OBJECTIVE: To evaluate acceptability of voluntary HIV counselling and testing (VCT) by pregnant women in the context of clinical trials assessing interventions to reduce mother-to-child transmission (MCT) of HIV in developing countries. METHODS: During September-October 1997, 13 studies located in West (Abidjan, Bobo Dioulasso), East (Nairobi, Mombasa, Dar Es Salaam, Blantyre, Lusaka, Harare) and South Africa (Soweto, Durban), and Thailand (Bangkok) were included in a cross-sectional mailing survey about the acceptability of VCT in antenatal clinics. Acceptance rate, return rate, overall acceptability of VCT (acceptance of both pre- and post-VCT sessions) were obtained using a standardized questionnaire. RESULTS: The median overall acceptability of VCT was 69% (range, 33-95%). Overall acceptability of VCT most frequently depended on return rates because acceptance rates were generally high. Where several studies were conducted in parallel in the same city or the same country, overall acceptability rates of HIV testing were generally comparable even if the intervention programmes differed. Overall acceptability rates of VCT were high in antenatal clinics where a particular effort in implementing VCT programmes had been made. CONCLUSIONS: This international survey shows that despite many obstacles, VCT is feasible and acceptable for pregnant women aiming to reduce their risk of transmitting HIV to their children.