Antibiotic regimens for suspected late onset sepsis in newborn infants

Abstract

Late onset neonatal sepsis (systemic infection after 48 hours of age) continues to be a significant cause of morbidity and mortality. Early treatment with antibiotics is essential as infants can deteriorate rapidly. It is not clear which antibiotic regimen is most suitable for initial treatment of suspected late onset sepsis. To compare the effectiveness and adverse effects of different antibiotic regimens for treatment of suspected late onset sepsis in newborn infants. The standard search strategy of the Cochrane Neonatal Review Group was used. This includes electronic searches of the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 4, 2004), MEDLINE (1966 - Dec 2004), EMBASE (1980 - Dec 2004) and CINAHL (1982 - Dec 2004), electronic abstracts of Pediatric Academic Society meetings (1996 - Dec 2004) and previous reviews including cross references (all articles referenced). Randomised and quasi randomised controlled trials comparing different initial antibiotic regimens in neonates with suspected late onset sepsis were evaluated. Both reviewer authors screened abstracts and papers against the inclusion criteria, appraised the quality of and extracted data from papers. For dichotomous outcomes, treatment effect was expressed as relative risk and risk difference with 95% confidence intervals. NNT was calculated for outcomes for which there was a statistically significant reduction in risk difference. Thirteen studies were identified as possibly eligible for inclusion. The majority of studies were excluded as they did not separate data for early and late onset infection. Two studies are still awaiting assessment. Only one small study, in 24 neonates, was included in this review. It compared beta-lactam therapy with a combination of beta lactam plus aminoglycoside. The study did not meet our prespecified criteria for good methodological quality. In babies with suspected infection there was no significant difference in mortality (RR 0.17, 95% CI 0.01 to 3.23) or treatment failure (RR 0.17, 95% CI 0.01 to 3.23). Antibiotic resistance was assessed and there were no cases in either group. There is inadequate evidence from randomised trials in favour of any particular antibiotic regimen for the treatment of suspected late onset neonatal sepsis. The available evidence is not of high quality. Although suspected sepsis and antibiotic use is common, quality research is required to specifically address both narrow and broad spectrum antibiotic use for late onset neonatal sepsis. Future research also needs to assess cost effectiveness and the impact of antibiotics in different settings such as developed or developing countries and lower gestational age groups.