RCT trial demonstrates: disappointing PrEP study outcomes probably down to poor adherence

The disappointingly varied results of recent clinical trials of antiretroviral prophylaxis for HIV prevention (PrEP) have not altogether dampened the interest in this intervention.  Contributions to STI journal include papers evaluating the potential acceptability of PrEP in various settings (see: Aghaizu & Nardone; Holt & De Wit).  Clearly, there remains a keen interest in the potential of this intervention, even though it remains very unclear what part it might have to play in future public health programs.

But why have the PrEP studies produced such varied results?  For example, levels of risk-reduction obtained in recent trials have been estimated as follows: CAPRISA 004 – 39%; iPrEX – 44%; CDC TDF2 – 62%; Partners-PrEP – 75%; FEM-PrEP – 0%; VOICE – 0%.  Poor adherence is the probably the most frequently alleged cause of this variation.  However, Hendrix & Bumpus, in an paper discussed by this blog (sti.blog/Hendrix&Bumpus, argue, on the basis of their pharmacokinetic study, that vaginal concentrations of the active drug achieved by the gel (at 130-170 times those achieved by the oral tablet) are at a level that ought, at least where the gel is used, to be proof against even relatively high levels of non-adherence – so the cause must be something else (but what?).

Now, however, Haberer & Bangsberg, reporting the results of a sub-study on adherence that took place within Partners-PrEP, claim to have evidence that points back in the direction of the more frequently alleged cause.  It suggests that PrEP achieves high-levels of biologic efficacy, so can be an effective intervention – in conditions replicating those of the sub-study.  But it also raises important questions about the kind of counselling and supervision that might be needed to make PrEP effective, if it ever became a major element of government public health HIV prevention programs.

Participants in the sub-study were a convenience sample of uninfected partners of HIV sero-discordant relationships from the three arms arm of Partners PrEP, totalling 1,147 partners, or 66% of the (initially) uninfected partners involved in Partners PrEP.  The sub-study participants had their adherence monitored both by unannounced pill-counts (UPC) and by electronic monitoring of pill-bottle opening (MEMS); where adherence was ≤80%, participants were enrolled in an intensive “adherence intervention”.  As a result, adherence in the sub-study seems to have been maintained at a remarkably high level – median 99.1% adherence as monitored by UPC, and 97.2% as monitored by MEMS.  It is interesting, however, and maybe significant as regards future PrEP implementation, that ≤80% adherence was observed at some point for as many as 25.8% of participants.

The headline finding of the sub-study is that of the 14 sero-conversions that took place over the duration of the study, all were from the placebo arm of the study (which was discontinued in July 2011).

The message seems to be that PrEP has biologic efficacy. Its effectiveness as a public health intervention, however, may depend on the feasibility of sustaining a level of support and counselling that comes anywhere near what was provided by the sub-study.  H&B recommend further research in order to identify key adherence counselling messages, establish counselling practices that are achievable in a “real-world” setting, and determine cost-effectiveness.

(Visited 61 times, 1 visits today)