|Home | About | Journals | Submit | Contact Us | Français|
HIV continues to spread rapidly with over 2.5 million new infections each year1. Since efficacious behavioral interventions have not resulted in durable declines in HIV incidence when scaled-up to achieve sufficient coverage in many populations, and it will take years for highly effective HIV preventive vaccines to demonstrate efficacy2-4, an increasing amount of attention has focused on whether available antiretroviral drugs could be used to slow the epidemic. For over a decade, increasingly well-tolerated highly active antiretroviral therapy (HAART) has dramatically changed HIV-associated morbidity and mortality and has improved the quality of life of HIV-infected individuals5, 6. Recent global initiatives have concentrated on expanding access to HIV treatment in resource-limited settings7; so, by the end of 2008, close to 4 million people were receiving ART8. In addition to the individual clinical benefits accrued from increased access to treatment, a growing group of researchers and public officials have suggested that ART maybe used to decrease HIV transmission globally9, 10. ART has already dramatically decreased mother-to-child HIV transmission (MTCT)11, and could conceivably be used to prevent the sexual transmission of HIV via reductions in genital tract HIV concentrations in individuals who are already infected12, 13, or as pre- or post- exposure prophylaxis for uninfected people exposed to HIV14 (See Figure 1).
However despite increasing drug availability, the effectiveness of ART for prevention may be limited by concurrent sexually transmitted infections (STIs) that increase infectiousness and susceptibility, non-adherence to therapy, drug-related toxicities, viral resistance, cost considerations, and risk compensation. The effects of HAART initiation can manifest differently in diverse social settings as sexual behavior involves concerns over pleasure and procreation. While early data from the developed world suggested that HAART could be associated with increased sexual risk15-18, more recent data from sub-Saharan Africa has suggested that wider access to HAART is not associated with increased sexual risk taking behaviors19-21. This paper reviews the potential uses of ART for HIV prevention and examines future directions for ART preventive interventions (See Figure 2).
HIV transmission remains a low probability but high consequence event, occurring in less than 1 in 100 contacts on average22-24, but the global pandemic is potentiated by the frequency of humans having sexual intercourse and factors that amplify infectiousness and susceptibility in specific settings. The per contact calculation is based on composite data, and transmission probabilities vary considerably during the course of the disease, with higher transmission probability in the acute and late phases of HIV infection, as a reflection of plasma and genital HIV concentrations25-29. The best empirical data related to our understanding of how viral suppression could lead to decreased HIV sexual transmission comes from the Rakai study conducted among Ugandan serodiscordant couples30, though this study was completed before generic ART had become widely accessible in Uganda. HIV was not transmitted in discordant couples when the infected partner had a plasma HIV RNA level of less than 400 copies/ml30. This association between viral load and the risk of HIV transmission among serodiscordant couples was confirmed in subsequent studies in Zambia and Thailand31, 32 in the pre-HAART era, and more recent data33 showed that HIV-infected partners in discordant relationships were substantially less likely to transmit HIV to their partners if they were on HAART.
However, other variables may affect genital tract HIV, besides systemic HIV burden. Mucosal HIV transmission is complex, since animal models suggest that either cell-free or cell-infected virus can replicate in a variety of host cells34-36. The minimum inoculum of HIV that can cause human infection remains to be elucidated37. Although some studies have documented HIV preferential binding to cervical and foreskin tissues through dendritic cells38, 39, women who have undergone hysterectomy and circumcised men can also be HIV-infected12, 40, so other urogenital cells can support HIV replication. The temporal window of opportunity for halting transmission through host defenses is very limited, as submucosal viral replication occurs within hours of exposure41. In the first study examining the relationship between treatment and the detection of genital tract HIV, the virus was more readily cultured from seminal plasma and leukocytes in participants with leukocytospermia and/or advanced disease stage, and less likely to be detected in semen among men taking zidovudine (AZT)42. Subsequent semen studies found that genital white blood cells were a significant source of viral burden43,44. These data suggest that factors that increase genital tract inflammation may potentiate infectiousness, and ART could decrease infectiousness, presumably by decreasing viral burden.
Several groups have demonstrated that HAART suppresses viral replication not only in the blood and lymphoid tissues, but also in the male and female genital tract12, 13, 45, 46. While treatment generally suppresses cell-free virus in the semen, many treated individuals can still harbor proviral CD4 cells in their semen44, 47. A further complication is that not all drug combinations of HAART may be equally effective in the genital and systemic compartments12, 48. Differential penetration of ART drugs in genital tissues may be a function of protein binding and other pharmacodynamic properties49. Highly protein-bound drugs, like protease inhibitors, achieve lower concentrations in genital tract secretions compared to blood plasma14, 50. Nucleoside analogues and non-nucleoside reverse transcriptase inhibitors achieve higher concentrations in genital tract secretions than blood plasma51-53, which might make them particularly effective in decreasing sexual transmission of HIV.
Based on these findings, questions arise as to when would be the best time to initiate HAART if a primary goal is to decrease infectiousness among those who could transmit HIV to others. One key period for treatment could be the acute phase of HIV infection when patients have elevated viremia for about three weeks, until host defenses suppress replication, creating a “viral set point” 29, 54, 55. When viral replication is unimpeded during acute HIV infection, individuals may have more than 1 million copies/ml of virus in their blood56, and the potential to infect many individuals54, 57. Studies from North America and Africa have shown between 40% and 50% of new transmissions were from recently infected patients54, 57. Through employing discordant HIV rapid test results and RNA pooling, almost 2% of STI clinic patients in Malawi were identified with acute HIV infection58. These findings suggest that promptly identifying “hot spot” microepidemics of newly infected persons may present an opportunity for early ART and behavioral interventions to slow the spread of HIV in high-risk settings59, 60. Studies are underway to determine if early identification and treatment of acutely infected individuals can have a public health impact on local epidemics, but other questions will take more time to address, such as whether treatment can be discontinued without being detrimental to the patient if it is initiated in the setting of acute HIV infection. Public health challenges to identifying patients with acute infection include very short duration, frequent lack of clinical detection, and the expense of performing RNA-pooling in the era of constrained health resources.
Acute HIV infection is followed by a longer period of chronic viral homeostasis in which an infected person may be asymptomatic with good systemic virological control. This period can be interrupted by STI infections, which may override the suppressive effects of ART in the genital tract by causing inflammation, potentiating local HIV replication27. Due to the relatively shorter periods of acute- and late-stage infection, the prolonged asymptomatic period, which may last more than a decade, may be the time in which substantial transmission events occur61-63. An analysis from Sub-Saharan Africa suggests that acutely infected persons play a major role in HIV transmission during early highly concentrated epidemics, while the contribution of chronically infected persons becomes more prominent during advanced and stabilized epidemics61. A recent projection suggested that annual HIV testing followed by immediate initiation of ART for all HIV-infected patients regardless of CD4 cell count could have a major impact by reducing the number of new HIV infections64.
So, can HAART effectively suppress genital tract HIV over a sufficiently long period of time to stop HIV sexual transmission?65. Studies from Taiwan and British Columbia, Canada have documented a greater than 50% reduction in anticipated number of incident HIV cases following the free provision of ART in 19979, 66. A study among 393 couples in the pre- and post-HAART eras in Spain observed an 80% reduction in HIV transmission following the introduction of HAART67. A recent observational study of discordant African couples showed substantially lower rates of HIV transmission among patients who were on HAART, but these individuals also tended to practice safer sex more than those who remained untreated33. In order to better understand the preventive effects of therapy in discordant couples, the Health Prevention Trials Network of the National Institutes of Health is in the midst of a randomized controlled trial to assess the impact of HAART on transmission among 1750 discordant couples in eight countries (HPTN052)68. HIV-infected partners in serodiscordant, monogamous relationships who have CD4 counts that would not warrant immediate initiation of HAART (ie >350 cells/ul) based on current guidelines are being randomized to either initiate HAART right away or are clinically monitored until their immunological or clinical status warrants treatment. This study will be fully enrolled in the coming year, and should provide important information shortly thereafter.
As more individuals live with HIV because of HAART, there is a larger pool of individuals who could transmit HIV, if HAART is not fully suppressive and they increase risk taking behavior. Studies on the impact of ART on sexual behavior among HIV-infected individuals have been inconsistent, and it is not clear whether the provision of ART may be associated with high risk sexual behavior in some subpopulations69, 70. Though data from injecting drug users and men who have sex with men (MSM) in the developed world suggest that high risk behavior can increase with ART17, 71, studies from resource-limited settings to date have not shown that the provision of ART does increase HIV risk taking behaviors72.
Despite the increasing interest in the evaluation of “test and treat” strategies, other genital tract factors could mitigate the public health impact of early initiation of HAART. For example, concurrent STIs can increase HIV transmission probabilities73, by attracting CD4 T-lymphocytes and releasing cytokines (TNF-alpha, IL-1) that enhance HIV transmission37. Bacterial STIs can increase genital tract HIV concentrations74, and interventions that provide treatment of these infections can decrease viral shedding27, 75-77. In the Rakai study, HSV-2 antibody negative participants with viral loads >38500 copies/ml and HSV-2 antibody positive participants with viral loads <1700 copies/ml had a similar probability of transmitting HIV to the uninfected partner78. Unfortunately, in many settings HSV-2 seroprevalence in young adult populations exceeds 50%79, and 2 studies have documented that thymidine-kinase inhibitor chemoprophylaxis does not decrease HIV transmission to uninfected patients80, 81. It is feasible that the removal of a pathogen that initiated a genitourinary inflammatory response may not be enough to restore the local cellular milieu.
Another concern about the wider use of ART for prevention has been the potential for the development and transmission of antiretroviral resistance if healthy individuals are expected to be fully adherent over many years. Multiple reports have now documented that resistant HIV has been sexually transmitted82. Though ART resistance has been demonstrated with increased frequency in newly diagnosed and treatment-naïve patients over time83, 84, recent findings suggest that ART-resistant HIV strains may have lower viral transmisability85. Thus, public health systems around the world will need to track some of the potential unintended consequences of earlier initiation of antiretroviral therapy for prevention, including risk compensation, trends in STI co-epidemics, and the prevalence and incidence of transmitted ART-resistant HIV, in order to fully understand the costs, as well as the benefits of treatment as prevention.
The evidence that suggests that ART can prevent HIV acquisition comes from the success of MTCT86, animal studies87, 88, and a case-control study of post-exposure prophylaxis (PEP) following needle stick injury in healthcare settings89. The Centers for Disease Control and Prevention (CDC) registry documented that individuals who took AZT monotherapy following occupational exposure were one fifth as likely to be HIV-infected as those who did not take medication89, 90. Data from the rhesus macaque model have suggested that 28 days of ART is needed for effective post-exposure prophylaxis88. Concerns have been raised that the use of PEP in non-occupational settings might result in increased risk taking in some populations, such as among MSM. In a Brazilian MSM study, high risk participants in a behavioral risk reduction study were educated about PEP while being counseled about safer sex, and participants were given 4 day “starter packs” of zidovudine/lamivudine. They were educated that if they were risky and began to take the medication, they needed to come back to the study site as soon as possible, so they could get the rest of the 28 day course that they would be expected to take. About a third of the 200 high risk men followed for 24 months (68 men) used PEP 109 times91, another third were risky, but did not use PEP, and the remainder heeded the counseling messages and did not engage in risk. The overall HIV incidence in the cohort was 2.9 per 100 person-years, but there were 10 infections among men who did not utilize PEP, compared to only one among the men who used PEP. Although HIV risk taking behavior may not be constant over time,92 most studies of PEP after sexual exposure have not demonstrated increases in risk taking behavior after PEP, and some have effectively included counseling so that PEP could provide an entry for intensified risk reduction interventions91, 93, 94.
A major impediment to the wider use of PEP in the past was the relative intolerability of some of the first-line recommended drugs, like azidothymidine and protease inhibitors95. However, the use of newer drugs like Tenofovir seems to be associated with increased tolerability and completion rates, although randomized controlled trials have never been done to compare PEP regimens head-to-head due to logistical issues, such as the huge sample size needed to compare two effective regimens. Tenofovir has many features that are desirable in a chemoprophylactic agent, including long intracellular half-life, activity in macrophages, and high concentrations in genital tissues96. A case-control study found that men who took dual therapy with tenofovir-emtricitabine were more likely to complete a 28-day day PEP regimen than historical controls taking 2 drug regimens containing AZT97. It is still not known if it is preferable to use 2 drugs versus 3 drugs for PEP, with the argument for 2 being increased tolerability and completion rates98, versus the view that because the person had already been exposed to HIV, more drugs would be extra protective against drug-selected or spontaneous mutant strains. Other newer drugs may offer opportunities for novel PEP strategies, because they are well-tolerated (e.g. Atazanavir or Raltegravir)99 or achieve high genital tract concentrations (e.g. Maraviroc)100. Although several lines of data suggest that PEP may decrease the likelihood of HIV transmission, in many settings it may be underutilized because of clinician concerns about risk compensation and cost, and because many at risk individuals are unfamiliar with its potential or how to access chemoprophylactic treatment.
In situations when the likelihood of exposure to HIV can be anticipated ahead of time, ART pre-exposure prophylaxis (PrEP) delivered either as oral therapy or as topical microbicide, could be a logical method of primary prevention. In the field of infectious disease preventive care, patients are routinely provided with prophylaxis prior to exposure when the risk of infection is imminent, including anti-tuberculosis therapy and anti-malarials. Simian-based data suggests that tenofovir-containing regimens protect against infection with rapid drug absorption and high drug levels remaining in intracellular genital tissues101, and with the possibility of intermittent dosing102. Over the past decade, animal studies have provided the basis for clinical PrEP research103, though concerns about access to optimal preventive services and medical treatment for vulnerable populations has impeded initial PrEP research104. Concerns about PrEP include possible behavioral disinhibition, ART cost, acquisition of resistant viral strains, treatment adherence, and chronic medication toxicities.
A phase II, randomized double-blinded placebo controlled trial was completed three years ago in Cameroon, Nigeria, and Ghana which demonstrated the safety of daily oral tenofovir compared to a placebo for HIV prevention among high-risk women also receiving HIV testing, counseling, and condoms105. After enrollment, 8 on-product seroconversions occurred with 2 in the women who took tenofovir and 6 on placebo. Although this difference was not statistically significant, the trend was in the right direction, and both groups of women had reduced their behavioral risk in the course of the study. The number of sex partners went down and the proportion of participants reporting condom use increased over time (52% at baseline to 95% at 12 months)105, 106. Despite concerns that PrEP could lead to behavioral disinhibition, an important finding in this study was that risk behavior decreased over time after initiating PrEP.
These reassuring data helped public health researchers feel comfortable in studying PrEP in efficacy trials in several high risk populations around the world. Over the next few years, data will be available to see whether oral tenofovir by itself, oral tenofovir co-formulated with emtricitabine, or topical tenofovir gel will be more effective than placebos among MSM in the Americas, Thailand and South Africa; among at risk women in sub-Saharan Africa; among HIV-discordant couples in Africa; and among Thai injection drug users. Prior studies have demonstrated that topical tenofovir is safe and well-tolerated, but in more than half of the women in a pharmacokinetic substudy, systemic levels of tenofovir were detected when administered topically to low risk women107. For some, this is good news, since systemic absorption means that significant genital tissue levels were achieved, but low concentrations might be less likely to be associated with clinical toxicities. For others, this would be a concern, since the levels were very low compared to the effects of taking a 300 mg oral dose. The ability of topical 1% tenofovir gel to deliver high drug levels to the genital tissues with lower systematic drug levels108 has led to the first study (known as the VOICE Trial) to compare daily use of 1% tenofovir gel to placebo gel and also compare oral tenofovir or tenofovir/emtricitabine compared to oral placebo109, 110. Assessment of the relative merits and limitations of oral versus topical PrEP agents will require careful anthropological work in addition to clinical trials in order to understand people’s beliefs about whether one approach is associated with more sexual pleasure, fewer systemic side effects, and perceptions of efficacy in one culture compared to another.
One broad area of concern is the development of drug resistance through the continued use of PrEP after becoming infected, either by a resistant virus, suboptimal adherence, or failure of the prophylactic regimen. In the early phase studies of therapeutic uses of tenofovir, when HIV-infected patients initiated therapy with tenofovir alone, no resistance was detected after 28 days111. In non-human primate-based studies, enriched resistant minor variants were detected in monkeys and tenofovir PEP was still partially effective compared to drug resistant SHIV112. One study noted no tenofovir resistance after tenofovir or tenofovir/emtricitabine failure113, while another study detected intermittent emtricitabine resistance114. When the virus develops tenofovir and emtricitabine resistance, the resistant virus is hypersusceptible to zidovudine, but diminishes the activity of other nucleoside reverse transcriptase inhibitors; there is no effect on other classes of drugs.
Two potential concerns emerge with tenofovir prophylaxis in light of current treatment regimens. First, in resource-limited settings, stavudine is routinely used as part of first-line drug treatment because it is inexpensive. Of concern is that clade C virus that is most common in these settings may preferentially select for K65R resistance after exposure to stavudine-containing HAART, which could create tenofovir resistance115, 116. Thus, wider stavudine use may be selecting for a larger pool of circulating strains with K65R, which potentially could result in failure of tenofovir-based prophylaxis. Another concern is that individuals on tenofovir-based PrEP will develop tenofovir resistance after becoming infected. However, mathematical modeling of HIV prophylaxis for primary prevention suggests that less than 1% of the predicted seroconversions would acquire or develop a tenofovir-resistant strain117, but in a world where tenofovir is being increasingly used for first line antiretroviral therapy, continued monitoring to assess the potential ecological effects of PrEP is warranted. Unlike clinical trials in which participants receive frequent antibody-based HIV testing, patients in clinical settings will require ongoing HIV testing to avoid substandard therapy if they potentially become HIV infected. In either of these situations, more individuals with tenofovir resistant virus could compromise their own available treatment choices and spread resistant virus to their partners.
Currently, clinical trials of PrEP are underway at multiple sites that will enroll >20,000 HIV-uninfected men and women from sites in Asia, South and North America, and Africa, and will address the role of continuous vs intermittent PrEP, topical vs oral PrEP, selection of specific drugs, and the influence of PrEP on risk practices118. Some of the first PrEP efficacy data may be available within the next year or two, so public health officials and clinicians will need to think about how to train providers to make the medications available to at risk populations and provide careful monitoring, and how to use PrEP to create “educable moments” that facilitate HIV prevention if PrEP “works.” Partial efficacy and potential toxicity management issues (e.g. what if tenofovir PrEP decreases the likelihood of HIV acquisition by 50% and a few participants develop renal failure?) are two examples of the kinds of challenges in the development of clear and succinct messages as clinical trial data matures. One community-based organization, the AIDS Vaccine Advocacy Coalition (www.avac.org) has developed a PrEPWatch feature that provides continuously updated information about the status of the clinical trials.
Many clinical questions will arise whatever the results of the first generation PrEP studies show (Figure 3). Other antiviral drugs are being considered for chemoprophylaxis, ranging from the oral CCR5 antagonist, Maraviroc, which achieves high genital tract levels, to UC-781 and TMC-120, which are poorly absorbable non-nucleoside reverse transcriptase agents and thus are being developed as topical microbicides. Injectable agents and compounds that can be delivered through a slow release ring may alter the need to be adherent to a stringent daily regimen. Some agents may be considered as optimal parts of PrEP regimens, such as lamivudine and emtricitabine, because the virus that is resistant to these drugs has decreased viral fitness 119, 120. So individuals who would become infected with this less virulent viral stain would be less likely to transmit to their own partners. Because of the long intracellular half-life of drugs like tenofovir, intermittent dosing strategies may make sense. Pre- and post-coital application of tenofovir gel is being studied in the Caprisa 004 study among high-risk women in South Africa.
One challenge of the use of ART for prevention is the possibility that at risk persons might obtain off-label drugs even in advance of any efficacy data being available. One recent report suggested that ART was being sold at clubs and self-administered prior to high risk sexual activity121, but other reports from San Francisco and Boston found minimal use among MSM in the past 2 years122, 123. Clearly, the potential for widespread and unregulated use is great, and the environment could change quickly once data become available to suggest a beneficial effect from antiretroviral PrEP. Early work suggests that context matters, i.e. the levels of interest in using PrEP were substantially affected by perceived efficacy, side effects, and cost122. Public health authorities, led by the CDC in the US and WHO globally, have begun to meet regularly to anticipate community responses and will be mobilized to work with clinicians, national governments, and representatives of high risk communities, once efficacy data become available124.
After nearly two decades of using HAART, clinicians and public health officials must consider how to optimally utilize these medicines in a focused manner to reduce the number of new infections. Major questions requiring operational research include long-term safety, adherence, benefits of different modes of drug delivery and dosing, and selection of resistant virus. Some of these issues will be addressed in the context of phase IV expanded safety studies, while others will require the strengthening of public health monitoring systems. Increasing efforts are now underway to monitor the emergence of acquired antiretroviral drug resistance in resource-limited settings via the WHO’s Global Network HIVResNet, which provides standardized tools, training, technical assistance, laboratory quality assurance, analysis of results and recommendations for guidelines and public health action125. The recent expansion of WHO treatment guidelines to initiate treatment at a higher CD4 cell count of 350 cells/ul rather than 200 cells/ul will mean that a larger pool of HIV-infected patients will be in need of treatment126. Recent curtailment in the US President’s Emergency Plan for AIDS Relief (PEPFAR) for HIV treatment could raise further ethical concerns regarding the need to provide life-saving medicine to those already infected with HIV relative to preventing future HIV infections127. Moreover, many resource-limited nations may have limited budgets for HIV treatment, and thus will have to carefully decide how to make best use of limited resources in order to also decrease the number of new infections.
The degree of public health benefits reaped through the use of ART for prevention will depend on the number of HIV-infected individuals treated, the ability to effectively engage individuals most likely to transmit HIV, the relative stage of a given epidemic, the efficacy of specific ART regimens to reduce viral load in the genital tract, the development of drug-resistant viral strains, and changes in risk taking behaviors that could compromise the protective effects of ART. While targeting ART preventive therapy to infectious individuals or individuals at greatest risk of acquiring HIV can be a major challenge, there is yet a lack of consensus over the preventive benefits of the widespread administration of ART for the general population. Mathematical models have suggested that the widespread provision of ART could substantially reduce HIV incidence, but this benefit could be potentially undermined by behavioral disinhibition128, 129.
In summary, the use of ART to reduce HIV transmission has moved to the forefront of public health approaches to HIV prevention because of the increased tolerability of the medications, decreased cost, expanded formulary, and limitations of other approaches. Due to clinical indications that earlier initiation of ART for infected individuals is warranted, optimizing the benefit will require attention to adherence, sustained access, behavioral risk reduction, and STI diagnosis and treatment. Although the use of ART for uninfected individuals holds great promise, public health authorities will need to assess the potential of local decreases in HIV infection relative to financial costs and ecological impact, if efficacy trial data shows benefit. The field is at an early stage, with major questions remaining to be answered, such as what is the least amount of medication that can be effective with pre-/post-coital dosing and what are the optimal routes of drug delivery (e.g. topical, oral, and injectable).
It is conceivable that in the future the ART formulary will consist of drugs to be utilized for specific preventive and therapeutic interventions, such as topical agents for stopping viral entry. An effective HIV vaccine is still years away3, but the utility of ARVs for prevention will need to be tempered by the fact that these agents are not likely to be 100% effective in the real world. Further studies in pharmacology, virology, and behavioral science will be needed to best understand the intended, and unintended, clinical consequences of increasingly using ART. The use of ART as a critical tool for HIV prevention will need to be part of a larger toolbox to reduce the number of new infections globally, including circumcision, prevention of MTCT, behavioral change, and treatment of STIs. It is possible that the life-saving agents that have transformed the natural history of HIV disease and the quality of life of infected patients may become part of HAARP: highly active antiretroviral prevention130, but their ultimate potential in preventing HIV transmission remains to be fully defined.