Research ajog.org

OBSTETRICS

Brain metabolite differences in one-year-oldinfants born small at term and associationwith neurodevelopmental outcomeRui V. Simoes, PhD; Monica Cruz-Lemini, MD, PhD; Nuria Bargallo, MD, PhD;Eduard Gratacos, MD, PhD; Magdalena Sanz-Cortes, MD, PhD

OBJECTIVE: We assessed brain metabolite levels by magnetic RESULTS: As compared with AGA controls, infants born small showed

resonance spectroscopy (MRS) in 1-year-old infants born smallat term, as compared with infants born appropriate for gestationalage (AGA), and their association with neurodevelopment at 2years of age.

STUDY DESIGN: A total of 40 infants born small (birthweight <10thcentile for gestational age) and 30 AGA infants underwent brain MRSat age 1 year on a 3-T scanner. Small-born infants were subclassifiedas late intrauterine growth restriction or as small for gestational age,based on the presence or absence of prenatal Doppler and birthweightpredictors of an adverse perinatal outcome, respectively. Single-voxelproton magnetic resonance spectroscopy (1H-MRS) data were ac-quired from the frontal lobe at short echo time. Neurodevelopment wasevaluated at 2 years of age using the Bayley Scales of Infant andToddler Development, Third Edition, assessing cognitive, language,motor, social-emotional, and adaptive behavior scales.

From BCNatal, Barcelona Center for Maternal-Fetal and Neonatal Medicine,Research Center, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDDiseases (CIBER-ER) (Drs Simões, Gratacós, and Sanz-Cortés); Fundació HoFetal Medicine Unit, Vall d’Hebron University Hospital, Universitat Autònoma dCentre de Diagnostic per la Imatge, Hospital Clínic and Medical Image platfo

Received Oct. 29, 2014; revised Feb. 6, 2015; accepted April 14, 2015.

This workwas funded by grants from theCerebra Foundation for the Brain InjuFundación Dexeus, Barcelona, Spain; and Agència de Gestió d’Ajuts Universnumber 928. M.S.-C. was supported by a Rio Hortega postdoctoral fellows

The authors report no conflict of interest.

Corresponding author: Rui V. Simões, PhD. portasferre@clinic.ub.es

0002-9378/$36.00 � ª 2015 Elsevier Inc. All rights reserved. � http://dx.doi.org/10.1

210.e1 American Journal of Obstetrics & Gynecology AUGUST 2015

significantly higher levels of glutamate and total N-acetylaspartate(NAAt) to creatine (Cr) ratio at age 1 year, and lower Bayley Scales ofInfant and Toddler Development, Third Edition scores at 2 years. Thesubgroup with late intrauterine growth restriction further showed lowerestimated glutathione levels at age 1 year. Significant correlationswere observed for estimated glutathione levels with adaptive scores,and for myo-inositol with language scores. Significant associationswere also noticed for NAA/Cr with cognitive scores, and for glutamate/Cr with motor scores.

CONCLUSION: Infants born small show brain metabolite differences at1 year of age, which are correlated with later neurodevelopment.These results support further research on MRS to develop imagingbiomarkers of abnormal neurodevelopment.

Key words: brain metabolism, intrauterine growth restriction, mag-netic resonance spectroscopy, small for gestational age

Cite this article as: Simoes RV, Cruz-Lemini M, Bargallo N, et al. Brain metabolite differences in one-year-old infants born small at term and association with neu-rodevelopmental outcome. Am J Obstet Gynecol 2015;213:210.e1-11.

mall-born infants are those with a

S birthweight <10th centile forgestational age. This represents a com-mon condition affecting up to 10% of alldeliveries at term1 and is associated witha higher risk for adverse neurologicaland cardiovascular outcome.2 Duringgestation, fetal smallness is identified byultrasound as an estimated fetal weight

<10th centile for gestational age, and athorough Doppler ultrasound assess-ment is necessary in such cases to eval-uate the severity. Among the late-onsetforms of fetal smallness, about two thirdspresent signs of severity, including pre-natal Doppler cerebroplacental ratio(middle cerebral to umbilical arterypulsatility index ratio <5th centile and/

Hospital Clínic and HospiIBAPS), University of Barspital Sant Joan de Déu (e Barcelona (Dr Cruz-Lemrm, IDIBAPS (Dr Bargalló)

redChild, Carmarthen,Waitaris i de Recerca (AGAUhip, Spain (CM10/00222).

016/j.ajog.2015.04.011

or mean uterine artery pulsatility index>95th centile and/or estimated fetalweight <3rd centile), and are associatedwith poorer perinatal outcome.3-5 It hasbeen suggested that these cases representtrue forms of placental insufficiencyand should be considered as late-onset intrauterine growth restriction(IUGR).6 The remaining one-third of

tal Sant Joan de Deu, Fetal i+D Fetal Medicinecelona, Centre for Biomedical Research on RareDr Simões); Department of Obstetrics, Maternal-ini); and Department of Radiology Hospital Clinic,, Barcelona, Spain.

les, UnitedKingdom;Obra Social “LaCaixa” andR) 2014 Suport a Grups de Recerca (SGR) grant

ajog.org Obstetrics Research

cases of late-onset fetal smallness,without signs of severity, should beconsidered as small for gestational age(SGA).7 While SGA fetuses have beencommonly defined as constitutionallysmall, there is no solid evidence to sup-port this and they may represent anotherform of pathological smallness. Regard-less of the clinical presentation, bothclinical groups of small fetuses are asso-ciated with poorer neurodevelopment asearly as the neonatal period and duringlater stages,8-12 affecting mostly func-tions associated with frontal networkingsuch as attention, creativity, language,memory performance, and learningabilities.8 The neurostructural and neu-rofunctional phenotypes associated withfetal smallness, and their relationships,are still only partially understood.

The study of brain metabolism is animportant tool to assess neurodevel-opment, given their intrinsic relationship.Previous studies have shown that smallfetuses have lower frontal lobe ratios of N-acetylaspartate (NAA, a neuronal marker)to choline compounds (Cho, marker ofcell membrane turnover) and lower ratiosof NAA to creatine (Cr, a potential glialmarker implicated in cellular ener-getics).13,14 These metabolic changes havebeen associated with disrupted brainmaturation13 but it remains unknown towhat extent they are a reflection of thebrain tissue exposure to an adverse inutero environment, and how they mayunderlie true neurostructural changes15-20

leading to brain remodeling and futurechanges in neurodevelopmental function.

In this study we addressed the hy-potheses that brain metabolite changeswere present in the frontal lobe of1-year-old small-born infants at term,and could be correlated with neuro-developmental outcome at 2 years of age.

MATERIALS AND METHODS

Study cohort and clinical perinataldataThis study is part of a larger prospectiveresearch program on growth restrictioninvolving fetal, and short- and long-termpostnatal, follow-up. The protocol usedwas approved by the local institutionalethics committee (review board 2010/5736) and all participants gave their

written informed consent. A consecutivesample of 70 neonates were prospectivelyrecruited at birth, all delivered at termfrom singleton pregnancies and pre-senting normal umbilical artery pulsa-tility index (<95th centile) at the time ofdelivery.21-23 Subjects were classified asappropriate for gestational age (AGA)(30 subjects) or small (40 subjects),based on their birthweight above orbelow the 10th centile,24 respectively. Inall cases, gestational age was correctedfrom fetal crown-rump length in the firsttrimester.25 Additionally, the small-bornneonate group was subclassified accord-ing to recent criteria proposed,7 consid-ering their birthweight and prenatalultrasoundDoppler parameters from thescan closest to the time of delivery. Thus,small-born infants were classified as late-onset IUGR (31 subjects) if signs ofseverity were present (cerebroplacentalratio<5th centile26 and/or mean uterineartery pulsatility index >95th centile26

and/or a birthweight <3rd centile7), oras SGA (9 subjects) when none of thesefactors were present. Middle cerebral,umbilical, and mean uterine artery pul-satility index parameters were measuredas reported earlier,3,26 using a SiemensSonoline Antares (Siemens, Erlangen,Germany) system with a 6 to 2MHzlinear curved-array transducer.Neonates with congenital malforma-

tions, chromosomal abnormalities, in-fections, chronic maternal pathology, ornoncephalic presentations were noteligible for this study. Maternal andperinatal data were prospectively recor-ded in most study patients.

Magnetic resonance acquisitionBrain magnetic resonance imaging wascarried out at 14 (�1.5) months of age,without sedation, during natural sleep.Data were acquired with a 3.0-T scanner(Tim Trio; Siemens Diagnostics Health-care, Erlangen, Germany) and a headmatrix radiofrequency coil was used.The total length of each magnetic reso-nance examination did not exceed 45minutes. Reference T1-weighted anat-omical images were acquired withmagnetization prepared rapid acquisi-tion gradient echo. Single-voxel protonspectroscopy (1H-MRS) was carried out

AUGUST 2015 Ameri

with point-resolved spectroscopy fromthe frontal lobe region (Figure 1, A),using the following parameters: 40 �20 � 20 mm3 voxel size, 2000-millisecond repetition time, 30-millisecond echo time, 98 averages,chemical shift selective water suppres-sion, and 3-minute 24-second acquisi-tion time. A reference spectrum with 16averages was also acquired, withoutwater suppression. Finally, T2-weightedimages (half-Fourier acquisition single-shot turbo spin-echo) were obtained;magnetic resonance spectroscopy (MRS)was repeated if artifacts were detected(eg, due to head movements). Structuralmagnetic resonance images were re-viewed for the presence of anatomicalabnormalities by an experienced neuro-radiologist blinded to groupmembership.

MRS postprocessingMRS data with gross visual artifacts and/or absence of an interpretable metabolicpattern27 were directly discarded. Theremaining MRS data were processedusing linear combination model-fitting(LCModel; S. Provencher Inc., http://s-provencher.com/pages/lcmodel.shtml).28

The basis sets used included a total of 23metabolites: alanine, aspartate, Cr andphosphocreatine (total Cr), gammaaminobutyric acid, glucose, glutamine,glutamate, glycine, phosphocholineand glycerophosphocholine (cholinecompounds, Cho), glutathione, myo-inositol, lactate, N-acetylaspartate, andN-acetylaspartyilglutamate (NAAG)(abbrieviated together as total NAA,NAAt), phosphoethanolamine, scyllo-inositol, taurine, threonine, valine,acetate, and ascorbate. Lipid andmacromolecule contributions were alsoincluded in the modeling. All MRS dataanalyzed had an estimated signal-to-noise ratio >10 and only spectrawith an estimated full width at halfmaximum <0.1 ppm were selected forfurther analysis. Those metabolite fit-tings with Cramér-Rao lower bounds(estimated SDs of the estimated con-centration)29 >50% were also discardedand only metabolite changes of at leasttwice their Cramér-Rao lower bounds(95% confidence interval) were selected,unless indicated otherwise. The data

can Journal of Obstetrics & Gynecology 210.e2

FIGURE 1MRS acquisition in 1-year-old infant brain

A, Point-resolved spectroscopy voxel positioned in frontal lobe, as seen on reference T1-weightedmagnetic resonance images (white rectangle; from left to right): sagittal, coronal, and axial views of

infant brain. B, Single voxel proton magnetic resonance spectroscopy (MRS) acquired with

30-millisecond echo time from appropriate-for-gestational-age infant. MRS pattern (in black ) is

shown with LCModel (S. Provencher Inc., http://s-provencher.com/pages/lcmodel.shtml) fitting

overlaid (in red ), including respective baseline correction (in gray ) and fitting residual (on top).

Spectral pattern assignments for major visible peaks (ppm): aspartate (3.89, 2.80, 2.65), choline

compounds (mostly phosphocholine and glycerophosphocholine: 3.94, 3.90, 3.87, 3.67, 3.64, 3.6,

3.21), total creatine (creatine and phosphocreatine: 3.93, 3.91, 3.03), gamma aminobutyric acid

(3.01, 2.28, 1.89), glutamine (3.75, 2.45, 2.12), glutamate (3.74, 2.35, 2.03), glycine (3.55),

glutathione (3.77, 2.97, 2.92, 2.56, 2.51, 2.15), myo-inositol (4.05, 3.61, 3.52, 3.26), and total

NAA, including N-acetylaspartate (2.00, 2.49, 2.67) and N-acetylaspartyilglutamate (1.88, 2.04,

2.2, 2.5, 2.72); lipids and macromolecules (0.9, 1.3, 2.0, 2.3, 2.8).

Simões. Brain MRS changes in infants born small at term. Am J Obstet Gynecol 2015.

Research Obstetrics ajog.org

were analyzed quantitatively, using thereference water scans (estimated mmol/kg water content in brain tissue), and asratios to total Cr and to Cho compounds(Cho) (Figure 1, B, metabolite assign-ments according to the literature30).

210.e3 American Journal of Obstetrics & Gynecol

Neurodevelopmental assessmentNeurodevelopmental outcome wasassessed at 23 (�1.5) months of age,using the Bayley Scales of Infant andToddler Development, Third Edition(BSID-III). Five distinct areas were

ogy AUGUST 2015

evaluated: cognitive, language, motor,social-emotional, and adaptive behavior.Each area consists of a score with anormalizedmean of 100� 15 (controls).Abnormal BSID-III was defined as ascore <85 (1 SD) in any of the 5scales.31,32 All developmental examina-tions were performed by a single expe-rienced psychologist, blinded to infantmedical history.

Statistical analysisQuantitative and qualitative parametersfrom the clinical/demographic data werecompared with the Student t test forindependent samples (or Mann-Whitney U test) and with the Pearsonc2, respectively. The Wilcoxon test wasused to compare paired groups withoutnormal distributions. MRS differencesbetween cases and controls, and betweenneurodevelopmental score groups, wereassessed using the Student t test for in-dependent samples and with generallinear model, adjusting by potentialconfounding factors: gestational age atbirth, sex, maternal breast-feeding >4months, age at magnetic resonance,and low socioeconomic status, whencomparing MRS differences betweensmall-born infants and controls; and ageat BSID-III test, when comparing MRSdifferences between infants with low andnormal BSID-III scores. Pearson andSpearman correlation tests were per-formed to assess univariate correlationsbetween metabolite levels and BSID-IIItest scores. All statistical analysis wereperformed with software (SPSS, version19.0; IBM Corp, Armonk, NY), con-sidering results significant at a P value< .05.

RESULTS

Infant cohort and perinatal clinicalobservationsAfter MRS postprocessing and qualityassessment, spectra from 59 infants wereselected for further analysis (33 smallborn, 83%; and 26 AGA, 87%) (Table 1).As expected, infants born small weredelivered earlier and had longer stays atthe neonatal intensive care unit thanAGAs. No signs of intracranial pathologywere found in any of the anatomicmagnetic resonance images. At 1 year of

TABLE 1Perinatal clinical and demographic observations, and BSID-III testscoresDemographic AGA (n [ 26) Small (n [ 33) P value

Maternal characteristics

Maternal age, y 31.8 � 7.2 32.6 � 4.8 .627a

Low maternal socioeconomic status, % 45.8 60.6 .269b

Maternal smoking during pregnancy, % 24.0 29.0 .672b

Perinatal data

GA at birth, wk 39.8 � 1.4 38.1 � 0.9 .000a

Cesarean delivery, % 22.7 45.2 .093b

Instrumental vaginal delivery, % 17.6 5.9 .157b

Gestational diabetes, % 8.3 0.0 .179b

Preeclampsia, % 0.0 6.3 .501b

Birthweight, g 3375.2 � 403.1 2304.9 � 263.8 .000a

Birthweight centile 50.7 � 31.3 2.1 � 2.6 .000c

Male sex, % 57.7 69.7 .339b

Umbilical artery pH <7.15, % 12.5 7.1 .552b

5-min Apgar score <7, % 0.0 0.0 e

Length, cm 49.9 � 1.5 45.5 � 2.2 .000c

Length centile 48.2 � 24.0 8.7 � 7.3 .000c

Cephalic perimeter, cm 34.2 � 0.8 32.3 � 1.2 .000a

Cephalic perimeter centile 34.2 � 19.7 12.7 � 16.9 .000c

Admission to NICU, d 0.1 � 0.4 1.2 � 2.4 .05c

Infant parameters at 12 mo

Age at MRI, wk 56.4 � 6.0 56.0 � 6.0 .800a

Weight, kg 9.8 � 1.0 8.5 � 1.2 .001a

Weight centile 38.4 � 29.9 13.0 � 21.2 .000c

Height, m 0.75 � 0.03 0.71 � 0.03 .000a

Height centile 51.2 � 29.0 20.8 � 19.6 .003c

Cephalic perimeter, cm 46.2 � 1.3 45.2 � 1.2 .016a

BMI 17.3 � 1.6 16.8 � 2.0 .352a

Breast-feeding >4 mo, % 47.1 69.6 .151b

BSID-III scores at 24 mo

Cognitive 101.8 � 15.0 99.0 � 17.1 .363d

Language 108.1 � 19.2 95.4 � 15.1 .223d

Motor 104.9 � 15.5 97.1 � 13.0 .091d

Social-emotional 109.3 � 25.4 94.1 � 24.8 .070d

Adaptive 103.9 � 13.6 97.6 � 15.9 .448d

Average score of all areas 106.0 � 13.2 96.5 � 12.6 .038d

Data are displayed as mean metabolite levels � SD.

AGA, appropriate for gestational age; BMI, body mass index; BSID-III, Bayley Scales of Infant and Toddler Development, ThirdEdition; GA, gestational age; MRI, magnetic resonance imaging; NICU, neonatal intensive care unit.

a Student t test for independent samples; b Pearson c2 test; c Mann-Whitney U test; d General linear model adjusting by sex,maternal low economic status, GA at delivery, breast-feeding >4 mo, and age at BSID-III as covariates.

Simões. Brain MRS changes in infants born small at term. Am J Obstet Gynecol 2015.

ajog.org Obstetrics Research

AUGUST 2015 Ameri

age, small-born infants were still signifi-cantly smaller than AGAs (height andweight centiles calculated as reported byothers33).With regard to brainMRS data,no significant differences were noticedbetween the 2 groups as far as signal-to-noise ratio (small born, 26 � 6.7;AGA, 28.4 � 5.9) and full width athalf maximum (small born, 0.049 �0.012; AGA, 0.047 � 0.012). In 8 (14%)examinations (6 small born and 2AGAs) the quantifications were limitedto metabolite ratios since the waterreference scan was not available. No sta-tistically significant differences in peri-natal clinical data were found betweencases with interpretable and uninter-pretable spectra.

Brain metabolite changes in infantsborn small at termEstimated metabolite levelsThe most consistent, significant differ-ence in small-born infants compared tocontrols was an increase in brain gluta-mate levels (Figure 2, A). This changewas also depicted in the mixed quanti-fication of glutamate and glutamine(glutamateþ glutamine; Figure 2, B). Anincrease in total NAA (NAAt) was alsonoticed in infants born small comparedto AGAs (Figure 2, C). After adjusting forpotential confounding variables, the in-crease in glutamate and glutamate þglutamine remained significant, whilethe NAAt increase was borderline sig-nificant (P ¼ .050) (Table 2).

Metabolite ratiosWhen considering metabolite ratios,small-born infants showed significantincreases in glutamate þ glutamine/Crand NAAt/Cr (Figure 2, D and E). Also,a slight but not significant increase inNAAt/Cho was detected (Figure 2, F)(Table 2).

Changes in SGA and late-onsetIUGR infantsWithin the group of 33 small-born in-fants with usableMRS data, 26 were late-onset IUGR and 7 were SGA. Increasedglutamate and NAAt levels wereobserved in late-onset IUGR infantswhen compared to controls (Table 3),reaching significance for the NAAt/Cr

can Journal of Obstetrics & Gynecology 210.e4

FIGURE 2Brain metabolite changes in small-born infants

Changes in estimated concentrations of A, glutamate (Glu), B, Glu and glutamine (Glx), and C, total NAA, and in ratios of D, Glx/creatine (Cr), E, NAAt/Cr,

and F, NAAt/choline (Cho). Data are displayed as mean � SD.

AGA, appropriate for gestational age.

*P < .05 (general linear model, adjusted by maternal low economic status, gestational age at delivery, breast-feeding >4 months, and age at magnetic resonance imaging as covariates).

Simões. Brain MRS changes in infants born small at term. Am J Obstet Gynecol 2015.

TABLE 2Brain metabolite levels in small-born compared to AGA infantsMetabolite AGA Small Change P valuea P valueb

Glutamatec 6.0 � 0.8 6.6 � 1.1 10% .041 .026

Glutamate þ glutaminec 6.5 � 1.0 7.2 � 1.4 11% .049 .032

NAAtc 4.4 � 0.8 4.7 � 0.5 8% .073 .05

Glutamate þ glutamine/Cr 1.5 � 0.3 1.7 � 0.4 7% .121 .035

NAAt/Cr 1.0 � 0.1 1.1 � 1.4 7% .095 .039

NAAt/Cho 4.1 � 0.5 4.4 � 0.7 7% .099 .432

Data are displayed as mean metabolite levels � SD for AGA and small-born infants and respective changes.

AGA, appropriate for gestational age; Cho, choline; Cr, creatine.

a Student t test; b General linear model adjusted by sex, maternal low economic status, gestational age at delivery, breast-feeding >4 mo, and age at magnetic resonance imaging as covariates; c Estimated concentrations (mmol/kg).

Simões. Brain MRS changes in infants born small at term. Am J Obstet Gynecol 2015.

Research Obstetrics ajog.org

ratio (Figure 3). Moreover, estimatedglutathione levels were significantlydecreased in late-onset IUGR infants(Table 3). Although these levels areclose to the glutathione concentrationsreported in adult brain by MRS,34,35

the estimated change detected(e12%) was slightly below the fittingerror ranges (Cramér-Rao lowerbounds, 8 � 2%). Metabolite changesin SGA infants were more variable,likely due to the lower sample sizes,but showed the same tendency as inlate-onset IUGRs when compared tocontrols. No significant metabolitedifferences were detected betweenSGA and late-onset IUGR subgroups.

Neurodevelopmental outcomeThe BSID-III test was performed in 70%of our initial infant cohort, including81% of the infants with quantifiableMRS data (31 small-born infants and

210.e5 American Journal of Obstetrics & Gynecol

18 AGA). Infants born small had lowerBSID-III scores than controls, reachingsignificance in the language area (P ¼.013) and close to significance in themotor and social-emotional domains

ogy AUGUST 2015

(P ¼ .063 and .057, respectively). Whenpotential confounding factors wereconsidered, the test average scoreremained significantly lower in cases,whereas a nonsignificant trend in motor

TABLE 3Brain metabolite levels in clinical subgroups of infants born smallMetabolite AGA SGA Change P valuea P valueb Late-onset IUGR Change P valuea P valueb

Glutamatec 6.0 � 0.8 7.1 � 1.4 18% .015 .115 6.4 � 0.9 7% .125 .111

Glutamate þ glutaminec 6.5 � 1.0 7.9 � 1.4 21% .015 .072 7.0 � 1.2 8% .154 .143

NAAtc 4.4 � 0.8 4.7 � 0.7 7% .389 .118 4.7 � 0.5 8% .076 .107

NAAt/Cr 1.0 � 0.1 1.0 � 0.2 3% .695 .126 1.1 � 0.1 7% .055 .004

Glutathionec 1.4 � 0.3 1.4 � 0.6 2% .845 .005 1.2 � 0.2 e12% .045 .641

Data are displayed as mean metabolite levels� SD and respective changes, for AGA infants and subgroups of small-born infants: SGA, without predictors of adverse perinatal outcome; and late-onset IUGR, with presence of predictors of adverse perinatal outcome (cerebral placental ratio <5th centile and/or mean uterine artery pulsatility index >95th centile and/or birthweight <3rdcentile).

AGA, appropriate for gestational age; Cr, creatine; IUGR, intrauterine growth restriction; SGA, small for gestational age.

a Student t test; b General linear model, sex, adjusted by maternal low economic status, gestational age at delivery, breast-feeding >4 mo, and age at magnetic resonance imaging as covariates;c Estimated concentrations (mmol/kg).

Simões. Brain MRS changes in infants born small at term. Am J Obstet Gynecol 2015.

ajog.org Obstetrics Research

and social-emotional scores was noticed(Table 1).

Brain metabolite levels andneurodevelopmental test scoresLinear correlationsThe strongest and most significant linearcorrelations were found for the esti-mated glutathione levels and adaptivescores: glutathione (R¼ 0.463; P¼ .003)(Figure 4, A) and glutathione/Cr ratio(R ¼ 0.346; P ¼ .021). Myo-inositolcorrelated significantly with language

FIGURE 3Brain NAAt/Cr changes insubgroups of small-born infants

Estimated brain NAAt/Cr changes in SGA and late-

onset IUGR. Data are displayed as mean � SD.

AGA, appropriate for gestational age; Cr, creatine; IUGR, intra-uterine growth restriction; SGA, small for gestational age.

*P < .05 (general linear model, adjusted by maternal low eco-nomic status, gestational age at delivery, breast-feeding >4months, and age at magnetic resonance imaging as covariates).

Simões. Brain MRS changes in infants born small at term.Am J Obstet Gynecol 2015.

scores (Figure 4, B), as well as the mixedcontribution myo-inositol and glycine(myo-inositol þ glycine, R ¼ 0.309;P ¼ .046).

Metabolite changes and poorneurodevelopmental outcomeA nonsignificant trend toward lowerestimated glutathione levels was noticedin infants with low adaptive scores. In-fants with low language scores hadsignificantly lower myo-inositol (andmyo-inositol þ glycine) levels, whichremained significant after adjusting forconfounding variables (Table 4). Signif-icantly higher NAA/Cr levels were asso-ciated with low cognitive scores andhigher glutamate/Cr levels noticed ininfants with low motor scores. Afteradjusting for the confounding variables,glutamate/Cr changes remained signifi-cant and NAA/Cr changes did not(Table 4).

COMMENT

We have detected significant increases infrontal lobe glutamate and NAA (NAAt/creatine) levels in small-born 1-year-oldinfants compared to AGA controls. Thesemetabolite changes were also observed inthe subgroup of late-onset IUGR infants,where lower estimated glutathione levelswere noticed. Small-born infants hadoverall lower neurodevelopmental scoresat 2 years of age. Estimated glutathioneand myo-inositol levels significantlycorrelated with adaptive and language

AUGUST 2015 Ameri

scores, respectively. Low language scoreswere significantly associated with lowermyo-inositol, whereas higher NAA/Crand glutamate/Cr ratios were found ininfants with low cognitive and motorscores, respectively.

Lower brain NAA/Cho levels havepreviously been detected in IUGR fe-tuses13 and in preterm infants, in whomthey were predictive of adverse neuro-developmental outcome at 2 years ofage.36-38 Lower levels of frontal lobeNAA/Cho detected in the brain of smallfetuses at term14 seem to be no longerpresent at 1 year of age, as they wereslightly increased in this specific popu-lation. Caution should be taken whencomparing these 2 studies though, sincethe specific MRS voxel position is ex-pected to be less precise in the fetal brain,due to unrestricted motion in utero,compared to the infant brain. Despitethe lack of comparable studies in small-born infants, our results are consistentwith the metabolite alterations reportedin different neurological conditions.IncreasedNAA and glutamate levels havebeen reported in the prefrontal cortex ofautistic children.39 Also, higher gluta-mate levels were found in autistic chil-dren,39 and associated with bipolar40,41

and attention deficit hyperactivity42

disorders. Lower brain glutathionelevels have been associated with chronichypoxia43 and a spectrum of psychiatricconditions in patients44,45 and animalmodels.46,47 Finally, lower myo-inositol

can Journal of Obstetrics & Gynecology 210.e6

FIGURE 4Correlation between estimated infant brain metabolite levels and Bayley Scales of Infant and ToddlerDevelopment, Third Edition scores

Linear fittings (straight line) and SD intervals (dashed lines). A, Glutathione (GSH) and adaptive scores: R ¼ 0.463; P ¼ .003. B, Myo-inositol (MI) and

language scores: R ¼ 0.333; P ¼ .031.

Simões. Brain MRS changes in infants born small at term. Am J Obstet Gynecol 2015.

Research Obstetrics ajog.org

levels have been detected in the frontallobe of depressed bipolar patients.48

The lower scores in BSID-III test thatwere detected in term small-born infantsare supported by previously reportedstudies,49 showing stronger significantdifferences in motor, cognitive, adaptive,and language scores. The associationsbetween specific metabolite changes andneurological deficits are too complex to

TABLE 4Metabolite changes based on BSID-IBSID-III test areas Metabolite

Adaptive Glutathionec

Glutathione/Cr

Language Myo-inositolc

Myo-inositol þ glyci

Cognitive NAA/Cr

Motor Glutamate/Cr

For each BSID-III test area, scores were classified as normal if �BSID-III, Bayley Scales of Infant and Toddler Development, Third

a Student t test; b General linear model, adjusted by maternal lc Estimated concentrations (mmol/kg).

Simões. Brain MRS changes in infants born small at term.

210.e7 American Journal of Obstetrics & Gynecol

address based on the exploratory natureof this study. However, taken togetherthey could reflect the functional trans-lation of impaired metabolic circuits,highlighting their role as potential imagebiomarkers.While NAA is the second most abun-

dant brain molecule and a well-established neuronal marker, alsoimplicated in the synthesis of myelin,

II test results at 2 years of ageNormal scores Low scores

1.3 � 0.3 1.1 � 0.3

0.3 � 0.1 0.2 � 0.1

3.2 � 1.1 2.3 � 0.9

nec 3.8 � 0.8 3.2 � 0.8

0.9 � 0.1 1.0 � 0.1

1.4 � 0.2 1.6 � 0.4

85 and low if <85. Data are displayed as mean metabolite levels �Edition; Cr, creatine.

ow economic status, gestational age at delivery, breast-feeding >4 m

Am J Obstet Gynecol 2015.

ogy AUGUST 2015

glutamate is the major excitatoryneurotransmitter in the human brain.NAA and glutamate are both synthesizedin neurons, as byproducts of the tricar-boxylic acid cycle. It could be hypothe-sized that, similarly to what has beenproposed in autism, an impairedbiosynthesis, transport, and/or con-sumption of NAA and glutamate byneurons and glial cells, as well as a

Change P valuea P valueb

e15% .068 .51

e17% .067 .546

e27% .016 .026

e17% .019 .012

10% .036 .115

14% .03 .033

SD and respective changes.

o, and age at magnetic resonance imaging as covariates;

ajog.org Obstetrics Research

disturbed brain energy metabolism,could explain their increased levels.39

The changes detected in NAA shouldreflect differences in neuronal densityand/or viability (maturation), whichcould explain to some extent the reducedstructure but increased organization ofbrain networks recently reported in 1-year-old IUGR-born infants.50 On theother hand, higher glutamate levelscould potentially derive from a hypo-functional dopaminergic system, assuggested for attention deficit hyperac-tivity.51 Moreover, glutathione is an anti-oxidant that protects cells from reactiveoxygen species. Hence, late-onset IUGRinfants may have less propensity to syn-thesize (and/or regenerate) glutathione,becoming more susceptible to reactiveoxygen species byproducts of oxygenmetabolism and therefore to mitochon-drial oxidative stress. Interestingly, gluta-thione was the only estimated metaboliteshowing modest but significant linearcorrelations with birthweights in thisstudy (Appendix; Supplementary Figure).Myo-inositol has been proposed as a glialmarker.52 It is the predominant isomer ofinositol and a precursor for inositolphosphates, which are intracellular mes-sengers implicated in several biologicalprocesses, such as nerve guidance53 andintracellular calcium regulation.54 There-fore, inositol is essential for neuronal ac-tivity and its depletion is associated withimpaired brain function.55

When both subgroups of small-borninfants were considered, metabolitechanges were present in those who pre-sented signs of an adverse perinataloutcome and in those who did not.Interestingly, while NAA changesshowed a linear trend across the 3 sub-groups, changes in glutamate (andmixed glutamine pool) were moreapparent in SGA cases whereas estimatedglutathione changes were mostly detec-ted in late-onset IUGR. Although nomajor conclusions can be drawn due tothe small sample size of the subgroupsstudied (especially SGA), we could inferthat signs of brain remodeling are pre-sent even in the less clinically severe formof smallness, coinciding with our previ-ous experience in term small fetuses.14

These results challenge the concept that

SGA fetuses should be considered asconstitutionally small, and are alignedwith reports showing signs of meta-bolic56 and cardiovascular57 remodelingin this clinical subgroup.The clinical relevance of our findings

relies on the possibility that specificbrain metabolite imbalances couldbecome early markers to identify infantsat risk for neurodevelopmental impair-ments during the first 2 years of life. Thisperiod represents an important windowof opportunity due to the high plasticityof the brain.58 Accordingly, postnatalclinical interventions have shown a sig-nificant positive impact on develop-mental outcome.59-61 Identifying infantswith signs of brain remodeling wouldallow public health resources to beprioritized on this population at risk.We acknowledge some limitations in

our work. The sample size of infantsrecruited per group (30-40) was chosenarbitrarily due to the exploratory natureof this study. A larger population cohortwill be required to investigate specificdifferences in SGA infants andstrengthen the associations be-tween brain metabolites and neuro-developmental outcome. Moreover,some methodological limitationsshould also be addressed. We acknowl-edge that the absence of outer volumesuppression pulses lead to partial con-taminations from adjacent tissues and/or subcutaneous fat (lipid-macromole-cule region at 1.3-0.9 ppm), while therelatively short repetition time used isexpected to induce a certain degree ofT1 saturation in our signals. Theselimitations were a compromise to ac-quire magnetic resonance data in atimely manner, since it is difficult forinfants to remain in natural sleep insidethe scanner for long periods. In anycase, this affected our estimations ofabsolute metabolite concentrations(not as precise as in other studies62,63)but essentially not the metabolite ratios,nor the purpose of our study: tocompare metabolite changes in casesand controls. Also, the glutathionechanges estimated here should beconfirmed by specific editing MRStechniques.34,64 Some significant dif-ferences detected in metabolite ratios

AUGUST 2015 Ameri

but not in absolute levels (eg, NAAt/Cr)could be explained either by the slightlylarger sample size and/or by possibleeffects of combined metabolite changesin the metabolites rated.

As for the strengths of this study, itrelies on a well-characterized cohort ofinfants, with good clinical follow-up. Toour knowledge, this is the first study ofinfants born small at term investigatingtheir brain metabolite alterations and itsassociation with neurodevelopmentaloutcome from 1-2 years of age. We havecarried out an extensive noninvasivecharacterization of brain MRS-detectablemetabolites in this population, usingestablished clinical protocols for mag-netic resonance acquisition and post-processing. Using a 3-T system enabledthe acquisition of good-qualityMRS data,with a low percentage of spectral datadiscarded (16%) and good fitting errorranges (estimated Cramér-Rao lowerbounds) for the major metaboliteschanges detected, namely: glutamate, 5�1%; glutamate þ glutamine, 6 � 1%;NAAt, 3 � 1%; and myo-inositol þglycine, 5 � 2%.

In conclusion, brain MRS detectschanges in small fetuses at 1 year of age.Metabolite alterations showed significantassociations with neurodevelopmentaltest scores, highlighting their potential asimage biomarkers to identify infants atrisk for an adverse neurological outcome.The pattern of metabolic changes wasdifferent to that previously reported interm fetuses, supporting the existenceand the need to clarify the developmentaltrajectories of abnormal neuro-development with perinatal origin. -

ACKNOWLEDGMENTS

The authors would like to thank Ms AngelaArranz, Dr Leandro Kesselman, Dr EduardoPosadas, and Dr Lucia Vazquez for their helporganizing the clinical information for this study.

REFERENCES

1. Marsal K. Intrauterine growth restriction. CurrOpin Obstet Gynecol 2002;14:127-35.2. Baschat AA. Neurodevelopment followingfetal growth restriction and its relationship withantepartum parameters of placental dysfunc-tion. Ultrasound Obstet Gynecol 2011;37:501-14.

can Journal of Obstetrics & Gynecology 210.e8

Research Obstetrics ajog.org

3. Cruz-Martinez R, Figueras F, Hernandez-Andrade E, Oros D, Gratacos E. Fetal brainDoppler to predict cesarean delivery for non-reassuring fetal status in term small-for-gestational-age fetuses. Obstet Gynecol2011;117:618-26.4. Doctor BA, O’Riordan MA, Kirchner HL,Shah D, Hack M. Perinatal correlates andneonatal outcomes of small for gestational ageinfants born at term gestation. Am J ObstetGynecol 2001;185:652-9.5. Figueras F, Eixarch E,Meler E, et al. Small-for-gestational-age fetuses with normal umbilicalartery Doppler have suboptimal perinatal andneurodevelopmental outcome. Eur J ObstetGynecol Reprod Biol 2008;136:34-8.6. Figueras F, Gratacos E. Stage-basedapproach to the management of fetal growthrestriction. Prenat Diagn 2014;34:655-9.7. Figueras F, Gratacos E. Update on the diag-nosis and classification of fetal growth restrictionand proposal of a stage-based managementprotocol. Fetal Diagn Ther 2014;36:86-98.8. Figueras F, Oros D, Cruz-Martinez R, et al.Neurobehavior in term, small-for-gestational ageinfants with normal placental function. Pediatrics2009;124:e934-41.9. Bassan H, Stolar O, Geva R, et al. Intrauterinegrowth-restricted neonates born at term orpreterm: how different? Pediatr Neurol 2011;44:122-30.10. Eixarch E, Meler E, Iraola A, et al. Neuro-developmental outcome in 2-year-old infantswho were small-for-gestational age term fetuseswith cerebral blood flow redistribution. Ultra-sound Obstet Gynecol 2008;32:894-9.11. Cruz-Martinez R, Figueras F, Oros D, et al.Cerebral blood perfusion and neurobehavioralperformance in full-term small-for-gestational-age fetuses. Am J Obstet Gynecol 2009;201:474.e1-7.12. Oros D, Figueras F, Cruz-Martinez R, et al.Middle versus anterior cerebral artery Dopplerfor the prediction of perinatal outcome andneonatal neurobehavior in term small-for-gestational-age fetuses with normal umbilicalartery Doppler. Ultrasound Obstet Gynecol2010;35:456-61.13. Story L, Damodaram MS, Allsop JM, et al.Brain metabolism in fetal intrauterine growth re-striction: a proton magnetic resonance spec-troscopy study. Am J Obstet Gynecol2011;205:483.e1-8.14. Sanz-Cortes M, Simoes RV, Bargallo N,Masoller N, Figueras F, Gratacos E. Proton mag-netic resonance spectroscopy assessment of fetalbrain metabolism in late-onset ’small for gesta-tional age’ versus ’intrauterine growth restriction’fetuses. Fetal Diagn Ther 2015;37:108-16.15. Egana-Ugrinovic G, Sanz-CortesM, Couve-Perez C, Figueras F, Gratacos E. Corpus cal-losum differences assessed by fetal MRI inlate-onset intrauterine growth restriction and itsassociation with neurobehavior. Prenat Diagn2014;34:843-9.16. Egana-Ugrinovic G, Sanz-Cortes M,Figueras F, Bargallo N, Gratacos E. Differences

210.e9 American Journal of Obstetrics & Gynecol

in cortical development assessed by fetal MRI inlate-onset intrauterine growth restriction. Am JObstet Gynecol 2013;209:126.e1-8.17. Egana-Ugrinovic G, Sanz-Cortes M,Figueras F, Couve-Perez C, Gratacos E. FetalMRI insular cortical morphometry and its asso-ciation with neurobehavior in late-onset small forgestational age fetuses. Ultrasound ObstetGynecol 2014;44:322-9.18. Sanz-Cortes M, Ratta GA, Figueras F, et al.Automatic quantitative MRI texture analysis insmall-for-gestational-age fetuses discriminatesabnormal neonatal neurobehavior. PLoS One2013;8:e69595.19. Sanz-Cortes M, Figueras F, Bonet-Carne E,et al. Fetal brain MRI texture analysis identifiesdifferent microstructural patterns in adequateand small for gestational age fetuses at term.Fetal Diagn Ther 2013;33:122-9.20. Sanz-Cortes M, Egana-Ugrinovic G,Zupan R, Figueras F, Gratacos E. Brainstemand cerebellar differences and their associationwith neurobehavior in term small-for-gestational-age fetuses assessed by fetalMRI. Am J Obstet Gynecol 2014;210:452.e1-8.21. Royal College of Obstetrics and Gynaecol-ogy. The investigation and management of thesmall-for-gestational-age fetus. Green-TopGuideline 31. London, UK: Royal College ofObstetrics and Gynaecology; 2002.22. Arduini D, Rizzo G. Normal values of pulsa-tility index from fetal vessels: a cross-sectionalstudy on 1556 healthy fetuses. J Perinat Med1990;18:165-72.23. American College of Obstetricians and Gy-necologists Committee on Obstetric Practice.Utility of antepartum umbilical artery Dopplervelocimetry in intrauterine growth restriction.ACOGCommittee opinion no. 188 (replaces no.116, November 1992). Int J Gynaecol Obstet1997;59:269-70.24. Figueras F, Meler E, Iraola A, et al.Customized birthweight standards for a Spanishpopulation. Eur J Obstet Gynecol Reprod Biol2008;136:20-4.25. Robinson HP, Fleming JE. A critical evalua-tion of sonar “crown-rump length” measure-ments. Br J Obstet Gynaecol 1975;82:702-10.26. Gomez O, Figueras F, Fernandez S, et al.Reference ranges for uterine artery mean pul-satility index at 11-41 weeks of gestation. Ul-trasound Obstet Gynecol 2008;32:128-32.27. Kreis R. Issuesof spectral quality in clinical 1H-magnetic resonance spectroscopy andagalleryofartifacts. NMR Biomed 2004;17:361-81.28. Provencher SW. Automatic quantitation oflocalized in vivo 1H spectra with LCModel. NMRBiomed 2001;14:260-4.29. Cavassila S, Deval S, Huegen C, vanOrmondt D, Graveron-Demilly D. Cramer-Raobounds: an evaluation tool for quantitation. NMRBiomed 2001;14:278-83.30. Govindaraju V, Young K, Maudsley AA.Proton NMR chemical shifts and coupling con-stants for brain metabolites. NMR Biomed2000;13:129-53.

ogy AUGUST 2015

31. Bayley N. Bayley scales of infant and toddlerdevelopment. San Antonio, TX: Harcourtassessment; 2006.32. Anderson PJ, De Luca CR, Hutchinson E,Roberts G, Doyle LW. Underestimation ofdevelopmental delay by the new Bayley-III scale.Arch Pediatr Adolesc Med 2010;164:352-6.33. Carrascosa Lezcano A, FerrandezLongas A, Yeste Fernandez D, et al. Spanishcross-sectional growth study 2008, part I:weight and height values in newborns of 26-42weeks of gestational age [in Spanish]. An Pediatr(Barc) 2008;68:544-51.34. Terpstra M, Henry PG, Gruetter R. Mea-surement of reduced glutathione (GSH) inhuman brain using LCModel analysis ofdifference-edited spectra. Magn Reson Med2003;50:19-23.35. Terpstra M, Vaughan TJ, Ugurbil K, Lim KO,Schulz SC, Gruetter R. Validation of glutathionequantitation from STEAM spectra against edited1H NMR spectroscopy at 4T: application toschizophrenia. MAGMA 2005;18:276-82.36. Kendall GS, Melbourne A, Johnson S, et al.White matter NAA/Cho and Cho/Cr ratios atMR spectroscopy are predictive of motoroutcome in preterm infants. Radiology2014;271:230-8.37. Bapat R, Narayana PA, Zhou Y, Parikh NA.Magnetic resonance spectroscopy at term-equivalent age in extremely preterm infants:association with cognitive and language devel-opment. Pediatr Neurol 2014;51:53-9.38. Van Kooij BJ, Benders MJ, Anbeek P, VanHaastert IC, De Vries LS, Groenendaal F. Cere-bellar volume and proton magnetic resonancespectroscopy at term, and neurodevelopment at2 years of age in preterm infants. Dev Med ChildNeurol 2012;54:260-6.39. Bejjani A, O’Neill J, Kim JA, et al. Elevatedglutamatergic compounds in pregenual anteriorcingulate in pediatric autism spectrum disorderdemonstrated by 1H MRS and 1H MRSI. PLoSOne 2012;7:e38786.40. Yuksel C, Ongur D. Magnetic resonancespectroscopy studies of glutamate-related ab-normalities in mood disorders. Biol Psychiatry2010;68:785-94.41. Gigante AD, Bond DJ, Lafer B, Lam RW,Young LT, Yatham LN. Brain glutamate levelsmeasured by magnetic resonance spectros-copy in patients with bipolar disorder: a meta-analysis. Bipolar Disord 2012;14:478-87.42. Altabella L, Zoratto F, Adriani W, Canese R.MR imaging-detectable metabolic alterations inattention deficit/hyperactivity disorder: frompreclinical to clinical studies. AJNR Am J Neu-roradiol 2014;35:S55-63.43. Carvalho C, Santos MS, Baldeiras I,Oliveira CR, Seica R, Moreira PI. Chronic hyp-oxia potentiates age-related oxidative imbalancein brain vessels and synaptosomes. Curr Neu-rovasc Res 2010;7:288-300.44. Gawryluk JW, Wang JF, Andreazza AC,Shao L, Young LT. Decreased levels of gluta-thione, the major brain antioxidant, in post-mortem prefrontal cortex from patients with

ajog.org Obstetrics Research

psychiatric disorders. Int J Neuro-psychopharmacol 2011;14:123-30.45. Ghanizadeh A, Akhondzadeh S, HormoziM,Makarem A, Abotorabi-Zarchi M, Firoozabadi A.Glutathione-related factors and oxidative stressin autism, a review. Curr Med Chem 2012;19:4000-5.46. Xian YF, Su ZR, Chen JN, et al. Iso-rhynchophylline improves learning and memoryimpairments induced by D-galactose in mice.Neurochem Int 2014;76C:42-9.47. Mehta KD, Mehta AK, Halder S, Khanna N,Tripathi AK, Sharma KK. Protective effect ofmelatonin on propoxur-induced impairment ofmemory and oxidative stress in rats. EnvironToxicol 2014;29:705-13.48. Silverstone PH,McGrathBM,KimH. Bipolardisorder and myo-inositol: a review of the mag-netic resonance spectroscopy findings. BipolarDisord 2005;7:1-10.49. Savchev S, Sanz-Cortes M, Cruz-Martinez R, et al. Neurodevelopmental outcomeof full-term small-for-gestational-age infants withnormal placental function. Ultrasound ObstetGynecol 2013;42:201-6.50. Batalle D, Eixarch E, Figueras F, et al.Altered small-world topology of structuralbrain networks in infants with intrauterinegrowth restriction and its association withlater neurodevelopmental outcome. Neuro-image 2012;60:1352-66.51. Hammerness P, Biederman J, Petty C,Henin A, Moore CM. Brain biochemical effects of

methylphenidate treatment using proton mag-netic spectroscopy in youthwith attention-deficithyperactivity disorder: a controlled pilot study.CNS Neurosci Ther 2012;18:34-40.52. Brand A, Richter-Landsberg C, Leibfritz D.Multinuclear NMR studies on the energy meta-bolism of glial and neuronal cells. Dev Neurosci1993;15:289-98.53. Xiang Y, Li Y, Zhang Z, et al. Nervegrowth cone guidance mediated by Gprotein-coupled receptors. Nat Neurosci2002;5:843-8.54. Joseph SK, Williamson JR. Inositol poly-phosphates and intracellular calcium release.Arch Biochem Biophys 1989;273:1-15.55. Ohnishi T, Murata T, Watanabe A, et al.Defective craniofacial development and brainfunction in a mouse model for depletion ofintracellular inositol synthesis. J Biol Chem2014;289:10785-96.56. Sanz-Cortes M, Carbajo RJ, Crispi F,Figueras F, Pineda-Lucena A, Gratacos E.Metabolomic profile of umbilical cord bloodplasma from early and late intrauterine growthrestricted (IUGR) neonates with and withoutsigns of brain vasodilation. PLoS One 2013;8:e80121.57. Crispi F, Figueras F, Cruz-Lemini M,Bartrons J, Bijnens B, Gratacos E. Cardiovas-cular programming in children born small forgestational age and relationship with prenatalsigns of severity. Am J Obstet Gynecol2012;207:121.e1-9.

AUGUST 2015 Americ

58. Dubois J, Dehaene-Lambertz G, Kulikova S,Poupon C, Huppi PS, Hertz-Pannier L. The earlydevelopment of brain white matter: a review ofimaging studies in fetuses, newborns and in-fants. Neuroscience 2014;276C:48-71.59. Maulik PK, Darmstadt GL. Community-based interventions to optimize early childhooddevelopment in low resource settings.J Perinatol 2009;29:531-42.60. Vanderveen JA, Bassler D, Robertson CM,Kirpalani H. Early interventions involving parentsto improve neurodevelopmental outcomes ofpremature infants: a meta-analysis. J Perinatol2009;29:343-51.61. Isaacs EB, Fischl BR, Quinn BT, ChongWK,Gadian DG, Lucas A. Impact of breast milk onintelligence quotient, brain size, andwhitematterdevelopment. Pediatr Res 2010;67:357-62.62. Kreis R, Hofmann L, Kuhlmann B,Boesch C, Bossi E, Huppi PS. Brain metabolitecomposition during early human brain develop-ment as measured by quantitative in vivo 1Hmagnetic resonance spectroscopy. MagnReson Med 2002;48:949-58.63. Pouwels PJ, Brockmann K, Kruse B, et al.Regional age dependence of human brain me-tabolites from infancy to adulthood as detectedby quantitative localized proton MRS. PediatrRes 1999;46:474-85.64. MescherM,Merkle H, Kirsch J, GarwoodM,Gruetter R. Simultaneous in vivo spectral editingand water suppression. NMR Biomed 1998;11:266-72.

an Journal of Obstetrics & Gynecology 210.e10

Research Obstetrics ajog.org

APPENDIX

SUPPLEMENTARY FIGURECorrelation between estimated infant brain GSH levels and fetal birthweight

Linear fittings (straight line) and SD intervals (dashed lines). A, Glutathione (GSH) (rho ¼ 0.281; P ¼ .046). B, GSH/creatine (rho ¼ 0.307; P ¼ .018).

Simões. Brain MRS changes in infants born small at term. Am J Obstet Gynecol 2015.

210.e11 American Journal of Obstetrics & Gynecology AUGUST 2015