+ MODEL

Pediatrics and Neonatology (2017) xx, 1e8

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: http://www.pediatr-neonatol.com

Original Article

Timing of sepsis is an important risk factor

for white matter abnormality in extremely
premature infants with sepsis
Ju Sun Heo a,b, Ee-Kyung Kim a,*, Young Hun Choi c,
Seung Han Shin a, Jin A Sohn a, Jung-Eun Cheon c, Han-Suk Kim a

a
Department of Pediatrics, Seoul National University College of Medicine, Seoul, South Korea
b
Department of Pediatrics, CHA Gangnam Medical Center, CHA University School of Medicine, Seoul,
South Korea
c
Department of Radiology, Seoul National University College of Medicine, Seoul, South Korea

Received Sep 13, 2016; received in revised form Mar 30, 2017; accepted Jul 28, 2017
Available online - - -

Key Words Background: Systemic infection is a major upstream mechanism for white matter abnormality
infant; (WMA). Our aim was to evaluate the risk factors for moderate-to-severe WMA in extremely pre-
extremely mature infants (gestational age < 28 weeks) with neonatal sepsis.
premature; Methods: Extremely premature infants with culture-proven sepsis between 2006 and 2015 in a
sepsis; tertiary neonatal intensive care unit were classified as having none-to-mild or moderate-to-
white matter severe WMA based on WM scores of brain magnetic resonance imaging at the term-
equivalent age. Various risk factors for WMA were analyzed.
Results: Sixty-three infants (87.5%) had none-to-mild WMA, and nine infants (12.5%) had
moderate-to-severe WMA. Multivariate logistic regression analysis revealed that postmenstrual
age (PMA) at sepsis diagnosis (OR: 0.640, 95% CI: 0.435e0.941, p Z 0.023) and PMA at sepsis
diagnosis <28 weeks (OR: 9.232, 95% CI: 1.020e83.590, p Z 0.048) were independently asso-
ciated with moderate-to-severe WMA. PMA at sepsis diagnosis had a significant negative corre-
lation with WM scores (r Z
0.243, p Z 0.039).
Conclusion: PMA at sepsis diagnosis might be an important risk factor for moderate-to-severe WMA
in extremely premature infants with postnatal sepsis, especially before PMA 28 weeks. Infants who
suffer from sepsis before PMA 28 weeks might need additional therapy for neuroprotection.
Copyright ª 2017, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/
4.0/).

* Corresponding author. Department of Pediatrics, Seoul National University College of Medicine, Seoul National University Children’s
Hospital, 101 Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. Fax: þ82 2 743 3455.
E-mail address: kimek@snu.ac.kr (E.K. Kim).

http://dx.doi.org/10.1016/j.pedneo.2017.07.008
1875-9572/Copyright ª 2017, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. This is an open access article under the CC BY-
NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article in press as: Heo JS, et al., Timing of sepsis is an important risk factor for white matter abnormality in extremely
premature infants with sepsis, Pediatrics and Neonatology (2017), http://dx.doi.org/10.1016/j.pedneo.2017.07.008
 + MODEL
2 J.S. Heo et al
1. Introduction
Although the survival of premature newborns has improved
over past decades, it has been accompanied by an increase in
the number of infants affected by long-term neuro-
developmental morbidities.1e3 The neuropathologies un-
derlying these morbidities are diverse, although the primary
type of lesion appears to be white matter abnormality
(WMA),4 which is characterized by deep focal areas of cystic
necrosis and more diffuse cell-specific WM injury.5e7 Impor-
tantly, the severity of WMA is strongly correlated with the
degree of neurodevelopmental impairment.8
One of the major upstream mechanisms of WMA is sys-
temic infection/inflammation,5,7,9 and the majority of
preterm infants develop at least one neonatal infection
during their hospital stay.10 In particular, early- or late-
onset sepsis, meningitis and necrotizing enterocolitis
(NEC) cause overwhelming systemic inflammation, often
resulting in brain injury.11 Multiple clinical studies have
demonstrated an association between WMA and postnatal
sepsis.12e17 Few reports, however, have described the risk
factors for WMA in the context of neonatal sepsis.
In this study, we analyzed data from extremely prema-
ture infants diagnosed with neonatal sepsis who underwent
brain magnetic resonance imaging (MRI) prior to discharge
to identify clinical risk factors for moderate-to-severe WMA
in septic neonates. Identifying the cause of WMA in infants
with sepsis will be important for predicting neuro-
developmental outcomes and may lead to more advanced
therapies for intervention and prevention.
2. Methods
2.1. Study design and population
The study subjects were extremely premature infants born
before 28 weeks of gestation who were admitted to the
Figure 1 Cohort flow diagram depicting study population selection. IVH, Intraventricular hemorrhage; MRI, Magnetic resonance
imaging; WMA, White matter abnormality.
Please cite this article in press as: Heo JS, et al., Timing of sepsis is an important risk factor for white matter abnormality in extremely
premature infants with sepsis, Pediatrics and Neonatology (2017), http://dx.doi.org/10.1016/j.pedneo.2017.07.008
neonatal intensive care unit of the Seoul National Univer-
sity Children’s Hospital between January 1, 2006 and
December 31, 2015. The medical records of all neonates
with a positive blood culture who underwent brain MRI
before discharge were retrospectively reviewed.
A total of 335 infants were born before 28 weeks of
gestation during the study period (Fig. 1). One hundred-four
infants (31.0%) were diagnosed with postnatal sepsis during
their hospital stay, and 74 infants (71.1%) underwent brain
MRI before discharge. Among the 30 infants who did not
undergo brain MRI, four had intraventricular hemorrhage
(IVH)  grade 3 (classified by Papile et al.,18) and three had
significantly increased echogenicity of WM according to
cranial sonography. Among the 74 infants who underwent
brain MRI, two were excluded from the study due to the
presence of IVH grade 4 not concurrent with sepsis. These
two cases had IVH grade 4 before the septic event. In total,
the medical records of 72 infants, including clinical in-
vestigations and treatment, were reviewed in this study.
2.2. Ethics statement
The study protocol was approved by the institutional review
board (IRB) of the Seoul National University Hospital (IRB
No. 1508-137-697) with a waiver of informed consent.
Patient records/information were anonymized and de-
identified prior to data analysis.
2.3. Clinical data collection
A trained neonatologist collected data from maternal and
infant medical charts. The data collected included vari-
ables that could be associated with WMA. The diagnosis of
neonatal sepsis required the isolation of a microorganism
from a blood culture and at least one of the following
clinical signs or symptoms: apnea, bradycardia, hypother-
mia (core temperature of <36.5  C), fever (core
 + MODEL
Sepsis and white matter abnormality
temperature of >38.0  C), hypoglycemia (blood glucose of
<40 mg/dL) and hyperglycemia (blood glucose of >140 mg/
dL). Apnea and bradycardia were included as clinical signs
of sepsis only when the episodes were newly developed or
the frequency and duration of the episodes increased.
Sepsis of coagulase-negative Staphylococcus (CONS) was
reviewed using the modified specific criteria of the Centers
for Disease Control and Prevention defined by Bizzarro
et al.19 Clinical chorioamnionitis (CAM) was diagnosed by
the presence of maternal fever (>37.8  C) and at least two
of the following criteria: maternal tachycardia
(>100 beats/min), maternal leukocytosis (white blood cell
count > 15,000 cells/mm3), uterine tenderness, fetal
tachycardia (>160 beats/min), or foul-smelling amniotic
fluid.20 Placental histological examinations were reviewed
by a specialized pathologist for a variety of lesions
(chorion, decidua, amnion and umbilical cord), including
histological inflammation (graded on a scale of 0e4).21 CAM
was histologically diagnosed as the presence of  grade 2 in
the chorion, decidua, or amnion, and histological funisitis
was diagnosed as  grade 1 in the umbilical cord. Postnatal
morbidities included treated patent ductus arteriosus (PDA)
(by medication, operation or both), moderate-to-severe
bronchopulmonary dysplasia (need for > 21% oxygen or
positive pressure at 36 weeks postmenstrual age (PMA) or
discharge), IVH  grade 2, retinopathy of prematurity
 stage 3 (classified by the International Committee for the
Classification of retinopathy of prematurity22), NEC  stage
2 (classified by modified Bell’s staging criteria23), surgical
NEC and hypotension (use of inotropics or systemic steroids
for low blood pressure). For sepsis-related factors, sepsis
was classified as early-onset (4 days of life), late-onset
(5e30 days of life), and late, late-onset (>30 days of
life).19 Multidrug-resistant bacteria were defined as organ-
isms resistant to three or more antimicrobial classes.
Combined meningitis was defined as bacterial growth in the
cerebrospinal fluid (CSF) and blood. Recurrent sepsis was
defined as more than two episodes of sepsis, including at
least one episode occurring before PMA of 28 weeks. The
presence of a central vascular catheter was included in the
study only when the device was emplaced before the onset
of sepsis and was in place at the time of positive blood
culture. Surgery as a potential risk factor for sepsis was
included only when the procedure occurred  seven days
before the onset of positive blood culture.19 PMA and
postnatal day (PND) at sepsis diagnosis and sepsis onset
time were recorded at the first event of sepsis.
2.4. MRI studies
In our unit, all infants with a birth weight below 1 kg were
recommended to undergo brain MRI scans, regardless of
sonographic findings. In cases of a birth weight above 1 kg,
brain MRIs were conducted when sonographic findings were
abnormal. MRI scans were obtained at term-equivalent age
or before discharge. MRI scans were acquired using a 1.5-
T scanner (Avanto; Siemens Healthcare, Erlangen, Ger-
many) and a specialized high-sensitivity neonatal array coil
built into an MR-compatible incubator with a ventilation
device (LMT nomag IC; Lammers Medical Technology,
Lübeck, Germany). Patients who were intubated could
Please cite this article in press as: Heo JS, et al., Timing of sepsis is an important risk factor for white matter abnormality in extremely
premature infants with sepsis, Pediatrics and Neonatology (2017), http://dx.doi.org/10.1016/j.pedneo.2017.07.008
3
undergo MRI using this ventilator device. The brain MRI
studies included the following sequences: axial T2-
weighted turbo spin-echo, axial fluid-attenuated inversion
recovery, axial T1-weighted spin-echo, three-dimensional
magnetization-prepared rapid acquisition with gradient
echo, and axial diffusion-weighted imaging sequences. The
detailed parameters for the sequences are presented in a
supplementary table (Table S1). When necessary, infants
were sedated according to institutional protocols.
A pediatric radiologist who was blinded to the clinical
condition of the subjects interpreted each MRI study. We
used a standardized scoring system consisting of three-
point scales.8,12,24 WMA was graded according to five scales
that assessed 1) the nature and extent of WM signal ab-
normality, 2) the loss in volume of the periventricular WM,
3) the extent of any cystic abnormalities, 4) ventricular
dilatation, and 5) thinning of the corpus callosum. Com-
posite WM scores were calculated and used to categorize
the infants based on the extent of cerebral abnormality.
WMA was categorized as none (score of 5e6), mild (score of
7e9), moderate (score of 10e12), or severe (score of
13e15). The infants were then grouped as either none-to-
mild or moderate-to-severe WMA.
2.5. Statistical analysis
The SPSS version 22.0 statistical software package (SPSS,
Inc., Chicago, IL, USA) was used for data analysis. Man-
neWhitney U test was used to compare continuous vari-
ables. Fisher’s exact test (two-sided) was used to compare
categorical variables. Logistic regression analysis with
backward stepwise selection was performed considering
collinearity to analyze the predictors of moderate-to-
severe WMA. WMA status was the dependent variable, and
all the variables with p-value lower than 0.2 in the Man-
neWhitney U test or Fisher’s exact test entered in the
regression model as independent variables. In the regres-
sion model, clinical or histological CAM, NEC, broncho-
pulmonary dysplasia and recurrent sepsis were considered
confounders. Variables were entered at an entry level of
significance of p < 0.1 and remained in the model at an exit
level of p < 0.05. A Pearson correlation analysis was per-
formed to detect the correlation between PMA at sepsis
diagnosis and WM scores. Statistical significance was
defined as p < 0.05.
3. Results
3.1. Moderate-to-severe white matter abnormality
Among the 72 infants included in this study, 63 (87.5%) had
none-to-mild WMA and nine (12.5%) had moderate-to-
severe WMA. Table 1 shows the characteristics of infants
with moderate-to-severe WMA. There were five infants with
IVH  grade 2. Two infants had IVH grade 4, and the others
had IVH grade 2. Gram-positive organisms were the most
common pathogens of sepsis (four methicillin-resistant
Staphylococcus aureus [MRSA] and two CONS). Gram-
negative organisms (one Enterobacter aerogenes, one
imipenem-resistant Acinetobacter baumannii, and one
Escherichia coli) and fungi (two Candida glabrata) were
 + MODEL
4 J.S. Heo et al

also identified as causative organisms. Only five infants

BPD, Bronchopulmonary dysplasia; BW, Birth weight; CONS, Coagulase-negative Staphylococcus; CSF, Cerebrospinal fluid; F, Female; GA, Gestational age; IRAB, Imipenem-resistant A.

enterocolitis; op, Operation; PDA, Patent ductus arteriosus; PMA, Postmenstrual age; PND, Postnatal day; ROP, Retinopathy of prematurity; VP, Ventriculo-peritoneal; WMA, White matter
baumannii; IVH, Intraventricular hemorrhage; M, Male; med, Medication; MRI, Magnetic resonance imaging; MRSA, Methicillin-resistant S. aureus; NA, Not available; NEC, Necrotizing
NEC (op), BPD, PDA (op), ROP
underwent spinal tap, and the results were all negative.

NEC, BPD, PDA (med þ op)

BPD, PDA (med), ROP (op)
Patient 5, who had the highest WM score, underwent

NEC (op), BPD, PDA (op),

BPD, PDA (op), ROP (op)
recurrent septic events due to MRSA and C. glabrata. The

BPD, PDA (med þ op),

IVH that developed during a septic event progressed to
grade 4, and the patient subsequently underwent a
Co-morbidities

ventriculo-peritoneal shunt operation. There were no lab-

BPD, ROP (op)

BPD, PDA (op)
oratory data on the CSF analysis for this patient.
PDA (med)

ROP (op)

ROP (op)
3.2. Risk factors for moderate-to-severe white
matter abnormality

The clinical characteristics of the infants by WMA status are

Negative
Negative

Negative

Negative
Negative
CSF test

listed in Table 2. The mean gestational age (GA) was 25þ3

results

weeks, and the mean birth weight was 727 g. GA, Apgar
NA

NA

NA

NA

score at 1 min, Apgar score at 5 min, PMA at sepsis diag-

nosis, and PND at sepsis diagnosis showed a tendency of a
at sepsis diagnosis, weeks and days)

negative association with moderate-to-severe WMA

(25þ6 and 16), CONS (29þ4 and 40)
Pathogens of sepsis (PMA and PND

(p < 0.2), whereas the incidence of male sex, IVH  grade

C. glabrata (24þ6 and 9), CONS

MRSA (26þ0 and 7), C. glabrata

2, PMA at sepsis diagnosis <28 weeks, and days of me-

E. aerogenosa (29þ4 and 27)

chanical ventilation showed a tendency of a positive asso-

ciation with moderate-to-severe WMA (Table 2). The
incidence of PMA at sepsis diagnosis <28 weeks was
MRSA (27þ5 and 27)

E. coli (26þ5 and 3)

significantly higher in infants with moderate-to-severe WMA

CONS (24þ3 and 8)
MRSA (25þ6 and 4)

MRSA (24þ1 and 6)

IRAB (24þ4 and 5)

(88.9 vs. 46.0%, p Z 0.028 by Fisher’s exact test), and WM

(26þ5 and 12)

scores were significantly higher in infants diagnosed with

sepsis before PMA 28 weeks than in infants diagnosed with
sepsis after PMA 28 weeks (mean  standard deviation,
Characteristics of nine infants with moderate-to-severe white matter abnormality.

7.3  2.47 vs. 6.2  1.40, p Z 0.031). For the multivariate

logistic regression analysis including independent variables
and confounders, PMA at sepsis diagnosis (OR: 0.640, 95%
CI: 0.435e0.941, p Z 0.023) and PMA at sepsis diagnosis
Other cerebral

<28 weeks (OR: 9.232, 95% CI: 1.020e83.590, p Z 0.048),

IVH Grade 2

IVH Grade 4

IVH Grade 4

IVH Grade 2

IVH Grade 2
eVP shunt

respectively, were retained as independently associated

findings

variables with moderate-to-severe WMA (Table 3). The

correlation analysis showed that the PMA at sepsis diagnosis
had a significant negative correlation with WM scores
(r Z
0.243, p Z 0.039) (Fig. 2).
MRI (PMA, weeks)
Scores of WMA in

4. Discussion
10 (36þ1)
12 (40þ1)
11 (47þ1)

10 (37þ2)
15 (47þ4)

11 (38þ5)

11 (35þ5)

11 (38þ6)
10 (36þ2)

In the present study, PMA at sepsis diagnosis was an

important risk factor for moderate-to-severe WMA. In
particular, if sepsis occurred before PMA 28 weeks, the risk
of moderate-to-severe WMA was potentially higher.
In recent studies using MRI to determine the spectrum of
GA 26þ3; F; singleton; BW 1020
GA 25þ6; M; singleton; BW 870

GA 23þ4; M; singleton; BW 620

GA 25þ3; F; singleton; BW 620

WMA in premature newborns, the incidence of focal cystic

GA 23þ6; M; twin; BW 730
GA 23þ5; M; twin; BW 650

GA 25þ1; M; twin; BW 820

GA 24þ0; M; twin; BW 730

GA 23þ3; M; twin; BW 550

necrosis with loss of all cellular elements has distinctly

General characteristics

decreased, whereas non-cystic WMA has become the most

prevalent pattern of brain injury in premature newborns.25
(GA, weeks; BW, g)

Non-cystic WMA is a cell-specific lesion with acute loss of

early differentiating oligodendrocytes, also known as pre-
myelinating oligodendrocytes (pre-OLs).7 Pre-OLs represent
a specific phase of the oligodendroglial lineage that is
highly susceptible to hypoxia-ischemia and systemic infec-
tion/inflammation.26 Pre-OLs are present as early as 18
abnormality.

weeks and comprise approximately 90% of the total oligo-

dendroglial population until approximately 28 weeks
Table 1
Patient

gestation.7,9,27 After 28 weeks gestation, expansion of the

immature OL population is prominent.27 Therefore, the
1
2
3

4
5

8
9

timing of insults to the developing brain is a critical factor

Please cite this article in press as: Heo JS, et al., Timing of sepsis is an important risk factor for white matter abnormality in extremely
premature infants with sepsis, Pediatrics and Neonatology (2017), http://dx.doi.org/10.1016/j.pedneo.2017.07.008
 + MODEL
Sepsis and white matter abnormality 5

Table 2 Comparison of the characteristics based on severity of white matter abnormality.

None-to-mild Moderate-to-severe p-value
(n Z 63) (n Z 9)
Demographics and perinatal factors
Gestational age, weeks, mean  SD 25þ4  1.473 24þ4  1.120 0.043a
Birth weight, grams, mean  SD 726  151 734  147 0.952a
Male, n (%) 31 (49.2) 7 (77.8) 0.158b
C-section delivery, n (%) 35 (55.6) 5 (55.6) >0.99b
Clinical or histological chorioamnionitis, n (%) 34 (54.0) 5 (55.6) >0.99b
Histological funisitis, n (%) 11 (17.5) 1 (11.1) >0.99b
Prenatal medications
Glucocorticoids, n (%) 52 (82.5) 6 (66.7) 0.363b
Antibiotics, n (%) 36 (58.1) 4 (44.4) 0.490b
Apgar score, 1 min, mean  SD 3.2  1.726 2.3  1.581 0.119a
Apgar score, 5 min, mean  SD 5.6  1.485 5.0  0.866 0.099a
Cord blood pH, mean  SD 7.303  0.089 7.282  0.118 0.883a
Cord blood base excess, mmol/L, mean  SD
3.4  3.965
5.5  5.848 0.466a
Postnatal factors
Treated PDA, n (%) 56 (88.9) 8 (88.9) >0.99b
Moderate to severe BPD, n (%) 48 (76.2) 8 (88.9) 0.673b
Use of systemic steroids for BPD, n (%) 9 (14.3) 2 (22.2) 0.619b
IVH  Grade 2, n (%) 13 (20.6) 5 (55.6) 0.038b
IVH Grade 2 12 (19.0) 3 (33.3)
IVH Grade 3 1 (1.6) 0 (0)
IVH Grade 4 0 (0) 2 (22.2)
ROP  stage 3, n (%) 31 (49.2) 6 (66.7) 0.480b
NEC  stage 2, n (%) 17 (27.0) 4 (44.4) 0.433b
Surgical NEC, n (%) 15 (23.8) 3 (33.3) 0.682b
Hypotension during hospitalization, n (%) 53 (84.1) 7 (77.8) 0.639b
Mechanical ventilation, days, mean  SD 48  34.375 68  28.403 0.114a
Hospital stay, days, mean  SD 114  34.065 131  44.011 0.250a
PMA at brain MRI, weeks, mean  SD 38þ5  2.784 39þ5  4.545 0.878a
Sepsis-related factors
PMA at sepsis diagnosis, weeks, mean  SD 28þ2  3.001 25þ6  1.783 0.011a
PMA at sepsis diagnosis <28 weeks, n (%) 29 (46.0) 8 (88.9) 0.028b
PND at sepsis diagnosis, days, mean  SD 19  17.044 10  9.488 0.054a
Sepsis onset time
Early onset sepsis, n (%) 7 (11.1) 2 (22.2) 0.312b
Late onset sepsis, n (%) 46 (73.0) 7 (77.8) >0.99b
Late, late onset sepsis, n (%) 10 (15.9) 0 (0.0) 0.343b
Pathogen
Gram positive, n (%) 46 (73.0) 6 (66.7) 0.701b
Gram negative, n (%) 13 (20.6) 3 (33.3) 0.406b
Fungus, n (%) 13 (20.6) 2 (22.2) >0.99b
MDR bacteria, n (%) 34 (54.0) 7 (77.8) 0.283b
Combined meningitis, n (%) 1/42 (2.4) 0/4 (0.0) >0.99b
Recurrent sepsis, n (%) 13 (20.6) 2 (22.2) >0.99b
Hypotension during sepsis, n (%) 27 (42.9) 3 (33.3) 0.667b
Central vascular catheter, n (%) 54 (85.7) 8 (88.9) >0.99b
Surgery, n (%) 10 (15.9) 0 (0.0) 0.343b
Duration of positive to negative conversion 88  51.025 72  38.502 0.438a
of blood culture results, hours, mean  SD
Duration from sepsis onset to use of susceptible 34  29.803 50  40.453 0.217a
antibiotics, hours, mean  SD
Bold values indicate p-value < 0.2.
BPD, Bronchopulmonary dysplasia; IVH, Intraventricular hemorrhage; MDR, Multi-drug resistant; MRI, Magnetic resonance imaging; NEC,
Necrotizing enterocolitis; PDA, Patent ductus arteriosus; PMA, Postmenstrual age; PND, Postnatal day; ROP, Retinopathy of prematurity;
SD, Standard deviation.
a
ManneWhitney U test.
b
Fisher’s exact test.

Please cite this article in press as: Heo JS, et al., Timing of sepsis is an important risk factor for white matter abnormality in extremely
premature infants with sepsis, Pediatrics and Neonatology (2017), http://dx.doi.org/10.1016/j.pedneo.2017.07.008
 + MODEL
6 J.S. Heo et al

Table 3 Logistic regression analysis for predicting moderate-to-severe white matter abnormality.
Variable Unadjusted model Adjusted modela
OR 95% CI p-value OR 95% CI p-value
PMA at sepsis diagnosis, weeks 0.637 0.432e0.939 0.023 0.640 0.435e0.941 0.023
PMA at sepsis diagnosis <28 weeks 9.379 1.107e79.486 0.040 9.232 1.020e83.590 0.048
CI, Confidence interval; OR, Odds ratio; PMA, Postmenstrual age.
a
Dependent variable þ independent variables þ confounders (including clinical or histological chorioamnionitis, necrotizing
enterocolitis  stage 2, bronchopulmonary dysplasia and recurrent sepsis).

contributing to WMA. Our data demonstrating that PMA at
sepsis diagnosis, particularly before 28 weeks, correlates
with moderate-to-severe WMA are consistent with these
findings. Although many animal studies correspond well
with this hypothesis,28e30 there have been few clinical
studies analyzing the relationship between the timing of
sepsis and the severity of WMA. The present study has
considerable value, in that the results suggest a significant
effect of the timing of systemic infection on WMA in
extremely preterm infants.
In the present study, hypotension during sepsis or hos-
pitalization was not correlated with moderate-to-severe
WMA. A potentiating interaction between hypoxia-ischemia
and systemic infection/inflammation has been demon-
strated in previous studies,31,32 although our results
contradict those results. We can think of two reasons for
this discrepancy. First, hypotension itself is not well
defined. Generally, most neonatologists consider hypoten-
sion present if the infant’s mean arterial pressure is less
than the GA in weeks.33,34 However, by this definition,
hypotension does not always accompany cerebral hypo-
perfusion. Lightburn et al.35 observed no significant dif-
ference in cerebral blood flow velocity between extremely
low birth weight infants with hypotension and infants with
normal blood pressure. Moreover, in a study by Bonestroo
et al.,36 antihypotensive treatment did not induce any
significant change in regional cerebral oxygen saturation or
fractional tissue oxygen extraction in hypotensive preterm
infants without PDA. Therefore, hypoperfusion may not
have occurred in the hypotensive infants in the present
study, resulting in no hypoxia-ischemia in the brain. Sec-
ond, the potentiation between systemic infection/inflam-
mation and hypoxia-ischemia depends on the relative
timing of the insults.7 In the rat pup model employed by
Eklind et al.,37 lipopolysaccharide enhanced vulnerability
during both the acute and chronic phases after adminis-
tration, whereas an intermediate interval between lipo-
polysaccharide treatment and hypoxia-ischemia resulted in
tolerance rather than potentiation. In the present study,
we did not consider the timing of the insults between

Figure 2 Scatter plot and linear line showing white matter scores as a function of postmenstrual age at sepsis diagnosis. PMA,
Postmenstrual age; WM, White matter.

Please cite this article in press as: Heo JS, et al., Timing of sepsis is an important risk factor for white matter abnormality in extremely
premature infants with sepsis, Pediatrics and Neonatology (2017), http://dx.doi.org/10.1016/j.pedneo.2017.07.008
 + MODEL
Sepsis and white matter abnormality
systemic infection and hypoxia-ischemia, and the time in-
tervals between these insults were likely heterogeneous.
According to the Fisher’s exact test in the present study,
IVH combined with sepsis was correlated with moderate-to-
severe WMA. When IVH develops, the amount of non-
protein-bound iron in the CSF increases, and this free iron
could be used to convert hydrogen peroxide to hydroxyl
radical via the Fenton reaction.38 Free radical generation is
the most important downstream mechanism of WM injury,
as free radical attack appears to be the principal final
common pathway to injury.7 In addition, increased
methemoglobin formation in the intraventricular space
following IVH induces expression of pro-inflammatory cy-
tokines.39 These cytokines potentiate maturation-
dependent toxicity, affecting pre-OLs.13 Therefore, we
conducted IVH grade 2, 3-adjusted analysis, and IVH grade
4 cases were excluded, unless occurring during the septic
events, due to significant impacts on WMA scores.
In the present study, recurrent postnatal sepsis was not
associated with moderate-to-severe WMA. Furthermore,
including NEC and prenatal CAM as additional inflammatory
events did not affect the results. In recurrent systemic
infection, the development of pre-OLs can be inhibited by
toxicity or by direct activation of receptors on pre-OLs such
as Toll-like receptors or interferon-g,7 and this inhibition of
pre-OL development may be associated with progressive
WM injury in preterm infants.15 However, in a study by Glass
et al.,15 seven of 12 infants with progressive WM injury
exhibited progression from none to minimal WM injury, and
no progression to moderate or severe injury was observed.
In the present study, we compared none-to-mild WMA with
moderate-to-severe WMA, as we could not investigate the
progression of WMA due to the lack of serial brain MRI
studies. The WM scores were higher in infants with recur-
rent postnatal sepsis than in infants without recurrent
sepsis, although this difference was not significant
(mean  standard deviation, 7.0  2.71 vs. 6.7  1.89,
p Z 0.847).
There were several limitations to this study. First, the
small sample size and relative rarity of moderate-to-severe
WMA limited our statistical analyses, which were insufficient
to confirm the effects of hypotension and recurrent sepsis on
WMA. Despite this limitation, PMA at sepsis diagnosis,
especially before 28 weeks, was identified as a risk factor for
moderate-to-severe WMA. Second, the data on combined
meningitis were imprecise, as many infants did not undergo
lumbar puncture based on the decision of the treating
physician. This lack of precision could explain why menin-
gitis was not associated with WMA. Third, several infants
experienced recurrent septic events. In the statistical
analysis, the pathogen, the presence of combined menin-
gitis, hypotension during sepsis, central vascular catheters,
surgery, duration of positive to negative conversion of blood
culture results, and duration from sepsis onset to use of
susceptible antibiotics were determined for each individual
episode of sepsis. These factors were different for each
episode in patients with recurrent sepsis, and therefore, it is
difficult to evaluate the precise risk factors in such cases.
Finally, we have no data on long-term neurodevelopmental
outcomes relative to the severity of WMA in our study.
Further investigations of the associations between WMA and
neurodevelopmental outcomes are needed.
Please cite this article in press as: Heo JS, et al., Timing of sepsis is an important risk factor for white matter abnormality in extremely
premature infants with sepsis, Pediatrics and Neonatology (2017), http://dx.doi.org/10.1016/j.pedneo.2017.07.008
7
5. Conclusions
This study suggests that PMA at sepsis diagnosis might be an
important risk factor for moderate-to-severe WMA in
extremely premature infants with neonatal sepsis. Infants
diagnosed with sepsis before PMA 28 weeks represent a
high-risk group for moderate-to-severe WMA. These findings
could be considered when determining the target of neu-
roprotective treatments in premature infants with sepsis.
Funding
This research was supported by the Basic Science Research
Program through the National Research Foundation of
Korea (NRF) funded by the Ministry of Education
(2012R1A1A2044109).
Conflicts of interest
None declared.
References
1. Moore T, Hennessy EM, Myles J, Johnson SJ, Draper ES,
Costeloe KL, et al. Neurological and developmental outcome in
extremely preterm children born in England in 1995 and 2006:
the EPICure studies. BMJ 2012;345:e7961.
2. Synnes AR, Anson S, Arkesteijn A, Butt A, Grunau RE, Rogers M,
et al. School entry age outcomes for infants with birth weight
800 grams. J Pediatr 2010;157:989e994.e1.
3. Wilson-Costello D, Friedman H, Minich N, Fanaroff AA, Hack M.
Improved survival rates with increased neurodevelopmental
disability for extremely low birth weight infants in the 1990s.
Pediatrics 2005;115:997e1003.
4. Shane AL, Hansen NI, Stoll BJ, Bell EF, Sánchez PJ,
Shankaran S, et al. Methicillin-resistant and susceptible
Staphylococcus aureus bacteremia and meningitis in preterm
infants. Pediatrics 2012;129:e914e22.
5. Volpe JJ. Postnatal sepsis, necrotizing entercolitis, and the
critical role of systemic inflammation in white matter injury in
premature infants. J Pediatr 2008;153:160e3.
6. 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;
147:609e16.
7. Volpe JJ, Kinney HC, Jensen FE, Rosenberg PA. The developing
oligodendrocyte: key cellular target in brain injury in the
premature infant. Int J Dev Neurosci 2011;29:423e40.
8. Woodward LJ, Anderson PJ, Austin NC, Howard K, Inder TE.
Neonatal MRI to predict neurodevelopmental outcomes in
preterm infants. N Engl J Med 2006;355:685e94.
9. Back SA. Cerebral white and gray matter injury in newborns:
new insights into pathophysiology and management. Clin Per-
inatol 2014;41:1e24.
10. 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;292:2357e65.
11. American College of Obstetricians and Gynecologists. ACOG
practice bulletin: clinical management guidelines for
obstetrician-gynecologists: number 38, September 2002. Peri-
natal care at the threshold of viability. Obstet Gynecol 2002;
100:617e24.
 + MODEL
8 J.S. Heo et al
12. Inder TE, Wells SJ, Mogridge NB, Spencer C, Volpe JJ. Defining
the nature of the cerebral abnormalities in the premature in-
fant: a qualitative magnetic resonance imaging study. J
Pediatr 2003;143:171e9.
13. Graham EM, Holcroft CJ, Rai KK, Donohue PK, Allen MC.
Neonatal cerebral white matter injury in preterm infants is
associated with culture positive infections and only rarely with
metabolic acidosis. Am J Obstet Gynecol 2004;191:1305e10.
14. Shah DK, Doyle LW, Anderson PJ, Bear M, Daley AJ, Hunt RW,
et al. Adverse neurodevelopment in preterm infants with
postnatal sepsis or necrotizing enterocolitis is mediated by
white matter abnormalities on magnetic resonance imaging at
term. J Pediatr 2008;153:170e5. 175.e1.
15. Glass HC, Bonifacio SL, Chau V, Glidden D, Poskitt K,
Barkovich AJ, et al. Recurrent postnatal infections are asso-
ciated with progressive white matter injury in premature in-
fants. Pediatrics 2008;122:299e305.
16. Hofer N, Müller W, Resch B. White matter damage and neonatal
sepsis. Acta Paediatr 2011;100:e1; author reply e1e2.
17. Chau V, Brant R, Poskitt KJ, Tam EW, Synnes A, Miller SP.
Postnatal infection is associated with widespread abnormal-
ities of brain development in premature newborns. Pediatr Res
2012;71:274e9.
18. Papile LA, Burstein J, Burstein R, Koffler H. Incidence and
evolution of subependymal and intraventricular hemorrhage: a
study of infants with birth weights less than 1,500 gm. J
Pediatr 1978;92:529e34.
19. Bizzarro MJ, Raskind C, Baltimore RS, Gallagher PG. Seventy-
five years of neonatal sepsis at Yale: 1928e2003. Pediatrics
2005;116:595e602.
20. Tita AT, Andrews WW. Diagnosis and management of clinical
chorioamnionitis. Clin Perinatol 2010;37:339e54.
21. Salafia CM, Minior VK, Rosenkrantz TS, Pezzullo JC, Popek EJ,
Cusick W, et al. Maternal, placental, and neonatal associations
with early germinal matrix/intraventricular hemorrhage in in-
fants born before 32 weeks’ gestation. Am J Perinatol 1995;12:
429e36.
22. International Committee for the Classification of Retinopathy
of Prematurity. The International classification of retinopathy
of prematurity revisited. Arch Ophthalmol 2005;123:991e9.
23. Walsh MC, Kliegman RM. Necrotizing enterocolitis: treatment
based on staging criteria. Pediatr Clin North Am 1986;33:
179e201.
24. Woodward LJ, Mogridge N, Wells SW, Inder TE. Can neuro-
behavioral examination predict the presence of cerebral injury
in the very low birth weight infant? J Dev Behav Pediatr 2004;
25:326e34.
25. 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:
155e64.
26. Favrais G, van de Looij Y, Fleiss B, Ramanantsoa N, Bonnin P,
Stoltenburg-Didinger G, et al. Systemic inflammation disrupts
Please cite this article in press as: Heo JS, et al., Timing of sepsis is an important risk factor for white matter abnormality in extremely
premature infants with sepsis, Pediatrics and Neonatology (2017), http://dx.doi.org/10.1016/j.pedneo.2017.07.008
the developmental program of white matter. Ann Neurol 2011;
70:550e65.
27. 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;21:1302e12.
28. Reddy K, Mallard C, Guan J, Marks K, Bennet L, Gunning M, et al.
Maturational change in the cortical response to hypoperfusion
injury in the fetal sheep. Pediatr Res 1998;43:674e82.
29. Hagberg H, Peebles D, Mallard C. Models of white matter
injury: comparison of infectious, hypoxic-ischemic, and exci-
totoxic insults. Ment Retard Dev Disabil Res Rev 2002;8:30e8.
30. Buser JR, Segovia KN, Dean JM, Nelson K, Beardsley D, Gong X,
et al. Timing of appearance of late oligodendrocyte pro-
genitors coincides with enhanced susceptibility of preterm
rabbit cerebral white matter to hypoxia-ischemia. J Cereb
Blood Flow Metab 2010;30:1053e65.
31. Wang X, Stridh L, Li W, Dean J, Elmgren A, Gan L, et al.
Lipopolysaccharide sensitizes neonatal hypoxic-ischemic brain
injury in a MyD88-dependent manner. J Immunol 2009;183:
7471e7.
32. Wang LW, Chang YC, Lin CY, Hong JS, Huang CC. Low-dose
lipopolysaccharide selectively sensitizes hypoxic ischemia-
induced white matter injury in the immature brain. Pediatr
Res 2010;68:41e7.
33. Seri I. Cardiovascular, renal, and endocrine actions of dopa-
mine in neonates and children. J Pediatr 1995;126:333e44.
34. Rozé JC, Tohier C, Maingueneau C, Lefèvre M, Mouzard A.
Response to dobutamine and dopamine in the hypotensive very
preterm infant. Arch Dis Child 1993;69:59e63.
35. Lightburn MH, Gauss CH, Williams DK, Kaiser JR. Cerebral blood
flow velocities in extremely low birth weight infants with hy-
potension and infants with normal blood pressure. J Pediatr
2009;154:824e8.
36. Bonestroo HJ, Lemmers PM, Baerts W, van Bel F. Effect of
antihypotensive treatment on cerebral oxygenation of preterm
infants without PDA. Pediatrics 2011;128:e1502e10.
37. 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;58:112e6.
38. Iida K, Takashima S, Ueda K. Immunohistochemical study of
myelination and oligodendrocyte in infants with periven-
tricular leukomalacia. Pediatr Neurol 1995;13:296e304.
39. Gram M, Sveinsdottir S, Ruscher K, Hansson SR, Cinthio M,
Akerström B, et al. Hemoglobin induces inflammation after
preterm intraventricular hemorrhage by methemoglobin for-
mation. J Neuroinflamm 2013;10:100.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
http://dx.doi.org/10.1016/j.pedneo.2017.07.008.