Background:  MDP 301 is a phase III, randomized, double-blind, placebo-controlled trial. The primary objective is to assess whether either 2% or 0.5% PRO2000 gel is effective in reducing vaginally acquired HIV infection. The study commenced in October 2005 and is conducted at 6 sites:  3 sites in South Africa, Zambia, Tanzania, and Uganda. A target of 9673 HIV^– women are to be enrolled and followed for 52 weeks at all sites, except Uganda, where follow-up is 2 years. This abstract describes the preliminary patterns of seroconversion.

Methods:  HIV Rapid testing is done at screening, weeks 12, 24, 40. and 52, (weeks 66, 76, 88, 100, 104 in Uganda). At these time-points, including at enrolment and week 4, serum is stored. Buffy coat is stored at enrolment and at weeks 24, 40, and 52 (and week 104, Uganda only). In the event of suspected seroconversion, the complete set of sera and buffy coats are sent to the central laboratory for confirmation. The algorithm involves testing complete sample sets. The serum samples are tested using Abbott AxSYM HIV Ag/Ab Combo enzyme-linked immunosorbent assay (ELISA) and BioRad Genetic Systems HIV-1 ELISA in parallel. Buffy coat samples are tested with the Roche qualitative DNA PCR Version1.5 assay. A redraw specimen is required to re-confirm the results.

Results:  Until September 2007, 166 seroconverters have been detected by Rapid testing. Of 166 participants, 154 (93%) have been tested by the central laboratory:  30 (24%) were found to be HIV⁺ at or before enrolment. Of these, 6 HIV antibody positive samples were missed by Rapid testing at screening and 24 sample sets were positive at enrolment. Of the 24 positive enrolment samples, 14 (58%) were classified as “acute infections,” 9 having a positive HIV Ag/Ab ELISA and polymerase chain reaction (PCR) result with negative HIV antibodies and 5 were positive by PCR only. Of the 124 remaining HIV seroconverters, 17 were diagnosed as “acute infections” with positive HIV Ag/Ab or PCR results.

Conclusions:  Accurate analysis of the endpoint requires exclusion of HIV infection at enrolment. To date, the testing algorithm has detected a high rate (24%) of HIV infections at the enrolment visit, more than half of which were “acute HIV infections” at that time-point. In addition, these results highlight that incorrect HIV Rapid test results occur, resulting in erroneous enrolments. The high proportion (93%) of tested sample sets allows for optimized review of the study data, which is critical for the safety analysis of the trial.