z‐scores of the median WHO growth standards).
Interventions	Intervention: Ringer's lactate (administered volume calculated "as per operational definition"; replacement scheduled for 6 hours) Control: 0.9% saline solution (administered volume calculated "as per operational definition"; replacement scheduled for 6 hours) Cointerventions: none reported
Outcomes	Mean serum bicarbonate (mEq/L), mean blood pH, mean serum potassium (mEq/L)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Lottery based".
Allocation concealment (selection bias)	Unclear risk	No information on allocation concealment.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information on blinding.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information on blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No losses of follow‐up reported.
Selective reporting (reporting bias)	Unclear risk	No previous trial registration and no access to prespecified protocol.
Other bias	Unclear risk	There was no description of the baseline characteristics of the participants per group, thus, not possible to evaluate how balanced the groups were. No funding statement.

BARF: Baxter Animated Retching Face; ED: emergency department; FLACC: face, legs, activity, cry, consolability pain assessment scale for children; IQR: interquartile range; RCT: randomized controlled trial; WHO: World Health Organization.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Akech 2014	Ineligible patient population (children with severe malnutrition).
Freedman 2013	Incorrect comparator; authors compared 2 different types of bolus, not comparing rehydration by 1 group or the other.
Golshekan 2016	Incorrect intervention (isotonic saline vs hypotonic saline).
Houston 2019	Incorrect intervention (slow vs fast rehydration) and different population (children with severe acute malnutrition).
Jucá 2005	Incorrect intervention (intervention was a polielectrolyte solution that had a strong ion difference (sodium – chloride) = 14 mEq/L, less than established for considering a balanced solution (> 20 mEq/L)).
Levy 2013	Incorrect intervention (infusion of 5% dextrose in normal saline solution vs normal saline solution). The 5% dextrose in normal saline solution did not meet the criteria to be considered a balanced solution.
Mahalanabis 1972	Incorrect comparator (comparing Ringer's lactate vs Ringer's lactate + hypotonic electrolyte solution in 5% glucose).
Neville 2006	Incorrect intervention (hypotonic saline vs isotonic saline).
Rahman 1988	Incorrect comparator (Dhaka solution vs Dhaka solution + dextrose).
Sendarrubias 2018	Incorrect comparator (0.9% saline vs glucose 2.5% serum (SGS 2.5%) solution).
Shaikh 2022	Ineligible. We identified that some results were identical to Mahajan 2012 suggesting this was copied from this paper. Therefore, we excluded this study, and wrote to the editors of the journal to request they flag this paper with an expression of concern.

Differences between protocol and review

We did not conduct the following analyses specified in the protocol (Florez 2020b), due to a low number of included studies:

• sensitivity analyses for primary outcomes;

• subgroup analyses for primary outcomes.

Contributions of authors

Conceiving and designing the review: IDF

Writing the protocol and providing intellectual content: IDF, JMS, GPG

Screening and reviewing the references: GPG, JMS

Extracting the data and assessing the risk of bias of the included studies: JMS, GPG, IDF

Entering data into Review Manager Web and conducting the analyses: IDF

Interpreting the results: IDF, JMS, GPG

Writing the first draft of the review: IDF

Critically reviewing of the manuscript for important intellectual content; JMS, GPG

All review authors read and approved the final review version.

Sources of support

Internal sources

• Liverpool School of Tropical Medicine, UK

External sources

• Foreign, Commonwealth and Development Office (FCDO), UK

  Project number 300342‐104

Declarations of interest

IDF: none.

JMS: none.

GPG: none.

New