HIV Prevention at CROI 2012
19th Conference on Retroviruses and Opportunistic Infections
March 5 - 8, 2012
Jared Baeten, MD PhD
Connie Celum, MD MPH
University of Washington
"Retention in the Partners PrEP Study was ≥95%......FTC/TDF, efficacy was 66% in women and 84% in men.....Adherence appears to be a critical factor for PrEP protection against HIV. Tenofovir levels in blood (plasma, cells) and tissue (biopsies of vaginal and rectal mucosa) have been measured in various studies to understand PrEP's protective effects and to measure adherence.....90% HIV protection efficacy was associated with a tenofovir diphosphate concentration of 16 fmol/million cells, consistent with less than daily......dosing. The model predicted that 76% efficacy was possible with twice-weekly dosing and >95% with 4-7 doses/week. These valuable data provide a framework for evaluation of intermittent dosing of PrEP and understanding how less-than-perfect adherence may still provide substantial HIV protection"
CROI: Tenofovir PrEP Reduces HIV Transmission Risk By Up To 99%: Adherence Critical to Protection From HIV in Partners PrEP Trial of TDF and TDF/FTC - written by Mark Mascolini - (03/7/11)
Over the past several years, it has become apparent that CROI is now a conference focused as much on prevention science as it is on pathogenesis and treatment. HIV prevention occupied a central place at CROI 2012, with important work presented relevant to populations both in the US and worldwide. As in previous years, many oral sessions were recorded and are available online (www.retroconference.org). We have cited specific links here for some of these - these contain both slides and video, but links with slides with audio only are available on the CROI website.
Antiretrovirals for HIV prevention: ART and PrEP
The hottest topic in HIV prevention research is the use of antiretrovirals for prevention: as antiretroviral treatment (ART) of HIV infected persons and as pre-exposure prophylaxis (PrEP) used by HIV uninfected persons. An excellent plenary on the opening morning by Wafaa El-Sadr sought to explore the incredible potential and many challenges of expanding treatment as an HIV prevention strategy (abstract 18)
A major focus of this CROI was sorting out the findings of different randomized clinical trials of PrEP. In the past few years, a number of PrEP trials have released results. CAPRISA 004 (reported at IAS 2010) tested peri-coitally applied 1% tenofovir vaginal gel among 889 South African women and found that tenofovir gel used in this fashion decreased HIV risk by 39% (95% CI 6-60%, p=0.017) (Abdool Karim et al., Science 2010). In late 2010, iPrEx, which evaluated daily oral emtricitabine/tenofovir (FTC/TDF) among 2499 men who have sex with men and transgender women from six countries, report that oral PrEP reduced HIV risk by 44% (95% CI 15-63%, p=0.005) (Grant et al New England Journal of Medicine 2010). In both trials, protection against HIV was substantially higher among those with high adherence.
In 2011, four PrEP trials released results. The Partners PrEP Study, which enrolled 4758 HIV serodiscordant couples from Uganda and Kenya and evaluated oral TDF and oral FTC/TDF as PrEP, and the TDF2 study, which tested oral FTC/TDF in a study of 1200 young heterosexuals from Botswana, both reported at IAS 2011 that PrEP significantly reduced HIV incidence in their populations. However, two other trials - FEM-PrEP and VOICE, both conducted among African women - had study arms stopped for futility (i.e., no HIV protection) in 2011: in the case of FEM-PrEP, oral FTC/TDF and in the case of VOICE, both oral TDF and daily vaginal tenofovir gel did not prevent HIV acquisition in those studies. The VOICE study is ongoing, with a study arm testing oral FTC/TDF, and full details of its findings will be available in early 2013.
With this as background, an exciting series of abstracts on PrEP for HIV prevention started out the first oral abstract session of CROI 2012 (http://app2.capitalreach.com/esp1204/servlet/tc?c=10164&cn=retro&s=20481&&dp=player.jsp&e=16609&mediaType=podiumVideo). The session included updated, final results from the Partners PrEP Study, analyses of blood tenofovir levels and their relationship to HIV protection from Partners PrEP and iPrEx, the first public presentation of the FEM-PrEP results, and new data from CAPRISA 004 and studies of new PrEP strategies.
Partners PrEP Study (Baeten, abstract 29). The Partners PrEP Study is a phase III, randomized, double-blind, placebo-controlled, three-arm trial of daily oral TDF and FTC/TDF PrEP for the prevention of HIV acquisition by HIV seronegative partner in heterosexual HIV serodiscordant partnerships (disclosure: we are the lead investigators of this study). In July 2011, the independent DSMB for the Partners PrEP Study recommended that the results of the study be reported, nearly 2 years earlier than expected, and the placebo arm discontinued, because of clear demonstration of HIV protection due to PrEP. These surprising results were reported in preliminary form at IAS 2011. At CROI 2012, finalized results through July 2011 follow-up in the study (i.e., when the placebo arm was stopped) were reported.
The Partners PrEP Study enrolled 4758 HIV serodiscordant couples, in which the HIV infected partner was not medically eligible for ART under the national guidelines of their country of residence at the time of study enrollment. HIV uninfected partners were randomized in an equal fashion to receive daily oral TDF, FTC/TDF, or placebo. Couples were followed for up to 36 months, for primary endpoints of HIV acquisition in the initially-seronegative partner and safety. The study was conducted at 9 clinical trial sites in rural and urban Kenya and Uganda. Most couples were married. For HIV infected partners, the average median CD4 count at enrollment was approximately 500 cells/μL. HIV uninfected participants were seen monthly, for HIV and pregnancy testing, safety monitoring, and individualized adherence counseling. HIV infected participants had quarterly visits for monitoring of HIV clinical status; during the study ~20% initiated ART. All participants received a comprehensive package of HIV prevention services, including individual and couples risk reduction counseling, condoms and condom counseling, and counseling regarding other HIV prevention services (e.g., male circumcision).
Retention in the Partners PrEP Study was ≥95% throughout the study follow-up period. Adherence to PrEP appeared to be high: 97% of dispensed pills were estimated to be taken, based on monthly pill counts of returned, unused study medication. In a poster from the Partners PrEP Study presented at CROI 2012, confirmation of high adherence was shown in a substudy that used objective measurement, including unannounced home visits for pill counts and MEMS caps (Bangsberg, abstract 1067).
In total, 82 HIV infections were observed: 17 among those assigned TDF, 13 among those randomized to FTC/TDF, and 52 among those receiving placebo, for an efficacy for HIV protection of 67% (95% CI 44-81%, p<0.0001) for TDF and 75% (95% CI 55-87%, p<0.0001) for FTC/TDF. The HIV protection effects of TDF versus FTC/TDF were not statistically different (p=0.23). Both TDF and FTC/TDF reduced HIV risk in both men and women, and the effects in women and men were statistically similar. For TDF, efficacy among women was 71% and was 63% among men; for FTC/TDF, efficacy was 66% in women and 84% in men - all were statistically significant. Additional subgroup analyses showed consistent HIV protection in groups defined by age, sexual behavior at baseline, country (Kenya or Uganda), and HIV disease markers of the HIV infected partner (CD4 count, plasma HIV viral load).
None of the participants who acquired HIV after enrollment were found to have virus that was resistant to TDF or FTC. There were 14 additional participants (5 in the TDF arm, 3 FTC/TDF, 6 placebo) who were found to have seronegative acute HIV infection at the time of study entry - thus, would could not have been protected by PrEP because they were already infected at the time of enrollment. Of these, two (one assigned TDF and one assigned FTC/TDF) developed virus that was resistant to the study medications: one a K65R mutation (resistance to TDF) and one a M184V mutation (resistance to FTC).
There were no statistically significant differences in the rate of serious adverse events and important laboratory adverse events between those receiving active PrEP and those receiving placebo. PrEP was also well tolerated - slightly more participants reported nausea during the first month of PrEP than those receiving placebo, that faded over time, but gastrointestinal symptoms like these occurred in <10% of participants overall. The study found no evidence of risk compensation - in fact, unprotected sex declined during follow-up, similarly across the three study arms. Additional safety information was presented in a poster that showed no statistically significant differences for PrEP versus placebo in terms of pregnancy incidence and outcomes (Mugo, abstract 1060).
In summary, the study demonstrated definitively reduced risk of HIV acquisition for African men and women taking TDF and FTC/TDF PrEP. The Partners PrEP Study is continuing: the DSMB recommended continued follow-up of the active PrEP arms and assignment of the placebo arm participants to active PrEP, to gather additional comparative information on the safety, efficacy, and tolerability of TDF and FTC/TDF as PrEP. The part of the study will finish at the end of 2012. Future work will target how to deliver PrEP to highest-risk couples to maximize benefits (Kahle, abstract 1102).
Tenofovir levels and HIV protection - Partners PrEP and iPrEx. Adherence appears to be a critical factor for PrEP protection against HIV. Tenofovir levels in blood (plasma, cells) and tissue (biopsies of vaginal and rectal mucosa) have been measured in various studies to understand PrEP's protective effects and to measure adherence. In the Partners PrEP Study, tenofovir plasma levels were compared for the 30 individuals who acquired HIV who had been assigned to active PrEP (17 TDF and 13 FTC/TDF) to a randomly-selected subset of 200 individuals from the active PrEP arms who did not acquire HIV (Donnell, abstract 30). A case-cohort design was used, which takes into account tenofovir measurement in multiple samples from participants over the course of study follow-up - in total, >1000 samples were tested. Among non-seroconverters, PrEP was found in 82% of samples - confirming high adherence in the study. In contrast, only 31% of seroconverters had tenofovir detected at their seroconversion visit, and only in 56% of samples prior to seroconversion. Taken together, having tenofovir detected was associated with an 86% relative risk reduction of HIV for TDF (95% CI 57-95, p<0.001) and 90% for FTC/TDF (95% CI 56-98%, p=0.002). Quantification of tenofovir levels in these samples showed that, when tenofovir was detected, it was present at high levels, consistent with daily dosing. Levels done over the course of study follow-up showed that most non-seroconverters took the medication consistently, with a minority who were consistently not taking PrEP. Together, these results are consistent with and reinforce the high levels of HIV protection demonstrated in the Partners PrEP Study.
In iPrEx, previous work has demonstrated tenofovir detection in 9% of seroconverters and 51% of non-seroconverters (Grant et al New England Journal of Medicine 2010) - consistent with the lower efficacy in that study compared to Partners PrEP but also consistent in that detection of tenofovir was associated with ~90% HIV protection. At CROI 2012, quantitative tenofovir diphosphate levels from peripheral blood mononuclear cells from seroconverting cases and non-seroconverting controls were assessed (Anderson, abstract 31LB).
In iPrEx, seroconverters had tenofovir detected more commonly prior to seroconversion compared to at the seroconversion visit, suggesting that HIV infection in the active arm of iPrEx occurred during periods of non-adherence. Levels were compared to those observed in a separate pharmacokinetics study (STRAND, abstract 995, CROI 2011) in which TDF was provided 2, 4, or 7 times per week by directly observed therapy. When compared to STRAND, iPrEx participants overall appeared to be using PrEP less than twice per week, with only ~18% of non-seroconverters (and no seroconverters) having dosing consistent with daily dosing. In a regression analysis of the full iPrEx cohort, HIV incidence in the active arm with undetectable drug was not different than placebo (suggesting the model was not contaminated by confounding) and that 90% HIV protection efficacy was associated with a tenofovir diphosphate concentration of 16 fmol/million cells, consistent with less than daily dosing. The model predicted that 76% efficacy was possible with twice-weekly dosing and >95% with 4-7 doses/week. These valuable data provide a framework for evaluation of intermittent dosing of PrEP and understanding how less-than-perfect adherence may still provide substantial HIV protection. Additional information relevant to these data were presented in a poster (Anderson, abstract 587).
FEM-PrEP primary results (Van Damme, abstract 32LB). FEM-PrEP was a randomized clinical trial of daily oral FTC/TDF or placebo, conducted among 2120 women from Kenya, South Africa, and Tanzania. In April 2011, the study DSMB recommended stopping of the trial because of no difference in HIV incident infections (at that time, there were 28 per arm). The majority of women were <25 years of age, ~13% reported exchange of sex for gifts or money, and ~15% had cervical chlamydial infection at baseline; notably, 70% at baseline felt themselves at little or no risk of acquiring HIV. A total of 73 HIV infections occurred in the study, of which 68 occurred after randomization and prior to product discontinuation and were included in the primary analysis: 33 in the FTC/TDF arm and 35 in the placebo arm (incidence 5 per 100 woman-years), translating to only a non-significant 6% reduction in HIV risk (p=0.8). There was no difference in creatinine or phosphorus adverse events between the study arms. Pregnancy incidence was high (~10% per year) but did not differ significantly by arm. No seroconverters developed TDF resistance but 5 were found to have resistance to FTC (M184V or M184I) - 2 of which were consistent with primary resistance, 2 with secondary (transmitted) resistance, and 1 of which occurred in the placebo arm; for all, resistance waned over time. Adherence by self-report and pill counts were high, but plasma drug levels revealed that only 15-26% of samples from seroconverters had tenofovir detected and only 26-38% of non-seroconverting controls. Thus, pill taking was very low in FEM-PrEP - too low to have sufficient statistical power to show whether PrEP was effective for HIV protection. Thus, the summary from the investigators was that PrEP adherence was too low to assess PrEP's HIV protection in FEM-PrEP.
Sensitive resistance testing in CAPRISA 004 (Wei, presented by Johnson, abstract 33). CAPRISA 004 reported results of cervicovaginal swab testing for virus potentially resistant to tenofovir among seroconverters from that study. A high-sensitivity method was employed, using a real-time assay for K65R and K70E validated against subtype C, with screening cut-offs of 0.3% of viruses (2 log10 lower than conventional bulk sequencing). 33 of 38 seroconverters had swabs available from whom 21 had virus able to be amplified. Only 1 individual had resistant virus detected - K65R mutation only, detected in only ~1% of viruses in the sample, which is consistent with background mutation rates and not necessarily selected by the product. This is in spite of tenofovir detected in samples from 26% of women (indicating drug pressure). Other work presented at CROI 2012 reinforced that no women in CAPRISA 004 had low-level resistance detected in blood plasma (Fischer, abstract 1063), including the one seroconverter who had very low cervicovaginal K65R detected.
Rectal tenofovir gel (McGowan, abstract 34LB). A phase I randomized, double-blind, placebo-controlled trial of tenofovir 1% gel for rectal use was reported (MTN 007). This study was designed to study the safety and acceptability of a reduced glycerin formulation (836 mOsmol/kg compared to 3111 with original tenofovir gel tested in the vaginal studies, like CAPRISA 004 and VOICE). This study enrolled 65 healthy, persons from the US (69% men, 31% women) with a history of receptive anal sex who were willing to be abstinent during follow-up. Participants were randomized to one of four study arms: 1% reduced glycerin tenofovir gel, placebo gel, no treatment, or 2% nonoxynol-9 (a "positive control" known to cause some epithelial disruption). Incident adverse events were mostly mild/moderate with no difference by arm but a trend to more adverse events, as expected, in the nonoxynol-9 arm. The reduced glycerin formulation of rectal tenofovir gel was well tolerated and acceptable. No significant harmful changes in a battery of intensive laboratory measures were observed for the tenofovir gel and the formulation will next move to phase II evaluation.
Rilpivirine long-acting (Jackson, abstract 35). Rilpivirine is a potent NNRTI being formulated as a long-acting nanosuspension for injectable, depot use as PrEP. Three doses were tested in this phase I study: 300, 600, and 1200 mg (20 women per dose, plus 6 men receiving the 600 mg dose). Blood, biopsy, and genital fluid were sampled, to 28 days. Prolonged plasma and genital tract exposure was achieved, with higher concentrations in female genital fluid than in plasma. These data support ongoing work to further evaluate safety, pharmacokinetics, and pharmacodynamics of this potentially powerful PrEP strategy.
Other studies of PrEP. A themed discussion of animal and human studies of PrEP focused on emerging opportunities (http://app2.capitalreach.com/esp1204/servlet/tc?c=10164&cn=retro&s=20481&&dp=player.jsp&e=16633&mediaType=podiumVideo). In a pharmacokinetics study in macaques (Radzio, abstract 1086), daily FTC/TDF achieved 10-fold lower concentrations of tenofovir in vaginal tissues compared to rectal tissues, consistent with what has been reported in biopsy studies in humans, and potentially explaining how vaginal transmission may be sensitive to PrEP non-adherence. Importantly, in macaque studies, where adherence is not a challenge, intermittent oral FTC/TDF provided 100% protection against SHIV vaginal challenge.
In a second macaque study (Kersh, abstract 1087), anti-SHIV T cell responses were explored in animals who had received PrEP. T cell responses were found in the majority of exposed, uninfected animals after receiving PrEP, suggesting that PrEP, by preventing infection in the face of viral challenge, may permit development of virus-specific T cell responses, in effect a "chemo-vaccination."
Intravaginal rings offer a strategy to deliver antiretroviral PrEP medications with lesser opportunities for adherence challenges than for daily oral pills or vaginal gels. A phase II assessment of dapivirine vaginal ring among African women, with use for 12 weeks, found the ring was well-tolerated, without safety concerns (Nel, abstract 1089). 3 seroconversions occurred, all in the placebo group (though this is too few to assess efficacy). Acceptability was high. Dapivirine vaginal ring is moving to phase III efficacy studies in 2012.
ART for prevention: community effects. In a session focused on HIV epidemiology (http://app2.capitalreach.com/esp1204/servlet/tc?c=10164&cn=retro&s=20481&&dp=player.jsp&e=16641&mediaType=podiumVideo), two abstracts reported on population-level studies to estimate HIV incidence reductions accompanying ART roll-out.
The first, from the Africa Center, described HIV incidence during ART roll-out since 2004 in rural KwaZulu-Natal, South Africa, where a series of annual population-based assessments of HIV incidence and care have been done (Tanser, abstract 136LB). In the area, community based-clinics have used nurses and counselors to deliver ART, using standard South African regimens, to >20,000 patients since 2004. Using 8 year of annual surveillance data and follow-up of 16,667 HIV repeat testers for HIV incidence, the authors tested the hypothesis that increasing coverage of ART among infected individuals would be accompanied by decreasing HIV incidence - in other words, a community-level test of effectiveness. Intensive geo-surveillance information is available for the Africa Center communities from 2004-2011, permitting detailed analyses linking the proportion of total HIV-infected persons on ART in a community (i.e., ART coverage) to HIV seroconversion risk for uninfected persons in those same communities. Crude HIV incidence was 2.64/100 person-years. By the end of 2011, an estimated 60% of HIV-infected persons met the new South African ART guidelines of initiating at CD4<350 cells/μL. Overall, the results demonstrated a 38% reduction in HIV seroconversion risk, that was statistically significant, for ART coverage in a community >30%; adjusted analysis showed a clear linear relationship between higher ART coverage and lower HIV acquisition risk. These results provide clear empirical evidence for a reduction in risk of acquiring HIV in communities with higher ART coverage of HIV infected people in a typical rural South African setting. An encouraging finding was that a community level impact on HIV incidence was observed in the context of 'real world' setting facing ART retention and adherence challenges, suggesting that HIV incidence is being impacted in the context of programmatic roll-out of ART in South Africa.
The second abstract, from the Rakai Community Cohort, assessed trends in ART, medical male circumcision, and voluntary counseling and testing in that intensively-studied community (Wawer, abstract 141). As these key services have rolled out, HIV incidence has declined -
6-fold in fact during the most recent periods (10-fold for men and 2-fold for women)-overall an encouraging finding but highlighting the need for additional prevention interventions for women. Service utilization has been rising but still sub-optimal, as indicated by an estimated 25-30% of HIV-uninfected non-Muslim adult men having been circumcised, emphasizing the need for demand generation for effective interventions.
Another abstract (Jain, abstract 143) described a study measuring community viral load during a rapid-testing campaign of >4300 persons in Mbarara district of southwestern Uganda in May 2011. The fingerprick viral load was highly correlated with the phlebotomy-obtained viral load (r=0.99) which is encouraging for the use of field projects measuring viral load without phlebotomy. Median CD4 was 400 and mean community viral was ~64,000 copies/mL with 37% having an undetectable viral load. Among those self-reporting being on ART, 83% had undetectable viral load, suggesting that ART implementation, even in this rural setting, was being done with high adherence.
HIV testing and linkages to care. HIV testing is critical to implementation of ART for prevention. A themed discussion of HIV testing and linkages to care focused on HIV testing as the first step in the cascade of linkages to HIV care, ART, and prevention interventions (http://app2.capitalreach.com/esp1204/servlet/tc?c=10164&cn=retro&s=20481&&dp=player.jsp&e=16630&mediaType=podiumVideo). One presentation focused on novel HIV testing strategies, and assessed the acceptability and ease of use of home self-testing for HIV among MSM (Katz, abstract 1131); 178 home tests were provided to 68 Seattle men, who reported that they were willing to use oral rapid antibody tests on oral fluids at home and found it easy to use. Another (Haukoos, abstract 1132) evaluated targeted HIV screening in an academic Denver ER, and found that ~20% of the ER patients identified as high risk agreed to HIV testing compared to 12% of the non-targeted ER attendees with a 4.7-fold greater detection of newly-identified HIV infected in the targeted testing group. Another ER study (McKellar, abstract 1133) evaluated routine opt-out HIV testing in 2 North Carolina ERs from 2008-2010 and found HIV testing acceptance rates of 64-69%, and high linkage to care although the number of newly identified HIV infected persons was small (7).
In rural Uganda, 5 day health campaign which included PIMA point of care CD4 testing for HIV-infected persons was reported (Chamie, abstract 1134). 140 were HIV-infected with routine referrals made for 132, of whom 57% were linked to care within 3 months, and enhanced referrals for 7 with advanced HIV, of whom 6 were linked to care. A new HIV diagnosis was associated with reduced odds of linkage and among those with new HIV diagnoses, those with higher CD4 count were less likely to link to care. The investigators concluded that routine referral to care is not sufficient to get high rates of linkage to HIV care.
Finally, results of a pilot of home-based HIV testing (HBCT) in rural South Africa were presented (van Rooyen, abstract 1135), which achieved 90% uptake of HIV testing and an HIV prevalence of 30%. HBCT identified HIV-infected persons with a median CD4 of 435, 1/3 were eligible for ART by the new South African guidelines, and linkage to care for all HIV-infected persons by 3 months was 94%. A related poster (Barnabas, abstract 687) reported on the performance of the PIMA point-of-care CD4 performed in the home in the KZN HBCT pilot compared to the gold standard of CD4 by FacsCount on a venous sample obtained by phlebotomy. The median CD4 was comparable (435 and 429, respectively) with a mean difference of 12 cells/μL, indicating high performance of the PIMA point-of-care CD4 in field settings, which facilitates linkages to HIV care at the time of HIV diagnosis.
Finally, a Tuesday afternoon symposium explored next steps in antiretrovirals for HIV prevention, including talks explaining the divergent results of PrEP trials (Baeten, abstract 67), exploring the hazards and potential of intermittent PrEP (Buchbinder, abstract 68), thinking forward to next PrEP strategies (Romano, abstract 69), and experience from the field on implementing ART for prevention (Asiimwe, abstract 70) (http://app2.capitalreach.com/esp1204/servlet/tc?c=10164&cn=retro&s=20481&&dp=player.jsp&e=16620&mediaType=podiumVideo).
Two abstracts, presented during the oral session on PrEP noted above
(http://app2.capitalreach.com/esp1204/servlet/tc?c=10164&cn=retro&s=20481&&dp=player.jsp&e=16609&mediaType=podiumVideo), assessed community-level impact of medical male circumcision, one of the resounding successes in HIV prevention. The first (Gray, abstract 36) evaluated HIV incidence in the Rakai Community Cohort Study (14,000 persons in 50 communities in rural Uganda) over time, as male circumcision was rolled out. Time periods between 1999 and 2003 (baseline/counterfactual) were compared to the period when Rakai conducted a clinical trial of male circumcision (2004-2006) and the period after the trial, when medical male circumcision services were rolled out in the community (2007-2011). Compared to the pre-medical male circumcision era, HIV incidence in non-Muslim men fell significantly once medical male circumcision rolled out (from 1.39 to 1.00% per year, p=0.002, a 27% reduction in adjusted analysis). Thus, in summary, roll-out of male circumcision was associated with a significant, population-level reduction in HIV incidence in men.
The second abstract, from Orange Farm in South Africa (where another of the original trials of circumcision for HIV prevention was conducted), assessed the relationship between medical male circumcision and prevalence of herpes simplex virus type 2 (HSV-2) (Auvert, abstract 37). In Orange Farm, voluntary medical male circumcision being rolled out to all adult men >15 years of age with community mobilization and outreach. In 2010-11, a cross-sectional survey among a random sample of 3356 men with questionnaire, genital exam to assess circumcision status, and HIV and HSV-2 serology. Circumcision prevalence was 53% (59% among younger men), compared with 11% before the trial was done there. HSV-2 prevalence was 16.5% in circumcised men and 31% in uncircumcised men, a 23% weighted prevalence reduction associated with circumcision, an effect seen for all age groups. Of note, circumcised men in this sample had half the prevalence of HIV (i.e., circumcision had a stronger relationship to HIV than to HSV-2). Thus, in summary roll-out of medical male circumcision in South Africa appears to have been accompanied by a population-level reduction in HSV-2 prevalence in men, which is an encouraging finding given the limited number of proven effective HSV-2 intervention strategies.
Epidemiology and HIV risk factors
A session on HIV epidemiology provided new insights into the epidemic in various settings
Seroadaptive behaviors and HIV risk in gay men. Seroadaptive sexual behavior - i.e., modifying sexual practices based on perceived HIV serostatus of sexual partners (serosorting - partner choice) or differences in risk of transmission by different sexual acts (seropositioning - sexual act choice) - has been debated for whether it provides some degree of HIV protection. . Data from 4 large cohorts of North American MSM (N=12,705, 663 seroconversions) were combined for an analysis of seroadaptive behaviors and HIV risk (Vallabhaneni, abstract 140). 12,705 MSM from North America. Sexual behavior was recorded twice-yearly. For 32% of the six-monthly intervals, unprotected anal intercourse with some seroadaptive practice was reported, and 62% of men reported some practice at some point in follow-up. The seroconversion rate was lowest for those with a single negative partner (HIV seroconversion occurring 0.25% of these visits), highest was unprotected anal intercourse with no seroadaptive practices (2.95%). Compared to no seroadaptive practices, serosorting (HR 0.62) and seropositioning (HR 0.26), being an exclusive top (0.17), and having a single negative partner (0.17) were all at lower risk. However, looking at this another way, compared to those reporting no unprotected anal intercourse, serosorting was still twice as risky (HR 2.03). Thus, serosorting is twice as risky as no unprotected anal intercourse but better than not seroadapting at all. Having a single negative partner was strongly protective. Together, these data reinforce condom use, limiting partner number, and, for men reporting unprotected sex, seroadaptation as a possible risk-reduction strategy.
Swaziland national prevalence survey (Justman 142). Swaziland has highest national prevalence in the world at 31%. In recent years, the government of Swaziland has scaled up a national combination prevention strategy, including a recent national campaign to massively scale up male circumcision. This study, called SHIMS (Swaziland HIV Incidence Measurement Survey) was a two-stage cluster sampling design study done in 2010 to assess national seroprevalence and establish a nationally representative seroincidence cohort to help, in future years, assess the impact of roll-out of HIV prevention. For the study, home HIV counseling and testing were done, on a sample weighted to achieve population representativeness and nonresponse. 14,950 households selected, 12,597 participated (84%),with 24,300 eligible members, of whom 16% had no contact and 10% refused (=a 74% participation rate, 82% for women and 65% for men). In total, 15% of men were circumcised. HIV prevalence was 23% for men and 38% for women, 31% total; 68% of HIV-infected women were aware of their status and 52% of men, and half of individuals who were aware were on ART. Men had lower prevalence below age 25, as has been seen elsewhere. Women peak prevalence was between ages 30 and 34 (54%), and men between ages 35 and 39 (48%). Notably, while the overall pattern was similar to a 2006 DHS study in the country, there was a lower prevalence in the 2010 survey for younger persons suggesting potentially falling incidence for younger persons. These nationally-representative data are incredibly valuable.
HIV risk in US women. The ISIS study enrolled 2099 women (88% black, 8% white, 12% Hispanic) at risk for HIV from 6 US metropolitan areas (Atlanta, Baltimore, New York, Newark, Raleigh-Durham, Washington DC) (Hodder, abstract 1048). 32 women (1.5%) were newly diagnosed as HIV infected at enrollment, and HIV incidence during the study was 0.24% per year, considerably higher than anticipated based on previous estimates of HIV incidence in the US. CDC recently announced a campaign to encourage testing in African-American women (http://hivtest.org/takecharge/).
Hormonal contraception and HIV risk. The relationship between hormonal contraceptive use and risk of HIV acquisition has been a topic of considerable interest and debate in the past year, with some studies suggesting increased risk (particularly for injectable contraceptives like depot medroxyprogesterone acetate [DMPA]) but others not finding increased risk. Contraceptives are critical to the health of women and families worldwide. The World Health Organization recently held a Technical Consultation to examine this issue and recommended that women using injectable contraceptives who are at high risk of HIV be especially counseled to use dual protection with condoms. Analyses of data from studies conducted for other primary purposes have been done; at CROI 2012, an analysis of such a study (the MIRA study, a randomized trial which failed to demonstrate that the diaphragm was protective against HIV prevention) was reported (McCoy, abstract 20LB http://app2.capitalreach.com/esp1204/servlet/tc?c=10164&cn=retro&s=20481&&dp=player.jsp&e=16608&mediaType=podiumVideo). In total, 4948 women participated in MIRA, from South Africa and Zimbabwe, the average age was 29. Contraceptive use was measured every 3 months, as was HIV status and sexual behavior. At baseline, 21% used combined oral contraceptives, 14% used progestin-only pills, and 26% used injectable contraception. A total of 274 HIV seroconversions were observed (incidence 4% per year). In multiple analyses, oral contraceptive pills (both combined and progestin-only) were not associated with increased HIV risk. Injectable contraception was associated with mildly elevated risk (adjusted HR 1.37, 95% CI 1.01-1.86, p=0.04) in the primary study model; however, the effect was somewhat attenuated depending on the analysis technique. Thus, in summary, the results of this study suggest possible increased risk of HIV acquisition with injectable contraception. More study of this critically important issue is a public health priority.