The trusted source for
healthcare information and
ABSTRACT & COMMENTARY
Synopsis: A meta-analysis of clinical trials suggests that, while the routine use of rG-CSF cannot be currently recommended for all neonates with early onset sepsis, its use appears promising in the management of septic, low birth weight, and/or neutropenic neonates.
Source: Bernstein HM, et al. Administration of recombinant granulocyte colony-stimulating factor to neonates with septicemia: A meta-analysis. J Pediatr. 2001;138:917-920.
A literature review of rg-csf for neonatal sepsis identified 5 studies (n = 20 to 44) deemed acceptable for inclusion in a meta-analysis to assess the effect of rG-CSF on mortality. Three studies were randomized and placebo-controlled, and 2 studies used historic controls. Two studies enrolled only patients weighing £ 1000 g, and 3 studies had no weight criterion. Four studies required neutropenia. The definition of early-onset sepsis was onset within 72 hours of birth in 3 studies and within 96 hours in 2 studies. None of the studies required a positive blood culture for diagnosis of sepsis. Among 73 rG-CSF recipients and 82 control subjects, mortality was lower among the rG-CSF recipients (odds ratio, 0.17; 95% CI, 0.03-0.70; P < .05). For subgroups of newborns < 2000 g or with neutropenia, the P value was .02. However, when the nonrandomized studies were excluded, the overall P value was .13.
Comment by Hal B. Jenson, MD, FAAP
Several small studies of rG-CSF for neonatal sepsis have been reported, which have varied by use of placebo controls or historic controls, birth weights and gestational ages of subjects, and severity of illness at study entry. These factors may confound the ability to perform post hoc meta-analysis. Perhaps the most significant shortcoming of these original studies, and subsequently this meta-analysis, is that a positive blood culture was not required for diagnosis of sepsis. The diagnosis of early-onset sepsis is often considered, especially in premature newborns, but infrequently confirmed by blood culture. It is likely that many cases of apparent sepsis in low-birth-weight premature newborns are the result of prematurity and may not result from bacterial infection. Also, this meta-analysis used an objective but crude outcome, of mortality, which might miss benefit of rG-CSF on reduced morbidity.
All 5 studies showed that rG-CSF was associated with at least a trend toward lower mortality, which persisted with meta-analysis of the 3 randomized, placebo-controlled studies. However, the low confidence interval (95% CI, 0.14-1.23) and the P value (P = .13) of meta-analysis of only these 3 studies suggest a low level of confidence in concluding benefit of rG-CSF in reducing mortality. Including the 2 studies with historic controls did show a significant reduction in mortality, but the inclusion of historic controls may bias the results and precludes routinely recommending rG-CSF for presumed early-onset neonatal sepsis at this time. Newborns with birth weight < 2000 g or with neutropenia appeared to benefit the most from rG-CSF. Additional studies or rG-CSF should incorporate these criteria into their design.