ARTICLE IN PRESS

THE JOURNAL OF PEDIATRICS • www.jpeds.com ORIGINAL

ARTICLES
Multiple Postnatal Infections in Newborns Born Preterm Predict Delayed
Maturation of Motor Pathways at Term-Equivalent Age with Poorer Motor
Outcomes at 3 Years
Torin J. A. Glass, MBBCh1,2, Vann Chau, MD1,2,3, Ruth E. Grunau, PhD3,4, Anne Synnes, MDCM, MHSc3,4, Ting Guo, PhD1,2,
Emma G. Duerden, PhD1,2, Justin Foong, MSc2, Kenneth J. Poskitt, MDCM5, and Steven P. Miller, MAS, MDCM1,2,3

Objectives To evaluate whether the number of postnatal infections is associated with abnormal white matter matu-
ration and poorer motor neurodevelopmental outcomes at 36 months of corrected age.
Study design A prospective longitudinal cohort study was undertaken of 219 newborns born preterm at 24-32
weeks of gestational age recruited between 2006 and 2013 with magnetic resonance imaging of the brain both
early in life and at term-equivalent age. Postnatal infection was defined as any clinical infection or positive culture
≥72 hours after birth. White matter maturation was assessed by magnetic resonance spectroscopic imaging, mag-
netic resonance diffusion tensor imaging, and tract-based spatial statistics. Neurodevelopmental outcomes were
assessed in 175 (82% of survivors) infants with Bayley Scales of Infant and Toddler Development-III composite
scores and Peabody Developmental Motor Scales at 35 months of corrected age (IQR 34-37 months). Infection
groups were compared via the Fisher exact test, Kruskal–Wallis test, and generalized estimating equations.
Results Of 219 neonates born preterm (median gestational age 27.9 weeks), 109 (50%) had no postnatal infec-
tion, 83 (38%) had 1 or 2 infections, and 27 (12%) had ≥3 infections. Infants with postnatal infections had more
cerebellar hemorrhage. Infants with ≥3 infections had lower N-acetylaspartate/choline in the white matter and basal
ganglia regions, lower fractional anisotropy in the posterior limb of the internal capsule, and poorer maturation of
the corpus callosum, optic radiations, and posterior limb of the internal capsule on tract-based spatial statistics analy-
sis as well as poorer Bayley Scales of Infant and Toddler Development-III (P = .02) and Peabody Developmental
Motor Scales, Second Edition, motor scores (P < .01).
Conclusions In newborns born preterm, ≥3 postnatal infections predict impaired development of the motor path-
ways and poorer motor outcomes in early childhood. (J Pediatr 2017;■■:■■-■■).

ewborns born at <32 weeks of gestational age (GA) are at a substantial risk of postnatal infection, with 20%-65% of

N newborns suffering from at least a single infection during this period of significant brain development.1-3 In a large
epidemiologic study, postnatal infection was found to double the risk of motor impairment and cerebral palsy and
greatly increase the risk of cognitive impairment.1 Similarly, a study of infants born very preterm followed to 9 years of age
showed that infants who had postnatal infection were more likely to have poorer motor development, cognitive delay, school
delay, attention–deficit hyperactivity disorder, and other mental health disorders.4 Although white matter injury is a known
complication of postnatal infection, the majority of infants with postnatal infec-
tion do not have punctate white matter injury on clinical neuroimaging.1,5-7 Ex-
perimental models of white matter injury with hypoxia–ischemia show inflammation
From the 1Department of Pediatrics (Neurology),
University of Toronto and the Hospital for Sick Children,
Toronto, Ontario; 2Neurosciences & Mental Health,
SickKids Research Institute, Toronto, Ontario, Canada;
Bayley-III Bayley Scales of Infant and Toddler Development-III 3BC Children’s Hospital Research Institute, Vancouver,
BPD Bronchopulmonary dysplasia British Columbia; 4Department of Pediatrics
CA Corrected age (Neonatology), University of British Columbia and BC
Women’s Hospital and Health Centre, Vancouver, British
Cho Choline Columbia; and 5Department of Radiology, University of
DTI Diffusion tensor imaging British Columbia and BC Children’s Hospital, Vancouver,
FA Fractional anisotropy British Columbia, Canada

GA Gestational age Supported by the Canadian Institutes of Health Research

(CIHR) operating grant MOP-79262 (to S.M.) and MOP-
IVH Intraventricular hemorrhage 86489 (to R.G.). S.M. receives support from the Bloorview
MRI Magnetic resonance imaging Children’s Hospital Chair in Pediatric Neuroscience. R.G.
MRSI Magnetic resonance spectroscopic imaging holds a Senior Scientist Salary Award from the BC
Children’s Hospital Research Institute. Research funding
NAA N-acetylaspartate for T.G. is provided by the University of British Columbia
PDMS-2 Peabody Developmental Motor Scales, Second Edition clinician investigator program.
PLIC Posterior limb of the internal capsule Portions of this study were presented at the Pediatric
PMA Postmenstrual age Academic Societies annual meeting, May 6-9, 2017, San
Francisco, California.
RR Risk ratio
TBSS Tract-based spatial statistics 0022-3476/$ - see front matter. © 2017 Elsevier Inc. All rights
TEA Term-equivalent age reserved.
https://doi.org10.1016/j.jpeds.2017.12.041

FLA 5.5.0 DTD ■ YMPD9674_proof ■ February 2, 2018

THE JOURNAL OF PEDIATRICS • www.jpeds.com
to have additive effects on the injury present.8 Microscopic white
matter injury, present on neuropathologic studies, has been
shown to be more widespread in the brains of infants with mac-
roscopic white matter injury, suggesting a more diffuse brain
injury may be present but below the resolution of current clini-
cal magnetic resonance imaging (MRI) techniques.9 Chau et al
reported that newborns born very preterm exposed to infec-
tion had reduced measures of white matter development on
MRI, even when they adjusted for white matter injury, which
was most prominent in brain regions important for motor and
cognitive development.6 This impairment in the develop-
ment of the white matter may reflect cell gliosis from injury
to the oligodendrocyte precursor cell following hypoxic–
ischemic and inflammatory events in which the brains of the
infants born preterm may be primed for further insults, in-
creasing their vulnerability to injury.10-12 Our aim was to in-
vestigate the impact of multiple postnatal infections on the
white matter development and outcomes of newborns born
very preterm with the hypothesis that a greater number of in-
fections would be associated with delayed white matter de-
velopment and poorer motor outcomes at 36 months of
corrected age (CA) compared with neonates without infection.
Methods
The study was approved by the University of British Columbia/
Children’s and Women’s Health Centre of British Columbia
Research Ethics Board, and informed consent was obtained
from parents/guardians before recruitment. Neonates 24-32
weeks of GA were recruited into a prospective longitudinal
cohort study at British Columbia Women’s Hospital from April
2006 to September 2013. Infants were excluded if they had clini-
cal evidence of a congenital malformation or syndrome, con-
genital infection, or evidence on ultrasound scan of a large
parenchymal hemorrhagic infarction (>2 cm), as these con-
ditions are strongly predictive of neurodevelopmental impair-
ments or early mortality. This cohort has been described
previously to address different hypotheses.6,13-17
Characteristics of infection were collected by systematic chart
review. Cultures that had multiple isolates of organisms con-
sistent with contamination were excluded. Culture-positive in-
fection was defined as any positive blood, urine, tracheal
aspirate, and/or cerebrospinal fluid culture associated with ≥5
days of antibiotic therapy; “clinical-only infection” was defined
as any instance in which there was a clinical concern for in-
fection with negative cultures in which the antibiotic treat-
ment duration was ≥5 days. A positive culture with the same
organism in separate locations or cultures at the same time was
considered a single infection. A positive tracheal aspirate culture
required a positive culture and ≥4 white blood cells per field.
An “early infection” was defined as any infection <72 hours
of postnatal age and “postnatal infection” as any infection ≥72
hours after birth with the number of infections calculated based
on the total number of postnatal infections up to 40 weeks of
postmenstrual age (PMA). Other clinical characteristics were
collected via chart review: histologic chorioamnionitis as con-
firmed by clinical pathology assessment, hypotension as any
2 Glass et al
FLA 5.5.0 DTD ■ YMPD9674_proof ■ February 2, 2018
Volume ■■ • ■■ 2017
treatment for low blood pressure, patent ductus arteriosus as
a defect requiring pharmacologic or surgical treatment, bron-
chopulmonary dysplasia (BPD) as requirement for oxygen
therapy beyond 36 weeks of PMA, and necrotizing enteroco-
litis as stages 2 and 3 of the Bell criteria.18
MRI of the Brain
MRI was performed early in life when neonates were clini-
cally stable and again at term-equivalent age (TEA), all without
pharmacologic sedation. At both time points, MRIs were carried
out on a Siemens 1.5 Tesla Avanto scanner using an MR-
compatible isolette (Lammers Medical Technology, Luebeck,
Germany) and specialized neonatal head coil (Advanced
Imaging Research, Cleveland, Ohio). Three-dimensional coronal
volumetric T1-weighted and axial fast-spin echo T2-weighted
images were taken. An experienced neuroradiologist, blinded
to the participant’s medical history, reviewed the images and
recorded the severity of white matter injury, intraventricular
hemorrhage (IVH), ventriculomegaly, and cerebellar hemor-
rhage according to scales previously described in this cohort.5,19
The most severe injury score seen on the preterm or term scan
was used in the structural MRI analysis, as the greatest sever-
ity of injury is most likely to adversely impact developmental
outcomes and to be associated with greater amounts of mi-
croscopic white matter injury not seen on MRI. White matter
injury volumes were calculated on the T1-weighted images with
voxels of abnormal T1 shortening identified as white matter
injury, reviewed by 2 neonatal neurologists, then manually seg-
mented with simultaneous coronal, sagittal, and axial views of
the brain via Display software (Montreal Neurology Institute
and Hospital, Montreal, Canada; http://www.bic.mni.mcgill.ca/
software/Display) as previously reported.20
Magnetic Resonance Spectroscopic Imaging
(MRSI)
MRSI was used as a measure of neuronal maturation using
quantitative metabolite ratios in 6 anatomical regions: the an-
terior, central and posterior white matter, the caudate, lenti-
form nucleus, and thalamus. To reflect the overall metabolism
of the regions, we analyzed the white matter regions as an
average of the 3 white matter regions with the basal ganglia
region analyzed as an average of the caudate, lentiform nucleus,
and thalamus regions.
Diffusion Tensor Imaging (DTI)
DTI is a measure of water diffusion in an ellipsoid space within
each 3-dimensional voxel of the MR image. Fractional anisot-
ropy (FA), the average directionality of diffusion, increases with
white matter maturation, reflecting the maturation of the oli-
godendrocyte lineage and early myelination.21,22 DTI param-
eters of FA, l1, l2, and l3 were acquired with a multirepetition,
single-shot echo planar sequence and excluded if significant
motion artifact was present.5 Region of interest analyses were
placed manually in 7 white matter anatomical regions (ante-
rior, central and posterior white matter regions, optic radia-
tions, splenium of the corpus callosum, genu of the corpus
callosum, and posterior limb of the internal capsule [PLIC])
■■ 2017
and 3 deep gray matter regions (caudate, lentiform nucleus,
and thalamus).
Tract-based spatial statistics (TBSS) was performed via func-
tional MRI of the brain software library (FSL; https://
fsl.fmrib.ox.ac.uk/fsl/fslwiki/FSL) as previously described.15,17
A diffusion tensor model was fit to the data at each voxel to
calculate the voxelwise FA with the spatial location of altera-
tions in diffusion measures done using the TBSS pipeline.23
The TBSS data were projected onto a mean FA tract skeleton
with the use of age-appropriate templates (preterm scans: 27-
29 weeks, 30-33 weeks; 34-36 weeks; term scans: 37-41 weeks,
≥42 weeks), which were then used to apply voxelwise regres-
sion cross-subject analysis. Cluster-size thresholding was applied
to the data in which the size of the cluster was determined by
500 permutations by the use of Randomise v.2.9 (FMRIB, Uni-
versity of Oxford, Oxford, UK) within FSL. A threshold of
P < .05 (95% percentile of the distribution) was set for the clus-
ters and corrected for multiple comparisons across space in
determining the association of FA and ≥3 infections across sub-
jects within each of the age-based templates and projected on
the white matter skeleton.
Developmental Follow-Up
At 36 months of CA, neurodevelopment was assessed with the
Bayley Scales of Infant and Toddler Development-III (Bayley-
III) cognitive, language, and motor composite scores and
Peabody Developmental Motor Scales, Second Edition (PDMS-
2), all with a mean of 100 and SD of 15.24,25 In addition to the
Bayley-III motor composite, the PDMS-2 total, gross, and fine
motor quotient values were used in the analyses as a more
robust assessment of motor impairment at 36 months of CA.25
Assessments were carried out by qualified therapists blinded
to the imaging findings of the participants. Socioeconomic
status was classified by the self-reported number of years of
maternal education. Cerebral palsy was defined as a con-
firmed diagnosis made by an experienced pediatrician by the
36-month assessment.
Statistical Analyses
Statistical analysis was performed with Stata V.14.2 (StataCorp,
College Station, Texas).26 Clinical and imaging characteristics
were compared with the Fisher exact for categorical and Kruskal–
Wallis tests for continuous data with a statistical significance
threshold of P < .05. The association between postnatal in-
fections and other clinical variables with neurodevelopmental
outcomes was tested with univariate logistic regression. The
mean values, averaged bilaterally, of FA and N-acetylaspartate
(NAA)/choline (Cho) were compared between neonates with
and without postnatal infection, in a generalized least squares
regression model for repeated measures, adjusting for PMA at
MRI scan and multiple regions of interest with a P < .05. A
log-transformed outcome variable for the NAA/Cho was used
to determine the percentage differences of the MR measures.6
An interaction term was examined in describing the imaging
relationships of postnatal infection modified by PMA at MRI
and considered P < .1 as significant. The 36-month outcomes
were compared with univariate analysis for multiple infec-
Multiple Postnatal Infections in Newborns Born Preterm Predict Delayed Maturation of Motor Pathways at
Term-Equivalent Age with Poorer Motor Outcomes at 3 Years
FLA 5.5.0 DTD ■ YMPD9674_proof ■ February 2, 2018
ORIGINAL ARTICLES
tions. An unadjusted risk ratio (RR) and adjusted Odds Ratio
(aOR) were calculated in assessing the 36-month outcomes as-
sociation with the groups of infections.
Results
Of the 234 neonates born 24-32 weeks of gestation recruited
into the cohort, 219 (94%) completed at least 1 MRI with a
median birth GA of 27.9 weeks (IQR 26.0-29.7 weeks). Early
MRI scans were completed at a median 32.1 weeks (IQR 30.4-
34 weeks), with TEA MRI scans in 184 (84%) at median 40.2
weeks (IQR 38.7-42.0 weeks). TEA MRIs were not done in 4
infants who died and in 31 infants who withdrew from the
study or whom did not return for a TEA MRI.
Of the 219 infants, 110 (50%) had 1 or more instances of
postnatal infection, and 109 (50%) had no infections in the
postnatal period. Because of the small number in the “clinical-
only infection” group, we grouped them with “culture-positive
infection” for the analysis and thus classified each infection event
as a “postnatal infection.” There were 54 (25%) infants with
1 postnatal infection, 29 (13%) had 2 infections, 19 (9%) had
3, 7 (3%) had 4, and 1 (0.5%) had 5 infections. Infants were
grouped according to the number of infections, with ≥3 in-
fections included together in the analysis due to the small
numbers in the 4- and 5-infection groups. One and two in-
fections were presented together in the Results section, as there
were few differences between the groups regarding clinical char-
acteristics, neuroimaging, and motor outcomes (Table I). There
were 46 (21%) infants with early infection, of whom 21 had
no other infection, and 2 who had positive cultures (1 sepsis,
1 lower respiratory culture), which were included in the sep-
ticemia and lower respiratory analyses, respectively. In total,
there were 202 distinct episodes of postnatal infection among
the total of 110 (50%) infants with infections. There was a
greater rate of septicemia and lower respiratory infections in
the ≥3 infections group as well as a greater rate of infection
with coagulase-negative Staphylococcus and Enterobacter species
organisms compared with the 1- or 2-infection groups (Table II;
available at www.jpeds.com). Neonatal characteristics more
common among those with greater numbers of infections in-
cluded lower GA at birth, lower birth weight, lower birth length,
smaller head circumference at birth, hypotension, patent ductus
arteriosus, BPD, and necrotizing enterocolitis stage ≥2 (Table I).
Magnetic Resonance Imaging
White matter injury was present on the first MRI in 31% of
infants (33 mild, 23 moderate, 12 severe) with 6 new in-
stances of white matter injury on the second scan (4 mild, 1
moderate, 1 severe). There were 74 cases of grades 2-4 IVH
and 14 of ventriculomegaly on the first scan, with 57 IVH
grades 2-4 and 8 ventriculomegaly present on the TEA scan.
Cerebellar hemorrhage was present in 28 subjects’ first scan,
with 17 present at TEA. At least 1 infection occurred before
the first MRI in 99 infants (90%). There was no increase in
white matter injury seen, either mild or moderate/severe with
increased numbers of infection, with no increase in white matter
injury when separate analysis was done for the different types
3
THE JOURNAL OF PEDIATRICS • www.jpeds.com Volume ■■

Table I. Demographics, clinical characteristics, and MRI findings of newborns 24-32 weeks of GA compared by number
of infections
Clinical and imaging No infection One infection Two infections Three or more infections
characteristics N = 109 N = 54 N = 29 N = 27 P value
Male 54 (50) 30 (56) 10 (35) 19 (70) .12
GA at birth, wk 29.4 (27.7-31.1) 27.3 (25.9-28.6) 25.6 (25.0-27.1) 25.7 (24.9-26.6) <.01
Birth weight, g 1190 (1020-1376) 909 (789-1190) 835 (663-950) 755 (630-896) <.01
Birth length, cm 38 (35-40) 35 (34-39) 33 (31-35) 33 (31-35) <.01
Birth head circumference, cm 27 (26-28) 25 (23-27) 24 (22-25) 24 (22-25) <.01
Twin birth 41 (38) 19 (35) 7 (24) 9 (33) .69
Antenatal MgSO4 24 (22) 13 (24) 5 (17) 5 (19) .97
Antenatal corticosteroids 11 (10) 7 (13) 2 (7) 4 (15) .75
Histologic chorioamnionitis 33 (31) 22 (42) 14 (48) 8 (31) .18
Early infection* 21 (19) 11 (20) 7 (24) 7 (26) .69
Hypotension 21 (19) 23 (43) 14 (48) 23 (85) <.01
Patent ductus arteriosus 35 (32) 27 (50) 21 (72) 25 (93) <.01
BPD 5 (5) 8 (15) 9 (31) 17 (63) <.01
NEC stage ≥2 1 (1) 1 (2) 2 (7) 4 (15) <.01
IVH grades 2-4 34 (32) 12 (34) 13 (54) 9 (35) .67
White matter injury 38 (35) 14 (26) 8 (28) 8 (31) .49
White matter injury volume, mm 35.8 (16.3-272.4) 41 (11-275) 53 (5-892) 17.9 (7.6-83.6) .60
Ventriculomegaly 20 (21) 11 (26) 6 (25) 11 (50) .07
Cerebellar hemorrhage 5 (5) 6 (17) 7 (29) 7 (27) <.01

MgSO4, magnesium sulfate; NEC, necrotizing enterocolitis.

Number (%) or median (IQR).
*Any infection <72 h of life.

of infection and organisms (all P > .05). Similarly, IVH sever- natal period was significantly associated with poorer Bayley-
ity and ventriculomegaly were not associated with infection III motor composite score and PDMS-2 total, gross, and fine
(all P > .05) (Table I). Cerebellar hemorrhage was more motor scores but not language and cognitive composite scores
common among those infants with postnatal infection (Table I). (Figure 2 and Table III [available at www.jpeds.com]). Un-
adjusted RR and aOR assessments of Bayley-III scores ≤85 and
MRSI and DTI PDMS-2 ≤80, considered clinically significant impairments,
The NAA/Cho ratio was lower over time in those infants with showed that a greater number of infections was significantly
≥3 infections in the white matter (coefficient −1.4%; 95% CI associated with lower Bayley-III motor composite scores and
−2.3% to −0.5%; P < .01) and basal ganglia (coefficient −0.7%;
95% CI −2.5% to −0.1%; P = .03), adjusted for white matter
injury. Those with ≥3 infections had lower mean FA over time
in the PLIC (coefficient −0.005; 95% CI −0.002 to −0.008,
P < .01). There were no other regions with significant differ-
ences in the FA over time (all P ≥ .10). There were no signifi-
cant differences in FA between the different types of infection
or organisms cultured. The TBSS analysis of the preterm MRIs
30-34 weeks of PMA revealed only delayed FA located within
the posterior corpus callosum. In contrast, on the term scans
at 37-42 weeks of PMA, lower FA was more widespread and
involved the complete corpus callosum, the optic radiations,
and PLIC (Figure 1; available at www.jpeds.com).

Developmental Outcomes
Thirty-six month CA outcome scores were available in 175
(82% of survivors) infants (median 35 months, IQR 34-37
months). There was 1 subject who died between the TEA MRI
and the 33-month assessment, with the remainder lost to Figure 2. Bayley-III outcomes at 36 months of CA by infec-
follow-up. There were no differences in the rates of white matter tion groups; no infections (light grey), 1 or 2 infections (medium
injury, IVH grade ≥2, or neonatal clinical factors (all P > .05) grey), 3 or more infections (dark grey). Motor, language, and
in the follow-up and nonfollow-up groups. Three infants were cognitive composite scores are reflected by the shading indi-
unable to complete the testing due to severe impairment and cated in the legend. The middle line and box reflect the median
and IQR, respectively, with the 1.5 IQR reflected by the whis-
thus were assigned scores of 49 (3.3 SDs from the mean). On
kers and open circles as any outliers. **P < .05 (all others P > .05).
univariate analysis, a greater number of infections in the neo-
4 Glass et al

FLA 5.5.0 DTD ■ YMPD9674_proof ■ February 2, 2018

■■ 2017
Table IV. Unadjusted RR and aOR values for poor Bayley-III and PDMS-2 scores in infants with ≥3 infections
Developmental assessment score Unadjusted RR
Bayley-III Motor composite <85 2.53 (1.10-5.85)
Bayley-III Language composite <85 5.86 (1.34-25.6)
Bayley-III Cognitive composite <85 0.55 (0.14-2.11)
PDMS-2 Total Motor <80 3.06 (1.27-7.4)
PDMS-2 Gross Motor <80 2.07 (1.06-4.03)
PDMS-2 Fine Motor <80 1.69 (0.58-4.97)
Odds (95% CI).
*Adjusted for GA, maternal education, cerebellar hemorrhage, and white matter injury volume.
poorer performance on the PDMS-2 total and gross motor
testing when adjusted for potential confounding factors, with
less impairment in fine motor scores (Table IV). There was no
significant increase in the incidence of cerebral palsy at 36
months of CA with multiple infections (P > .05) (Table III).
Discussion
Using multimodal MRI methods and follow-up assessments
in a cohort of newborns born very preterm, we showed that
≥3 postnatal infections are associated with delays in brain matu-
ration, particularly within white matter fiber pathways and sub-
cortical nuclei implicated in motor function, and with poorer
motor outcomes at 36 months of CA. Findings were associ-
ated with alterations in brain maturation over time, with dif-
ferences most prevalent on the term-equivalent MRI scans
reflected in involvement of the PLIC, corpus callosum, and optic
radiations. This finding is in agreement with previous re-
search suggesting that poorer motor outcomes and cerebral
palsy are more likely in infants born preterm with reduced FA
in these regions, implicating these regions in motor develop-
ment and highlighting the utility of advanced MRI tech-
niques in predicting motor outcomes.15,27
Newborns born very preterm are at risk for multiple post-
natal infections due to many factors. The innate and adap-
tive immune systems that respond to infectious antigens are
significantly impaired in newborns born preterm, leaving the
infant more prone to infections.28,29 Infants born preterm also
have reduced transplacental maternal antibody transfer and
immature protein recognition receptors important for bacte-
rial antigen recognition and antimicrobial immune response.30,31
The consequence of this is an inappropriate inflammatory re-
sponse to infections with activation of Toll-like receptors in
the innate immune system that, when associated with hypoxia–
ischemia, can result in neural injury.32 This may also lead to
priming of the immune system for subsequent events mani-
fested by greater concentrations of inflammatory markers and
free radicals in the cerebrospinal fluid, as seen in infants with
white matter injury.33-35 Neuropathology analysis of infants with
severe white matter injury has showed that tumor necrosis
factor-a and interleukin-1b levels are elevated in infants with
infection, with greater levels compared with hypoxia–ischemia
alone, further supporting their combined effects.35
In a large epidemiologic study of >6000 infants born at ex-
tremely low birth weight (<1000 g), all types of infection, not
Multiple Postnatal Infections in Newborns Born Preterm Predict Delayed Maturation of Motor Pathways at
Term-Equivalent Age with Poorer Motor Outcomes at 3 Years
FLA 5.5.0 DTD ■ YMPD9674_proof ■ February 2, 2018
ORIGINAL ARTICLES
P value aOR* P value
.02 1.99 (1.2-3.2) <.01
<.01 1.68 (0.94-3.00) .08
.37 0.55 (0.16-1.83) .33
<.01 1.93 (1.21-3.09) <.01
.03 1.84 (1.18-2.85) <.01
.33 1.59 (0.88-2.90) .13
just sepsis or meningitis, increased the risk of severe cogni-
tive and motor impairment.1 We also found no major differ-
ence in the outcomes between the different types or organisms
of infection. We did not observe an increased risk for severe
impairment with increasing number of infections, with a low
rate of cerebral palsy overall. Rand et al showed that long-
term neurodevelopmental outcome at 9 years was more likely
to be delayed in infants with postnatal infection but with no
difference in severe motor outcomes.4 Furthermore, in a small
subset of infants, Rand et al also reported that ≥2 infections
increased the risk for motor impairment 2-fold (RR 5.7 vs 2.6)
in addition to increasing the risk of cognitive impairment (RR
2.1 vs 1.3).4 Additive effects are supported by the literature on
subsequent events, as synthesized by Khwaja and Volpe, in which
the brain of the infant born preterm is sensitized following an
initial event of infection or hypoxia–ischemia and has a greater
likelihood to be injured following subsequent events of in-
flammation and/or hypoxia–ischemia by a stimulus with a lower
threshold than was necessary for the initial response.8,36-38
The brain of the newborn born preterm also has an imma-
ture blood–brain barrier that makes it especially vulnerable
to hypoxia–ischemia injury, particularly of the oligodendro-
cyte precursor, the prominent cell injured in white matter
injury of the infant born preterm.11,39-41 In hypoxia–ischemia
models of preterm brain injury, a significant increase in
proinflammatory markers is seen, suggesting that even in the
absence of infection, a significant immune system response
occurs.42,43 Of our infants in the ≥3 infections group, 78%
had a coagulase-negative Staphylococcus infection, which is
not known to cause direct cerebral inflammation but occurs
frequently in conjunction with hypotension.44 In addition, in
our group of infants with ≥3 infections, 85% had at least a
single event of hypotension, as opposed to 45% and 19% in
those with 1 or 2 and no infections, respectively. Previous
work in rats that used a lipopolysaccharide injection with
hypoxia–ischemia injury model showed a greater duration
of hypoxia was required to induce injury in older rats than
younger rats, indicating the importance of the developmen-
tal window on the systemic response and the vulnerability of
the preterm brain to hypoxic–ischemic injury.45 How hypoxia–
ischemia and systemic inflammation interact in the genesis
of neonatal brain injury and dysmaturation, or gliosis, re-
quires further study.12,46
Another potential contributor to brain dysmaturation or
gliosis includes injury resulting from hyperoxia with hypocarbia.
BPD is a known complication of chronic ventilation in the
5
THE JOURNAL OF PEDIATRICS • www.jpeds.com
infant born preterm, with hyperoxia a well-established caus-
ative factor in the development of BPD.47 Of our infants with
≥3 infections in our cohort, 65% required long-term oxygen
therapy compared with 5% in the no-infection group.
Hyperoxia and hypocarbia are potent vasoconstrictors impor-
tant in the pressure–passive autoregulatory system of the
preterm brain and are independent risk factors in the devel-
opment of white matter injury.48 The impacts with which these
factors also contribute to the poorer outcomes seen in infants
with multiple infection also need further study.
Multiple postnatal infections have been shown to be asso-
ciated with progressive white matter injury on subsequent
MRIs.5,7 Within our cohort, we found no difference in the
rate or volume of MRI-detected white matter injury,
ventriculomegaly, or IVH and the numbers or types of infec-
tions. There was an increased rate of cerebellar hemorrhage
among those infants with ≥3 infections, which may reflect their
lower GA at birth, making them more vulnerable to cerebel-
lar hemorrhage. This has potential implications on the out-
comes analysis because cerebellar hemorrhage is associated with
poor neurologic outcomes.49 Adams et al, in a smaller subset
of the current cohort, showed that postnatal infection was as-
sociated with slower increase in corticospinal FA over time com-
pared with newborns not infected.16 Our study expands these
findings, showing that maturational delays in infants with ≥3
more infections included the corpus callosum and white matter
regions including the PLIC on both traditional DTI and TBSS
analyses with supportive findings on MRSI analyses.6
Follow-up at 36 months of CA revealed that infants with
≥3 infections had poorer motor development than infants with
fewer infections. Compared with other large cohorts of new-
borns born preterm with infection, we did not find increased
risk of cognitive impairment in those infants with greater
numbers of infections.1,4 Some differences in previous cohorts
may be due to improvements in clinical practices and neona-
tal intensive care unit care in recent decades. In addition, the
majority of infants in this cohort are from high-resource set-
tings in which better cognitive and language outcomes are
expected.50,51 The differences from previous literature, with a
predisposition for motor tracts and motor impairments seen
in our study, also may be a result of the changes seen in the
distribution of white matter injury, from PVL in older studies
to more diffuse white matter injury in contemporary studies
in which diffuse white matter injury results in less-pronounced
axonal changes.9 As the motor tracts are some of the first areas
to myelinate, they also are the most susceptible to injury in
the preterm brain.52,53
As described previously by Chau et al, many of the infants
who had multiple infections would have had later MRI scans
due to taking longer to become clinically stable, which would
result in differences in postnatal age at the MRI between the
infection groups.6 This would, however, favor the delayed images
appearing more mature in the multiple-infection groups, un-
derestimating the extent of the injury and time interaction. It
also should be considered that the “clinical-only” infections
could be the result of a virus, which would also have impli-
cations on white matter development, immune system devel-
6 Glass et al
FLA 5.5.0 DTD ■ YMPD9674_proof ■ February 2, 2018
Volume ■■
opment, and expression of proinflammatory markers. However,
we did not find neuroimaging abnormalities consistent with
known descriptions of viral infections, nor were there any posi-
tive viral cultures. Overall, our findings support the need for
further monitoring of infants with postnatal infections and con-
tinued assessment of how to improve the care of these infants.
Three or more postnatal infections are associated with delays
in MRI metrics of brain maturation, particularly in areas of
motor function, with poorer motor outcomes at 36 months
of CA. These results highlight the vulnerability of the preterm
brain to multiple postnatal infections and supports the need
to better understand the interaction of hypoxia–ischemia with
inflammation within the neonatal intensive care unit. Fur-
thermore, data suggest the need for a personalized approach
to infection control in those infants with 1 or 2 infections. Pre-
venting further infections in this vulnerable group of preterm
newborns may have the potential to improve outcomes. More
research is needed on the function and role of the develop-
ing immune system and the impact of hypoxia–ischemia and
infection on proinflammatory responses, as well as investi-
gating potential therapies to prevent the detrimental impacts
of multiple postnatal infections in the preterm newborn. ■
We thank the children and their parents who generously committed their
time and energy to this study. We also thank the therapists and physi-
cians of the Neonatal Follow-up Program at BC Women’s Hospital for
their invaluable work assessing these children. Finally, we thank Janet
Rigney, Sandra Belanger, RN, and Mark Chalmers, RRT, for their re-
search support and work on this study.
Submitted for publication Sep 15, 2017; last revision received Nov 21, 2017;
accepted Dec 15, 2017
Reprint requests: Steven P. Miller, MAS, MDCM, Department of Pediatrics
(Neurology), The Hospital for Sick Children, University of Toronto, 555
University Ave, Room 6546 Hill Wing, Toronto, ON, M5G 1X8, Canada. E-mail:
steven.miller@sickkids.ca
References
1. Stoll BJ, Hansen NI, Adams-Chapman I, Fanaroff AA, Hintz SR, Vohr B,
et al. Neurodevelopmental and growth impairment among extremely low-
birth-weight infants with neonatal infection. JAMA 2004;292:2357-65.
2. Adams-Chapman I, Stoll BJ. Neonatal infection and long-term
neurodevelopmental outcome in the preterm infant. Curr Opin Infect Dis
2006;19:290-7.
3. Orsi G, d’Ettorre G, Panero A, Chiarini F, Vullo V, Venditti M. Hospital-
acquired infection surveillance in a neonatal intensive care unit. Am J Infect
Control 2009;37:201-3.
4. Rand K, Austin N, Inder T, Bora S, Woodward L. Neonatal infection and
later neurodevelopmental risk in the very preterm infant. J Pediatr
2016;170:97-104.
5. Chau V, Poskitt KJ, McFadden DE, Bowen-Roberts T, Synnes A, Brant
R, et al. Effect of chorioamnionitis on brain development and injury in
premature newborns. Ann Neurol 2009;66:155-64.
6. Chau V, Brant R, Poskitt KJ, Tam EWY, Synnes A, Miller SP. Postnatal
infection is associated with widespread abnormalities of brain develop-
ment in premature newborns. Pediatr Res 2012;71:274-9.
7. Glass HC, Bonifacio SL, Chau V, Glidden D, Poskitt K, Barkovich AJ, et al.
Recurrent postnatal infections are associated with progressive white matter
injury in premature infants. Pediatrics 2008;122:299-305.
8. Eklind S, Mallard C, Leverin A-L, Gilland E, Blomgren K, Mattsby-
Baltzer I, et al. Bacterial endotoxin sensitizes the immature brain to hypoxic-
ischaemic injury. Eur J Neurosci 2001;13:1101-6.
■■ 2017
9. Buser JR, Maire J, Riddle A, Gong X, Nguyen T, Nelson K, et al. Ar-
rested preoligodendrocyte maturation contributes to myelination failure
in premature infants. Ann Neurol 2012;71:93-109.
10. Hagberg H, Mallard C. Effect of inflammation on central nervous system
development and vulnerability. Curr Opin Neurol 2005;18:117-23.
11. Wang L, Tu Y, Huang C, Ho C. JNK signaling is the shared pathway
linking neuroinflammation, blood-brain barrier disruption and
oligodendroglialapoptosis in the white matter injury of the immature brain.
J Neuroinflammation 2012;9:175.
12. Fleiss B, Gressens P. Tertiary mechanisms of brain damage: a new hope
for treatment of cerebral palsy? Lancet Neurol 2012;11:556-66.
13. Chau V, Synnes A, Grunau RE, Poskitt KJ, Brant R, Miller SP. Abnor-
mal brain maturation in preterm neonates associated with adverse de-
velopmental outcomes. Neurology 2013;81:2082-9.
14. Brummelte S, Grunau RE, Chau V, Poskitt KJ, Brant R, Vinall J, et al. Pro-
cedural pain and brain development in premature newborns. Ann Neurol
2012;71:385-96.
15. Duerden E, Foong J, Chau V, Branson H, Poskitt K, Grunau R, et al. Tract-
based spatial statistics in preterm-born neonates predicts cognitive and
motor outcomes at 18 months. AJNR Am J Neuroradiol 2015;36:1565-
71.
16. Adams E, Chau V, Poskitt KJ, Grunau RE, Synnes A, Miller SP.
Tractography-based quantitation of corticospinal tract development in
premature newborns. J Pediatr 2010;156:882-8.
17. Glass TJA, Chau V, Gardiner J, Foong J, Vinall J, Zwicker JG, et al. Severe
retinopathy of prematurity predicts delayed white matter maturation and
poorer neurodevelopment. Arch Dis Child Fetal Neonatal Ed
2017;102:F532-7.
18. Bell MJ, Ternberg JL, Feigin RD, Keating JP, Marshall R, Barton L, et al.
Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clini-
cal staging. Ann Surg 1978;187:1-7.
19. Miller SP, Ferriero DM, Leonard C, Piecuch R, Glidden DV, Partridge JC,
et al. Early brain injury in premature newborns detected with magnetic
resonance imaging is associated with adverse early neurodevelopmental
outcome. J Pediatr 2005;147:609-16.
20. Guo T, Duerden E, Adams E, Chau V, Branson H, Chakravarty M, et al.
Quantitative assessment of white matter injury in preterm neonates. Neu-
rology 2017;88:1-9.
21. Miller SP, Vigneron DB, Henry RG, Bohland MA, Ceppi-Cozzio C,
Hoffman C, et al. Serial quantitative diffusion tensor MRI of the prema-
ture brain: development in newborns with and without injury. J Magn
Reson Imaging 2002;16:621-32.
22. Drobyshevsky A, Song SK, Gamkrelidze G, Wyrwicz AM, Derrick M, Meng
F, et al. Developmental changes in diffusion anisotropy coincide with im-
mature oligodendrocyte progression and maturation of compound action
potential. J Neurosci 2005;25:5988-97.
23. Smith SM, Jenkinson M, Johansen-Berg H, Rueckert D, Nichols TE, Mackay
CE, et al. Tract-based spatial statistics: voxelwise analysis of multi-
subject diffusion data. Neuroimage 2006;31:1487-505.
24. Bayley N. The Bayley Scales of Infant Development III. New York (NY):
Psychological Corporation; 2005.
25. Folio MR, Fewell RR. Peabody Developmental Motor Scales (PDMS-2).
2nd ed. Austin (TX): PRO-ED; 2000.
26. StataCorp. Stata statistical software. Release 14. College Station (TX):
StataCorp LP; 2015.
27. Huppi PS, Murphy B, Maier SE, Zientara GP, Inder TE, Barnes PD, et al.
Microstructural brain development after perinatal cerebral white matter
injury assessed by diffusion tensor magnetic resonance imaging. Pediat-
rics 2001;107:455-60.
28. Cuenca A, Wynn J, Moldawer L, Levy O. Role of innate immunity in neo-
natal infection. Am J Perinatol 2013;30:105-12.
29. Lavoie P, Huang Q, Jolette E, Whalen M, Nuyt A, Audibert F, et al. Pro-
found lack of interleukin (IL)–12/IL-23p40 in neonates born early in ges-
tation is associated with an increased risk of sepsis. J Infect Dis
2010;202:1754-63.
30. Kan B, Razzaghian HR, Lavoie PM. An immunological perspective on neo-
natal sepsis. Trends Mol Med 2016;22:290-302.
Multiple Postnatal Infections in Newborns Born Preterm Predict Delayed Maturation of Motor Pathways at
Term-Equivalent Age with Poorer Motor Outcomes at 3 Years
FLA 5.5.0 DTD ■ YMPD9674_proof ■ February 2, 2018
ORIGINAL ARTICLES
31. Wynn J, Scumpia P, Winfield R, Delano M, Kelly-Scumpia K, Barker T,
et al. Defective innate immunity predisposes murine neonates to poor
sepsis outcome but is reversed by TLR agonists. Blood 2008;112:1750-8.
32. Lehnardt S, Massillon L, Follett P, Jensen FE, Ratan R, Rosenberg PA, et al.
Activation of innate immunity in the CNS triggers neurodegeneration
through a Toll-like receptor 4-dependent pathway. Proc Natl Acad Sci USA
2003;100:8514-9.
33. Ellison VJ, Mocatta TJ, Winterbourn CC, Darlow BA, Volpe JJ, Inder TE.
The relationship of CSF and plasma cytokine levels to cerebral white matter
injury in the premature newborn. Pediatr Res 2005;57:282-6.
34. Inder T, Mocatta T, Darlow B, Spencer C, Volpe JJ, Winterbourn C. El-
evated free radical products in the cerebrospinal fluid of VLBW infants
with cerebral white matter injury. Pediatr Res 2002;52:213-8.
35. Kadhim H, Tabarki B, Verellen G, De Prez C, Rona AM, Sebire G. In-
flammatory cytokines in the pathogenesis of periventricular leukomalacia.
Neurology 2001;56:1278-84.
36. Leviton A, Fichorova R, O’Shea T, Kuban K, Paneth N, Dammann O, et al.
Two-hit model of brain damage in the very preterm newborn: small for
gestational age and postnatal systemic inflammation. Pediatr Res
2013;73:362-70.
37. Hagberg H, Gressens P, Mallard C. Inflammation during fetal and neo-
natal life: implications for neurologic and neuropsychiatric disease in chil-
dren and adults. Ann Neurol 2012;71.
38. Khwaja O, Volpe JJ. Pathogenesis of cerebral white matter injury of pre-
maturity. Arch Dis Child Fetal Neonatal Ed 2008;93:F153-61.
39. Gilles FH, Leviton A, Kerr CS. Endotoxin leukoencephalopathy in the
telecephalon of the newborn kitten. J Neurol Sci 1976;27:183-91.
40. Back SA, Luo NL, Borenstein NS, Volpe JJ, Kinney HC. Arrested oligo-
dendrocyte lineage progression during human cerebral white matter de-
velopment: dissociation between the timing of progenitor differentiation
and myelinogenesis. J Neuropathol Exp Neurol 2002;61:197-211.
41. Back SA, Luo NL, Borenstein NS, Levine JM, Volpe JJ, Kinney HC. Late
oligodendrocyte progenitors coincide with the developmental window of
vulnerability for human perinatal white matter injury. J Neurosci
2001;21:1302-12.
42. Albertsson A, Bi D, Duan L, Zhang X, Leavenworth J, Qiao L, et al. The
immune response after hypoxia-ischemia in a mouse model of preterm
brain injury. J Neuroinflammation 2014;11.
43. Fleiss B, Tann C, Degos V, Sigaut S, Steenwinckel JV, Schang A, et al.
Inflammation-induced sensitization of the brain in term infants. Dev Med
Child Neurol 2015;57:17-28.
44. Mallard C, Wang X. Infection-induced vulnerability of perinatal brain
injury. Neurol Res Int 2012;2012:102153.
45. Eklind S, Mallard C, Arvidsson P, Hagberg H. Lipopolysaccharide induces
both a primary and a secondary phase of sensitization in the developing
rat brain. Pediatr Res 2005;58:112-6.
46. Back SA, Miller SP. Brain injury in premature neonates: a primary ce-
rebral dysmaturation disorder? Ann Neurol 2014;75:469-86.
47. Buczynski BW, Maduekwe ET, O’Reilly MA. The role of hyperoxia in the
pathogenesis of experimental BPD. Semin Perinatol 2013;37:69-78.
48. Shankaran S, Langer JC, Kazzi SN, Laptook AR, Walsh M, National In-
stitute of Child H, et al. Cumulative index of exposure to hypocarbia and
hyperoxia as risk factors for periventricular leukomalacia in low birth weight
infants. Pediatrics 2006;118:1654-9.
49. Tam E, Rosenbluth G, Rogers E, Ferriero D, Glidden D, Goldstein R, et al.
Cerebellar hemorrhage on MRI in preterm newborns associated with ab-
normal neurological outcome. J Pediatr 2011;158:245-50.
50. Cusson R. Factors influencing language development in preterm infants.
J Obstet Gynecol Neonatal Nurs 2003;32:402-9.
51. Gross S, Mettelman B, Dye T, Slagle T. Impact of family structure and
stability on academic outcome in preterm children at 10 years of age. Pe-
diatrics 2001;138:169-75.
52. Sie LT, van der Knaap MS, van Wezel-Meijler G, Valk J. MRI assessment
of myelination of motor and sensory pathways in the brain of preterm
and term-born infants. Neuropediatrics 1997;28:97-105.
53. Welker KM, Patton A. Assessment of normal myelination with mag-
netic resonance imaging. Semin Neurol 2012;32:15-28.
7
THE JOURNAL OF PEDIATRICS • www.jpeds.com Volume ■■

Figure 1. TBSS model for preterm (30-34 weeks of GA) and

term (37-42 weeks of GA) model for infants with ≥3 infec-
tions (blue) compared with the no-infection group on a white
matter (yellow) skeleton map. The number of infants with ≥3
infections is indicated by the ‘N =,’ with the denominator re-
flecting the total number of infants.

Table II. Infection location and organism characteristics divided by the number of infection groups
One infection Two infections Three or more infections
Infection location or organism N = 54 N = 29 N = 27 P value
Clinical-only infection 7 (13) 5 (17) 8 (30) .21
Septicemia 16 (30) 17 (59) 17 (63) <.01
Urinary tract infection 24 (44) 17 (59) 15 (56) .41
Lower respiratory infection 6 (11) 10 (34) 16 (59) <.01
Meningitis 1 (2) 1 (3) 0 (0) >.99
Coagulase-negative Staphylococcus 23 (43) 17 (59) 21 (78) .01
Enterococcus species 8 (15) 7 (24) 7 (26) .41
Escherichia coli 5 (9) 2 (7) 6 (22) .17
Klebsiella species 4 (7) 5 (17) 6 (22) .13
Enterobacter species 1 (2) 5 (17) 9 (33) <.01
Candida species 1 (2) 2 (7) 4 (15) .08
Other pathogens 7 (13) 11 (38) 9 (33) .02

Number (%).

Table III. Bayley-III and PDMS-2 outcomes at 36 months of CA compared by number of infections
No infection One infection Two infections Three or more infections
Developmental assessment and diagnosis N = 93 N = 42 N = 24 N = 22 P value
Bayley-III Motor composite score 103 (97-115) 103 (94-110) 100 (91-107) 93 (76-107) .02
Bayley-III Language composite score 109 (103-118) 112 (103-115) 106 (83-115) 100 (94-112) .18
Bayley-III Cognitive composite score 100 (95-110) 105 (95-105) 100 (95-105) 100 (90-110) .57
PDMS-2 Total Motor 96 (90-102) 94 (88-97) 93 (88-97) 87 (74-94) <.01
PDMS-2 Gross Motor 96 (89-102) 94 (87-98) 91 (85-96) 84 (72-94) <.01
PDMS-2 Fine Motor 100 (94-103) 94 (91-97) 97 (88-103) 93 (85-97) <.01
Cerebral palsy 5 (5) 3 (6) 3 (12) 1 (4) .70

Median (IQR) or number (%).

7.e1 Glass et al

FLA 5.5.0 DTD ■ YMPD9674_proof ■ February 2, 2018