3. FROM THE AMERICAN ACADEMY OF PEDIATRICS
PATHOGENESIS AND most of these women with histologic The major risk factors for early-onset
EPIDEMIOLOGY OF EARLY-ONSET chorioamnionitis do not have a positive neonatal sepsis are preterm birth,
SEPSIS placental culture.3 The incidence of clin- maternal colonization with GBS, rupture
ical chorioamnionitis varies inversely of membranes >18 hours, and mater-
Before birth, the fetus optimally is
with gestational age. In the National nal signs or symptoms of intra-amniotic
maintained in a sterile environment.
Institute of Child Health and Human infection.14–16 Other variables include
Organisms causing early-onset sepsis
Development Neonatal Research Net- ethnicity (ie, black women are at higher
ascend from the birth canal either
work, 14% to 28% of women delivering risk of being colonized with GBS), low
when the amniotic membranes rupture
preterm infants at 22 through 28 weeks’ socioeconomic status, male sex, and
or leak before or during the course of
gestation exhibited signs compatible low Apgar scores. Preterm birth/low
labor, resulting in intra-amniotic infec-
with chorioamnionitis.4 The major risk birth weight is the risk factor most
tion.2 Commonly referred to as “cho-
factors for chorioamnionitis include closely associated with early-onset sep-
rioamnionitis,” intra-amniotic infection
low parity, spontaneous labor, longer sis.17 Infant birth weight is inversely
indicates infection of the amniotic fluid,
length of labor and membrane rupture, related to risk of early-onset sepsis.
membranes, placenta, and/or decidua.
multiple digital vaginal examinations The increased risk of early-onset sep-
Group B streptococci (GBS) can also sis in preterm infants is also related to
(especially with ruptured membranes),
enter the amniotic fluid through occult complications of labor and delivery
meconium-stained amniotic fluid, internal
tears. Chorioamnionitis is a major risk and immaturity of innate and adaptive
fetal or uterine monitoring, and pres-
factor for neonatal sepsis. Sepsis can immunity.18
begin in utero when the fetus inhales ence of genital tract microorganisms
or swallows infected amniotic fluid. (eg, Mycoplasma hominis).5
The neonate can also develop sepsis in At term gestation, less than 1% of DIAGNOSTIC TESTING FOR SEPSIS
the hours or days after birth when women with intact membranes will The clinical diagnosis of sepsis in the
colonized skin or mucosal surfaces are have organisms cultured from amni- neonate is difficult, because many of
compromised. The essential criterion otic fluid.6 The rate can be higher if the signs of sepsis are nonspecific and
for the clinical diagnosis of chorio- the integrity of the amniotic cavity is are observed with other noninfectious
amnionitis is maternal fever. Other compromised by procedures before conditions. Although a normal physical
criteria are relatively insensitive. When birth (eg, placement of a cerclage or examination is evidence that sepsis is
defining intra-amniotic infection (cho- amniocentesis).6 In women with pre- not present,19,20 bacteremia can occur
rioamnionitis) for clinical research term labor and intact membranes, the in the absence of clinical signs.21 Avail-
studies, the diagnosis is typically based rate of microbial invasion of the amni- able diagnostic testing is not helpful in
on the presence of maternal fever of otic cavity is 32%, and if there is pre- deciding which neonate requires em-
greater than 38°C (100.4°F) and at least term premature rupture of membranes pirical antimicrobial therapy but can
two of the following criteria: maternal (PPROM), the rate may be as high as assist with the decision to discontinue
leukocytosis (greater than 15 000 cells/ 75%.7 Many of the pathogens recovered treatment.22
mm3), maternal tachycardia (greater from amniotic fluid in women with pre-
than 100 beats/minute), fetal tachycar- term labor or PPROM (eg, Ureaplasma Blood Culture
dia (greater than 160 beats/minute), species or Mycoplasma species) do A single blood culture in a sufficient
uterine tenderness, and/or foul odor of not cause early-onset sepsis.8–10 How- volume is required for all neonates
the amniotic fluid. These thresholds are ever, both Ureaplasma and Myco- with suspected sepsis. Data suggest
associated with higher rates of neo- plasma organisms can be recovered that 1.0 mL of blood should be the
natal and maternal morbidity. from the bloodstream of infants whose minimum volume drawn for culture
Nonetheless, the diagnosis of cho- birth weight is less than 1500 g.11 When when a single pediatric blood culture
rioamnionitis must be considered even a pathogen (eg, GBS) is recovered from bottle is used. Dividing the specimen in
when maternal fever is the sole abnor- amniotic fluid, the attack rate of neo- half and inoculating aerobic and an-
mal finding. Although fever is common natal sepsis can be as high as 20%.12 aerobic bottles is likely to decrease the
in women who receive epidural anes- Infants born to women with PPROM sensitivity. Although 0.5 mL of blood
thesia (15%–20%), histologic evidence who are colonized with GBS have an has previously been considered ac-
of acute chorioamnionitis is very com- estimated attack rate of 33% to 50% ceptable, in vitro data from Schelonka
mon in women who become febrile when intrapartum prophylaxis is not et al demonstrated that 0.5 mL would
after an epidural (70.6%).3 Furthermore, given.13 not reliably detect low-level bacteremia
PEDIATRICS Volume 129, Number 5, May 2012 1007
Downloaded from pediatrics.aappublications.org at Hacettepe University on May 9, 2012
4. (4 colony-forming units [CFU]/mL or Body Surface Cultures infants with “traumatic taps” (or
less). 23 Furthermore, up to 25% of Bacterial cultures of the axilla, groin, nonbacterial illnesses), the mean
infants with sepsis have low colony and the external ear canal have a poor number of white blood cells in un-
count bacteremia (≤4 CFU/mL), and positive predictive accuracy. They are infected preterm or term infants was
two-thirds of infants younger than 2 expensive and add little to the evalu- consistently <10 cells/mm3.44–50 Cell
months of age have colony counts <10 ation of an infant with possible bac- counts 2 standard deviations from the
CFU/mL.24,25 Neal et al demonstrated terial sepsis.34,35 mean were generally less than 20
that more than half of blood specimens cells/mm3.46 In a study by Garges et al,
inoculated into the aerobic bottle were Tracheal Aspirates the median number of white blood cells
less than 0.5 mL.26 A study by Connell in infants who were born at greater
Cultures and Gram stains of tracheal
et al indicated that blood cultures with than 34 weeks’ gestation and had
aspirate specimens may be of value if
an adequate volume were twice as bacterial meningitis was 477/mm3.43
obtained immediately after endotra-
likely to yield a positive result.27 A blood In contrast, the median number of white
cheal tube placement.36 Once an infant
culture obtained through an umbilical blood cells in infants who were born at
has been intubated for several days,
artery catheter shortly after placement less than 34 weeks’ gestation and had
tracheal aspirates are of no value in
for other clinical indications is an ac- meningitis was 110/mm3.51 Infants with
the evaluation of sepsis.37
ceptable alternative to a culture drawn meningitis attributable to Gram-negative
from a peripheral vein.28 The risk of pathogens typically have higher CSF
Lumbar Puncture
recovering a contaminant is greater white blood cell counts than do infants
with a blood culture drawn from an The decision to perform a lumbar punc-
with meningitis attributable to Gram-
umbilical vein.29 There are, however, ture in a neonate with suspected early-
positive pathogens.52 Adjusting the
data to suggest that a blood culture onset sepsis remains controversial. In
CSF white blood cell count for the
drawn from the umbilical vein at the the high-risk, healthy-appearing in-
number of red blood cells does not
time of delivery using a doubly clam- fant, data suggest that the likelihood
of meningitis is extremely low.38 In the improve the diagnostic utility (loss of
ped and adequately prepared segment sensitivity with marginal gain in speci-
of the cord is a reliable alternative to infant with clinical signs that are thought
to be attributable to a noninfectious ficity).53 In addition, the number of bands
a culture obtained peripherally.30
condition, such as respiratory distress in a CSF specimen does not predict
syndrome, the likelihood of meningitis meningitis.54 With a delay in analysis
Urine Culture (>2 hours), white blood cell counts
is also low.39 However, in bacteremic
A urine culture should not be part of the infants, the incidence of meningitis may and glucose concentrations decrease
sepsis workup in an infant with suspected be as high as 23%.40,41 Blood culture significantly.55
early-onset sepsis.31 Unlike urinary tract alone cannot be used to decide who Protein concentrations in uninfected,
infections in older infants (which are needs a lumbar puncture, because term newborn infants are <100 mg/
usually ascending infections), urinary blood cultures can be negative in up dL.44–50 Preterm infants have CSF pro-
tract infections in newborn infants are to 38% of infants with meningitis.42,43 tein concentrations that vary inversely
attributable to seeding of the kidney The lumbar puncture should be per- with gestational age. In the normogly-
during an episode of bacteremia. formed in any infant with a positive cemic newborn infant, glucose con-
blood culture, infants whose clinical centrations in CSF are similar to those
Gastric Aspirates course or laboratory data strongly in older infants and children (70%–80%
The fetus swallows 500 to 1000 mL of suggest bacterial sepsis, and infants of a simultaneously obtained blood
amniotic fluid each day. Therefore, if who initially worsen with antimicro- specimen). A low glucose concentration
there are white blood cells present in bial therapy. For any infant who is is the CSF variable with the greatest
amniotic fluid, they will be present in critically ill and likely to have cardio- specificity for the diagnosis of menin-
gastric aspirate specimens at birth. vascular or respiratory compromise gitis.43,51 Protein concentrations are
However, these cells represent the ma- from the procedure, the lumbar punc- higher and glucose concentrations are
ternal response to inflammation and ture can be deferred until the infant is lower in term than in preterm infants
have a poor correlation with neonatal more stable. with meningitis. However, meningitis
sepsis.32 Gram stains of gastric aspirates Cerebrospinal fluid (CSF) values indic- occurs in infants with normal CSF
to identify bacteria are of limited value ative of neonatal meningitis are con- values, and some of these infants have
and are not routinely recommended.33 troversial. In studies that have excluded high bacterial inocula.43,51
1008 FROM THE AMERICAN ACADEMY OF PEDIATRICS
Downloaded from pediatrics.aappublications.org at Hacettepe University on May 9, 2012
5. FROM THE AMERICAN ACADEMY OF PEDIATRICS
Peripheral White Blood Cell Count 28 through 36 weeks’ gestation, and Counts obtained 6 to 12 hours after
and Differential Count 500/mm3 in infants born at <28 weeks’ birth are more likely to be abnormal
Total white blood cell counts have little gestation. Peak values occurred at 6 to than are counts obtained at birth, be-
value in the diagnosis of early-onset 8 hours after birth; the lower limits of cause alterations in the numbers (and
sepsis and have a poor positive pre- normal at that time were 7500/mm3, ratios) of mature and immature neutro-
dictive accuracy. 56,57 Many investi- 3500/mm3, and 1500/mm3 for infants phils require an established inflammatory
gators have analyzed subcomponents born at >36 weeks’ gestation, 28 to response. Therefore, once the decision is
of the white blood cell count (neutrophil 36 weeks’ gestation, and <28 weeks’ made to start antimicrobial therapy
indices)—absolute neutrophil count, gestation, respectively.61 It is notewor- soon after birth, it is worth waiting 6 to
absolute band count, and immature to thy that the study by Schmutz et al was 12 hours before ordering a white blood
total neutrophil (I/T)ratio—to identify performed at 4800 feet above sea level, cell count and differential count.68,69
infected infants. Like most diagnostic whereas that of Manroe et al was per-
formed at 500 feet above sea level. Platelet Counts
tests for neonatal sepsis, neutrophil in-
dices have proven most useful for ex- The absolute immature neutrophil count Despite the frequency of low platelet
follows a similar pattern to the absolute counts in infected infants, they are a
cluding infants without infection rather
neutrophil count and peaks at approx- nonspecific, insensitive, and late indica-
than identifying infected neonates. Neu-
imately 12 hours of life. The number of tor of sepsis.70,71 Moreover, platelet
tropenia may be a better marker for
immature neutrophils increases from a counts are not useful to follow clinical
neonatal sepsis and has better speci-
maximal value of 1100 cells/mm3 at response to antimicrobial agents, be-
ficity than an elevated neutrophil count,
birth to 1500 cells/mm3 at 12 hours of cause they often remain depressed for
because few conditions besides sepsis
age.58 Absolute immature counts have days to weeks after a sepsis episode.
(maternal pregnancy-induced hyper-
tension, asphyxia, and hemolytic dis- a poor sensitivity and positive predic- Acute-Phase Reactants
ease) depress the neutrophil count of tive accuracy for early-onset sepsis.22
Furthermore, if exhaustion of bone mar- A wide variety of acute-phase reactants
neonates.58 The definitions for neutro-
row reserves occurs, the number of im- have been evaluated in neonates with
penia vary with gestational age,58–61 suspected bacterial sepsis. However, only
type of delivery (infants born by cesar- mature forms will remain depressed.64
C-reactive protein (CRP) and procalcito-
ean delivery without labor have lower The I/T ratio has the best sensitivity of nin concentrations have been investiga-
counts than infants delivered vagi- any of the neutrophil indices. However, ted in sufficiently large studies.72,73 CRP
nally),61 site of sampling (neutrophil with manual counts, there are wide concentration increases within 6 to 8
counts are lower in samples from interreader differences in band neu- hours of an infectious episode in neo-
arterial blood),62 and altitude (infants trophil identification.65 The I/T ratio is nates and peaks at 24 hours.74,75 The
born at elevated altitudes have higher <0.22 in 96% of healthy preterm infants sensitivity of a CRP determination is
total neutrophil counts).63 In late pre- born at <32 weeks’ gestational age.66 low at birth, because it requires an
term and term infants, the definition Unlike the absolute neutrophil count inflammatory response (with release
for neutropenia most commonly used and the absolute band count, maximum of interleukin-6) to increase CRP con-
is that suggested by Manroe et al normal values for the I/T ratio occur at centrations.76 The sensitivity improves
(<1800/mm3 at birth and <7800/mm3 birth (0.16) and decline with increasing dramatically if the first determination
at 12–14 hours of age).58 Schmutz et al postnatal age to a minimum value of is made 6 to 12 hours after birth. Benitz
reinvestigated these reference ranges 0.12.58 In healthy term infants, the 90th et al have demonstrated that excluding
using modern cell-counting instrumen- percentile for the I/T ratio is 0.27.59 a value at birth, 2 normal CRP deter-
tation in 30 254 infants born at 23 to 42 A single determination of the I/T ratio minations (8–24 hours after birth and
weeks’ gestation.61 Infants with diagnoses has a poor positive predictive accuracy 24 hours later) have a negative pre-
known to affect neutrophil counts (eg, (approximately 25%) but a very high dictive accuracy of 99.7% and a nega-
those born to women with pregnancy- negative predictive accuracy (99%).66 tive likelihood ratio of 0.15 for proven
induced hypertension or those with The I/T ratio may be elevated in 25% to neonatal sepsis.76 If CRP determina-
early-onset sepsis) were excluded. In 50% of uninfected infants.67 tions remain persistently normal, it is
this study, the lower limits of normal Exhaustion of bone marrow reserves strong evidence that bacterial sepsis is
for neutrophil values at birth were will result in low band counts and lead unlikely, and antimicrobial agents can be
3500/mm3 in infants born at >36 weeks’ to falsely low ratios. The timing of the safely discontinued. Data are insufficient
gestation, 1000/mm3 in infants born at white blood cell count is critical. 68 to recommend following sequential CRP
PEDIATRICS Volume 129, Number 5, May 2012 1009
Downloaded from pediatrics.aappublications.org at Hacettepe University on May 9, 2012
6. concentrations to determine the dura- Furthermore, scores obtained in the meningitis attributable to GBS is trea-
tion of antimicrobial therapy in an infant first several hours after birth have been ted for a minimum of 14 days.88 Other
with an elevated value (≥1.0 mg/dL). shown to have poorer sensitivity and focal infections secondary to GBS (eg,
Procalcitonin concentrations increase negative predictive value than scores cerebritis, osteomyelitis, endocarditis)
within 2 hours of an infectious episode, obtained at 24 hours of age.67 Sepsis are treated for longer durations.88 Gram-
peak at 12 hours, and normalize within screening panels commonly include negative meningitis is treated for
2 to 3 days in healthy adult volunteers.77 neutrophil indices and acute-phase re- minimum of 21 days or 14 days after
A physiologic increase in procalcitonin actants (usually CRP concentration). The obtaining a negative culture, whichever
concentration occurs within the first positive predictive value of the sepsis is longer.88 Treatment of Gram-negative
24 hours of birth, and an increase in screen in neonates is poor (<30%); meningitis should include cefotaxime
serum concentrations can occur with however, the negative predictive accuracy and an aminoglycoside until the results
noninfectious conditions (eg, respira- has been high (>99%) in small clinical of susceptibility testing are known.87,88
tory distress syndrome).78 Procalcitonin studies.22 Sepsis screening tests might be The duration of antimicrobial therapy
concentration has a modestly better of value in deciding which “high-risk” in infants with negative blood cultures
sensitivity than does CRP concentration healthy-appearing neonates do not need is controversial. Many women receive
but is less specific.73 Chiesa and col- antimicrobial agents or whether therapy antimicrobial agents during labor as
leagues have published normal values can be safely discontinued. prophylaxis to prevent early-onset GBS
for procalcitonin concentrations in term infections or for management of sus-
and preterm infants.79 There is evidence TREATMENT OF INFANTS WITH pected intra-amniontic infection or
from studies conducted in adult pop- SUSPECTED EARLY-ONSET SEPSIS PPROM. In those instances, postnatal
ulations, the majority of which focused blood cultures may be sterile (false
on patients with sepsis in the ICU, that In the United States, the most common negative). When considering the dura-
significant reductions in use of anti- pathogens responsible for early-onset tion of therapy in infants with negative
microbial agents can be achieved in neonatal sepsis are GBS and Escherichia blood cultures, the decision should
patients whose treatment is guided by coli.17 A combination of ampicillin and include consideration of the clinical
procalcitonin concentration.80 an aminoglycoside (usually gentamicin) course as well as the risks associated
is generally used as initial therapy, and with longer courses of antimicrobial
this combination of antimicrobial agents agents. In a retrospective study by Cor-
Sepsis Screening Panels also has synergistic activity against dero and Ayers, the average duration of
Hematologic scoring systems using GBS and Listeria monocytogenes.82,83 treatment in 695 infants (<1000 g)
multiple laboratory values (eg, white Third-generation cephalosporins (eg, with negative blood cultures was 5 ±
blood cell count, differential count, and cefotaxime) represent a reasonable al- 3 days.89 Cotten et al have suggested
platelet count) have been recommen- ternative to an aminoglycoside. However, an association with prolonged adminis-
ded as useful diagnostic aids. No matter several studies have reported rapid tration of antimicrobial agents (>5 days)
what combination of tests is used, the development of resistance when cefo- in infants with suspected early-onset
positive predictive accuracy of scoring taxime has been used routinely for the sepsis (and negative blood cultures)
systems is poor unless the score is treatment of early-onset neonatal sep- with death and necrotizing enterocoli-
very high. Rodwell et al described a sis,84 and extensive/prolonged use of tis.90 Two recent papers also support
scoring system in which a score of 1 was third-generation cephalosporins is a risk this association.91,92
assigned to 1 of 7 findings, including factor for invasive candidiasis.85 Be-
abnormalities of leukocyte count, total cause of its excellent CSF penetration,
neutrophil count, increased immature empirical or therapeutic use of cefo- PREVENTION STRATEGIES FOR
polymorphonuclear leukocyte (PMN) taxime should be restricted for use in EARLY-ONSET SEPSIS
count, increased I/T ratio, immature to infants with meningitis attributable to The only intervention proven to decrease
mature PMN ratio >0.3, platelet count Gram-negative organisms.86 Ceftriax- the incidence of early-onset neonatal
≤150 000/mm3, and pronounced degen- one is contraindicated in neonates sepsis is maternal treatment with
erative changes (ie, toxic granulations) because it is highly protein bound intrapartum intravenous antimicro-
in PMNs. 81 In this study, two-thirds and may displace bilirubin, leading to a bial agents for the prevention of GBS
of preterm infants and 90% of term risk of kernicterus. Bacteremia without an infections.93 Adequate prophylaxis is
infants with a hematologic score identifiable focus of infection is generally defined as penicillin (the preferred
≥3 did not have proven sepsis. 81 treated for 10 days.87 Uncomplicated agent), ampicillin, or cefazolin given for
1010 FROM THE AMERICAN ACADEMY OF PEDIATRICS
Downloaded from pediatrics.aappublications.org at Hacettepe University on May 9, 2012
7. FROM THE AMERICAN ACADEMY OF PEDIATRICS
≥4 hours before delivery. Erythromycin hours of life. Approximately 1% of infants starting antimicrobial agents after cul-
is no longer recommended for prophy- will appear healthy at birth and then tures have been obtained.
laxis because of high resistance rates. develop signs of infection after a vari-
In parturients who have a nonserious able time period.21 Every critically ill Challenge 2: Identifying
penicillin allergy, cefazolin is the drug infant should be evaluated and receive Healthy-Appearing Neonates With
of choice. For parturients with a history empirical broad-spectrum antimicrobial a “High Likelihood” of Early-Onset
of serious penicillin allergy (anaphy- therapy after cultures, even when there Sepsis Who Require Antimicrobial
laxis, angioedema, respiratory com- are no obvious risk factors for sepsis. Agents Soon After Birth
promise, or urticaria), clindamycin is The greatest difficulty faced by clini- This category includes infants with 1 of
an acceptable alternative agent, but cians is distinguishing neonates with the risk factors for sepsis noted pre-
only if the woman’s rectovaginal GBS early signs of sepsis from neonates viously (colonization with GBS, prolonged
screening isolate has been tested and with noninfectious conditions with rel- rupture of membranes >18 hours, or
documented to be susceptible. If the atively mild findings (eg, tachypnea with maternal chorioamnionitis). GBS is not
clindamycin susceptibility is unknown or without an oxygen requirement). In a risk factor if the mother has received
or the GBS isolate is resistant to clin- this situation, data are insufficient to adequate intrapartum therapy (penicil-
damycin, vancomycin is an alternative guide management. In more mature lin, ampicillin, or cefazolin for at least
agent for prophylaxis. However, nei- neonates without risk factors for in- 4 hours before delivery) or has a ce-
ther clindamycin nor vancomycin has fection who clinically improve over the sarean delivery with intact membranes
been evaluated for efficacy in pre- first 6 hours of life (eg, need for oxygen in the absence of labor.93 The risk of
venting early-onset GBS sepsis in is decreasing and respiratory distress infection in the newborn infant varies
neonates. Intrapartum antimicrobial is resolving), it is reasonable to with- considerably with the risk factor pres-
agents are indicated for the following hold antimicrobial therapy and monitor ent. The greatest risk of early-onset
situations93: the neonates closely. The 6-hour win- sepsis occurs in infants born to women
1. Positive antenatal cultures or molec- dow should not be considered absolute; with chorioamnionitis who are also
ular test at admission for GBS (ex- however, most infants without infec- colonized with GBS and did not receive
cept for women who have a cesarean tion demonstrate some improvement intrapartum antimicrobial agents. Early-
delivery without labor or membrane over that time period. Any worsening of onset sepsis does occur in infants who
rupture) the infant’s condition should prompt appear healthy at birth.21 Therefore,
2. Unknown maternal colonization sta-
tus with gestation <37 weeks, rup-
ture of membranes >18 hours, or
temperature >100.4°F (>38°C)
3. GBS bacteriuria during the current
pregnancy
4. Previous infant with invasive GBS
disease
Management guidelines for the new-
born infant have been published93 and
are available online (http://www.cdc.
gov/groupbstrep/guidelines/index.html).
CLINICAL CHALLENGES
Challenge 1: Identifying Neonates
With Clinical Signs of Sepsis With FIGURE 1
a “High Likelihood” of Early-Onset Evaluation of asymptomatic infants <37 weeks’ gestation with risk factors for sepsis. aThe diagnosis
of chorioamnionitis is problematic and has important implications for the management of the
Sepsis Who Require Antimicrobial newborn infant. Therefore, pediatric providers are encouraged to speak with their obstetrical
Agents Soon After Birth colleagues whenever the diagnosis is made. bLumbar puncture is indicated in any infant with
a positive blood culture or in whom sepsis is highly suspected on the basis of clinical signs, re-
Most infants with early-onset sepsis sponse to treatment, and laboratory results. IAP, intrapartum antimicrobial prophylaxis; WBC, white
exhibit abnormal signs in the first 24 blood cell; Diff, differential white blood cell count.
PEDIATRICS Volume 129, Number 5, May 2012 1011
Downloaded from pediatrics.aappublications.org at Hacettepe University on May 9, 2012
8. some clinicians use diagnostic tests initiating antimicrobial treatment gen- CONCLUSIONS
with a high negative predictive accuracy erally decreases with increasing num-
The diagnosis and management of neo-
as reassurance that infection is not bers of risk factors for infection and
nates with suspected early-onset sepsis
present (allowing them to withhold greater degrees of prematurity. Sug-
are based on scientific principles mod-
antimicrobial agents). The decision of gested algorithms for management of
ified by the “art and experience” of the
whether to treat a high-risk infant healthy-appearing, high-risk infants are
practitioner. The following are well-
depends on the risk factors present, shown in Figs 1, 2, and 3. Screening
established concepts related to neo-
the frequency of observations, and blood cultures have not been shown to
natal sepsis:
gestational age. The threshold for be of value.21
1. Neonatal sepsis is a major cause of
morbidity and mortality.
2. Diagnostic tests for early-onset
sepsis (other than blood or CSF cul-
tures) are useful for identifying in-
fants with a low probability of sepsis
but not at identifying infants likely to
be infected.
3. One milliliter of blood drawn before
initiating antimicrobial therapy is
needed to adequately detect bacter-
emia if a pediatric blood culture bot-
tle is used.
4. Cultures of superficial body sites,
FIGURE 2
Evaluation of asymptomatic infants ≥37 weeks’ gestation with risk factors for sepsis. The diagnosis
a gastric aspirates, and urine are of
of chorioamnionitis is problematic and has important implications for the management of the no value in the diagnosis of early-
newborn infant. Therefore, pediatric providers are encouraged to speak with their obstetrical onset sepsis.
colleagues whenever the diagnosis is made. bLumbar puncture is indicated in any infant with
a positive blood culture or in whom sepsis is highly suspected on the basis of clinical signs, re- 5. Lumbar puncture is not needed in
sponse to treatment, and laboratory results. WBC, white blood cell; Diff, differential white blood cell all infants with suspected sepsis (es-
count.
pecially those who appear healthy)
but should be performed for infants
with signs of sepsis who can safely
undergo the procedure, for infants
with a positive blood culture, for in-
fants likely to be bacteremic (on the
basis of laboratory data), and infants
who do not respond to antimicrobial
therapy in the expected manner.
6. The optimal treatment of infants with
suspected early-onset sepsis is
broad-spectrum antimicrobial agents
(ampicillin and an aminoglycoside).
Once the pathogen is identified,
antimicrobial therapy should be
FIGURE 3
Evaluation of asymptomatic infants ≥37 weeks’ gestation with risk factors for sepsis (no narrowed (unless synergism is
chorioamnionitis). aInadequate treatment: Defined as the use of an antibiotic other than penicillin, needed).
ampicillin, or cefazolin or if the duration of antibiotics before delivery was <4 h. bDischarge at 24 h
is acceptable if other discharge criteria have been met, access to medical care is readily accessible, 7. Antimicrobial therapy should be
and a person who is able to comply fully with instructions for home observation will be present. If discontinued at 48 hours in clinical
any of these conditions is not met, the infant should be observed in the hospital for at least 48 h and
until discharge criteria are achieved. IAP, intrapartum antimicrobial prophylaxis; WBC, white blood situations in which the probability
cell; Diff, differential white blood cell count. of sepsis is low.
1012 FROM THE AMERICAN ACADEMY OF PEDIATRICS
Downloaded from pediatrics.aappublications.org at Hacettepe University on May 9, 2012
9. FROM THE AMERICAN ACADEMY OF PEDIATRICS
LEAD AUTHOR Rosemarie C. Tan, MD, PhD Ann L. Jefferies, MD – Canadian Paediatric Society
Richard A. Polin, MD Kasper S. Wang, MD Rosalie O. Mainous, PhD, RNC, NNP – National
Kristi L. Watterberg, MD Association of Neonatal Nurses
COMMITTEE ON FETUS AND Tonse N. K. Raju, MD, DCH – National Institutes
NEWBORN, 2011–2012 FORMER COMMITTEE MEMBER of Health
Lu-Ann Papile, MD, Chairperson Vinod K. Bhutani, MD
Jill E. Baley, MD FORMER LIAISON
William Benitz, MD LIAISONS William Barth, Jr, MD – American College of
Waldemar A. Carlo, MD CAPT Wanda Denise Barfield, MD, MPH – Centers Obstetricians and Gynecologists
James Cummings, MD for Disease Control and Prevention
Praveen Kumar, MD George Macones, MD – American College of STAFF
Richard A. Polin, MD Obstetricians and Gynecologists Jim Couto, MA
REFERENCES
1. Escobar GJ. The neonatal “sepsis work-up”: 11. Goldenberg RL, Andrews WW, Goepfert AR, grams at birth: A population-based study.
personal reflections on the development of et al The Alabama Preterm Birth Study: Pediatrics. 2000;106(2 pt 1):256–263
an evidence-based approach toward newborn umbilical cord blood Ureaplasma ure- 20. Buckler B, Bell J, Sams R, et al. Unnecessary
infections in a managed care organization. alyticum and Mycoplasma hominis cul- workup of asymptomatic neonates in the
Pediatrics. 1999;103(1, suppl E):360–373 tures in very preterm newborn infants. era of group B streptococcus prophylaxis
2. Polin RA, St Geme JW III. Neonatal sepsis. Am J Obstet Gynecol. 2008;198(1):43.e1–43. [published online ahead of print August 22,
Adv Pediatr Infect Dis. 1992;7:25–61 e5 2010]. Infect Dis Obstet Gynecol. doi:
3. Riley LE, Celi AC, Onderdonk AB, et al. Asso- 12. Benitz WE, Gould JB, Druzin ML. Risk factors 21. Ottolini MC, Lundgren K, Mirkinson LJ,
ciation of epidural-related fever and non- for early-onset group B streptococcal sep- Cason S, Ottolini MG. Utility of complete
infectious inflammation in term labor. Obstet sis: estimation of odds ratios by critical blood count and blood culture screening to
Gynecol. 2011;117(3):588–595 literature review. Pediatrics. 1999;103(6). diagnose neonatal sepsis in the asymp-
4. Stoll BJ, Hansen NI, Bell EF, et al; Eunice Available at: www.pediatrics.org/cgi/content/ tomatic at risk newborn. Pediatr Infect Dis
Kennedy Shriver National Institute of Child full/103/6/e77 J. 2003;22(5):430–434
Health and Human Development Neonatal 13. Newton ER, Clark M. Group B streptococcus 22. Gerdes JS. Clinicopathologic approach to
Research Network. Neonatal outcomes of and preterm rupture of membranes. Obstet the diagnosis of neonatal sepsis. Clin Per-
extremely preterm infants from the NICHD Gynecol. 1988;71(2):198–202 inatol. 1991;18(2):361–381
Neonatal Research Network. Pediatrics. 14. Schuchat A, Zywicki SS, Dinsmoor MJ, et al. 23. Schelonka RL, Chai MK, Yoder BA, Hensley D,
2010;126(3):443–456 Risk factors and opportunities for preven- Brockett RM, Ascher DP. Volume of blood
5. Tita AT, Andrews WW. Diagnosis and man- tion of early-onset neonatal sepsis: a multi- required to detect common neonatal patho-
agement of clinical chorioamnionitis. Clin center case-control study. Pediatrics. 2000; gens. J Pediatr. 1996;129(2):275–278
Perinatol. 2010;37(2):339–354 105(1 pt 1):21–26 24. Dietzman DE, Fischer GW, Schoenknecht FD.
6. Gibbs RS, Duff P. Progress in pathogenesis 15. Schrag SJ, Hadler JL, Arnold KE, Martell- Neonatal Escherichia coli septicemia—
and management of clinical intraamniotic Cleary P, Reingold A, Schuchat A. Risk fac- bacterial counts in blood. J Pediatr. 1974;
infection. Am J Obstet Gynecol. 1991;164(5 tors for invasive, early-onset Escherichia 85(1):128–130
pt 1):1317–1326 coli infections in the era of widespread 25. Kellogg JA, Ferrentino FL, Goodstein MH,
7. Romero R, Quintero R, Oyarzun E, et al. intrapartum antibiotic use. Pediatrics. 2006; Liss J, Shapiro SL, Bankert DA. Frequency of
Intraamniotic infection and the onset of 118(2):570–576 low level bacteremia in infants from birth
labor in preterm premature rupture of the 16. Martius JA, Roos T, Gora B, et al. Risk fac- to two months of age. Pediatr Infect Dis J.
membranes. Am J Obstet Gynecol. 1988;159 tors associated with early-onset sepsis in 1997;16(4):381–385
(3):661–666 premature infants. Eur J Obstet Gynecol 26. Neal PR, Kleiman MB, Reynolds JK, Allen SD,
8. DiGiulio DB, Romero R, Kusanovic JP, et al. Reprod Biol. 1999;85(2):151–158 Lemons JA, Yu PL. Volume of blood sub-
Prevalence and diversity of microbes in the 17. Stoll BJ, Hansen NI, Sánchez PJ, et al; mitted for culture from neonates. J Clin
amniotic fluid, the fetal inflammatory response, Eunice Kennedy Shriver National Institute of Microbiol. 1986;24(3):353–356
and pregnancy outcome in women with pre- Child Health and Human Development Neo- 27. Connell TG, Rele M, Cowley D, Buttery JP,
term pre-labor rupture of membranes. Am J natal Research Network. Early onset neonatal Curtis N. How reliable is a negative blood
Reprod Immunol. 2010;64(1):38–57 sepsis: the burden of group B Streptococcal culture result? Volume of blood submitted
9. DiGiulio DB, Romero R, Amogan HP, et al. and E. coli disease continues. Pediatrics. for culture in routine practice in a children’s
Microbial prevalence, diversity and abun- 2011;127(5):817–826 hospital. Pediatrics. 2007;119(5):891–896
dance in amniotic fluid during preterm 18. Wynn JL, Levy O. Role of innate host 28. Pourcyrous M, Korones SB, Bada HS, Patter-
labor: a molecular and culture-based in- defenses in susceptibility to early-onset son T, Baselski V. Indwelling umbilical arte-
vestigation. PLoS ONE. 2008;3(8):e3056 neonatal sepsis. Clin Perinatol. 2010;37 rial catheter: a preferred sampling site for
10. Viscardi RM. Ureaplasma species: role in (2):307–337 blood culture. Pediatrics. 1988;81(6):821–825
diseases of prematurity. Clin Perinatol. 19. Escobar GJ, Li DK, Armstrong MA, et al. 29. Anagnostakis D, Kamba A, Petrochilou V,
2010;37(2):393–409 Neonatal sepsis workups in infants >/=2000 Arseni A, Matsaniotis N. Risk of infection
PEDIATRICS Volume 129, Number 5, May 2012 1013
Downloaded from pediatrics.aappublications.org at Hacettepe University on May 9, 2012
10. associated with umbilical vein catheterization. 42. Stoll BJ, Hansen N, Fanaroff AA, et al. To tap at: www.pediatrics.org/cgi/content/full/123/
A prospective study in 75 newborn infants. or not to tap: high likelihood of meningitis 6/e967
J Pediatr. 1975;86(5):759–765 without sepsis among very low birth weight 55. Rajesh NT, Dutta S, Prasad R, Narang A. Effect
30. Polin JI, Knox I, Baumgart S, Campman E, infants. Pediatrics. 2004;113(5):1181–1186 of delay in analysis on neonatal cerebrospinal
Mennuti MT, Polin RA. Use of umbilical cord 43. Garges HP, Moody MA, Cotten CM, et al. fluid parameters. Arch Dis Child Fetal
blood culture for detection of neonatal bac- Neonatal meningitis: what is the correlation Neonatal Ed. 2010;95(1):F25–F29
teremia. Obstet Gynecol. 1981;57(2):233–237 among cerebrospinal fluid cultures, blood 56. Christensen RD, Rothstein G, Hill HR, Hall RT.
31. Visser VE, Hall RT. Urine culture in the cultures, and cerebrospinal fluid parame- Fatal early onset group B streptococcal
evaluation of suspected neonatal sepsis. ters? Pediatrics. 2006;117(4):1094–1100 sepsis with normal leukocyte counts.
J Pediatr. 1979;94(4):635–638 44. Shah SS, Ebberson J, Kestenbaum LA, Hodinka Pediatr Infect Dis. 1985;4(3):242–245
32. Vasan U, Lim DM, Greenstein RM, Raye JR. RL, Zorc JJ. Age-specific reference values for 57. Engle WD, Rosenfeld CR. Neutropenia in
Origin of gastric aspirate polymorphonuclear cerebrospinal fluid protein concentration in high-risk neonates. J Pediatr. 1984;105(6):
leukocytes in infants born after prolonged neonates and young infants. J Hosp Med. 982–986
rupture of membranes. J Pediatr. 1977;91(1): 2011;6(1):22–27 58. Manroe BL, Weinberg AG, Rosenfeld CR,
69–72 45. Byington CL, Kendrick J, Sheng X. Normative Browne R. The neonatal blood count in
33. Mims LC, Medawar MS, Perkins JR, Grubb cerebrospinal fluid profiles in febrile in- health and disease. I. Reference values for
WR. Predicting neonatal infections by evalu- fants. J Pediatr. 2011;158(1):130–134 neutrophilic cells. J Pediatr. 1979;95(1):89–98
ation of the gastric aspirate: a study in two 46. Ahmed A, Hickey SM, Ehrett S, et al. Cere- 59. Schelonka RL, Yoder BA, desJardins SE, Hall
hundred and seven patients. Am J Obstet brospinal fluid values in the term neonate. RB, Butler J. Peripheral leukocyte count
Gynecol. 1972;114(2):232–238 Pediatr Infect Dis J. 1996;15(4):298–303 and leukocyte indexes in healthy newborn
34. Choi Y, Saha SK, Ahmed AS, et al. Routine 47. Bonadio WA, Stanco L, Bruce R, Barry D, Smith term infants. J Pediatr. 1994;125(4):603–606
skin cultures in predicting sepsis pathogens D. Reference values of normal cerebrospinal 60. Mouzinho A, Rosenfeld CR, Sánchez PJ, Risser
among hospitalized preterm neonates in fluid composition in infants ages 0 to 8 weeks. R. Revised reference ranges for circulating
Bangladesh. Neonatology. 2008;94(2):123– Pediatr Infect Dis J. 1992;11(7):589–591 neutrophils in very-low-birth-weight neonates.
131 48. Nascimento-Carvalho CM, Moreno-Carvalho Pediatrics. 1994;94(1):76–82
35. Evans ME, Schaffner W, Federspiel CF, Cotton OA. Normal cerebrospinal fluid values in 61. Schmutz N, Henry E, Jopling J, Christensen
RB, McKee KT, Jr, Stratton CW. Sensitivity, full-term gestation and premature neo- RD. Expected ranges for blood neutrophil
specificity, and predictive value of body nates. Arq Neuropsiquiatr. 1998;56(3A):
concentrations of neonates: the Manroe
surface cultures in a neonatal intensive 375–380
and Mouzinho charts revisited. J Perinatol.
care unit. JAMA. 1988;259(2):248–252 49. Martín-Ancel A, García-Alix A, Salas S, Del 2008;28(4):275–281
36. Sherman MP, Goetzman BW, Ahlfors CE, Castillo F, Cabañas F, Quero J. Cerebrospi-
62. Christensen RD, Rothstein G. Pitfalls in the
Wennberg RP. Tracheal asiration and its nal fluid leucocyte counts in healthy neo-
interpretation of leukocyte counts of newborn
clinical correlates in the diagnosis of nates. Arch Dis Child Fetal Neonatal Ed.
infants. Am J Clin Pathol. 1979;72(4):608–611
congenital pneumonia. Pediatrics. 1980;65 2006;91(5):F357–F358
63. Lambert RM, Baer VL, Wiedmeier SE, Henry
(2):258–263 50. Kestenbaum LA, Ebberson J, Zorc JJ, Hodinka
E, Burnett J, Christensen RD. Isolated ele-
37. Srinivasan HB, Vidyasagar D. Endotracheal RL, Shah SS. Defining cerebrospinal fluid
vated blood neutrophil concentration at
aspirate cultures in predicting sepsis in white blood cell count reference values in
altitude does not require NICU admission if
ventilated neonates. Indian J Pediatr. 1998; neonates and young infants. Pediatrics. 2010;
appropriate reference ranges are used.
65(1):79–84 125(2):257–264
J Perinatol. 2009;29(12):822–825
38. Johnson CE, Whitwell JK, Pethe K, Saxena K, 51. Smith PB, Garges HP, Cotton CM, Walsh TJ,
64. Christensen RD, Rothstein G. Exhaustion of
Super DM. Term newborns who are at risk Clark RH, Benjamin DK Jr. Meningitis in
mature marrow neutrophils in neonates
for sepsis: are lumbar punctures necessary? preterm neonates: importance of cerebro-
with sepsis. J Pediatr. 1980;96(2):316–318
Pediatrics. 1997;99(4). Available at: www. spinal fluid parameters. Am J Perinatol.
pediatrics.org/cgi/content/full/99/4/e10 2008;25(7):421–426 65. Schelonka RL, Yoder BA, Hall RB, et al. Dif-
ferentiation of segmented and band neu-
39. Eldadah M, Frenkel LD, Hiatt IM, Hegyi T. 52. Smith PB, Cotten CM, Garges HP, et al. A
trophils during the early newborn period.
Evaluation of routine lumbar punctures in comparison of neonatal Gram-negative rod
J Pediatr. 1995;127(2):298–300
newborn infants with respiratory distress and Gram-positive cocci meningitis. J Per-
syndrome. Pediatr Infect Dis J. 1987;6(3): inatol. 2006;26(2):111–114 66. Lloyd BW, Oto A. Normal values for mature
243–246 53. Greenberg RG, Smith PB, Cotten CM, Moody and immature neutrophils in very preterm
40. Isaacs D, Barfield CP, Grimwood K, McPhee MA, Clark RH, Benjamin DK Jr. Traumatic babies. Arch Dis Child. 1982;57(3):233–235
AJ, Minutillo C, Tudehope DI; Australian Study lumbar punctures in neonates: test per- 67. Gerdes JS, Polin RA. Sepsis screen in neo-
Group for Neonatal Infections. Systemic formance of the cerebrospinal fluid white nates with evaluation of plasma fibronec-
bacterial and fungal infections in infants blood cell count. Pediatr Infect Dis J. 2008; tin. Pediatr Infect Dis J. 1987;6(5):443–446
in Australian neonatal units. Med J Aust. 27(12):1047–1051 68. Newman TB, Puopolo KM, Wi S, Draper D,
1995;162(4):198–201 54. Kanegaye JT, Nigrovic LE, Malley R, et al; Escobar GJ. Interpreting complete blood
41. May M, Daley AJ, Donath S, Isaacs D; Aus- American Academy of Pediatrics, Pediatric counts soon after birth in newborns at risk
tralasian Study Group for Neonatal Infec- Emergency Medicine Collaborative Research for sepsis. Pediatrics. 2010;126(5):903–909
tions. Early onset neonatal meningitis in Committee. Diagnostic value of immature 69. Rozycki HJ, Stahl GE, Baumgart S. Impaired
Australia and New Zealand, 1992–2002. Arch neutrophils (bands) in the cerebrospinal sensitivity of a single early leukocyte count
Dis Child Fetal Neonatal Ed. 2005;90(4):F324– fluid of children with cerebrospinal fluid in screening for neonatal sepsis. Pediatr
F327 pleocytosis. Pediatrics. 2009;123(6). Available Infect Dis J. 1987;6(5):440–442
1014 FROM THE AMERICAN ACADEMY OF PEDIATRICS
Downloaded from pediatrics.aappublications.org at Hacettepe University on May 9, 2012
11. FROM THE AMERICAN ACADEMY OF PEDIATRICS
70. Manzoni P, Mostert M, Galletto P, et al. Is 79. Chiesa C, Natale F, Pascone R, et al. C re- 87. Nizet V, Klein JO. Bacterial sepsis and men-
thrombocytopenia suggestive of organism- active protein and procalcitonin: reference ingitis. In: Remington JS, Klein JO, Wilson
specific response in neonatal sepsis? intervals for preterm and term newborns Christopher B, Nizet V, eds. Infectious Dis-
Pediatr Int. 2009;51(2):206–210 during the early neonatal period. Clin Chim eases of the Fetus and Newborn Infant. 7th
71. Guida JD, Kunig AM, Leef KH, McKenzie SE, Acta. 2011;412(11-12):1053–1059 ed. Philadelphia, PA: Saunders; 2010:222–275
Paul DA. Platelet count and sepsis in very low 80. Hayashi Y, Paterson DL. Strategies for re- 88. Pickering LK, Baker CJ, Kimberlin DW, Long
birth weight neonates: is there an organism- duction in duration of antibiotic use in SS, eds. Red Book: 2009 Report of the
specific response? Pediatrics. 2003;111(6 pt hospitalized patients. Clin Infect Dis. 2011; Committee on Infectious Diseases. 28th ed.
1):1411–1415 52(10):1232–1240 Elk Grove Village, IL: American Academy of
72. Vouloumanou EK, Plessa E, Karageorgopoulos 81. Rodwell RL, Leslie AL, Tudehope DI. Early Pediatrics; 2009
DE, Mantadakis E, Falagas ME. Serum diagnosis of neonatal sepsis using a he- 89. Cordero L, Ayers LW. Duration of empiric
procalcitonin as a diagnostic marker for matologic scoring system. J Pediatr. 1988; antibiotics for suspected early-onset sep-
neonatal sepsis: a systematic review and 112(5):761–767 sis in extremely low birth weight infants.
meta-analysis. Intensive Care Med. 2011;37
82. Baker CN, Thornsberry C, Facklam RR. Syn- Infect Control Hosp Epidemiol. 2003;24(9):
(5):747–762
ergism, killing kinetics, and antimicrobial 662–666
73. Benitz WE. Adjunct laboratory tests in the
susceptibility of group A and B streptococci. 90. Cotten CM, Taylor S, Stoll B, et al; NICHD
diagnosis of early-onset neonatal sepsis.
Antimicrob Agents Chemother. 1981;19(5): Neonatal Research Network. Prolonged du-
Clin Perinatol. 2010;37(2):421–438
716–725 ration of initial empirical antibiotic treat-
74. Gabay C, Kushner I. Acute-phase proteins and
83. MacGowan A, Wootton M, Bowker K, Holt ment is associated with increased rates of
other systemic responses to inflammation.
HA, Reeves D. Ampicillin-aminoglycoside in- necrotizing enterocolitis and death for
N Engl J Med. 1999;340(6):448–454
teraction studies using Listeria monocy- extremely low birth weight infants. Pedi-
75. Philip AG. Response of C-reactive protein
togenes. J Antimicrob Chemother. 1998;41 atrics. 2009;123(1):58–66
in neonatal Group B streptococcal in-
(3):417–418 91. Kuppala VS, Meinzen-Derr J, Morrow AL,
fection. Pediatr Infect Dis. 1985;4(2):145–
148 84. Bryan CS, John JF, Jr, Pai MS, Austin TL. Schibler KR. Prolonged initial empirical
Gentamicin vs cefotaxime for therapy of neo- antibiotic treatment is associated with ad-
76. Benitz WE, Han MY, Madan A, Ramachandra
P. Serial serum C-reactive protein levels in natal sepsis. Relationship to drug resistance. verse outcomes in premature infants.
the diagnosis of neonatal infection. Pedi- Am J Dis Child. 1985;139(11):1086–1089 J Pediatr. 2011;159(5):720–725
atrics. 1998;102(4). Available at: www.pedi- 85. Manzoni P, Farina D, Leonessa M, et al. Risk 92. Alexander VN, Northrup V, Bizzarro MJ.
atrics.org/cgi/content/full/102/4/e41 factors for progression to invasive fungal Antibiotic exposure in the newborn in-
77. Dandona P, Nix D, Wilson MF, et al. Procalci- infection in preterm neonates with fungal tensive care unit and the risk of necrotiz-
tonin increase after endotoxin injection in colonization. Pediatrics. 2006;118(6):2359– ing enterocolitis. J Pediatr. 2011;159(3):
normal subjects. J Clin Endocrinol Metab. 2364 392–397
1994;79(6):1605–1608 86. Bégué P, Floret D, Mallet E, et al. Pharma- 93. Centers for Disease Control and Prevention.
78. Lapillonne A, Basson E, Monneret G, Bienvenu cokinetics and clinical evaluation of cefo- Prevention of perinatal group B strepto-
J, Salle BL. Lack of specificity of procalcitonin taxime in children suffering with purulent coccal disease—revised guidelines from
for sepsis diagnosis in premature infants. meningitis. J Antimicrob Chemother. 1984; CDC, 2010. MMWR Recomm Rep. 2010;59
Lancet. 1998;351(9110):1211–1212 14(suppl B):161–165 (RR-10):1–36
PEDIATRICS Volume 129, Number 5, May 2012 1015
Downloaded from pediatrics.aappublications.org at Hacettepe University on May 9, 2012