The risk of significant neurocognitive disabilities in preterm survivors is well recognised, particularly under 26 weeks gestation [1, 2]. Although many factors are associated with an increased risk of neurocognitive impairment, postnatal growth failure is now recognized as an important and potentially reversible risk [3–5]. Suboptimal growth is common in very low birthweight infants (VLBWI) [6, 7] especially in those under 26 weeks . Head growth is an especially important measure of growth failure because it correlates with brain growth . Hack et al showed that subnormal head size at 8 months was predictive of poorer verbal and performance IQ scores at 3  and 8 years . Brain growth by 28 days after birth and the expected date of delivery are key predictors of long-term brain growth [12, 13].
Early postnatal growth failure or extrauterine growth restriction describes the severe nutritional deficit that develops in preterm infants in the first few weeks of life [3, 4]. The deficit refers to the gap between the energy and protein (and other nutrients) required to mimic fetal growth rates and the energy and protein that is actually delivered to the preterm infants. Current recommendations suggest a calorie intake of 120kcal/kg/day and a minimal protein intake of 2.5-3 g/kg/day. These are estimates based on matching fetal growth in utero  but do not take into account other factors that may increase individual infant requirements (such as catch-up growth, sepsis and chronic respiratory disease) and therefore increase the risk of postnatal growth failure . Indeed, postnatal malnutrition may be inevitable based on current recommendations [16, 17] and is exacerbated by huge variations in neonatal nutritional practice [18–21].
Very preterm infants have a gut that is too immature to digest milk in sufficient quantity to meet nutritional requirements. Virtually all preterm infants <29 weeks gestation and <1200 g require parenteral nutrition (PN) for a period that depends on gestation birthweight and other morbidities. The mean duration of PN (>75% all nutrition) in these infants (survivors) is 15.6 days  increasing to 20.8 days for infants <700 g . Given these infants have the highest incidence of early and late growth failure and long term neurocognitive disability, effective PN delivery is essential to avoid major early nutritional deficits in these infants.
Inadequate and/or inconsistent nutritional strategies are one barrier to effective PN delivery but there are others. The most important is metabolic "intolerance". Early concerns about amino acid tolerance  continue to have profound effects on nutritional policies . More recent evidence evaluating neonatal amino acid PN formulations, suggests amino acids can be rapidly introduced without metabolic complications [24–28] even in sick infants  and without causing acidosis . This is essential if fetal protein accretion rates are to be matched and the large protein deficits which are routinely encountered in the first week of life are to be avoided . Recommended maximum protein intake is 4 g/kg/day .
Optimal utilisation of protein for preterm infant growth depends on an adequate non-protein energy intake. A minimum of 20-25kcal/g protein is required [22, 32] indicating that 100-120kcal/kg/day is needed to achieve maximal protein accretion . Glucose and lipid infusion rates needed to achieve this may not be tolerated, especially in the first week, leading to hyperglycaemia and hyperlipidaemia. Increasing protein intake without providing an adequate non-protein calorie intake may result in growth failure and increased blood levels of urea and amino acids . Carbohydrate may be the major determinant of optimal growth in preterm infants  and should account for 60-75% calories . Glucose intolerance can be managed with reducing intake but is routinely managed effectively with an insulin infusion [36, 37] although the long term risks and benefits of this approach are still unknown.
Postnatal growth can be improved with increased macronutrient intake [38–40] but evidence for an effect long-term neurodevelopment is more limited. Early introduction of amino acids  can also improve short term postnatal growth but in this study , persistent differences in head circumference did not translate into altered neurodevelopment outcome. Tan et al  did not show improved neurodevelopmental outcome with increased macronutrient intake but did not achieve the differences in nutritional intake expected. A correlation between protein and energy deficit (first 28 days) head growth at 36 weeks CGA was demonstrated and energy deficit (28 days) was associated with worse neurodevelopmental outcome at 3 months . Early nutritional intake of a cohort of extremely low-birthweight survivors  has been correlated with 18 month neurodevelopmental outcomes. This suggested that early head growth failure may have a lasting effect on neurocognitive ability even if there was subsequent catch up growth before term. Provisional reports from other population-based cohort studies have supported this association  suggesting a change in head circumference z-score of -1.4 between birth and 28 days. This is consistent with our own audit findings and those of Tan et al (unpublished data) suggesting head growth failure reaches a nadir at approximately day 28. However, evidence linking early nutritional intervention with improved early head (and then ultimately neurodevelopmental outcome) is still lacking.
The final barriers to effective early nutrient delivery in the very preterm infant are PN prescription, formulation and administration. The conventional neonatal PN strategy has been based on individualised neonatal PN (iNPN) prescription and formulation to address the rapidly changing and variable fluid and electrolyte needs characteristic of the very preterm infant. This can be at the expense of early nutritional strategy when the evidence base supports early and consistent macronutrient delivery. Poor neonatal PN prescribing practice contributes to poor nutrition [45, 46] and computer aided prescribing  can improve protein and energy intake [48, 49]. However, iNPN has other limitations. Although iNPN prescription is flexible, the manufactured individualized PN bag is not so rapid responses to changes in fluid and electrolyte requirements after manufacture is not possible. When Tan et al  compared 2 iNPN regimens with a 30% difference in prescribed macronutrient content, the difference in actual energy and protein intake was <15%. This inefficiency in PN delivery was due to co-administration of other drug infusions, fluid restriction and changing electrolyte requirements. Thus, maximising nutritional intake in very preterm infants cannot be guaranteed by simply increasing the macronutrients in the PN formulation.
Standardising neonatal PN has been considered as an alternative to iNPN regimens  but has receive scant attention in published guidelines [31, 51]. Early evidence suggested iNPN was required to meet the complex of the preterm infant . Although some recent studies concur [49, 52] increasing evidence suggests that with careful attention to local workload and PN prescribing practice most infants can be managed on a standard PN formulation [53–60] sometimes with improved macronutrient intake. Standardised PN solutions that allow some flexibility with electrolytes can overcome the variability in preterm electrolyte needs . Increasing the concentration of neonatal PN (reducing the volume) has the potential to maintain nutritional intake in the face of fluid restriction and multiple drug infusions. Conventionally, stability and osmolality concerns have limited this approach, but current guidelines have virtually no evidence base. High osmolality of aqueous PN solutions can be offset by concurrent administration of intravenous lipids and dextrose.
Using the standardisation and concentration concepts, the preterm infant's competing needs for extreme flexibility for fluid and electrolyte management versus consistent optimal nutritional delivery can be accommodated in a "two compartment" PN model. We developed a standardised concentrated neonatal PN (scNPN) regimen that comprised a relatively inflexible (protected) nutrition compartment (85 ml/kg/day aqueous PN and 15 ml/kg/day intravenous lipid) and a highly flexible supplementary fluid compartment (usually 50 ml/kg/day). This supplementary compartment is then reduced or increased as total fluid requirements demand. Unexpected electrolyte derangement is corrected using standardised electrolyte infusions that replace part of the supplementary infusion as required. All standardized drug infusions are managed in the same way. Changes in infusion rate are titrated against the supplementary infusion not the nutrition compartment. Finally, early introduction of enteral feeds results in the reduction of the supplementary infusion until the enteral feed rate exceeds the supplementary infusion rate. Only then is PN reduced. This system allows maximum flexibility of fluid, electrolyte and drug infusion management with minimal impact on nutrient delivery.
We have shown the scNPN system of PN delivery is more effective at delivering protein, with >90% infants receiving >90% prescribed protein . This lead to a 20% increase in the first 14 day protein intake when compared to a nutritionally identical iNPN regimen . Significant cost reductions were also achieved (38%) similar to those reported for other standardised regimens . There are no randomised controlled trials comparing standardised versus individualised neonatal PN, probably because logistics and patient safety considerations make this unfeasible in the complex very preterm population. However, given the potential benefits of the scNPN, a randomised controlled trial comparing the existing scNPN regimen with one where macronutrient content was maximised (scNPNmax) is desirable.
We speculate that the scNPN and scNPNmax regimens will provide efficient macronutrient delivery in the early neonatal period. We propose that optimising early protein and energy intake will partially correct early head growth failure characteristic of infants <29 weeks gestation. This could have implications for long term neurodevelopment. We hypothesise that the 30% increase in protein and calories achieved by the scNPNmax regimen will lead to a significant improvement in head growth velocity over the first 28 days of life.
To compare the two allocation groups with respect to the rate of head growth from measurement made at enrolment to a measurement made between 27 and 29 completed days after birth (i.e. change in head circumference/(time of last measurement-time of first measurement)
To compare the two allocation groups with respect to the following:
growth measures ( 7, 14, 21 28 completed days and at 36 weeks corrected gestational age (CGA):
occipitofrontal head circumference (OFC), weight, mid-upper arm circumference (MUAC) and lower leg length (LLL)
modelling of weekly head growth, protein and calorie intake data
the efficiency of nutrient delivery (including protocol violations).
Nutritional intake at 7, 14, 21 and 28 days
energy, protein, fat, glucose (including energy and protein deficits)
predicted iNPN intakes based on mathematical model
the tolerance to each regimen by identifying abnormalities (and any required clinical interventions) in the following:
Nutritional tolerance (first 28 days or duration of PN):
protein: daily serum urea, metabolic acidosis, amino acid profile day 7 and 21.
fat: weekly triglyceride profile, hyperlipidaemia
glucose: hypo/hyperglycaemia (including insulin use)
Biochemical tolerance (first 28 days or duration of PN):
Use of supplementary electrolyte infusions
other recognised PN complications
Vascular access device usage and non-infective complications
Neurodevelopmental outcome at 2 years (assessed using Bayley III scales)