Professional Documents
Culture Documents
Sepsis Neonatal
Sepsis Neonatal
newborn infants. Although the incidence of sepsis in term and late preterm infants is low,
the potential for serious adverse outcomes is of such great consequence that caregivers
should have a low threshold for evaluation and treatment for possible sepsis in neonates.
●Late preterm infants (also called near-term infants) are those born from 34
through 36 completed weeks of gestation
Infants with early-onset sepsis typically present with symptoms during their birth
hospitalization. Term infants with late-onset sepsis generally present to the outpatient
setting or emergency department unless comorbid conditions have prolonged the birth
hospitalization. The approach to evaluation of neonates in the outpatient setting is
discussed separately.
●Vertical transmission, resulting in initial neonatal colonization that evolves into later
infection
●Horizontal transmission from contact with care providers or environmental sources
Disruption of the intact skin or mucosa, which can be due to invasive procedures (eg,
intravascular catheter), increases the risk of late-onset infection.
EPIDEMIOLOGY — The overall incidence of neonatal sepsis ranges from one to five
cases per 1000 live births. Estimated incidence rates vary based on the case definition
and the population studied. Globally, neonatal sepsis and other severe infections were
responsible for an estimated 430,000 neonatal deaths in 2013, accounting for
approximately 15 percent of all neonatal deaths.
Rates of neonatal sepsis increase with decreasing gestational age. The incidence of early-
onset sepsis in the United States has decreased primarily due to reduction in group B
streptococcal (GBS) infections, owing to the use of intrapartum antibiotic prophylaxis (IAP).
The estimated incidence of sepsis (both early- and late-onset) in term neonates is one to
two cases per 1000 live births. In a prospective national surveillance study (2006 to 2009),
the incidence of early-onset sepsis (defined as positive blood or cerebrospinal fluid [CSF]
cultures) was 0.98 cases per 1000 live births; the rate among infants with birth weight
>2500 grams was 0.57 per 1000.
The incidence is higher in late preterm than term infants. In an observational cohort study
(1996 to 2007), the reported incidences of early- and late-onset sepsis (defined as positive
blood culture) in late preterm neonates were 4.4 and 6.3 per 1000, respectively.
Early-onset GBS infection rates in the United States reported through the Centers for
Disease Control and Prevention's (CDC) Active Bacterial Core Surveillance Report have
declined from 0.6 per 1000 live births in 2000 to 0.21 per 1000 live births in 2015. Late-
onset GBS infection rates have remained relatively stable in the same interval (0.4 per
1000 live births in 2000 and 0.32 per 1000 live births in 2015).
Black race has been identified as an independent risk factor for early- and late-onset GBS
sepsis. Reasons for the disproportionately high disease burden among black populations
cannot be fully explained by prematurity, adequacy of prenatal care, or socioeconomic
status [11
ETIOLOGIC AGENTS — Group B Streptococcus (GBS) and Escherichia coli are the most
common causes of both early- and late-onset sepsis, accounting for approximately two-
thirds of early-onset infection [12,22,23].
Other bacterial agents associated with neonatal sepsis include (table 1):
The patterns of pathogens associated with neonatal sepsis have changed over time as
reflected by longitudinal databases from single tertiary centers. The incidence of early-
onset GBS has declined by 80 percent in the United States with the use of intrapartum
antibiotic prophylaxis (IAP). IAP appears to also reduce the risk of early-onset E.
coli infection:
●Candida
MATERNAL RISK FACTORS — The following maternal factors are associated with an
increased risk of sepsis, particularly group B Streptococcus (GBS) infection.
●Chorioamnionitis – Chorioamnionitis may reflect intrauterine onset of infection.
Consultation with obstetric providers to determine suspicion for chorioamnionitis is an
important aspect of neonatal management
●Maternal GBS colonization and other findings that increase the risk of GBS infection
in the neonate, including any of the following:
●Membrane rupture ≥18 hours – The risk of proven sepsis increases 10-fold to 1
percent when membranes are ruptured beyond 18 hours [31].
GBS screening and maternal intrapartum antibiotic prophylaxis (IAP) reduces the risk of
GBS infection but does not eliminate it. In a prospective national surveillance study (2006
to 2009), the GBS screening culture was negative in 81 percent of mothers of term infants
with early-onset GBS [18]. Approximately one-half of infants who developed early-onset
sepsis were born to mothers who received intrapartum antibiotics.
Because the signs and symptoms of sepsis can be subtle and nonspecific, it is important
to identify neonates with risk factors for sepsis and to have a high index of suspicion for
sepsis when an infant deviates from his or her usual pattern of activity or feeding [9].
●Fetal and delivery room distress – The following signs of fetal and neonatal
distress during labor and delivery may be early indicators of neonatal sepsis:
•Apgar score ≤6, which is associated with a 36-fold increased risk of sepsis.
•Jaundice – 35 percent
•Hepatomegaly – 33 percent
•Vomiting – 25 percent
•Diarrhea – 11 percent
EVALUATION AND INITIAL MANAGEMENT — Neonates with signs and symptoms of
sepsis (table 2) require prompt evaluation and initiation of antibiotic therapy [6,9]. Because
the signs and symptoms of sepsis are subtle and nonspecific, laboratory testing is
performed in any infant with identifiable risk factors and/or signs and symptoms concerning
for sepsis. This approach is consistent with guidelines published by the American
Academy of Pediatrics (AAP) and the Centers for Disease Control and Prevention (CDC)
[6,37].
Early-onset sepsis — Evaluation for neonatal early-onset sepsis (EOS) includes all of the
following:
●Review of the pregnancy, labor, and delivery, including risk factors for sepsis and
the use and duration of maternal intrapartum antibiotic prophylaxis (IAP)
●Laboratory testing
The extent of the diagnostic evaluation for sepsis is directed by the infant's symptoms and
maternal risk factors.
Risk estimators — Multivariate predictive models for risk of EOS have been developed
and validated in clinical use, including the EOS calculator [29,38-40]. These tools can be
used to estimate the risk of early-onset sepsis in individual patients based on risk factors
(eg, newborn clinical condition, highest intrapartum maternal temperature, maternal group
B Streptococcus [GBS] status, the administration of maternal IAP, gestational age,
duration of rupture of membranes). The estimated risk of sepsis can be used to help guide
decision-making regarding diagnostic evaluation and empiric antibiotic treatment. The
threshold used to trigger evaluation and empiric treatment varies according to the specific
model used. EOS risk calculators are not valid for preterm infants (<34 weeks gestation),
and they do not apply to late-onset sepsis.
Symptomatic neonates — Infants with signs and symptoms of sepsis (table 2) should
undergo a full diagnostic evaluationand should receive empiric antibiotic treatment
(algorithm 1).
●Blood culture.
●Lumbar puncture (LP) (if the infant is clinically stable enough to tolerate the
procedure).
●C-reactive protein (CRP) and/or procalcitonin (PCT) levels – These tests are not
routinely required but may be helpful in determining length of therapy if followed
serially. (See 'Other inflammatory markers' below.)
Late-onset sepsis — Infants presenting with signs and symptoms at ≥7 days of age
should undergo prompt evaluation and empiric antibiotic treatment. A full diagnostic
evaluation should be performed. In addition to the tests described above for early-onset
sepsis, the following should also be obtained:
●Urine culture
●Cultures from any other potential foci of infection (eg, tracheal aspirates if intubated,
purulent eye drainage, or pustules)
Infants with late-onset sepsis generally present to the outpatient or emergency department
setting unless comorbid conditions have prolonged the birth hospitalization.
●Ill-appearance
●Cerebrospinal fluid (CSF) pleocytosis (white blood cell [WBC] count of >20 to
30 cells/microL
The empiric antibiotic regimen should include agents active against GBS and other
organisms that cause neonatal sepsis (eg, E. coli and other gram-negative pathogens).
The combination of ampicillin and gentamicin or ampicillin and a third-generation
cephalosporin (eg, cefotaxime, if available) are appropriate regimens that provide empiric
coverage for these organisms until culture results are available. Ampicillin and gentamicin
is generally preferred; however, local antibiotic resistance patterns must be considered. In
the era of GBS IAP, approximately 30 percent of early-onset sepsis is due to ampicillin-
resistant, gram-negative organisms [14]. The addition of a third-generation cephalosporin
to the empiric treatment of early-onset sepsis is warranted among neonates with
suspected meningitis
LABORATORY TESTS — The goals of the diagnostic evaluation are to identify and treat
all infants with bacterial sepsis and minimize the treatment of patients who are not
infected. Laboratory assessment includes cultures of body fluids that confirm the presence
or absence of a bacterial pathogen and other studies that are used to evaluate the
likelihood of infection.
•Volume of blood – The optimal volume of blood is based on the weight of the
infant. A minimum blood volume of 1 mL is desirable for optimal detection of
bacteremia when a single blood culture bottle is used [6]. At the author's
institution, the suggested optimal volume is 2 mL for infants weighing ≤3 kg and
3 mL for those who weigh >3 to 5 kg. Dividing this volume into two aliquots to
inoculate an anaerobic as well as the aerobic culture bottle is discouraged, as it
is likely to decrease the sensitivity. Anaerobic cultures are generally not
necessary, and in vitro data suggest that small sample volumes do not reliably
detect low levels of bacteremia [41].
LP is indicated in neonates with clinical findings concerning for sepsis (table 2). In addition,
LP should be performed in infants with any of the following [6]:
Cerebrospinal fluid (CSF) should be sent for Gram stain, routine culture, cell count with
differential, and protein and glucose concentrations. The interpretation of CSF needs to
account for variations due to gestational age, chronologic age, and birth weight (table 3).
Blood culture may be negative in as many as 38 percent of infants with meningitis [6,43].
In a retrospective study, 8 of the 36 term infants with meningitis had no symptoms
referable to the central nervous system and had sterile blood cultures [44]. In addition,
three infants with both positive CSF and blood cultures were asymptomatic.
The clinical features and diagnosis of neonatal bacterial meningitis are discussed
separately.
Urine culture — Urine culture obtained by catheter or bladder tap should be included in
the sepsis evaluation for infants one week of age or older. A urine culture need not be
routinely performed in the evaluation of an infant ≤6 days of age, because a positive urine
culture in this setting is a reflection of high-grade bacteremia rather than an isolated
urinary tract infection [6,45].
Other cultures — In patients with late-onset infection, cultures should be obtained from
any other potential foci of infection (eg, purulent eye drainage or pustules).
Tracheal aspirates can be of value if obtained immediately after intubation [6]. However,
they may reflect lower respiratory tract colonization rather than indicating a causative
pathogen in an infant who has been intubated for several days.
In infants with early-onset infection, Gram stains of gastric aspirates are of limited value as
are bacterial cultures of body surface sites (eg, axilla, groin, and external ear canal) [6].
Complete blood count — A complete blood count (CBC) is used to evaluate the
likelihood of sepsis in a neonate with risk factors or signs of infection. Abnormal findings
on a CBC cannot be used to establish the diagnosis of sepsis.
Two large multicenter studies have evaluated the diagnostic value of CBCs in early-onset
neonatal sepsis [46,47]. These studies found that low white blood cell (WBC)
count (<5000/microL), absolute neutropenia (ANC <1000 neutrophils/microL), relative
neutropenia (ANC <5000 neutrophils/microL), and elevated I/T ratio were associated with
culture-proven sepsis. However, none of the tests were sufficiently sensitive to reliably
predict neonatal sepsis.
CBCs obtained 6 to 12 hours after delivery are more predictive of sepsis than those
obtained immediately after birth, because the WBC and ANC normally increase during the
first six hours of life [6,46,48].
The following neutrophil indices are used to determine the likelihood of infection:
●I/T ratio − An elevated I/T ratio (≥0.2) has the best sensitivity of the neutrophil
indices for predicting neonatal sepsis and can be helpful as an initial screen when
used in combination with risk factors and/or other tests [7,49,50]. A normal I/T ratio
can help rule out sepsis; however, an elevated value is not highly predictive of sepsis
and may be observed in 25 to 50 percent of uninfected infants [6].
In a study of 3154 neonates who underwent evaluation for early-onset sepsis with
blood culture and two serial WBC measurements, the I/T ratio was ≥0.2 in all 142
neonates with culture-positive or clinical sepsis, as well as in 1473 neonates without
infection [51].
The I/T ratio is limited by the wide range of normal values, which reduces its positive
predictive value, especially in asymptomatic patients [52]. Inter-reader differences in
band neutrophil identification with manual differential counts is another limitation [6].
In addition, exhaustion of the bone marrow reserves, which may occur during critical
illness, will result in low band counts and lead to falsely low ratios
●Absolute neutrophil count − Although both elevated and low neutrophil counts can
be associated with neonatal sepsis, neutropenia has greater specificity, as few
conditions other than sepsis and pre-eclampsia depress the neutrophil count of
neonates [6].
Neutrophil counts vary with gestational age (counts decrease with decreasing
gestational age), type of delivery (counts are lower in infants born by cesarean
delivery), site of sampling (counts are lower in arterial than in venous samples),
altitude (counts are higher at elevated altitudes), and timing after delivery (counts
increase during the first six hours of life).
The lower limit of a normal neutrophil count for infants >36 weeks of gestation
is 3500/microL at birth and 7500/microL six to eight hours after delivery (figure 1) [53].
For infants born at 28 through 36 weeks of gestation, the lower limits of normal
neutrophil counts at birth and at six to eight hours after birth
are 1000/microL and 1500/microL, respectively (figure 2).
Late-onset sepsis — CBCs are frequently used to support the diagnosis of late-onset
sepsis. In this setting, CBCs are less variable than in the first days of life. However, WBC
indices still perform poorly in identifying neonates with late-onset sepsis.
In a study of 37,826 neonates (mostly infants continuously hospitalized from birth) who
underwent evaluation for late-onset (defined as 4 to 120 days) sepsis with blood culture
and CBC, abnormal WBC (<1000 or >50,000/microL), high absolute neutrophil count
(ANC) (>17,670/microL), elevated I/T ratio (≥0.2), and low platelet
count (<50,000/microL) were associated with culture positivity [54]. However, sensitivity
was inadequate to reliably predict late-onset sepsis.
Screening protocols used to identify serious bacterial infections (SBIs) in febrile infants two
to three months of age are inadequate in neonates, as they fail to accurately identify
neonates with SBIs [44
A single measurement of CRP soon after birth is not a useful marker in the diagnosis
of neonatal sepsis. However, sequential assessment of CRP values may help support
a diagnosis of sepsis. If the CRP level remains persistently normal
(<1 mg/dL[10 mg/L]), neonatal bacterial sepsis is unlikely [6].
CRP levels can be helpful in guiding the duration of antibiotic therapy in suspected
neonatal bacterial infection. Infants with elevated CRP levels that decrease to
<1 mg/dL (10 mg/L) 24 to 48 hours after initiation of antibiotic therapy typically are not
infected and generally do not require further antibiotic treatment if cultures are
negative. However, routine use of serial CRP measurements can be associated with
longer length of hospital stay [57].
An elevated CRP level alone does not justify continuation of empiric antibiotics for
more than 48 hours in well-appearing infants with negative culture results [58].
Additional evaluation may be warranted to investigate alternative explanations for
persistently elevated CRP values.
PCT seems to have some utility in guiding the duration of antibiotic therapy in
neonates with suspected sepsis. A multicenter randomized controlled trial evaluated
the use of serial PCT testing to guide antibiotic therapy in 1710 term and late preterm
neonates with suspected early-onset sepsis [62]. Infants were assigned to one of two
treatment protocols in which duration of antibiotic therapy was guided by risk
classification based on clinical examination findings, blood culture results, WBC, and
CRP, with or without serial PCT measurements. Duration of antibiotic therapy was
reduced in the PCT group compared with the control group (55.1 versus 65.0 hours,
respectively).The rate of reinfection was low in both groups (0.7 percent for the PCT
group and 0.6 percent for the control group) and there was only one death (in the
control group). However, the study protocol's suggested duration of antibiotic therapy
for infants with negative cultures was fairly liberal (eg, up to 72 hours for infants
categorized as "low risk"; five to seven days for those in the "medium risk" category);
these are longer than what some experts would advocate. In addition, the suggested
duration of antibiotic therapy was overruled by the treating clinician in 45 percent of
infants in the PCT group and 41 percent of those in the control group. The findings
reinforce the concept that when serial PCT levels are obtained, they should be used
in conjunction with other clinical indicators of sepsis and should not be the sole basis
of decision-making. Additional studies are required to more confidently establish the
role of PCT as an adjunctive measure for decisions regarding length of antibiotic
treatment in neonatal sepsis.
Culture-proven sepsis — The isolation of pathogenic bacteria from a blood culture is the
gold standard to confirm the diagnosis of neonatal sepsis. A positive blood culture is
diagnostic of sepsis when a bacterial pathogen is isolated (table 1). Isolation of skin flora
(eg, diphtheroids and coagulase-negative staphylococci [CoNS]) in culture suggests
contamination rather than infection, although CoNS can be pathogenic in patients with
indwelling vascular catheters or other devices [9].
Probable sepsis — In some cases, a pathogen may not be isolated in culture, yet the
neonate has a clinical course that is concerning for sepsis (eg, ongoing temperature
instability; ongoing respiratory, cardiocirculatory, or neurologic symptoms not explained by
other conditions; or ongoing laboratory abnormalities suggestive of sepsis [ie,
cerebrospinal fluid (CSF) pleocytosis, elevated ratio of immature to total neutrophil counts,
or elevated C-reactive protein]).
A composite of observational assessment and serial laboratory testing is typically used to
make a diagnosis of probable sepsis [49]. The criteria used are usually broad, in an
attempt to ensure that all infected infants are identified but at the cost of testing and
treating a number of uninfected infants. There is no consensus definition for the diagnosis
of probable sepsis in neonates [2].
Alternative diagnoses (table 4) should also be entertained when an infant with suspected
sepsis has negative cultures. (See "Management and outcome of sepsis in term and late
preterm infants", section on 'Probable but unproven sepsis' and 'Differential
diagnosis' below.)
Infection unlikely — Infants with mild and/or transient symptoms (ie, fever alone or other
symptoms that quickly resolve) who remain well-appearing with normal laboratory values
and negative cultures at 48 hours are unlikely to have sepsis. Empiric antibiotic therapy
should be discontinued after 48 hours in these neonates [6,69].
In a retrospective study of 2785 newborns who underwent evaluation for sepsis based on
clinical symptoms (54 percent) or risk factors (46 percent), 22 infants (0.8 percent) were
found to have culture-proven sepsis and 40 (1.4 percent) had probable sepsis (ie, culture-
negative clinical sepsis) [7].
It is often difficult to differentiate neonatal sepsis from other conditions. However, given the
morbidity and mortality of neonatal sepsis, empiric antibiotic therapy should be provided
(after cultures are obtained) to infants with suspected sepsis pending definitive culture-
based diagnosis.
Here are the patient education articles that are relevant to this topic. We encourage you to
print or e-mail these topics to your patients. (You can also locate patient education articles
on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)
●Basics topics
●Although the incidence of sepsis in term and late preterm infants is low
(approximately one to six cases per 1000 births), the potential for serious adverse
outcomes, including death, is of such great consequence that caregivers should have
a low threshold for evaluation and treatment for possible sepsis.
●Group B Streptococcus (GBS) and Escherichia coli are the most common bacteria
causing neonatal sepsis.
●Maternal risk factors for neonatal sepsis in term and late preterm infants include
chorioamnionitis, intrapartum maternal temperature ≥38°C (100.4°F), delivery at <37
weeks gestation, maternal GBS colonization, and prolonged rupture of membranes
(≥18 hours).
●Clinical manifestations are nonspecific and include fetal distress; low Apgar score;
temperature instability (usually fever); respiratory and cardiocirculatory symptoms
(most commonly respiratory distress and tachycardia); neurologic symptoms
(irritability, lethargy, poor tone, and seizures); and gastrointestinal abnormalities (poor
feeding, vomiting, and abdominal distension
●The differential diagnosis of neonatal sepsis includes other systemic infections and
noninfectious conditions including respiratory diseases (eg, transient tachypnea of the
newborn and respiratory distress syndrome); cardiac diseases (eg, congenital heart
disease and supraventricular tachycardia); neurologic injury (eg, from anoxia or
hemorrhage); inborn errors of metabolism; and neonatal abstinence syndrome).