Reducing the prescription of antibiotics in newborn infants suspected for having an early onset sepsis.

Due to suspected Early Onset Sepsis (EOS) newborn infants receive frequently
antibiotics which have several side effects. EOS, defined as a sepsis starting
within the first 2 days of life, is a severe and life-threatening disease. The
mortality when untreated is around 100%. The first clinical symptoms of a
sepsis are non-specific and may be subtle. The presence or absence of maternal
risk factors like maternal fever urinary tract infection, chorioamnionitis and
prolonged rupture of membranes may be related to EOS, but do not predict the
occurrence of an EOS.

Common laboratory parameters are not sensitive enough to predict or exclude the
diagnosis of EOS (Fowlie 1998). The golden standard to detect a neonatal sepsis
is a blood culture. A positive blood culture is a definite proof of infection,
a negative blood culture has a high negative predictive value (Kaiser 2002).
The result of a blood culture is definite only after 48-72 hours. Given the
difficulty of a clinical diagnosis of EOS, potential fatal consequences of
delayed treatment and the slow result of the blood culture, immediately
starting antibiotic therapy is the recommended approach in all infants with
suspected EOS. This approach will cause, in retrospect, *unnecessary* treatment
in many infants. All these infants are treated with broad spectrum antibiotics
for at least 2-3 days. These antibiotics are given intravenously and require
that the infant is admitted at a pediatric ward.

The wide-spread use of antibiotics has drawbacks. It will lead to the emergence
of resistant bacteria in the NICU (De Man 2000). In the newborn itself it may
lead to an abnormal colonization (of the gastro-intestinal tract), development
of resistant bacteria, and potentially to a higher incidence of yeast
infections during the newborn period (Fanaro 2003, Saiman 2000). Although there
is large variation in the incidence of multiresistant bacteria between several
countries, the incidence has increased the previous decades and the use of
antibiotics is considered as a risk factor (Quinn 1994, Burwen 1994, Asensio
2000). Restriction of the use of antibiotics can lead to a reduction of
antibiotic resistant micro-organisms (White 1997, Seppala 1997).

Since the 90s promising studies with cytokines, especially IL-6 (Buck 1994, De
Bont 1994, Kuster 1998, Kallman 1999, Mehr 2001, Gonzales 2003,
Verboon-Maciolek 2006, Ng 2006), IL-8 (Franz 1999, Franz 2001, Nupponen 2001,
Mehr 2001, Krueger 2001, Santana 2001, Franz 2004, Verboon-Maciolek 2006),
IL-10 (Ng 2003, Ng 2006) and TNF-α (De Bont 1994, Ng 1997, Silveira 1999,
Dollner 2001, Ng 2003), were published. The sensitivity of a single cytokine
measurement direct post partum for the diagnosis of an EOS was higher than the
commonly used diagnostic tests such as C-reactive protein (CRP) and white blood
cell (WBC) count. However, most of the studies were limited to term infants and
had a relatively small sample size. Franz et al (2004) described in a multi
center randomized intervention study that with an extra IL-8 measurement
directly after birth the use of antibiotics for EOS was reduced from 50% to 36%
in infants suspected for an EOS.

Newborn infants have often one or more risk factors for an infection whereas in
only a few infants a sepsis occurs. E.g., only one of the 100 infants who are
colonized by a Group B Streptococcus (GBS) develops a GBS-sepsis. It is
possible that a genetic susceptibility for EOS exists, i.e. a DNA-polymorphism,
rendering some patients at risk for EOS.

Our preliminary data shows that the prescription rate of antibiotics for
suspected EOS versus blood culture proven EOS was high. We aim to develop a new
diagnostic model to reduce unnecessary antibiotic treatment which can be tested
in the near future.

Reducing the prescription of parenteral antibiotic therapy has several benefits
for the individual infant and the neonatal intensive care unit as a whole.
Antibiotic treatment induces altered colonization patterns and overgrowth of
fungal and gram negative micro-organisms might occur. These changes in
colonization render newborns treated with antibiotics more susceptible for gram
negative and fungal infections. Broad spectrum antibiotics might select
resistant micro-organisms in the NICU. Decreasing unnecessary antibiotic
treatment in preterm and term infants could reduce the incidence of late onset
bacterial and fungal sepsis and reduce the rates of outbreaks with multi
resistant bacteria. When an infant has an infection with a (multi) resistant
bacteria it has - apart from clinical consequences - logistic consequences:
Intensive care wards are closed in order to reduce the risk of transmission of
these multi resistant bacteria.

Due to prevention of unnecessary use of parenteral antibiotics in term infants,
these infants will not be admitted at a pediatric ward and not be separated
from their parents, which is an emotional and economic profit.

Non-therapeutic observational study.

1.) During a three year period infants born in participating hospitals and
admitted to a pediatric or maternity ward in whom an EOS cannot be excluded
will be included in this study. Direct post partum blood will be collected from
cord blood (NS) and in blood taken for clinical purpose (T1) for cytokine
measurements. Four hours post partum (T2), 24 hours post partum (T3) and 48
hours post partum (T4) the cytokine measurements will be repeated when blood is
taken for clinical diagnostics. CRP will also be measured from the blood taken
direct post partum, 24 hours post partum and 48 hours post partum. The
cytokines and CRP measurements will not have influence on the antibiotic
therapy. Retrospectively the course of the cytokines will be analyzed with the
results of the blood culture and the clinical course. As outcome there will be
1.) proven EOS (positive blood culture), 2.) clinical EOS (sick, but negative
blood culture) 3.) No EOS but suspected EOS direct post partum, 4.) No EOS.
Potential confounders and clinical parameters of the mother and infant will be
registered to determine their influence on the course of cytokines.
In addition, umbilical cord blood will be analyzed for DNA polymorphism,
putatively involved in EOS.

If serum remains this serum will be kept anonymously for usage in a later stage
for improving reference values of diagnostic tests in newborn infants.

2.) Twenty newborn infants who will be admitted direct post partum to a
pediatric ward for glucose measurements and who are not suspected for an EOS
will be included. Cytokine levels will be measured in cord blood and in blood
taken for glucose controls direct post partum, 4 hours post partum, 24 hours
post partum and 48 hours post partum.

This research needs 0.2 ml blood on four different moments from the infant. On
these moments blood is already taken on clinical grounds. The blood sample will
be obtained from a central line, or via vena/ capillary puncture: in total 0.8
ml extra blood will be taken, which has no consequences for the newborn infant.