Balanced Crystalloid vs Normal Saline in Pediatric Acute Gastroenteritis

Overview

Acute gastroenteritis (AGE) is among the most common reasons for paediatric emergency visits. Children with significant dehydration often require intravenous (IV) fluid therapy. Two main types of IV crystalloid solutions are currently used in clinical practice: 0.9% sodium chloride (normal saline, NS) and balanced crystalloids such as Isolyte-S, which contain acetate and gluconate as bicarbonate precursors. Normal saline has a high chloride content (154 mEq/L), which may worsen the metabolic acidosis already present in many children with acute gastroenteritis. Balanced crystalloids have a chloride content closer to that of plasma (98 mEq/L) and additionally contain acetate and gluconate, which are metabolised in peripheral tissues to consume hydrogen ions and thereby raise serum bicarbonate - a mechanism distinct from simply avoiding chloride overload. This study prospectively observes and compares early biochemical and clinical outcomes in children with acute gastroenteritis who receive one of these two fluid types as part of their routine clinical care. The treating physician independently decides which fluid to use; the research team does not influence this decision and does not order any additional tests or procedures. Laboratory values used as outcomes are drawn solely from blood tests obtained as part of standard care. The primary aim is to determine whether, at approximately 4 hours after IV fluid start, serum bicarbonate has changed more in children who received a balanced crystalloid compared with those who received normal saline. Secondary aims include comparing blood pH, chloride levels, need for additional IV boluses, time to first oral fluid intake, hospitalisation rate, and 72-hour return visits.

BACKGROUND AND SCIENTIFIC RATIONALE Acute gastroenteritis causes gastrointestinal bicarbonate loss through diarrhoea, combined with reduced oral intake and, in severe cases, lactate accumulation from hypoperfusion, resulting in metabolic acidosis. The conventional choice, 0.9% sodium chloride (NS), carries a supraphysiological chloride load (154 mEq/L versus plasma reference of 98-103 mEq/L). This reduces the strong ion difference (SID), independently precipitating hyperchloraemic metabolic acidosis and failing to correct pre-existing acidosis. Isolyte-S has a chloride content of 98 mEq/L and contains acetate (27 mEq/L) and gluconate (23 mEq/L) as bicarbonate precursors. Acetate is rapidly metabolised in cardiac muscle, skeletal muscle, and renal cortex independently of hepatic function (half-life approximately 30 minutes) via: CH3COO- + O2 + H+ → 2CO2 + 2H2O, consuming 1 mEq of H+ per mEq of acetate. Gluconate undergoes hepatic oxidation via the pentose phosphate pathway with analogous H+ consumption. The total buffer capacity of Isolyte-S is 50 mEq/L - approximately twice that of Ringer's lactate (28 mEq/L). This active acid-neutralisation capacity is mechanistically distinct from simply avoiding chloride overload. EVIDENCE BASE The 2023 Cochrane systematic review (Florez et al., 5 RCTs, 465 children) demonstrated improved bicarbonate and pH with balanced crystalloids but judged certainty as low to very low and explicitly called for new adequately powered studies. The meta-analysis by Lehr et al. (2022) reported a pooled delta-HCO3 of +1.60 mmol/L (95% CI 0.04-3.16) favouring balanced fluids in critically ill children. The only RCT using Plasma-Lyte A versus NS in paediatric AGE (Allen et al., 2016) reported delta-HCO3 of +1.6 versus 0.0 mEq/L at 4 hours (p=0.004) but was underpowered for clinical outcomes. No published study has investigated Isolyte-S (acetate/gluconate-buffered, 50 mEq/L total buffer capacity) specifically in children with AGE. STUDY DESIGN Single-centre, prospective, non-interventional observational cohort study conducted in the paediatric emergency department of Aydin Adnan Menderes University Research and Training Hospital, Turkey (KAD-KLVZ-25). No randomisation, allocation, or investigator-initiated prescribing occurs. The treating physician's independent clinical decision determines group assignment before research consent and enrolment. STATISTICAL FRAMEWORK Sample size: 180 total enrolments (anticipated), yielding 126 evaluable participants assuming 70% T4 laboratory availability. Power calculation: minimum clinically important difference delta = 1.5 mmol/L, SD = 3.0 mmol/L, alpha = 0.05 (two-sided), power = 0.80. Primary analysis: PS-IPTW (propensity score inverse probability of treatment weighting) combined with IPOW (inverse probability of observation weighting) to simultaneously address confounding by indication and informative T4 laboratory missingness. A pre-specified covariate list is fixed prior to data lock. Supporting analysis: multivariable linear regression with robust standard errors (HC3). Pre-specified mechanistic mediation analysis: causal mediation framework (R package mediation) to partition the total effect on delta-HCO3 into the indirect effect mediated via delta-Cl (hyperchloraemic mechanism) and the direct effect attributable to acetate/gluconate buffering (ACME and ADE). This analysis is exploratory and will be reported as mechanism-consistent association, not causal inference. Seven pre-specified sensitivity analyses are defined in the statistical analysis plan. Reporting standard: STROBE statement for observational studies. Software: R version 4.3 or later. DATA COLLECTION Data are recorded on a standardised case report form at three time points: T0 (baseline, at IV fluid start), T4 (3-6 hour window, only if routine laboratory results are available), and 72 hours (electronic health record review for return visits). No study-mandated procedures, tests, or visits are added to the routine care pathway. T4 laboratory values are ascertained exclusively from blood gas or electrolyte results obtained during routine clinical management.