Position statement on the use of antiretroviral therapy to reduce <scp>HIV</scp> transmission, <scp>J</scp>anuary 2013: The <scp>B</scp>ritish <scp>HIV A</scp>ssociation (<scp>BHIVA</scp>) and the <scp>E</scp>xpert <scp>A</scp>dvisory <scp>G</scp>roup on <scp>AIDS</scp> (<scp>EAGA</scp>)

Abstract

It must be noted that no single prevention method can completely prevent HIV transmission. ART reduces the risk of transmission only of HIV. Irrespective of ART, condoms remain the most effective way to prevent the spread of other STIs. The above position statement by BHIVA and EAGA summarizes extensive discussion about various aspects of the scientific data and it was felt that some explanatory notes would be helpful, particularly where there are areas of controversy. The mechanisms by which condoms and ART prevent HIV transmission are fundamentally different. Condoms prevent contact with genital fluids and their efficacy is reduced by factors that compromise the integrity of the physical barrier, such as non-use, slippage and breakage. ART prevents HIV transmission by stopping viral replication and lowering the amount of virus within the genital compartment; its value will be reduced by nonadherence, poor absorption and the presence of other STIs. The observed reduction in HIV transmission between couples (assumed to be having vaginal sex) in the HIV prevention trials network (HPTN) 052 trial 1 was 96%, when the HIV-positive partner took ART. We do not yet know, however, how ART use affects HIV transmission between couples in ‘real-world’ settings outside a clinical trial. Conversely, there has never been a randomized controlled trial of the efficacy of condom use vs. no use. However, several meta-analyses of observational and cohort studies of HIV infection in couples who maintained 100% condom use have found that this strategy is about 80% (79–93%) effective in reducing HIV infections 2. It must be noted, though, that it is not possible to make a direct comparison of these two strategies: HPTN 052 was a prospective randomized controlled trial enrolling HIV-serodiscordant couples where HIV transmission was the primary outcome, whereas the condom evaluation was a meta-analysis of multiple observational studies, and as such may underestimate the effect of condoms. BHIVA and EAGA believe that giving an actual figure for the risk of transmission for one episode of sex in a serodiscordant couple is not currently meaningful for an individual and that any figure proposed would be misleading, for the reasons outlined below. In the absence of such a figure, BHIVA and EAGA have therefore adopted the term ‘extremely low’ whilst recognizing the difficulty inherent in the imprecise nature of such a term. The studies conducted to date in heterosexual serodiscordant couples indicate that there have been no confirmed transmissions from people whose HIV infection is virologically undetectable (< 50 copies/mL). The small number of documented HIV transmissions in these studies occurred from HIV-positive individuals who had only recently started therapy and in whom, therefore, it is unlikely that an undetectable HIV viral load had been achieved or sustained for the 6-month time period recommended by this statement. However, to be certain that the risk of transmission approaches zero in defined circumstances, a much larger number than the 1763 serodiscordant couples enrolled in HPTN 052 would have to be studied. This will be tested in the PARTNER study 9. There was a 96% reduction in HIV transmission risk demonstrated in the HPTN 052 study, which can be considered as ‘extremely low risk’. Within the study partnerships, there was only one genotypically confirmed HIV transmission from an HIV-infected participant on ART. In this case, an individual randomized to immediate ART had not yet achieved an undetectable viral load at the time of viral transmission. The BHIVA and EAGA statement requires evidence of confirmed HIV viral load < 50 copies/mL for 6 months, which would exclude a comparable risk to that observed in the trial, hence justifying the ‘extremely low’ statement; although this does not mean zero risk. The nature of sexual exposure does influence the actual risk of acquisition/transmission. The actual relative risk for each individual sex act is not certain, as multiple factors are at play 3-5. Biologically, the integrity of the exposed mucosal surface is important as well as the presence of concomitant mucosal infections. The latter influence membrane integrity, attract inflammatory target cells and affect HIV shedding by the genital tract. Estimates of risk of HIV acquisition per coital act are largely influenced by log10 viral load of the HIV-infected partner; whilst the majority of these data are from African heterosexual couples reporting vaginal sex 3, 4, 6, the assumption from these trials is that the majority of sex acts were vaginal sex. The concept that the HIV-positive partner's viral load is the key determinant of risk of transmission is pertinent for all sex acts, although the absolute risk is affected by the nature of exposure. Because the risk of transmission through anal sex is higher than that through vaginal sex 7, and because of the lack of high-grade evidence that ART prevents viral transmission through this route, it is not possible at this time to confirm the same level of protection by ART as for vaginal sexual exposure. The data overall show that, for each log10 increase in plasma HIV-1 RNA, the per-act risk of transmission increased 2.9-fold [95% confidence interval (CI) 2.2–3.8] 5, 6. Whilst HIV viral load is the most significant contributor to risk of onward viral transmission, there is an order of magnitude difference in risk of transmission between insertive and receptive anal sex, with transmission by receptive anal sex around 10-fold more likely than transmission by insertive sex 5-7. Insertive anal sex carries a similar level of risk to insertive or receptive vaginal sex (estimated at 5–6/10 000 exposures), whilst receptive anal sex carries an estimated 10-fold higher risk of viral transmission (estimated at 50/10 000 exposures) 5-8. UK data from Fisher et al. 8 correlated the risk of onward transmission via anal sex to viral load, recent HIV infection and recent STI (rate ratio 5.32; 95% CI 2.51–11.29). The PARTNER study 9 will evaluate risk of viral transmission through anal sex when the HIV-positive partner is on suppressive ART, although this study will not report until 2015. For sexual transmission of HIV, reduction in genital tract HIV viral shedding is a critical factor that determines the overall risk of onward viral transmission. Generally, the plasma level of HIV RNA is a good surrogate marker for genital tract viral burden, but this is not always the case. HIV resides within anatomical ‘sanctuary sites’, meaning sites that are not directly part of the blood system, where local drug exposure and viral dynamics may differ significantly. Antiretroviral drug penetration varies by gender and may be drug (as opposed to class) specific, with high inter-individual variability 10. The HPTN 052 study reported a 96% reduction in the risk of onward viral transmission, measured from the time of randomization into the study, when the plasma viral load of the HIV-infected partner was below the limit of detection 1. For individuals initiating ART, it can be anticipated that the majority of people starting an effective regimen based on their pretreatment viral genotype (i.e. their virus is sensitive to all the drugs taken) will achieve an undetectable viral load within 6 months of initiating therapy 11. Overall, there are insufficient data to clearly respond to this question. Models have explored the role of specific STIs and onward HIV transmission risk. Observational and biological data provide compelling evidence of the importance of STIs in HIV transmission, but only one of nine STI treatment intervention trials has shown an effect 12. Overall, trial evidence strongly supports the concept that STI treatment reduces HIV infection. However, issues in trial design and conduct, including HIV epidemic phase, STI prevalence, efficacy of the intervention (especially in the herpes trials 12, 13) and statistical power, have affected five of the six trials. In addition, these studies were undertaken prior to general availability of ART, and the significantly higher impact of HIV viral burden on risk of transmission than that conferred by the STI(s) probably also explains the lack of overall efficacy of STI treatment. In the pivotal HPTN 052 study of serodiscordant heterosexual couples, 28% of transmissions were not from the enrolled infected partner 1. This was demonstrated by the absence of a virological link between the newly acquired infection and the partner who was infected at enrolment. This provides conclusive evidence that there were concurrent partnerships, not just the one protected by ART but at least one other that was unprotected, reinforcing the need for condom use outside long-term partnerships. The vast majority of HIV transmission in the UK is via casual sex and sex between new partners. If individuals plan on having sex without condom use, it is important to consider waiting until the HIV and STI statuses and the HIV viral load of the sexual partner(s) are known. If an HIV-positive individual chooses to rely on ART as the main method of prevention of HIV transmission, then viral load monitoring should be recommended more frequently than every 6 months. Full adherence to ART with continued suppression of plasma viral load is critical for the strategic use of ART to continue to prevent onward transmission. Stopping ART is usually accompanied by a significant increase in HIV viral load and hence an increase in the risk of onward sexual transmission. If ART is stopped for any reason, continued use of other prevention strategies is required to reduce the risk of transmission.