Reports of modest success from two trials (2010) of tenofovir as pre-exposure prophylaxis against HIV (PrEP)– CAPRISA 004 (1) (topical gel/women) relative risk reduction (RRR) 39%; iPrEX (2) (oral tablet/MSM) RRR 44% – have inspired recent contributions to STI journal.  In particular, there have been attempts to model the possible impact on the epidemic of incorporating PrEP in the toolbox of interventions (STI 2011; 87: suppl. 1, see especially: http://sti.bmj.com/content/87/Suppl_1/A36.1.abstract?sid=dc8129a6-8215-4699-a8f5-050cd87a8ff9; http://sti.bmj.com/content/87/Suppl_1/A350.2.abstract?sid=dc8129a6-8215-4699-a8f5-050cd87a8ff9): also studies evaluating the potential acceptability of PrEP in various settings (see: http://sti.bmj.com/content/early/2012/09/25/sextrans-2012-050648.abstract?sid=dc8129a6-8215-4699-a8f5-050cd87a8ff9; http://sti.bmj.com/content/88/4/258.abstract?sid=dc8129a6-8215-4699-a8f5-050cd87a8ff9).

Further tenofovir PrEP trials have produced results that are puzzlingly varied, and, given the hopes placed in PrEP as a preventative tool, somewhat disappointing:  CDC TDF2 (3) (oral tablet/heterosexual partners of men/women of indeterminate HIV status) RRR 62%; Partner’s PrEP (4) (oral tablet/heterosexual partners of HIV+ men/women) RRR 75%;  FEM-PrEP (5) (oral tablet/women) discontinued for futility; VOICE (topical gel/women) discontinued for futility. This has prompted some reflection on the efficacy of large RCT of the kind, with the discussion focussing on the difficulties around adherence (http://sti.bmj.com/content/87/Suppl_1/A196.3.abstract?sid=778e710d-42a6-4de3-a7e3-b9cd3f22a57e).  Crucially, Hendrix discusses the problem of disentangling the drug-related factors from issues of adherence (http://sti.bmj.com/content/87/Suppl_1/A1.4.abstract?sid=dc8129a6-8215-4699-a8f5-050cd87a8ff9).

One complicating factor in assessing tenofovir PrEP trial results is the use of two “routes” – oral tablet and topical gel – likely to produce concentrations of the drug at different anatomical sites.  This issue has now been taken up in an important recent paper (Hendrix & Bumpus) reporting the results of Microbicide Trials Network (MTN) study 001 (http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0055013).  In discussing pharmacokinetic issues, this paper also deals with the role of adherence as an explanatory variable in a way that casts an altogether fresh light on the issues Hendrix tackles in the above-cited paper (http://sti.bmj.com/content/87/Suppl_1/A1.4.abstract?sid=dc8129a6-8215-4699-a8f5-050cd87a8ff9).

MTN-001 is a crossover study in which all participants were assigned to a randomized sequence of daily tenofovir orally vaginally, or both orally and vaginally in three 6-week periods.  The key finding is that the concentration of the drug in its active form tenofovir disphosphate in vaginal mucosa achieved by the gel is 130-173 times greater than that achieved by the tablet; whereas the concentration in the serum and peripheral blood mononuclear cells (PMBC) achieved by the gel is respectively 58 and 56 times less.

It has generally been assumed that it is vaginal tissue concentration of the drug which gives protection.  One possible conclusion we might draw from these findings is that vaginal tissue is not, in fact, the critical site of ARV action.  The problem is that CAPRISA 004 (1) demonstrates that vaginal gel is having some effect through the vaginal compartment; yet that effect – given that it exists – is not at the level that might have been anticipated given 1. the increased HIV protection, and, especially, 2. increased toleration of missed doses, that we would expect from such a concentration. The second point is particularly important given the weight attached in this discussion to adherence as an explanatory variable.  With the half-life of elimination of tenofir diphosphate an estimated 90 hours, women having dosed once with gel could stop all dosing for more than 2 weeks before their vaginal tissue concentration fell to the vaginal concentrations achieved in oral dosing!  Hence, non-adherence appears not to be a plausible as a sole explanatory variable.  Something else must be going on.  But what?  The authors suggest, first, the possibility of tissue toxicity either from the tenofovir or the gel vehicle; second, some other factor that we are as yet unable to determine.

The authors conclude, naturally, that identification of the unknown variable is essential to improving future PrEP development…

In the wake of the 2010 results of the Pre-Exposure Prophylaxis Initiative (iPrEx) study – a multi-national trial of daily oral tenofovir/emtricitabine to prevent acquisition of HIV among men who have sex with men (MSM) indicating a 44% reduction in risk –  there have been attempts to model the potential health impact and cost-effectiveness of Pre-Exposure

Prophylaxis interventions among MSM, according to various scenarios of prioritization of high-risk groups (e.g. Cremin, Garnett et al.: http://sti.bmj.com/content/87/Suppl_1/A36.1.abstract?sid=a288a0b2-af68-40d5-8b61-1ea27b632395). A recent modelling study based on epidemiological data of MSM in Lima, Peru – Gomez, Hallett et al. – previewed for STI journal last year as a late breaker abstract (http://sti.bmj.com/content/87/Suppl_1/A350.2.abstract?sid=a288a0b2-af68-40d5-8b61-1ea27b632395), has now been published in PLoS Medicine.  It brings out with particular clarity the potential impact on cost-effectiveness of focussing limited resources on a relatively narrow, high-risk sector of the MSM population such as trans-women and sex-workers http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed.1001323.

Why is this interesting?  Gomez, Hallett et al. make an apt comparison of the total cost of introducing PrEP in Peru on a sufficient scale to reduce new infections by a third – US$20 million p.a. over 10 years – and total current HIV spending p.a. in Peru (prevention, treatment and care) – US$40 million p.a..  In short, given an insufficiency of resources, it is not enough that an intervention be cost-effective, it must also be affordable. This model considers relatively modest levels of PrEP MSM coverage of 5% (low) and 20% (high), respectively – and for each of these levels of coverage looks at the impact on cost-effectiveness of different levels of prioritization. The cost of ART is not taken into account in the calculation of costs averted.

Practically all the PrEP interventions considered fall below the WHO threshold of cost-effectiveness (3 times per cap. GDP per DALY averted), but a low coverage and high prioritization intervention would meet even the more stringent World Bank cost-effectiveness threshold of <$500 per DALY averted.  Such an intervention has a cost-effectiveness that is comparable with the cost-effectiveness of STI treatment interventions, even if it is less cost-effective that voluntary counselling and testing and condom provision (Aldridge, Miranda et al.: http://researchonline.lshtm.ac.uk/4538/1/1471-2458-9-352.pdf).

Such considerations may one day prove very useful to governments seeking to make policy decisions on resource allocation that are in line with UNAIDS guidelines, which stress both the importance of cost-effectiveness and the potential of prioritizing high-risk groups (http://www.unaids.org/en/media/unaids/contentassets/documents/unaidspublication/2011/JC2216_WorldAIDSday_report_2011_en.pdf).

The general message of the paper seems to be that PrEP is no panacea, but could in future make a contribution as part of a combination package even in limited resource settings, especially if policy makers consider the option of a low coverage narrowly focussed on high-risk groups. Yet, for the present, the feasibility of PrEP as a future public health intervention seems fraught with uncertainty.  Firstly its effectiveness  is hard to determine where levels of adherence are also uncertain  (http://sti.bmj.com/content/87/Suppl_1/A1.4.abstract?sid=2ffb6e56-e489-49f5-8ea0-eb291723644f).  Secondly, there is  the impact of risk compensation behaviour (http://sti.bmj.com/content/88/4/258.abstract?sid=2ffb6e56-e489-49f5-8ea0-eb291723644f), and of pill sharing (http://sti.bmj.com/content/87/Suppl_1/A35.2.abstract?sid=2ffb6e56-e489-49f5-8ea0-eb291723644f).  Both of these would be likely to ensue with the attendant risks of heightened drug resistance, where PrEP is deployed in limited-resource settings.

One area in which the recent wide scale deployment of ARV could have the greatest impact is that of mother-to-child HIV transmission (PMTCT).  The UNAIDS goal is for total elimination of PMTCT by 2015 – though this is scarcely feasible (STI 2010:86 Suppl.II – especially M. Mahy et al. (http://sti.bmj.com/content/86/Suppl_2/ii48.full?sid=50f44956-1266-4dd0-bed1-08ebd39bb60a)).  A component of the PMTCT risk is that of breastfeeding – and a complicating factor here is the need to balance the dangers of PMTCT against the protective effects of breastfeeding in low- medium- resource settings against many of the causes of infant mortality.

Results of the 48-week follow up of the Breastfeeding, Anti-retrovirals and Nutrition (BAN) randomized control trial, conducted 2004-2010 on 2639 breast-feeding HIV+ mothers in Malawi, published in last month’s The Lancet (http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(12)60321-3/fulltext) indicate some of the complexities involved in making recommendations for resource-poor settings – let alone estimating what it would take to eliminate PMTCT (see above: M. Mahy et al.).

This study is particularly interesting in view of the influence of early findings of this trial in bringing about important changes in the WHO guidelines (2010) on which projections for HIV reduction (STI 2010:86 Suppl.II) were based.  Broadly, changes included: 1. an extension of ART to cover six-months of exclusive breastfeeding; 2. dropping of the earlier (2006) recommendations concerning rapid weaning and no mixed feeding (http://www.avert.org/hiv-breastfeeding.htm).  Since then (2011) the Kesho Bora study, evaluating of maternal HAART (during pregnancy and breastfeeding) against single dose infant nevirapine, estimated the relative benefit of HAART at  43% (http://sti.bmj.com/content/88/4/313.full?sid=e656a189-e88b-46c9-859f-f719ff4ebf3b.

What more do we learn from the details of the 48-week follow-up in last month’s The Lancet?  The relative advantage of ART until weaning at 28 weeks over no ART (4% infant sero-conversion as against 7%) is comparable to earlier findings.  A surprise, however, is that 30% of sero-conversions took place after the 28 weeks at which breastfeeding was supposed to have stopped (9 in the maternal ART; 13 in the infant nevirapine; 6 in the control group).  This brings home the difficulty of making recommendations that can be applied in contexts where cultural expectations (e.g. mixed feeding or wet nursing of infants, see http://www.time.com/time/health/article/0,8599,1878917,00.html), social pressures or practical constraints must vary widely. The trial also gives details of neonatal and infant morbidity and mortality from any cause until 48 weeks.  Mortality stood at 3% for all groups.

In summary, the reduction of sero-conversion by 3% with ART prophylaxis is appreciable but not overwhelming. It has to be set in the context of high levels of infant mortality from causes (e.g. diarrhoea) from which breastfeeding offers considerable protection, and an evident difficulty of complying with recommendations for weaning in low-income settings (which, there is every reason to suppose, will be replicated in non-trial situations).  This tends to endorse the thinking behind the 2010 guidelines: namely, that ART prophylaxis is strongly to be recommended for the duration of breast-feeding, but that rapid weaning, and no subsequent mixed feeding, is unlikely to confer any health advantage.

Mary Mahy, Nathan Shaffer et al., “What will it take to achieve virtual elimination of mother-to-child transmission of HIV? An assessment of current progress and future needs”, STI 2010:86 Suppl. II

D. J. Jamieson, C. Van der Horst et al., “Maternal and infant antiretroviral regimens to prevent postnatal HIV-1 transmission: 48-week follow-up of the BAN randomised controlled trial”, The Lancet, online pub., 26^th April 2012

M. Desai, “Maternal HAART during breastfeeding confers a 43% RR reduction in mother-to-child transmission of HIV-1”, Clinical Round-up, STI 2011:87:3

One potential approach to HIV prevention attracting much current interest is the use of “pre-exposure prophylaxis” (PrEP). This involves taking anti-retroviral drugs (ARV) before exposure to HIV in order to reduce the risk of transmission. In parallel with clinical trials currently underway to establish the efficacy of PrEP, an important body of research is investigating its acceptability for potential users. A recent study of PrEP acceptability among “at risk” populations in Lima, Peru (female sexual workers (FSM); men who have sex with men (MSM); male-to-female transgendered persons (TG)) complements on-going clinical trials of Truvada as PrEP in the city. It also claims to be the first PrEP acceptability study undertaken outside the US.
The study recruited 45 participants, and involved the generation of material for analysis through focus groups, as well as “conjoint analysis”. The latter technique, more familiar in the context of marketing research, confronts participants with a number of hypothetical products (8 in this study) composed of various combinations of pre-determined attributes. So, the “products” of this analysis were distinguished on the basis of the attribute of “cost” as expensive ($250) or cheap ($10); on the basis of the attribute of efficacity as more efficacious (95%) or less ($70) – and so forth for 7 pre-determined key attributes. Participants were invited to score 8 products, and subsequent analysis allows us to establish the impact of attributes on acceptability.
It is no surprise that cost, efficacy, and the presence/absence of side-effects (in that order) prove to be the most important attributes. The interest of this study may lie elsewhere. First there is the issue of potential up-take. While the highest acceptability score of 83% suggests the potential given the optimal drug, the authors suggest that the average acceptability of 53% may provide a more realistic indication of the likely take-up of any real-life product. In other words, PrEP does not look like a panacea – if we were tempted to expect one. Secondly, findings of the study point to the potential importance of sensitivity to target groups and micro-environments in the roll-out any PrEP intervention. For example, the reactions of participants tend to reveal marked disparities between the FSW participants and the MSM/TG participants in their view of the feasibility of “daily dosing”, or of the likelihood of increased risk-taking behaviour as a direct result of PrEP. There are also indications that very local environmental factors can impact on user preferences, e.g. the prevalence of small family-operated drugstores in Lima on the preference for PrEP being dispensed by health-care professionals rather than pharmacists.

J.T. Galea et al., “Acceptability of pre-exposure prophylasis as an HIV prevention strategy: barriers and facilitators to pre-exposure prophylaxis uptake among at-risk Peruvian populations”, International Journal of STD & AIDS 2011: 22: 256-262

http://ijsa.rsmjournals.com/cgi/content/abstract/22/5/256