icon-    folder.gif   Conference Reports for NATAP  
  18th CROI
Conference on Retroviruses
and Opportunistic Infections
Boston, MA
February 27 - March 2, 2011
Back grey_arrow_rt.gif
HIV Prevention at CROI 2011
  Jared Baeten, MD PhD
Connie Celum, MD MPH
University of Washington
Continuing the trend of the past few years, HIV prevention occupied a central place at CROI 2011, with important work presented relevant to populations both in the US and worldwide. Arguably, CROI is now a conference focused as much on prevention science as it is on pathogenesis and treatment. As in previous years, many oral sessions were recorded and are available online (www.retroconference.org/2011/).
Antiretrovirals for HIV prevention: treatment, PEP, 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- and post-exposure prophylaxis (PrEP and PEP). A major focus of this CROI was on use of antiretrovirals as PrEP. Since CROI 2010, two important PrEP trials were completed: CAPRISA 004, which tested peri-coitally applied 1% tenofovir vaginal gel among 889 South African women (Abdool Karim et al., Science 2010), and iPrEx, which evaluated daily oral emtricitabine/tenofovir among 2499 men who have sex with men and transgender women from six countries (Grant et al New England Journal of Medicine 2010). In CAPRISA 004, tenofovir gel reduced HIV acquisition risk by 39% (95% CI 6-60%, p=0.017). In iPrEx, oral PrEP reduced HIV risk by 44% (95% CI 15-63%, p=0.005). In both trials, protection against HIV was substantially higher among those with high adherence: 54% efficacy in CAPRISA 004 among those with ≥80% adherence and 73% efficacy in iPrEx for those with ≥90% adherence, arguing that the success of PrEP as a public health intervention will necessitate motivating high adherence in PrEP users. Ongoing clinical trials are continuing to evaluate oral and vaginal tenofovir-based PrEP in heterosexual and injecting drug user populations, with results expected over the next two years, and new PrEP agents in the pipeline.
An excellent preconference symposium on biostatical methods in trial design - aimed at a broad audience - is worth watching on the web (abstracts 14, 15, 16). As PrEP trials come to completion and new ones are potentially initiated, the veracity, applicability, and impact of their results rests in large part on rigorous and appropriate design. One plenary (Celum, abstract 120) provided a comprehensive overview of the PrEP field.
LINK to [excellent] Plenary on Wednesday: Drugs for Prevention-Topical and Systemic PrEP. Connie Celum,Univ of Washington, Seattle, US http://app2.capitalreach.com/esp1204/servlet/tc?c=10164&cn=retro&s=20445&&dp=player.jsp&e=13736&mediaType=podiumVideo
A number of abstracts reported updated and new data from the iPrEx trial. The primary trial analysis included data through May 1, 2010, but the study cohort continued blinded study medication through August 2010, and participants were then followed for an additional 8 weeks after stopping study drug, during which time a substantial number of additional infections (and thus, additional information on the efficacy of PrEP) accumulated (Grant, abstract 92). In total, 131 infections while on PrEP (83 placebo, 48 active), corresponding to a 42% decrease in HIV risk as a result of PrEP (95% CI 18-60%, p=0.002). The protective effect of PrEP was durable through 144 weeks of study follow-up. During the post-PrEP 8 weeks of follow-up, 2 additional infections were observed in the placebo arm and 4 in the active PrEP arm, a difference that was not statistically significant and that suggests that PrEP does not "mask" HIV infection with delayed seroconversion. In contrast to CAPRISA 004, in which tenofovir gel reduced acquisition of genital herpes by 51%, iPrEx found that oral emtricitabine-tenofovir did not reduce herpes acquisition (Lama, abstract 1002). The gel formulation delivers very high concentrations of the active drug to the genital tract - in laboratory studies, tenofovir only has activity against the genital herpes virus at high concentrations, like those achieved with topical but not oral dosing. One important study outcome was the degree which the iPrEx study resulted in 'collateral benefits' in the trial communities, including increased HIV testing and counseling for an often-marginalized population, advocacy and community building, increased condom use, and research capacity building at the study sites. The iPrEx study has now transitioned to an open label phase, in which all participants will be offered emtricitabine-tenofovir PrEP for 18 months, which will provide new information about safety, use, counseling, and adherence in the context of known efficacy of emtricitabine-tenofovir PrEP. This new phase of research is essential to understand whether and how PrEP can be implemented more widely for HIV prevention in this population.
The protective effect of PrEP in iPrEx was greater with higher adherence: 16% protection for those men who were <50% adherent, 34% protection for 50-90% adherence, and 68% protection for >90% adherence in the updated analysis. Two accompanying abstracts explored adherence using blood levels of tenofovir and emtricitabine. First, the team validated an assay for measurement of these compounds in stored samples (Anderson, abstract 96LB). Then, 179 samples, from a random selection across all adherence strata at the study week 24 visit were tested. The detection rate was 50% for the chemically active form of tenofovir and 62% for emtricitabine. Detection was significantly higher for men from the US vs. non-US sites (97% vs. 50%, p<0.0001) and for older vs. younger men (73% for ≥25 years vs. 44% <25 years, p<0.001). Sample collection practices, height, and kidney function did not explain these differences. The authors interpreted these findings as showing that 40-50% of study participants took PrEP less than the protocol-specified daily dosing. In the second abstract (Amico, abstract 95LB), drug levels were correlated with other measures of adherence during the trial: interview with participants, computer-assisted interview (which had been hypothesized as less likely to result in over-reporting of adherence than face-to-face interview), and pill counts in the study pharmacy. The conclusion was that all measures appeared to substantially over-report adherence when assessed versus blood detection of the study drug. Low reported adherence was uncommon but was predictive of no exposure. Better measurement and understanding of motivators for adherence will be needed to maximize the potential benefits of PrEP for HIV prevention.
There was no increased risk of serious adverse events and laboratory safety abnormalities in iPrEx as a result of PrEP. There was a greater proportion reporting headache (4% vs. 3%), nausea (2% vs. <1%), and weight loss (2% vs. 1%) in the active vs. placebo arm, although these side effects were mild, rare, time-limited, and mostly present in the first few weeks after starting PrEP. Abnormal kidney function measurements were extremely rare - involving only 5 participants taking PrEP - and all resolved when PrEP was stopped.
Loss of bone mineral density (BMD) is a potential concern with tenofovir - in HIV+ persons taking tenofovir for treatment, BMD declines slightly, although any clinical significance of this decline has not been established (Jules: question?). In the iPrEx trial, 503 participants underwent serial bone density scans during follow-up (Mulligan, abstract 94LB). In the first 24 weeks, there was a greater decrease in BMD for active PrEP vs. placebo (p=0.001), corresponding to a <1% difference in BMD. This difference did not worsen during longer follow-up. There was no difference in fracture risk, although the duration of the study was not long enough to reliably assess this outcome. Notably, average bone density in iPrEx participants was somewhat lower than expected for the population of young healthy men who were the study participants. In another abstract on BMD and PrEP use (Liu, abstract 93), 184 men from San Francisco who participated in a phase II study of oral tenofovir PrEP had serial BMD measured. Initiation of tenofovir was associated with small decline in BMD, with a ~1%. As in iPrEx, there was no difference in fractures. Longer studies of BMD in persons on longer-term PrEP will be needed.
In iPrEx, there were no cases of HIV resistance in persons who acquired HIV while on PrEP, but emtricitabine (but not tenofovir) resistance was seen in two persons who had acute HIV infection at the time of study enrollment (and were thus in the "window period" before seroconversion). Additional testing of iPrEx samples with ultra-sensitive resistance testing confirmed these results (Liegler, abstract 97LB). The two cases of emtricitabine resistance faded and became undetectable within 6 months after stopping PrEP.
Several mathematical modeling studies assessed the potential impact of eventual roll-out of PrEP. One modeled HIV antiretroviral resistance in the context of PrEP and ART roll-out (Abbas, abstract 98LB), finding that most resistance in communities would be from treatment failure and nonadherence, not PrEP, and that the biggest HIV prevention benefit could come from concurrent ART and PrEP implementation, rather than either strategy alone (from Jules: I suspect drug resistance could be a concern with non-adherence). Most resistance from PrEP would be from inadvertent PrEP use by persons with established HIV infection, emphasizing importance of HIV testing and strategies to reduce the potential for black market use. A second model (Hallett, abstract 99LB) assessed the relative costs and prevention benefits of antiretroviral-based prevention for HIV serodiscordant couples, including PrEP for the uninfected partner until the time the infected partner starts ART (at a CD4 count of 200 or 350) versus early ART (at CD4 counts of 350 or 500) for the infected partner. PrEP was found to be more cost-effective than early ART, if the cost of PrEP was less than ART and if targeted to couples with highest risk behaviors. Finally, a third abstract modeled PrEP cost-effectiveness in a hypothetical roll-out campaign South Africa (Walensky, abstract 37LB). General PrEP roll out was determined to be potentially "very cost effective" by WHO criteria, even more so among higher risk populations, such as young women with high HIV incidence (as were recruited in the CAPRISA 004 trial), if PrEP were to have higher efficacy than seen in trials to date, or if costs of PrEP were to go down. PrEP would be cost-saving if costs were to decrease substantially, efficacy was higher, and targeting to highest-risk populations could be done. The authors summarized that PrEP could represent excellent comparative value for money.
Over the past several CROI conferences, non-human primate model studies have provided important information about the potential for antiretroviral PrEP and PEP to prevent HIV infection; CROI 2011 was no different. In one study (Cong, abstract 31), emtricitabine-tenofovir PrEP offered complete protection against SHIV rectal challenge, even in the context of a emtricitabine-resistant challenge virus. The results imply that exposure to drug-resistant virus will not necessarily lead to PrEP failure. The mechanism for the finding is not known; emtricitabine-resistant viruses are hypersusceptible in vitro to tenofovir, which might explain the findings. In a second study (Dobard, abstract 30), a different antiretroviral - raltegravir, an integrase inhibitor - was formulated into a topical vaginal gel, for use as PEP in a macaque model. In vitro testing demonstrated that the drug inhibited virus for >10 hours post-infection, suggesting it would be an effective PEP agent. In a low-dose vaginal challenge model, gel containing raltegravir was given 3 hours after virus challenge. Five of six animals were protected over 20 challenges (p<0.005 versus 0 of 6 control animals). Thus, this opens the door for a new class of PEP/PrEP agents.
Two studies reported new data regarding the safety, pharmacokinetics, and usability of tenofovir gel. The first (Anton, abstract 34LB) was a phase I trial of rectally-applied 1% tenofovir gel, with the formulation of the gel matching what had been tested for vaginal use (the osmolarity of the vaginal product is significantly greater than an ideal rectal product). The study procedures were intensive - 12 visits with 8 sigmoidoscopies over 3 1/2 months for each of the 18 participants, including 7 days of active or placebo gel use, pharmacokinetics studies after oral tenofovir and gel dosing, and intensive safety monitoring (against placebo gel). There were more adverse events during the active gel period than during in the placebo period, mostly gastrointestinal events, most grade 1. There were also more complaints from participants. The rectal dosing generated high concentrations of active tenofovir in rectal biopsy samples and levels correlated with HIV infectivity in ex vivo testing. Very little tenofovir was absorbed into the bloodstream. The results emphasize the need to develop a rectal-specific microbicide. In the second study (Hendrix, 35LB), the safety, acceptability, and pharmacokinetics of vaginal tenofovir gel were compared with oral tenofovir and with dual dosing of both oral and gel in a large (144 participants), multinational, cross-over study (MTN 001), in which each method was used for 6 weeks. Nausea, diarrhea, and headache were more common with oral than vaginal dosing, but all were uncommon. Vaginal dosing achieved vaginal tissue concentrations of the active form of tenofovir (tenofovir diphosphate) that were 100 times that from oral dosing, and there was no additive effect of dual dosing; however, the necessary concentration of tenofovir diphosphate in tissues to prevent HIV infection is not known. Oral dosing was preferred to vaginal dosing in US women; gel and pills were equally acceptable in African women. Self-reported adherence was high but drug levels showed that potentially 40-60% of women were not using the study products consistently. These results provide critical information, both related to pharmacokinetics and adherence, for future PrEP studies.
Mathematical modeling studies have indicated 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, if they are effectively linked to HIV care, and if ART is initiated (together often called the 'test and treat' strategy). Several posters, discussed together in one session, assessed community viral load - i.e., an aggregate measure of all viral load results available in a community - as a marker for HIV transmission reduction as a result of ART use. In San Francisco (Das, abstract 1022), there has been a decline in community viral load between 2004 and 2009, with a concurrent decline in newly diagnosed HIV cases. In both Washington DC (Castel, abstract 1023) and New York City (Laraque, abstract 1024) community viral load has also decreased, with significant differences between neighborhoods. Modeling of MSM in the Netherlands indicates that immediate ART treatment needs to be coupled to behavior change in order to observe a sustained reduction in community infectiousness (van Sighem, abstract 483). The potential for test and treat interventions to reduce community viral load among injection drug users, in conjunction with drug use interventions, was reported from the ALIVE cohort (Kirk, abstract 484). Modeling of HIV transmission and the impact of ART programs in Africa indicates that ART reduces HIV incidence and that the majority of infections were caused by undiagnosed persons with HIV in the non-acute phase, underscoring the importance of increasing access to HIV testing (Vimalanand, abstract 482).
Prevention of mother-to-child transmission
Antiretrovirals have long been the mainstay of efforts to prevent mother-to-child transmission of HIV, and new results presented at CROI were no exception. Two large and important studies provided new information. The first (HPTN 046, Coovadia, abstract 123LB), assessed extended post-natal nevirapine prophylaxis from 6 weeks to 6 months for infants born to HIV+ women in South Africa, Tanzania, and Uganda. Extended prophylaxis offered significant HIV transmission benefit - at 6 months, transmission rates were 1.4% in the nevirapine group and 3.4% in the placebo group (p=0.03). The benefit was greatest for infants born to women with CD4 counts ≥350 who were not on ART. The second study (HPTN 040, Nielsen-Saines, abstract 124LB) assessed post-exposure prophylaxis provided to infants born to women who did not receive ART prior to labor, due to late HIV diagnosis. Neonatal dual antiretroviral PEP (zidovudine plus 3 doses of nevirapine) or 3-drug PEP (zidovudine plus 2 weeks of lamivudine+nelfinavir) PEP were both superior to 6 weeks of zidovudine alone. The 2-drug regimen had less toxicity, although resistance testing is not yet completed for infants experiencing breakthrough infection.
The 5th N'Galy-Mann lecture - celebrating the lives of international collaborators - was given by Anthony Harries, from the International Union Against TB and Lung Disease (abstract 18). The talk addressed HIV treatment, care, and prevention on a broad scale, informed by the author's experiences working in Malawi. One encouraging note: Malawi proposes in 2011 to initiate all HIV-infected pregnant women on ART, regardless of CD4 count, perhaps offering a bridging mechanism to expand antiretroviral therapy to those with CD4 counts >200 (as called for by recent WHO guidelines, but implemented in few countries, due to resource constraints)..
HIV prevention in the US
New HIV cases continue to accumulate in the US, in spite of access to treatment, prevention and care. In a plenary talk on the first morning of the conference, Jonathan Mermin from the US CDC laid out approaches to 'high-impact prevention' for the US, maximizing effects, advancing science, addressing costs and cost-effectiveness, and taking continuous evaluation into account (abstract 19). Another exceptional plenary, on Tuesday, provided an in-depth assessment of cost-effectiveness analysis for HIV care and prevention, which was relevant to both US and international settings (Walenksy, abstract 74).
Two studies from the CDC assessed HIV among men who have sex with men in the US. In a venue-based sample of 8153 men from 21 cities, 1562 were infected and 273 were assessed as likely being recent infections, based on reported negative testing within a year (Heffelfinger, abstract 130). Age ≥20 and black, Hispanic, or other non-white race/ethnicity were associated with greater likelihood of new HIV infection. Among ~1200 black and Latino men who have sex with men from 3 US cities (Millett, abstract 131LB), 12% were HIV+ and unaware of their status. Beliefs that intraracial mixing reduced HIV and low perceived risk were associated with being HIV+ and unaware. Missed opportunities for testing for men engaged in care were seen - black MSM who were 'out' to their providers were more likely to have unrecognized HIV infection.
Sexually transmitted infections, circumcision, and other factors influencing HIV transmission
Epidemiologic studies continued to advance knowledge about risk factors for HIV and new avenues for prevention. An afternoon symposium on responding to risk in different populations (abstract 66-69) provided excellent overview talks.
In a prospective HIV-1 incidence cohort among high-risk men who have sex with men and female commercial sex workers from coastal Kenya (Okuku, abstract 29), genital hygiene was protective against genital herpes acquisition in the men (aIRR 0.3, with a dose-response as the frequency of genital washing increased). In contrast, intravaginal cleansing in women was associated with increased HSV-2 risk (aIRR 1.9). Whether genital washing protects against sexually transmitted infections and HIV in men thus remains a topic of investigation.
In updated data from the Rakai, Uganda male circumcision study (which, along with trials from Kisumu, Kenya and Orange Farm, South Africa conclusively demonstrated that male circumcision reduces a man's risk of acquiring HIV by half), observational follow-up of the trial participants for up to nearly 5 years after the trial ended sound that the protective effect of circumcision has persisted, even increased - to 73% (Kong abstract 36). To date, 80% of men in the uncircumcised control arm have taken up circumcision since the trial ended, without evidence of behavioral risk increase as a result of circumcision.
Among ~3300 HIV serodiscordant couples participating in a multinational prospective study, the risk of HIV infection per sexual act was calculated (Hughes, abstract 135). The per-act risk of HIV was 1.9 per 1000 for male-to-female transmission and 1.0 for female-to-male transmission; however, the per-act risk for male-to-female and female-to-male transmission were not statistically different after adjusting for plasma viral load and other factors. This study provides some of the strongest observational data about the efficacy of condom use, which reduced HIV risk by ~80%. Plasma viral load was the strongest predictor of HIV risk, and sexually transmitted infections increased HIV susceptibility while male circumcision decreased risk. These results reiterate that key modifiable factors - reducing plasma HIV levels (with ART), use of condoms, male circumcision, and treatment of genital infections - drive HIV risk.
In another study of HIV serodiscordant couples (Bachanas, abstract 136), >3500 HIV+ persons in care in Kenya, Namibia, and Tanzania (42% male, 64% on ART) were interviewed about HIV testing of their sexual partners. 68% reported their partner had been tested for HIV, and 95% of those knew their partner's status (28% were HIV negative). However, overall, 36% did not know their partner's HIV status, and over half (54%) had a partner of negative or unknown status. 77% reported consistent condom use, which was less likely those seeking pregnancy. Interestingly, women were less likely to report condom use, even those with known HIV negative partners, than men. Integrating partner/couple testing and ongoing counseling about safer conception are opportunities to bring prevention into care settings. In a fascinating demography analysis (Tanser, abstract 137), population-based HIV testing in a rural area in KwaZulu-Natal, South Africa identified transmission clusters, highest in peri-urban high density populations near the national road to Mozambique. Overall, 40% of new HIV infections occurred in an area occupying just 8% of the study surface area. Thus, geographically-targeted prevention could be highly effective, even in a generalized epidemic.