You are here

Article Text

Article menu
Original article
Early growth and neurodevelopmental outcome in very preterm infants: impact of gender
  1. A Frondas-Chauty1,2,
  2. L Simon1,
  3. B Branger3,
  4. G Gascoin4,
  5. C Flamant1,2,
  6. P Y Ancel5,
  7. D Darmaun2,6,
  8. J C Rozé1,2,3
  1. 1Department of Neonatal Medicine, University Hospital of Nantes, Nantes, France
  2. 2INRA UMR 1280, Physiologie des Adaptations Nutritionnelles, IMAD, CRNH Ouest, Nantes, France
  3. 3Loire Infant Follow-Up Team (LIFT) Network, Pays de Loire, France
  4. 4Department of Neonatal Medicine, University Hospital of Angers, Angers, France
  5. 5Institut National de la Santé et de la Recherche Médicale, Epidemiological Research Unit on Perinatal Health and Women's and Children's Health, Tenon Hospital, Paris, France
  6. 6Nantes University, IMAD, Nantes, France
  1. Correspondence to Dr Anne Frondas-Chauty, CHU de Nantes, Service de Réanimation néonatale, 38 bd Jean Monnet, 44093 Nantes Cedex 1, France; anne.frondas{at}chu-nantes.fr

Abstract

Background and objective Nutrition in the neonatal unit may impact the neurological outcome of very preterm infants, and male preterms are more likely to suffer neonatal morbidity and adverse neurological outcomes. We hypothesised that growth during hospitalisation would impact neurological outcome differently, depending on infant gender.

Methods Surviving infants born between 1 January 2003 and 31 December 2009 with a gestational age <33 weeks, and enrolled in Loire Infant Follow-up Team, a regional cohort in western France, qualified for the study. Growth during neonatal hospitalisation was assessed by the change in weight z-score between birth and discharge, and infants where ranked into 5 classes, depending on their change in z-score (<−2, −2 to −1.01, −1 to −0.51, −0.50 to 0.01 and ≥0), the last class being the reference. The main outcome criterion was neurodevelopmental outcome at 2 years of corrected age. For each class of changes in weight z-score, crude or adjusted OR for non-optimal outcome was calculated for each gender, and compared between genders.

Results 1221 boys and 1056 girls were included. Gender and early growth interact, (p=0.02). Moreover when change in weight z-score varied from <−2 to (−0.50 to −0.01), adjusted OR for non-optimal outcome varied from 3.2 (1.5–6.8) to 2.2 (1.2–4.1) in boys versus 1.8 (0.7–4.2) to 0.95 (0.4–1.9) in girls. For each class, the OR was significantly higher in boys.

Conclusions In very preterm infants, male neurodevelopment appears to be much more sensitive than female to poor postnatal growth.

  • Growth
  • Neurodevelopment
  • Neonatology
  • Outcomes research
  • Nutrition

https://doi.org/10.1136/archdischild-2013-305464

Statistics from Altmetric.com

    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.

    What is already known on this topic

    • Early growth in very preterm babies influences neurodevelopmental outcome.

    What this study adds

    • Male neurodevelopment appears to be much more sensitive than female to poor postnatal growth.

    • This effect appears as soon as change of weight Z-score is under than −0.01 during hospitalisation.

    Introduction

    Extrauterine growth restriction commonly occurs in neonatal units.1 Very preterm infants indeed frequently suffer cumulative energy and protein deficits, resulting in a loss of at least one weight z-score in the 1st month of life.2 Over the last decade, evidence has accumulated to suggest a strong correlation between the growth performance of very preterm infants in the neonatal unit, and neurological outcome after a year of corrected age.3 Ehrenkranz demonstrated that achieving a weight gain between 12 g/kg/d and 21.2 g/kg/d during hospital stay significantly reduced the risk of neurological impairment, and abnormal mental and psychomotor developmental indexes at 18 months of corrected age.4 Current nutritional procedures for extreme premature infants therefore have become more ‘aggressive’,5 in an effort to curb cumulative energy.6

    Among cohorts of preterm infants, boys have long been known to be more sensitive to neonatal morbidity and adverse neurological outcome.7 Moreover, recent studies documented a lower fat-free mass in boys, compared with girls, at the time of hospital discharge.8 ,9 The question thus arises, whether the relation between extrauterine growth restriction on neurodevelopment may differ, depending on gender.

    The aim of the current study was to determine in a relatively large regional cohort of very preterm babies less than 33 weeks of gestation, whether the rate of initial growth in the neonatal unit affected neurological outcome at 2 years of age differently depending on infant gender.

    Methods

    Study population and data collection

    Data source constituted the Loire Infant Follow-up Team (LIFT)10 cohort which is a cohort of infants born in one region (Pays de la Loire, a region in western France) and enrolled in the regional follow-up network. All preterm infants less than 33 weeks are included in the cohort. Enrolment is initiated during neonatal hospitalisation in the five neonatal intermediate or/and intensive care units of the region. Approximately 95% of all surviving infants <33 SA were enrolled in the LIFT cohort. Infants excluded were principally those born in regional hospitals but living in another geographical area, and infants whose parents refused the inclusion in the cohort.

    Among surviving, very preterm infants with a gestational age <33 weeks born between 1 January 2003 and 31 December 2009 and enrolled in LIFT cohort, we included all infants discharged between 34 weeks and 44 weeks of corrected age, and whose anthropometric measurements at birth and at discharge and date of discharge were known. We excluded those with a genetic syndrome. The cohort was registered at the French Commission Nationale de l'Informatique et des Libertés (no 851117) ethics committee. The children’s parents provided written informed consent at enrolment. At the time of inclusion, data from the medical chart were collected; this data were checked for accuracy at regular intervals by a quality assurance engineer.

    Developmental assessment at 2 years

    The infants were evaluated at 2 years of corrected age. Children were assessed for either optimal or non-optimal neurodevelopmental outcome, after physical examination by a trained paediatrician and a psychomotor evaluation by a LIFT network psychologist. Neuromotor function was regarded as non-optimal in cases of cerebral palsy or when physical examination revealed milder signs of abnormal movement (phasic stretch in the triceps surae muscle and imbalance of passive axial tone with predominance of extensor tone) during independent walking. Psychomotor evaluation was assessed with the revised Brunet-Lézine test.11 This test is administered by specialised psychologists to assess four developmental areas: movement/posture, coordination, language and socialisation. Each of these domains is graded separately, and when combined yield a global developmental quotient. The mean and maximal global developmental scores are 100 and 140, respectively, and values <85 were used to define non-optimal psychomotor development. Infants who were not able to perform the Brunet-Lézine test due to severe neurological impairment were considered as having non-optimal psychomotor development. In the absence of psychological evaluation, neurodevelopmental outcome was assessed with the Ages and Stages Questionnaire (ASQ)12 which is a parent-completed questionnaire that assesses development in the five areas, that is, communication, gross motor, fine motor, problem solving and personal-social skills. Parents filled the ASQ form while in the waiting room before the child's physical examination was performed. The maximal overall ASQ score is 300, and a score of <185 was considered as non-optimal13 (an ASQ score of <185 had a specificity of 0.9 and a positive likelihood ratio of 8 to predict a Brunet-Lézine test less than 85).13 Children with a non-optimal neuromotor and/or psychomotor assessment were regarded as having overall a ‘non-optimal neurodevelopmental outcome’. Children without documented physical examination or psychomotor assessment were considered non-assessable at 2 years, except for infants with severe neurological disability.

    Growth assessment

    To assess growth, we used measurements performed at birth and discharge. We calculated z-score by using λ-μ-σ method (LMS),14 as follows:Embedded Image

    where L, M and S represent respectively λ, μ and σ corresponding to gestational age to calculate birth weight z-score, and to postmenstrual age at discharge to calculate discharge weight z-score.

    Thus, we used reference growth curves for which LMS parameters have been published for weight, height and head circumference. For birth and discharge (up to 41 weeks of postmenstrual age) measurements, we used Olsen's preterm infant growth chart.15 For the few preterm infants discharged after 41 weeks of postmenstrual age and for follow-up measurements, we used WHO growth curves.16 To calculate propensity score by binary logistic regression, we needed a binary variable of interest. Thus we considered a premature infant had experienced ‘adequate or appropriate extrauterine growth’ when the change in weight z-score between birth and discharge was equal to or higher than zero, the change being calculated as follows:Embedded Image

    Statistical analysis

    Means and SDs are reported for continuous variables and frequencies for categorical variables. Analysis of variance (ANOVA) and χ2 or Fisher's test, if necessary, were used to compare infant characteristics and 2-year outcome according to variation of weight z-score during hospitalisation, and to compare data of neonatal hospitalisation between infants assessed or not at 2 years of age.

    Neurodevelopmental outcome at 2 years of age was treated as a binary variable: having or not having a ‘non-optimal neurodevelopmental outcome’.

    The propensity score method was used to reduce bias in assessing the relation between variation of weight z-score during hospitalisation and neurodevelopmental outcomes at 2 years of age.17 ,18 The propensity score is defined as a conditional probability, between 0 and 1, that a subject will have ‘appropriate growth’ during hospitalisation, based on an observed group of covariates. This score is then used just as if it were the only confounding covariate. Thus, the collection of predictors is collapsed into a single predictor, which may better adjust covariates between the groups and reduce bias. Several full non-parsimonious logistic regression models were developed to derive a propensity score for appropriate growth (ie, change of weight z-score during hospitalisation ≥0). The models included true confounders: variables that are potentially associated with growth in the neonatal unit and outcome. These variables included: characteristics of the newborns (gender, gestational age (GA), z-score of birth weight, region of birth), characteristics of pregnancy (antenatal corticosteroids, multiple pregnancy, hypertension during pregnancy) and characteristics of neonatal hospitalisation (Apgar score, intubation at birth, nosocomial infection, haemodynamic shock, patent ductus arteriosus treated with ibuprofen or surgery, severe cranial ultrasound/MRI abnormalities, bronchopulmonary dysplasia defined by supplemental oxygen requirement at 28 days or 36 weeks, length of parenteral nutrition, breastfeeding at discharge). Patients with missing data were excluded from multivariate analysis. The Hosmer-Lemeshow goodness-of-fit test and the area under the receiver operating characteristic (ROC) curve were used to assess model fit. The ROC curve plots the sensitivity of propensity score as a function of (1 − specificity) for appropriate extrauterine growth.

    First, statistical interaction analysis was deployed to identify the interaction effect between genders and appropriate growth in hospitalisation. Second we studied by gender, crude associations between variation of weight z-score during hospitalisation and non-optimal neurodevelopment assessment and then the same associations after adjustment for the propensity score. We used logistic regression models for univariate and multivariate analyses. We estimated the crude and adjusted OR for propensity score, and its 95% CI, of a non-optimal neurodevelopmental score, at 2 years corrected age in LIFT cohort, associated with growth during hospital stay. Growth was studied in clusters and as a continuous variable. Third, we compared between genders adjusted OR for propensity score, of a non-optimal neurodevelopmental score in each cluster of variation of weight z-score with this formula:Embedded Image

    Fourth we performed several sensitivity analyses. To facilitate comparisons, we used weight z-score as a continuous variable. We calculated the adjusted ORs for a non-optimal neurodevelopment at 2 years of age in female and male preterm infants, per one weight z-score lost between birth and hospital discharge in the whole cohort, and in subgroups: infants who were discharged before 42 weeks, infants with a gestational age less than 29 weeks and more than 28 weeks, and among preterm infants with and without developmental quotient (DQ) at 2 years of age.

    Finally to evaluate the impact of missing data, we performed multiple imputations. Ten imputed data sets were generated for male and female genders.

    All p values were based on two-sided tests. All analyses were performed using SPSS V.15.0.

    Results

    Study population and follow-up

    Of the 2889 preterm infants born before 33 weeks of gestation during the 7 years of the study period, 2760 were enrolled in follow-up (figure 1). Among these 2760 infants, 2299 were discharged between 35 weeks and 44 weeks, with anthropometric measurements at discharge recorded in the database. Four hundred and forty-seven infants were excluded because of data not registered: birth weight (24 infants), weight at discharge (144 infants) and date of discharge (279 infants). Fourteen infants were excluded because discharge was after 44 weeks. The 842 infants born during the study period but not assessed at 2 years, had a better growth during hospitalisation (weight z-score change of −0.7 (±0.78)), in comparison with −0.85 (±0.83) in the 2047 infants assessed at 2 years of age. Moreover infants not assessed had shorter oxygen therapy and parenteral nutrition, less intrauterine growth retardation, but less antenatal corticotherapy and breastfeeding at discharge (table 1).

    Figure 1
    Figure 1

    Cohort profile.

    View this table:
    Table 1

    Comparison between infants assessed or not at 2 years

    Among the population included, 22 infants died after discharge or had a genetic syndrome. A total of 2277 infants with 1056 girls and 1221 boys were finally included in this analysis (table 2 and web only tables 2B and 2C). Neurological assessment was missing for 230 infants (10%). For psychomotor evaluation of the 2047 infants assessed at 2 years, 1164 infants underwent a Brunet-Lézine test, 1833 infants had an ASQ and 1121 infants had both. One hundred and seventy-one infants underwent physical examination but no psychomotor evaluation because this was impossible due to their severe neurological disability (n=63) or because calculation of the DQ was not feasible during psychomotor evaluation (n=108). Finally psychomotor evaluation was realised by an ASQ instead of a Brunet-Lézine test for 400 female and 483 male infants.

    View this table:
    Table 2

    Demographic and clinical characteristics of very preterm infants according to changes in weight z-score between birth and discharge

    Growth and outcome

    An inefficient growth during hospitalisation is associated with a non-optimal neurological outcome at 2 years of age. Girls had more adequate growth. We observed interaction between gender and growth during hospitalisation (p=0.02).

    Risk factors for appropriate growth during hospitalisation

    Propensity score for appropriate growth was calculated for 2277 infants. The propensity scores ranged from 0.001 to 0.681. The Hosmer-Lemeshow test was 6.9, p=0.546. The ROC curve area was 0.78±0.013. So the model seems to be appropriate. Variables associated with satisfactory growth were intrauterine growth retardation, geographical origin (1) and length of parenteral nutrition between 4 days and 12 days. Infants with early intubation, severe cranial ultrasound or MRI abnormalities, and breastfeeding at discharge had a risk of loss of weight z-score during hospitalisation. ‘Gestational age below 30 weeks’ and severe neonatal events such as nosocomial infection, haemodynamic shock, treatment with ibuprofen, surgery for patent ductus arteriosis or bronchopulmonary dysplasia, were not significant factors in determining growth during hospitalisation (table 3)”.

    View this table:
    Table 3

    Associations between all the variables included in calculation of propensity score and changes in weight z-score ≥0 during neonatal hospitalisation

    Growth and neurodevelopmental outcome by gender, after adjustments

    For boys, a loss of weight z-score was associated with a ‘non-optimal neurological outcome’ (table 4). This association was still found after adjustment on gestational age and z-score of birth weight and after adjustment on the propensity score. Adjusted OR ranged from 2.1 to 3.2. For girls an inefficient growth had a lower effect that did not reach statistical significance, on neurological outcome at a corrected age of 2 years. Comparison of adjusted OR between genders was statistically significant for each cluster of weight z-score (table 4).

    View this table:
    Table 4

    Crude and adjusted for propensity score ORs for non-optimal neurodevelopment at 2 years of age in female and male preterm infants, according to weight z-score change between birth and hospital discharge

    Sensitivity analysis in subgroups shows that growth was significantly associated with a ‘non-optimal neurological outcome’ when growth was studied as a continuous variable, or if preterm was more than 28 weeks of gestational age, or if we take into account only preterm infants discharged before 42 weeks. Multiple imputations did not modify this effect (table 5).

    View this table:
    Table 5

    Sensitivity analyses

    Discussion

    In a regional cohort of very preterm infants, we found an association between growth velocity during the stay in neonatal unit and neurological outcome at 2 years of corrected age. Furthermore, the effect of early growth was significantly stronger in boys than girls, with a higher risk of non-optimal neurological outcome when weight z-score declined during hospitalisation in neonatal unit. Ehrenkranz was first to document the relation of weight gain (expressed in g/kg/d) on neurological outcome.4 Recently, Belfort confirmed such a relation: a gain of weight z-score between 1 week of postnatal age and term is associated with improved neurological outcome at 18 months of corrected age, as assessed by the motor and psychomotor development indexes of Bayley Scales of Infant Development.19 At 5 years of corrected age, this lack of early weight gain during the first few weeks of life remains an important predictor of cognitive impairment in extreme premature babies, despite the lack of impact on brain size, estimated from changes in head circumference.20

    Boys thus are more sensitive to poor growth in terms of neurological outcome. In the very preterm population, boys have poorer neonatal profile as they are more prone to sepsis, major surgery and intraventricular haemorrhage.7 Male sex, even after multiple adjustments nevertheless remains an independent and important risk factor for poor neurological and respiratory outcome at follow-up.21 Kent described this gender difference for neurological outcome until 27 weeks of gestation. Boys may be more vulnerable to the stress associated with birth. Studies on placentae of mothers who delivered between 24 weeks and 31 weeks gestational age described sex-specific alterations of pro-oxidant/antioxidant balance with a predominantly pro-oxidant status in placentae of male infants.22 Later on, nutrition and growth could affect boys more dramatically than girls. Lucas studied IQ at 8 years in premature babies who had been fed either standard or enriched enteral formulae. A negative effect of unenriched formula on IQ was observed only in boys.23 Similarly, docosahexaenoic acid supplementation of preterm formulae improved neurological outcome only in girls.24 The brain undergoes rapid growth during hospitalisation. A suboptimal nutrition could alter its organisation with a particular sensitivity in boys. Using cerebral tractography, Liu described gender differences in language and motor-related fibres.25 A specific sensitivity of those fibres to nutrition during the third trimester of pregnancy could explain gender differences in the relationship between initial postnatal growth and neurological outcome for very preterm infants.

    Growth velocity in neonatal unit depends on adverse medical events and nutritional intake. The loss of weight z-score in our study was associated with postnatal intubation and severe cerebral lesions. Surprisingly, neither gestational age, nor severe neonatal events such as infection or bronchopulmonary dysplasia were associated with non-optimal growth. In our study preterm infants born small for gestational age or receiving intravenous infusions for a longer time were surprisingly less exposed to a loss in z-score weight. Likewise Senterre and Rigo26 described a better gain of weight z-score in small for gestational age premature infants. Such better outcome may be due to the fact that prescribing physicians may pay particular attention to the nutritional management of the more frail preterm infants, and try their best to optimise their nutritional protocol, which would explain the progress made over the last 10 years in addressing the postnatal cumulative nutritional deficit.6 More ‘aggressive’ nutrition with enhanced protein and energy intakes increases fat-free mass accretion27 so growth during hospitalisation may need to be assessed, not just based on weight z-score, but in terms of ‘quality of growth’. Studies of body composition performed in premature babies document fat mass excess9 but the putative effect of such ‘abnormal’ body composition on neurological outcome at 2 years remains unknown.27 Breastfeeding paradoxically results in a loss of weight z-score during hospitalisation, but a better neurodevelopment at 2 years or 5 years.28

    The main limitation of the current study stems from its observational nature. Nevertheless, the relation of growth during neonatal hospitalisation was found to depend, not on gestational age, but on gender. Another limitation was that analyses were performed with data from 71% of the entire population born during inclusion period. Nevertheless the sample infant population assessed at 2 years of age had more severe neonatal events and less appropriate growth that the entire population enrolled. The third limitation was the age at which neurological outcome was evaluated. We defined non-optimal neurological assessment with neuromotor and psychomotor evaluation at 2 years of corrected age, an age at which neither language nor fine motor disability and praxis ability can be examined. So we may misjudge the neurological effect of a poor initial growth. Nevertheless this study was population based in a large cohort of very preterm infants, with the same neurological assessment for boys and girls. The gender-specific sensitivity of early growth therefore is a robust finding, consistent with gender-specific differences already described for other neonatal events.21

    In conclusion, inadequate growth in the neonatal unit is associated with neurological outcome, with gender specificity. The mechanisms responsible for such specificity are incompletely understood. Regardless of its mechanisms, such findings suggest that when conducting randomised studies in neonatology, it may be necessary to stratify randomisation according to infant gender. In clinical practice, growth in the neonatal intensive care unit needs to be followed with intense scrutiny, particularly in male infants. Whether specific nutritional strategies should be developed for male infants would warrant further investigation.

    References

    1. ↵
      1. Clark RH,
      2. Thomas P,
      3. Peabody J
      . Extrauterine growth restriction remains a serious problem in prematurely born neonates. Pediatrics 2003;111(5 Pt 1):986–90.
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Embleton NE,
      2. Pang N,
      3. Cooke RJ
      . Postnatal malnutrition and growth retardation: an inevitable consequence of current recommendations in preterm infants? Pediatrics 2001;107:270–3.
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Ramel SE,
      2. Demerath EW,
      3. Gray HL,
      4. et al
      . The relationship of poor linear growth velocity with neonatal illness and two-year neurodevelopment in preterm infants. Neonatology 2012;102:19–24.
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Ehrenkranz RA,
      2. Dusick AM,
      3. Vohr BR,
      4. et al
      . Growth in the neonatal intensive care unit influences neurodevelopmental and growth outcomes of extremely low birth weight infants. Pediatrics 2006;117:1253–61.
      OpenUrlAbstract/FREE Full Text
    5. ↵
      1. De Curtis M,
      2. Rigo J
      . The nutrition of preterm infants. Early Hum Dev 2012;88(Suppl 1):S5–7.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Senterre T,
      2. Rigo J
      . Reduction in postnatal cumulative nutritional deficit and improvement of growth in extremely preterm infants. Acta Paediatr 2012;101:e64–70.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Kent AL,
      2. Wright IM,
      3. Abdel-Latif ME
      . Mortality and adverse neurologic outcomes are greater in preterm male infants. Pediatrics 2012;129:124–31.
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Simon L,
      2. Borrego P,
      3. Darmaun D,
      4. et al
      . Effect of sex and gestational age on neonatal body composition. Br J Nutr 2013;109:1105–8.
      OpenUrlCrossRefPubMed
    9. ↵
      1. Gianni ML,
      2. Roggero P,
      3. Piemontese P,
      4. et al
      . Body composition in newborn infants: 5-year experience in an Italian neonatal intensive care unit. Early Hum Dev 2012;88(Suppl 1):S13–17.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Leroux BG,
      2. N'Guyen The Tich S,
      3. Branger B,
      4. et al
      . Neurological assessment of preterm infants for predicting neuromotor status at 2 years: results from the LIFT cohort. BMJ Open 2013;3. pii: e002431.
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Josse D
      . Revised Brunet-Lézine Scale of psychomotor development of first childhood [in French]. ECPA, ed. Paris, France, 2011.
    12. ↵
      1. Skellern CY,
      2. Rogers Y,
      3. O'Callaghan MJ
      . A parent-completed developmental questionnaire: follow up of ex-premature infants. J Paediatr Child Health 2001;37:125–9.
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Flamant C,
      2. Branger B,
      3. Nguyen The Tich S,
      4. et al
      . Parent-completed developmental screening in premature children: a valid tool for follow-up programs. PLoS One 2011;6:e20004.
      OpenUrlCrossRefPubMed
    14. ↵
      1. Cole TJ
      . The LMS method for constructing normalized growth standards. Eur J Clin Nutr 1990;44:45–60.
      OpenUrlPubMedWeb of Science
    15. ↵
      1. Olsen IE,
      2. Groveman SA,
      3. Lawson ML,
      4. et al
      . New intrauterine growth curves based on United States data. Pediatrics 2010;125:e214–24.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      WHO Child Growth Standards. Methods and development: length/height-for-age, weight-for-age, weight-for-height and body mass index-for age. Geneva: World Health Organization, http://www.who.int/childgrowth/standards/Technical_report.pdf (accessed 11 July 2013).
    17. ↵
      1. Rubin DB
      . Estimating causal effects from large data sets using propensity scores. Ann Intern Med 1997;127(8 Pt 2):757–63.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. D'Agostino RB Jr.
      . Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17:2265–81.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Belfort MB,
      2. Rifas-Shiman SL,
      3. Sullivan T,
      4. et al
      . Infant growth before and after term: effects on neurodevelopment in preterm infants. Pediatrics 2011;128:e899–906.
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Franz AR,
      2. Pohlandt F,
      3. Bode H,
      4. et al
      . Intrauterine, early neonatal, and postdischarge growth and neurodevelopmental outcome at 5.4 years in extremely preterm infants after intensive neonatal nutritional support. Pediatrics 2009;123:e101–9.
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Peacock JL,
      2. Marston L,
      3. Marlow N,
      4. et al
      . Neonatal and infant outcome in boys and girls born very prematurely. Pediatr Res 2012;71:305–10.
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Stark MJ,
      2. Hodyl NA,
      3. Wright IM,
      4. et al
      . Influence of sex and glucocorticoid exposure on preterm placental pro-oxidant-antioxidant balance. Placenta 2011;32:865–70.
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      1. Lucas A,
      2. Morley R,
      3. Cole TJ
      . Randomised trial of early diet in preterm babies and later intelligence quotient. BMJ 1998;317:1481–7.
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Makrides M
      . DHA supplementation during the perinatal period and neurodevelopment: Do some babies benefit more than others? Prostaglandins Leukot Essent Fatty Acids 2013;88:87–90.
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Liu Y,
      2. Metens T,
      3. Absil J,
      4. et al
      . Gender differences in language and motor-related fibers in a population of healthy preterm neonates at term-equivalent age: a diffusion tensor and probabilistic tractography study. AJNR Am J Neuroradiol 2011;32:2011–16.
      OpenUrlAbstract/FREE Full Text
    26. ↵
      1. Senterre T,
      2. Rigo J
      . Optimizing early nutritional support based on recent recommendations in VLBW infants and postnatal growth restriction. J Pediatr Gastroenterol Nutr 2011;53:536–42.
      OpenUrlCrossRefPubMed
    27. ↵
      1. Costa-Orvay JA,
      2. Figueras-Aloy J,
      3. Romera G,
      4. et al
      . The effects of varying protein and energy intakes on the growth and body composition of very low birth weight infants. Nutr J 2011;10:140.
      OpenUrlCrossRefPubMed
    28. ↵
      1. Roze JC,
      2. Darmaun D,
      3. Boquien CY,
      4. et al
      . The apparent breastfeeding paradox in very preterm infants: relationship between breast feeding, early weight gain and neurodevelopment based on results from two cohorts, EPIPAGE and LIFT. BMJ Open 2012;2:e000834.
      OpenUrlAbstract/FREE Full Text

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

      Files in this Data Supplement:

    Footnotes

    • Contributors AF-C carried out the initial analyses, drafted the initial manuscript and approved the final manuscript as submitted. LS, CF and PYA carried out the initial analyses, reviewed and revised the manuscript and approved the final manuscript as submitted. BB coordinated and supervised data collection in all sites, revised the manuscript and approved the final manuscript as submitted. GG supervised data collection in one site, reviewed and revised the manuscript and approved the final manuscript as submitted. DD conceptualised and designed the study, drafted the initial manuscript and approved the final manuscript as submitted. JCR conceptualised and designed the study, carried out the initial analyses, drafted the initial manuscript and approved the final manuscript as submitted.

    • Funding The LIFT cohort is supported by grants from the Regional Health Agency of Pays de la Loire with a grant of 350 000 euros each year.

    • Competing interests None.

    • Ethics approval The cohort was registered at the French CNIL (Commission Nationale de l'Informatique et des Libertés no 851117) ethics committee.

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

    Linked Articles

    • Fantoms
      Highlights from this issue
      Martin Ward Platt
      Archives of Disease in Childhood - Fetal and Neonatal Edition 2014; 99 F347-F347 Published Online First: 12 Aug 2014. doi: 10.1136/archdischild-2014-307225