Drug Levels Are Key to Effectiveness of Tenofovir Vaginal Gel for HIV Prevention

alt

Although the VOICE study failed to show an overall protective effect for tenofovir vaginal gel, further analysis suggests that women who had detectable blood drug levels were at lower risk of HIV infection, according to an analysis described in the June 29 advance edition of Journal of Infectious Diseases. Another recent analysis from the CAPRISA 004 trial showed that higher levels of tenofovir in the genital tract were associated with greater reduction in risk.

Oral pre-exposure prophylaxis (PrEP) using tenofovir/emtricitabine (Truvada) is highly effective at preventing HIV infection among people who take it as directed, especially in studies of men who have sex with men and heterosexual couples. But studies of prevention for young women have been disappointing, and new methods that are easier to use are needed.

VOICE Analysis

James Dai from Fred Hutchinson Cancer Research Center in Seattle and colleagues performed further statistical analysis of data from the VOICE trial, looking at the effect of various confounding factors that may have influenced the results.

The VOICE study included more than 5000 women in South Africa, Uganda, and Zimbabwe. They were randomly assigned to use daily Truvada, daily tenofovir alone, a 1% tenofovir vaginal gel microbicide applied daily, or placebo pills or gel.

As previously reported, none of these methods led to a significant reduction in the risk of HIV infection in an intent-to-treat analysis, which includes everyone assigned to use a method whether they did so or not. Although the participants reported good adherence, tests of blood tenofovir levels told a different story. A qualitative analysis explored some of the reasons why most women did not use the pills or gel as directed.

Dai and his team developed a regression analysis strategy to evaluate whether conventional covariate adjustment had succeeded in removing the effect of confounding variables such as age and frequency of unprotected sex.

After adjusting for predictors of HIV risk at baseline, the researchers determined that confounding effects were nearly eliminated when comparing placebo recipients to women who had blood plasma drug levels indicating tenofovir gel adherence -- either ever having a detectable tenofovir level or having a detectable level at month 3 of the study.

This suggests that the observed reduction in new HIV infections among the adherent women in VOICE was attributable to a real prevention effect of the gel, and not due to selection bias or chance alone.

While they used the VOICE study as an example of the method, they suggested that this approach should be generally applicable to other placebo-controlled PrEP trials.

CAPRISA 004 Analysis

A related analysis, described in the July 1 Journal of Acquired Immune Deficiency Syndromes, looked at pharmacokinetic and pharmacodynamic characteristics of tenofovir gel in relation to its protective effect against HIV.

Like the VOICE study, the CAPRISA 004 trial tested a 1% tenofovir vaginal gel, but in this trial it was applied before and after sex rather than daily. As previously reported, the gel reduced the risk of HIV infection by 39% overall and by 54% when used consistently.

Salim Abdool Karim from Columbia University's Mailman School of Public Health, and colleagues compared tenofovir levels in blood plasma and cervical-vaginal fluid (CVF) from 34 women in the study who became infected with HIV and 302 randomly selected women who remained uninfected. In total, they analyzed 336 CVF samples, 55 plasma samples, and 23 paired cervical and vaginal tissue samples obtained from biopsies.

The researchers detected tenofovir in genital tract samples from 8 infected women (24%) and 119 uninfected women (39%), a difference that fell short of statistical significance. Among women with detectable genital tract tenofovir, the median CVF concentration was 97% lower in infected compared with uninfected women (476 vs 13,821 ng/mL), but again the difference was not significant. 

However, women who became infected were less than half as likely as uninfected women to have CVF tenofovir concentrations above 100 ng/mL (15% vs 33%; odds ratio [OR] 0.35), and this difference did reach statistical significance. The difference was even greater using a higher threshold of 1000 ng/mL (9% vs 26%, respectively; OR 0.27).

Turning to blood drug levels, none of the infected women and just 17% of the uninfected women had detectable plasma tenofovir. Tenofovir concentrations were below 1 ng/mL in all women.

A tenofovir concentration of >100 ng/mL in CVF was associated with 65% protection against HIV infection, whereas a >1000 ng/mL concentration correlated with 76% protection, the study authors concluded. Protection was not observed in women with no detectable tenofovir in their genital tract, while the detection of any genital tenofovir (>1 ng/mL) was associated with 53% effectiveness.

"Tenofovir was seldom detected systemically [in blood plasma] in the CAPRISA 004 trial, confirming that intermittent topical application achieves limited and short-lived systemic tenofovir exposure," they added in their discussion. In contrast, biopsy assays showed that 30% of women had detectable tenofovir diphosphate (the active form of the drug) in their vaginal and cervical tissue samples, even though biopsies were done a median of 43 days after the last gel application.

Unfortunately, as reported at this year's Conference on Retroviruses and Opportunistic Infections, a larger follow-on study of the same tenofovir vaginal gel used before and after sex (FACTS 001) saw no overall difference in HIV infection rates between women using the active gel versus placebo gel, again due to low adherence. But among the minority of women who did achieve high (>80%) adherence, the risk of infection fell by 57%.

Together, these studies confirm that good adherence to a prevention method is critical for protection against HIV. They also add further evidence that higher tenofovir levels or more consistent administration may be necessary to protect women from exposure via vaginal sex, compared to what is needed to protect gay men from exposure via anal sex.

7/9/15

References

JY Dai, CW Hendrix, BA Richardson, et al. Pharmacological Measures of Adherence and Risk of HIV Acquisition in the VOICE Study. Journal of Infectious Diseases. June 29, 2015 (Epub ahead of print).

ADM Kashuba, TN Gengiah, L Werner, SS Abdool Karim, et al. Genital Tenofovir Concentrations Correlate With Protection Against HIV Infection in the CAPRISA 004 Trial: Importance of Adherence for Microbicide Effectiveness. Journal of Acquired Immune Deficiency Syndromes 69(3):264-269. July 1, 2015.