Potential impact of HIV among IDUs on heterosexual transmission in Asian settings: scenarios from the Asian Epidemic Model

https://doi.org/10.1016/S0955-3959(02)00209-8Get rights and content

Abstract

Introduction: An important public health question is the degree to which HIV epidemics among injecting drug users (IDUs) can contribute to more generalised spread of the virus heterosexually among non-IDUs. Mathematical models can be of use in understanding the complexities of such epidemiological problems. Methods: In this paper, the Asian Epidemic Model (AEM) is used to explore different scenarios for potential HIV transmission by IDUs to non-injecting sexual partners and beyond. The model is applied to a ‘general Asian’ setting, in which (1) there is the potential for sex work to drive a self-sustaining heterosexual epidemic, and (2) the vast majority of IDUs are male and do not engage in sex work themselves, but a substantial minority of the IDUs do purchase sexual services from sex workers. Within such a setting, three different scenarios are modeled, and within each scenario, the level of HIV prevalence among IDUs is varied, to see the impact of preventing HIV among IDUs at different stages. Results: The number of HIV infections attributable to IDUs over a 15-year period was greatest when the epidemic among IDUs started in advance of a generalised heterosexual epidemic, with a 171% increase in infections if the IDU prevalence was not kept under control. Once a heterosexual epidemic was underway, controlling HIV prevalence among IDUs still resulted in a meaningful reduction in infections (23%), including among non-IDUs. Once the heterosexual epidemic started to decline, the relative proportion of infections attributable to IDUs increased, creating a situation for IDUs to feed into new epidemics. Discussion: In the early stages of a heterosexual epidemic, maintaining HIV prevalence at low levels among IDUs can buy critical time to address the potential for rapid heterosexual transmission. In more advanced heterosexual epidemics, preventing HIV among IDUs can still avert a substantial number of infections.

Introduction

Multi-person use (sharing) of needles and syringes for injecting drugs is a relatively efficient mode of transmitting HIV (Kaplan & Heimer, 1992). This has led to extremely rapid transmission of HIV among injecting drug users (IDUs) in many places, with incidence rates of 20% per year or higher (Des Jarlais & Friedman, 1998, Ball, Rana & Dehne, 1998, UNAIDS/WHO, 2001). However, sexually active IDUs also have the potential to transmit HIV to or contract HIV from their sexual partners. Therefore, an important public health question is the degree to which spread of HIV among IDUs can contribute to more generalised spread of the virus heterosexually among non-IDUs.

Both the rapidity and extent of sexual transmission of HIV from IDUs to non-injecting sexual partners have varied greatly in different settings. In North America and Western Europe, IDUs have been the predominant source of heterosexual transmission of HIV by infection of their sexual partners (Lansky, Fleming, Byers, Karon & Wortley, 2001, Centers for Disease Control and Prevention, 2001), but sexual transmission from IDUs to their partners has not led to a broader heterosexual epidemic in any country in North America or Western Europe. In these countries rates of partner exchange and levels of STDs that facilitate HIV transmission are apparently too low to support a general heterosexual epidemic. With each new heterosexual infection generating substantially less than one additional heterosexual infection (in modeling terms R0<1), HIV transmission through heterosexual routes alone cannot be sustained (May & Anderson, 1987).

However, in other settings, such as Eastern Europe and Asia, where sex work transactions have been the dominant source of heterosexual transmission of HIV, the role of injecting drug use in generalised spread is less clear and depends on the sexual networks (both commercial and non-commercial) of IDUs, the levels of other STDs, and the sexual partner exchange rates of both IDUs and their sexual partners, including sex workers. In many Asian countries (e.g. Thailand, Cambodia, India, etc.) it is clearly established that prevailing conditions can produce self-sustaining heterosexual epidemics, and while the Eastern European situation is less clear at present, the recent upswing in sex work in the region raises serious concerns.

Since the HIV is transmitted so efficiently through multi-person use of injecting equipment, epidemics among IDUs often grow with extreme rapidity after introduction of the virus. Such rapid growth of HIV in injecting populations has been seen across Asia. For example, in Manipur, India, HIV prevalence among IDUs increased from 0% prevalence in September 1989, to 50% prevalence 6 months later (Naik et al., 1991). Later studies in Manipur have found HIV prevalence rates from approximately 40% to over 80% in different samples and at different times (Manipur State AIDS Control Society, 2001, Manipur State AIDS Control Society, 2002). HIV has clearly spread from IDUs to their non-injecting regular sexual partners in Manipur with HIV infection among wives of IDUs rising from 6% in 1991 to 45% in 1997 (Chakrabarti et al., 2000, Panda et al., 2000). There are also indications of spread from IDUs to the larger heterosexual population in Manipur with recent sentinel surveillance among attendees in antenatal clinics finding rates from 1 to 2% (National AIDS Control Organization, 2001).

Rapid HIV spread also occurred among IDUs in Bangkok, Thailand with prevalence increasing from approximately 2% to over 40% in 7 months in 1988 (Kitayaporn et al., 1994, Weniger et al., 1991). A generalised heterosexual epidemic also occurred in Thailand as a whole, with substantial infection rates among sex workers first being detected in the North in 1989. It does not appear that the initial spread of HIV among IDUs in Bangkok directly caused the generalised epidemic, since the initial HIV infection among IDUs was with subtype B, while the generalised epidemic was dominated by subtype E (Ou et al., 1993). However, the IDU epidemic likely did interact with the generalised heterosexual epidemic, and continues to do so today, in ways that have not been fully documented. The early dominance of subtype B in Bangkok was not observed in the North, where subtype E dominated among IDUs from the start (Subbarao et al., 1998). Furthermore, over time, E came to be dominant among IDUs, even in Bangkok (Vanichseni et al., 2001, Wasi et al., 1995) with those infected with E being younger and more likely to be single when compared with those infected with B, a possible indication of linkages among the sexual and injecting drug use epidemics.

Sexual behaviour studies throughout the region have established that sexual relations between sex workers and injecting drug use do exist. In Bangladesh the behavioural surveillance system found that 33% of street-based IDUs sampled in Dhaka and Rajshahi had visited sex workers in the last month, but only one-fifth of them had used condoms. A quarter (23.1%) of IDUs sampled in Hanoi and one-fifth (20.3%) in Da Nang, Vietnam had visited sex workers in the last year, with 17% having had unprotected sex in that time frame (Tung et al., 2001). And in Jakarta, Indonesia, one-third had done so. These are not isolated figures—this pattern of a strong overlap between those injecting and those offering sex work services is seen throughout the region. At the same time, while there are a few places where sex workers are reported to inject, e.g. parts of Vietnam, to date this has not been a common practice among sex workers in the region. The present paper, therefore, focuses on the impact that HIV among non-sex worker IDUs has on heterosexual epidemics in the region and the impacts of IDUs who are sex workers will be addressed in future research.

These country examples and the sexual behaviour data gathered hint at the potential complexity of the interactions among HIV epidemics among IDUs and those among other populations. Unfortunately, the epidemiological data needed to characterise these relations and assess their effects on HIV spread has not yet been collected. Yet the lack of effective prevention efforts for IDUs allows HIV to continue spreading rapidly among IDUs in many areas in Eastern Europe and Asia (UNAIDS/WHO, 2001). The current situation clearly requires immediate and intensive efforts to reduce needle-borne transmission of HIV among IDUs. However, the extent to which HIV will spread to non-injecting sexual partners of IDUs, including sex workers, and then further to other sexual partners of the IDUs’ sexual partners is clearly a critical question in assessing the future of HIV epidemics in these areas. Given the well demonstrated potential for extensive sex work related transmission in many parts of Asia and the rapidly evolving sex work situation in Eastern Europe, assessing the impact of the relationship between injecting drug use and sex work on the epidemic becomes even more important.

Mathematical models can be of use in understanding the complexities of epidemiological problems such as this one—the transmission of HIV from IDUs to non-injecting sexual partners and its influence on larger scale epidemics. They can provide insights into the current status of an epidemic, the sources of new HIV infections, and how an epidemic evolves over time. Models can help to identify key variables—for which modest changes can produce large effects on the epidemic. Finally, modeling can help identify areas where better data collection is urgently needed.

In this paper, the Asian Epidemic Model (AEM; Thai Working Group on HIV/AIDS Projection, 2001) is used to explore different scenarios for potential HIV transmission by IDUs to non-injecting sexual partners and beyond. The model is applied to a ‘general Asian’ setting, in which (1) there is the potential for sex work to drive a self-sustaining heterosexual epidemic, and (2) the vast majority of IDUs are male and do not engage in sex work themselves, but a substantial minority of the IDUs do purchase sexual services from sex workers. Within such a setting, three different scenarios are modeled, and within each scenario, the level of HIV prevalence among IDUs is allowed to vary, to see the impact of preventing HIV among IDUs at various stages of the heterosexual epidemic. The first scenario addresses the question of what can happen if an epidemic among IDUs starts before the heterosexual epidemic takes off?’ The second scenario explores what can happen if an IDU epidemic starts after heterosexual transmission is already underway. And the third scenario, addresses what can happen if an IDU epidemic starts (or continues) after a heterosexual epidemic is brought under control’. The data needed as behavioural and epidemiological inputs for the AEM are taken from a variety of sources in India.

Section snippets

Description of the AEM

In surveys done around Asia, a significantly larger proportion of men than women in the region report premarital and extramarital sex (Brown, Mulhall & Sittitrai, 1994, Brown & Sittitrai, 1995). This creates a gender imbalance—far more men are seeking sexual contacts than women are available to provide them. As a consequence, a small number of sex workers have large numbers of clients per night and sex work has become an important component of sexual cultures in most Asian countries. Since

Data sources

The behavioural inputs for these scenarios and the reported HIV trends used to adjust the transmission parameters to obtain a good model fit were based on data from several Indian cities to create a composite city for this paper. Use of these data sources allowed us to set the model parameters and formulate a baseline scenario from which to evaluate the impacts of the different scenarios used in this exercise. Observed prevalence trends among sex workers, IDUs and general population males and

Results

The three scenarios described in the methods section were modeled, using different behaviours and varying the level of HIV among IDUs as described above (very low, low, high and very high). The number of IDUs was fixed at 38 000 and the proportion of IDUs who visit sex workers was fixed at 10% for all scenarios, which is roughly consistent with results seen in the general male population.

Importance of IDU related heterosexual spread in different scenarios

In most Asian epidemics, patterns of sex work have the greatest influence on the extent of HIV spread, with the greatest number of adult infections occurring through sex worker/client transmission in the earlier stages and client/spouse or regular partner transmission in the later stages. Although ultimately the spread of epidemics in general populations depends on the level of sexual risk behaviour in these populations, the modeled scenarios here show that even under conditions where sex work

Acknowledgements

This work was supported by Family Health International (FHI), with funding provided by the United States Agency for International Development (USAID). Views expressed do not necessarily represent the views of FHI, USAID, Beth Israel Medical Center, the East West Center or SHARAN.

References (38)

  • D.C. Des Jarlais et al.

    Fifteen years of research on preventing HIV infection among injecting drug users: what we have learned, what we have not learned, what we have done, what we have not done

    Public Health Reports

    (1998)
  • D.C. Des Jarlais et al.

    Maintaining low HIV seroprevalence in populations of injecting drug users

    Jouranl of American Medical Association

    (1995)
  • G.R. Eldred Tellis et al.

    Rapid situation assessment of drug use in Mumbai

    (2000)
  • E.H. Kaplan et al.

    A model-based estimate of HIV infectivity via needle sharing

    Journal of Acquired Immune Deficiency Syndromes

    (1992)
  • D. Kitayaporn et al.

    HIV-1 incidence determined retrospectively among drug users in Bangkok, Thailand [see comments]

    AIDS

    (1994)
  • A. Lansky et al.

    A method for classification of HIV exposure category for women without HIV risk information

    MMWR

    (2001)
  • Status report 2000–2001

    (2001)
  • Epidemiological analysis of HIV/AIDS in Manipur-September, 1986 to February, 2002

    (2002)
  • R.M. May et al.

    Transmission dynamics of HIV infection

    Nature

    (1987)
  • Cited by (0)

    View full text