Influence of Empirical Antibiotic (AB) Policy for Suspected Early-Onset Infection on the Bacterial Epidemiology in a Neonatal Unit

BACKGROUND: The worldwide epidemiology of causative bacteriae in early-onset infection has been remarkably stable along the last decades.Widespread use of broad-spectrum ABs for early-onset infection in neonates (N) can select multi-resistant bacteriae.
OBJECTIVE: We postulated that switching from Ampicillin (A) to Penicillin G (PG) for suspected early-onset infection in our three levels neonatal units, wathever the gestational age, could modify the selective pressure among Enterobacteriacae (EB) and decrease emergence of resistant inducible bacteriae.
DESIGN/METHODS: This is a prospective bacteriological survey of 22 months comparing three successive time-periods (P1,2,3) reflecting a progressive switching from Ampicillin-Amikacin (A-Ak) to Penicillin G-Amikacin (PG-Ak). It accounts for respectively A-Ak and PG-Ak (%) : P1 : (66 and 34); P2 : (48 and 52); P3 : (12 and 88). Systematic bacterial screening included blood culture and multiple swabs on admission. Multiple swabs and faeces culture were repeated once a week. Additional cultures were taken when clinically indicated. Antibiograms were done with the Viteck2 System (Biomerieux). Data were collected on the Bacteriological logicial Infopartner R (Nancy).
RESULTS: Antenatal AB policy was not modified along the survey. We included 334, 340 and 306 N for P1, P2 and P3, respectively with the same proportion of very preterms. The neonatal morbidity was similar between the three Ps. In respect to the bacteriological screening, no major differences in the proportion of bacterial groups were observed. However, the proportion of inducible EB significantly decreased between the P1 and the P3 (39.7% vs 21.7%, p