Abstract
We thank Hari Prasath C and Anil Kumar for their thoughtful and detailed appraisal of our prospective study on oral-first rehydration for moderate neonatal hypernatremic dehydration (NHD). 1 Their commentary engages seriously with both the strengths and the limitations of our hypernatremic dehydration treatment (HDT) clinic protocol, and we welcome the opportunity to respond to the points raised.
The commentators correctly identify the central methodological advance of our study: the replacement of sodium-level thresholds as the sole determinant of rehydration route with a structured, clinically stratified protocol. Their observation that group membership – reflecting protocol-guided decision-making – was the strongest independent predictor of intravenous fluid use in our multivariate analysis (OR 28.267, 95% CI 10.321–53.69, p < 0.001) accurately captures our primary message. A haemodynamically stable neonate who accepts feeding and maintains urine output should not be routed to intravenous rehydration solely on the basis of an admission sodium of 160 mEq/L; clinical context and a structured escalation framework are the operative variables.
Regarding the correction rate, we fully acknowledge that the median sodium reduction rate of 0.66 mEq/L per hour in our prospective group, while significantly safer than the pre-protocol rate of 1.05 mEq/L per hour still exceeds the widely recommended ceiling of 0.5 mEq/L per hour. 2 We emphasise, however, that this figure represents a median across the entire cohort, including patients who received orogastric and intravenous rehydration. The subset managed with pure oral rehydration – the majority – had slower, more gradual sodium correction, which is consistent with the physiological basis for preferring the enteral route. We agree with the commentators that the current study is not powered to detect low-frequency neurological events and that the absence of seizures cannot be interpreted as definitive proof of safety at this correction rate. Prospective randomised trials with predefined neurological endpoints are needed, and we explicitly called for these in our original publication. Refinement of oral rehydration volumes per kilogram per hour across sodium strata within the moderate range is indeed the next iteration priority.
On the absence of standardised neurological assessment at discharge, this is a genuine limitation we acknowledge. Our primary safety endpoint was seizure-freedom, which, while clinically meaningful, does not capture subclinical cerebral injury detectable on cranial ultrasound or formal neurological scoring. We agree that future protocol-based studies should incorporate cranial ultrasound and structured neurological assessment as predefined safety co-endpoints, and we intend to include these in our ongoing follow-up work with this cohort.
The commentators raise a valid concern regarding the operationalisation of step-transition criteria – specifically, the absence of explicit time windows governing progression from oral to orogastric feeding and from orogastric to intravenous fluids. In practice, transitions were guided by attending neonatologist assessment at scheduled clinical reviews, but we recognise that this approach limits reproducibility in centres with different nurse-to-patient ratios and monitoring resources. A time-anchored escalation matrix is a constructive suggestion that we will incorporate into future protocol iterations and publications.
We are particularly pleased that the commentators highlight our identification of serum uric acid as an independent predictor of intravenous rehydration need. This finding was unexpected and, to our knowledge, has not previously been described in moderate NHD as a triage variable. As the commentators note, elevated serum uric acid in the dehydrated neonate reflects reduced renal urate clearance and increased purine catabolism under conditions of volume depletion, and may convey information about metabolic severity not captured by serum sodium or creatinine alone. 3 We concur that prospective validation of serum uric acid as a triage biomarker in moderate NHD is warranted and plan to pursue this in a larger cohort.
Finally, the commentators correctly identify the absence of a structured lactation support component as a gap in our protocol. NHD arises almost exclusively from inadequate breastfeeding, and the prevention of recurrence is directly modifiable through lactation counselling. Lactation support was provided as part of routine care during the study period, but it was not formalised as a protocol element and was not systematically documented. We agree that any protocol addressing NHD should incorporate a concurrent, standardised lactation intervention as an integral component, not an adjunct. The contextual factors operating in our population – including a high caesarean section rate that may delay lactogenesis – make population-specific lactation strategies particularly important, and we will address this in future protocol revisions.
In conclusion, we thank Hari Prasath C and Anil Kumar for a rigorous and constructive appraisal. Their commentary strengthens the evidence base for the limitations that must be addressed before oral-first rehydration protocols can be adopted as standard of care for moderate NHD, and we are in broad agreement with the research priorities they identify: adequately powered randomised trials with predefined neurological and neurodevelopmental endpoints, refined volume titration, and integrated lactation support.
