Intravenous rehydration of malnourished children with acute gastroenteritis and severe dehydration: A systematic review

Background: Rehydration strategies in children with severe acute malnutrition (SAM) and severe dehydration are extremely cautious. The World Health Organization (WHO) SAM guidelines advise strongly against intravenous fluids unless the child is shocked or severely dehydrated and unable to tolerate oral fluids. Otherwise, guidelines recommend oral or nasogastric rehydration using low sodium oral rehydration solutions. There is limited evidence to support these recommendations. Methods: We conducted a systematic review of randomised controlled trials (RCTs) and observational studies on 15 th June 2017 comparing different strategies of rehydration therapy in children with acute gastroenteritis and severe dehydration, specifically relating to intravenous rehydration, using standard search terms. Two authors assessed papers for inclusion. The primary endpoint was evidence of fluid overload. Results: Four studies were identified, all published in English, including 883 children, all of which were conducted in low resource settings. Two were randomised controlled trials and two observational cohort studies, one incorporated assessment of myocardial and haemodynamic function. There was no evidence of fluid overload or other fluid-related adverse events, including children managed on more liberal rehydration protocols. Mortality was high overall, and particularly in children with shock managed on WHO recommendations (day-28 mortality 82%). There was no difference in safety outcomes when different rates of intravenous rehydration were compared. Conclusions: The current ‘strong recommendations’ for conservative rehydration of children with SAM are not based on emerging evidence. We found no clinical trials providing a direct assessment of the current WHO guidelines, and those that were available suggested that these children have a high mortality and remain fluid depleted on current therapy. Recent studies have reported no evidence of fluid overload or heart failure with more liberal rehydration regimens. Clinical trials are urgently required to inform guidelines on routes and rates of intravenous rehydration therapy for dehydration in children with SAM.


Introduction
Severe acute malnutrition (SAM) is directly responsible for over 500,000 child deaths per year and plays a significant contributing factor in the deaths of many more 1 . The World Health Organization (WHO) recommends ten key steps for inpatient management of complicated SAM, and suggest that strict adherence to this protocol should reduce mortality to less than 10% 2,3 . With regard to cases complicated by diarrhoea, the 2003 WHO Guidelines for the inpatient treatment of severely malnourished children states that 'Dehydration, usually resulting from profuse watery diarrhoea (3 or more per day), is often difficult to diagnose in malnourished children because the clinical signs usually relied on to diagnose dehydration are similar to those found in severe wasting without dehydration 2 .' Nevertheless, a prospective study of unselected Kenyan children hospitalised with SAM published in 2012 showed that diarrhoea (defined as three or more watery stools) complicated 49% of cases at admissions to hospital and developed in another 16% following admission with case fatalities of 21% and 18% respectively compared with 12% in cases without diarrhoea (χ 2 = 17.6 p<0.001) 4,5 . Furthermore, whilst the guidelines indicate that signs of severe dehydration (sunken eyes and decreased skin turgor) in SAM are unreliable markers of hydration status (and challenges arise with recognition of dehydration in children with kwashiorkor) the Kilifi study showed that signs of severe dehydration were more common in those with diarrhoea (35% versus 8%) and was a risk factor for mortality (crude odds ratio 1.7; 95% confidence interval (CI) 1.1, 2.6; p = 0.012). In the multivariate model, bacteraemia (odds ratio 6.7 (95% CI 2.5-17.8 p<0.001) and hyponatraemia (odds ratio 4.9 (95% CI 2.2-11.1 p<0.001) were key risk factors for mortality 5 .
Fluids are the mainstay of treatment for dehydration; however, WHO rehydration strategies in severely malnourished children are extremely cautious. The guidelines advise oral or nasogastric rehydration using ReSoMal (rehydration solution for malnutrition, a hypo-osmolar solution with lower sodium and higher levels of potassium than standard oral rehydration solution) and strongly recommends against the use of intravenous fluids unless the child is shocked. For the shocked child (or child with severe dehydration who is unable to tolerate oral fluids), WHO guidelines recommend 15mls/Kg of intravenous fluid over 1 hour (repeated once if necessary), and a 10ml/Kg blood transfusion over 3 hours if there is no subsequent improvement (Table 1) 3,6 . These are based on concerns regarding susceptibility to fluid overload in children with malnutrition and concerns regarding the use of additional sodium. In the section of the 2003 guidelines (under Step 4: Correct Underlying Electrolyte Imbalance/), the recommendations state 'All severely malnourished children have excess body sodium even though plasma sodium may be low (giving high sodium loads will kill') 2 . There is no published physiological evidence to support this contention.
Researchers observing children with SAM have postulated that this group have compromised cardiovascular function and are therefore susceptible to fluid overload. Therefore, clinical practice has avoided aggressive fluid therapy in this group of children.
WHO guidelines referring to management of children with SAM were revisited in 2013 and remain unchanged with no mention of any of emerging new data 6 . The recommendations continue to be based on expert opinion (strong recommendations based on weak level of evidence) rather than clinical evidence. Therefore, we have conducted an independent systematic review of the current available evidence underlying use of intravenous rehydration for children with dehydration and SAM.

Objectives
To conduct a critical appraisal of available evidence on the safety of intravenous (IV) rehydration therapy for treatment of severe dehydration in children with SAM.

Methods
We did not publish a protocol prior to conducting this review. A search of online literature was performed. There were predetermined criteria, as detailed below for eligibility of studies, data outcomes, and an assessment of risk of bias and study method quality in each of the identified studies. English search terms were used.

Selection criteria
Population. Children aged 0 to 12 years with SAM who had received IV fluids for management of severe dehydration secondary to gastroenteritis or shock. We used the WHO definitions for malnutrition (weight for height (WHZ) <-3, mid-upper arm circumference (MUAC) <115mm or oedema consistent with kwashiorkor), gastroenteritis (three or more loose watery stools per day) and for severe dehydration (presence of two of the following signs: reduced skin turgor, sunken eyes, inability to drink, lethargy or reduced consciousness) 3 . We excluded studies with chronic or persistent diarrhoea lasting ≥ 14 days.

Intervention and comparison.
All studies that evaluated intravenous fluids were included with a specific focus on safety. Studies were excluded if they considered rehydration in children without malnutrition, or only considered rehydration via the oral or nasogastric route.

Outcome.
Clinical studies that reported on any outcomes were included. The primary outcomes for this review were incidence of fluid overload following fluid administration (defined as clinical or echocardiographic evidence of heart failure or pulmonary oedema or the need for diuretics), as an indicator for safety, and urine output in response to rehydration therapy as a marker for efficacy. Secondary outcomes of interest were all-cause mortality, and other assessments of IV fluid safety and frequency of other fluid related adverse events: neurological compromise (defined as any reported seizure activity, altered consciousness or unequal pupils); cardiovascular compromise (defined as any worsening of blood pressure, bradycardia or tachycardia); or/and development or worsening of hyponatraemia (since WHO recommends hypo-osmolar intravenous fluids for resuscitation).

Study design.
Randomised-controlled trials (RCTs) and observational studies were included. The authors screened the results of the literature search for studies that met the inclusion criteria as determined by the PICOS outline (see Figure 1).

Study selection
The search produced 599 studies (see Figure 1.) After screening and evaluation, four studies were identified that investigate use of intravenous fluid treatment in children with severe malnutrition complicated by dehydration, incorporating a total of 883 patients. All four of these studies were conducted in low resource settings (Kenya 7,8 , Uganda 8 and Bangladesh 9,10 ). Two of these studies were randomised controlled trials 7,10 , and two were observational studies 8,9 . Two considered different rates of rehydration 7,8 and both incorporated serial assessment of haemodynamic responses, urinary output, with one including sequential detailed echocardiographic and electrocardiographic assessment before and after receipt of fluids. The third study evaluated the use of a standardised protocol for management of dehydration 9 and the fourth trial treated all children with severe dehydration with IV fluids (assessing safety) and then randomised children to one of three oral rehydration solutions (ORS) 10 . One study included children with cholera only 10 (see Box 1).
There was some heterogeneity in the population eligibility criteria, sample size, and methods employed by each study, and in their results. Table 2 and Table 3 show the setting, methodology and features of the included studies and their results. One further study was identified on clinicaltrials.gov that assessed two rates of IV rehydration in children with SAM. This study was conducted and completed in Bangladesh. The corresponding author was contacted and indicated that the report is currently under review for publication 11 . The only indication for intravenous infusion in a child with SAM is shock OR a child with severe dehydration and who cannot be rehydrated orally or by nasogastric tube. These children should receive 15ml/Kg/hour of either ½ strength Darrow's + 5% dextrose OR Ringers lactate +5% dextrose. Children should be monitored every 5-10minutes for signs of over-hydration and congestive heart failure. If there is no improvement, a blood transfusion (10ml/Kg over at least 3hours) should be given.

Evidence for intravenous rehydration
This review includes 266 children with cholera out of a possible total of 802 (33%) from two studies that identified children with cholera (Alam et al and Ahmed et al) 9,10 . Ahmed did not perform any sub-analyses on children with cholera. Relevant findings from Alam 2009: 149 (85%) children presented with severe dehydration and required IV rehydration (mean amount of IV fluid required was 103 ml/Kg (95%CI 96-109)) No significant difference in baseline electrolyte abnormalities No children died in this study and no children developed signs of fluid overload No child developed signs of hyponatraemia (not specified) All children were clinically rehydrated within 6 hours although 31% did not pass urine within this period Rice-ORS group had significantly less stool output Implications in practice There is just one study evaluating safety of intravenous rehydration in children with SAM and cholera. This study rehydrated 149 children with a mean amount of 103ml/Kg of 'cholera saline' (sodium 133 mmol/L, potassium 13 mmol/L, chloride 98 mmol/L) and did not report any adverse outcome from this treatment i.e. no fluid overload, no significant difference in dysnatraemia, mortality or fluid related adverse effects. Guidelines for cholera in children with SAM remain extremely conservative and are at risk of undertreating children.

Risk of bias
The quality of each of the included studies was assessed for risk of bias using the Cochrane collaboration's tool in order to evaluate validity 12 . Three of the studies had a low overall risk of bias 7,8,10 , while the remaining study had a moderate risk due to limitations in the methodology 9 .

Urine output
Three studies reported on urine output in response to rehydration therapy. Alam et al. measured urine every 6 hours (collected in urine collectors) and reported that 31% of children remained oliguric at 6 hours, and 12% by 12 hours 10 . However, it should be noted that this included all children, of which 147 had received intravenous fluids for 4 to 6 hours and 26 that received only ORS. Akech et al. measured urine output hourly for the first 8 hours (using urinary catheters) and reported frequency of oliguria (<1ml/ kg/hour) at 8 hours (which also by the study's definition included resolution of shock and absence of oliguria). Oliguria was found to be common in both arms, but more in the group receiving half strength Darrow's and 5% dextrose (9/22, 41%) than in those receiving Ringers lactate (3/25, 12%), p=0.05 7 . Obonyo et al. measured urine output using urinary catheters and reported 'satisfactory urine output' (>1ml/Kg/hour) in 8/11 (73%) of children group 1 and 7/9 (78%) children in group 2 8 .

Secondary outcomes Mortality
All of the studies reported on mortality rates. One study in Asia reported no deaths in 175 severely malnourished children included in the trial 10 , whilst the other three reported mortality rates ranging from 9% to 81.8%. Obonyo et al. reported mortality at 48 hours and 28 days and found a universally high mortality between both groups: 35% and 81.8%, respectively, in Group 1 (receiving WHO management), and 44% and 55.6% respectively in Group 2 (receiving rehydration alone). Of note was that in children fulfilling the WHO shock criteria none survived 8 . Akech et al. 7 also found a high overall mortality (31/61, 51%) with no statistically significant difference between the two arms in a Phase II trial of 61 children; 39% of these deaths occurred in the first 24 hours and 52% within 48 hours. Ahmed et al. reported 30 (9%) case fatalities in the standardised protocol 9 , in which children received IV rehydration over a longer duration i.e. more than 3-6 hours, and use of IV fluids was avoided where possible (children also received a full package of care including IV antibiotics for a minimum of 48 hours, vitamin A prophylaxis, and standardised management for hypoglycaemia)  due to aspiration of feeds and two dying in the community. Group 2: 3 early deaths due to cardiovascular collapse and 1 due to respiratory arrest (severe acidosis and hypoxaemia). Two late deaths occurred at day 4 (1 respiratory arrest associated with metabolic complications and 1 in the community) All fatal events were judged as unrelated to fluid challenges or the IV fluid regime, and largely were secondary to underlying comorbidities Bradycardia at admission (10% of patients) and low systolic blood pressure and persistent weak pulse were associated with early death, the majority within 48hr of admission.

Akech et al., 2010 7
Low and 12% at 12 hours All of the children were clinical rehydrated within 6hours and 31% did not pass urine within this period however by 12 hours 88% of the children produced urine and by 24hours all had produced urine. Hypernatraemia or severe hyponatraemia and severe hypo or hyperkalaemia was not observed in any child Groups statistically differed in stool output and in ORS and water intakes.
The difference was entirely contributed by rice-ORS group children who had the greatest reduction in stool output and the ORS volume taken was less in the rice-ORS group and water intake greater. There was no statistical difference in duration of cholera. None of the children died in this study There was significant between-study heterogeneity in mortality rates. Mortality was greatest when signs of shock were observed 8 . There was no statistical difference reported when different rates of IV rehydration were adopted.

Hyponatraemia
The definitions of hyponatraemia varied across the studies, so summary data could not be generated.

Neurological complications
No studies reported any incidences of seizures. However, it is unclear if this outcome was predefined for inclusion in any of these studies.
(2017) published mean age adjusted heart rate and systolic and diastolic blood pressures of survivors and fatalities at baseline, and between 30 minutes and then 8 hour intervals up to 48 hours. They found that a low systolic blood pressure and a persistent weak pulse after IV fluid administration were associated with death in 11/20 (55%) patients. Longitudinal monitoring demonstrated decreases in heart rate in both groups over the first 48 hours with no significant differences between mortalities and survivors. Diastolic blood pressure was lower in patients in the rehydration only group who died before 4 hours, suggesting uncorrected hypovolaemia. No evidence for biventricular heart failure was found. In this study, children were monitored additionally using cardiac biomarkers (Troponin I, a highly sensitive biochemical marker of myocardial damage, and brain natriuretic peptide (BNP), a marker of volume expansion and pressure loading, which has previously been used to predict outcome in paediatric heart failure 13-15 at admission and 48 hours. Troponin I remained within the normal range in both groups. BNP was elevated (>300pg/ml) in 6/10 (60%) in group-1 and 2/9 (22%) at baseline and remained elevated in 4/7 (57%) patients in Group 1 but not in any patients in Group 2. The interpretation of this is challenging and requires further exploration.

Discussion
All studies in the review included children with SAM, diarrhoea and some level of dehydration ranging from 41% with any dehydration 9 to 100% with severe dehydration 8 . Two focused on shock (since this is the only indication for use of fluid therapy recommended by WHO), whilst two studies extended the management to children without strict WHO shock or rehydration regimes that are currently recommended by WHO in order to generate pilot data on safety and preliminary data on outcomes (efficacy) 7,8 . Critical to the evidence underpinning current treatment recommendations and contrary to the accepted opinion that the malnourished heart is susceptible to failure if intravenous fluids are administered, none of the studies demonstrated any evidence of fluid overload in their respective cohorts. These findings highlight the lack of available data to support current guidelines recommending conservative fluid therapy in rehydration of children with SAM and pave the way for further research to inform future practice. This body of evidence undermines the physiological rationale of fluid overload secondary to myocardial dysfunction as a mechanism of death during rehydration in paediatric SAM, and lends support to the notion that this group of children are likely to remain underfilled when following current guidelines. The most compelling data suggesting superiority of one strategy over another with regards to mortality come from Ahmed et al. There is no compelling evidence to support the conservative approach to rehydration across the spectrum of disease encompassed by the four studies identified in this systematic review. Fluid overload as a consequence of IV fluids in SAM complicated by dehydration has not been detected.

Application of guidelines
In practice, the WHO guidelines are complex, challenging to follow and open to wide interpretation (see Table 1) (3); the clinician must decide whether to rehydrate the child over 4 or 10 hours or any interval between the two limits, and at what point the child is eligible for intravenous rather than oral rehydration. Additionally, the feasibility of alternating F75 and ReSoMal every 30 minutes in a busy, pressurised and low resource setting is limited 6 .
For children with shock, the fluid regimen is restricted to a maximum of two slow IV infusions over a total of two hours, with advice to transfuse if no improvement. The recommendation of a transfusion to manage hypovolaemic shock in the context of severe dehydration is extremely concerning as it will not replete the volume (water and electrolytes) loss. As detailed above, the evidence for this level of caution with IV fluid administration is limited and may risk persistent hypovolaemia and lethal shock. Additionally, there are no compelling data that support the superiority of oral over IV rehydration in non-shocked children with severe dehydration.
Difficulties also arise when the child is near to the threshold for SAM and may be misclassified due to dehydration. Prospective research has found that 20% of children hospitalised with acute gastroenteritis (AGE) and severe dehydration (10% or more loss of body weight) temporarily fulfil anthropometric criteria for SAM (MUAC <11.5cm or WHZ <-3SD), but following rehydration they are reclassified as undernourished 27 . Thus, the current recommendations have much wider implications, with potentially 20% of children with severe dehydration secondary to AGE and without SAM receiving low volume low sodium rehydration, which may explain the poor outcomes that have been observed in the large case-control study Global Enteric Multicentre study (GEMS). This was conducted in Africa and Asia, and showed that patients with moderate/severe gastroenteritis are 8.5 times more likely to die than non-gastroenteritis controls 28,29 . A third of the fatalities occurred < 7 days following hospitalisation -indicating that current management strategies may not be working in practice for children with and without SAM and emphasising the need to revisit the current guidelines.

Conclusions
The WHO guidelines for rehydration in SAM are not supported by high quality evidence. There is also no evidence to support concerns regarding fluid overload in this population on the contrary, haemodynamic evaluation indicates that these children remain under-filled. In the light of recent famines across Africa and the current cholera outbreaks, we are extremely concerned that children with SAM (and potentially 20% of non-SAM children) are having intravenous rehydration fluids withheld on the grounds of strong recommendations based on low quality of evidence. New evidence on myocardial function and from clinical trials suggests they should be afforded the same standard of care as children without malnutrition. Further research should urgently evaluate safety and efficacy of a more aggressive approach to rehydration and support development of evidence based, user-friendly and directive guidelines.

Competing interests
No competing interests were disclosed. Rehydration in children with SAM is fraught, extremely difficult in field/emergency conditions and frankly unclear as to best practice. This review is sorely needed, so the conclusion that more research is required was disappointing.

Grant information
Despite a large number of studies found through an initial search -only four heterogeneous studies passed through the filters, assessing complications of the use of IV fluids in SAM children. The studies were a mixture of RCTs and observational, as well as varied in their inclusion criteria, treatment protocols and end points.
It was difficult as a reviewer to understand the descriptions of the different studies due to this variation. Overall there appeared to be no association between the rate of fluid resuscitation and signs of fluid overload, although one study (observational -Ahmed ) showed a higher mortality in the high volume et al groups. However, the pesentation of the studies in the paper makes it a little hard to compare the results effectively. In addition, most studies identified were small, often with only sub-groups eligible for inclusion, which means that extended analysis is not possible. Only 2 of the 4 studies assessed children who meet the WHO criteria for shock.
Although all were in resource-limited contexts, none were in emergency settings (Kenya, Uganda and Bangladesh). One of the studies from Bangladesh looked at children with cholera, rehydrated at a mean rate of 18-25ml/kg/hr (NB WHO recommendation is 30ml/kg over 2 hours i.e. 15ml/kg/hr approximately). My understanding is that the pattern of fluid loss in cholera is different from other forms of GI loss and mechanism of dehydration -and it would have been good to see some acknowledgement of this in the discussion section, especially since the mortality outcomes in this paper were better than the others.
Similarly -it would have been useful for the authors to discuss the possible effect of the different mechanisms by which malnutrition occurs -particularly acute on chronic scenarios.
However -this is clearly a significant clinical issue with scanty evidence and the possibility of high mortality. Perhaps the authours could have also searched the grey literature in case of reports from agencies/NGOs etc that share their experience tthrough alternative means.
Overall -the authors have given us cause for thought. Even though the sample size is not large enugh to permit extended analysis and the studies are heterogenous -they do suggest that that WHO guidelines are too narrow and cautious. There is definitely enough evidence presented here to warrant a proper study of the dehydration/shock management in SAM children -and in fact perhaps several studies given are too narrow and cautious. There is definitely enough evidence presented here to warrant a proper study of the dehydration/shock management in SAM children -and in fact perhaps several studies given the variety of mechanisms of dehydration (e.g. cholera vs other gastroenteritis) as well as degrees of shock.
I congratulate the authors on identifying an important problem with a very limited evidence base in the face of possibly misguided WHO protocols.
Are the rationale for, and objectives of, the Systematic Review clearly stated? Yes

Are the conclusions drawn adequately supported by the results presented in the review? Yes
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. The aim of this review was to investigate the available evidence for the WHO recommendations of cautious intravenous rehydration in children with severe acute malnutrition and severe dehydration. Caution is advised because of the assumptions that malnourished children have excess body sodium and reduced left ventricular mass and function. The authors found only four heterogeneous studies. There was no evidence for harm from fluid overload or heart failure.
Reading the descriptions of individual studies was difficult because they differed in design, populations, interventions and outcomes, so I reviewed the original studies. One observational study specifically addressed the question of heart failure and fluid resuscitation. The study was small (n=20) and found no association between heart failure and the rate of fluid resuscitation. The second observational (also sequential) study compared up to 30 to 100 ml/kg fluid resuscitation . There were no patients with heart failure in either group. However it is unclear how this was defined and sought. The higher volume group had higher mortality: 17 vs 9% with volume of intravenous fluid an unadjusted risk factor for death. Thus undetected heart failure causing death cannot be excluded. One randomised, controlled trial (RCT) comparing two different fluid regimens: isotonic fluid given relatively rapidly and hypotonic fluid (as per WHO recommendations) given relatively slowly. This study did not have fluid overload/cardiac failure as its primary outcome but rather shock resolution. No patients developed clinical signs of pulmonary 1 2 3 its primary outcome but rather shock resolution. No patients developed clinical signs of pulmonary oedema. As a pilot study it lacked power, and the presence of clinically-apparent pulmonary oedema is a crude sign of cardiac failure, but is clinically relevant and easily detectable. The second RCT randomised severely malnourished cholera patients to one of three oral rehydration solutions but intravenous resuscitation was given to all those severely dehydrated (n = 149) at a mean volume of 100 ml/kg. The rate of administration was not stated. The study specifically defined and sought signs of heart failure. No child developed such signs.
Houston formulated a specific question, whether intravenous fluid resuscitation in severely et al malnourished children with diarrhoeal dehydration or shock was associated with heart failure. They used a search strategy including multiple electronic databases and reviewing the reference lists in the resultant studies. However they did not hand search the key, relevant journals, searched the non-English literature or the "grey" literature e.g. conference abstracts, doctoral theses. The inclusion criteria were explicit and risk of bias was assessed. Which author(s) selected and assessed the studies was not stated, neither was how disagreement was resolved.
That said, the authors have identified an important problem lacking evidence. The little available evidence suggests that a cautious approach to fluid resuscitation in severely malnourished, dehydrated or shocked children may be misguided, denying these children urgently needed intravenous fluids. They correctly conclude that further research in this area is needed urgently. Severe acute malnutrition (SAM) when compounded by severe gastroenteritis, with or without shock, is daunting to treat. WHO in an effort to simplify treatment has an overall management plan based on several studies of haemodynamic and cardiac function in children with acute SAM. The studies were small, ranging from 21 to 63 enrollees and are now rather old (1977)(1978)(1979)(1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992)(1993)(1994). In those studies of children with severe gastroenteritis, blood volume, cardiac mass, left ventricular function, cardiac output and stroke volume were all impaired in children with SAM when compared with findings in well-nourished children. These results led WHO experts to recommend extreme caution when giving fluids to children with SAM and to avoid IV fluids if at all possible. More recent studies have reported similar findings but have related them to the body surface area or body mass and found them to be proportionately normal.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
Hypovolaemia requires fluid replacement, and if shock is present most emergency paediatricians would want to replace fluids aggressively. This is at odds with WHO advice.
The FEAST study did not enrol children with either severe malnutrition or gastroenteritus: the caution for fluids in non hypovolaemic shock expressed after that study therefore cannot be expanded to include children with SAM.
There is also the question of what type of fluid to use for IV administration. WHO recommends half strength Darrow's solution with 5% dextrose or Ringer's lactate with 5% on the grounds that children with SAM have a total body sodium greater than normal. These fluids are hypo-osmolar and would not be chosen as replacement fluids for fluid loss or shock in any other emergency situation.
In this review Houston searched for relevant articles on the use of IV fluids in children with SAM plus et al gastroenteritis, with the aim of finding if there were any complications, especially of fluid overload. The literature is sparse and they were able to identify only four articles, two RCTs and two observational trials. Even these four are heterogenous in aims, inclusion criteria, study treatment protocols, observations and end points; but they do all include some children with SAM and severe dehydration with or without shock who received IV fluid administration. I found reading about the different studies quite confusing as the studies were so different in design and objectives; I had to go to the original papers for more detail. If table 2 had more details in the column concerning the sample size of each study it would help. Most studies are too small to allow for much analysis and in several studies only a proportion of the children qualify for inclusion in this review. For instance Akech enrolled 61 children but only 26 had severe dehydration. Different fluid regimens and et al rates of fluid administration were used, and the aetiologies of diarrhoea varied among studies. Importantly, none of the children developed fluid overload.
In the study of children with cholera, 147 patients were given a mean of 103 ml of IV fluid over a 4-6 hr period (25-18ml/kg/hr, in contrast to the WHO protocol's recommendation for SAM with severe diarrhoea period (25-18ml/kg/hr, in contrast to the WHO protocol's recommendation for SAM with severe diarrhoea of 30ml/kg maximum in 2 hours. The mechanism of gastrointestinal fluid loss in cholera is different from mechanisms in other causes of gastroenteritis, and most clinicians would reduce their caution and give extra fluids to this group of patients. Among the cholera patients in this study, 147 children received IV fluids and there were no deaths. This is very different from the case fatality rates of 9-81% in the other three studies. The causes of malnutrition are multiple and complex. Malnutrition is seldom simply due to poor intake except in famine conditions; even then infections both acute (e.g. chest infection) and chronic (e.g. TB, HIV) may compound the problem. Gastroenteritis also differs in cause, acuteness of onset and duration. A difficult clinical scenario is acute on chronic diarrhoea in a severely malnourished, probably septic child.
Where does this leave us? These studies, though they only include small numbers of children, all suggest that WHO guidelines for how to give IV fluid to a severely dehydrated child with SAM are too cautious. There have been enough pilot studies to justify a large, well run study of the subject. It needs to be large enough to answer questions not only as to how much fluid should be given but also what fluid, at what rate and by which route. The sample size must allow for sub-analyses of SAM children with and without oedema, shock, degrees of anaemia and other confounders.

Is the statistical analysis and its interpretation appropriate? Yes
Are the conclusions drawn adequately supported by the results presented in the review? Yes No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. We thank the two reviewers for their supportive comments and overviews.