Asia and Africa: On Different Trajectories?

See allHide authors and affiliations

Science  25 Jun 2004:
Vol. 304, Issue 5679, pp. 1932-1938
DOI: 10.1126/science.304.5679.1932

Asia soon could have more HIV-infected people than sub-Saharan Africa. But that doesn't mean it will have African-scale epidemics

From the western deserts of India to the eastern ports of China, HIV has raced through the same high-risk group: injecting drug users (IDUs). There is no better way to spread the virus than sharing contaminated needles, and one IDU community after another in Asia has seen prevalence rates rocket from zero to 50% in as short as 6 months. Add high rates of infection among IDUs to a booming sex industry and huge populations of migrant workers, and you have the ingredients for a potential disaster: Some AIDS experts even predict that Asia will have more HIV-infected people in 2010 than the 30 million or so in sub-Saharan Africa today.

Enlightened care.

“Anita,” a native of Nepal, says she was sold to a brothel in India when she was 14. She receives free health care and HIV prevention services at a nearby clinic for sex workers run by the National AIDS Research Institute in Pune.

But there's an asterisk to this grim scenario: The scale of Asia's epidemic has more to do with the size of the continent's population than the explosive spread of the virus. In the past few years, it has become clear that the epidemics in Asia and Africa are surprisingly different. HIV has not spread rapidly by heterosexual sex to cause an African-styled “generalized” epidemic anywhere in Asia, and many epidemiologists now believe it is unlikely to do so in most parts of the continent.

Yes, HIV has spread quickly in high-risk groups and made inroads into the population at large, infecting an estimated 7 million Asians in all. But no Asian country has reported a prevalence rate in adults higher than 4% (Cambodia, 1999), and relatively few prenatal clinics have found prevalence rates above 1%—the standard measure for a generalized epidemic. In contrast, 12 African countries now have prevalence rates in adults of 10% or more; South Africa reported that 26.5% of pregnant women in 2002 were infected with HIV.

Could that happen in Asia? “Much of Asia has the potential for an epidemic where 2% to 3% of adults become infected, and by my standards, that's an extremely serious epidemic,” says Tim Brown, a physicist by training who lives in Bangkok and does HIV modeling for the East-West Center. “But will we see 10% to 15% prevalence in Asia outside of a few isolated areas? Absolutely not.”

Like Brown, Christopher Beyrer, an epidemiologist at the Johns Hopkins Bloomberg School of Public Health in Baltimore, Maryland, stresses that a low prevalence rate in a large country can spell disaster. “One or two percent of the Indian or Chinese populations is an enormous number of people,” says Beyrer. But he also shares the emerging perspective that Asia will not go the way of sub-Saharan Africa. “We've been looking, waiting for evidence of that for a number of years, but—happily—it's not panning out,” says Beyrer, who has closely studied HIV's spread in Thailand, Myanmar (formerly Burma), China, and Malaysia.

To many who combat epidemics, this means there is a tremendous opportunity to derail the virus by aggressively targeting high-risk groups—IDUs, sex workers, and gay men, who often are ostracized by societies for legal and moral reasons. “If we focus our attention there, we can really do a lot of good work,” says Elizabeth Pisani, an epidemiologist with Family Health International (FHI) based in Jakarta, Indonesia.

Pisani decries the hype that has surrounded projections of rampant viral spread through general populations in Asia. “With HIV,” she says, “the facts speak for themselves.” And she says that the “obsession with generalized epidemics” has clouded attempts to discern the actual epidemiological patterns. “In the early to mid-1990s, we didn't really understand the shape that epidemics were taking in Asia,” she says.

The spread of HIV of course differs dramatically from country to country in Asia, as well as within different regions of specific countries. “It's very difficult to speak about ‘the Asian epidemic,’” says Peter Piot, head of the Joint United Nations Programme on HIV/AIDS (UNAIDS). “Whatever we come up with, we always find a big exception in Asia.”

Yet in the past few years, epidemiologists have stepped up attempts to characterize HIV's spread across Asia as a whole, forcing many to rethink how the virus moves around the continent and, ultimately, how to design targeted prevention programs. Beyrer and his co-workers used molecular biology to track how various HIV subtypes spread out from heroin-producing Myanmar and Laos along drug-trafficking routes in India, China, and Vietnam. Brown spearheaded the development of a fascinating computer model that analyzes the specific characteristics of Asian epidemics. And Pisani and two dozen colleagues in an informal network called Monitoring the AIDS Pandemic have just completed a detailed report on AIDS in Asia that incorporates this new understanding of the epidemic across the continent. They plan to release it in July at the XV International AIDS Conference in Bangkok. Its title: “Face the Facts!”

Misleading generalizations?

By 1998, James Chin, an epidemiologist at the University of California, Berkeley, had become a thorn in the side of many colleagues who accused him of downplaying the potential of HIV to spread in Asia. That year, Chin, who started the global HIV/AIDS surveillance program at what's now UNAIDS, co-authored a paper that said, “HIV prevalence rates in the total sexually active population of most Asian-Pacific countries will not, in our opinion, ever reach 0.5%.”

Chin had long questioned the potential of HIV to spread by heterosexual sex everywhere outside Africa. “I don't think most of the world's heterosexual population has sufficient sexual risk behaviors in terms of networks to drive any type of epidemic of sexual transmission,” says Chin, now a consultant to the World Health Organization and other groups. “You get pockets here and there in the sex workers, and that obviously has to be looked at. And you get epidemics in injecting drug users who share blood where HIV has penetrated into IDU networks. That will continue. But generalized transmission of HIV is a very inappropriate epidemiologic term.”

Seeing red?

As the spread of HIV has turned Africa more red (adult prevalence >20%), Asia has become darker blue (2% to 5% adult prevalence). The scales, at least for now, remain dramatically different.


In contrast to Chin's analyses, the U.S. National Intelligence Council published a report in September 2002 that roiled health officials in China and India. The report predicted that by 2010, HIV infections in China could total up to 15 million. India, the report warned, could have 4% of its population infected—a staggering 25 million people. “Those numbers were plucked out of the air,” charges Chin.

Richard Feachem, an epidemiologist who heads the Global Fund to Fight AIDS, Tuberculosis, and Malaria, says he views those gloomy forecasts in precisely the opposite way from Chin. “It's a huge mistake to say that this can't happen,” says Feachem. “You have to do a heck of a lot to attenuate this epidemic. It would be wise to assume the worst rather than best. I see 15 years of optimistic assumptions about Asia that have turned out not to be true. So I say let's make more pessimistic assumptions, and if we're proved to be wrong, everyone will be very happy.”

Low-risk high.

This heroin user in Manipur, India, has a clean syringe and needle provided by a needle-exchange program that former users run.

Piot of UNAIDS, who played a pioneering role in characterizing the African epidemic, further cautions people not to jump to conclusions about the potential spread in Asia. “The major difference when you compare the African and Asian epidemic is the speed,” says Piot. “It still can go to double-digit prevalence in some countries, and certainly some of the Indian states,” he warns. And Swarup Sarkar, a UNAIDS epidemiologist based in Bangkok, worries, too, about what he calls “subnational” epidemics. “There are subnational epidemics at 5% prevalence or more going on, and we're ignoring this because of low national prevalence overall,” says Sarkar, who has worked extensively in India and several other Asian countries.

Whether the epidemic in Asia follows the path Chin predicts or one that more closely resembles an African epidemic rests largely on the rate of spread in the general population by heterosexual sex. Although epidemiologists wish they had more data, many now say that marked distinctions characterize sexual behavior patterns in Asia and Africa. “From all behavioral studies, there's a big difference in the number of partners that men and women in Asia report having, and that's not true in Africa,” says Piot. In other words, relatively few females in Asia, other than sex workers, have multiple partners. As Beyrer puts it, “For the great majority of women in Asia, it's still virginity until marriage and monogamy afterward.”

Piot also notes that in some hard-hit countries in Africa, studies have reported that many HIV-infected older men have sex with younger women, who become infected themselves and then spread the virus to men their own age. “The more we look at it, intergenerational sex is a big, big driver,” says Piot. “If everybody would have sex with someone from their own age group, the epidemic would die out.” In Asia, says FHI's Pisani, “when women do have nonmarital sex, it tends to be with people of their own age.”

In addition to sexual practices, circumcision plays a role in the spread of HIV, says medical anthropologist Daniel Halperin, who works on HIV/AIDS at the U.S. Agency for International Development. He points to several studies that link the lack of circumcision to Africa's worst HIV epidemics. “Every country in the world that's even getting very close to double-digit prevalence has lack of male circumcision and multiple-sexual- partner patterns among both genders,” says Halperin. He notes that the four countries in Asia that have high circumcision rates—the Philippines, Pakistan, Bangladesh, and Indonesia—have among the lowest HIV prevalence rates in the region.

Fast tracks.

HIV races through communities of injecting drug users (right), and a molecular analysis of the virus reveals that subtypes—the strains, designated alphabetically—travel from heroin-producing Myanmar and Laos along drug-trafficking routes.

Risk reduction.

This free needle program run by Daytop in Kunming, China, remains a rarity in its home country—and throughout Asia.

A model epidemic

East-West's Brown and co-worker Wiwat Peerapatanapokin decided to investigate the dynamics of HIV's spread in Asia and to test Chin's theory. This mathematical exercise, called the Asian Epidemic Model, charts how various behaviors alter the movement of the virus as it crisscrosses between the general population, IDUs, sex workers, and their clients.

Brown cautions that “modeling is still more art than science, especially given the huge data gaps in the region.” Still, by adjusting various parameters, they have arrived at several provocative conclusions that many colleagues have studied closely (see page 1934). UNAIDS's Piot goes so far as to say that when it comes to describing the forces that spread HIV in Asia, “the person [who] I believe has the best vision on this is Tim Brown.”

IDUs play a starring role in the model. Thailand, China, Myanmar, India, Vietnam, Nepal, Indonesia, and Malaysia all have seen the rapid spread of HIV through their IDU populations. “The nidus of the epidemic in mainland Asia are the IDUs,” says Roger Detels, an epidemiologist at the University of California, Los Angeles (UCLA), who works in several Asian countries and has trained dozens of scientists from the region. “The only countries where the first major epidemic was not in IDUs were Laos and Cambodia.”

The model shows how IDUs dramatically speed an epidemic. “It really can accelerate things by 20 to 30 years,” says Brown. Typically, the IDU population “seeds” sex workers, the other main high-risk group, who provide a “bridge” into the general population. “Sex workers and clients really form a feedback loop,” explains Brown. “It's the rate at which you get feedback between those two populations that determines how quickly the epidemic rises.”

This dynamic also explains one of the more curious aspects of the Asian AIDS epidemic: Cambodia has the highest documented adult prevalence rate (2.7% in 2001), but hardly any IDUs. Brown explains this by comparing Cambodia to its next-door neighbor, Thailand. In the early 1990s, about 20% of men in both countries reported hiring sex workers. “In Thailand and Cambodia the epidemic was kind of a foregone conclusion because the level of risk was so high,” says Brown. While IDUs fueled the epidemic in Thailand, their role became overwhelmed because sex workers had so many clients. “Clients and sex workers drive the epidemics in Asia,” says Brown, and ultimately, it's the total number of men who visit sex workers that determines how quickly HIV spreads. “If you double the number of clients, you have a larger number of men who are exposed to HIV,” says Brown. “So it builds larger epidemics.”

Reaching out.

Billboard in Taunggyi, Myanmar, mixes sex, drugs, and rock ‘n’ roll to spread the AIDS prevention message.

In one particularly vivid simulation, Pisani used the Asian Epidemic Model to analyze Jakarta with and without IDUs. Jakarta did not detect an HIV-infected IDU until 1998, and the prevalence rate in that community by 2001 had jumped to 48%. Despite consistently low condom use, the HIV prevalence rate among sex workers in Jakarta stayed low for years. But studies show that as the prevalence rate rose in IDUs—one in five of whom report paying for sex—it bridged into the sex worker population, which had a 7% prevalence rate in one 2003 survey. Pisani says this has created “a critical mass of infections” in the sex industry that now will spread more widely. If risk behaviors do not change, she estimates that Jakarta will have 110,000 infections by 2010, one-third of those in IDUs. But without any IDUs, the model suggests that Jakarta would have had only 3000 infections in 2010. “If we had prevented the epidemic among drug users, we would have prevented the whole epidemic,” says Pisani. “It should wake us up to the prevention opportunities where HIV among drug users is still low. Bangladesh, Pakistan, the Philippines—they don't have to go the way of Jakarta.”

Impassioned debates surround several of the assumptions in Brown's model. The Global Fund's Feachem questions the generalization that Asian women have fewer multiple partners. “I wish it were true,” he says. Epidemiologist Wu Zunyou of China's Center for Disease Control and Prevention in Beijing notes that one study of sexual behavior in China, India, Russia, Zimbabwe, and Peru found that the Chinese reported lower rates of extramarital sex—but their rates of sexually transmitted diseases were the same or higher. “Chinese enjoy the doing, not the talking,” says Wu. UNAIDS's Piot questions whether circumcision has much power. “If lack of circumcision would be a big driver, Asia should have a big, big epidemic,” he says. And Pisani says the Asian Epidemic Model is “not stable enough to run over long periods.”

Writing on the wall.

Prevention messages such as this one at an AIDS clinic in Kolkata, India, aim to derail HIV with knowledge of how it is transmitted.

But there is wide consensus on one prediction that Brown makes: HIV potentially could establish a severe epidemic on the island of New Guinea, which is half Papua and half Indonesia. “If there's one place in Asia and the Pacific that has the potential for a more African-style epidemic, that's it,” says Brown.

Like countries in sub-Saharan Africa that have substantial HIV epidemics, both men and women in New Guinea report having multiple sexual partners, and the island has virtually no IDUs. “It's probably the hottest epidemic in the whole region,” says Carol Jenkins, a medical anthropologist who lived in Papua New Guinea for 15 years and conducted HIV and behavioral surveys. “Premarital sex has no negative connotation whatsoever.” According to one study of unmarried women between 15 and 24, 41% were sexually active and 30% of that group had sex with men 10 years older than themselves. Although the most recent adult prevalence rate in Papua New Guinea remains a relatively low 0.7%, one prenatal clinic last year reported an HIV prevalence rate of 2.5%. “It may look like Africa in 5 years,” says Jenkins.


In 1997, Beyrer and Yu Xiao-Fang of Hopkins made a startling discovery with their Chinese colleagues Chen Jie and Lui Wei of the Guangxi Provincial Health and Anti-Epidemic Station. A molecular analysis of HIV-infected blood samples from IDUs in Guangxi revealed that two cities located only a few hundred kilometers apart—as the crow flies—had completely different strains, or subtypes, of the virus. Over the next year Beyrer and Yu traveled extensively through Guangxi, and the perplexing discovery soon made perfect sense: One city had a north-south highway running through it, while the other had an east-west highway—and no road directly connected the two locations. “It was really clear that these viruses were following these highways that were distinct,” says Beyrer. “It appeared that the viruses were spreading along heroin trafficking routes through chains of injectors.”

The researchers expanded their collaboration to look at HIV subtypes found in IDUs from other areas of China and from Myanmar, India, and Vietnam. As they explained in a paper published in AIDS in January 2000, the analysis indicated that specific subtypes moved from heroin-producing Myanmar and Laos to other countries along trafficking routes. “Heroin routes end up being routes for spread and diversification of the virus, which is bad news for the populations along them,” says Beyrer.

To Beyrer and other epidemiologists, such studies hold a key to preventing the spread of HIV in Asia, which requires an intimate understanding of high-risk groups. “It doesn't matter what your culture is—whether it's Uygur, Muslim, Burmese tribal, or Chinese hipster—if you're on this heroin route, you need to implement prevention,” says Beyrer. “And along these heroin routes, there was a glaring absence of prevention and treatment.” Similarly, studies in Indonesia and Thailand both have shown that many IDUs become infected by sharing needles in prisons, where precious little HIV prevention work takes place. “Too many policymakers operate under the view that injecting epidemics don't have anything to do with anything else,” says Brown.

As research clearly has shown in Europe, Australia, and the United States, a “harm- reduction package” that offers clean needles, substitution drugs such as methadone, health care, and counseling can powerfully thwart HIV. But prevention campaigns in Asian countries largely have steered clear of IDUs and other high-risk populations, including clients of sex workers, gay and transgendered men (many of whom are bisexual), and female sex workers who inject drugs. Outside Hong Kong, no locality has a wide-scale, comprehensive harm-reduction program for IDUs: Thailand recently targeted users in its stepped-up war on drugs, China still sends them to compulsory treatment centers, and one treatment camp in India even chains them to their beds (Science, 23 April, p. 509). Government health workers in China and Myanmar shy away from distributing condoms to sex workers. Out of nearly 400 sites that tracked HIV's spread around India in 2002, only three explicitly measured the prevalence rate in gay men, and China had none.

But as Thailand and Cambodia have shown, Asian countries can stage effective prevention campaigns. Both Cambodia and Thailand famously succeeded at slowing their epidemics by launching massive education and condom campaigns targeted at sex workers and their clients (Science, 19 September 2003, p. 1658). Without those campaigns, Brown's model suggests that these two countries may well have Africa-scaled epidemics today. “Thailand and Cambodia without interventions would have reached 10% to 15% adult prevalence by now,” says Brown. “And the reasons they didn't is because condom use went up to 80% to 90%, and men using sex workers dropped in half.”

Staging effective prevention campaigns requires constant reevaluation, too, as many Asian countries are undergoing momentous transitions. “With a large middle class coming up in India, China, and Indonesia, it's difficult to believe that sexual behavior is not changing,” says UNAIDS's Sarkar.

Already, studies in both Japan and Thailand have found significant increases in the number of women who report having premarital sex and multiple partners. Masahiro Kihara of the Kyoto University School of Public Health and co-workers wrote in the February 2003 issue of the Journal of the Acquired Immune Deficiency Syndromes that over the past decade, multiple partnerships were “dramatically elevated in young women.” In the March 2003 issue of AIDS and Behavior, a group led by Fritz van Griensven, an epidemiologist at the U.S. Centers for Disease Control and Prevention branch in Thailand, describes a survey in which 43% of unmarried female vocational students in northern Thailand between 15 and 21 reported having had sexual intercourse.

FHI's Pisani urges people to interpret these trends with caution. An increase in the number of young women who have sex with men their own age could prevent the spread of HIV. Pisani well realizes that this is a discomfiting message. “We're hamstrung by our own touchy-feely rhetoric, and we sacrifice public health,” she says. “The guys are going to have sex anyway. If there are no ‘nice girls’ to have sex with, they're going to have sex with prostitutes.”

Balancing the scientific findings with the public health needs presents a terrifically tricky equation that no computer model can solve. “You don't want to underestimate the threat,” says UCLA's Detels. “When you're doing public health and you want to introduce interventions, you want to introduce anxiety so the public will do something. But you don't want too much anxiety or they'll say, ‘Hell, I'm already infected.’”

Pisani says she's “greatly concerned” that an exaggerated emphasis on the threat of generalized epidemics in Asia has taken people's eyes off the target. “It makes it harder for us to do the job that we need to do,” she says. “We've painted it as though it only counts if it's a generalized epidemic. In Asia we can save tens of thousands of lives and a lot of misery if we have the courage to focus our interventions where they need to be focused. And in general, we know where the risk is concentrated in Asia.”

View Abstract

Stay Connected to Science

Navigate This Article