In 2001 France issued a new set of guidelines for the use of antenatal antibiotics (AA). These guidelines recommended intrapartum antimicrobial prophylaxis (IAP) to prevent group B streptococcal (GBS) disease and AA to prolong pregnancy in the event of preterm premature rupture of membranes (AA for PPROM). This study aims to determine the effects of AA, recommended by national guidelines, on the incidence and distribution of pathogens in early-onset neonatal sepsis (EONS). We performed a population-based, prospective, observational study of level II and III perinatal centres throughout the region of Alsace, a northeastern area of France, between March 2004 and February 2005. The study population included all neonates with confirmed or probable EONS, who were treated with antibiotics for at least 5 days. We analysed exposure to AA, as well as clinical and microbiological data obtained from medical records. A total of 20 131 neonates were born during the study period, and 217 were included in the study. Of these, 24 subjects had confirmed sepsis, 140 had probable sepsis and 53 had possible EONS. The overall incidence of confirmed EONS was 1.19 per 1000 births. The infecting bacteria was GBS in 15 of 24 (62.5%) confirmed EONS cases (incidence: 0.75 per 1000 births) and in 81 of 140 (58%) probable sepsis cases. Escherichia coli was identified in 6 of 24 (25%) cases of confirmed EONS (incidence: 0.3 per 1000 births) and in 30 of 140 (21%) cases of clinical sepsis. Among E. coli infections (n= 36), amoxicillin resistance (n= 18) was statistically linked with AA use (P = 0.045). This link was significant in cases of PPROM (P = 0.015), but not when IAP was administered to prevent GBS disease (P = 0.264). IAP was not performed in 18 of 60 (30%) cases and 32 of 93 (34%) cases, despite positive screening or the presence of risk factors for EONS, respectively. Group B streptococcus remains the predominant pathogen in the era of AA. Aminopenicillin-resistant E. coli infections seem to be linked to prolonged AA in cases of PPROM and appear to preferentially affect preterm infants. Therefore, postnatal treatment strategies should consider this possible effect. Our data indicate that the current policy of GBS maternal prophylaxis is not associated with an excessive risk of pathogen resistance. Considering the high incidence of GBS EONS in our region, possible progress could result from better observance of guidelines. These results strengthen the need for continuation of surveillance.