The FEM-PrEP study

Just five months after the announcement of iPrEx’s promising results, a large trial of pre-exposure prophylaxis in women was closed in April 2011. The investigators concluded that even if it continued for its originally planned duration, the FEM-PrEP trial was highly unlikely to show a significant protective effect of Truvada against HIV infection in this population.

However the trial organisers, Family Health International, stressed that based on the interim trial results, it could not be determined whether or not Truvada was effective in the prevention of HIV infection in women. Further analyses (ongoing at the time of writing) are required before any conclusions can be made.

The FEM-PrEP study recruited 1951 HIV-negative women aged 18 to 35 at risk of HIV infection in South Africa, Kenya and Tanzania. Women were randomised to receive daily oral Truvada (tenofovir and FTC in one pill) or placebo. On average, participants took the study drug for 12 months and attended study visits for 14 months.

Among all trial participants the approximate annual rate of new HIV infections was 5% per year, and a total of 56 new infections had been recorded up to February 2011. These were equally distributed between the Truvada and placebo groups (28 in each arm).

The decision to halt recruitment early indicated that the Independent Data Monitoring Committee, which scrutinised the trial, came to the conclusion that even if another 2000 women were to be enrolled onto the trial and followed for a year each (as planned), it was highly unlikely that any protective effect of Truvada that might emerge would be enough to demonstrate a statistically significant benefit overall.

This might be because the magnitude of effect seen in the first 1951 women was somewhat too small to be statistically significant, or because there was no substantial difference, and more analysis of the trial data will be needed.

In particular it is unclear whether the women in the two study arms had a similar total exposure to their assigned study medication: an equal number of infections at month 12 may not translate into an equal rate of infection, because women in one group may have a higher number of total months covered by study medication.

The investigators said they planned to concentrate further analyses on two linked areas: a detailed investigation of adherence, and of drug levels in women taking Truvada.

Overall adherence (based on participant’s self-report) was 95%, but at the time of writing, there are no data on differences in adherence between study arms, differences in adherence over time or differences in adherence according to the adherence measures used.

Monitoring will look into whether drug levels in blood are matched by levels in the vagina. Studies have found that oral tenofovir tends to reach higher levels in rectal tissues than in vaginal tissue. Although a study presented at the 2011 Conference on Retroviruses and Opportunistic Infections (CROI) found adequate levels in vaginal tissue, presenter Craig Hendrix warned that “what is ‘enough’ for prevention is yet to be defined”.¹

There were significantly more pregnancies in women taking Truvada than placebo. This could explain the lack of efficacy. Pregnancy excluded women from the trial, which would mean women on Truvada spent more time not taking pills, on average, than women on placebo. Any protective effect of Truvada may be so attenuated that it was not demonstrably better than placebo.

The more troubling possibility is that Truvada had an unexpected drug interaction with the women’s contraception (all participants were provided with free contraceptives, in most cases injectable or oral hormonal contraceptives). If tenofovir or FTC altered levels of contraceptive, this could also alter the thickness of the vaginal mucous membrane and thus alter a woman’s vulnerability to HIV. However, no such interaction has been described.

The third possibility is that side-effects amongst women taking Truvada might have caused their adherence to be poorer than in women taking placebo; the iPrEx study found a higher rate of nausea in men taking Truvada in the first month.

The result may indicate that antiretroviral drugs may show different levels of effective prevention in different populations, social groups and locations. This may be driven by differences in adherence or sexual behaviour. It therefore underlines the importance of testing new prevention technologies in a variety of different populations.