Early postnatal prophylactic granulocyte-macrophage colony
stimulating factor (GM-CSF) corrects neutropenia but does not lower
risk for sepsis or improve survival and short-term outcomes in
extremely preterm neonates, according to the results of a single-
blind, multicenter, randomized controlled trial reported in the
January 17 issue of The Lancet.

"Systemic sepsis is a major cause of death in preterm neonates,"
write Robert Carr, from the Department of Haematology, Guy's and St
Thomas' Hospital, King's College, London, United Kingdom, and
colleagues. "There are compelling theoretical reasons why treatment
with haemopoietic [CSFs] might reduce sepsis and improve outcomes,
and as a consequence these agents have entered into use in neonatal
medicine without adequate evidence. We assessed whether [GM-CSF]
administered as prophylaxis to preterm neonates at high risk of
neutropenia would reduce sepsis, mortality, and morbidity."

The trial of GM-CSF administered as prophylaxis for reduction of
sepsis in extremely preterm neonates who were small for gestational
age (the PROGRAMS trial) took place in 26 centers between June 2000
and June 2006. Within 72 hours of birth, 280 neonates of 31 weeks'
gestation or less and below the 10th percentile for birth weight were
randomly assigned either to treatment with GM-CSF, 10 µg/kg per day
subcutaneously for 5 days, or to standard care.

The treating clinicians completed a detailed clinical record form
daily from recruitment to day 28. The main study endpoint was sepsis-
free survival to 14 days from trial entry, and analysis was by
intention to treat.

During the first 11 days, infants treated with GM-CSF had a
significantly more rapid increase in neutrophil counts after trial
entry than did control infants (difference between neutrophil count
slopes, 0.34 ¡Ñ 109/L/day; 95% confidence interval [CI], 0.12 ¡X
0.56). However, sepsis-free survival was not significantly different
for all infants (93 of 139 treated infants, 105 of 141 control
infants; difference, −8%; 95% CI, −18 to 3). A meta-analysis
including data from this study and from previous published
prophylactic trials showed no benefit in survival.

"Early postnatal prophylactic GM-CSF corrects neutropenia but does
not reduce sepsis or improve survival and short-term outcomes in
extremely preterm neonates," the study authors write.

Study limitations include time to enrollment delayed by the need for
informed parental consent and inability to determine the effect of GM-
CSF as adjuvant treatment for established sepsis.

"We believe that, before embarking on future single agent clinical
trials, we should consider if this is too simplistic an approach to
sepsis prevention in the preterm infant, whose immune system appears
compromised in many different ways," the study authors
conclude. "Knowledge of the functional characteristics of neonatal
innate immunity remains limited and needs continued research effort.
Successful future stratagems will need a wider view of their
antibacterial defences."

In an accompanying comment, Frank Shann, MD, from the Intensive Care
Unit, Royal Children's Hospital, in Melbourne, Australia, notes that
about 90% of babies receive at least 1 off-label or unlicensed drug
during a stay in intensive care, and that some of these untested
treatments are ineffective or even harmful.

"Although leucopheresis is far more difficult than giving injections
of G-CSF or GM-CSF, the procedure is much easier to do than in the
early 1980s; perhaps the time is right for a large randomised trial
in neonates," Dr. Shann writes. "G-CSF and GM-CSF should no longer be
used routinely in neonatal intensive care without further evidence
from randomised trials. Unfortunately, funding is difficult to get
for trials of treatments that will be used in very few patients, or
that are not under patent."

Action Medical Research supported the trial. The authors and Dr.
Shann have disclosed no relevant financial relationships.

Lancet. 2009;373:188¡V190, 226¡V233.