Objectives: To identify factors affecting recruitment into a randomised trial involving HIV infected pregnant women. To compare the population enrolled in a Phase III trial to that enrolled during the same period in a parallel longitudinal cohort, and to assess the generalisability of the trial results.
Material and methods: The PACTG 316 trial was initiated in the USA and subsequently extended to Europe. Based on their involvement in an ongoing European cohort study, clinicians in 27 antenatal HIV reference centres in Italy, Spain, Sweden, UK, Belgium, Germany, Switzerland, Netherlands and Denmark were asked to participate. In centres with local, national and American approval, eligible women attending antenatal clinics were offered the chance to take part in the trial, which aimed to evaluate the additional use of nevirapine during labour and neonatally to reduce HIV mother-to-child transmission (MTCT). HIV-infected women who did not enrol in the trial were enrolled in the European Collaborative Study (ECS). Reasons for non-randomisation were recorded. Clinical and laboratory information on mother-child pairs were collected according to a standard protocol.
Results: Between February 1999 and June 2000, 247 women were enrolled in the ECS cohort and 118 were randomised in the 316 trial. Reasons for non-randomisation included the presence of the placebo arm, randomisation procedures and delays in obtaining approval from the various regulatory bodies. Women in the trial were younger, and their HIV disease was less advanced than those included in the ECS. The MTCT rate in the ECS was higher than in the trial.
Conclusions: Differences between women who participated in the trial and those who did not had an effect on the absolute vertical transmission rate, but not on the relative effectiveness of the intervention assessed within the trial. Extrapolation of the trial MTCT rates to the general HIV infected population may be inappropriate.