University of California, San Francisco Logo

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

Home > Global Health Literature Digest > Efficacy of Carraguard
Efficacy of Carraguard for prevention of HIV infection in women in South Africa: a randomized, double-blind, placebo-controlled trial
Global Health Sciences Literature Digest
Published September 7, 2009
Journal Article

Skoler-Karpoff S, Ramjee G, Ahmed K, et al. Efficacy of Carraguard for prevention of HIV infection in women in South Africa: a randomized, double-blind, placebo-controlled trial. Lancet 2008;372:1977-87.

In Context

HIV incidence and prevalence is higher among women than men in sub-Saharan Africa.(1,2) Social, cultural, and financial factors limit women's use of many existing HIV prevention strategies, such as abstinence, mutual monogamy, and consistent use of condoms.(3) Thus, there is a need for a female-controlled method to prevent sexual transmission of HIV. Although previous trials of vaginal microbicides have not been found to reduce HIV infections and may increase infections in women,(4,5,6,7) Carraguard, a carrageenan (seaweed)-based microbicide, has been shown to prevent mucosal transmission of HIV both in the presence and absence of semen and to be safe when used vaginally.


To measure the efficacy of Carraguard at reducing sexually acquired HIV infections among women


Clinical sites in KwaZulu-Natal (one site) and in the Western Cape (two sites), South Africa

Study Design

Randomized, double-blind, placebo-controlled trial


HIV-negative, sexually active women


HIV incidence and adjusted hazards ratio for time to seroconversion


Between March 17, 2004 and March 31, 2007, 6,202 sexually-active, HIV-negative women aged 16 years and older, who were neither pregnant nor planning a pregnancy in the next two years, were not four weeks or less post-pregnancy, had not had a Papanicolaou test graded as carcinoma, did not inject drugs, and were not in other clinical or HIV prevention trials were recruited from local health clinics, shopping centers, churches, taxi ranks, and other community venues and entered into the trial. Women were randomized using a 1:1 ratio to receive Carraguard or placebo gels. They were provided HIV risk reduction counseling along with condoms, instructed to use them with the gels during each act of vaginal intercourse, and asked to return applicators at follow-up visits. After the baseline clinic visit, participants returned at one month, three months, and then every three months thereafter for a minimum of nine and a maximum of 24 months. At each visit the women were tested for HIV, pregnancy, and other sexually transmitted infections (STIs). Treatment was provided as needed. At the time of screening and yearly thereafter, women underwent a pelvic and speculum examination and a Papanicolaou test. Women who became pregnant or HIV-infected were discontinued. At baseline all participants had blood drawn and stored for HIV PCR DNA testing of the women who seroconverted within three months of enrollment. Laboratory data from seroconverters was reviewed by three HIV diagnostic experts who were blinded to data other than that from the laboratory.

Randomization was done by a statistician not connected with the trial who produced a computer-generated block randomization scheme which was stratified by site and assigned each participant to gel A (Carraguard) or gel B (placebo). Gel tubes were identical and had bar codes that were used to document assignment and to track distribution and returns of gel and gel applicators.

Adherence was measured by self-report and staining of applicators to determine if there had been vaginal insertion of the applicator. The number of applicators returned unopened, the number opened and inserted (based upon staining), and the number opened but not inserted was tracked.

Women were asked about the number of acts of vaginal intercourse in the previous two weeks and, for the last act only, asked if gel was used and if a condom was used. The use of gel/condoms at this sex act was used as a proxy for all acts in that time period. The number of sex acts during which gel was used was estimated by dividing the average number of applicator insertions per week (based on results from staining the applicator) by the average number of sex acts per week (as reported by participants).

The time to HIV infection was compared between Carraguard and placebo groups using the Kaplan-Meier product limit method. Differences in the time to seroconversion were measured using the log-rank test and stratified by site. The date of seroconversion was taken to be the midpoint between the first visit at which the participant tested HIV-positive and the previous visit when the HIV test result was negative. The time to seroconversion was defined as the difference in the number of days between the calculated seroconversion date and the enrollment data plus one. Participants who completed the study and remained HIV-uninfected were censored at the last date they were HIV-negative. Women who were lost to follow-up were censored at the date of their last HIV-negative test.

The hazards ratio of HIV seroconversion comparing Carraguard to placebo was done using a Cox proportional hazards model that controlled for age, site, presence of STI at screening, baseline coital frequency, and baseline condom use. A second model was tested for site interaction. The Wilcoxon test was used to conduct a sensitivity analysis of early differences in the distribution of time to seroconversion.

Safety outcomes included incidence of abnormal genital epithelium and STIs. Distribution of these outcomes in the treatment and placebo groups was compared using the log-rank test. Differences in the proportions of outcomes between the groups were also compared.

Safety analyses were done using the intention-to-treat population while the efficacy outcomes were done using the efficacy population (a subset of the intention-to-treat population). The efficacy population excluded those for whom outcome data were not available (e.g., HIV-positive at enrollment, no HIV tests after enrollment, and those who returned all applicators unopened). Safety was monitored by an independent safety monitoring board who reviewed the data at three interim points.


A total of 9,564 women were screened, of which 6,202 were enrolled and randomized. Study completion in the two groups was similar (69% in the Carraguard group and 68% in the placebo group). Loss to follow-up occurred in 14% of the women (420 in the intervention and 418 in the control group). Demographic characteristics, sexual behaviors, and contraceptive use were similar in the two groups.

There were 285 seroconversions 134 of which occurred in the Carraguard group and 151 in the placebo group. The time to seroconversion among 3,011 women in the Carraguard and 2,994 women in the placebo groups was similar (3.3/100 women-years and 3.8/100 woman-years, respectively: log-rank test, P=0.30 and relative hazard 0.87, 95% confidence interval [CI] 0.69-1.09). Use of gel, as based upon applicator testing, occurred in 41.1% of sex acts in the Carraguard group and 43.1% of sex acts in the placebo group. Of the 72 (2%) women in the Carraguard group and the 78 (3%) women in the placebo who reported serious adverse events, only one event was determined to be related to gel and this was from the placebo gel.


Carraguard was not effective in preventing HIV through vaginal intercourse.

Quality Rating

This was a high-quality study in that it was truly random; b) there was proper concealment of allocation to study arms; c) loss to follow-up was accounted for and was not a large proportion of participants; and d) an intention-to-treat approach was used in the analysis.

Programmatic implications

The negative findings from this study do not provide for any change in HIV prevention strategies. This study adds to the large body of evidence that has not found vaginal microbicides to be effective in reducing HIV transmission. As such, additional studies that can address the need for a female-controlled methods to prevent acquisition of HIV are needed.


  1. UNAIDS/WHO. AIDS epidemic update. Geneva: UNAIDS/World Health Organization, 2007 Accessed September 8, 2008.
  2. Rehle T, Shisana O, Pillay V, Zuma K, Puren A, Parker W. National HIV incidence measures-new insights into the South African epidemic. S Afr Med J 2007;97:194-9.
  3. Johnson KE, Quinn TE. Update on male circumcision: prevention success and challenges ahead. Curr Infect Dis Rep 2008;10:243-51.
  4. Van Damme L, Ramjee G, Alary M, et al. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: a randomised controlled trial. Lancet 2002;360:971-7.
  5. Feldblum PJ, Adeiga A, Bakare R, et al. SAVVY Vaginal Gel (C31G) for prevention of HIV infection: a randomized controlled trial in Nigeria. PLoS ONE 2008;3:1474.
  6. Peterson L, Nanda K, Opoku BK, et al. SAVVY (C31G) Gel for prevention of HIV infection in women: a phase 3, double-blind, randomized, placebo-controlled trial in Ghana. PLoS ONE 2007;2:1312.
  7. Van Damme L, Govinden R, Mirembe F, et al. Lack of effectiveness of cellulose sulfate gel for the prevention of vaginal HIV transmission. N Engl J Med 2008;359:463-72.