Microbicides could prevent more infections in men than in women

Microbicides intended for vaginal use by women that contain non-nucleoside (NNRTI) antiretroviral drugs could end up preventing more infections in men than women, a mathematical model devised by biostatisticians from universities in Sydney, Montreal, Philadelphia and Los Angeles has found.

This paradoxical outcome might arise, the researchers say, because use of microbicides containing NNRTIs could result in a significant number of NNRTI-resistant HIV infections, especially if there is significant absorption into the body of women using them. Furthermore; because drug-resistant virus is often less fit than wild-type virus, female-to-male transmission would be less likely to occur. Men would also benefit more directly if the microbicide was ‘bi-directional’ and provided some protection for a man having sex with a women with HIV who was using it.

Microbicides containing antiretrovirals are being developed as a new generation of products following the failure of earlier types of microbicides. Antiretrovirals not in use for treatment of people with HIV are being investigated for their preventive potential when delivered within a gel for vaginal or rectal use. However, there are concerns that if the antiretroviral drug is absorbed into the body at low levels, the drug could give rise to resistance in people who are already infected with HIV. This is especially true for NNRTIs, where resistance to the currently licensed drugs in this class is quick to develop.

The authors of the study developed this model because, they say, they are concerned that the randomised placebo-controlled trials of microbicides containing antiretrovirals that are ongoing or about to start – see this report – will be unable to establish whether use of the microbicide by women who unwittingly have HIV would generate significant amounts of resistance. That is because the studies will test women for HIV every month, reducing the risk that they will be exposed to sub-optimal drug levels and so develop resistance.

But in the real world a microbicide that is systemically absorbed could lead to significant levels of resistance in women who use it if they are already infected with HIV when they begin to use the product – or become infected despite using it.

Because of this, the authors say, tests of systemic absorption should be conducted in all microbicide efficacy trials and “careful and regular population-based monitoring” of NNRTI resistance should be conducted if microbicides are licensed and used in public health programmes.

Results

The surprising conclusions derived from the modelling exercise were originally aired at the Microbicides 2008 conference back in February but the full results, published recently in the Proceedings of the National Academy of Sciences, explain the model behind the findings, the assumptions built into it, and the public health implications of its predictions.

The model finds, for instance, that if a microbicide were ‘high-risk’ and highly systemically absorbed (meaning that clinically significant levels of NNRTI would be reached in the blood 50-90% of the time women used them), men in the general population would benefit more than women in terms of infections prevented if the microbicide stopped less than half of male-to-female HIV transmissions; if women’s adherence was less than 57%; and, even if women were 100% adherent, if the microbicide’s efficacy for women was less than 45%.

However even if a microbicide is ‘low-risk’ and poorly absorbed (meaning that clinically significant levels of NNRTI would be reached in the blood only 1-3% of the time women used them), men would benefit more than women if adherence was very low in women (less than 10%) or efficacy was low (less than 18%). If women were 100% adherent to a low-risk microbicide men would still benefit if it were less than 55% effective and, if they only used it half the time, less than 45% effective.

However if a microbicide containing antiretrovirals were both very efficacious (as they are expected to be) and poorly absorbed, then women would benefit more than men. For instance, if a microbicide stopped two-thirds of male-to-female HIV transmissions, then women would benefit more than men as long as they used the microbicide at least once in four to five episodes of sexual intercourse.

The mathematical model used predicts results for two time-frames; firstly, during an actual randomised controlled trial (RCT) like the International Partnership for Microbicides one; and secondly, the public health effects in terms of HIV infections ten years after such a microbicide is licensed. For the first scenario the authors entered into the model 15 different variables including transmission probabilities, condom use, number of sex acts and partners, microbicide adherence rates, the ‘window period’ likely after an HIV test, and the degree of systemic absorption of the microbicide.

In the second scenario, they added 13 more variables including the degree to which the microbicide protected men, initial HIV prevalence in the population, the average rate at which drug resistance develops in the population (ranging from six months to never) and the degree to which the presence of NNRTI-resistance mutations affect viral fitness and transmissibility.

For this important last variable the authors assumed that drug resistant virus would never be transmitted more than 50% as often as wild-type virus. While studies of NNRTI-resistant viruses have shown that some are as ‘fit’ as wild-type virus, other studies have shown that even in these cases, drug resistant virus is usually quickly ‘outgrown’ by wild-type and is therefore less likely to be transmitted.

The model finds that during the RCT an average of 36 cases of HIV (out of 10,000 participants) would be prevented by both high-and low-risk microbicides but that whereas only one case of HIV would be prevented for every one case of resistant HIV in a high-risk microbicide, 36 cases would be prevented with a low-risk microbicide.

After ten years in a public health setting, the paradoxical effects of viral fitness would mean that on average more infections would be prevented with a high-risk microbicide than a low-risk one; an 11% reduction in women and a 14% reduction in men. But even with a low-risk microbicide, men would benefit slightly more than women with an 8% reduction in HIV cases in men and a 7% one in women. There would, however, be ten times as many cases of drug resistant virus being transmitted from women to men than vice versa.

The average number of women with HIV who had NNRTI mutations in the population after ten years would be 22.4% with a high-risk microbicide and 4.8% with a low-risk one. However the authors point out that treatments for people with NNRTI resistance are becoming more available even in the developing world and that by the time of this scenario, the use of ‘second-generation’ NNRTIs that are effective against one or two NNRTI mutations might be widespread.