Skip to main content

Foundational Trials

GASTROSAM

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Intravenous Rehydration for Severe Acute Malnutrition with Gastroenteritis
  • Acronym: GASTROSAM
  • Year: 2025
  • Journal published in: New England Journal of Medicine
  • Citation: Maitland K, Ouattara SM, Sainna H, et al. Intravenous Rehydration for Severe Acute Malnutrition with Gastroenteritis. N Engl J Med. 2025;393(13):1257-1268.

Context & Rationale

  • Background
    • Children admitted with complicated severe acute malnutrition (SAM) frequently have acute gastroenteritis, dehydration, and major baseline electrolyte derangements.
    • International guidance historically promoted cautious rehydration in SAM (predominantly oral/nasogastric strategies), with restricted use of intravenous (IV) fluids, reflecting concern about cardiac dysfunction and iatrogenic fluid overload.
    • Despite this caution, observational series from low-resource inpatient settings reported high mortality among children with SAM complicated by diarrhoea and dehydration, raising concern that undertreatment of hypovolaemia/hypoperfusion might contribute to deaths.
    • Evidence directly comparing standard-volume IV rehydration versus cautious oral/nasogastric strategies in WHO-defined severe dehydration in SAM was limited, and clinical equipoise remained.
  • Research Question/Hypothesis
    • In hospitalised children with SAM, acute gastroenteritis, and WHO-defined severe dehydration, does IV rehydration with Ringer’s lactate (100 ml/kg) reduce early mortality (by 96 hours) compared with a cautious oral/nasogastric rehydration strategy, without increasing fluid overload?
  • Why This Matters
    • Rehydration strategy is a high-frequency, high-stakes decision in paediatric wards managing SAM in low- and middle-income countries.
    • Clarifying safety and efficacy of IV rehydration in severe dehydration could shift practice away from dogma-driven restriction towards physiology- and outcomes-informed care.
    • Any demonstrated trade-offs (e.g., sodium correction versus potassium depletion) would have immediate implications for monitoring and electrolyte replacement.

Design & Methods

  • Research Question: Among children with SAM, acute gastroenteritis, and WHO-defined severe dehydration, is standard-volume IV rehydration with Ringer’s lactate superior (mortality by 96 hours) and safe (no excess fluid overload) versus a cautious oral/nasogastric rehydration strategy?
  • Study Type: Multicentre, open-label, randomised, factorial, superiority trial on paediatric hospital wards in 6 hospitals (Kenya, Uganda, Niger, Nigeria); allocation 2:1:1 (oral strategy:rapid IV:slow IV).1
  • Population:
    • Setting: inpatient paediatric wards in sub-Saharan Africa (6 hospitals across 4 countries).
    • Inclusion: age 6 months to 12 years.
    • Inclusion: severe acute malnutrition (weight-for-height z-score < −3, mid-upper-arm circumference < 11.5 cm, and/or bilateral pedal oedema).
    • Inclusion: acute gastroenteritis (≥3 loose stools per day).
    • Inclusion: severe dehydration per WHO clinical criteria (≥2 signs, including lethargy/unconsciousness, sunken eyes, inability to drink/drinking poorly, or skin pinch returning very slowly).
    • Exclusion: known congenital or rheumatic heart disease.
    • Exclusion: diarrhoea duration >14 days.
  • Intervention:
    • Intravenous rehydration with Ringer’s lactate 100 ml/kg, delivered either as “rapid” (over 3 hours if <12 months; over 6 hours if ≥12 months) or “slow” (over 8 hours).
    • Shock management within IV arms: “rapid” arm permitted Ringer’s lactate bolus 20 ml/kg; “slow” arm did not include boluses.
    • Factorial component: 1:1 assignment to ReSoMal vs standard WHO oral rehydration solution for children without SAM (administered orally/nasogastrically or after completion of IV fluids).
  • Comparison:
    • Cautious oral/nasogastric rehydration strategy: oral rehydration solution 5 ml/kg every 30 minutes for 2 hours, then 5–10 ml/kg hourly for 4–10 hours, alternating hourly with F-75 therapeutic milk.
    • Shock management in oral-strategy arm: Ringer’s lactate bolus 15 ml/kg for participants with shock at admission or shock developing during treatment.
    • Factorial component: 1:1 assignment to ReSoMal vs standard WHO oral rehydration solution for children without SAM.
  • Blinding: Open-label for bedside clinicians and participants (treatment allocation known); laboratory testing was performed with blinding to treatment assignment.
  • Statistics: A total of 272 children with severe dehydration were required to detect a 30% relative reduction in mortality at 96 hours (assumed 58% with oral strategy vs 41% with IV rehydration) with 80% power at a two-sided 5% significance level; primary analysis was intention-to-treat (site-adjusted models reported).1
  • Follow-Up Period: Intensive clinical monitoring for the first 8 hours; biochemical sampling at randomisation, 8 hours, and 24 hours; mortality assessed at 96 hours and 28 days (with additional day-7 follow-up).

Key Results

This trial was not stopped early. Enrolment reached the prespecified sample size; safety was reviewed by an independent data monitoring committee during recruitment.

Outcome Intravenous rehydration (pooled) Oral rehydration strategy Effect p value / 95% CI Notes
Death by 96 hours 9/134 (7%) 11/138 (8%) RR 1.02 95% CI 0.41 to 2.52; P=0.69 Rapid: 5/67 (7%); Slow: 4/67 (6%); site-adjusted Mantel–Haenszel RR reported.
Death by 28 days 14/134 (10%) 17/138 (12%) HR 0.85 95% CI 0.41 to 1.78; P=Not reported Rapid: 8/67 (12%); Slow: 6/67 (9%).
Suspected pulmonary oedema 0/134 (0%) 0/138 (0%) Not estimable Not reported Prespecified safety outcome.
Signs consistent with heart failure 0/134 (0%) 0/138 (0%) Not estimable Not reported Prespecified safety outcome.
Serious adverse event 24/134 (18%) 32/138 (23%) OR 0.73 95% CI 0.40 to 1.32; P=Not reported Rapid: 14/67 (21%); Slow: 10/67 (15%).
Deterioration in level of consciousness or seizures (suspected cerebral oedema) 6/134 (4%) 10/138 (7%) OR 0.54 95% CI 0.24 to 1.23; P=Not reported Clinically suspected; no neuroimaging confirmation reported.
Severe hyponatraemia at 8 hours 20/128 (16%) 58/129 (45%) OR 0.23 95% CI 0.13 to 0.41; P=Not reported Baseline severe hyponatraemia was common across groups.
Severe hypokalaemia at 8 hours 57/127 (45%) 40/128 (31%) OR 1.79 95% CI 1.07 to 3.01; P=Not reported Directionally consistent with potassium-free crystalloid exposure.
Severe hyponatraemia at 24 hours 21/127 (17%) 35/129 (27%) OR 0.53 95% CI 0.29 to 0.98; P=Not reported Sodium correction faster in IV arms.
Severe hypokalaemia at 24 hours 36/125 (29%) 26/128 (20%) OR 1.59 95% CI 0.89 to 2.83; P=Not reported Persistent potassium vulnerability in IV arms.
Change in sodium from baseline to 8 hours (mmol/L) 7.5 ± 5.0 1.9 ± 6.0 Mean difference 5.7 95% CI 4.5 to 7.0; P=Not reported Rapid: 7.6 ± 4.9; Slow: 7.5 ± 5.2.
Change in potassium from baseline to 8 hours (mmol/L) −0.1 ± 0.9 0.2 ± 0.8 Mean difference −0.3 95% CI −0.5 to −0.2; P=Not reported Rapid: −0.2 ± 0.9; Slow: −0.1 ± 0.8.
  • Mortality by 96 hours was similar between pooled IV rehydration (9/134; 7%) and the oral strategy (11/138; 8%), with a wide confidence interval around the site-adjusted RR.
  • No cases of clinically suspected pulmonary oedema or heart failure were reported in any group (0/272), directly challenging the central safety concern that has driven restrictive IV practice.
  • Electrolyte trajectories diverged: IV rehydration produced faster sodium correction but higher early severe hypokalaemia, framing electrolyte replacement/monitoring as a key safety co-intervention.

Internal Validity

  • Randomisation and allocation:
    • Randomisation used sealed, opaque envelopes with block sequences and site stratification (2:1:1 allocation), supporting allocation concealment at the point of enrolment.
  • Dropout or exclusions:
    • Vital status was available for 271/272 participants at 96 hours and 270/272 at 28 days, indicating minimal loss to follow-up for the primary outcome.
  • Performance and detection bias:
    • Open-label delivery introduces risk for co-intervention and ascertainment bias in subjective outcomes (e.g., clinical diagnosis of oedema/heart failure/cerebral oedema).
    • Primary outcome (mortality) is objective, and laboratory analyses were performed with blinding to allocation, reducing detection bias for biochemical endpoints.
  • Protocol adherence and separation of the variable of interest:
    • Receipt of IV fluids within 24 hours: oral strategy 30/135 (22%) vs pooled IV 125/126 (99%).
    • Time to initiation of IV fluids (minutes): oral strategy 123 (IQR 55–245) vs pooled IV 14 (IQR 8–26).
    • Nasogastric tube insertion: oral strategy 126/135 (93%) vs pooled IV 82/126 (65%).
    • Vomiting after nasogastric tube insertion: oral strategy 96/135 (71%) vs pooled IV 65/126 (52%).
    • The 22% crossover from oral strategy to IV fluids represents clinically appropriate rescue but likely attenuates between-group contrast for efficacy endpoints.
  • Baseline characteristics:
    • Groups were broadly comparable (median age 13–14 months; weight ~6.0 kg; kwashiorkor present in ~48–51%).
    • Profound baseline biochemical risk was common: baseline severe hyponatraemia 71/135 (53%) in oral group and 30/67 (45%) and 32/67 (48%) in rapid/slow IV groups; baseline severe hypokalaemia 59/135 (44%) in oral group and 33/67 (49%) and 27/67 (40%) in rapid/slow IV groups.
  • Heterogeneity:
    • Multicentre across 4 countries; analyses adjusted for trial site, but patient mix and care processes across hospitals can still introduce clinical heterogeneity.
  • Timing and dose:
    • IV dosing was prespecified at 100 ml/kg of Ringer’s lactate delivered over 3–6 hours (“rapid”) or 8 hours (“slow”), while the oral strategy delivered frequent ORS aliquots with F-75 alternation over 4–10 hours.
    • Bolus policies differed by arm (oral: 15 ml/kg for shock; rapid IV: 20 ml/kg for shock; slow IV: no bolus), potentially influencing early haemodynamic rescue in the sickest participants.
  • Outcome assessment:
    • Mortality outcomes are objective and hard to bias; adverse events relied on clinical diagnosis and thus remain more vulnerable to open-label ascertainment differences.
  • Statistical rigor:
    • Intention-to-treat analysis with site adjustment was reported; confidence intervals for mortality were wide, reflecting limited precision for detecting modest differences in outcomes.

Conclusion on Internal Validity: Overall, internal validity appears moderate: randomisation and follow-up were strong, outcomes were largely objective, but open-label delivery and substantial rescue IV use in the oral-strategy arm reduce separation and constrain inferential certainty for efficacy.

External Validity

  • Population representativeness:
    • Participants were typical of complicated SAM admissions in sub-Saharan Africa (median age ~13 months; high prevalence of kwashiorkor; frequent baseline hyponatraemia/hypokalaemia and acidaemia).
    • Exclusions (known congenital/rheumatic heart disease; diarrhoea >14 days) may limit applicability to children with chronic diarrhoea or coexistent structural cardiac disease.
  • Applicability:
    • The trial was delivered in hospitals with research capacity for frequent monitoring and serial biochemistry; real-world generalisability depends on whether similar monitoring and electrolyte replacement can be implemented at scale.
    • Where laboratory testing is unavailable, the observed potassium risk with IV rehydration underscores the need for pragmatic protocols for supplementation and clinical surveillance.

Conclusion on External Validity: Findings are likely generalisable to inpatient paediatric wards managing complicated SAM in similar low-resource settings, but translation to facilities without reliable monitoring/electrolyte replacement may be constrained by implementation capacity.

Strengths & Limitations

  • Strengths:
    • Directly addresses a long-standing, practice-defining safety concern with a pragmatic randomised design in high-burden settings.
    • Clinically meaningful primary endpoint (96-hour mortality) with minimal loss to follow-up.
    • Frequent early clinical monitoring and blinded laboratory endpoints enabled detailed safety phenotyping (fluid overload and electrolyte trajectories).
    • Included two IV delivery rates (rapid vs slow), supporting internal comparison of infusion speed within the broader IV strategy.
  • Limitations:
    • Open-label delivery and clinically triggered rescue IV use in the oral-strategy arm reduce treatment separation and may dilute efficacy contrasts.
    • Safety outcomes relied on clinical recognition (no routine imaging), and subtle pulmonary or cardiac overload could be under-recognised.
    • Electrolyte effects (notably hypokalaemia) highlight reliance on laboratory monitoring and supplementation infrastructure that may not be universally available.
    • Multicentre design improves generalisability, but outcomes may still be influenced by site-level variation in supportive care and staffing.

Interpretation & Why It Matters

  • Clinical implication
    In hospitalised children with SAM and WHO-defined severe dehydration, IV rehydration with Ringer’s lactate (100 ml/kg) did not demonstrate excess clinical fluid overload and produced similar short-term mortality to a cautious oral/nasogastric strategy, supporting IV rehydration as a plausible option when severe dehydration is present.
  • Safety trade-off
    IV therapy corrected hyponatraemia more quickly but increased early severe hypokalaemia, implying that “IV fluids alone” is not the intervention; safe implementation requires deliberate potassium strategy and monitoring.
  • Methodological signal
    The absence of overt fluid overload with higher-volume IV fluids challenges a central mechanistic assumption underpinning restrictive guidance and re-orients future trials towards pragmatic effectiveness endpoints and scalable co-interventions (electrolyte replacement, monitoring algorithms).

Controversies & Subsequent Evidence

  • “Data versus dogma” framing and implications for guidance:
    • An accompanying editorial highlighted that GASTROSAM provides direct experimental evidence against the long-held presumption that IV rehydration in SAM inevitably precipitates fluid overload, while emphasising the importance of system readiness (monitoring and electrolyte management) for safe adoption.2
  • Electrolyte safety debate (potassium risk with balanced crystalloids):
    • Correspondence raised concern that potassium-free balanced solutions (Ringer’s lactate) may worsen hypokalaemia in a population with profound baseline deficits, arguing that fluid selection and/or potassium supplementation strategy should be explicit when generalising results to routine wards.3
  • Generalisability and “trial care” versus routine ward care:
    • Correspondence emphasised that improved outcomes and safety may depend on broader care-system strengthening (staffing, monitoring, access to electrolytes and therapeutic feeds), and queried whether outcomes can be replicated where these elements are inconsistent.4
  • Why equipoise existed pre-trial:
    • Pre-trial systematic review evidence supporting or refuting IV strategies in malnourished children with dehydration was limited, with uncertainty regarding both efficacy and safety; this uncertainty underpinned the justification for randomisation.5
  • Positioning relative to earlier African paediatric fluid evidence:
    • A prior phase II rehydration trial in severe paediatric dehydration (non-SAM focus) informed feasibility and safety monitoring frameworks for fluid-delivery rate and physiologic response, but did not resolve the SAM-specific safety concern addressed by GASTROSAM.6
    • Broader African paediatric fluid-resuscitation evidence (e.g., FEAST) remains mechanistically relevant when extrapolating fluid bolus concepts across syndromes, but differs importantly from GASTROSAM in population (SAM with diarrhoeal dehydration) and in the intervention (rehydration rather than bolus resuscitation for presumed septic shock).7
    • Earlier SAM-focused physiologic and pilot interventional studies informed safety concerns and trial design (cardiac function, shock phenotypes), providing context for why IV restriction persisted despite limited definitive evidence.8
  • Unreported factorial component in the index report:
    • The index manuscript described factorial assignment to ReSoMal versus standard WHO oral rehydration solution for children without SAM, but comparative outcomes for this component were not reported in the primary results tables.

Summary

  • GASTROSAM randomised 272 children with SAM, acute gastroenteritis, and WHO-defined severe dehydration to a cautious oral/nasogastric strategy versus IV rehydration with Ringer’s lactate (rapid or slow 100 ml/kg).
  • There was no demonstrable difference in 96-hour mortality (7% pooled IV vs 8% oral; RR 1.02; 95% CI 0.41 to 2.52; P=0.69) or 28-day mortality (10% pooled IV vs 12% oral; HR 0.85; 95% CI 0.41 to 1.78).
  • No cases of clinically suspected pulmonary oedema or heart failure were reported in any group (0/272), challenging the central safety rationale for IV restriction in SAM.
  • IV rehydration accelerated sodium correction and reduced severe hyponatraemia, but increased early severe hypokalaemia, identifying potassium management as a core implementation requirement.
  • Open-label delivery and substantial rescue IV use in the oral strategy arm (22% received IV fluids within 24 hours) likely reduced treatment separation for efficacy endpoints.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • In the index trial, electrolyte abnormalities were frequent at baseline (e.g., severe hyponatraemia and severe hypokalaemia), and electrolyte trajectories differed materially by rehydration strategy; any implementation should pair fluid protocols with a defined potassium and monitoring approach.

Overall Takeaway

GASTROSAM is landmark because it directly tested, in high-burden inpatient settings, a central safety dogma that has constrained IV rehydration in children with SAM. Although it did not demonstrate a mortality advantage for IV rehydration, it provided reassuring absence of overt fluid overload while simultaneously revealing a clinically important electrolyte trade-off (faster sodium correction but more hypokalaemia) that should shape how any practice change is operationalised.

Overall Summary

  • IV rehydration (Ringer’s lactate 100 ml/kg) in SAM with WHO severe dehydration showed similar mortality to a cautious oral/nasogastric strategy and no observed fluid-overload signal.
  • Electrolyte effects are pivotal: IV corrected hyponatraemia faster but increased early severe hypokalaemia.
  • Practice impact depends less on “IV versus oral” and more on bundled delivery: monitoring, potassium strategy, and context-appropriate implementation.

Bibliography