Definitions of bloodstream infection in the newborn

Khalid N. Haque, FRCP (Lond, Edin, Ire), FRCPCH

Objective: To develop definitions of bloodstream infections in
the newborn that would enable clinicians to identify infection
early, so patients can be enrolled in clinical trials. The definitions
should be useful for surveillance and epidemiologic purposes.
Method: Search of EMBASE, MEDLINE, and Cochrane Library
using age and English language limited key words sepsis, septi-
cemia, and shock. Extensive study of textbook of neonatology and
discussions with experts in the field.
Results: The search identified >2,000 references. The most
appropriate were selected and reviewed. Definitions of blood-
stream infection were developed after consultation with an inter-
national faculty.
Conclusion: Current definitions of neonatal infection (and as-
sociated categories) used by neonatal clinicians and researchers
have been either adapted/modified from definitions developed for
adults or generated by individuals to suit their local needs or the
needs of a particular study. It is clear that definitions generated
for adults are not applicable to children or to newborn infants. In
addition, developing and using unique definitions to suit individ-
ual or local needs make comparisons of outcome data and result
of studies very difficult. This article proposes a set of definitions
that are based as much as possible on current evidence. These
definitions may be applicable widely for daily management of an
infant with an infection and for research and epidemiologic stud-
ies. (Pediatr Crit Care Med 2005; 6[Suppl.]:S45–S49)
KEY WORDS: definitions; neonatal infections; neonatal sepsis

The beginning of wisdom is calling things by their right name.—Chinese Proverb

I n September 2004, the Interna-
tional Sepsis Forum convened a
consensus conference to establish
a series of definitions for a large
number of infections in children and in
the newborn. The goals of these defini-
tions were as follows:
1. Identify infection early so that optimal
therapy can be instituted.
2. Identify infections so that patients can
be enrolled in clinical trials
3. Be useful for surveillance and epide-
miologic studies
This submission deals with blood-
stream infections (BSIs) in the newborn
including term and premature infants. It
briefly outlines the burden of the disease
in both the developing and the developed
world, describes the search strategy used,
discusses the definitions currently in
common use, and reviews the difficulties
From the University of London and Epsom and St.
Helier University Hospitals, NHS Trust, Carshalton, UK.
This work was supported by the Mannion Family
Fund—Center for the Critically Ill Child, Division of
Critical Care Medicine at Children’s Hospital Boston,
the PALISI Network, and the ISF.
Copyright © 2005 by the Society of Critical Care
Medicine and the World Federation of Pediatric Inten-
sive and Critical Care Societies
DOI: 10.1097/01.PCC.0000161946.73305.0A
in their application to meet the previ-
ously mentioned objectives. Descriptions
of specific definitions of pediatric infec-
tions, systemic inflammatory response
syndrome (SIRS), sepsis, severe sepsis,
and septic shock produced by the Inter-
national Consensus Conference in 2002
are in press; they do not include defini-
tions of BSI in the newborn. An attempt
is made to suggest definitions for neo-
nates based on available evidence and to
provide an explanation and the rationale
behind the variables used to support the
suggested definitions.
BSI reflects a highly complex process
that is dependent on a series of interac-
tions between a pathogen and the host.
Our understanding of this process has
improved considerably in recent years
and is still evolving. It is now accepted,
for example, that the traditional concepts
of BSI do not adequately reflect events in
vivo and may even be an oversimplifica-
tion of the processes involved. This im-
provement in our understanding has
highlighted the fact that the definitions
of sepsis used in adults or even in chil-
dren cannot be extrapolated to neonates.
For example, in adults and even in older
children, great reliance is placed on a
positive blood culture for diagnosis of
BSI. Although this may be valid where
large volumes of blood or multiple sam-
ples can be taken, this is not possible in
the newborn, where the average volume
of blood placed in a blood culture bottle is
⬍0.5 mL (1, 2), making this test mark-
edly less sensitive.
Decision making in clinical medicine
requires quantitative reasoning based on
the probability of the disease being
present. This in itself is based on the
correct interpretation of clinical signs
and symptoms and results from labora-
tory investigations plus clinicians’ per-
ceived ratio between harm and benefit of
the treatment. However, at the end of the
day the ultimate decision made by an
astute clinician must be based on
Chachmah or wisdom.
Burden of Disease
Throughout the world, 1.6 million
neonates die every year from infection
(3). Although most of these deaths are
in developing countries, where neona-
tal mortality from sepsis may be as high
as 60% (4), the incidence of infection in
the developed world is also high at 2.2–
8.6 per 1000 live births (5). In fact, 48%
of all infections occur in children of ⬍1
yr of age, and just over half (27%) occur
in the newborn period. Despite under-
reporting, which is common, between

Pediatr Crit Care Med 2005 Vol. 6, No. 3 (Suppl.) S45

33% and 66% of babies admitted to
neonatal intensive care units are diag-
nosed to have infection some time dur-
ing their stay in the neonatal units (5–
7). The mortality rate from severe
infection in newborns remains high at
between 20% and 40% (6, 8), and there
has been no reduction in mortality at-
tributed to infection over the last 2–3
decades (9) despite the use of broad-
spectrum antibiotics and supportive
care. It is also being recognized in-
creasingly that infection in the new-
born infant, even away from the brain,
considerably increases the chances of
white matter damage and long-term
disability (10, 11). Thus, it is important
for clinicians to make the diagnosis of
infection early and accurately so that
they can institute appropriate treat-
ment. Early and accurate diagnosis is
also essential to maintain surveillance,
which is important to prevent infection
and to give clinicians the ability to en-
roll babies into clinical and epidemio-
logic studies.
Search Strategy and Comment
EMBASE, MEDLINE, and Cochrane
Library were searched using age and En-
glish language limited key words sepsis,
septicemia, and septic shock. This identi-
fied ⬎2,000 references. These diagnoses
can be accessed by ⬎400 codes using the
International Classification of Diseases
(version 10).
In reviewing a large selection of the
references it became clear that establish-
ing a working definition of BSI in the
newborn was an inherently imperfect
process due to lack of consensus, an im-
perfect understanding of the pathophysi-
ology of infection, and the development
of new and better (more sensitive and
specific) laboratory tests.
Currently, there is no evidence to sug-
gest that one definition is superior to
another. Despite published definitions of
sepsis in adult patients (12), in one recent
survey of adult critical care physicians,
71% cited no common definition of sepsis
used clinically (13). The main reason for
this lack of consensus on definitions is
that they are produced by individuals or
group of individuals with completely dif-
ferent mindsets (e.g., microbiologists, ep-
idemiologists, critical care physicians,
and generalists) or are produced with
specific research projects in mind.
S46
Current Definitions
Most clinicians use definitions either
modified or adapted from the adult liter-
ature, where there is still lack of consen-
sus at the bedside (13). This lack of con-
sensus on the definition of BSI may also
be due to lack of specific and sensitive
clinical and laboratory parameters. Clini-
cians, fearing adverse effects from infec-
tion, institute antibiotic therapy in a
large number of cases when they are un-
certain of the diagnosis of infection. In a
recent report of a large cohort of very low
birth weight babies from a large number
of neonatal units in the United States,
Stoll et al. (8) reported that although the
actual number of very low birth weight
newborns with culture proven infection
was only 1.9%, 50% of the babies in the
cohort were prescribed antibiotics by the
clinicians on the basis of clinical and risk
factors.
Other factors that add to clinicians’
uncertainty are the differing pathophysi-
ology, exposure, and susceptibility of the
newborn to infection. For example, the
immature immune system in the new-
born, in particular the preterm, responds
differently than that of an adult or even
an older child to the same offending
pathogens. Thus, the signs and symptoms
of infection in the newborn are very non-
specific and the same is true for activa-
tion of markers of sepsis. To add to this
complexity is the recent recognition of
the genetic variation in response to infec-
tion; for example, some babies have a
predominantly interleukin-8 response to
infection, whereas others exhibit a pre-
dominantly granulocyte colony-stimulat-
ing factor response (14).
Most current definitions are based on
the seminal paper by Bone et al. (15), who
defined infection into a) infection and
bacteremia; b) SIRS; c) sepsis; d) severe
sepsis; and e) septic shock. Hayden (16),
using the same definitions as Bone,
adapted them for children by giving age-
specific values; temperature ⬎37.9°C or
⬍36°C, heart rate of ⬎180 beats/min,
respiratory rate ⬎60 breaths/min, and
white blood cell count of either ⬎12,000/
mm3 or ⬍4000/mm3 or presence of
ⱖ10% immature (band forms) neutro-
phils. Since there was no consensus on
the numbers suggested by Hayden, they
were modified to ⬎2 SD above normal for
age. Immediately, the problem arose that
for many of the variables, standard cen-
tiles did not exist for newborns. Samson
et al. (17) suggested using values ⬎90th
Pediatr Crit Care Med 2005 Vol. 6, No. 3 (Suppl.)
centile for age in children. Barton et al.
(18), in a large study on the use of re-
combinant human protein C, used a com-
bination of criteria, using Hayden’s crite-
ria for SIRS and temperature while using
Samson’s recommendations of ⬎90th
centile for heart and respiratory rate.
Others (19, 20) have added to the debate
by adding variables like the Glasgow
Coma Scale, metabolic acidosis, and de-
layed capillary refill time to define septic
shock. This highlights the fact that infec-
tion, far from being a homogeneous con-
dition reflecting simply the presence of
pathogens in blood (bacteremia, viremia),
is actually a continuum of phases from
infection to sepsis, severe sepsis, septic
shock, and ultimately multiple-organ
failure and death. The difficulty for the
clinician is to determine in which phase
the patient is at any given moment as the
patient may move from one phase to an-
other in either direction imperceptibly
(Fig. 1).
Definitions
Sepsis. Sepsis has been variously de-
fined as “the presence of pathogenic
microorganism or their toxins in tis-
sues or blood” or as “a harmful or dam-
aging host systemic response to in-
fection when the host response to
infection becomes enhanced or deregu-
lated.” The International Sepsis Defini-
tion Conference (ISDC) in 2001 (12)
defined sepsis as “the clinical syndrome
defined as the presence of both infec-
tion and systemic inflammatory re-
sponse syndrome.” The ISDC definition
requires a clearer understanding of
SIRS, which can be triggered by events
other than sepsis (e.g., trauma). Al-
though the concept of SIRS is widely
understood and accepted for adults, its
value for children in particular neo-
nates is yet to be evaluated (23). This is
particularly so in the light of our recent
understanding that, due to various de-
ficiencies in the host defense mecha-
nisms, the fetus and the neonate may
respond very differently than an adult
given the same pathogenic insult. For
example, early-onset (⬍72 hrs) infec-
tion in neonates is due to fetal inflam-
matory response syndrome, in which
the fetus or the newborn responds to
either an ascending infection from the
birth canal (chorioamnionitis) or a he-
matogenous spread from maternal in-
fection. It is not clear why only some
fetuses should generate such a powerful
response to infection when it is not in
their interest as it increases the risk of
cerebral palsy in survivors by nearly
nine-fold (24).
Severe Sepsis. Severe sepsis is defined
as sepsis complicated by organ dysfunc-
tion and hypotension.
Septic Shock. In children and neo-
nates, septic shock is usually character-
ized by tachycardia (heart rate ⬎180/
min) with signs of decreased perfusion
(differently measured, i.e., increased cap-
illary refill time ⬎3 secs, and hypoten-
sion ⱖ2 SD below normal range for age)
requiring fluid and inotropic therapy.
Categories of Bloodstream Infection
in the Newborn. These categories are pre-
sented in Table 1. In neonatal care, nos-
ocomial infections constitute nearly 75%
of all bacterial infections and are mainly
due to Gram-positive organisms (6), al-
though in some reports from the United
States, Gram-negative organisms are in-
creasingly being cultured (25). Mortality
from nosocomial infection is low in the
newborn, but recent evidence suggests
that nosocomial infection may not
merely prolong hospital stay but may also
cause long-term neurologic morbidity
(11).
There is no consensus but consider-
able debate in the literature as to the Figure 1. Suggested continuum of infection in the newborn. WBC, white blood cell count; CRP,
predictive value of the clinical variables C-reactive protein; IL, interleukin; rRNA, recombinant RNA; PCR, polymerase chain reaction; FIRS,
used to diagnose infection. For example, fetal inflammatory response syndrome. From Refs. 21 and 22.
most textbooks and clinical studies sug-
gest that a respiratory rate ⬎60 breaths/
min is abnormal, but studies (22, 23, 26 –
Table 1. Categories of bloodstream infection (BSI)
28) have clearly shown that healthy
babies may have a respiratory rate ⬎60 Primary BSI: BSI in patients without an identifiable focus of infection (e.g., group B Streptococcus
breaths/min. However, a respiratory rate or Escherichia coli infection)
⬎60 breaths/min when accompanied by Secondary BSI: BSI caused by pathogens related to infection at another site. (e.g., pneumonia)
grunting or chest recession or desatura- Early-onset BSI: Infection that occurs within the first 72 hrs of life usually reflecting vertical
tions is always abnormal (22, 23, 26 –28), transmission (e.g., group B Streptococcus infection)
Late-onset BSI: Infection that occurs after 72 hrs of life usually reflecting horizontal transmission
and this is the justification for suggesting
(e.g., urinary tract infection).
this as a diagnostic criterion for sepsis. I Bacteremia/infection: “A pathological process caused by the invasion of normally sterile tissue or
recognize that the same debate could be fluid or body cavity by pathogenic or potentially pathogenic organisms”
had for each and every item chosen in Proven BSI: A positive blood culture or a positive PCR in the presence of clinical signs and
Table 2 (29 –31). symptoms of infection (see Table 2 for signs and symptoms)
Blood culture is considered by many Probable BSI: Presence of clinical signs and symptoms of infection (Table 2) and at least two
as the gold standard; however, the reli- abnormal laboratory results when blood culture is negative
Possible BSI: Presence of clinical signs and symptoms of infection (Table 2) plus raised CRP or
ability of blood culture for small blood
IL-6/IL-8 level in the absence of a positive blood culture
samples (frequent in newborn care) has No BSI: Absence of clinical signs and symptoms of infection and abnormal laboratory results.
not been well evaluated (1). Fischer et Nosocomial BSI: An infection that occurs ⱖ48 hrs after admission in a baby who did not have
al. (32) estimated that if 1 mL of blood evidence of an infection on admission, characterized by growth of a pathogen not related to
is sent for culture (in practice much infection at another site from one blood culture or a positive PCR in the presence of clinical
smaller volumes are sent from newborn features of infection (Table 2)
and premature babies), then the sensi-
PCR, polymerase chain reaction; CRP, C-reactive protein; IL, interleukin.
tivity of this test is only around 30 –
40%, and this rises to 70 – 80% when
3 mL of blood is sent for culture. The add to the sensitivity of the test (33). the diagnosis entirely on blood culture
alternative of taking multiple small- The challenge, therefore, is to think results to establish definitions that are
volume blood culture samples does not beyond the current paradigm of basing useful to clinicians for making a diag-

Pediatr Crit Care Med 2005 Vol. 6, No. 3 (Suppl.) S47

Table 2. Suggested diagnostic criteria for sepsis in neonates fection. I recommend greater use of these
tests because predictive value for small-
Clinical variables
volume blood culture (often still used as a
Temperature instability
Heart rate ⬎SD above normal for age (ⱖ180 beats/min, ⱕ100 beats/min) gold standard) and other individual tests
Respiratory rate (⬎60 breaths/min) plus grunting/recession or desaturationsa is poor.
Lethargy/altered mental status
Glucose intolerance (plasma glucose ⬎10 mmol/L)
Feed intolerance CONCLUSION
Hemodynamic variables
BP 2 SD below normal for age.
Establishing a working definition of
Systolic pressure ⬍50 mm Hg (newborn day 1) BSI in the newborn is fraught with diffi-
Systolic pressure ⬍65 mm Hg (infants ⱕ1 month) culties due to lack of consensus, con-
Tissue perfusion variables stantly changing understanding of the
Capillary refill ⬎3 secs pathophysiology of infection, and the lack
Plasma lactate ⬎3 mmol/L
Inflammatory variables of new and sensitive laboratory tests.
Leukocytosis (WBC count ⬎34,000 ⫻ 109/L) Based on a review of the evidence and the
Leukopenia (WBC count ⬍5,000 ⫻ 109/L) consensus of experts at an international
Immature neutrophils ⬎10% meeting, diagnostic criteria for BSI in the
Immature: total neutrophil ratio ⬎0.2
Thrombocytopenia ⬍100,000 ⫻ 109/L
newborn are suggested based on the clin-
CRP ⬎10 mg/dL or ⬎2 SD above normal value ical presentation and the results of spe-
Procalcitonin ⬎8.1 mg/dL or 2 SD above normal value cific laboratory tests. The success or fail-
IL-6 or IL-8 ⬎70 pg/mL ure of these suggestions will depend on
16 S PCR: positive their validation in clinical studies (21).
BP, blood pressure; WBC, white blood cell; CRP, C-reactive protein; IL, interleukin; PCR, poly-
merase chain reaction. REFERENCES
a
From Refs. 21 and 22. Modified from Refs. 29 –31.
1. Brown DR, Kutler D, Rai B: Perinatal/
neonatal clinical presentation: Bacterial con-
Table 3. Sensitivity, specificity, and positive and negative predictive value of some laboratory tests used centration and blood volume required for a
in the diagnosis of infection in the newborn (Ref. 4) positive blood culture. J Perinatol 1995; 15:
157–159
Test Sensitivity Specificity PPV NPV 2. Weinstein MP, Towns ML, Quartey SM, et al:
the clinical significance of positive blood cul-
Blood culture 11–38 68–100 90–100 72–100 ture in 1990’s: A prospective comprehensive
WBC ⬍5000 ⬎30,000 17–90 31–100 50–86 60–89 evaluation of the microbiology, epidemiology
I/T ratio ⬎0.02 81 45 23 92 and outcome of bacteremia and fungemia in
CRP ⬎10 mg/L 37 95 63 87 adults. Clin Infect Dis 1997; 24:584 – 602
IL-8 ⬎70 pg/mL 77 76 42 94 3. WHO Health Bulletin. Basel, Switzerland,
I/T ratio ⬎0.02 ⫹ CRP ⬎ 10 mg/L 89 41 24 94 World Health Organization, 2002
IL-8 ⬎ 70 pg/mL ⫹ CRP ⬎10 mg/L 91 74 43 98 4. Haque KN: In: Textbook of Paediatrics. Sixth
16 S PCRa 96.0 99.4 88.9 99.8
Edition, McIntosh N, Stenson B (Eds). Edin-
burgh, Livingston, 2003
PPV, positive predictive value; NPV, negative predictive value; WBC, white blood cell count; I/T,
5. Isaac D, Barfield C, Clothier T, et al: Late
immature/total neutrophil ratio; CRP, C-reactive protein; IL, interleukin; PCR, polymerase chain
onset infections of infants in neonatal units.
reaction.
a J Paediatr Child Health 1996; 32:158 –161
From Ref. 35. All values are percentages.
6. Haque KN, Khan A, Kerry S, et al: Pattern of
culture proven neonatal sepsis in a district
nosis of infection at the bedside by us- combinations of some of the laboratory general hospital in the UK. Inf Control Hosp
Epidemiol 2004; 25:759 – 64
ing predictive clinical variables and tests currently used. The variables cho-
7. Sanghvi KP, Tudehope DL: Neonatal bac-
highly sensitive and specific laboratory sen for inclusion in Table 3 reflect
terial sepsis in a NICU: A 5 year analysis.
tests. This is a step change from the those that are likely to give the highest J Paediatr Child Health 1996; 32:333–338
recommendations that sepsis can only yield. 8. Stoll BJ, Hansen N, Fanaroff AA, et al:
be diagnosed when infection is con- There is considerable evidence that in- Changes in pathogens causing early onset
firmed by a positive blood culture (34). terleukin-6, interleukin-8, and a positive sepsis in very low birth weight infants. N Eng
Currently, there is not a single lab- polymerase chain reaction are better J Med 2002; 347:240 –247
oratory test that has receiver operating markers of infection (35–37) than the to- 9. Bell R, Glinianaia SV, Rankin J, et al: Chang-
characteristic value ⬎0.8. Thus, it re- tal leukocyte count, C-reactive protein, or ing pattern of perinatal death, 1982–2000: A
mains important for clinicians to use a procalcitonin. I recognize that PCR may retrospective cohort study. Arch Dis Child
Fetal Neonatal Ed 2004; 89:F531–F536
combination of clinical variables and not be available at many institutions;
10. Dammann O, Kuban KC, Leviton A: Perinatal
laboratory tests to get the greatest pre- thus, I suggest using a combination of
infection, fetal inflammatory response, white
dictive value. Table 2 suggests diagnos- C-reactive protein (⬎10 mg/L) and either matter damage and cognitive limitations in
tic criteria for sepsis in the newborn, interleukin-6 or interleukin-8 (⬎70 pg/ children born preterm. Ment Retard Dev Dis-
and Table 3 provides the specificity, mL), which gives a negative predictive ability Res Rev 2002; 8:46 –50
sensitivity, and positive and negative value of 98%. With the use of this com- 11. Stoll B, Hansen NI, Adams-Chapman I, et al:
predictive value of individual tests and bination, clinicians can safely exclude in- Neurodevelopmental and growth impair-

S48 Pediatr Crit Care Med 2005 Vol. 6, No. 3 (Suppl.)

 ment among extremely low-birth-weight in-
fants with neonatal infection. JAMA 2004;
292:2357–2365
12. Levy MM, Mitchell P, Marshall JC, et al: 2001
SCCM/ESICM/ACCP/ATS/SIS International
Sepsis Definitions Conference. Crit Care Med
2003; 31:1250 –1256
13. Ramsay G, Gerlach H, Levey MM, et al: An
international sepsis survey: A study of doc-
tor’s knowledge and perception about sepsis.
Crit Care Med 2003: 31:1250 –1256
14. Vecchio AD, Laforgia N, Capasso M, et al: The
role of molecular genetics in the pathogene-
sis and diagnosis of neonatal sepsis. Clin
Perinatol 2004; 31:53– 67
15. Bone RC, Sprung CL, Sibbald WJ: Definitions
for sepsis and organ failure. Crit Care Med
1992; 20:724 –726
16. Hayden WR: Sepsis terminology in paediat-
rics. J Pediatr 1994; 124:657– 658
17. Samson LM, Allen UD, Creery WD, et al:
Elevated interleukin-1 receptor antagonist
levels in paediatric sepsis syndrome. J Pedi-
atr 1997; 131:587–591
18. Barton P, Kalil AC, Nadel S, et al: Safety,
pharmacokinetics and pharmacodynamics of
drotrecogin alfa (activated) in children with
severe sepsis. Pediatrics 2004; 113:7–17
19. Krafte-Jacobs B, Brilli R: Increased circulat-
ing thrombomodulin in children with septic
shock. Crit Care Med 1998; 26:933–938
20. Carcillo JA, Fields AI, Task Force Members:
Clinical practice parameters for hemody-
namic support of pediatric and neonatal pa-
Pediatr Crit Care Med 2005 Vol. 6, No. 3 (Suppl.)
tients in septic shock. Crit Care Med 2002;
30:1365–1378
21. Morley CJ, Thornton AJ, Cole TJ, et al: Symp-
toms and signs in infants younger than six
months of age correlated with severity of
their illness. Pediatrics 1991; 88:1119 –1124
22. Morley CJ, Thornton AJ, Fowler MA, et al:
respiratory rate and severity of illness in ba-
bies under six months old. Arch Dis Child
1990; 65:834 – 837
23. Martinot A, Leclerc F, Cremer S, et al: Sepsis
in neonates and children: Definitions, epide-
miology, and outcome. Pediatr Emerg Care
1997; 13:277–281
24. Wheater M, Rennie JM: Perinatal infection is
an important risk factor for cerebral palsy in
very low-birth-weight infants. Dev Med Child
Neurol 2000; 42:364 –367
25. Nambiar S, Singh N: Change in epidemiology
of health care associated infections in NICU.
Pediatr Infect Dis J 2002; 21:839 – 42
26. Hewson PH, Humphries SM, Roberton DM, et
al: Markers of serious illness in infants under
six months old presenting to a children’s hos-
pital. Arch Dis Child 1990; 65:750 –756
27. Bonadio WA, Hennes H, Smith D, et al: Re-
liability of observations in distinguishing in-
fection outcomes of febrile young infants.
Pediatr Infect Dis J 1993; 12:111–114
28. Richardson DK, Gray JE, McCormick MC, et
al: Score for neonatal acute physiology: A
physiologic severity index for neonatal inten-
sive care. Pediatrics 1993; 91:617– 623
29. Rudolph CD, Rudolph AM (Eds): Rudolph’s
Paediatrics. 21st Edition. New York,
McGraw-Hill, 2002
30. Graner JS, Jarvis WR, Emori TG, et al: CDC
definitions for nosocomial infections. Is J In-
fect Control 1988; 15:128 –140
31. Ennui A, Rey C, Concha A, et al: Comparison
of procalcitonin with CRP and serum amy-
loid for the early diagnosis of bacterial sepsis
in critically ill neonates and children. Inten-
sive Care Med 2001; 27:211–215
32. Fischer JE, Seifarth FG, Baenziger O, et al:
Hindsight judgment on ambiguous episodes
of suspected infection in critically ill chil-
dren. Eur J Pediatr 2003; 162:840 – 843
33. Wiswell TE, Hachey J: Multiple site blood
cultures in the initial evaluation for neonatal
sepsis during the first week of life. Pediatr
Infect Dis 1991; 10:365–369
34. Sibbald WJ, Doig G, Inman KJ: Sepsis, SIRS
infection. Intensive Care Med 1995; 21:
299 –301
35. Jordan JA, Durso MB: Comparison of 16Sr-
RNA gene PCR and BACTEC 9240 for detec-
tion of neonatal bacteremia. J Clin Microbiol
2000; 38:2574 –78
36. Buck C, Bundschu J, Gallati H, et al: Inter-
leukin-6: A sensitive parameter for the early
diagnosis of neonatal bacterial infection. Pe-
diatrics 1994; 93:54 –58
37. Martin H, Olander B, Norman M: Reactive
hyperemia and interleukin-6, interleukin-8
and tumor necrosis factor alpha in the diag-
nosis of early-onset neonatal sepsis. Pediat-
rics 2001; 108:E61
S49