RESUMO
OBJECTIVE: Difficulty in the diagnosis of neonatal sepsis has lead to the practice of superficial cultures. The usefulness of this practice has been criticized repeatedly. PATIENTS AND METHODS: Results of 3881 cultures performed on 735 newborns (age < or = 48 hr) with early infection risk were reviewed. The types of samples chosen were urine (UR), gastric aspirate (GA), pharyngeal (PS), external ear (ES), umbilical swab (US), meconium (MC) and blood. BBV for sepsis for the different samples is calculated, as well as which ones better reflect vertical transmission. RESULTS: At the time of the study, 342 newborns (46.5%) were already colonized. Bacteria most frequently isolated were: E. coli (92), PCN staphylococci (85), enterococci (83) and S. agalactiae (70). Twenty-nine newborns had bacteriemia, with S. agalactiae being the leading cause (16). ES correlated best with positive blood cultures (83.3%). UR culture sensitivity was significantly lower than that of all other samples. The Highest PPVs were for GA, PS and US. NPV was high and similar for all samples (98-99%). ES and US best reflected vertical transmission (p < 0.0001). MC provided the highest number of positive mixed cultures, most of them difficult to evaluate. CONCLUSIONS: Body surface sample advantages are: 1) The possibility of discarding a vertically transmitted infection. 2) Quicker positive results than those in blood whenever liquid blood culture media are employed. 3) The possible identification of the etiologic agent when the mother has been given antibiotics intrapartum. 4). Information about dangerous colonization without infection. In order not to lose this useful information, but to save cost and effort, we advise that the number of surface cultures be reduced, keeping only ES and PS?