Seminars in Fetal & Neonatal Medicine xxx (2016) 1e6

Contents lists available at ScienceDirect

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier.com/locate/siny

Review

Gut bacteria and late-onset neonatal bloodstream infections in

preterm infants
Phillip I. Tarr a, *, Barbara B. Warner b
a
Division of Gastroenterology, Hepatology, and Nutrition, Pathobiology Research Unit, Department of Pediatrics, Washington University School of Medicine,
St Louis, MO, USA
b
Fetal Center, Division of Newborn Medicine, Washington University School of Medicine, St Louis, MO, USA

s u m m a r y
Keywords: Late-onset neonatal bloodstream infections remain challenges in neonatology. Hand hygiene, line care,
Late-onset neonatal bloodstream infections and judicious use of indwelling lines are welcome interventions, but might not reduce the incidence of
Gram-negative bacteria
late-onset neonatal bloodstream infections from bacteria originating in the gut. Accumulating data
Gram-positive bacteria
Gut microbes
suggest that many pathogens causing late-onset neonatal bloodstream infections are of gut origin,
including Gram-positive cocci. In addition to the host-canonical paradigm (i.e., all bacteria have equal
risk of invasion and bloodstream infections are functions of variable infant susceptibility), we should now
consider bacteria-canonical paradigms, whereby late-onset neonatal bloodstream infection is a function
of colonization with a specific subset of bacteria with exceptional invasive potential. In either event, we
can no longer be content to reactively approach late-onset neonatal bloodstream infections; instead we
need to reduce the occurrences of these infections by broadening our scope of effort beyond line care,
and determine the pre-invasive habitat of these pathogens.
© 2016 Published by Elsevier Ltd.

1. Introduction cerebral palsy, independent of intracranial structural abnormalities

Late-onset neonatal bloodstream infections in preterm infants
are illnesses (i) in which a credible pathogen is recovered from the
blood, and (ii) that occur after the first 72 h of age [1e5]. The Gram-
negative bacilli, Gram-positive cocci, and fungi that represent the
majority of infections have differing pre-invasion habitats in the
body, control strategies, treatments, and prognoses when they
cause extraintestinal infections.
Infants who have had culture-proven late-onset neonatal
bloodstream infections have higher mortality than those who have
not had these infections [1,6]. Late-onset neonatal bloodstream
infections in very low birth weight infants bestow independent
risks to long-term child development, in addition to the well-
recognized risks bestowed by brain injury, bronchopulmonary
dysplasia, and retinopathy of prematurity [2]. Indeed, it has been
calculated that a single late-onset neonatal bloodstream infection
in a preterm infant approximately quadruples the likelihood of
* Corresponding author. Address: Department of Pediatrics, Washington
University School of Medicine, Campus Box 8208, 660 S. Euclid, St Louis, MO 63110,
USA. Tel.: þ1 314 286 2848.
E-mail address: tarr@wustl.edu (P.I. Tarr).
[7]. Neonatal infections are also associated with lower Bayley Scale
of Infant Development II scores, worse psychomotor development,
abnormal vision, and lesser occipital frontal circumferences, among
those infants who survive their episode of late-onset neonatal
bloodstream infections and who are discharged to home [3].
Our understanding of late-onset neonatal bloodstream
infections is severely limited by their unpredictable onset in infants
at risk. In this review, we emphasize data from human cohorts, and
cite animal data only if they might illuminate human biology
relevant to late-onset neonatal bloodstream infections.
2. Prospects for preventing late-onset neonatal bloodstream
infections beyond line care and hand hygiene
Lessons can be learned from attempts to control early-onset
(occurring <72 h after birth) infection with Streptococcus aga-
lactiae (Group B streptococcus (GBS)). Screening for maternal
colonization with GBS, and the treatment of mothers and their
newborns with parenteral antibiotics (usually b-lactam agents),
have reduced the incidence of early-onset bloodstream infections
with GBS [8,9]. In contrast, the incidence of late-onset neonatal
bloodstream infections caused by GBS is increasing [10,11], and can

http://dx.doi.org/10.1016/j.siny.2016.06.002
1744-165X/© 2016 Published by Elsevier Ltd.

Please cite this article in press as: Tarr PI, Warner BB, Gut bacteria and late-onset neonatal bloodstream infections in preterm infants, Seminars
in Fetal & Neonatal Medicine (2016), http://dx.doi.org/10.1016/j.siny.2016.06.002
2 P.I. Tarr, B.B. Warner / Seminars in Fetal & Neonatal Medicine xxx (2016) 1e6
exceed the incidence of early-onset GBS bloodstream infections
[12]. It is logical, therefore, to extend successful strategies to reduce
early-onset GBS infections to late-onset neonatal bloodstream in-
fections, caused by GBS as well as by other pathogens. However, it is
important to note that prevention strategies focused on early-onset
GBS rely on one critical assumption: the mother at risk, by virtue of
gut or birth canal colonization with GBS, may be identified. Once
this risk is identified, it can be managed (perinatal and usually
parenteral antibiotics). We do not yet know if this strategy can be
extended to late-onset neonatal bloodstream infections, but data
are emerging that might inform the discussion.
Late-onset neonatal bloodstream infections are generally caused
by species whose members are well represented in the gut and
elsewhere in and on the body in health. There are, therefore, two
distinct paradigms for the pathobiology that underlies the occur-
rence of a late-onset neonatal bloodstream infection (Box 1). The
first paradigm, and one that is highly ingrained among
neonatologists, assumes that risk for a late-onset neonatal blood-
stream infection is a function of specific host factors that increase
susceptibility. Such variably at-risk hosts are then exposed to a
population of microbes that have constant pathogenic potential. In
this “host-canonical” situation, invasion of the bloodstream is
considered to be driven by physical breaches in the integrity of the
mucosa or the integument, or by impaired or poorly developed
innate or acquired host defense mechanisms. According to this first
model, late-onset neonatal bloodstream infections would be best
prevented by attention to host biology (though recognizing that
beyond line care and hand hygiene, protection interventions are
not well validated). The second paradigm assumes that among
preterm infants, there is a fairly constant level of impairment of
host defense against bacteria that can invade the bloodstream. In
this “bacteria-canonical” situation, late-onset neonatal blood-
stream infections result when a specific member (i.e., a species,
serotype, pathotype, or clade) of a genus that is widely found in or
on infants in the healthy state, colonizes a host and then invades
the bloodstream, because that specific member has exceptional
invasive potential.
It is quite plausible that all very preterm infants are at high risk
for late-onset neonatal bloodstream infections, by virtue of their
poorly developed immune systems, increased gut permeability,
and the near-universal use of indwelling lines (at least tempo-
rarily), and that this risk is further stratified according to gesta-
tional age at birth, day of life, and certain co-morbidities. Such a
non-dissimilar high risk background would tend to support the
bacteria-canonical paradigm, within defined subgroups (e.g.,
gestational age at birth, age, etc.). That is to say, late-onset neonatal
bloodstream infections occur only if a rare member of a common
bacterial taxon colonizes the infant and if that colonizer has path-
ogenic potential. Additional data support this concept. For example,
even though there are many genera of bacteria in and on the bodies
of preterm infants, the number of identified genera in blood cul-
tures is highly circumscribed. Specifically, Gram-negative bacilli
(largely Escherichia coli, Klebsiella, Serratia, Enterobacter, and Pseu-
domonas spp.), GBS, enterococci, Staphylococcus aureus, and
coagulase-negative Staphylococci account for almost all late-onset
neonatal bloodstream infections [14,15]. Interestingly, with few
modifications, these pathogens account for the bulk of bloodstream
infections at all ages, and in diverse settings, even though they are
not the most prevalent members of the gut bacterial communities.
If the bacterial canonical paradigm explains even a subset of
late-onset neonatal bloodstream infections, it would be logical that
by reducing the rate, duration, and consequences of colonization by
highly virulent bacteria, there will be great benefit to very low birth
weight infants. If the host-canonical paradigm is more appropriate,
Please cite this article in press as: Tarr PI, Warner BB, Gut bacteria and late-onset neonatal bloodstream infections in preterm infants, Seminars
in Fetal & Neonatal Medicine (2016), http://dx.doi.org/10.1016/j.siny.2016.06.002
Box 1
Host-canonical vs bacteria-canonical paradigms for late-onset
neonatal bloodstream infections.
Host-canonical risk for late-onset neonatal bloodstream
infections
There are major inter-individual variations in host risk.
There are minor variations in bacterial populations to which
the hosts are exposed, at least at taxonomic levels of genus
and higher. Therefore, specific host risks ordain late-onset
neonatal bloodstream infections in infants exposed to a
set of bacteria with similar pathogenic potential.
Observations in support of a host-canonical paradigm for
late-onset neonatal bloodstream infections:
1 Line care checklists and protocols, and augmented hand
hygiene reduce late-onset neonatal bloodstream infec-
tion incidence. This is a host-specific intervention.
2 Children “age-out” of a period of high risk for late-onset
neonatal bloodstream infections. This suggests matura-
tion of a host-defense process.
Bacteria-canonical risk for late-onset neonatal bloodstream
infections
There are inter-individual variations in host risk, but most
differences are based on gestational age and illness
severity. Hence, among infants of the same gestational age,
day of life, and with a similar set of general co-morbidities,
there is a fairly consistent inter-child risk from mucosal and
integumentary barrier defects, and from poorly developed
immune systems. However, there may be major variations
in bacterial populations that hosts encounter. If a preterm
infant is colonized with an organism with highly virulent
potential, even though it might belong to a taxon that is
often regarded as “commensal” or harmless, bloodstream
infections occur. In this scenario, the specific organism or-
dains most of the risk.
Observations in support of a bacteria-canonical paradigm
for late-onset neonatal bloodstream infections:
1 The number of taxa that inhabit the gut is large, whereas
the number of species that cause late-onset neonatal
bloodstream infections (and, indeed, bloodstream in-
fections at all ages) is much more circumscribed. This
demonstrates, at least at the species level, variant path-
ogenic potential.
2 Considerable molecular epidemiologic data suggest that
species causing many episodes of late-onset neonatal
bloodstream infections, namely E. coli and Group B
streptococcus, belong to phylogenetic subsets that are
associated with extra-intestinal pathogenicity.
3 One technique to interdict outbreaks of late-onset
neonatal bloodstream infections is to isolate infected in-
fants, even though the pathogens in question may belong
to organisms that, at a genus and species level, are
frequently found in the gut.
4 Bacteria that invaded the bloodstream of patients in the
study by Carl et al. [13] were rarely found in “overlap” or
random controls. The inferred relative risk for invasion by
these specific organisms is high.
 P.I. Tarr, B.B. Warner / Seminars in Fetal & Neonatal Medicine xxx (2016) 1e6
there are some tools to achieve the goal of reducing the incidence of
late-onset neonatal bloodstream infections, but most such at-
tempts to prevent these infections focus on central line care and
hand hygiene [16e18]. The only intervention to prevent blood-
stream infections caused by organisms that do not dwell on the
skin is isolation of children during apparent outbreaks. However,
the vast majority of late-onset neonatal bloodstream infections do
not occur in nosocomial clusters, at least not those that are
recognized in real-time, and isolation is an intervention that
actually endorses the concept of the bacteria-canonical paradigm.
Whereas it is not surprising that Gram-negative bacteria that
reside in the gut cause Gram-negative bacterial bloodstream in-
fections, there has been remarkably little emphasis in the field of
neonatology on the threat posed by this biomass. However, accu-
mulating data now compel us to consider the infant gut as a critical
habitat of organisms that subsequently invade the bloodstream of
their hosts. These data build on two complementary contributions
from the internal medicine literature. First, resistance patterns of
colonizing bacteria in the gut may be used to tailor and optimize
empiric antibiotics during clinical deterioration, a strategy
employed mostly in intensive care units [19e21]. Second, selective
bowel decontamination, a practice more common in Europe than in
the USA and most frequently used in adult intensive care settings, is
effective at reducing mortality and nosocomial infections including
bloodstream infections [22]. Parenthetically, selective bowel
decontamination does not favor the emergence of antibiotic-
resistant pathogens in these settings [23,24].
3. The gut as the pre-invasion habitat/reservoir of bacterial
bloodstream invaders
Several observations support the concept that the gut is the
reservoir of Gram-negative pathogens before late-onset neonatal
bloodstream infections are clinically apparent. Though it is
intuitive that increased gut permeability would permit the extra-
intestinal dissemination (translocation) of bacteria harbored by
this organ, it was not until Graham et al. demonstrated that,
among 20 episodes of bloodstream infections in 15 infants
caused by Gram-negative bacilli, 19 of the bloodstream isolates
could be paired to bacteria in stool samples that had been ob-
tained before the episode of sepsis. Putative identity was
assigned according to identical pulsed-field gel electrophoresis
(PFGE) patterns [25]. PFGE is a macrogenomic microbial finger-
printing technique that differentiates isolates within the same
species. Antibiotic susceptibility (to gentamicin) was similarly
concordant between the bloodstream and fecal surveillance
isolates. In a subsequent study, this group demonstrated that 58
of 59 episodes of Gram-negative bacillary bloodstream infections
were caused by bacteria that could be matched to organisms of
the same genus and species, based on concordant gentamicin
susceptibility/resistance profiles [26].
With the recent explosion of interest in the gut microbiome,
attempts have been made to determine if there exists an identifi-
able community of bacteria that predispose to late-onset neonatal
bloodstream infections. In general, these studies have suggested
potentially different microbial populations prior to the onset of
sepsis, with risks including low microbial diversity [27]. Taft et al.
[28] identified a paucity of actinobacteria in early life and under-
representation of Pseudomonades in samples just before late-onset
sepsis, but no highly compelling concordance between blood-
stream isolate and fecal sequences at the level of species. Shaw et al.
[29] identified overrepresentation of fecal aerobic bacteria and
facultative anaerobes compared to infants who did not have
bloodstream infections. Gram-negative bloodstream infections
Please cite this article in press as: Tarr PI, Warner BB, Gut bacteria and late-onset neonatal bloodstream infections in preterm infants, Seminars
in Fetal & Neonatal Medicine (2016), http://dx.doi.org/10.1016/j.siny.2016.06.002
3
were associated with microbial gut predominance of Enter-
obacteriaceae. They also identified in many subjects a concordance
between antibacterial resistance of the bloodstream isolate and
that of bacteria from the stool prior to culture, but did not report
more specific characterization of the pathogens and the stool
isolates.
Our personal view of the origin of late-onset neonatal
bloodstream infections has been influenced by Carl et al. [13].
This paper precisely established identity between specific bac-
teria in the stool and subsequent invasion of the blood by these
strains. Specifically, in a cohort of 217 preterm infants in a single
neonatal intensive care unit, 11 children whose bloodstream in-
fections were caused by E. coli, Serratia marcescens, Klebsiella
pneumoniae, and methicillin-resistant Staphylococcus aureus
(MRSA) also had sufficient stool prior to the event that was stored
and available for microbiologic study. Of these 11 cases, seven
excreted a cognate pathogen (E. coli, S. marcescens, and GBS)
before clinical deterioration was apparent, and the culture was
obtained. Isogenicity between stool and blood isolates was
established by whole pathogen sequencing, which offers a
considerable degree of confidence that the gut and the blood-
stream isolate were of the same origin. Asymptomatic carriers of
isogenic GBS and S. marcescens were also identified in time- and
physical-proximity to cases, but, overall, excretion of these
particular agents was rare. Hence, there appear to be
microclusters of colonization within the neonatal intensive care
unit (NICU) of infants' gut with pathogens that can invade the
blood. Such a finding is in accordance with data from a large
German network [30] suggesting that microclusters of late-
onset neonatal bloodstream infections occur, but that they are
often not identified as outbreaks in real-time.
A summary of studies that have attempted to match bacteria in
the blood with bacteria in the stool before the late-onset neonatal
bloodstream infection episode occurred is provided in Table 1.
Most studies of the role of the gut as the habitat/reservoir for
pathogens that subsequently cause late-onset neonatal blood-
stream infections exclude Gram-positive bacteria. However, MRSA
[33,34] and methicillin-susceptible S. aureus [35] are found in the
gut [36] and stool has been recommended as an optimal specimen
for their surveillance [37,38] in preterm infants. Coagulase-negative
Staphylococci, which are bona fide pathogens in preterm infants
[6,14,39], also have a gastrointestinal reservoir [40]. Staphylococcus
as a genus is well represented in the stools of preterm infants,
especially in the first week of life [41], an observation that also
demonstrates that the gut harbors Gram-positive bacteria with
particular ability to invade the bloodstream, even though such or-
ganisms are generally considered to colonize the skin. Finally, the
origins of GBS that cause late-onset neonatal bloodstream in-
fections remain enigmatic. Intra-institutional outbreaks of late-
onset GBS bloodstream infections are rare [42e44], suggesting
that the bloodstream isolate is probably acquired by infants during
birth. However, Carl et al. [13] found fecal excretion of GBS by in-
fants who subsequently developed a late-onset neonatal blood-
stream infection with an isogenic GBS, after an earlier-in-life
interval in which GBS was not recoverable from the stool, and, on
occasion, the same GBS colonized infants in the vicinity of the cases.
These data now prompt us to reconsider in-neonatal intensive care
unit (NICU) acquisition of the pathogen, and the infant gut as the
pre-invasive habitat.
Interestingly, many of these studies identified prior antibiotic
use (though variably defined) as a risk factor for bloodstream
infection. No doubt sicker children in general receive more antibi-
otics and are more prone to late-onset neonatal bloodstream in-
fections. However, animal data [45] support the role of antibiotics
4 P.I. Tarr, B.B. Warner / Seminars in Fetal & Neonatal Medicine xxx (2016) 1e6

Table 1
Studies attempting to link bloodstream infections to pre-infection gut bacteria.

Study Technology applied Findings Comments

Almuneef et al. [31] Antibiogram and bacterial identification
of individual isolates from stool
Graham et al. [25] Prior to bloodstream infection, infants' stool
isolates were saved and these organisms were
compared (antibiograms, species) to the
subsequent bloodstream isolate from these infants
Smith et al. [26] Prospective stool culture-based isolates
were saved and compared to subsequent
bloodstream isolates by gentamicin
susceptibility profiling
Stewart et al. [32] 454 pyrosequencing and denaturing
gradient gel electrophoresis analysis of
stools prior to the development of late-onset
neonatal bloodstream infection
Taft et al. [28] 16S rRNA gene sequencing of stool prior to
the development of late-onset neonatal
bloodstream infection, supplemented by stool
shotgun sequencing
Carl et al. [13] Stools stored culture-based recovery of cognate
pathogen in stools collected prior to the
onset of episodes of late-onset neonatal
bloodstream infections in pre-term infants.
Antibiograms and other phenotypes, and whole
genome sequencing, were used to confirm or
refute identity between blood and fecal isolates.
Shaw et al. [29] 16S rRNA gene sequencing from fecal DNA.
Some samples underwent bacterial
isolation attempts.
Two of three E. coli bloodstream infections
were matched to fecal isolates by
non-sequence typing methodology
(antibiograms).
There was considerable concordance
between the bloodstream isolates
and bacteria in stool prior to the infection
Isolates from the surveillance
culture obtained a mean of six
days before the episode were highly
predictive of gentamicin-resistant
subsequent infections (100% sensitivity,
98% specificity, 94% positive predictive value)
Five infants developed late-onset neonatal
bloodstream infections, and two of these
had NEC. In one child there was an
apparent match between fecal
carriage of coagulase-negative staphylococci
and subsequent bloodstream infection.
Bacteria were compared at the genus and
species level based on sequence data from
the stool and compared to the specific
isolate from the blood
Isogenicity was demonstrated between
bloodstream and pre-bloodstream
infection isolates of pathogens in
7 of 11 intensively studied children.
Staphylococci in stool preceded staphylococcal
bloodstream infection. Antibiograms
of blood and fecal isolates were concordant.
Simultaneous carriage by
infants without bloodstream infection
of similar organisms as those causing
bloodstream infections was demonstrated.
Sole focus was on Gram-negative bacteria
Associations between fecal carriage and
bloodstream isolation were not as strong
for gentamicin-susceptible Gram-negative
pathogens
No attempt to recover bloodstream
isolate from stool
Many episodes were associated with
prior fecal excretion, but the degree
of comparison is limited because the
fecal analysis was not based on isolated
pathogen sequencing
These organisms were
quite rare in two different control groups.
Colonization was demonstrated for variable
intervals prior to clinical deterioration.
GBS unexpectedly was harbored in the gut.
Bloodstream infections are preceded by
overrepresentation of taxa in the gut that
are plausibly related to that event,
including staphylococcal colonization.

in promoting bacterial translocation through the colon wall to
extra-intestinal sites, so a direct causal association cannot be
excluded.
4. Probiotics to prevent late-onset neonatal bloodstream
infections
purity of probiotic formulations and the risk of serious neonatal
systemic infections from probiotics [49,50], we cannot endorse
this intervention to prevent late-onset neonatal bloodstream
infections.
5. The future

Probiotics offer a possible intervention to prevent late-onset
neonatal bloodstream infections, especially those of gut origin.
Two recent meta-analyses have addressed the efficacy of pro-
biotics in preventing late-onset neonatal bloodstream infections.
Zhang et al. [46] proposed that probiotics might prevent blood-
stream infections in low birthweight infants, but no effect was
seen in children weighing <1000 g at birth. Rao et al. also
concluded that probiotics had a preventive effect [47] and like-
wise found that infants in a subgroup analysis (those born at <28
weeks' gestation and/or weighing <1000 g at birth) did not
benefit from this intervention. Late-onset neonatal bloodstream
infections were an outcome in the large probiotic study recently
reported by Costeloe et al. [48]. This study failed to demonstrate
any benefit of Bifidobacterium breve BBG-001 in preventing late-
onset neonatal bloodstream infections. As with all meta-
analyses of studies of probiotics, conclusions are necessarily
limited by the various probiotics chosen in these interventions,
differing dosage regimens and durations, and single versus
combination probiotics. In view of continuing concerns about
Worldwide, neonatologists approach late-onset neonatal
bloodstream infections in a reactive “rule out sepsis” mode. It is
critical that neonatologists remain vigilant and respond to the
earliest signs of bloodstream infections in all preterm infants, as
these are not normal hosts. However, in many cases, the specific
pathogens that subsequently invade the bloodstream may be found
in the stool prior to clinical signs of systemic infection. Moreover,
accumulating data suggest that colonization with specific blood-
stream invaders appears to be relatively rare among infants at risk.
Hence, such colonization might be an ultimate biomarker of soon-
to-be-apparent sepsis.
Ideally, colonization with bacteria with enhanced capability of
invading the host would be avoided altogether in NICUs, but a sec-
ondary goal should be the prevention of systemic dissemination
among the colonized. If further studies confirm these findings, then it
will be appropriate to consider bacteria-canonical approaches to
prevent late-onset neonatal bloodstream infections. Specifically, all
preterm infants will be considered to be at similar risk according to
their degree of prematurity, and risk reduction will be focused on the

Please cite this article in press as: Tarr PI, Warner BB, Gut bacteria and late-onset neonatal bloodstream infections in preterm infants, Seminars
in Fetal & Neonatal Medicine (2016), http://dx.doi.org/10.1016/j.siny.2016.06.002
 P.I. Tarr, B.B. Warner / Seminars in Fetal & Neonatal Medicine xxx (2016) 1e6
events that lead to acquisition or invasion of the bloodstream of the
subset of infants harbouring a bacterium, generally in the gut, with
exceptional pathogenic potential. Such bacteria-canonical ap-
proaches will quite likely involve microbial monitoring of the hosts,
and require more extensive knowledge of the genomics of neonatal
pathogens than exists now, so as to decide appropriate taxonomic
targets at the level of species, serotype, pathotype, or clade, as
appropriate, on a pathogen-by-pathogen basis within a genus.
Possible risk mitigation strategies among infants colonized by an
organism with exceptional pathogenic potential might include at-
tempts to alter the gut microbial community by diet or probiotics,
implementation of selective bowel decontamination, administration
of antimicrobial peptides, or strategic use of short-term parenteral
antibiotics in the host (the perinatal GBS prevention model).
Interventions to prevent late-onset neonatal bloodstream in-
fections that occur according to the host-canonical paradigm
should focus on repair or avoidance of defects in barrier integrity,
and of acceleration of maturation of host defense mechanisms.
These goals are less easily defined and might be more difficult to
achieve than would be the avoidance of colonization with a specific
pathogen, and prevention of bloodstream invasion among the
colonized. However, our advancing understanding of content and
biological effects of neonatal gut bacterial communities, and of
accelerating efforts to understand and to optimize neonatal nutri-
tion, will, we hope, soon optimize preterm infant gut health
strategies.
Certainly, the relative importance of host susceptibility and of
bacterial-specific processes could vary according to the physiologic
state of the particular infant, and the genomic content of the
invading pathogen. In any event, it is clear that expansion of
bloodstream infection prevention practices to include sources
beyond central lines will be needed to diminish the burden of late-
onset neonatal bloodstream infections.
Practice points
 Line care and hand hygiene have considerable merit, but
these interventions are probably not adequate by them-
selves to prevent bloodstream infections from bacteria
that invade the host from the gut.
Research directions
 How can we identify bacteria that have exceptional ability
to invade the bloodstream, among the greater bacterial
community in the preterm infant gut?
 How can we prevent colonization of preterm infants with
bacteria that have exceptional ability to invade the
bloodstream?
 How can we confine such pathogens to the gut, among
infants who are colonized?
Acknowledgements
We are grateful to Ms Maida Redzic for assistance with manu-
script preparation, Drs Michael Carl and Carey-Ann Burnham for
helpful conversations, and to Dr Chrisoph Haertel for critical
reading of the manuscript.
Please cite this article in press as: Tarr PI, Warner BB, Gut bacteria and late-onset neonatal bloodstream infections in preterm infants, Seminars
in Fetal & Neonatal Medicine (2016), http://dx.doi.org/10.1016/j.siny.2016.06.002
5
Conflict of interest statement
None declared.
Funding sources
This work was supported by NIH Grants UH3AI083265,
P30DK052574 (for the Biobank Core) and the Children's Discovery
Institute of Washington University and the St Louis Children's
Hospital.
References
[1] Bakhuizen SE, de Haan TR, Teune MJ, van Wassenaer-Leemhuis AG, van der
Heyden JL, van der Ham DP, et al. Meta-analysis shows that infants who have
suffered neonatal sepsis face an increased risk of mortality and severe com-
plications. Acta Paediatr 2014;103(12):1211e8.
[2] Bassler D, Stoll BJ, Schmidt B, Asztalos EV, Roberts RS, Robertson CM, et al.
Using a count of neonatal morbidities to predict poor outcome in extremely
low birth weight infants: added role of neonatal infection. Pediatrics
2009;123(1):313e8.
[3] Benjamin Jr DK, Stoll BJ, Fanaroff AA, McDonald SA, Oh W, Higgins RD, et al.
Neonatal candidiasis among extremely low birth weight infants: risk factors,
mortality rates, and neurodevelopmental outcomes at 18 to 22 months. Pe-
diatrics 2006;117(1):84e92.
[4] Schlapbach LJ, Aebischer M, Adams M, Natalucci G, Bonhoeffer J, Latzin P, et al.
Impact of sepsis on neurodevelopmental outcome in a Swiss National Cohort
of extremely premature infants. Pediatrics 2011;128(2):e348e57.
[5] Klinger G, Levy I, Sirota L, Boyko V, Lerner-Geva L, Reichman B, et al. Outcome
of early-onset sepsis in a national cohort of very low birth weight infants.
Pediatrics 2010;125(4):e736e40.
[6] Ohlin A, Bjorkman L, Serenius F, Schollin J, Kallen K. Sepsis as a risk factor for
neonatal morbidity in extremely preterm infants. Acta Paediatr 2015;104(11):
1070e6.
[7] Wheater M, Rennie JM. Perinatal infection is an important risk factor for ce-
rebral palsy in very-low-birthweight infants. Dev Med Child Neurol
2000;42(6):364e7.
[8] Phares CR, Lynfield R, Farley MM, Mohle-Boetani J, Harrison LH, Petit S, et al.
Epidemiology of invasive group B streptococcal disease in the United States,
1999e2005. JAMA 2008;299(17):2056e65.
[9] Bauserman MS, Laughon MM, Hornik CP, Smith PB, Benjamin Jr DK, Clark RH,
et al. Group B Streptococcus and Escherichia coli infections in the intensive
care nursery in the era of intrapartum antibiotic prophylaxis. Pediatr Infect Dis
J 2013;32(3):208e12.
[10] Alexander VN, Northrup V, Bizzarro MJ. Antibiotic exposure in the newborn
intensive care unit and the risk of necrotizing enterocolitis. J Pediatr
2011;159(3):392e7.
[11] Cordero L, Rau R, Taylor D, Ayers LW. Enteric gram-negative bacilli blood-
stream infections: 17 years' experience in a neonatal intensive care unit. Am J
Infect Control 2004;32(4):189e95.
[12] Lean WL, Kamlin CO, Garland SM, Jacobs SE. Stable rates of neonatal sepsis in a
tertiary neonatal unit. J Paediatr Child Health 2014;51(3):294e9.
[13] Carl MA, Ndao IM, Springman AC, Manning SD, Johnson JR, Johnston BD, et al.
Sepsis from the gut: the enteric habitat of bacteria that cause late-onset
neonatal bloodstream infections. Clin Infect Dis 2014;58(9):1211e8.
[14] Boghossian NS, Page GP, Bell EF, Stoll BJ, Murray JC, Cotten CM, et al. Late-
onset sepsis in very low birth weight infants from singleton and multiple-
gestation births. J Pediatr 2013;162(6):1120e4. 4 e1.
[15] Padula MA, Dewan ML, Shah SS, Padula AM, Srinivasan L, McGowan KL, et al.
Risk factors associated with laboratory-confirmed bloodstream infections in a
tertiary neonatal intensive care unit. Pediatr Infect Dis J 2014;33(10):
1027e32.
[16] Zachariah P, Furuya EY, Edwards J, Dick A, Liu H, Herzig CT, et al. Compliance
with prevention practices and their association with central line-associated
bloodstream infections in neonatal intensive care units. Am J Infect Control
2014;42(8):847e51.
[17] Kaplan HC, Lannon C, Walsh MC, Donovan EF, Ohio Perinatal Quality C. Ohio
statewide quality-improvement collaborative to reduce late-onset sepsis in
preterm infants. Pediatrics 2011;127(3):427e35.
[18] Schulman J, Stricof R, Stevens TP, Horgan M, Gase K, Holzman IR, et al.
Statewide NICU central-line-associated bloodstream infection rates decline
after bundles and checklists. Pediatrics 2011;127(3):436e44.
[19] Harris AD, McGregor JC, Johnson JA, Strauss SM, Moore AC, Standiford HC,
et al. Risk factors for colonization with extended-spectrum beta-lactamase-
producing bacteria and intensive care unit admission. Emerg Infect Dis
2007;13(8):1144e9.
[20] Blot S, Depuydt P, Vogelaers D, Decruyenaere J, De Waele J, Hoste E, et al.
Colonization status and appropriate antibiotic therapy for nosocomial
bacteremia caused by antibiotic-resistant gram-negative bacteria in an
intensive care unit. Infect Control Hosp Epidemiol 2005;26(6):575e9.
6 P.I. Tarr, B.B. Warner / Seminars in Fetal & Neonatal Medicine xxx (2016) 1e6
[21] Reddy P, Malczynski M, Obias A, Reiner S, Jin N, Huang J, et al. Screening for
extended-spectrum beta-lactamase-producing Enterobacteriaceae among
high-risk patients and rates of subsequent bacteremia. Clin Infect Dis
2007;45(7):846e52.
[22] Price R, MacLennan G, Glen J, Su DC. Selective digestive or oropharyngeal
decontamination and topical oropharyngeal chlorhexidine for prevention of
death in general intensive care: systematic review and network meta-anal-
ysis. BMJ 2014;348:g2197.
[23] Plantinga NL, Bonten MJ. Selective decontamination and antibiotic resistance
in ICUs. Crit Care 2015;19:259e65.
[24] Daneman N, Sarwar S, Fowler RA, Cuthbertson BH, Su DCSG. Effect of selective
decontamination on antimicrobial resistance in intensive care units: a sys-
tematic review and meta-analysis. Lancet Infect Dis 2013;13(4):328e41.
[25] Graham 3rd PL, Della-Latta P, Wu F, Zhou J, Saiman L. The gastrointestinal
tract serves as the reservoir for Gram-negative pathogens in very low birth
weight infants. Pediatr Infect Dis J 2007;26(12):1153e6.
[26] Smith A, Saiman L, Zhou J, Della-Latta P, Jia H, Graham 3rd PL. Concordance of
gastrointestinal tract colonization and subsequent bloodstream infections
with gram-negative bacilli in very low birth weight infants in the neonatal
intensive care unit. Pediatr Infect Dis J 2010;29(9):831e5.
[27] Madan JC, Salari RC, Saxena D, Davidson L, O'Toole GA, Moore JH, et al. Gut
microbial colonisation in premature neonates predicts neonatal sepsis. Archs
Dis Childh Fetal Neonatal Ed 2012;97(6):F456e62.
[28] Taft DH, Ambalavanan N, Schibler KR, Yu Z, Newburg DS, Deshmukh H, et al.
Center variation in intestinal microbiota prior to late-onset sepsis in preterm
infants. PloS One 2015;10(6):e0130604.
[29] Shaw AG, Sim K, Randell P, Cox MJ, McClure ZE, Li MS, et al. Late-onset
bloodstream infection and perturbed maturation of the gastrointestinal
microbiota in premature infants. PloS One 2015;10(7):e0132923.
[30] Haertel C, Faust K, Avenarius S, Bohnhorst B, Emeis M, Gebauer C, et al.
Epidemic microclusters of blood-culture proven sepsis in very-low-birth
weight infants: experience of the German Neonatal Network. PloS one
2012;7(6):e38304.
[31] Almuneef MA, Baltimore RS, Farrel PA, Reagan-Cirincione P, Dembry LM.
Molecular typing demonstrating transmission of gram-negative rods in a
neonatal intensive care unit in the absence of a recognized epidemic. Clin
Infect Dis 2001;32(2):220e7.
[32] Stewart CJ, Marrs EC, Nelson A, Lanyon C, Perry JD, Embleton ND, et al.
Development of the preterm gut microbiome in twins at risk of necrotising
enterocolitis and sepsis. PloS one 2013;8(8):e73465.
[33] Boyce JM, Havill NL, Otter JA, Adams NM. Widespread environmental
contamination associated with patients with diarrhea and methicillin-resis-
tant Staphylococcus aureus colonization of the gastrointestinal tract. Clin
Infect Dis 2007;28(10):1142e7.
[34] Donskey CJ. The role of the intestinal tract as a reservoir and source for
transmission of nosocomial pathogens. Clin Infect Dis 2004;39(2):219e26.
[35] Squier C, Rihs JD, Risa KJ, Sagnimeni A, Wagener MM, Stout J, et al. Staphy-
lococcus aureus rectal carriage and its association with infections in patients
in a surgical intensive care unit and a liver transplant unit. Infect Control Hosp
Epidemiol 2002;23(9):495e501.
Please cite this article in press as: Tarr PI, Warner BB, Gut bacteria and late-onset neonatal bloodstream infections in preterm infants, Seminars
in Fetal & Neonatal Medicine (2016), http://dx.doi.org/10.1016/j.siny.2016.06.002
[36] Singh K, Gavin PJ, Vescio T, Thomson Jr RB, Deddish RB, Fisher A, et al.
Microbiologic surveillance using nasal cultures alone is sufficient for detection
of methicillin-resistant Staphylococcus aureus isolates in neonates. J Clin
Microbiol 2003;41(6):2755e7.
[37] Heinrich N, Mueller A, Bartmann P, Simon A, Bierbaum G, Engelhart S. Suc-
cessful management of an MRSA outbreak in a neonatal intensive care unit.
Eur J Clin Microbiol Infect Dis 2011;30(7):909e13.
[38] Nakao A, Ito T, Han X, Lu YJ, Hisata K, Tsujiwaki A, et al. Intestinal carriage of
methicillin-resistant Staphylococcus aureus in nasal MRSA carriers hospital-
ized in the neonatal intensive care unit. Antimicrob Resist Infect Control
2014;3:14e24.
[39] Stoll BJ, Gordon T, Korones SB, Shankaran S, Tyson JE, Bauer CR, et al. Late-
onset sepsis in very low birth weight neonates: a report from the National
Institute of Child Health and Human Development Neonatal Research
Network. J Pediatr 1996;129(1):63e71.
[40] Aujoulat F, Roudiere L, Picaud JC, Jacquot A, Filleron A, Neveu D, et al. Tem-
poral dynamics of the very premature infant gut dominant microbiota. BMC
Microbiology 2014;14:325e39.
[41] La Rosa PS, Warner BB, Zhou Y, Weinstock GM, Sodergren E, Hall-Moore CM,
et al. Patterned progression of bacterial populations in the premature infant
gut. Proc Natl Acad Sci USA 2014;111(34):12522e7.
[42] MacFarquhar JK, Jones TF, Woron AM, Kainer MA, Whitney CG, Beall B, et al.
Outbreak of late-onset group B Streptococcus in a neonatal intensive care
unit. Am J Infect Control 2010;38(4):283e8.
[43] Elling R, Hufnagel M, de Zoysa A, Lander F, Zumstein K, Krueger M, et al.
Synchronous recurrence of group B streptococcal late-onset sepsis in twins.
Pediatrics 2014;133(5):e1388e91.
[44] Al-Maani A, Streitenberger L, Clarke M, Yau YC, Kovach D, Wray R, et al.
Nosocomial transmission of group B streptococci proven by positive envi-
ronmental culture. Oman Med J 2014;29(5):376e9.
[45] Knoop KA, McDonald KG, Kulkarni DH, Newberry RD. Antibiotics promote
inflammation through the translocation of native commensal colonic bacteria.
Gut 2015:1e10.
[46] Zhang GQ, Hu HJ, Liu CY, Shakya S, Li ZY. Probiotics for preventing late-onset
sepsis in preterm neonates: a PRISMA-compliant systematic review and
meta-analysis of randomized controlled trials. Medicine 2016;95(8):e2581.
[47] Rao SC, Athalye-Jape GK, Deshpande GC, Simmer KN, Patole SK. Probiotic
supplementation and late-onset sepsis in preterm infants: a meta-analysis.
Pediatrics 2016;137(3):1e16.
[48] Costeloe K, Hardy P, Juszczak E, Wilks M, Millar MR, Probiotics in Preterm
Infants Study Collaborative G. Bifidobacterium breve BBG-001 in very preterm
infants: a randomised controlled phase 3 trial. Lancet 2016;387(10019):
649e60.
[49] Vallabhaneni S, Walker TA, Lockhart SR, Ng D, Chiller T, Melchreit R, et al.
Notes from the field: Fatal gastrointestinal mucormycosis in a premature
infant associated with a contaminated dietary supplementeConnecticut,
2014. Morb Mortal Wkly Rep 2015;64(6):155e6.
[50] Brecht M, Garg A, Longstaff K, Cooper C, Andersen C. Lactobacillus sepsis
following a laparotomy in a preterm infant: a note of caution. Neonatology
2016;109(3):186e9.