The association between neonatal sepsis and labour epidural analgesia (EA) was evaluated. In a retrospective matched case-control study, 453 women delivering with EA (EA group) were matched on parity with 453 women with deliveries without EA (non-EA group). Significantly more neonates born in the EA-group had fever ≥ 38.0°C (11.6% vs 1.8%, p < 0.001) at birth. The overall incidence of neonatal sepsis, based on clinical symptoms and defined as proven (by a positive blood culture) or suspected (no positive blood culture), was significantly higher in the EA group (6.0% vs 2.2%; p = 0.002), but the incidence of proven neonatal sepsis alone was not (0.4% vs 0%; p = 0.250). EA turned out to be an independent risk factor for neonatal sepsis (adjusted OR 2.43, 95% CI 1.15-5.13; p = 0.020). However, in the EA group as well as the non-EA group, the incidence of neonatal sepsis was significantly higher in mothers with intrapartum fever compared with afebrile mothers (11.0% vs 2.9% in the EA group; p = 0.004; 8.2% vs 1.3% in the non-EA group; p = 0.006). Therefore we conclude, that the positive association between neonatal sepsis and labour EA is possibly mediated by maternal intrapartum fever.