icon- folder.gif   Conference Reports for NATAP  
  16th CROI
Conference on Retroviruses and Opportunistic Infections Montreal, Canada
February 8-11, 2009
Back grey_arrow_rt.gif
HIV Prevention at CROI 2009
  Jared Baeten, MD PhD
Connie Celum, MD MPH
University of Washington
HIV prevention has occupied an increasingly central place at CROI in recent years, with this year's conference no exception. Interesting and important observational and interventional studies were presented, relevant to a variety of settings and risk groups worldwide.
Topical Microbicides
One of the most exciting presentations at this year's CROI was the final results of HPTN 035, a phase IIb, double-blind, controlled clinical trial evaluating the safety and efficacy of the topical vaginal microbicides BufferGel and PRO2000 for prevention of HIV acquisition (Abdool Karim, abstract 48LB). The study, run by the US National Institutes of Health HIV Prevention Trials Network and Microbicides Trials Network, enrolled 3087 HIV-negative women at sites in Malawi, South Africa, the United States, Zambia, and Zimbabwe. In a unique 4 arm design, participants were randomized to one of the two active products, to a placebo gel, or to a "no gel" arm that permitted assessment of any protective benefit of the placebo as a mechanical barrier in preventing HIV. The two microbicides chosen were not "specific" to HIV in their mechanism of action; BufferGel was postulated to act as a pH buffer in the vagina in the presence of semen, while PRO2000 is a synthetic polyanionic polymer that is hypothesized to interfere with HIV attachment to host cells. Both products had a host of pre-clinical and early clinical studies of safety and surrogates of efficacy completed before they were moved to full-scale testing in this trial. During an average follow-up of just over 20 months, the investigators achieved a remarkable 94% retention rate, comparable across all four arms. Gel use was reported to be high - 81% of sex acts - and importantly, was similar across the three gel groups. Condoms were used with 74% of sex acts, with slightly higher use in the no gel arm (81% of acts) compared with the gel arms (71-72% of acts). In intent-to-treat analysis, ∼30% fewer HIV infections occurred among women randomized to PRO2000 compared with those either of the control arms (Table), an effect that was short of statistical significance. BufferGel neither increased nor decreased HIV risk. In a per-protocol analysis, excluding protocol-required time off of study product (mostly due to pregnancy), the reduced HIV risk for PRO2000 was essentially the same. The rates of HIV acquisition were similar for women randomized to placebo gel and to the no gel arms.


In subgroup analyses, there was a suggestion of a "dose-response": for Pro 2000, with 9% efficacy for women who were "low" users (less than the median of 85% of sex acts with gel use), compared with 44% efficacy for those with higher than the median use of gel. Among women who had both low condom use and high gel use (those for whom the gel might be expected to have the highest efficacy), PRO2000 had a 78% decrease in HIV risk compared with placebo gel.
Both products were safe, with no increased risk of systemic or local toxicity compared with placebo. The 11% annual pregnancy rate was comparable across the study arms, and confirmed that neither active product have contraceptive properties. Neither product reduced the risk of other sexually transmitted infections, including most notably herpes simplex virus type 2 (HSV-2), although animal model studies had suggested that PRO2000 might also be effective in preventing HSV-2 infection.
Thus, PRO2000 0.5% vaginal gel showed 30% efficacy in the prevention of HIV acquisition in women, just shy of statistical significance, an effect that was similar when compared against both placebo gel and non-gel control arms, while BufferGel showed no efficacy. Results were corroborated by subgroup analyses that evaluated efficacy by gel and condom use.
What are the next steps? The modest effect (30% HIV reduction overall) and non-statistical significance of the findings of HPTN 035 are not sufficient to move PRO2000 0.5% gel at this time from an investigational stage to licensure and global distribution to women at risk of HIV. However, the Medical Research Council of the United Kingdom is currently conducting a larger phase III efficacy trial of 0.5% PRO2000, with results expected late this year. Equally positive findings in that study could push for revised consideration of the future of this specific product. In any case, though, these are very encouraging results for the microbicide field, with a well-conducted trial that provided substantial evidence for the concept that a topical product can reduce HIV acquisition in women. Previously, several large efficacy studies of other surfactant and non-specific microbicide products found no benefit in preventing HIV acquisition. The positive signal seen in this trial, coupled with high retention and adherence to the gel products, gives new hope to those working to develop a topical, female-controlled HIV prevention product.
Pre-Exposure Prophylaxis (PrEP)
Pre-exposure prophylaxis (PrEP) refers to HIV uninfected persons using an antretroviral agent (e.g., oral or formulated as a topical microbicide) prior to high-risk exposures in an effort to block initial HIV infection, a situation analogous to malaria prophylaxis in travelers to malaria-endemic areas. There are currently eight safety and efficacy trials of antitretroviral PrEP ongoing or planned worldwide, a sign of high scientific enthusiasm for this novel potential HIV prevention strategy. While the first PrEP efficacy results are not expected until at least 2010, there was still much discussion of PrEP at this year's CROI.
In a forward-thinking plenary (abstract 73), Dr. Sharon Hillier, from the University of Pittsburgh, reviewed the rationale for PrEP, summarized current efficacy trials, and anticipated needs for future implementation of PrEP, if successful. Current PrEP trials are using oral tenofovir (Viread), oral co-formulated emtricitabine-tenofovir (Truvada), or intravaginal tenofovir gel, the last an indication of the direction that the topical microbicide field has taken into incorporating antiretroviral agents into its products. The scale of the current trials is large, with >20,000 participants anticipated to enroll across the eight trials; however, each trial is distinct in its study population and route of HIV exposure, including men who have sex with men, injection drug users, high- and low-risk women, and HIV serodiscordant couples. Thus, each trial will contribute distinct but complementary information about PrEP safety and efficacy in different risk groups, the efficacy and acceptability of oral and topical PrEP, the incidence of drug-resistant virus in breakthrough infections, and needs for safety monitoring and other factors that will be essential for PrEP implementation to be accepted for different populations. A long-term goal is to test additional PrEP agents (e.g., antiretroviral medications that do not overlap with those used for HIV treatment) and intermittent dosing schedules that might reduce cost and dosing burden potentially without compromising efficacy.
In two abstracts, investigators from the US Centers for Disease Control and Prevention presented data from non-human primate models mimicking the effect of PrEP on mucosal HIV transmission. The first (Dobard, abstract 46) explored the efficacy of 1% tenofovir gel and 5% emtricitabine / 1% tenofovir dual-agent gel in a vaginal exposure model. Pig-tailed macaques underwent twice-weekly, low-dose vaginal challenge with a pathogenic simian-HIV virus for 10 weeks, with gel applied just 30 minutes before each challenge. Among 9 animals assigned placebo gel, 8 became infected, after a median of 4 challenges. In contrast, both active gels provided 100% protection across 20 mucosal challenges, a finding that was highly statistically significant. Tenofovir and emtricitabine were measurable in a majority of plasma samples collected after gel application and just before viral challenge, although at low concentrations, suggesting minimal absorption. Thus, 1% tenofovir gel, alone or in combination with emtricitabine, conferred complete protection against repeated vaginal exposure in this animal model. Encouragingly, a single gel application shortly before exposure was sufficient for protection, suggesting intermittent or relatively unplanned use may work. These results are extremely encouraging for those planning studies of antiretroviral-containing topical microbicides.
The second abstract (Garcia-Lerma, abstract 47) evaluated the efficacy of oral emtricitabine/tenofovir PrEP when provided at different times prior to viral exposure, a situation somewhat akin to intermittent PrEP use. The emtricitabine/tenofovir dose was equivalent to human Truvada dosing. A repeat rectal exposure model, with low-dose viral challenge done weekly for 14 weeks, was used. Oral PrEP significantly reduced the risk of viral infection. The timing of PrEP dosing related to its efficacy: 31 of 32 control animals became infected, compared to half of animals given the first PrEP dose either 2 hours before or 2 hours after viral exposure, 2 of 6 animals given PrEP 7 days prior to exposure, and only 1 of 6 animals in two groups given PrEP 1 or 3 days prior to exposure. These data suggest oral emtricitabine/tenofovir PrEP maintains high efficacy even if dosed 3 days prior to exposure, suggesting daily use may not be required. Tenofovir has a long intracellular half-life, which likely accounts for this effect. However, in this study, dosing just 2 hours before (or even 2 hours after) an exposure was not as effective, which may reflect time needed to absorb the oral dose, transport to tissues, and phosphorylate to the active moiety of tenofovir.
A few additional important findings were reported from the oral PrEP study. Compared with controls, animals who received PrEP had blunted viremia during primary infection - by 1.5 log10 copies/mL - suggesting that if lower viral set point is also observed in humans who have breakthrough infections on PrEP, then this could reduce HIV transmission during acute infection and potentially lead to slower disease progression in those who become infected in spite of PrEP. Importantly, no viral resistance was detected in animals with breakthrough infections - clearly a central question for ongoing field studies of PrEP. Finally, a poster from this group (Owen, abstract 1072) demonstrated that animals with breakthrough infections do not have delayed HIV seroconversion - important as the goal is to detect HIV seroconversion promptly on PrEP, to limit time on mono or dual PrEP therapy if infection occurs and thus avoid selection of drug-resistant viruses.
These studies continue a series of carefully conducted and innovative experiments using oral and topical PrEP in an animal model system (including abstracts 32LB and 986 from CROI 2008). Overall, oral and vaginal PrEP using tenofovir or combination emtricitabine/tenofovir had high efficacy in preventing infection in this macaque mucosal challenge system. These findings are very encouraging and support ongoing clinical trials of oral and vaginal antiretroviral PrEP for HIV prevention.
Antiretroviral Treatment for HIV Prevention
A prime predictor of HIV transmission risk in epidemiologic studies is the level of HIV in blood. When HIV infected persons are treated with antiretroviral therapy (ART), plasma and genital viral concentrations are decreased, frequently to very low levels, which is anticipated to substantially reduce the infectiousness of HIV infected persons. Recent mathematical modeling studies have argued that increasing the number of persons on ART could substantially reduce new HIV transmissions, particularly if the number of persons who know their HIV serostatus is substantially increased and they are effectively linked to HIV care and ART is initiated (called the 'test and treat' strategy). Two abstracts provided data on ART and HIV transmission from large observational studies conducted among HIV serodiscordant couples. The first (Reynolds, abstract 52a) followed 205 HIV couples from the Rakai district Uganda, among whom 20 included an HIV-infected partner who started ART. There were 34 transmissions (rate 8.6 transmissions per 100 person-years) among those couples in which the HIV-infected partner was not on ART versus no transmissions in the 20 couples on treatment. The second study (Sullivan 52bLB) followed a much larger cohort - 2993 couples from Rwanda and Zambia. All incident cases of HIV were matched by sequencing, and only HIV transmission events that occurred from the HIV+ partner were analyzed. HIV incidence was 3.4 per 100 person-years for couples not on ART, versus only 0.7 cases per 100 person-years among couples taking ART (relative risk 0.21, 95% confidence interval 0.08-0.59, p<0.001). The ultimate impact of ART on HIV spread is still a topic of debate. In a symposium on the global epidemic, Dr. Christopher Fraser, from Imperial College in London, reviewed the epidemiologic evidence supporting use of wide-spread ART to reduce HIV transmission (abstract 14). While the mathematical modeling data are encouraging that treatment may have a substantive impact on HIV prevalence, the assumptions underlying some of the modeling work are notable - high coverage and efficacy and low behavior change among those taking ART (also called "risk disinhibition") are required. Indeed, two additional abstracts presented this year (Sheth, abstract 50 and Marcelin, abstract 51) described occasional detection of HIV RNA in semen of men on ART, potentially suggesting that suppression of viral replication in plasma may not be sufficient to completely eliminate transmission risk. Empiric evidence of the effect of ART on HIV transmission continues to be required, including data from the ongoing HIV Prevention Trials Network (HPTN) 052 study, an ongoing clinical trial assessing the effect of early versus standard-of-care initiation of antiretroviral treatment on HIV transmission within discordant couples in which the HIV infected partner has a CD4 count between 350 and 550 cells/μL.
Observational epidemiologic data relevant to HIV prevention
Interesting epidemiologic observations were presented about men who have sex with men (MSM), an under-recognized risk group, from three African countries: Malawi, Namibia and Botswana (Beyrer, abstract 172). The findings of this study suggest that there are concentrated HIV epidemics within populations of African MSM, with high HIV prevalence (17% overall) and high frequency of risk behaviors including concurrent partnerships and bisexual partnerships. A prospective observational cohort study among HIV-infected women found that hormonal contraceptive use was not associated with more rapid HIV disease progression (Stringer, abstract 175). Observational data associating hormonal contraception with HIV transmission risk and with HIV disease progression have not been consistent across studies, and continued well-designed, large analyses remain needed. The results of this study are important and reassuring, as effective family planning is essential for women with and at-risk for HIV.
A large series of studies over the past decade continue to define the optimal antiretroviral strategy for prevention of mother-to-child transmission of HIV, particularly in resource-constrained settings. In a plenary presentation (abstract 127), Dr. Jeff Stringer, from the University of Alabama at Birmingham and the Centre for Infectious Disease Research in Zambia, discussed the challenges of prevention of transmission of HIV via breastfeeding. In many high-prevalence areas, PMTCT coverage remains less than optimal, and strategies to reduce breast milk HIV transmission are particularly needed. Current work is focusing on two key strategies: maternal ART for women requiring treatment for their own health (e.g., because of low CD4 counts) and extended infant antiretroviral prophylaxis during a portion of breastfeeding for children born to women who do not yet require ART.
The benefits of the latter were recently reported in a clinical trial from Malawi, that demonstrated 14 weeks of infant nevirapine or nevirapine plus zidovudine reduced postnatal HIV transmission from 10.6% to 5.2-6.4% among infants followed for 9 months (Kumwenda et al. N Engl J Med 2008). A follow-up observational analysis from this same clinical trial population was presented at this year's CROI (Taha, abstract 92) and detailed the benefits of maternal ART for further reducing HIV transmission through breast milk. Among women in the trial, 14% initiated ART during the postnatal period, beginning in 2006 when wide-spread ART became available in Malawi. The rate of HIV transmission after 14 weeks was 3.7% among infants of women ineligible for ART versus 2.1% for infants of ART treated women and 10.5% for infants of women who were eligible but did not receive ART. Thus, ART substantially reduced HIV transmission through breast milk when provided to women eligible to receive ART for their own health. A lingering question from this work is what degree of benefit ART might have in preventing post-natal HIV transmission when provided to women who do not otherwise qualify for ART according to national guidelines. One ongoing study (the Breastfeeding, Antiretrovirals, and Nutrition [BAN] Study) is evaluating the effect of maternal ART versus infant ART prophylaxis during the breastfeeding period, which will be informative for strategies to further reduce MTCT transmission during the post-natal period in areas where ART resources are limited.
HIV Testing in High-Prevalence Settings
A key component in implementing sexual and perinatal HIV prevention strategies, including microbicides, PrEP, ART, and PMTCT, as discussed above, is knowledge of one's HIV serostatus. Relevant to this, in a late breaker presentation (Mohammed, abstract 137LB) reported data from the recently-reported Kenya AIDS Indicator Survey (KAIS), a population-based serosurvey of HIV-1 and STI prevalence in the East African country. In total, 9691 households were included, including 15,853 individuals. HIV prevalence was 7.1% (8.4% for women and 5.4% for men), corresponding to an estimated 1.3 million infections nationwide among those aged 15-64. Twenty-six percent of men and 45% of women had ever been tested for HIV, reflecting a nearly decade-long program promoting HIV testing in this country. Still, 56% of HIV infected persons had never been tested and only 16% knew they were positive. Knowledge of serostatus was directly related to being on cotrimoxazole prophylaxis or ART. An estimated 217,000 individuals in Kenya are thought to be in need of ART (based on a CD4 cut-off for initiation of 250 cells/μL, the current Kenya national guideline). Nearly half (44%) of HIV infected persons had an HIV uninfected spouse, which is consistent with recent data from several African countries suggesting high risk in stable partnerships, suggesting there are an estimated 350,000 HIV discordant couples (6% of all couples in the country) currently in Kenya. Thus, lack of knowledge of HIV status is the strongest barrier to prevention, care, and treatment in Kenya (and likely in most high prevalence countries) and continued access to testing must be a continued priority.
Two abstracts from Uganda addressed the potential for home-based HIV testing to bridge some of this gap. The first (Gupta, abstract 138) offered home versus clinic-based HIV counseling and testing (HCT) to family members of HIV infected persons enrolled in a randomized trial of home versus clinic-based delivery of ART in Jinja, Uganda. Home-based HCT had significantly higher uptake than clinic-based HCT (56% vs. 11% of household members), with 89% acceptance of HCT among family members found by home-based testing teams. Half of stable sexual partners of HIV infected persons were HIV seronegative. Thus, home-based testing was estimated to detect twice as many HIV infected persons and three times the number of HIV discordant couples as facility-based HCT. The second abstract (Nuwaha, abstract 139) presented follow-up data from a district-wide home-based HCT program from southwestern Uganda presented at last year's CROI (abstract 129LB). That program achieved 63% district-wide coverage, testing 205,979 persons with 93% acceptance (i.e., of those at home when the testing team arrived) and documenting 4.9% prevalence, and identifying 860 discordant couples. This year's abstract described pre- and post-program surveys about testing, disclosure, stigma, and risk behavior, administered to a sample of the population. After the home-based testing program, the prevalence of those ever having been tested for HIV rose from 20% to 63% (p<0.01), with significant increases in HIV disclosure (including to a spouse - from 50% to 64%), decreases in measures of HIV stigma, and increase in self-reported condom use for HIV-infected men (from 6% to 55%).
Many questions remain about the role for home-based HCT in high prevalence settings, including frequency of testing and implementation in more concentrated urban areas and in settings with higher stigma. Nonetheless, the work presented here reinforces the place of HIV testing as central to HIV prevention and linkages of HIV positive persons to care and treatment.