Article Text

Download PDFPDF

Do recommended protein intakes improve neurodevelopment in extremely preterm babies?
  1. E A Cester1,2,
  2. F H Bloomfield1,3,
  3. J Taylor4,
  4. S Smith4,
  5. B E Cormack1,3
  1. 1Liggins Institute, University of Auckland, Auckland, New Zealand
  2. 2Neonatal Care Unit, University of Turin, Torino, Italy
  3. 3Newborn Services, Auckland City Hospital, Auckland, New Zealand
  4. 4Child Development Unit, Auckland City Hospital, Auckland, New Zealand
  1. Correspondence to B E Cormack, Neonatal Dietitian, Auckland City Hospital, Level 8. Room 81.038 Support Building, Auckland City Hospital, Private Bag, Auckland 920 24, New Zealand; bcormack{at}


Objective To determine whether achieving recommended protein intakes for extremely low birthweight (ELBW; birth weight <1000 g) babies, resulting in better growth, improves neurodevelopmental outcomes.

Design A prospective cohort study of ELBW babies before and after the introduction of a new nutritional policy designed to meet international consensus protein recommendations. Forty-five children born ‘before’ and 42 born ‘after’ the policy change were assessed at 2 years’ corrected age (CA). Associations between nutritional intakes, growth and neurodevelopmental outcome (Bayley Scales of Infant and Toddler Development, Third edition (Bayley-III), motor and sensory impairment) were assessed using univariate and multivariate analyses.

Results Bayley-III cognitive (mean (SD) 96 (12) vs 96 (15)), motor (96 (13) vs 95 (15)) or language scores (89 (11) vs 91 (17)) were not different between the ‘before’ and ‘after’ cohorts. In the ‘before’ cohort, motor scores were positively associated with enteral nutrition intakes and growth velocity. Neither were sensory impairments different between groups (visual impairment 4 vs 2, hearing impairment 2 vs 0) nor was the gross motor function classification score (any cerebral palsy 2 vs 1).

Conclusions In this prospective cohort study, increasing intravenous and enteral protein intakes to recommended levels in the first month after birth was not associated with improved cognitive, language or motor scores or decreased sensory impairments at 2 years’ CA despite significantly improved early growth and reduced postnatal faltering growth. Appropriate randomised controlled trials are needed to answer definitively whether higher early protein intakes improve neurodevelopmental outcome in this population.

  • Nutrition
  • Growth
  • Neurodevelopment
  • Neonatology

Statistics from

What is already known on this topic

  • Growth, especially in head circumference, in extremely low birthweight (ELBW) babies is highly correlated with neurodevelopmental outcomes and postnatal faltering growth in this population is common.

  • International consensus protein intake recommendations have increased in the past two decades and may decrease faltering postnatal growth. However, whether increased protein intakes in ELBW improve neurodevelopmental outcome is unknown.

What this study adds

  • Despite achieving recommended protein intakes, which resulted in better early growth, there was no difference in neurodevelopmental outcome at 2 years of age.

  • Randomised controlled trials are required to answer definitively the impact of early protein intakes on neurodevelopmental outcome.


Survival in extremely low birthweight (ELBW; birth weight <1000 g) preterm infants has increased substantially in recent decades, but this remains a fragile population at greater risk of long-term sequelae, especially neurodevelopmental impairment and cognitive delay.1 As a population, preterm babies are growth restricted at birth compared with their gestational-age-matched intrauterine peers who go on to be born at term.2 Furthermore, extrauterine growth restriction, or postnatal faltering growth, is a common feature in ELBW infants with many dropping 1 or 2 SD scores for weight, length and head circumference between birth and hospital discharge.3 ,4 Faltering growth in the postnatal period is associated with poorer long-term neurodevelopmental outcomes.5 ,6 Indeed, growth overall, but especially in head circumference, is highly correlated with neurodevelopmental outcomes7 with, for example, examination performance and IQ in adolescents born preterm best predicted by white matter volume.8

To address the widespread postnatal faltering growth in ELBW, international recommendations for protein intakes have been increased progressively9 in recognition of the fact that many ELBW babies develop a negative nitrogen balance in the first weeks after birth that is difficult to redress.10 However, a recent Cochrane review reported on the lack of available evidence for the benefits of early administration of amino acids on mortality, early or late growth and on neurodevelopment.11 It still remains unclear whether, and to what extent, higher protein intakes in early postnatal life could affect neurodevelopment in ELBW infants.

To determine whether the achievement of internationally recommended intakes would improve growth, we previously have reported the growth outcomes from a prospective cohort study of 100 ELBW infants, 50 before and 50 after the implementation of a new nutritional policy designed to ensure ELBW nutritional intakes met international consensus recommendations.12 We observed that babies in the second cohort, after the introduction of the new guidelines, had an increased growth velocity (GV) resulting in a significantly reduced downward crossing of centiles for the three anthropometric indices. We now report the neurodevelopmental outcomes for these babies at 2 years’ corrected age (CA) to determine whether the achievement of recommended nutrient intakes, with the attendant improved growth, was associated with improved neurodevelopmental outcomes.


Characteristics of the prospective cohorts, nutrition policy and its modifications and growth measurements have been described previously.12 Briefly, the first cohort were born between April 2006 and July 2007, following which changes to the nutrition policy were developed and implemented, with the second cohort of babies followed prospectively between August 2009 and November 2010. Inclusion criteria were birth weight <1000 g, admission within 24 h from birth and stay in our neonatal intensive care unit for at least 30 days. Triplets and babies with inborn errors of metabolism were excluded. Nutritional policy changes12 included the commencement of intravenous nutrition, lipid and enteral feeding on day 1 (if not contraindicated), increasing the protein concentration of the standard intravenous nutrition solution to provide a maximum of 4 g/kg/day of protein rather than 3.8 g/kg/day, a change in human milk fortifier (HMF) to provide an additional 1 g protein/100 mL of milk rather than 0.8 g/100 mL, HMF added once enteral feeds reached 5 mL per feed rather than once full enteral feeds were attained, use of preterm formula rather than term formula in babies whose mothers did not provide expressed breast milk and use of a preterm formula with a higher concentration of protein (2.5 vs 2.0 g/100 mL). For both cohorts, data concerning total enteral and intravenous intakes were recorded daily in the first 4 weeks. Mean energy and protein intakes were calculated for weeks 1–4 based on actual intakes. Energy and protein intakes were calculated using preterm transitional breast milk composition for the first 2 weeks (65 kcal and 1.5 g protein/100 mL) and mature breastmilk composition thereafter (72 kcal and 1.4 g protein/100 mL).13 The nutritional guideline changes resulted in increased protein intake (mean (SD) 3.8 (0.3) vs 3.3 (0.4) g/kg/day and increased GV (19.5 (5.0) vs 16.2 (5.4) g/kg/day) over the first month after birth as reported previously.12

A Bayley Scales of Infant and Toddler Development, third edition (Bayley-III) assessment was performed as close as possible to 24 months’ CA14 together with a paediatric assessment. The psychologist performing the assessment was blinded as to which cohort babies were in. Outcome categories were assigned according to the criteria shown in table 1.

Table 1

Outcome categories for infants <30 months of age (modified from refs 15 ,16)

Data concerning the need for speech and language therapy, physiotherapy, occupational therapy, neurodevelopment therapy, early teacher intervention and dietitian follow-up were collected retrospectively from our psychologist reports. Cerebral palsy or motor impairment was also defined using the Gross Motor Function Classification Score (GMFCS).17

Data analysis was performed with JMP V.10 (SAS Institute, Cary, North Carolina, USA). Continuous data were tested for normality, transformed where appropriate and analysed with unpaired t test or repeated measures analysis of variance. Nominal variables were assessed by Fisher's exact test. Linear regression was used to investigate the potential confounding effects of sex and small-for-gestational-age (SGA; <10th percentile on the UK-WHO growth chart) status on Bayley-III scores. Associations between nutrient and energy intakes during the first 4 weeks, growth and Bayley-III scores initially were investigated by univariate analysis. Significant variables (p<0.05) were entered into a stepwise forward regression model, with an F to enter of 2.5.

The Human Disability and Ethics Committee confirmed that ethical approval was not required for this study.


Characteristics of the overall cohort (100 babies) have been reported previously.12 Patient characteristics and morbidities of the surviving babies who were assessed at 2 years’ CA were similar in the two cohorts (table 2).

Table 2

Baseline characteristics of babies assessed at 2 years’ corrected age in the two cohorts: before and after the change in nutritional guidelines

Of the original 50 babies in the cohort before the change in nutrition policy, three died and two did not attend follow-up; in the ‘after’ cohort, one baby died, six babies declined to attend any further follow-up and one was lost to follow-up. Eight babies in each group did not have a Bayley-III assessment at 2 years’ CA and two babies in the ‘after’ cohort were not evaluated because of severe impairment. Thus, Bayley-III scores were available for 37 babies in the ‘before’ cohort and for 32 babies in the ‘after’ cohort, with assessments conducted at a mean (SD) CA of 24 (4) months.

No differences were observed between the two groups in the cognitive, motor or language scores (cognitive 96 (12) vs 96 (15), p=0.8; motor 96 (13) vs 95 (15), p=0.7; language 89 (11) vs 91 (17), p=0.6). Linear regression models demonstrated that motor scores were significantly less (p=0.03) in male SGA babies (84 (5)) than in male AGA (99 (3)) or in female babies (SGA: 101 (4); AGA: 94 (3)). Outcome category, sensory impairment, the need of special support and the GMFCS were available for 45 babies of the ‘before’ cohort and for 42 babies of the ‘after’ cohort and were not different between groups (table 3).

Table 3

Functional impairment and need for special intervention of the babies assessed at 2 years’ corrected age in the two cohorts: before and after the change in nutritional guidelines

There was a trend (p=0.05) for a greater need for an early intervention teacher among the babies in the ‘before’ cohort. Cognitive and language scores were not associated with nutrient intakes in either cohorts or overall. In the ‘before’ cohort only, there was a significant positive, but fairly weak, correlation between motor score and GV in the first 30 days after regaining birth weight (table 4).

Table 4

Variables significantly associated with Bayley-III motor score in ‘before’ cohort

There were no significant correlations between enteral and intravenous nutrition intakes and Bayley-III scores in the ‘after’ cohort (data not shown). Motor scores of babies in the ‘before’ cohort were positively associated with enteral, but not intravenous, nutrition intakes and with GV (table 4).

Commencement of enteral feeding on the first day after birth was associated with higher motor scores (101 (13) vs 91 (11) in those who did not commence enteral feeds on day 1; p<0.01). These associations remained significant after adjusting for sex, birthweight z-score and days of ventilation, included as a proxy for illness severity.


In this prospective cohort study, a significant increase in protein intake in the first month after birth did not alter Bayley-III scores at 2 years’ CA, despite improved early growth. We also did not find any difference in cerebral palsy or sensory deficits between the two groups, consistent with other reports in the literature.20 ,21 High protein nutrition and its effect on neurodevelopment remains a controversial issue and the literature is not homogeneous in its findings. An observational study by Stephens et al22 of 148 ELBW infants found an independent positive correlation between protein intake in the first week after birth and Bayley II Mental Development Index (MDI) score at 18 months’ CA. In Stephens’ study, the majority of energy and protein intakes during the first week after birth were intravenous and it should be noted that some babies received very low protein intakes, in contrast to the babies we report here. This also is true for other papers reporting the relationship between neurodevelopment outcome and intravenous protein intakes.20–22 Blanco et al21 reported a lower Bayley II MDI score at 18 months’ CA in infants who received 2 g/kg/day intravenous protein in the first week increasing to 4 g/kg/day compared with their counterparts who received 0.5 g/kg/day increasing to a maximum of 3.5 g/kg/day. This difference was no longer found at the age of 2 years. Of note is the fact that the higher amino acid group actually had poorer growth; thus, although this study appears to contradict the hypothesis that greater protein intakes in early life improve neurodevelopmental outcome, it is consistent with the hypotheses that in extremely preterm babies, protein intakes are probably inadequate for optimal growth23 and that early growth and neurodevelopmental outcomes in this population are linked. In contrast, a study comparing Bayley II MDI score at 18 months’ CA between ELBW infants who received at least 3 g/kg/day of protein in the first 5 days after birth and those who received <3 g/kg/day found no differences between the two groups.20 In this study, enteral nutrition contributed to approximately 5% of the total energy intake by day 7. In our study, approximately 25% of week 1 energy intake was from enteral nutrition in both cohorts. Early enteral nutrition also is advocated to improve outcomes24 ,25 and this factor must be considered when interpreting studies in which enteral nutrition was either restricted to minimal enteral feeds or significantly delayed. We found a significant correlation between motor scores and enteral intakes during week 1 after birth, but only in the ‘before’ cohort. One explanation might be that healthier infants are more likely to receive more enteral nutrition and these infants would be expected to have a greater likelihood of better Bayley-III scores. However, this association remained after adjusting for variables such as birthweight z-score and duration of ventilation and a similar relationship was not found in the ‘after’ cohort, despite enteral intakes not being different between the two cohorts during the first week after birth.12 An alternative explanation might be that in babies in whom total protein intake is insufficient, improving enteral intake improves outcome whereas once total protein intake is sufficient this is no longer the case.

Extremely preterm babies frequently suffer both fetal and postnatal growth restriction4 ,26–28 with faltering postnatal growth being very common4 ,29 due to the difficulties in providing nutrition sufficient to support the very rapid rate of growth that fetuses of equivalent gestational age undergo in utero.4 ,30 Both fetal growth restriction and postnatal faltering growth are associated with adverse neurodevelopmental outcome.5 ,31 Thus, it seems likely that suboptimal nutrition contributes to both the faltering postnatal growth and the impaired neurodevelopmental outcomes that are seen commonly in ELBW babies.23 ,32 Faltering growth in preterm babies is different from the classical faltering growth of poorly nourished term infants where weight is affected first followed by length with sparing of head growth until severe malnutrition is present. In preterm babies, length and head growth appear to be affected much more quickly with z-scores falling by 1–2 SDs for all measures of growth during the first month of postnatal life.12 ,33 This early effect on head growth likely reflects the extremely rapid brain growth that occurs at the gestational ages ELBW babies are receiving suboptimal nutrition as head circumference correlates with brain volume.34 There is an increase in total brain tissue volume of 22 mL each week from 29 to 41 weeks’ gestation,35 resulting in a fourfold increase in cerebral cortical volume between 28 and 40 weeks35–37 and a >30-fold increase in cerebellar cortex surface area between 24 weeks and term.38 ,39 This very rapid development may also result in greater vulnerability to damage from less than optimal nutrition. Indeed, in adolescents born preterm, performance in final school examinations and IQ score is best predicted by the total volume of white matter in the brain detected on MRI scans at a mean age of 16 years, irrespective of the presence or severity of other brain abnormalities.8 However, despite our data demonstrating improved head growth in the ‘after’ cohort, consistent with the results of a recent randomised controlled trial which demonstrated that protein intakes equivalent to those in our ‘before’ group resulted in improved head circumference at 28 days compared with the control group that received lower protein intakes,40 we did not find improved neurodevelopmental outcomes as assessed by Bayley-III scores. This could be due to the observational, cohort nature of the study or because there is no beneficial effect for neurodevelopment of increasing protein intake in extremely preterm babies above a certain threshold and, perhaps, protein intakes in both the ‘before’ and ‘after’ groups were above this threshold. Finally, 2-year outcomes are quite early in the context of extremely preterm birth; it may be possible that any beneficial effects are not manifest until older ages. Limitations of this study are the potential bias inherent in non-randomised studies, as well as the non-contemporaneous nature of the two cohorts, meaning that other changes in management practices over the duration of the study that we have not identified could also affect outcomes. The outcome data reported here were collected retrospectively and some data, such as the need for special intervention, were provided by the parents. Thus, these data need to be interpreted with caution and further studies should be performed to confirm the relationship between the need for an early intervention teacher and early nutritional intakes in ELBW babies.

We acknowledge that this follow-up study is underpowered to detect a clinically meaningful difference in Bayley-III score. Thus, although we have not found an effect of increased protein intake on Bayley-III scores at the age of 2 years in this cohort study, this does not mean that there is no effect. For example, to detect a five-point Bayley-III score difference with 90% power at a 5% level of significance would require 190 babies per group with a total of 380 subjects. A more meaningful difference in neurodevelopmental outcome may be survival free of disability at 2 years’ CA. To detect an absolute difference of 15% in survival free of disability between the two groups at 2 years’ CA requires a total of 430 babies (n=215 in each group) with 85% power and a 5% level of significance (two-sided), taking into account a 10% loss to follow-up rate and a rate of 50% in the control group.

A randomised controlled trial of increased protein intakes in early life with neurodevelopmental outcome as the primary outcome is required to answer definitively whether increased protein intakes in the first weeks after birth in extremely preterm babies improves survival free from neurodisability. Such a trial (the ProVIDe study) is now recruiting (ACTRN126 12001084875).



  • Contributors EAC wrote the study proposal, collected and analysed the data and wrote the first draft of the manuscript. JT and SS performed the Bayley-III assessments and advised on the study proposal. BEC and FHB advised on the study proposal, ethics approval and supervised the data analysis. All authors contributed to and approved the final manuscript.

  • Competing interests None.

  • Ethics approval The Northern Regional Ethics Committee (Auckland, New Zealand).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.