OBJECTIVE: Evaluate the mode of delivery of HIV-infected women and the risk of mother-to-child transmission. PATIENTS AND METHODS: A retrospective study conducted on HIV-infected women who delivered at the maternity ward of Bichat Hospital in Paris between 1st January 2000 and 31(st) December 2004. Pregnancy care, antiretroviral therapy, decision of the mode of delivery and neonate treatment were conformable to the French recommendations. RESULTS: The analysis was performed on 332 cases out of 358 pregnancies followed during this period. 75% received a Highly Active Anti Retroviral Therapy (HAART), 24% an AZT monotherapy and 1% did not receive any antiretroviral treatment. Plasmatic HIV viral load was under the level of detectability (50 copies/ml) for 64,6% of women under HAART and 28,7% of women under AZT monotherapy. Only 31,7% of women under HAART delivered vaginally. 44,7% of women under HAART with undetectable viral load at the moment of delivery delivered vaginally. 59,5% of women who were allowed to deliver vaginally had finally a vaginal delivery. 332 women gave birth to 341 babies with 9 twin pregnancies and one still-birth at 22 WA. Out of these 340 babies, 3 babies whose mother received HAART were HIV infected (2 in utero and 1 per-partum). DISCUSSION AND CONCLUSION: The reasons why only one third of HIV-infected women could deliver vaginally in this study are primarily the persistence of a detectable HIV viral load under HAART. Women's choice of the mode of delivery comes next, which depends on the quality of the counselling about the benefits and risks of the cesarean section in the context of HIV infection. The third reason is obstetrical contra indications to vaginal delivery in the context of HIV infection. In the future, it is possible to reduce the incidence of cesarean section in HIV-infected women by elevating the level of HIV plasmatic viral load which allowed vaginal delivery (1000 copies/ml), by improving the observance to antiretroviral treatment, by adaptating antiretroviral medications posology using determination of serum protease inhibitors concentration and by modifying obstetrical management with less restrictive contra indications to vaginal delivery. However the impact of prophylactic cesarean section when plasmatic HIV viral load is undetectable must still be evaluated.