A panel discussion at the Microbicides 2006 Conference focused on how STIs could affect microbicide research, looking at modelling projections of how a microbicide with anti-STI activity could affect STI transmission, the need to monitor anti-HSV-2 in microbicide studies and how complicated it may be to measure a microbicide's activity against an STI.
Most of the microbicides described currently in clinical efficacy studies have activity in laboratory or animal models against other sexually transmitted infections (STIs) besides HIV. Although this could potentially increase the benefit to study participants randomised to the microbicide arm (and for users of the product if it ever reaches the market), it also increases the complexity of conducting and interpreting the ongoing clinical trials.
STIs could affect HIV microbicide trials in a number of ways. For instance, a number of epidemiological studies have shown that STIs, especially, gonorrhoea, chlamydia, bacterial vaginosis (BV) and trichomoniasis, and perhaps most importantly, Herpes Simplex Virus-type 2 (HSV-2) can increase the risk of HIV. Preventing STI transmission with a microbicide could in turn impact HIV transmission.
In fact, according to modelling projections presented by Anna Foss, a research fellow at the London School of Hygiene and Tropical Medicine, even a microbicide with relatively low efficacy against STI transmission and no direct activity against HIV might reduce HIV transmission in areas where there is a low to medium prevalence of the STI.
The projections were made with a new model that reflects how microbicide-related changes in gonorrhoea transmission could impact HIV transmission. According to this model, having gonorrhoea doubles the risk of HIV transmission. The model was then fitted with actual behavioural and biological data drawn from Cotonou, Benin. For example, in Benin, and in the model, condoms are used consistently 56% of the time and there is a 41% HIV prevalence among sex workers and so on. The model looks at three different background gonorrhoea prevalence scenarios low (SW 8.6%, clients 2.3%), medium (sex workers 21%, clients 5.4%) and high (sex workers 40%, clients 9.7%).
Using this model, Ms Foss and colleagues made projections of what would happen if, with all other things such as condom use and risk-taking behaviour, remaining constant, microbicides with varying efficacies against gonorrhoea and HIV are used in at least 50% of the non-condom protected sex acts. Some of her key findings include:
- If a microbicide with 50% anti-HIV and anti-gonorrhoea efficacy were used, the relative reduction in HIV-risk among sex workers would be about 24-33% across all gonorrhoea-prevalence scenarios.
- A microbicide without any HIV-efficacy but a 50% efficacy against gonorrhoea could reduce HIV risk by up to 12%.
- When there is a high background gonorrhoea prevalence, it would be hard for a moderately effective anti-gonorrhoea microbicide to have an impact on either gonorrhoea or HIV transmission, since the risk of gonorrhoea is very high. In this context, there would be a substantial additional benefit of reducing HIV risk if the microbicide was also highly effective against gonorrhoea.
This model only uses one STI but in the real world, the number of prevalent STIs is greater and their interactions are much more complex (people at high risk for STIs often seek out treatment as well). Nevertheless, these projections suggest that the impact that microbicides have in the ongoing trials could be context-specific and will vary from site to site, and that, to some extent, it will be influenced by the prevalence of various STIs and the activity of microbicides against them. It also suggests that the microbicides currently in clinical efficacy trials could have benefits which the ART-based microbicides will not have since they usually have little or no activity against STIs.
It also suggests that there may be some complicated data coming out of the ongoing microbicide studies. Notably, modelling projections show that even if you have a microbicide that is 50% effective against HIV, the actual efficacy in the trial arm may be less depending upon the background prevalence and incidence of STIs that upregulate transmission of HIV. Thus, STIs have the ability to act as a “confounder” in the interpretation of microbicide trial results.
“Clinical trials should prospectively assess biological factors that could modify or confound the effect of candidate microbicides on HIV acquisition” said Professor Connie Celum of the University of Washington. “Secondly — and we need to think about this with each individual microbicide — is to look at STIs that could represent primary or secondary efficacy endpoints.”
Prof. Celum is the primary investigator on two studies looking at acyclovir’s impact on HIV transmission, and so she focused primarily on HSV-2. Like HIV, HSV-2 also affects women disproportionately, especially in Africa where the HSV-2 prevalence in some settings is over 60%. In fact, a recent systematic review of 18 longitudinal studies looking at HSV-2 and HIV acquisition has generated estimates that as many as 38-69% of new HIV infections in women could be in part due to HSV-2 infection (Freeman AIDS 2006). HSV-2 could also be involved in the high prevalence of bacterial vaginosis and other infections.
A microbicide with anti-HSV-2 efficacy could offer an enormous benefit to women and would be likely to increase whatever anti-HIV efficacy the product has. Fortunately, all of the gel microbicides in clinical efficacy studies have in vitro activity against HSV-2 (though some are more potent than others) but moving from in vitro to clinical assessments might work better in theory than in practice.
According to Prof. Celum, “where we are looking at HIV as the primary efficacy endpoint, we need to assess HSV-2 due to potential confounding of the effect of candidate microbicides on HIV acquisition. To do that, we need to determine HSV-2 serostatus at enrolment. For those who are HSV-2 negative on enrolment, we need to assess HSV-2 serostatus at the end of the trial, and look back through time at when they acquired it through stored samples."
Looking at microbicides' direct activity in preventing HSV-2 acquisition as a primary or secondary efficacy endpoint in an HIV-microbicide study could be complicated, however.
“I think we need to think really critically about this,” said Prof. Celum, “because if trials enrol HIV-negative women — that’s our primary enrolment criteria — you have to anticipate that you’ll have about a 50% HSV-2 seroprevalence, so if you want to have HSV-2 as a primary endpoint, you’ll have to either change your inclusion criteria or have a higher acquisition rate [of HSV-2] to give you the power to look at this.”
“If you do want to look at this as a primary endpoint, and enrol, therefore, HIV-negative, HSV-2 negative women, you’re going to increase your screen-to-enrolled ratio by at least two, probably more like three, and you’re going to decrease the HIV incidence and probably increase the cost of the trial.”
It might also be possible to look for reduction in HSV-2 shedding in HSV-2 positive women, but this is more complex because the gel could possibly interfere with HSV DNA PCR measurements (see below) and “HSV-2 PCR isn’t really well-validated as a surrogate marker for HSV-2 transmission,” Prof. Celum said.
“Everything has a cost, particularly in large phase 2/3 clinic trials. We need to be more strategic in the use of resources and weigh carefully design issues, selection criteria, power, lab and clinical costs,” said Prof. Celum.
Challenges at the reference lab
One of the issues to which Prof. Celum alluded, is that the use of laboratory STI tests could be altered by the microbicide. According to Dr Khatija Ahmed, primary investigator for the Carraguard study at the University of Limpopo site, measuring the effects of microbicides against STIs in the laboratory comes with its own set of challenges.
“The vagina is a very complex environment,” she began. Any microbicide that is added to the vagina could affect its chemistry and potentially render the usual techniques used to measure STIs obsolete." These challenges are not insurmountable, but “we need to get some standardised techniques and models that we can use in all the microbicide research that is ongoing,” she said.
She listed several concerns about measuring a microbicide's effect on STI incidence:
**The method of action against STIs differ from one microbicide to another.
**Many patients have asymptomatic infections, which raises the issue of when to test.
**What sample collection devices should be used (cotton swabs/dacron)?
**What types of samples (swabs, urine, or washings)?
**What methodology is used prior to collection?
**How are the specimens transported to the laboratory (is refrigeration necessary?)
**Standard testing methods differ for STIs (for example, culture or serology or molecular techniques).
**How might the microbicide or the gel that it is formulated interfere with the test?
“When you put microbicide into the vagina, there is always some residual microbicide that is left. The quantity of this microbicide depends on various factors, including the interval between insertion and specimen collection. If the participant has microbicide inserted 24 hours before or 8 hours before, there is going to be an interference with the test; the volume inserted — is it 2 ml is it 4 ml — all those things could interfere,” she said. Other factors include the viscosity of the microbicide, how it is inserted (the type of applicator), the number of insertions (is there a cumulative effect?), the nature of the specimen being collected and the type of test that needs to be performed.
For example, studies have shown that Carraguard, BufferGel, Cellulose Sulfate, Acidform and even placebo gels can interfere with some nucleic-acid amplification techniques (such as PCR-based tests). This could affect the choice of test used, or necessitate modifications to the test procedure (such as changing the transport medium, collection device or adding a wash step).
The STI testing techniques have now been modified in the phase III Carraguard and Cellulose Sulfate studies, but she is unsure of what work has been done in adjusting the laboratory methodology in other ongoing efficacy studies.
“Prior to testing of STIs in microbicide trials, in-vitro and in-vivo testing must be done on each individual microbicide and the placebo to determine the optimal test and whether modifications are required to the procedure. If this is not done, the secondary endpoint data cannot be validated,” she concluded.
Ahmed K. Laboratory testing of sexually transmitted infections in microbicide clinical trials. Microbicides 2006 Conference, Cape Town, panel discussion talk, 2006.
Celum C. Microbicides and STDs: role of HSV-2. Microbicides 2006 Conference, Cape Town, panel discussion talk, 2006.
Foss A et al. The importance of microbicide STI-efficacy in reducing HIV-risk: model projections. Microbicides 2006 Conference, Cape Town, OB19, 2006.