In August 1995, the results of a landmark study demonstrated that community-based treatment of symptomatic sexually transmitted infections (STIs) could dramatically reduce HIV incidence.1 Strong biological and epidemiological evidence had long pointed to a critical role for STI control in HIV prevention.2 But the magnitude of the impact shown in this randomized controlled trial in the Mwanza region of Tanzania -- a 40 percent reduction in new HIV cases -- was striking.
Less than three years later, the findings from a second community-based randomized trial of the impact of STI treatment on HIV incidence delivered yet another surprise. This study, which offered treatment for common curable STIs to all adults in ten communities in Uganda's Rakai district, found that the intervention had reduced some STIs, but had had no impact on the rate of new HIV infections.3
These apparently contradictory results have understandably left many people confused about the role of STI control in HIV prevention. Many assumed that one of the studies must be wrong. Others argued that discrepancies in the findings are a result of distinct differences between the epidemics in the two study sites. The debate continues, as scientists, health care providers, HIV program managers and international donors seek to understand the implications of the Mwanza and Rakai studies.
Understanding Rakai
The eagerly awaited results of the Rakai trial came as a shock to many who had expected them to clarify the additional HIV prevention benefit of treating all STIs in a community, rather than just those in people who sought treatment for symptoms. In order to ensure coverage of asymptomatic STIs in these communities with high STI prevalence, the study had offered directly observed home-based antibiotic treatment to everyone 15 to 59 years of age, regardless of symptoms, clinical signs or STI treatment-seeking behavior.
Conducted by a team of U.S. and Ugandan researchers, the Rakai study involved more than 12,000 participants in ten clusters of four to seven villages. Every ten months participants in half the communities received single doses of three antibiotics, while those in the other half received a treatment for intestinal worms, an iron-folate tablet and a low-dose multivitamin. Pregnant women in both groups who tested positive for syphilis were treated. All had agreed to participate with the understanding that their communities had an equal chance of being assigned to the intervention or the control group.
Originally designed to provide five rounds of mass STI treatment, the trial was halted after three rounds when an analysis of the data showed conclusively that the intervention was having no effect on HIV incidence (the rate of new infections occurring during the study period). Although the levels of some STIs were significantly lower in the intervention group than in the control group, the incidence of HIV-1 was 1.5 per 100 person years in both groups.
Several explanations have been proposed for the discrepancy between these findings and the substantial reductions in new HIV cases that occurred in Mwanza as a result of improved STI treatment. One is that the Rakai intervention decreased the prevalence of some but not all the STIs that facilitate HIV transmission. After three rounds of mass treatment, only syphilis and trichomoniasis levels were significantly lower in the intervention group than in the control group. However, the authors of the Rakai study note that significant reductions in the prevalence of gonorrhea, chlamydial infection, trichomoniasis and bacterial vaginosis (BV) (an infection that is related to sexual activity through not sexually transmitted) were achieved among pregnant women enrolled in the study, with no effect on their acquisition of HIV.
Another possible confounding factor in Rakai was the high prevalence of two infections -- bacterial vaginosis (BV) and herpes -- for which there are no highly effective, affordable treatments. At baseline half of the women in the study had BV, which is difficult to cure and has been linked with increased risk of HIV. Moreover, 42 percent of genital ulcers in the Rakai study population were positive for herpes simplex virus-2 (HSV-2), which can be treated but not cured with expensive antiviral therapy that was not given in this study. Only 7 percent of the genital ulcers were caused by syphilis or chancroid. No studies have explicitly examined the role of herpes in HIV transmission, and the Rakai results certainly point to the need for such research.
The limited effect of the mass treatment on gonorrhea and chlamydial infection, the high prevalence of BV and herpes, and other factors may have contributed to the striking difference between the findings in Mwanza and Rakai. Perhaps the most important factor, however, is the one proposed by Penelope Hitchcock and Lieve Fransen.4 In their commentary on the Rakai results, published in the same issue of The Lancet, Hitchcock and Fransen note that the seemingly contradictory results from the two studies are actually consistent with a widely accepted model of the forces necessary to establish an HIV epidemic.
In this model, STIs have their greatest impact on HIV transmission during the early stages of an epidemic, when HIV spreads primarily among groups with high rates of sexual partner change and STIs. Since people who have many sex partners tend to engage in fewer acts of intercourse with each of those partners, their risk of transmitting an HIV infection is actually quite low. This is because HIV is not transmitted efficiently during its long latency period -- unless one or both partners have an STI. But those with multiple partners are more likely to have an STI, which increases the efficiency of HIV transmission and the likelihood that the virus will be acquired during a few sexual encounters with an HIV-positive partner.
In a more mature epidemic, STIs are less important in driving the epidemic. This is because HIV is spreading throughout the general population, among people who have fewer sex partners but engage in many more acts of intercourse per partner. If one partner has HIV, eventually, after many acts of unprotected sex, transmission will occur.
When the Mwanza study began in 1992, the HIV epidemic there was in an early phase, with an HIV prevalence rate of 4 percent among the general population. The Mwanza results, therefore, reflect the impact of STI prevention and treatment in an immature epidemic where STIs were still a driving force in the spread of HIV. In Rakai, on the other hand, HIV prevalence had already reached 16 percent by the time the trial began in 1994. In such a mature epidemic, STIs are still an important risk factor for individuals, but their contribution to the spread of HIV among the population may be negligible.
Lessons from Rakai and Mwanza
A closer analysis of the differences between Rakai and Mwanza, then, points to the urgency of improving STI treatment and prevention for populations with low HIV prevalence but high rates of other STIs. Providing effective STI services to these groups could save millions of lives -- if we act now, before HIV becomes well established and that window of opportunity is closed forever.
Many countries worldwide, from India to Madagascar to Jamaica, can benefit from this lesson. But even in countries with high HIV prevalence, the epidemic is not uniform. In most countries with mature epidemics, for example, adolescents have high rates of STIs and low rates of HIV and should be targeted with accessible, user-friendly STI services.
It's important to remember that the Rakai study confirmed that STI treatment is an extremely effective way of improving the health of mothers and children. Unpublished data from Rakai also show that STIs do increase an individual's risk of acquiring or transmitting HIV. Although the Rakai results suggest that STI control is most effective as an HIV prevention strategy in an immature epidemic, they certainly should not be used to justify continued neglect of STI services in resource-constrained areas with high rates of HIV. Effective, affordable STI services are needed everywhere -- not only to help contain HIV, but also to reduce the tremendous burden of morbidity and mortality caused by STIs themselves.
Next Steps
What constitutes effective STI services? Rakai is less helpful on that question. What is often forgotten in the debate over the Mwanza and Rakai studies is that both interventions had limited success in controlling STIs. Mwanza seems to have reduced HIV incidence primarily by reducing the duration of STIs, while showing a reduction in symptomatic urethritis and syphilis incidence. The mass treatment in Rakai, on the other hand, did not produce significant reductions in most STIs, except among pregnant women, whose sexual behavior is unlikely to be representative of the general population.
The failure of these interventions to achieve STI control is hardly surprising. Mwanza demonstrated both the value and the limitations of the syndromic approach to STI management. Because this approach uses clinical algorithms based on the constellation of patient symptoms and clinical signs, it is only effective when people have, recognize and seek appropriate treatment for symptoms. STIs are asymptomatic in most women and many men, and many of those who do have symptoms seek care in the informal sector.
In Rakai, on the other hand, the lack of effective, routine STI treatment between mass treatment rounds appears to have allowed reintroduction of STIs into the study communities, either by travelers who acquired these infections elsewhere or by community members who did not participate in the study. Although the study achieved impressive levels of participation -- 80 percent of eligible residents who were present in the community during each round received mass treatment -- this still meant that 30 percent of all those eligible throughout the study were not covered because they were not present.
Mass treatment is by definition an emergency measure, and is most cost-effective when applied to groups with high STI prevalence and rates of partner change high enough to maintain epidemic growth of STIs -- what are known as "core groups." In an FHI study in South Africa, for example, presumptive treatment and peer education targeting women in communities where miners live away from home in single-sex hostels resulted in significant reductions in STIs among the women and the miners.5
Clearly such presumptive treatment cannot stand alone as an STI control intervention, even in core groups. Implicit in the logic of this strategy for reducing STI prevalence is the need for maintenance strategies. To sustain lower rates, more long-term control measures, such as primary prevention and improved case management, must be in place.6
Improving STI services for women is particularly challenging because asymptomatic infections are so common in women and because of the social and cultural barriers they face in seeking treatment for symptomatic STIs. Syndromic management is of limited value, even for women who experience symptoms, because current algorithms for vaginal discharge do not work very well. There is a critical need for inexpensive, easy-to-use diagnostic tests for gonococcal and chlamydial infections to improve case management and identify asymptomatic infections. In the meantime, the best strategies for controlling STIs in women are syphilis screening and treatment of all pregnant women who test positive for syphilis, stronger efforts to refer the partners of male STI patients for treatment, and concerted efforts to make existing services more comfortable and convenient for women.
Because of the rapport they have built with sexually active women over the years, family planning clinics are often hailed as the solution to many of these challenges. The experience to date, however, has been disappointing. In most settings STI programs should not be implemented through reproductive health clinics alone because of the weakness of the tools available for managing STIs in women and because these programs lack experience in reaching men and core groups.
Family planning and reproductive health programs certainly have an important role to play in counseling clients about STIs, promoting condom use and referring clients for diagnosis and treatment of STIs, as outlined in the U.S. Agency for International Development's technical guidance on integrated services.7 And in some selected settings, integration of syndromic management may be appropriate. Whatever the setting, it is important to remember that there is no uniform approach to STI management and prevention for women or men. The best approach in a given setting will depend on the resources available -- including staff skills and time -- and the populations affected.
We are beginning to better understand what is appropriate in different settings, and there is much we can do now with the imperfect tools at our disposal to improve STI prevention and treatment. First and foremost, we need a stronger commitment to improving services at the places where people first seek treatment of STIs, whether that be a pharmacy, a physician's office or a public health clinic. At the same time, research is urgently needed to identify better tools -- from more affordable, appropriate diagnostic tests to innovative strategies for treating asymptomatic infections.
The Rakai results have highlighted some important unanswered questions for further research -- questions about the role of asymptomatic, difficult-to-treat and incurable STIs in HIV transmission and about the most effective mix of interventions for controlling the spread of STIs, including HIV, in different populations. Nevertheless, what is clear is that management and prevention of STIs remains critical as a public health measure in its own right and as one of the most effective ways to reduce HIV transmission among many of the populations most at risk.
Gina Dallabetta, MD, is director of technical support in FHI's HIV/AIDS Prevention and Care Department.
References
- H Grosskurth, F Mosha, J Todd, et al. 1995. Impact of improved treatment of sexually transmitted diseases on HIV infection in rural Tanzania: randomised controlled trial. The Lancet 346:530-536.
- DT Fleming and JN Wasserheit. 1999. From epidemiological synergy to public health policy and practice: the contribution of other sexually transmitted diseases to sexual transmission of HIV infection. Sexually Transmitted Infections 75:3-17.
- Wawer M, NK Sewankambo, D Serwadda, et al. 1999. Control of sexually transmitted diseases for AIDS prevention in Uganda: a randomised community trial. The Lancet 353:525-535.
- P Hitchcock, L Fransen. 1999. Preventing HIV infection: lessons from Mwanza and Rakai. The Lancet 353:513-514.
- R Steen, B Vuylsteke, A DeCoito, et al. Evidence of declining STD prevalence in a South Africa mining community following a core group intervention. Sexually Transmitted Infections, in press.
- R Steen and G Dallabetta. The use of epidemiologic mass treatment and syndrome management for sexually transmitted disease control. 1999. Sexually Transmitted Diseases (April suppl.):S12-S20.
- United States Agency for International Development. 1998. Integration of family planning/MCH with HIV/STD prevention, Priority for primary prevention with a focus on high transmitters. Programmatic technical guidance. December 23, 1998.
