Introduction: Without promptly started antibiotic therapy, early neonatal bacterial infections incur a significant mortality. Superficial bacteriologic samples at birth have in France a real place for the diagnosis and the decision to treat a neonate.
Objectives: In order to limit their indication and their choice, the aim of this article was to describe the proportion of neonates with samples and to determine the diagnostic value of the gastric aspirate, the ear swab and the placental sample.
Methods: Neonates born in the CHRU of Lille in 2005 and staying in the maternity ward were prospectively included. Criteria for samples, type of samples and diagnosis taken were noted. Sensibility, specificity, positive and negative predictive values and likelihood ratios for a positive test and a negative test were calculated.
Results and conclusion: This study included 3918 neonates; 1.7% (65 children) were infected according to our criteria; 42.3% received bacteriologic samples. In accordance with the Anaes guidelines (2002), if mothers were Group B Streptococci positive and received intrapartum antibiotics (up to 2 injections) or did not have any screening test without any other indication of samples, the neonate did not have to receive bacteriologic samples. The gastric aspirate was the best exam thanks to the excellent negative predictive value of its direct examination: 99.4% (IC 95%: 98.8-99.7), its high likelihood ratio for a positive test: 10.04 (IC 95%: 8.29-12.15) and its low likelihood ratio for a negative test: 0.16 (IC 95%: 0.09-0.29); this sample could restrict the antibiotics' ratio given to the neonate. Placental sample could be taken only in certain indications.