High-Energy and -Protein Diet Increases Brain and Corticospinal Tract Growth

DOI: 10.1542/peds.2007-1267 2008;121;148Pediatrics

and Janet A. EyreLyvia Dabydeen, Julian E. Thomas, Tessa J. Aston, Hilary Hartley, Sunil K. Sinha

in Term and Preterm Infants After Perinatal Brain InjuryHigh-Energy and -Protein Diet Increases Brain and Corticospinal Tract Growth

http://pediatrics.aappublications.org/content/121/1/148.full.htmllocated on the World Wide Web at:

The online version of this article, along with updated information and services, is

of Pediatrics. All rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Boulevard, Elk Grove Village, Illinois, 60007. Copyright 2008 by the American Academy published, and trademarked by the American Academy of Pediatrics, 141 Northwest Pointpublication, it has been published continuously since 1948. PEDIATRICS is owned, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

at Indonesia:AAP Sponsored on March 7, 2014pediatrics.aappublications.orgDownloaded from at Indonesia:AAP Sponsored on March 7, 2014pediatrics.aappublications.orgDownloaded from

ARTICLE

High-Energy and -Protein Diet Increases Brain andCorticospinal Tract Growth in Term and PretermInfants After Perinatal Brain InjuryLyvia Dabydeen, MB, BSa, Julian E. Thomas, MDa, Tessa J. Aston, MSca, Hilary Hartley, MSca, Sunil K. Sinha, MD, PhDb,

Janet A. Eyre, MBChB, DPhila

aDevelopmental Neuroscience, School of Clinical Medical Sciences (Child Health), University of Newcastle Upon Tyne, Newcastle Upon Tyne, United Kingdom;bDepartment of Paediatrics and Neonatology, James Cook University Hospital, Middlesbrough, United Kingdom

The authors have indicated they have no nancial relationships relevant to this article to disclose.

ABSTRACT

OBJECTIVE.Our hypothesis was that infants with perinatal brain injury fail to thrive inthe first postnatal year because of increased energy and protein requirements fromdeficits that accumulated during neonatal intensive care. Our aim was to assesswhether dietary energy and protein input was a rate-limiting factor in brain and bodygrowth in the first year after birth.

METHODS.We conducted a prospective, double-blind and randomized, 2-stage groupsequential study and controlled for gestation, gender, and brain lesion. Neonateswith perinatal brain damage were randomly allocated to receive either a high-(120% recommended average intake) or average (100% recommended averageintake) energy and protein diet. The study began at term and continued for 12months. Three-day dietary diaries estimated energy and protein intake. The primaryoutcome measure was growth of occipitofrontal circumference. Other measureswere growth of axonal diameters in the corticospinal tract, which were estimated byusing transcranial magnetic stimulation, weight gain, and length.

RESULTS. The study was terminated at the first analysis when the 16 subjects hadcompleted the protocol, because the predetermined stopping criterion of 1 SDdifference in occipitofrontal circumference at 12 months corrected age in thosereceiving the higher-energy and -protein diet had been demonstrated. Axonal di-ameters in the corticospinal tract, length, and weight were also significantly in-creased.

CONCLUSIONS. These data support our hypothesis that infants with significant perinatal brain damage have increasednutritional requirements in the first postnatal year and suggest that decreased postnatal brain growth may exacerbatetheir impairment. There are no measures of cognitive ability at 12 months of age, and whether there will be anyimprovement in the status of these children, therefore, remains to be shown.

INFANTS WITH SIGNIFICANT brain injury commonly suffer from growth faltering. Although nonnutritional factorsrelated to neurologic pathophysiology will have an impact on growth,1 the pattern of their early growth failure istypical of chronic undernutrition, where body mass is lost before length and brain growth is compromised, suggestingthat the early nutritional needs of these infants are not being met.2,3 The growth faltering begins very early, beforethe development of abnormal neurologic signs; thus, dysphagia is unlikely to be a major factor initially.2,4 It is nowappreciated that critically ill neonates accumulate deficits in energy and protein during intensive care, which are notrecovered by the time of discharge.58 For preterm and term infants, both the accumulated total energy and proteindeficits predict the degree of growth faltering during the acute hospital admission.7 Furthermore, it is increasinglyappreciated that, to achieve appropriate growth rates after discharge, the dietary intake of these infants must beincreased above recommended average requirements9 to meet not only their needs for normal maintenance andgrowth but also that required to catch up the energy and protein deficits.10 Our hypothesis is that, in infants who

www.pediatrics.org/cgi/doi/10.1542/peds.2007-1267

doi:10.1542/peds.2007-1267

KeyWordsnutrition, brain growth, corticospinal tract,perinatal brain injury, neonatalencephalopathy, white matter injury,human, randomized, double-blinded

AbbreviationsEARestimated average requirementOFCoccipitofrontal circumferenceTMStranscranial magnetic stimulationCMCDcentral motor conduction delay

Accepted for publication Jul 3, 2007

Address correspondence to Janet A. Eyre,MBChB, DPhil, Sir James Spence Institute ofChild Health, Royal Victoria Inrmary, QueenVictoria Road, Newcastle Upon Tyne NE1 4LP,United Kingdom. E-mail: [email protected]

PEDIATRICS (ISSN Numbers: Print, 0031-4005;Online, 1098-4275). Copyright 2008 by theAmerican Academy of Pediatrics

148 DABYDEEN et al at Indonesia:AAP Sponsored on March 7, 2014pediatrics.aappublications.orgDownloaded from

suffered significant perinatal brain injury, failure to meettheir increased energy and protein requirements ac-quired during the acute illness contributes significantlyto the growth faltering that occurs in the first 6 to 12months after discharge from hospital. Thus, a compo-nent of the early growth faltering arises from relativeundernutrition and may be preventable.

It has long been recognized that the most strikingconsequence of undernutrition during the first 6 to 12months after birth is permanently reduced brain size,1117

associated with a thinner cerebral cortex,18 diminishednumber of neurons,19 reduced myelination,20 poor den-dritic arborization, and changes in the microscopic fea-tures of dendritic spines, such as a reduction in theirwidth and number.21,22 Numerous studies have docu-mented the relationship between subnormal headgrowth and such adverse neurodevelopmental outcomesas decreased perceptual motor skills, general cognitiveability, language, academic achievement, adaptive be-havior, and higher parental ratings of attention prob-lems. When it has been tested, associations betweensubnormal head circumference and adverse develop-mental outcomes remain significant despite controllingfor sociodemographic and neonatal risk factors and formajor neurosensory impairment.2333 Early postnatalbrain growth seems to be the most sensitive period forlater IQ. In children born at term, IQ scores at 8 years arehighest in children whose heads grew most during thefirst year, even after adjusting for confounders.34,35 Headgrowth after infancy is not associated with later IQ scoresand does not compensate for poorer growth in the firstyear of life.34 Findings from studies of very low birthweight infants also suggest that the critical period forcatch-up brain growth, in terms of later intelligence,may be confined to the first year of life.28,32

If our hypothesis is correct, failure in the first yearafter birth to meet the additional nutritional require-ments of children who have suffered acute perinatalbrain injury is likely to not only compromise their over-all growth but also growth of the brain, thereby com-pounding their impairment. The aim of our study was toassess whether a high-energy and -protein diet wouldlead to significantly greater brain and body growth in thefirst postnatal year for infants who suffered significantperinatal brain injury.

METHODSWe undertook a prospective, randomized, and double-blinded comparison of the growth of the brain and over-all body growth in the first 12 months after term ininfants with acute perinatal brain injury fed either a dietthat met recommended estimated average requirements(EARs) for energy (average-energy group) or a high-energy diet with a target energy input of 120% EAR(high-energy group).9 For both groups, the target fortheir protein/energy ratio was 2.5 g/420 kJ (100 kcal) to3.6 g/420 kJ (100 kcal), as recommended by the expertpanel for the American Society for Nutritional Sciences.36

Ethical approval was obtained according to the Declarationof Helsinki from the ethical committees of the participatingcenters, as was written informed consent from the par-

ent(s). To achieve double-blinding, only the pediatric nu-tritional team composed of a consultant specializing ingastroenterology and childhood nutrition (Dr Thomas) andpediatric dieticians (Ms Aston and Ms Hartley) was awareof subject allocation. The remainder of the research teamand the families were blinded to subject allocations, whichwere not revealed to the investigators until after the prin-cipal data analyses were performed.

Subject RecruitmentSubjects were recruited by the research associate (DrDabydeen) while inpatients in 1 of the 4 level 3 neonatalintensive care nurseries in north east England. Subjectswere allocated to treatment groups by minimization, amethod of ensuring excellent balance between groupsfor several prognostic factors, even in small samples.With minimization, the group allocated to the next en-rolled participant depends on the characteristics of thoseparticipants already enrolled. The aim is that each allo-cation should minimize the imbalance across multiplefactors.37 If the parent(s) gave consent for inclusion, theinfants were allocated to be fed to a target nutritionalinput of either 100% or 120% of the estimated averageenergy requirement for age and birth weight centile.9 DrThomas was responsible for minimization, which wascomputer generated and controlled for 3 prognostic fac-tors: gestation (32 or 32 weeks), gender, and brainlesion.

There were 2 inclusion criteria: severe neonatal en-cephalopathy38 and/or gestation of 32 weeks withwhite matter disease.39 Subjects were excluded if theyhad congenital malformations, chromosomal abnormal-ities, or significant chronic illnesses (ie, pulmonary, car-diac, renal, or gastrointestinal) or had taken medicationaffecting growth and, therefore, would be expected tohave atypical postdischarge growth.

Children with severe neonatal encephalopathy wereidentified clinically based on their history, electroen-cephalogram findings, and clinical signs.38 To identifysubjects with white matter disease, all of the infants bornat 32 weeks of gestation had ultrasound scans per-formed by a consultant radiologist using a 7.5-MHztransducer at postnatal days 1 to 3 and 7 to 10 and at3weeks after birth. White matter disease was defined asevidence of multiple, bilateral echolucencies, character-istic of cystic periventricular leukomalacia and/or intra-ventricular hemorrhage with parenchyma echodensitiesor lucencies consistent with parenchymal infarctionand/or nonprogressive ventricular enlargement, definedas 1 lateral ventricle greater than the 99th percentile,without an increased rate of head growth.39

NutritionThe target nutritional energy and protein inputs werecomputed throughout the 12 months according to theinfants age and birth weight percentile.9 For weightreference standards, the revised United Kingdom 1990reference data (version 1996/1) were used.40 The parentsand the dietician agreed on individualized feeding plansbased on the childs target energy and protein input and

PEDIATRICS Volume 121, Number 1, January 2008 149 at Indonesia:AAP Sponsored on March 7, 2014pediatrics.aappublications.orgDownloaded from

ensuring a balanced intake of vitamins and minerals. Adietician (Ms Aston or Ms Hartley) contacted the familyweekly and visited the family in the home as required toensure that these targets were being achieved. The strat-egies used to achieve the targets included increasing feedvolumes, altering food texture and thickness, increasingparental feeding skill, and correcting the feeding positionof the infant. If these measures failed, energy and pro-tein supplementation of feeds was introduced. Ethicalapproval did not allow for invasive interventions, suchas gastrostomy feeding, and such interventions re-mained the decision of the clinical team involved witheach childs care. The difference between target andintake was monitored weekly, using parental 24-hourdietary recall. On the basis of these data, feeding planswere continuously adjusted. A formal 3-day, prospec-tively collected food diary was used to estimate nutri-tional intake near term and at 3-monthly intervals. Im-mediately before each diary, a dietician visited the hometo provide training in the estimation of the volumes andthe description of the food and fluids consumed andwasted. During the visit the dietician observed a feed toconfirm the accuracy of the estimations. The parentswere provided with record sheets and chose 3 days whenthe children were eating their usual diet. For childrenfed infant formula, the volume consumed was recorded.Two infants (1 in each intervention group; Table 1) werepartially breastfed, but breastfeeds were not included inthe target intake or in the food diary. For those taking amixed diet, the parents also gave a full description of thefoods offered, including keeping food labels. For home-prepared food, parents provided recipes and describedthe cooking methods. Immediately after completion ofeach diary, the dietician revisited the home to reviewand clarify the record. The forms were then coded by thedieticians, and the daily intakes of energy, protein, andnutrients were computed using a food database (Micro-diet, Downlee Systems Limited, High Peak, United King-dom [www.microdiet.co.uk]).

DeprivationTo look for possible socioeconomic differences betweenthe groups that might confound the findings, the depri-

vation rating of the subjects was determined using theTownsend Deprivation Scale and the ward in which theywere resident. The Townsend Deprivation Scale is par-ticularly suitable for our study, because it is based ondata from the north of England and provides an index ofmaterial deprivation for all 678 wards in which aresubjects could have been resident, derived from 4 vari-ables: unemployment, car ownership, housing tenure,and household overcrowding.41

OutcomeMeasuresAll of the outcome measurements were made by DrDabydeen and Dr Thomas, who were blind to subjectallocation. Measurements were made at baseline (term)and final measurements at 12 months; intermediatemeasurements were also made to provide informationon the pattern of growth in the first 12 months. Headcircumference and weight were, therefore, also mea-sured at 3 monthly intervals (Fig 3 A and B), corticospi-nal tract axonal diameter was also estimated at 4 and 8months (Fig 3C), and length was also measured at 6months. SD (z) scores for anthropometric measures, de-rived from the British 1990 growth reference, which wasrevised in September 1996, were used so that age andgender data could be combined.40

Weight was measured to the nearest 10 g, with thechild unclothed, by using a portable digital electronicscale. Length was measured using a horizontal stadiom-eter accurate to 1 cm. For both weight and length, 3measurements were made, and the mean was calcu-lated.

Brain GrowthTwo measures for brain growth were used. The first wasoccipitofrontal circumference (OFC), because it is a val-idated indicator of brain volume, weight, and DNA con-tent in newborns, children, and adults.17,4245 The secondwas axon diameter growth in the corticospinal tract. Thiswas chosen because it can be measured noninvasivelywith transcranial magnetic stimulation (TMS),46 andaxon diameter growth is a marker for growth of thepyramidal neuron as a whole, because there is a positive

TABLE 1 Characteristics of Subjects

Group

High Energy Average Energy No Consent

No. 8 8 19Deprivation index mean (/ 95% condence limits) 1.9 (/ 1.1) 1.3 (/ 1.1) 2.0 (/0.7)

Range 3.40 to 5.36 2.20 to 6.04 3.39 to 5.59Male, n (%) 4 (50) 5 (63) 10 (53)Partially breastfed, n (%) 1 (13) 1 (13) 2 (11)Died during the study, n (%) 0 0 2 (11)White matter damage, n (%) 5 (63) 6 (75) 13 (68)Cystic periventricular leukomalacia, n (%) 2 (25) 3 (38) 5 (26)Intraventricular hemorrhage with parenchymal

echodensities/lucencies, n (%)3 (38) 3 (38) 6 (32)

Nonprogressive ventricular enlargement, n (%) 0 0 2 (10)Severe neonatal encephalopathy, n (%) 3 (38) 2 (25) 6 (32)


linear correlation between axonal diameter and somasize and the horizontal spread of the dendritic tree inlayer 5 pyramidal neurons of the motor cortex.4750

Occipitofrontal CircumferenceOFC was measured by using a flexible, nonstretchabletape scaled to 1 mm. The tape was placed superior to thesupraorbital ridge and adjusted around the occiput untila maximum circumference was obtained from 3 mea-surements.

Corticospinal Axonal DiameterTMS (MagStim Company Ltd, Whitland, Wales) wasused to estimate the conduction delay within the corti-cospinal tract following previously published methods.46

A figure-8 coil, with each circle having a diameter of 55mm (SPC-ENG 8618; MagStim Company Ltd), was usedto excite corticospinal neurons. TMS was applied duringthe spontaneous contraction of biceps. Electromyogramwas recorded bilaterally from biceps using miniaturized,skin-mounted differential amplifiers. A 3-dB bandpassof 5 to 1500 Hz was applied, and the signals were sam-pled at 5 KHz and stored on computer. The onset latencyof the motor-evoked potentials in biceps was defined aswhen the electromyogram of biceps clearly deviated byeye from background activity. Total motor conductiondelay was estimated from the shortest onset latency of20 motor-evoked potentials at a stimulation intensity of1.2 times the threshold or at the maximum stimulatoroutput. Magnetic stimulation over the C5 vertebra ex-cited spinal motor roots. The longest onset latency of 20responses in biceps estimated peripheral motor conduc-tion delays. Subtraction of peripheral from total motorconduction delays estimated central motor conductiondelays (CMCDs).

The corticospinal pathway length to C5 was estimatedfrom the distance from the vertex to vertebra promi-nens, which we have demonstrated previously to be 1.3times the corticospinal pathway length.51 The maximumconduction velocity of corticospinal axons projecting toC5 was calculated by dividing this distance by the con-duction delay of corticospinal axons projecting to bicepspinal motoneurons (CMCD for biceps minus 1 millisec-ond for spinal transsynaptic delay).46 The maximum di-ameter was then determined using the ratio between theconduction velocity of myelinated corticospinal axonsand their diameters of 5.2 m seconds1 m1, derivedby using invasive measurements in subhuman primates,including developing primates.52

Statistical AnalysisThe study was designed to test the 1-sided hypothesisthat brain growth for those fed the higher-energy dietwould be greater than that of those fed the average-energy diet.53 We decided a clinically significant effectwould be a 0.5-SD increase in OFC. There was evidence,however, that there might be a more substantial effect,because additional nutrition given early in developmentto preterm infants increased the OFC by 1 SD andreduced the incidence of cerebral palsy at the age of 7 to

8 years by ninefold.54,55 Therefore, for ethical consider-ations, a 2-stage, 1-sided, group-sequential design wasadopted with a prespecified stopping criterion of a1 SDincrease in head circumference.53 The first-stage analysiswas specified to occur when 8 subjects had been re-cruited to each group, giving an 80% power at the .05level of detecting a 1-SD increase in OFC at 12 monthscorrected age. If the study then continued, the finalanalysis was specified to take place when 32 subjects hadbeen recruited into each group, giving an 80% power ofdetecting a 0.5-SD increase at the .05 level. The studydesign allowed for the possibility that our first-stageanalysis may produce significant results; thus, minimi-zation was chosen as the most suitable tool to achieve abalance of critical prognostic variables between smallgroups. The only statistical comparisons made werebased on the a priori hypotheses. The 2 groups werecompared by analysis of covariance, examining baselinecorrected data, with birth weight z score included as acovariant to control for the effect of extreme outliers.56

Data from children who were eligible to participate butwhose parents refused consent have been included forcomparison in the graphs but were not included in thestatistical analyses.

RESULTSThe study was stopped at the first-stage analysis becausethe prespecified stopping criterion of a 1-SD increasein OFC at 12 months of age had been demonstrated inthe high-energy group compared with the average-en-ergy group (Fig 3A).

Characteristics of the SubjectsForty-three infants were considered for inclusion. Eightwere excluded because of chronic lung disease. The par-ents of 35 infants were approached for consent, of whom16 gave consent; 5 were term infants (birth weight zscore: mean: 0.07; median: 0.08; range: 1.52 to 1.67) and 11 were preterm infants (gestation: mean: 28weeks; median: 28 weeks; range: 2331 weeks; birthweight z score: mean: 0.27; median: 0.04; range:1.37 to 0.91). The parents of 19 declined, and theseinfants formed the no-consent group. Six were term(birth weight z score: mean: 0.38; median: 0.5;range:1.59 to 1.57) and 13 were preterm (gestation:mean: 28 weeks; median: 27 weeks; range: 2431weeks; birth weight z score: mean: 0.52; median: 0.50;range:1.59 to 1.57). Mortality in the first year was 6%,representing 2 infants, both from the no-consent group.

All 16 of the subjects recruited completed the studyprotocol, and all were included in the analysis. The 19who declined consent agreed to weight and OFC databeing collected, and 3 also consented to TMS studies.

The characteristics of recruited subjects by group al-location (high-energy group and average-energy group)and those whose parents declined consent (no-consentgroup) are summarized in Table 1 and Fig 1A. There wasno significant difference between the groups on the levelof deprivation (Table 1; P .64). There were no signif-icant differences in gestational age at birth (Fig 1A; P


.53), the number of days the infants received assistedventilation while in intensive care (mean 95% confi-dence limits: high-energy group: 13.6 3.5; average-energy group: 9.12 1.70; P .62), or in baseline entryanthropometric measures between the groups (Table 1;Fig 1A; birth weight: P .98; baseline weight: P .85;birth OFC: P .33; baseline OFC: P .27). The weightz scores were significantly lower at discharge from thehospital compared with that at birth for both groups (Fig1B; paired t test: high-energy group: P .047; average-energy group: P .01).

Estimated Nutritional IntakeAll of the children were fed orally, and none had agastrostomy inserted during the period of the study.Figure 2 shows the estimated energy intake and theprotein/energy ratios achieved. The mean energy intakefor the average-energy group remained close to the tar-get of 100% EAR. The mean energy intake of the high-energy group was also close to the target (mean: 119%)for the first 6 months. It then fell progressively to a meanof 101% EAR by 12 months corrected age (Fig 2A). Themean protein/energy ratios remained within our targetrange of 2.5 g/420 kJ (100 kcal) to 3.6 g/420 kJ (100kcal) throughout the first 12 months (Fig 2B).

Occipitofrontal CircumferenceThe high-energy group had significantly greater headcircumference z scores at 12 months (Fig 3A). All 3 of thegroups showed an initial drop in the OFC z scores in thefirst 6 months. Thereafter, the high-energy groupshowed an increase in OFC z scores, whereas the aver-age-energy group showed a continuing decline. The no-

consent group showed the most rapid decline in OFC zscores.

WeightThe z scores for weight for the high-energy group weregreater throughout the study than those of the average-energy group (Fig 3A). The differences were significant at 3months and 6 months. The no-consent group had thelowest-weight z scores throughout the study (Fig 3B).

LengthThe high-energy group maintained a normal length (meanz score 95% confidence limits: 6 months:0.15 0.55;12 months: 0.31 0.58), whereas the average-energygroup showed faltering in linear growth (mean z score 95% confidence limits: 6 months: 1.34 0.52; 12months: 0.98 0.0.60). The differences between thehigh-energy group and the average-energy group weresignificant (6 months: P .019; 12 months: P .04). Nomeasures of length were made in the no-consent group.

Corticospinal Axonal DiameterAll 3 of the groups had similar maximum axonal diam-eters near term (Fig 3C). The high-energy group showedthe greatest rate of growth so that at 7.5 and 12 monthscorrected age, their axonal diameters are significantlylarger than those in the average-energy group. The 3children studied in the no-consent group showed littleincrease in axonal diameter.

DISCUSSIONThis is the first double-blinded, randomized, and prospec-tive study to assess the effect of dietary supplementation onthe growth of human infants who have been critically ill inthe neonatal period and suffered parenchymal brain in-jury. A previous study of supplemental nutrition in un-selected premature infants54,55 and animal experiments hadsuggested that the effect might be large, and indeed it was,with a 1-SD increase in head size and corticospinal ax-onal diameter at age 1 year with significantly greaterweight and length gains also observed in the group fed ahigh-energy and -protein diet compared with those fed anaverage-energy and -protein diet.

FIGURE 1A, Gestational age, weight, andOFC at birth and at baseline for the 2 intervention groups.The data are graphed as mean and 95% CLs for themean. B, Comparison of the weight zscores at birth and at baseline for the 2 intervention groups. Circles indicate the high-energy group; squares, average-energy group; Triangles, no-consent group.

FIGURE 2Estimated energy (A) and protein intake (B) (1 kcal 4.2 kJ) for the 2 study groups. Intakeis expressed as the percentage of the EARs for age and birthweight centile of the subject.Data are graphed as mean and 95% CLs for the mean. Filled circles indicate the high-energy group; lled squares, average-energy group.


Both groups had significantly lower weight z scores ondischarge from the hospital than at birth and so had a realneed for catch-up growth in the first year. Despite theestimated mean energy intake being maintained at orgreater than their estimated average energy requirementsfor age and birth weight percentile, rather than showingcatch-up growth, the children in both experimental groupsshowed progressive weight faltering when their energyintake was close to 100% EAR (Figs 2 and 3; from birth inthe average-energy group and from 6 months in the high-energy group). This supports our hypothesis that thesechildren required a greater-than-average energy and pro-tein intake just to achieve appropriate growth rates, letalone catch-up growth in the first year.

It is possible that energy and protein intake were signif-icantly overestimated by parents; however, we believe thisis unlikely, because before weaning, the parents were sim-ply required to record the volume of feed consumed. In thelater 6 months, after the introduction of solid food, weminimized the possibility of overestimation by carefultraining and by observing feeds to confirm parental esti-mates at the start of each 3-day diary.

For both groups, weight z scores were significantlylower at discharge from the hospital compared with atbirth, providing strong evidence for significant energyand protein deficits accumulated during their acute ill-nesses (Fig 1B).7 It is conceivable also that repair of acutebrain injury requires additional energy over and abovethat needed for normal brain growth. As far as we areaware, there have been no studies either in humans orin animal models that have addressed this issue. Finally,the infants recovering from acute brain injury may havedysregulation of central energy homeostasis. There issome evidence to support this in that both term andpreterm infants without brain damage, when offeredcalorically dense feeds, consume lower volumes thanthose offered less energy-dense feeds. Thus, there is littledifference in the overall energy intake, implying that theneuroendocrine control of energy intake is mature be-fore term.5760 In contrast, the infants in our high-energygroup maintained an increased energy input for the first6 months despite being fed calorically dense feeds (Fig2A), suggesting they had impairment of, or delay in, thematuration of energy homeostasis.

The observed growth benefit in those fed the high-energy and -protein diet may have resulted from eitherincreased energy or protein intakes (Fig 2).61,62 It is aca-demic to try and argue for or against either, because pro-tein and energy needs are reciprocally limiting. If energyintake is insufficient, protein is used as an energy source,and the nitrogen balance becomes less positive. Increasingthe caloric intake will spare the protein loss and improvenitrogen retention, but with limited protein intake, theprotein retention reaches a plateau, and the energy excessis used for only fat deposition.8,36 It was for these reasonsthat our target protein/energy ratio in both our experimen-tal groups was between 2.5 g/420 kJ (100 kcal) and 3.6g/420 kJ (100 kcal), as recommended by the expert panelof the American Society for Nutritional Sciences.36 In-creased intakes of other dietary constituents, such as zinc,calcium, phosphorus, and vitamins, may also have contrib-uted.8,63 However, the intakes of vitamins, minerals, andessential fatty acids for those in both intervention groupsfar exceeded reference nutrient intake norms. These factorsare unlikely, therefore, to be rate limiting when comparinggrowth between the 2 intervention groups but may wellhave been important factors when considering the failureto thrive observed in the no-consent group when com-pared with both intervention groups.

It is likely that the support and education provided tothe family by a dietary therapist going regularly into thehome also has a beneficial effect. This does not, how-ever, explain the difference between the 2 interventiongroups, because there were no significant differencesbetween the groups in the hours of therapy time (me-

FIGURE 3OFC (A), weight (B), andmaximum axonal diameter in the corticospinal projection to themotoneurons of biceps (C) in the 2 intervention groups. Filled circles indicate the high-energy group; lled squares, average-energy group; open triangles, no-consent group.The numbers above each graph are the P values for the comparison between the 2intervention groups at each time point. Weight and OFC are expressed as z scores cor-rected for baseline at term. Data are graphed as mean and 95% CLs for the mean exceptfor the axonal diameters for the no-consent group, which are individual values for the 3subjects joined by a dotted line. The hashed line indicates the mean values for axondiameter obtained in our previous studies of normal subjects by using TMS. The starsrepresent data obtained by direct postmortemmeasurement obtained at the level of thepyramid in a neurologically normal subject at term and at 4 and 8 months (reported byVerhaart77).


dian [range]: contacts: high-energy group: 38 h [1480h]; average-energy group: 28 h [1078 h]; duration ofeach home visit: high-energy group: 0.97 h [0.72]1.23h]; average-energy group: 1.01 h [0.841.22 h).

The growth of both our average- and high-energygroups was better than that described in 2 previousstudies of the early growth of similar children with peri-natal brain injury. In contrast, the pattern of growth ofthe 19 infants in our no-consent group was very similar,with mean weight z scores falling to 2 by 12 months ofage (Fig 3).2,64 It is noteworthy that, despite severe fail-ure to thrive, none of the children in the no-consentgroup were referred by their clinicians to a specialistnutritional service during the period of the study or hadgastrostomies placed. We hypothesize that the progres-sive onset of feeding difficulties compromised the intakeof the children in the no-consent group and that theirintake increasingly did not even meet the average rec-ommended energy intake, as has been described by Sul-livan et al65 in older children with cerebral palsy.

There was a striking difference in the rate of headcircumference growth between the 2 study groups, withthe high-energy group having significantly greater headcircumferences at 12 months than the average-energygroup. The no-consent group showed the greatest falter-ing of head circumference growth. None of the 3 groupsmaintained their birth or enrollment z score for OFC;this is not surprising, because all either had ultrasoundevidence of white matter loss or had suffered a verysevere encephalopathy, likely to lead to neuronal loss.Head circumference is an excellent predictor of brainvolume, weight, and DNA content.17,4245 Children with-out brain damage who die during the first year of lifewith severe undernutrition have significantly reducedOFCs, total brain weight, and RNA and DNA content.1417

Postnatal catch-up growth in OFC in small-for-gesta-tional-age infants only occurs if adequate nutrition isachieved during the first year.29,66,67 The children in theaverage-energy group and the no-consent group arelikely, therefore, to have permanent reductions in brainvolume, weight, and cell number in comparison withthe high-energy group.

TMS revealed nearly normal growth of corticospinalaxonal diameters in the high-energy group (Fig3C),whereas it was significantly reduced in the average-energy group. Disturbingly, the 3 children studied fromthe no-consent group demonstrated almost no growth inmaximum axonal diameter. Prolonged CMCDs havebeen reported previously in undernourished children,consistent with decreased axonal conduction velocitiesand diameters,68 and, as in the present study, the degreeof prolongation was related to the severity of growthfaltering. Undernourishment during early developmentin the rat leads to permanent reductions in corticospinaltract axonal diameters, implying that the reduced axonaldiameters observed in our study at 12 months of agemay persist into adulthood.69,70 Because axonal diametergrowth is a marker for growth of the neuron as a whole,these data imply decreased neuronal growth in the av-erage-energy group and the no-consent group relative tothe high-energy group.4750 Consistent with our findings,

decreased soma size, dendritic arborization of corticalpyramidal neurons, dendritic spine number, and syn-apse/neuron ratio have been found after undernourish-ment during development in animals.71 Similar changesare observed in histopathological studies of the pyrami-dal neurons of the motor cortex in children who die afterundernutrition in the first year after birth.21,22 Thus, wepropose that the reduced axonal diameters of the sub-jects in the average-energy group and the no-consentgroup are markers for decreased pyramidal neuron somasize, dendritic arborization, and synapse number andindicate that undernutrition during the first 12 monthsleads to exacerbation of the original neurologic deficit.

CONCLUSIONSAn implicit assumption by many clinical caregivers isthat abnormalities in growth and body composition ininfants with significant brain injury are because of un-alterable aspects of the disease process and, thus, evenvery severe failure to thrive is tolerated, as is clearlydemonstrated in our no-consent group. Our randomizedand double-blinded study establishes that a componentof their growth failure in the first year is preventablewith early intervention by a skilled nutritional team. Thelong-term benefits of increased nutrition and increasedbrain and body growth for children with significant peri-natal brain injury are unknown, because previous stud-ies of postnatal nutrition and neurodevelopmental out-come have predominantly excluded such infants.However, randomized intervention studies investigatingthe benefits of nutritional supplementation for childrenwho are undernourished or at risk of undernourishmenthave demonstrated long-term benefits for both motordevelopment and academic achievement if the interven-tion begins before 12 months of age.7275 All of theinfants in our study had very significant brain damage,and the majority are likely to have significant neuro-logic sequelae.61,76 Neurodevelopmental tests are notsensitive enough to distinguish between degrees ofseverity of impairment in infancy and early childhood.We have, therefore, elected to wait and perform de-tailed reassessments of the children at the age of 8years when we can perform MRI scans without theneed for a general anesthetic and can assess theirmotor and cognitive outcomes in detail with appro-priately sensitive tests. Nonetheless, the benefits ofsupplemental nutrition in terms of body and braingrowth indicate that additional studies are required,directly measuring the energy and protein balance ofhigh-risk infants with parenchymal brain injury todefine an optimal diet that meets their nutritionalneeds during this critical period of brain growth.

ACKNOWLEDGMENTSFunding for this study was obtained from the NewcastleHealth Care Charity and the Wellcome Trust.

We thank the parents who willingly consented totheir childs involvement in the study and the consult-ants at the NICUs who agreed for their patients to beincluded in the study.


REFERENCES1. Stevenson R, Roberts C, Vogtle L. The effects of non-nutritional

factors on growth in cerebral palsy. Dev Med Child Neurol. 1995;37:124130

2. Rogers B, Andrus J, Msall M, et al. Growth of preterm infantswith cystic periventricular leukomalacia. Dev Med Child Neurol.1998;40:580586

3. Reilly S, Skuse D. Characteristics and management of feedingproblems of young children with cerebral palsy. Dev Med ChildNeurol. 1992;34:379388

4. Bertino E, Coscia A, Mombro M, et al. Postnatal weight in-crease and growth velocity of very low birthweight infants.Arch Dis Child Fetal Neonatal Ed. 2006;91:349356

5. Embleton N, Pang N, Cooke R. Postnatal malnutrition andgrowth retardation: an inevitable consequence of current rec-ommendations in preterm infants? Pediatrics. 2001;107:270273

6. Hulst J, Joosten K, Zimmermann L, et al. Malnutrition incritically ill children: from admission to 6 months after dis-charge. Clin Nutr. 2004;23:223232

7. Hulst J, van Goudoever J, Zimmermann L, et al. The effect ofcumulative energy and protein deficiency on anthropometricparameters in a pediatric ICU population. Clin Nutr. 2004;23:13811389

8. Rigo J, Senterre J. Nutritional needs of premature infants:current issues. J Pediatr. 2006;149:S80S88

9. Dietary reference values for food energy and nutrients for theUnited Kingdom. Report of the Panel on Dietary ReferenceValues of the Committee on Medical Aspects of Food Policy.Rep Health Soc Subj (Lond). 1991;41:1210

10. Carver J. Nutrition for preterm infants after hospital discharge.Adv Pediatr. 2005;52:2347

11. Galler J, Shumsky J, Morgane P. Malnutrition and brain de-velopment. In: Allan Walker W, Watkins J, eds. PaediatricNutrition. New York, NY: Decker; 1996:196212

12. Stoch M, Smythe P. 15-year developmental study on effects ofsevere undernutrition during infancy on subsequent physicalgrowth and intellectual functioning. Arch Dis Child. 1976;51:327336

13. Stoch M, Smythe P, Moodie A, Bradshaw D. Psychosocialoutcome and CT findings after gross undernutrition duringinfancy: a 20 year developmental study. Dev Med Child Neurol.1982;24:419436

14. Dobbing J. Undernutrition and the developing brain. Arch DisChild. 1970;120:411415

15. Winick M. Malnutrition and brain development. J Pediatr.1969;74:667679

16. Winick M, Rosso P. The effect of severe malnutrition on cel-lular growth of the human brain. Pediatr Res. 1969;3:181184

17. Winick M, Rosso P. Head circumference and cellular growth ofthe brain in normal and marasmic children. J Pediatr. 1969;74:774779

18. Dobbing J, Sands J. Vulnerability of developing brain. IX. Theeffect of nutritional growth retardation on the timing of the braingrowth-spurt. Biol Neonate. 1971;19:363378

19. Dobbing J, Hopewell J, Lynch A. Vulnerability of developingbrain. VII. Permanent deficit of neurons in cerebral and cerebellarcortex following early mild undernutrition. Exp Neurol. 1971;32:439447

20. Krigman M, Hogan E. Undernutrition in the developing rat:effect upon myelination. Brain Res. 1976;107:239255

21. Benitez-Bribiesca L, De la Rosa-Alvarez I, Mansilla-Olivares A.Dendritic spine pathology in infants with severe protein-calorie malnutrition. Pediatrics. 1999;104(2). Available at:www.pediatrics.org/cgi/content/full/104/2/e21

22. Cordero M, DAcuna E, Benveniste S, Prado R, Nunez J, Co-lombo M. Dendritic development in the neocortex of infants

with early postnatal undernutrition. Pediatr Neurol. 1993;9:457464

23. Pryor H, Thelander H. Abnormally small head size and intellectin children. J Pediatr. 1968;73:593598

24. Nelson K, Deutschberger J. Head size at one year as a predictorof four-year IQ. Dev Med Child Neurol. 1970;12:487495

25. Dolk H. The predictive value of microcephaly during the firstyear of life for mental retardation at seven years. Dev Med ChildNeurol. 1991;33:974983

26. Ounsted M, Moar V, Scott A. Head circumference and devel-opmental ability at the age of seven years. Acta Paediatr Scand.1988;77:374379

27. Gross S, Oehler J, Eckerman C. Head growth and developmen-tal outcome in very-low-birth-weight infants. Pediatrics. 1983;71:7075

28. Hack M, Breslau N. Very low birth weight infants: effects ofbrain growth during infancy on intelligence quotient at 3 yearsof age. Pediatrics. 1986;77:196202

29. Hack M, Breslau N, Weissman B, Aram D, Klein N, BoraweskiM. Effect of very low birth weight and subnornal head size oncognitive abilities at school age. N Engl J Med. 1991;325:231237

30. Kitchen W, Doyle L, Ford G, Callanan C, Rickards A, Kelly E.Very low birth weight and growth to age 8 years: II. Headdimensions and intelligence. Am J Dis Child. 1992;146:4650

31. Cooke R, Foulder-Hughes L. Growth impairment in the verypreterm and cognitive and motor performance at 7 years. ArchDis Child. 2003;88:482487

32. Stathis S, OCallaghan M, Harvey J, Rogers Y. Head circumfer-ence in ELBW babies is associated with learning difficulties andcognition but not ADHD in the school-aged child. Dev Med ChildNeurol. 1999;41:375380

33. Peterson J, Taylor H, Minich N, Klein N, Hack M. Subnormalhead circumference in very low birth weight children: neona-tal correlates and school-age consequences. Early Hum Dev.2006;82:325334

34. Gale C, OCallaghan F, Bredow M, Martyn C; Avon LongitudinalStudy of Parents and Children Study Team. The influence of headgrowth in fetal life, infancy, and childhood on intelligence at theages of 4 and 8 years. Pediatrics. 2006;118:14861492

35. Fisch R, Bilek M, Horrobin J, Chang P-N. Children with superiorintelligence at 7 years of age. Am J Dis Child. 1976;130:481487

36. Klein C. Nutrient requirements for preterm infant formulas. JNutr. 2002;132:1395s1577s

37. Altman D. Treatment allocation by minimisation. BMJ. 2005;330:843

38. Sarnat H, Sarnat M. Neonatal encephalopathy following fetaldistress. A clinical and electroencephalographic study. ArchNeurol. 1976;33:696705

39. Kuban K, Sanocka U, Leviton A, et al. White matter disordersof prematurity: association with intraventricular hemorrhageand ventriculomegaly. The Developmental Epidemiology Net-work. J Pediatr. 1999;134:539546

40. Preece M, Freeman J, Cole T. Sex differences in weight ininfancy: Published centile charts for weight have been updated.BMJ. 1996;313:1486

41. Phillimore P, Beattie A, Townsend P. Widening inequality ofhealth in northern England, 198191. BMJ. 1994;308:11251128

42. Bartholomeusz H, Courchesne E, Karns C. Relationship betweenhead circumference and brain volume in healthy normal tod-dlers, children, and adults. Neuropediatrics. 2002;33:239241

43. Cooke R, Lucas A, Yudkin P, Pryse-Davies J. Head circumfer-ence as an index of brain weight in the fetus and newborn.Early Hum Dev. 1977;1:145149

44. Bray O, Shields W, Wolcott G, Madsen J. Occipitofrontal head


circumference, an accurate measure of intracranial volume.J Pediatr. 1969;78:904908

45. Lindley A, Benson J, Grimes C, Cole T III, Herman J. Therelationship in neonates between clinically measured headcircumference and brain volume estimated from head CT-scans. Early Hum Dev. 1999;56:1729

46. Eyre J, Miller S, Clowry G, Conway E, Watts C. Functionalcorticospinal projections are established prenatally in the hu-man foetus permitting involvement in the development ofspinal motor centres. Brain. 2000;123:5164

47. Sakai H. Morphological properties of fast and slow PT cells inthe cat revealed by intracellular pressure injection of HRP [inJapanese]. Nippon Seirigaku Zasshi. 1982;44:199204

48. Sakai H, Woody C. Relationships between axonal diameter,soma size, and axonal conduction velocity of HRP-filled, pyra-midal tract cells of awake cats. Brain Res. 1988;460:17

49. Yamamoto T, Samejima A, Oka H. Morphology of layer Vpyramidal neurons in the cat somatosensory cortex: an intra-cellular HRP study. Brain Res. 1987;437:369374

50. Deschenes M, Labelle A, Landry P. Morphological character-ization of slow and fast pyramidal tract cells in the cat. BrainRes. 1979;178:251274

51. Eyre JA, Miller S, Ramesh V. Constancy of central conductiondelays during development in man: investigation of motor andsomatosensory pathways. J Physiol (Lond). 1991;434:441452

52. Olivier E, Edgley S, Armand J, Lemon R. An electrophysiolog-ical study of the postnatal development of the corticospinalsystem in the Macaque monkey. J Neurosci. 1997;17:267276

53. Demets D, Ware J. Group sequential methods for clinical trialswith a one-sided hypothesis. Biometrika. 1980;67:651660

54. Lucas A, Morley R, Cole T, Gore S, Lucas P. Early diet inpreterm babies and developmental status at 18 months. Lancet.1990;335:14771481

55. Lucas A, Morley R, Cole T. Randomised trial of early diet inpreterm babies and later intelligence quotient. BMJ. 1998;28:14811487

56. Twisk J, Proper K. Evaluation of the results of a randomizedcontrolled trial: how to define changes between baseline andfollow-up. J Clin Epidemiol. 2004;57:223228

57. Carver J, Wu P, Hall R, et al. Growth of preterm infants fednutrient-enriched or term formula after hospital discharge.Pediatrics. 2001;107:683689

58. Cooke R, Griffin I, McCormick K, et al. Feeding preterm infantsafter hospital discharge: effect of dietary manipulation on nu-trient intake and growth. Pediatr Res. 1998;43:355360

59. Fomon S, Filer L, Ziegler E, Bergmann K, Bergmann R. Skimmilk in infant feeding. Acta Paediatr Scand. 1977;66:1730

60. Fomon S, Filmer LJ, Thomas L, Anderson T, Nelson S. Influ-ence of formula concentration on caloric intake and growth ofnormal infants. Acta Paediatr Scand. 1975;64:172181

61. Pinto-Martin J, Riolo S, Cnaan A, Holzman C, Susser M, Pan-eth N. Cranial ultrasound prediction of disabling and nondis-abling cerebral palsy at age two in a low birth weight popula-tion [published correction appears in Pediatrics. 2001;108:238].Pediatrics. 1995;95:249254

62. Ernst K, Radmacher P, Rafail S, Adamkin D. Postnatal malnu-trition of extremely low birth-weight infants with catch-upgrowth postdischarge. J Perinatol. 2003;23:2003

63. Castillo-Duran C, Rodriguez A, Venegas G, Alvarez P, Icaza G.Zinc supplementation and growth of infants born small forgestational age. J Pediatr. 1995;127:206211

64. Reilly S, Skuse D, Poblete X. Prevalence of feeding problemsand oral motor dysfunction in children with cerebral palsy: acommunity survey. J Pediatr. 1996;129:877882

65. Sullivan P, Juszezak E, Lambert B, Rose M, Ford-Adams M,Johnson A. Impact of feeding problems on nutritional intakeand growth: Oxford Feeding Study II. Dev Med Child Neurol.2002;44:461467

66. Georgieff M, Hoffman J, Pereira G, Bernbaum J, Hoffman-Williamson M. Effect of neonatal caloric deprivation on headgrowth and 1-year developmental status in preterm infants.J Pediatr. 1985;107:581587

67. Brandt I, Sticker EJ, Lentze MJ. Catch-up growth of headcircumference of very low birth weight, small for gestationalage preterm infants and mental development to adulthood.J Pediatr. 2003;142:463470

68. Tamer S, Misra S, Jaiswal S. Central motor conduction time inmalnurished children. Arch Dis Child. 1997;77:323325

69. Sima A, Sourander P. The effect of pre- and postnatal under-nutrition on axonal growth and myelination of central motorfibres. Acta Neuropathol (Berl). 1978;42:1518

70. Quirk G, Mejia W, Hesse H, Su H. Early malnutrition followedby nutritional restoration lowers the conduction velocity andexcitability of the corticospinal tract. Brain Res. 1995;670:277282

71. Salas M, Diaz S, Nieto A. Effects of neonatal food deprivationon cortical spines and dendritic development of the rat. BrainRes. 1974;73:139144

72. Super C, Herrera M, Mora J. Long-term effects of food supple-mentation and psychosocial intervention on the physicalgrowth of Colombian infants at risk of malnutrition. Child Dev.1990;61:2949

73. Jahari A, Saco-Pollitt C, Husaini M, Pollitt E. Effects of anenergy and micronutrient supplement on motor developmentand motor activity in undernourished children in Indonesia.Eur J Clin Nutr. 2000;54(suppl 2):S60S68

74. Freeman H, Klein R, Townsend J, Lechtig A. Nutrition andcognitive development among rural Guatemalan children.Am J Public Health. 1980;70:12771285

75. Waber D, Vuori-Christiansen L, Ortiz N, et al. Nutritional sup-plementation, maternal education, and cognitive developmentof infants at risk of malnutrition. Am J Clin Nutr. 1981;34(suppl4):807813

76. Thornberg E, Thiringer K, Odeback A, Milsom I. Birthasphyxia: incidence, clinical course and outcome in a Swedishpopulation. Acta Paediatr. 1995;84:927932

77. Verhaart J. Hypertrophy of the pes pedunculi and pyramid asa result of degeneration of the contralateral corticofugal fibretracts. J Comp Neurol. 1950;92:115


ServicesUpdated Information &

mlhttp://pediatrics.aappublications.org/content/121/1/148.full.htincluding high resolution figures, can be found at:

References

ml#ref-list-1http://pediatrics.aappublications.org/content/121/1/148.full.htat:This article cites 75 articles, 16 of which can be accessed free

Citations

ml#related-urlshttp://pediatrics.aappublications.org/content/121/1/148.full.htThis article has been cited by 7 HighWire-hosted articles:

Subspecialty Collections

ubhttp://pediatrics.aappublications.org/cgi/collection/smoking_sSmoking

abuse_subhttp://pediatrics.aappublications.org/cgi/collection/substance_Substance Abuse

al_surgery_subhttp://pediatrics.aappublications.org/cgi/collection/neurologicNeurological Surgery

_subhttp://pediatrics.aappublications.org/cgi/collection/neurologyNeurology

y_subhttp://pediatrics.aappublications.org/cgi/collection/neonatologNeonatology

orn_infant_subhttp://pediatrics.aappublications.org/cgi/collection/fetus:newbFetus/Newborn Infantthe following collection(s):This article, along with others on similar topics, appears in

Permissions & Licensing

mlhttp://pediatrics.aappublications.org/site/misc/Permissions.xhttables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,

Reprints http://pediatrics.aappublications.org/site/misc/reprints.xhtml

Information about ordering reprints can be found online:

rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Grove Village, Illinois, 60007. Copyright 2008 by the American Academy of Pediatrics. All and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elkpublication, it has been published continuously since 1948. PEDIATRICS is owned, published, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

at Indonesia:AAP Sponsored on March 7, 2014pediatrics.aappublications.orgDownloaded from

DOI: 10.1542/peds.2007-1267 2008;121;148Pediatrics

and Janet A. EyreLyvia Dabydeen, Julian E. Thomas, Tessa J. Aston, Hilary Hartley, Sunil K. Sinha

in Term and Preterm Infants After Perinatal Brain InjuryHigh-Energy and -Protein Diet Increases Brain and Corticospinal Tract Growth

rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Grove Village, Illinois, 60007. Copyright 2008 by the American Academy of Pediatrics. All and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elkpublication, it has been published continuously since 1948. PEDIATRICS is owned, published, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

at Indonesia:AAP Sponsored on March 7, 2014pediatrics.aappublications.orgDownloaded from

High-Energy and -Protein Diet Increases Brain and Corticospinal Tract Growth

Documents