Article infectious diseases

Antimicrobial Therapy and Late Onset Sepsis

Alison Chu, MD,* Joseph Educational Gaps
R. Hageman, MD,†
1. In an infant with a positive blood culture with coagulase-negative Staphylococci, how
Michael Schreiber, MD,‡
does the clinician distinguish a true infection from a likely contaminant?
Kenneth Alexander,
2. What is the evidence behind the use of certain antimicrobial agents in late onset
MD, PhDx
sepsis?

Author Disclosure Abstract

Drs Chu, Hageman, Late onset sepsis infections contribute a significant proportion of the morbidity and
mortality of hospitalized infants, especially in very low birth weight infants. Although
Schreiber, and
it is fairly clear which infants are at higher risk of developing sepsis, it is less clear
Alexander have
whether a standard for diagnostic evaluation exists and is being used consistently across
disclosed no financial institutions. In the current setting of changing epidemiology and emergence of anti-
relationships relevant biotic-resistant organisms, it is important to evaluate the antimicrobial agents used
to this article. This for empirical therapy and to emphasize the importance of antimicrobial stewardship.
commentary does not In addition, it is imperative to evaluate possible methods for prevention of these
infections.
contain a discussion of
an unapproved/
investigative use of
Objectives After completing this article, readers should be able to:
a commercial product/
device. 1. Recognize risk factors for late onset sepsis in hospitalized neonates.
2. Describe some important causative organisms causing late onset sepsis.
3. Recognize that there are controversies in the diagnostic evaluation in late onset
sepsis.
4. Understand the principles underlying antibiotic choice in empirical antibiotic therapy.

Introduction
Late onset sepsis (LOS) is a common clinical challenge for neonatologists. Although ad-
vances in neonatal practice have led to improved infant survival, infections continue to ac-
count for a significant proportion of morbidity and mortality in newborns, especially very
low birth weight (VLBW) infants (£1500 g). (1) VLBW in-
fants who develop LOS are significantly more likely to die
Abbreviations and to have impaired neurodevelopmental outcomes than
those who are not infected. (2) The evolving epidemiology
CoNS: coagulase-negative Staphylococci and the continuing emergence of antibiotic resistance among
CRP: C-reactive protein organisms that cause LOS require ongoing evaluation of
EOS: early onset sepsis the antibiotics used to treat suspected LOS. This article
GA: gestational age provides an overview of the risk factors associated with, etio-
IgG: immunoglobulin G logic agents identified in, diagnostic evaluation of, and anti-
LOS: late onset sepsis biotic therapy for LOS. In this context, we focus the rest
LP: lumbar puncture of our discussion of LOS within the population of hospital-
MRSA: methicillin-resistant Staphylococcus aureus ized, primarily premature, infants.
VLBW: very low birth weight

*Department of Neonatology, Pritzker School of Medicine, University of Chicago, Chicago, IL.

†
Department of Pediatrics, NorthShore University HealthSystem, Evanston, IL, and Professor Emeritus of Pediatrics, Feinberg School
of Medicine, Northwestern University, Chicago, IL.
‡
Department of Pediatrics and Professor of Pediatrics, Pritzker School of Medicine, University of Chicago, Chicago, IL.
x
Division of Pediatric Infectious Disease and Professor of Pediatrics, Pritzker School of Medicine, University of Chicago, Chicago, IL.

e94 NeoReviews Vol.13 No.2 February 2012

 infectious diseases late onset sepsis

Background and Definitions past decade. In 2007, the reported incidence of LOS
The gold standard test for sepsis in neonates has been a in VLBW infants in the United States was near 20%,
positive blood culture (1) along with clinical signs and whereas EOS in VLBW infants was w2%. (6) A large
symptoms of sepsis. A single positive blood culture in study of infants born in Israel from 1995 through
an asymptomatic infant may be a false positive and, de- 1998 showed similar rates of LOS. (5) Additional strat-
pending on the organism, may have little clinical signifi- ification revealed that 56% of infants born at a GA of
cance. LOS has varying definitions, with some groups 24 to 25 weeks, compared with 9% of infants with a
including infections occurring 48 hours after birth, (3) GA >34 weeks, had at least 1 episode of LOS, and 53%
72 hours after birth, (2) or anytime 4 to 7 days after birth of neonates with birth weight <750 g had ‡1 episodes
(4) with or without clinical symptoms. (5) On the other of LOS, compared with close to 17% of those born weigh-
end, some clinicians define LOS up to 30 days of life, ing 1250 to 1500 g.
whereas others may include any infections occurring be-
fore discharge from the hospital. Whereas early onset sepsis Comorbidities and Invasive Care
(EOS) is assumed largely to be due to vertical transmis- Smaller infants are usually at greater risk for developing
sion, LOS has been attributed to nosocomial or horizontal complications of their prematurity. Many investigators
acquisition. have found increased risk of LOS in infants with patent
ductus arteriosus, necrotizing enterocolitis, and chronic
lung disease. (2,5) Infants with these clinical problems
Risk Factors for LOS often require more invasive care during their hospitaliza-
Neonates in general are at increased risk for infection tion. Several studies have found an increased risk for LOS
because of their immature immune defenses, including in VLBW infants who required prolonged periods of me-
fragile cutaneous barriers, and relative immune tolerance. chanical ventilation, central catheter use, and parenteral
(1) Among hospitalized infants, certain risk factors confer nutrition. (2,4) In contrast, a case-control study per-
greater risk for LOS and comprise a cycle of increased sus- formed in the United Kingdom in 2011 concluded that,
ceptibility to infection (Fig). when controlled for GA, the only independent risk fac-
tor for Gram-negative LOS was the duration of paren-
Birth Weight and Gestational Age teral nutrition administration. (7) Other investigators,
Many studies have verified that the rate of neonatal infec- after controlling for birth weight and other interventions,
tion varies inversely with gestational age (GA) and birth found an increased risk of Candida infection in infants
weight of the infant. (2) The increased risk of LOS seen born at younger gestational ages (<26 weeks) or who
in VLBW infants holds true across countries and over the had abdominal surgery. (8)

Genetics
Little is known about the contribution of genetics to li-
ability for neonatal LOS. A retrospective study spanning
10 years compared sepsis concordance rates between
monozygotic and dizygotic twins. By logistic regression
analysis, it was determined that almost half of variance
in risk of LOS was due to genetic factors alone, with
the remaining half attributable to residual environmental
factors. These authors also cite investigations that dem-
onstrate positive associations between polymorphisms
in the genes encoding proinflammatory cytokines, for
example, tumor necrosis factor a and b, interleukin-6,
and susceptibility to and severity of sepsis. (9) Studies
like this suggest a significant role of genetic susceptibil-
ity to infection and impart a strong impetus for further
research.

Causative Organisms
Figure. Factors that confer a greater risk for LOS in the When choosing empirical antibiotic agents in suspected
neonate. sepsis, clinicians must balance concern for the most

NeoReviews Vol.13 No.2 February 2012 e95

infectious diseases late onset sepsis

prevalent organisms against concern for the less commonly may overestimate the true infection rate. In a study of
identified Gram-negative bacteria and fungal organisms, 629 infants, those classified as having CoNS were no
which have higher associated mortality rates. (2) With this more likely to die than patients who were uninfected.
in mind, we discuss aspects of some of the important or- (2) In another study of 16,629 infants, infants with clin-
ganisms to consider in LOS (Table). ical sepsis in combination with culture-positive CoNS
infection had lower mortality rates than infants with clin-
Selected Strains of Gram-Positive Bacteria ical sepsis but negative blood cultures. (13) This lack of
The majority (w45%–75%) of pathogens responsible for CoNS-related mortality might be interpreted to suggest
LOS are Gram-positive bacteria. (10) The most common either that most CoNS “infections” are not true infec-
organism isolated in LOS, coagulase-negative Staphylo- tions or that true CoNS infections are not associated with
cocci (CoNS), is also the overall least virulent. (4) Groups mortality.
from England, Israel, and the United States have all re- Methicillin-resistant Staphylococcus aureus (MRSA) has
ported similar rates (47%–54%) of LOS infections sec- recently emerged as another increasingly prevalent or-
ondary to CoNS in the past decade. (2,3,5) However, ganism identified in LOS. A study using data from the
given the low virulence of CoNS, and its ubiquity in the National Nosocomial Infections Surveillance system from
environment, it is frequently difficult to distinguish true 1995 to 2004 showed that, of all reported S aureus in-
infection from specimen contamination. The diagnosis fections, 23% were MRSA. This study also showed that
of a true CoNS bacteremia relies on 2 cultures obtained the incidence of MRSA LOS increased from 0.7 per
from different sites within 24 hours growing the same 10,000 patient days in 1995 to 3.1 in 2004, a 308% in-
organism with the same sensitivities, although this is crease. (14)
rarely done in practice. (11) In response to this problem,
members of the Vermont Oxford Network used their Gram-Negative Bacteria
database to create criteria for CoNS sepsis. (12) How- LOS due to Gram-negative organisms is associated with
ever, even using criteria such as these, published studies higher mortality. (2) Increasing antibiotic resistance is also

Table. Epidemiology of Causative Organisms in LOS in Neonates*

Stoll 2002 Vergagno 2011 Makhoul 2002
(U.S.) (2) (England) (3) (Israel) (5)
Gram-positive organisms 70.2% — 55.4%
Staphylocoocus, coagulase negative 47.9% 54% 47.4%
S aureus 7.8% 18% 3.9%
Enterococcus spp. 3.3% 16% 2.9%
Group B Streptococcus 2.3% 8% 0.3%
Other 8.9% — 1.1%
Gram-negative organisms 17.6% — 31.2%
E coli 4.9% 13% 2.8%
Klebsiella 4.0% — 14.7
Pseudomonas 2.7% 5% 4.2%
Enterbacter 2.5% 21% 3.8%
Serratia 2.2% — —
Acinetobacter — 2% 2.3%
Other 1.4% — 3.6%
Fungal organisms 12.2% 9%† 11.1%
C albicans 5.8% — 2.6%
Candida parapsilosis 4.1% — 1.9%
Other 2.3% — 6.6%
Mixed organisms — — 1.8%

A dash (—) indicates a value not specifically reported in the study. Study population: Stoll (2): VLBW infants (401–1500 g); average age of first LOS
infection ¼ 22 days of age; LOS defined as infection after 3 days old. Vergagno (3): preterm and term infants; median age at first LOS infection ¼ 18 days
of age (range 3–386 days); LOS defined as infection after 48 hours old. Makhoul (5): VLBW infants (<1500 g); no data provided on age at first LOS
infection; LOS defined as infection after 72 hours.
*Values are reported as the percentage of all culture-proven LOS infections in each cohort.
†
Reported as Candida species.

e96 NeoReviews Vol.13 No.2 February 2012

 infectious diseases late onset sepsis

an increasing problem in gram negative bacteria causing evaluation, >12 hours after initial evaluation, or seri-
LOS. In the United States, Escherichia coli has been ally in a survey of treatment practices from 2002. (15)
reported to be the most common Gram-negative rod We discuss the utility of and evidence behind using
causing LOS (Table 1). However, a longitudinal study CRP in the evaluation of sepsis in the final paper of this
of infections showed that the proportion of LOS cases series.
due to E coli has declined steadily since 1958. (4) Other
Gram-negative organisms commonly reported in LOS Lumbar Punctures
from studies outside the United States include Klebsiella, Whereas lumbar puncture (LP) may be more routinely
Enterobacter, and Serratia. Although less common, LOS completed as part of the febrile neonate workup in pre-
caused by Pseudomonas aeruginosa carries the highest viously healthy full-term infants, there are no standard-
mortality risk among premature infants, with reported ized guidelines for when to perform LP in hospitalized
rates of 45% to 74%. (10) VLBW infants with suspected LOS. In a retrospective
study of infants treated with antibiotics after 72 hours
Yeasts of age, the safety of using a clinical algorithm in determin-
Yeasts account for 7% to 20% of LOS infections. (5) Al- ing whether an LP should be done was evaluated. The
though less common than bacterial infections, blood- authors suggested that, in sick infants with presumed
stream infections due to yeast carry significant mortality LOS, neurologic signs, a positive blood culture, the lack
risk, and should, therefore, be considered in ill infants of localizing signs of infection with the presence of risk
as a possible etiology for LOS. Candida species, most factors (eg, mechanical ventilation, presence of a central
often Candida albicans and, C parapsilosis, are the most catheter, VLBW) are all reasons to perform an LP. Using
commonly encountered fungal organisms affecting pre- these criteria in infants with suspected LOS, this study
mature infants diagnosed with LOS, (4) found that 71% of evaluations included an LP, and
late onset meningitis was diagnosed in <2% of all evalua-
Aspects of the Diagnostic Evaluation of LOS tions. No significant differences in short- or long-term
Blood Cultures morbidity were found between infants who did and did
Neonatologists use the positive blood culture to confirm not receive LP, including infants with diagnosed menin-
the diagnosis of sepsis. Although no consensus guidelines gitis. However, it is not stated whether the study was
exist for the number of blood cultures that should be ob- powered to detect differences in outcomes, especially
tained before initiation of empirical antibiotic therapy, given the low incidence of meningitis in this cohort.
83% of clinicians polled on their practices reported that (17) It should also be noted that in the scenario of a pos-
they only drew one blood culture when no central venous itive blood culture, the importance of performing an LP
catheter or a catheter without blood return was present. may be dependent on the organism identified, given that
(15) One study examining the utility of one versus two isolates such as group B Streptococcus or Candida portend
blood cultures (drawn from two peripheral sites, drawn such poor outcomes if spread to the CSF.
within 15–30 minutes of each other and inoculated into
aerobic bottles) obtained 269 pairs of blood cultures in Treatment
the evaluation of sepsis in neonates. These investigators Because a thorough review of the mechanisms and cov-
found that of the 9% of infants diagnosed with culture- erage of commonly used antibiotics is presented in the
proven sepsis, every single episode had two positive blood accompanying article in this issue, which concerns EOS,
cultures with the same organism and sensitivities. The au- we focus on a discussion of how empirical antibiotic ther-
thors concluded that a single blood culture drawn with apy differs in LOS.
‡1 mL of blood resulted in no loss of accuracy in the di-
agnosis of sepsis in neonates. (16) However, the use of Empirical Antibiotic Therapy
more than one blood culture can be used to differentiate The goal of antibiotic use before the identification of
between true CoNS infection from contamination. (11) the infectious microbe, or empirical antibiotic therapy,
is to eradicate harmful organisms as early in the clinical
Use of C-Reactive Protein course as possible. However, the life-saving potential of
The use of C-reactive protein (CRP) to differentiate true antibiotic use in infants at high risk for infection must
infection from contamination, especially with positive be balanced against the possible negative consequences
CONS blood cultures, is still debated. Approximately of widespread use in low-risk infants. Antibiotic therapy
30% of NICU clinicians reported drawing a CRP at initial can alter the neonatal microbiome, potentially making

NeoReviews Vol.13 No.2 February 2012 e97

infectious diseases late onset sepsis

the infant more susceptible to opportunistic infection, Antifungal Therapy

and lead to increased incidence of antibiotic-resistant or- The rationale for treatment of Candida infections in neo-
ganisms. In a retrospective cohort study, the administra- nates is largely based on data in adults, because there is
tion of prolonged initial empirical antibiotic therapy a lack of well-designed studies on antifungal therapy in
(‡5 days; negative cultures) in the first week of life in neonates. The most commonly used agents are ampho-
VLBW infants was associated with an increase in LOS. tericin B, which binds to ergosterol, leading to altered
(18) Therefore, the choice of antibiotic therapy for sus- cell permeability inducing pore formation and cell death,
pected sepsis should be tailored for the most likely organ- and fluconazole, which binds and inhibits production
isms with the highest mortality risk, with consideration of ergosterol. The safety and efficacy of newer antimicro-
to local resistance patterns. No consensus guidelines exist bials such as echinocandins are being studied in neonates.
as to specific antibiotic regimens. (19) (26) Caspofungin is an echinocandin in which the prin-
Vancomycin is commonly used in LOS, likely because cipal mechanism of action is the noncompetitive inhibi-
the majority of LOS infections are caused by CoNS and tion of b-(1,3)-D-glucan synthase, an essential enzyme
the concern for rising rates of b-lactamase antibiotic re- responsible for fungal cell wall synthesis.
sistance in LOS isolates including MRSA. One large mul-
ticenter study reported that 44% of VLBW infants were Duration of Empirical Antibiotic Therapy
treated with vancomycin after day 3, and perhaps more Although no actual guidelines exist on duration of empir-
importantly, 30% of babies without proven infection re- ical therapy, a 2002 survey of treatment practices revealed
ceived this drug. (2) The Centers for Disease Control and that the majority of clinicians would discontinue antibiot-
Prevention and others have recommended against empir- ics after 2 to 3 days of negative cultures. (15) This prac-
ical vancomycin therapy to prevent the emergence and tice seems reasonable on the basis of published studies
spread of vancomycin-resistant strains. (20) Given the examining the time to positive cultures in LOS. One
relatively low reported virulence of Gram-positive organ- study found that of positive cultures in LOS evaluations,
isms in LOS (21) and lack of data to suggest decreased 98% had a time to detection £48 hours. Of those that
rates of fulminant sepsis with the use of vancomycin, (22) were positive later than 48 hours in this study, 7 of 8 grew
this seems to represent a reasonable approach. (2,16) CoNS, and 4 were contaminants. (27) A more recent
A recent quality improvement initiative using a guideline study in Belgium found that the median time to positive
to reduce vancomycin use in 2 NICUs with low MRSA cultures in LOS was 21 hours and that the median time to
infection rates demonstrated significant reduction in positive cultures for Gram-negative and Gram-positive
vancomycin exposure without an increase in attributable organisms was 11 hours and close to 24 hours, respec-
morbidity or mortality. (23) However, the widespread tively. (28) However, in a critically ill infant, prolonged
use of vancomycin may be explained by the concern for courses of antibiotics are often used. It is important
impaired neurodevelopmental outcomes reported in in- in these cases of presumed “culture-negative sepsis” to
fants with CoNS LOS. (24) As a result, the decision avoid erroneous attribution and to reevaluate the infant
as to whether to include vancomycin as part of empirical for possible noninfectious etiologies for their clinical
therapy pending culture results rests with the clinician. worsening.
The temptation to use third-generation cephalospor-
ins seems unwarranted on the basis of recent literature. Antimicrobial Stewardship
Data published in 2011 from the British Health Protec- On the basis of this evidence, it is critical to emphasize the
tion Agency’s surveillance program showed that 95% rational use of antibiotics. It seems that oxacillin/nafcillin
of organisms in LOS were susceptible to gentamicin with plus gentamicin is a reasonable regimen for empirical an-
either flucloxacillin or amoxicillin and that only 79% were tibiotic therapy in LOS, although the specific agents cho-
susceptible to cefotaxime monotherapy. (25) Although sen should take into account local resistance patterns.
cefotaxime has been frequently considered for empirical The use of vancomycin should be initiated if, in the judg-
therapy with its broader spectrum of coverage, it actually ment of the clinician, the infant is critically ill and the pos-
provides less coverage for the relevant disease-causing or- tulated infecting organism may be a methicillin-resistant
ganisms and has greater potential to promote antibiotic strain of a Gram-positive organism. Third-generation
resistance and increase the risk for fungal infections. An cephalosporin use should be discouraged outside of sus-
important caveat to this rule would be in suspected men- pected meningitis. Antibiotic therapy should be narrowed
ingitis, in which cefotaxime has superior cerebrospinal as much as possible once an organism is identified. In ad-
fluid penetration. dition to careful selection of what agent to use, duration

e98 NeoReviews Vol.13 No.2 February 2012

 infectious diseases late onset sepsis

of empirical antibiotic therapy should be limited to 2 to low birth weight, invasive interventions, as well as comor-
3 days if cultures are negative. It is also important to avoid bid conditions of prematurity. The predominant etiologic
treating colonization (positive endotracheal cultures with- agents of LOS include coagulase-negative Staphylococ-
out evidence of pneumonia) and prophylactic antibiotic cus, MRSA, methicillin-sensitive Staphylococcus aureus,
use for invasive devices. (21,29) E coli, and Candida. Although diagnostic practices vary,
antibiotics remain the mainstay of treatment, but the reg-
imen of empirical therapy is debated and ever evolving.
Preventions to Decrease LOS Rate
Institutions should establish epidemiologic surveillance
Despite being well studied and the only established means
programs and partner with their infectious disease col-
for preventing transmission of infectious organisms among
leagues to devise appropriate empiric antibiotic regimens
hospitalized patients, hand hygiene is poorly adhered to,
specific to each NICU. Several studies have evaluated the
with recently reported rates of 70% in NICUs. (30) Other
possible use of antimicrobials as preventative strategies
potentially promising interventions are being studied, and
for LOS. Because there is significant cost resulting from
here we report recent literature on the use of antimicrobial
sometimes inappropriate antimicrobial use, it is impor-
agents in prevention of LOS and a multicenter trial therapy
tant to develop and implement antimicrobial stewardship
with intravenous polyvalent immunoglobulin G (IgG).
programs.
Antimicrobial Prophylaxis
As emphasized earlier, Gram-positive infections consti-
tute a large portion of LOS infections. On the basis of
studies in pediatric oncology patients, a recent evaluation American Board of Pediatrics Neonatal-Perinatal
of 85 VLBW infants has demonstrated the efficacy of van- Medicine Content Specifications
comycin catheter locks in reducing the incidence of infec- • Know the clinical manifestations,
tion from 42% to 17%. (31) Given this and other studies, laboratory features, and differential
diagnosis of neonatal sepsis.
there seems to be some reduction in the incidence of
• Understand the treatment and
CoNS LOS without increase in resistant organisms, but complications of sepsis.
there is no evidence of a decrease in morbidity or mortality • Know the infectious agents that cause
with vancomycin prophylaxis. (32) Antifungal prophylaxis neonatal sepsis.
has also been studied for prevention of Candida infections. • Know the maternal, perinatal, and neonatal risk factors for
neonatal sepsis.
Although topical nystatin and prophylactic fluconazole ap-
• Know the epidemiology, prevention, and pathogenesis of
pear to provide significant reduction in Candida LOS, this neonatal infection with Staphylococcus aureus and
strategy should be balanced against the low reported rates Staphylococcus epidermidis.
of invasive candidiasis in VLBW infants. (10,26) • Know the management, including understanding of antibiotic
resistance, and complications of neonatal infection with
Staphylococcus aureus and Staphylococcus epidermidis.
Intravenous Immunoglobulin
The previously reported benefits of intravenous infusion
of IgG in the prevention and treatment of neonatal sepsis
in premature infants resulted in a multicenter, prospec- References
tive, randomized trial of more than 3400 infants. The in- 1. Higgins RD, Baker CJ, Raju TN. Executive summary of the
vestigators randomized infants receiving antibiotics for workshop on infection in the high-risk infant. J Perinatol. 2010;30
(6):379–383
suspected or proven infection to either receiving two in-
2. Stoll BJ, Hansen N, Fanaroff AA, et al. Late-onset sepsis in very
fusions of polyvalent IgG or matching placebo infusions low birth weight neonates: the experience of the NICHD Neonatal
48 hours apart. This study demonstrated no differences Research Network. Pediatrics. 2002;110(2 pt 1):285–291
between groups in death or major disability at age 2 years; 3. Vergnano S, Menson E, Kennea N, et al. Neonatal infections in
no significant differences in rates of EOS or LOS were England: the NeonIN surveillance network. Arch Dis Child Fetal
Neonatal Ed. 2011;96(1):F9–F14
reported. (33)
4. Bizzarro MJ, Raskind C, Baltimore RS, Gallagher PG. Seventy-
five years of neonatal sepsis at Yale: 1928-2003. Pediatrics. 2005;
116(3):595–602
Conclusions 5. Makhoul IR, Sujov P, Smolkin T, Lusky A, Reichman B.
LOS in the neonate continues to contribute significantly Epidemiological, clinical, and microbiological characteristics of
to the morbidity and mortality of hospitalized infants late-onset sepsis among very low birth weight infants in Israel:
worldwide. Risk factors for LOS include prematurity, a national survey. Pediatrics. 2002;109(1):34–39

NeoReviews Vol.13 No.2 February 2012 e99

infectious diseases late onset sepsis

6. Fanaroff AA, Stoll BJ, Wright LL, et al. Trends in neonatal 20. Malik RK, Montecalvo MA, Reale MR, et al. Epidemiology
morbidity and mortality for very low birth weight infants.” Am J and control of vancomycin-resistant enterococci in a regional neo-
Obstet Gynecol. 2007;196:147.e1–147.e8 natal intensive care unit. Pediatr Infect Dis J. 1999;18(4):352–356
7. Samanta S, Farrer K, Breathnach A, Heath PT. Risk factors for 21. Isaacs D. Rationing antibiotic use in neonatal units. Arch Dis
late onset Gram-negative infections: a case-control study. Arch Dis Child Fetal Neonatal Ed. 2000;82(1):F1–F2
Child Fetal Neonatal Ed. 2011;96(1):F15–F18 22. Karlowicz MG, Buescher ES, Surka AE. Fulminant late-onset
8. Shetty SS, Harrison LH, Hajjeh RA, et al. Determining risk sepsis in a neonatal intensive care unit, 1988–1997, and the impact
factors for candidemia among newborn infants from population- of avoiding empiric vancomycin therapy. Pediatrics. 2000;106(6):
based surveillance: Baltimore, Maryland, 1998–2000. Pediatr Infect 1387–1390
Dis J. 2005;24(7):601–604 23. Chiu CH, Michelow IC, Cronin J, Ringer SA, Ferris TG,
9. Bizzarro MJ, Jiang Y, Hussain N, Gruen JR, Bhandari V, Zhang H. Puopolo KM. Effectiveness of a guideline to reduce vancomycin use
The impact of environmental and genetic factors on neonatal late- in the neonatal intensive care unit. Pediatr Infect Dis J. 2011;30(4):
onset sepsis. J Pediatr. 2011;158(2):234–238, e1 273–278
10. Downey LC, Smith PB, Benjamin DK Jr. Risk factors and 24. Stoll BJ, Hansen NI, Adams-Chapman I, et al; National
prevention of late-onset sepsis in premature infants. Early Hum Institute of Child Health and Human Development Neonatal
Dev. 2010;86(suppl 1):7–12 Research Network. Neurodevelopmental and growth impairment
11. Zaidi AKM, Harrell LJ, Rost JR, Reller LB. Assessment of among extremely low-birth-weight infants with neonatal infection.
similarity among coagulase-negative staphylococci from sequential JAMA. 2004;292(19):2357–2365
blood cultures of neonates and children by pulsed-field gel 25. Muller-Pebody B, Johnson AP, Heath PT, Gilbert RE,
electrophoresis. J Infect Dis. 1996;174(5):1010–1014 Henderson KL, Sharland M; iCAP Group (Improving Antibiotic
12. Vermont Oxford Database Manual of Operations. Release 2.0. Prescribing in Primary Care). Empirical treatment of neonatal
Burlington, VT: Vermont Oxford Network; 1993 sepsis: are the current guidelines adequate? Arch Dis Child Fetal
13. Jean-Baptiste N, Benjamin DK Jr, Cohen-Wolkowiez M, et al. Neonatal Ed. 2011;96(1):F4–F8
Coagulase-negative staphylococcal infections in the neonatal in- 26. Zaoutis T, Walsh TJ. Antifungal therapy for neonatal candidi-
tensive care unit. Infect Control Hosp Epidemiol. 2011;32(7): asis. Curr Opin Infect Dis. 2007;20(6):592–597
679–686 27. Kaiser JR, Cassat JE, Lewno MJ. Should antibiotics be dis-
14. Lessa FC, Edwards JR, Fridkin SK, Tenover FC, Horan TC, continued at 48 hours for negative late-onset sepsis evaluations in
Gorwitz RJ. Trends in incidence of late-onset methicillin-resistant the neonatal intensive care unit? J Perinatol. 2002;22(6):445–447
Staphylococcus aureus infection in neonatal intensive care units: data 28. Guerti K, Devos H, Ieven MM, Mahieu LM. Time to positivity
from the National Nosocomial Infections Surveillance System, of neonatal blood cultures: fast and furious? J Med Microbiol. 2011;
1995–2004. Pediatr Infect Dis J. 2009;28(7):577–581 60(Pt 4):446–453
15. Rubin LG, Sánchez PJ, Siegel J, Levine G, Saiman L, Jarvis WR; 29. Patel SJ, Oshodi A, Prasad P, et al. Antibiotic use in neonatal
Pediatric Prevention Network. Evaluation and treatment of neonates intensive care units and adherence with Centers for Disease Control
with suspected late-onset sepsis: a survey of neonatologists’ practices. and Prevention 12 Step Campaign to Prevent Antimicrobial Re-
Pediatrics. 2002;110(4):e42–e49 sistance. Pediatr Infect Dis J. 2009;28(12):1047–1051
16. Sarkar S, Bhagat I, DeCristofaro JD, Wiswell TE, Spitzer AR. 30. Pessoa-Silva CL, Hugonnet S, Pfister R, et al. Reduction of
A study of the role of multiple site blood cultures in the evaluation health care associated infection risk in neonates by successful hand
of neonatal sepsis. J Perinatol. 2006;26(1):18–22 hygiene promotion. Pediatrics. 2007;120(2):e382–e390
17. Flidel-Rimon O, Leibovitz E, Eventov Friedman S, Juster- 31. Garland JS, Alex CP, Henrickson KJ, McAuliffe TL, Maki DG.
Reicher A, Shinwell ES. Is lumbar puncture (LP) required in every A vancomycin-heparin lock solution for prevention of nosocomial
workup for suspected late-onset sepsis in neonates? Acta Paediatr. bloodstream infection in critically ill neonates with peripherally
2011;100(2):303–304 inserted central venous catheters: a prospective, randomized trial.
18. Kuppala VS, Meinzen-Derr J, Morrow AL, Schibler KR. Pro- Pediatrics. 2005;116(2):e198–e205
longed initial empirical antibiotic treatment is associated with ad- 32. Bell SG. Vancomycin prophylaxis for late-onset sepsis in very
verse outcomes in premature infants. J Pediatr. 2011;159(5): low and extremely low birth weight neonates. Neonatal Netw.
720–725 2008;27(5):351–354
19. Gordon A, Jeffery HE. Antibiotic regimens for suspected late- 33. Brocklehurst P, Farrell B, King A, et al; INIS Collaborative
onset sepsis in newborn infants. Cochrane Database Syst Rev.2005; Group. Treatment of neonatal sepsis with intravenous immune
(3):CD004501 globulin. N Engl J Med. 2011;365(13):1201–1211

e100 NeoReviews Vol.13 No.2 February 2012

 infectious diseases late onset sepsis

NeoReviews Quiz
New minimum performance level requirements
Per the 2010 revision of the American Medical Association (AMA) Physician’s Recognition Award (PRA) and credit system,
a minimum performance level must be established on enduring material and journal-based CME activities that are certified
for AMA PRA Category 1 CreditTM. In order to successfully complete 2012 NeoReviews articles for AMA PRA Category 1
CreditTM, learners must demonstrate a minimum performance level of 60% or higher on this assessment, which measures
achievement of the educational purpose and/or objectives of this activity. Starting with 2012 NeoReviews, AMA PRA
Category 1 CreditTM can be claimed only if 60% or more of the questions are answered correctly. If you score less than 60%
on the assessment, you will be given additional opportunities to answer questions until an overall 60% or greater score is
achieved.

1. Late onset sepsis, attributed to nosocomial or horizontal transmission, is common among very low birth weight
(VLBW) infants (<1500g). Of the following, according to the National Institutes of Child Health and Human
Development Neonatal Research Network, the incidence of late onset sepsis among VLBW infants in the United
States, is closest to
A. 10 percent
B. 20 percent
C. 30 percent
D. 40 percent
E. 50 percent

2. The risk of late onset sepsis from gram-negative microorganisms is increased among VLBW infants who require
invasive care during their hospitalization. Of the following, according to a case-control study performed in the
United Kingdom in 2011, the only independent risk factor for gram-negative late onset sepsis after controlling
for gestational age is the duration of
A. Chest tube placement
B. Ductus arteriosus ligation
C. Intravascular catheterization
D. Mechanical ventilation
E. Parenteral nutrition

3. A 28-day-old preterm infant, who has chronic lung disease and need for support with mechanical ventilation,
is suspected to have late onset sepsis. In choosing the antibiotics for treatment, you review the profile of
microorganisms likely to cause late onset sepsis in preterm infants in your nursery. Of the following, the most
common microorganism isolated in late onset sepsis is
A. Candida albicans
B. Coagulase negative Staphylococcus
C. Enterococcus spp
D. Escherichia coli
E. Group B Streptococcus

4. A 14-day-old preterm infant, who weighed 650g at birth at an estimated gestational age of 24 weeks, is
suspected to have late onset sepsis. The infant has clinical evidence of rapidly worsening shock. The blood
culture is positive for gram-negative microorganisms. Of the following, the gram-negative microorganism that
carries the highest mortality risk among preterm infants is
A. Acinetobacter baumanni
B. Enterobacter cloacae
C. Klebsiella pneumoniae
D. Pseudomonas aeruginosa
E. Serratia marcescens

5. A 6-week-old preterm infant, who weighed 1,520g at birth at an estimated gestational age of 32 weeks, is
brought to the emergency department after an acute life-threatening event. The infant is suspected to have

NeoReviews Vol.13 No.2 February 2012 e101

infectious diseases late onset sepsis

late onset sepsis and meningitis. You choose to treat with an antibiotic that has superior cerebrospinal
fluid penetration. Of the following, the antibiotic most likely to provide effective coverage for suspected
meningitis is
A. Cefotaxime
B. Fluconazole
C. Gentamicin
D. Nafcillin
E. Vancomycin

e102 NeoReviews Vol.13 No.2 February 2012