One of the main obstacles to HIV prevention in Africa remains the insufficient number of HIV tests performed. The low percentage of individuals aware of their serostatus is due in part to the insufficient availability of HIV testing, but also to individuals' refusal to have the HIV test. Because affordable treatment is now available, it is possible to implement on a much greater scale programs of prevention of mother-to-child HIV transmission, accompanied by the expansion of prenatal HIV testing. It is therefore important to understand the reasons women refuse these tests. Here we analyse the women who refused the offer of prenatal HIV testing in the DITRAME Plus research program, intended to prevent mother-to-child transmission in Abidjan from 2001 through 2005. Three groups of women were followed for two years after they were offered HIV counselling and testing during an antenatal consultation: 347 HIV-infected women, 393 seronegative women, and 62 women who refused HIV testing. Nine months after delivery, HIV testing was again offered to the latter group. We collected quantitative data on social and demographic characteristics, sexual behavior, and communication with their male partners about STIs, HIV, and HIV testing, before and after the pregnancy. In-depth interviews were conducted with 15 women who refused HIV-testing. We sought to determine if their sociodemographic and behavioural profile was closer to that of HIV-positive or seronegative women, to assess the effects of prenatal counselling and the offer of testing on their attitudes about HIV risk, and to measure the percentage of women who accepted testing when the offer was repeated 9 months postpartum. Women who refused HIV-testing had a sociodemographic profile similar to that of the women who accepted testing and were seronegative. They did not have more at-risk behaviours. These women offered several reasons for their refusal, including avoidance of the anguish of a positive test result and the desire to ask their husbands first. Among the women who initially refused prenatal testing, only 23% had discussed STI/HIV issues with their male partner; after the testing offer, more than 90% suggested that their partner have an HIV test. Finally, 20% accepted the postpartum test offer, and those whose male partner had an HIV test were four times more likely than the others to accept (RR = 4.05 [1.55-10.58]). This study shows that prenatal counselling and the offer of HIV testing have beneficial effects on HIV prevention within the couple, even for women who refuse testing. It also points out that the decision to accept testing may take time and that repeating the offer is worthwhile. Finally, our results confirm the importance of the role of the regular partner in the acceptance of HIV testing, and reinforce the relevance of a couple-centred approach to voluntary counselling and testing.