OBJECTIVE: To determine the uptake and acceptability of different methods of a universal offer of voluntary HIV testing to pregnant women. DESIGN: Randomised controlled trial involving four combinations of written and verbal communication, followed by the direct offer of a test. The control group received no information and no direct offer of a test, although testing was available on request. SETTING: Hospital antenatal clinic covering most of the population of the city of Edinburgh. SUBJECTS: 3024 pregnant women booking at the clinic over a 10 month period. MAIN OUTCOME MEASURES: Uptake of HIV testing and women's knowledge, satisfaction, and anxiety. RESULTS: Uptake rates were 6% for those in the control group and 35% for those directly offered the test. Neither the style of leaflet nor the length of discussion had an effect on uptake. Significant independent predictors of uptake were a direct test offer; the midwife seen; and being unmarried, previously tested, and younger age. Knowledge of the specific benefits of testing increased with the amount of information given, but neither satisfaction nor anxiety was affected by the type of offer. CONCLUSIONS: The universal offer of HIV testing is not intrusive and is acceptable to pregnant women. A policy of offering the HIV test to all women resulted in higher uptake and did not increase anxiety or dissatisfaction. Uptake depends more on the midwife than the method of offering the test. Low uptake rates and inadequate detection of HIV infection point to the need to assess a more routine approach to testing.