University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Home > Global Health Literature Digest > Effectiveness and Safety
Effectiveness and safety of vaginal microbicide 1% tenofovir gel for prevention of HIV infection in women
Global Health Sciences Literature Digest
Published August 6, 2010
Journal Article

Abdool Karim Q, Abdool Karim SS, CAPRISA 004 Study Group. Effectiveness and safety of vaginal microbicide 1% tenofovir gel for prevention of HIV infection in women [Abstract TUSS0502]. 18th International AIDS Conference, Vienna, Austria, July 20, 2010.

Kashuba AD, Abdool Karim SS, Kraft E, et al. Do systematic and genital tract tenofovir concentrations predict HIV seroconversion in the CAPRISA 004 tenofovir gel trial? [Abstract TU220503]. 18th International AIDS Conference, Vienna, Austria, July 20, 2010.

Sokal D, Karim Q, Omar Z, et al. Safety of 1% tenofovir vaginal microbicide gel in South African women: results of the CAPRISA 004 trial [Abstract TUSS0504]. 18th International AIDS Conference, Vienna, Austria, July 20, 2010.

In Context

Vaginal microbicides have long been considered an important theoretical component of comprehensive HIV prevention, but success in clinical trials has been elusive.(1, 2) Several compounds have been brought to Phase IIb and III trials; none have been proven efficacious and at least one, containing the spermicide nonoxynol-9, has been proven harmful.(3) Recent trials have studied formulations that include antiretroviral drugs based in part on their success when administered orally to prevent transmission from mother to child.(4) The CAPRISA 004 trial, conducted in South Africa, is the first Phase IIb trial to have available results, which were presented at the 18th International AIDS Conference in Vienna in late July 2010.


To compare the efficacy and safety of 1% tenofovir intravaaginal gel with placebo gel for the prevention of sexually transmitted HIV


Randomized, placebo-controlled, double-blind Phase IIb trial


Participants were HIV-uninfected women who are sexually active (having had two episodes of sexual intercourse in prior 30 days), 18 years of age or older, and from one urban and one rural site in KwaZulu Natal, South Africa. From earlier studies, this population was known to be at elevated risk for HIV infection. Forty-nine percent of the study population had negative results of herpes simplex type 2 (HSV-2) serology tests.


Participants were tested twice at baseline with rapid HIV antibody tests and HIV polymerase chain reaction (PCR). Participants were randomized to receive either 1% tenofovir gel or an identical placebo. Participants were instructed to insert the prefilled applicators of gel intravaginally up to 12 hours before sex, as soon as possible but within 12 hours after sex, and no more than twice within a 24-hour period. Women were followed for up to 40 months and were seen for monthly pregnancy tests, quarterly pelvic examinations, and blood draws at 3, 12, and 24 months and at study exit. Endpoints were HIV infection (as measured by PCR); HSV-2 seroconversion; and, in a subset of women, serum, genital tract, and vaginal and cervical (by tissue biopsy) levels of tenofovir and its metabolically active form, tenofovir diphosphate.


Eight hundred eighty-nine participants were randomized, 445 to the tenofovir arm and 444 to the placebo arm; 843 (95%) completed the study. The groups were comparable on a number of sexual risk factors; the mean number of sexual partners was 3.0 in the treatment group and 3.6 in the placebo group (P=0.78), and only approximately 30% in either group used condoms always. By intention-to-treat analysis, 38 women in the treatment group became HIV infected (incidence 5.6 per 100 person-years) compared to 60 in the placebo group (9.1 per 100 person-years, relative risk [RR]=0.61, 95% confidence interval [CI]: 0.4-0.96). This corresponds to a 39% lower risk of becoming infected. Adherence was clearly important. Efficacy decreased from 54% among those who used gel >80% of the time to 38% among those who used it 50%-80% of the time to 28% among those who used it <50% of the time. Among the 38 women in the treatment arm who became infected with HIV, the log mean viral load was higher, but not significantly so, than viral load among infected women in the placebo arm (4.65 vs. 4.30, P=0.15). No participant who had seroconverted in the tenofovir arm had evidence of tenofovir resistance. There were no hepatic adverse events, including among those with hepatitis B infection; no safety concerns in pregnancy; and no increase in HIV risk behavior while using the gel. There was a small increase in mild diarrhea in treatment arm participants (17% vs. 11%, P=0.015). Of 434 participants uninfected with HSV-2 at the start of the trial, 208 were randomized to the tenofovir arm and 226 to the placebo arm; 426 completed the study. There were 29 seroconversions in the treatment arm and 58 in the placebo arm (incidence 9.9 vs. 20.2 per 100 person-years, RR=0.49, 95% CI: 0.30-0.78). In the subset of women in whom tenofovir levels were measured, a greater proportion of those who did not seroconvert had measurable tenofovir than had those who seroconverted.


1% tenofovir gel was safe and efficacious in preventing HIV and HSV-2 infection in highly at-risk women in South Africa. Note, however, that the upper relative risk bound for HIV was 0.96, indicating a possible protective effect as low as 4% per year. Adherence was essential; efficacy was almost twice as high among women who used the gel >80% of the time compared to those who used it <30%.

Study Quality

This was a high-quality study. The methods for randomization and blinding were described adequately, as was the description of participant enrolment and retention.

Programmatic Implications

The CAPRISA 004 is a landmark prevention trial, and its implications are potentially sweeping. This is the first study that has proven in principle that microbicides can lead to a significant decrease in HIV and HSV-2 infection. Several speakers at the conference, however, suggested that results from additional studies must be considered before adopting this technology wholesale. Also, as noted above, the effect of treatment is highly adherence dependent, suggesting that adherence or the potency and/or durability of the formulation must increase to maximize this important intervention's public health impact. If the full potential of this intervention is realized, mathematical models presented later at the conference calculated, a microbicide with an efficacy of 40% could avert 271,000 new infections in South Africa if coverage was 50% and 460,000 if coverage was 80% over the next 10 years.(5)


  1. Morris GC, Lacey CJ. Microbicide and HIV prevention: lessons from the past, looking to the future. Curr Opin Infect Dis 2010 Feb;23(1):57-63.
  2. Wilkinson D, Tholandi M, Ramjee G, Rutherford GW. Nonoxynol-9 spermicide for prevention of vaginally acquired HIV and other sexually transmitted infections: systematic review and meta-analysis of randomised controlled trials including more than 5000 women. Lancet Infect Dis 2002 Oct;2(10):613-7.
  3. Van Damme L, Govinden R, Mirembe FM, et al Lack of effectiveness of cellulose sulfate gel for prevention of HIV transmission. N Engl J Med 2008;359:463-72.
  4. Mnyani C, McIntyre J. Preventing mother-to-child transmission of HIV. BJOG 2009;116 (suppl 1):71-6.
  5. Williams BG, Abdool Karim S, Gouws E, Abdool Karim Q. Impact of tenofovir gel on the HIV epidemic in South Africa: a mathematical model to estimate the effect of the CAPRISA 002 microbicide trial results [Abstract LBPE27]. 18th International AIDS Conference, Vienna, Austria, July 23, 2010.