An effective vaccine for HIV continues to elude scientists 25 years after the first appearance of AIDS. In its absence, the development of vaginal microbicides has increasingly been viewed as a priority in efforts to contain the AIDS epidemic, especially among women [1–3]. The first-generation microbicides may only be partially effective; efficacy can be as low as 35% . However, mathematical modeling has shown that even a microbicide with low efficacy can reduce HIV exposure risk substantially, as long as condom migration (called condom replacement by Smith et al.), which refers to a decrease in condom usage in response to the availability of microbicide, is low. Define the condom replacement rate to be the percentage of partnerships that stop using condoms for every 1% of partnerships that use microbicides. The analysis by Foss et al. showed that if microbicides are half as efficacious as condoms then HIV incidence does not change if the condom replacement rate is 0.5 but it rises if the decrease in condom usage exceeds this rate. For microbicides that have the same efficacy as condoms, incidence level does not change if the condom replacement rate is 1. The same implications follow from the models in  and .
However, these previous analyses did not take into consideration the fact that the effectiveness of a microbicide in preventing male-to-female HIV transmission may differ from its effectiveness in blocking transmission in the other direction, and that the effect of condom replacement on HIV prevalence over time depends critically on microbicide efficacy in preventing transmission in both directions. In particular, these previous studies significantly underestimated the effect of condom replacement if microbicide efficacy in preventing transmission in one direction is substantially lower than its efficacy in preventing transmission in the other direction. For example, the introduction of microbicides that are just as efficacious as condoms in preventing male-to-female transmission would unambiguously increase the endemic HIV prevalence level in women if the condom replacement rate is as low as 0.67 when the efficacy of microbicides in blocking female-to-male transmission is substantially lower than that of condoms (Fig. 1c).
The effects of condom replacement when microbicide efficacy varies with the direction of transmission can be described using a mathematical model consisting of two differential equations that describe how the disease prevalence among men (PM) and women (PW), respectively, change over time:
where the female-to-male transmission probability of the disease per partnership is βM, the male-to-female transmission probability per partnership is βW, the length of time an individual is sexually active in the population is 1/δ, a behavioral parameter describing men's level of risk exposure in a partnership is rM and the analogous behavioral parameter for women is rW, and c is the rate with which individuals change partners. The behavioral parameters rM and rW depend on the usage levels of condoms and microbicides, as well as their efficacies in preventing transmission of the disease. Assuming that condoms are equally effective in blocking male-to-female and female-to-male transmissions and that the effects of condoms and microbicides are independent,
where εc denotes the efficacy of condoms, εM is the efficacy of microbicides in blocking female-to-male transmission, and εW is the microbicide efficacy in blocking transmission in the other direction. In addition, αn is the fraction of partnerships in which neither microbicides nor condoms are used, αc is the fraction of partnerships in which only condoms are used, αM is the fraction of partnerships in which only microbicides are used, and αb is the fraction of partnerships in which both forms of protection are used. Assuming that αb = 0, the steady-state endemic prevalence among women increases (decreases) subsequent to the introduction of microbicides if
is positive (negative), where αBc and αAc, respectively, denote the fraction of partnerships using condoms before and after microbicides became available. The parameter ρ gives the condom replacement rate [(αBc − αAc)/αM]. The steady-state prevalence in women does not change if the above expression is 0. Notice from this expression that, assuming microbicides are just as effective as condoms in blocking male-to-female transmission, a condom replacement rate of 1 will not change the steady-state prevalence in women if εW = εM, but it will unambiguously increase prevalence among women as long as εW > εM. The same conclusion applies to the steady-state endemic prevalence in men (formula not shown). The prevalence level among women over time subsequent to the introduction of microbicides is shown in Fig. 1 for different values of εM and αM, assuming that αBc = 0.4, αAc = 0.2, and that microbicides are just as effective as condoms in blocking male-to-female transmission. The initial prevalence level in the figure is the endemic steady-state prevalence level in the absence of microbicides. As shown in the figure, a condom replacement rate of 1 will unambiguously increase the prevalence level in women if microbicides are not as effective as condoms in preventing female-to-male transmission. In fact, for the parameter values used, if microbicides are completely ineffective in blocking female-to-male transmission, then a condom replacement rate that is as low as 0.67 can increase prevalence among women. If microbicides are half as effective as condoms in blocking female-to-male transmissions, then a condom replacement rate of 0.8 would cause the steady-state prevalence level to rise slightly in women. If microbicide efficacy in blocking male-to-female transmission is half that of condoms and if microbicides have no effect on the infectiousness of infected women, then a condom replacement rate of 0.3 would be enough to increase the steady-state prevalence levels in both men and women (data not shown).
It has been noted that microbicides are likely to be most effective at the population level if they can prevent transmission to women as well as female-to-male transmission . A corollary to this claim is that even if microbicides are highly efficacious in blocking transmission to women the effect of condom replacement will be significant if microbicides have limited ability to reduce the infectiousness of women who are already infected. Therefore, mathematical models that focus only on the impact of microbicides on the disease incidence in women, or do not take into account how differences in microbicide efficacy by direction of transmission can affect prevalence levels in men and women over time, will underestimate the impact of condom replacement and provide an overly optimistic projection of the impact of microbicides on HIV prevalence.
I would like to thank Ellen M. Palmer for her comments and suggestions, as well as for numerous discussions concerning the topic of this article.
1. Weber J, Desai K, Darbyshire J. The development of vaginal microbicides for the prevention of HIV transmission. PLoS Med 2005; 2:0392–0395.
2. Quinn T, Overbaugh J. HIV/AIDS in women: an expanding epidemic. Science 2005; 308:1582–1583.
3. Minnis A, Padian N. Effectiveness of female controlled barrier methods in preventing sexually transmitted infections and HIV: current evidence and future research directions. Sex Transm Infect 2005; 81:193–200.
4. Foss A, Vickerman P, Heise L, Watts C. Shifts in condom use following microbicide introduction: should we be concerned? AIDS 2003; 17:1227–1237.
5. Smith R, Bodine E, Wilson D, Blower S. Evaluating the potential impact of vaginal microbicides to reduce the risk of acquiring HIV in female sex workers. AIDS 2005; 19:413–421.
6. Karmon E, Potts M, Getz W, Microbicides. HIV: help or hindrance? J Acquir Immun Defic Syndr 2003; 34:71–75.
7. McCormack S, Hayes R, Lacey C, Johnson A. Microbicides in HIV prevention. Br Med J 2001; 322:410–413.
8. Pinkerton S, Abramson P. Effectiveness of condoms in preventing HIV transmission. Soc Sci Med 1997; 44:1303–1312.
9. Hearst N, Chen S. Condom promotion for AIDS prevention in the developing world: is it working? Stud Family Plann 2004; 35:39–47.