This study was conducted to evaluate the accuracy and kinetics of interleukin 8 (IL-8) as a test for early-onset bacterial infections (EOBI) in neonates and to examine whether IL-8 would allow "unnecessary" antibiotic therapy to be reduced. First, IL-8 was measured 378 times on admission, along with C-reactive protein (CRP), immature to total neutrophil ratio (IT ratio) and blood cultures, in full-term and preterm neonates with suspected EOBI. Combined culture-proven and clinical EOBI were detected on admission by the combination of IL-8 > or = 70 pg ml(-1) and/or CRP > 10 mg l(-1) with 92% sensitivity and 74% specificity. An increased IL-8 was found in 62% of the infected infants, while CRP was still normal. In a second study period, IL-8 was determined prospectively in 331 infants who presented clinical signs of EOBI or had a birth history of amniotic infection. Antibiotic therapy was restricted to those infants with suspected infection who also had an increased IL-8 and/or CRP (n = 158). Another 39 infants received antibiotics on the basis of clinical signs despite negative IL-8 and CRP. Of 150 non-infected infants in whom IT ratio, IL-8 and CRP were available, treatment would have been indicated for 93 infants based on IT ratio and/or CRP (n = 77) or clinical signs (n = 16), but was only initiated in 55 infants based on IL-8 and/or CRP (n = 28) or clinical signs (n = 27), an apparent reduction in "unnecessary" antibiotic therapy of 40%.
Conclusion: The combination of IL-8 and CRP is a reliable test for the diagnosis of EOBI and may be helpful in enabling antibiotic therapy to be reduced in newborn infants.