Epidemics in African countries at war
In 1996, UNAIDS stated that ‘HIV spreads fastest in conditions of poverty, powerlessness and social instability – conditions that are often at their most extreme during emergencies.’[i] However, evidence shows otherwise. After 1984, when HIV tests have been available to track HIV prevalence, seven African countries have had continuing or new civil disorders or wars that lasted for a decade or longer. In all seven, HIV epidemics grew slowly or not at all during the war.[ii] Moving counterclockwise from West Africa, these countries are Sierra Leone, Liberia, DRC, Angola, Mozambique, Somalia, and Southern Sudan.
Sierra Leone’s civil war began in 1991, when rebels from the Revolutionary United Front entered from Liberia. The rebels were notorious for chopping off hands to terrorize the population. For 11 years, the rebels lived off profits from diamond mines. Within months after the war ended in 2002, a survey found that 0.6 percent of adolescents and adults aged 12-49 years had lost a limb during the war, and 6 percent of women reported forced sex by someone other than their regular partner during the past year.[iii] The same survey found that only 0.9 percent of adults and adolescents were HIV-positive.
Liberia’s civil war began in 1989, when Charles Taylor’s National Patriotic Front entered from Cote d’Ivoire. Taylor seized power in 1990. Subsequently, war continued off and on until 2003, when Taylor resigned as president and moved to Nigeria. Surveys before the war found little HIV infection. HIV prevalence increased during the war, but remained low. A national survey in 2006-07 found 1.5 percent adult HIV prevalence.
In DRC, President Mobutu’s grip on power slipped with age and with the end of the cold war. In 1991, soldiers rioted to protest low pay, and foreigners fled the breakdown of civil order. In 1996, Laurent Kabila’s rebels overthrew Mobutu. Civil war resumed in 1998 with troops from Angola, Rwanda, Uganda, Zimbabwe and other African countries attacking or supporting Kabila’s government. After Joseph Kabila replaced his assassinated father in 2001, the central government negotiated and extended its authority over most of the country through early 2009. A 1997 study in three cities found that ‘HIV seroprevalence rates remain relatively low and stable’ despite ‘social disruption, the rapid decline in health-care provision, and the decrease in funding in health education programmes.’[iv] As already reported, HIV prevalence fell from the late 1980s among women in urban antenatal clinics, and a national survey in 2007 found 1.3 percent HIV prevalence among adults.
Angola’s civil war lasted from independence in 1975 until the death of the rebel leader, Jonas Savimbi, in 2002. In that year, the median HIV prevalence in sentinel surveys among women attending antenatal clinics in six provinces was 2.4 percent.[v] In 2004, Angola’s vice-minister of health noted that ‘inland provinces that have been more protected by the effect of war have lower figures of [HIV] prevalence.’[vi] Angola’s HIV prevalence was significantly lower than in neighboring countries at peace to the south and east.
Civil war in Mozambique lasted from 1976 to 1992. During much of that time, the South African government supported rebels to block Mozambique’s aid to South Africa’s anti-apartheid activists. Rebel attacks ‘destroyed, looted, or forced to close’ 46 percent of rural clinics and health posts through 1988.[vii] Mozambique’s HIV prevalence stayed low during the war, but took off when peace was achieved. Among women attending an antenatal clinic in Maputo (the capital), HIV prevalence rose from 1.2 percent in 1992, when the war ended, to 5.8 percent in 1996, and then to 21 percent in 2004 (see Figure 6.1). Parallel increases occurred throughout the country.
The secession of northern Somalia in 1988 sparked Somalia’s civil war. In the north, Somaliland emerged as a new state in 1991, which as of 2007 was not recognized by the world community. In the south, militias overthrew the government in 1991. Fighting among militias, Islamic organizations, and foreign forces continued into 2009. Sentinel surveys in urban antenatal clinics have consistently reported low HIV prevalence among women, and UNAIDS estimated 0.5 percent adult HIV prevalence in Somalia in 2007 (see the Statistical Annex).
Southern Sudanese have fought for autonomy or secession for decades. The first north-south civil war began in 1956 and lasted until 1972. The second began in 1983 and ended in 2005. (Another civil war in Darfur, in western Sudan, continued into 2007.) In 2002-03, surveys assisted by the US CDC measured HIV prevalence in parts of Southern Sudan controlled by the rebels. Findings from these surveys – particularly, adult HIV prevalence of 0.4 percent in Rumbek town and 0.7 percent in rural areas of Yei County[viii] – give a good indication of low HIV prevalence in the south’s overwhelmingly rural population at the end of the war.
Two studies published in 2007 present similar evidence for other African countries with shorter or more limited wars. A study assisted by the UN High Commission on Refugees found that data from Burundi, Rwanda, and north Uganda, as well as from several countries considered in previous paragraphs, ‘did not show an increase in prevalence of HIV infection during periods of conflict...’[ix] The other study developed an index of armed conflict for 37 African countries during 1991-2000. For 15 countries with no armed conflict, average HIV prevalence in urban women in 2000 (averaging each country’s median HIV prevalence in urban sentinel antenatal clinics) was more than twice as high as the average for 22 African countries with armed conflict during the decade.[x]
During wars, UNAIDS recognizes ‘a strong likelihood that AIDS control activities, whether undertaken by national governments or NGOs, will have been severely disrupted or have broken down altogether.’[xi] Hence, it is noteworthy that HIV prevalence in African countries after long-term wars has been lower than in neighboring countries at peace. Certainly, HIV prevention programs have stopped some transmission (for example, by providing condoms). However, the failure of these programs to consistently outperform the ‘no program’ situation in countries at war shows that their positive impacts are overwhelmed by other factors.
One proposed explanation for low HIV prevalence after long-term wars is that wars isolate people, reducing their sexual contacts with HIV-positive outsiders. Although wars no doubt reduce sexual contacts with visiting truckers, they likely increase contacts with visiting soldiers. The net impact on sexual exposures to HIV is not clear. Another possible explanation is that wars slow HIV transmission through unsafe healthcare by closing government, mission, and private clinics. This interpretation fits events in Yambuku, DRC, where closing the mission hospital – which had spread Ebola through contaminated injections – stopped the 1976 Ebola outbreak.
Low-level generalized epidemics in Africa
As of 2007, almost 300 million people – more than a third of Africa’s population – lived in countries with less than 2 percent HIV prevalence among adults. In six countries – Gambia, Niger, Madagascar, Mauritania, Senegal, and Somalia – less than 1 percent of adults were HIV-positive. HIV prevalence in some of these countries is comparable to what is found in concentrated epidemics in the Americas and Europe. However, these are generalized, not concentrated epidemics – women are infected as often or more often than men (Table 6.5). For example, Senegal’s 2002 national survey found that 0.9 percent of women were infected vs. 0.4 percent of men. As in many other African countries, HIV prevalence among women in Ethiopia, Guinea, Liberia, and Niger increased with education and wealth.
Table 6.5: Distribution of HIV infections in selected African countries with low-level generalized epidemics
Sources: See the Statistical Annex.
In African countries with low HIV prevalence, infections notably do not concentrate in promiscuous adults. Information on sexual behavior and HIV prevalence is available from national surveys for nine countries with adult HIV prevalence ranging from 0.7 percent to 2.2 percent (Benin, Burkina Faso, DRC, Ethiopia, Ghana, Guinea, Liberia, Mali, and Niger). In these nine countries, women who reported 0-1 sexual partners in the past year had 91-100 percent of HIV infections in women (see Table 7.1). In the same nine countries, men who reported 0-1 sexual partners had 56-96 percent of HIV infections in men. Does sexual transmission of HIV from sex workers to clients to wives drive these low-level generalized epidemics? The evidence does not agree. In the same nine countries, 60-85 percent of HIV-positive wives had husbands who were HIV-negative (see Table 7.5).
Even so, non-IDU sex workers often have high HIV prevalence in low-level generalized epidemics. For example, nine sentinel surveys in Ethiopia during 1990-2005 reported 36-74 percent of sex workers to be HIV-positive.[xii] In generalized epidemics, women in sex work are on the front lines – they are exposed to both blood and sex risks. From available evidence, no one can show what proportions of their HIV infections come from which risks.
Outside of Africa, the oldest generalized epidemics were recognized in Haiti and in other small countries in the Caribbean region from the 1980s. Currently nine small countries in the region (Bahamas, Barbados, Belize, Dominican Republic, Haiti, Guyana, Jamaica, Suriname, and Trinidad and Tobago) have low-level generalized or mixed epidemics, with estimated adult HIV prevalence ranging from 1 percent to 3.0 percent (in the Bahamas).
Aside from Africa and the Caribbean, low-level generalized or mixed epidemics have been recognized in only 5 countries: Cambodia, India, Myanmar, and Thailand in Asia; and Papua New Guinea in Oceania. Across these 5 countries, estimated adult HIV prevalence ranges from 0.3 percent to 1.5 percent, but because of India’s large population, these five countries collectively account for 3.4 million of the estimated 28 million increase in global HIV infections from 1988 to 2007. In some other countries with very low HIV prevalence, such as Nepal, information is too limited to reliably characterize their epidemics as concentrated or very low-level mixed or generalized epidemics.
The Caribbean
After UNAIDS’s 2002 estimate put Haiti’s adult HIV prevalence as high as 6.1 percent, a national survey in 2005-06 found only 2.2 percent of adults to be HIV-positive. Taken together, the nine countries in the Caribbean with at least 1 percent estimated adult HIV prevalence, have several hundred thousand HIV infections, less than 1 percent of the world total.
Comparable or higher HIV prevalence in women vs. men shows these are generalized or mixed, not concentrated, epidemics. In contrast to Africa, HIV prevalence in the Dominican Republic (though not in Haiti) is higher in rural than in urban adults, and falls with education and wealth (Table 6.6).
Generalized or mixed epidemics in nine small countries in the Caribbean are growing slowly, if at all. Other regional countries have concentrated epidemics, or possibly mixed epidemics, with low HIV prevalence. In Guatemala and Panama, for example, surveys among women attending antenatal clinics seldom find more than 1 percent to be HIV-positive.[xiii]
Table 6.6: Distribution of HIV infections in Asian and Caribbean countries* with low-level generalized epidemics
* All countries for which data are available from national surveys. † From any to not more than 7 years of education. Sources: See the Statistical Annex.
India
After the discovery of HIV in India in 1986, HIV spread rapidly. By 1990, expanded testing found 13-55 percent HIV prevalence in seven sites spanning thousands of kilometers, including sites testing sex workers, IDUs, and people seeking treatment for sexually transmitted diseases.[xiv] India’s National AIDS Control Organization estimated 200,000 HIV infections in 1990, increasing to 3.5 million in 1998.[xv] In 2005, UNAIDS estimated that India had 5.7 million infections – more than any other country in the world. However, estimates for 1998 and 2005 were too high.[xvi] A 2005-06 national survey found 0.28 percent HIV prevalence in adults.[xvii] Based on the survey, but with some adjustments, the government of India and UNAIDS agreed on an estimate of 0.36 percent adult HIV prevalence and 2.5 million infections in 2006.[xviii]
The HIV epidemic varies regionally across India, a large country, with a population greater than all of sub-Saharan Africa. Most HIV infections are found in four southern states with a fourth of India’s population – Andhra Pradesh, Karnataka, Maharashtra, and Tamil Nadu. In these states, adult HIV prevalence in the 2005-06 survey ranged from 0.34-0.96 percent. In three eastern states with 0.5 percent of India’s population – Mizoram, Manipur, and Nagaland – HIV prevalence is also relatively high, due in part to IDUs injecting heroin from Myanmar. The 2005-06 survey found 1.1 percent adult HIV prevalence in Manipur. Other states, with collectively three-quarters of India’s population, have an adult HIV prevalence of only 0.12 percent.
Overall, India appears to have a low-level mixed epidemic, with some states having more concentrated or more generalized epidemics. An estimated 1.6 men are infected for every woman, a sex ratio that is intermediate between concentrated and generalized epidemics. Among women, education beyond the primary level reduces risk. On the other hand, greater wealth appears to increase risk among women, except that the highest wealth quintile has lower HIV prevalence (see Table 6.6).
Many AIDS experts have proposed that heterosexual sex work accounts for most infections in India, as sex workers infect men who infect wives. A World Bank model showing sex workers driving India’s HIV epidemic assumed that 15 percent of men visited sex workers an average of 50 times per year.[xix] Overwhelming evidence shows much more conservative sexual behavior.[xx] Moreover, in the 2005-06 survey, 39 percent of married, HIV-positive women had HIV-negative husbands (see Chapter 7). Commercial sex no doubt contributes to India’s HIV epidemic – but how much does it contribute?
Thailand
In 1987-88, the HIV epidemic exploded among IDUs in Bangkok. In the next year, it exploded among (non-IDU) women in sex work in the far north. From 1988 to 1989, HIV prevalence among brothel-based sex workers in Chiang Mai city in north Thailand increased from 0.4 percent to 44 percent.[xxi] Two studies among sex workers in northern Thailand during 1989-93 show 3.6-17 percent contracting HIV each month (after setting aside those who were already infected).[xxii] These are among the highest rates of HIV incidence ever observed in any population. Because studies ignored medical injections and other blood exposures, no one knows how much blood exposures contributed to these extraordinarily high rates of HIV incidence.
Unlike Africa, where HIV hit hardest among well-off people in capital cities, Thailand’s epidemic was most intense in rural areas of remote provinces in the far north and in other border and coastal provinces. HIV prevalence among Thai army recruits from several districts in the far north approached 20 percent during 1991-93, compared to 2-4 percent for other regions.[xxiii] In the mid-1990s, HIV prevalence in pregnant women attending antenatal care peaked over 2 percent for the nation.
From these peaks, Thailand’s HIV prevalence fell dramatically by 2003 to 0.5 percent in army recruits and just over 1 percent in pregnant women.[xxiv] Because studies continued to ignore blood exposures, there is no basis to allocate credit for reduced HIV transmission across multiple changes, including more condom use and safer healthcare. Currently, Thailand appears to have a mixed epidemic, with an estimated 1.5 men infected for every woman, and MSMs and IDUs accounting for substantial numbers of infections.
Cambodia, Myanmar, and Papua New Guinea
The Paris Peace Agreement in 1991 ended decades of war in Cambodia and led to an influx of aid. Changes that came with peace appear to have ignited Cambodia’s HIV epidemic. At the end of the war in 1991, HIV prevalence in blood donors in Cambodia was only 0.1 percent. During the late 1990s, many surveys among sex workers (especially brothel-based sex workers), police, military personnel, and patients seeking treatment for sexually transmitted disease reported HIV prevalence over 5 percent.[xxv] A national survey in 2005 found 0.6 percent HIV prevalence among both men and women, describing a low-level generalized epidemic. Among women, HIV prevalence increased with wealth, but decreased with education (see Table 6.6).
IDUs led the expansion of Myanmar’s HIV epidemic in the early 1990s, with later spread among sex workers and others. From the late 1990s, almost 2 percent of army recruits have been HIV-positive, and comparable rates of HIV prevalence have been found in pregnant women. These rates stabilized after 2000. In 2007, UNAIDS estimated 0.7 percent of adults to be HIV-positive. As of early 2007, a UN website attributed 30 percent of new HIV infections to IDU,[xxvi] which describes a mixed epidemic.
In Papua New Guinea, HIV prevalence among pregnant women attending selected urban antenatal clinics passed 1 percent in 2003-04. UNAIDS estimated adult HIV prevalence at 1.5 percent in 2007 (see the Statistical Annex). As of early 2009, there is no indication that the expansion of Papua New Guinea’s epidemic has been curbed.
Many people who write about international health and development link HIV with poverty. A 2002 article in the American Journal of Public Health generalized that ‘In all societies, regardless of their degree of development or prosperity, the HIV/AIDS epidemic…now affects almost exclusively the most marginalized sectors of society...’[xxvii] A 2004 essay in Lancet declared that ‘poverty reduction will undoubtedly be at the core of a sustainable solution to HIV/AIDS.’[xxviii]
In a recent book, Eileen Stillwagon presents evidence that malnutrition and diseases linked to poverty – such as malaria, tuberculosis, and parasite infections – increase people’s susceptibility to HIV infection, and facilitate HIV transmission from those who are already infected to others.[xxix] For example, schistosomiasis, a parasite infection, causes genital sores that increase risks to acquire and to transmit HIV infection.
Poverty may well impact sexual behavior and biological conditions in ways that at times favor HIV transmission. However, national surveys as well as other studies[xxx] show that people with more wealth are often more likely to be HIV-positive in Africa and also in some Asian countries. Apparently, whatever effects poor nutrition and poverty-related diseases might have to favor HIV transmission, these effects are often overwhelmed by other factors.
The exposures that lead to generalized epidemics appear to vary according to political boundaries. None of the world’s richest and most developed countries has a generalized epidemic. But many of the poorest countries similarly do not have generalized epidemics. For adults who are neither MSMs nor IDUs, high risk to acquire HIV infection is not explained by poverty, but appears to be linked in some other way to living in specific countries in Asia, the Caribbean, and especially Africa.
From 1988 through 2007, the estimated number of HIV infections in countries with generalized epidemics increased from 2.5 million to 26 million. This increase occurred in the face of well-funded HIV prevention programs.
Because HIV prevention programs to date have not consistently stopped generalized epidemics, it is clear that standard programs come with no guarantees for success. There is thus a risk that new generalized epidemics might emerge. Which countries are at risk? Some expert predictions have been far off the mark. For example, in 2002 a US Intelligence Community Assessment predicted 50-75 million HIV infections in China, India, Ethiopia, Russia, and Nigeria by 2010.[xxxi] As of 2009, approximately 7 million people were infected in these five countries, generalized epidemics had not emerged in China or Russia, and low-level epidemics continued in Ethiopia and India.
Probably the best basis for prediction is simply to extrapolate from what has happened to date. For example, in Bangladesh and Indonesia, HIV infections have been observed among IDUs for years without spreading into the general population. In such countries, the conditions – whatever they are – may not support HIV transmission among the general population. On the other hand, African wars appear to suppress generalized epidemics, while Mozambique’s epidemic take-off coincided with the end of its civil war in 1992. Thus, African countries with peace after many years of war – such as Angola, Liberia, Sierra Leone, and Southern Sudan – may be the most likely venues for new and fierce generalized epidemics. But without knowing the factors that allow HIV to spread through the general population, such ‘blind’ extrapolations provide unreliable projections – and no guidance for how to prevent or to restrain generalized epidemics.
[i] UNAIDS. Guidelines for HIV Interventions in Emergency Settings. Geneva: WHO, 1996. p. 2.
[ii] Gisselquist D. ‘Impact of long-term civil disorders and wars on the trajectory of HIV epidemics in sub-Saharan Africa’, J Soc Aspects HIV/AIDS, 2004, 1: 114-27.
[iii] Kaiser R, Spiegel P, Salama P, et al. HIV/AIDS seroprevalence and behavioral risk factor survey in Sierra Leone, April 2002. Atlanta: CDC, no date.
[iv] Mulanga-Kabeya C, Nzilambi N, Edidi B, et al. ‘Evidence of stable HIV seroprevalence in selected populations in the Democratic Republic of the Congo’, AIDS, 1998, 12: 905-10. p. 908.
[v] WHO, Regional Office for Africa (WHO/AFRO). HIV/AIDS Epidemiological Surveillance Update for the WHO Africa Region. Zimbabwe: WHO/AFRO, 2003.
[vi] Health Systems Trust. ‘Angola: HIV infection rate for pregnant women at 2.8%.’, Health Systems Trust, 2004. Available at:
http://www.hst.org.za/news/20040592 (accessed 9 September 2007).
[vii] Finnegan W. A Complicated War: The harrowing of Mozambique. Berkeley: University of California Press, 1992. p. 24.
[viii] Kaiser R, Kedamo T, Lane J, et al. ‘HIV, syphilis, herpes simplex virus 2, and behavioral surveillance among conflict-affected populations in Yei and Rumbek, southern Sudan’, AIDS, 2006, 20: 942-4.
[ix] Spiegel PB et al. ‘Prevalence of HIV infection’. p. 2187.
[x] Strand RT, Dias LF, Bergstrom S, et al. ‘Unexpected low prevalence of HIV among fertile women in Luanda, Angola. Does war prevent the spread of HIV?’, Int J STD AIDS, 2007, 18: 467-71.
[xi] UNAIDS. Guidelines for HIV interventions in emergency settings. p. 2.
[xii] UNAIDS, WHO. Ethiopia: Epidemiological Fact Sheets on HIV/AIDS and Sexually Transmitted Infections, December 2006. Geneva: WHO, 2006.
[xiii] UNAIDS, WHO. Guatemala: Epidemiological Fact Sheets on HIV/AIDS and Sexually Transmitted Infections, December 2006. Geneva: WHO, 2006; UNAIDS, WHO. Panama: Epidemiological Fact Sheets on HIV/AIDS and Sexually Transmitted Infections, December 2006. Geneva: WHO, 2006.
[xiv] UNAIDS, WHO. India: Epidemiological Fact Sheets on HIV/AIDS and Sexually Transmitted Infections, 2004 update. Geneva: WHO, 2004.
[xv] National AIDS Control Organization (NACO). ‘An overview of the spread and prevalence of HIV/AIDS in India’. Available at:
http://nacoonline.org/facts_overview.htm (accessed 9 September 2007).
[xvi] Dandona L, Lakshmi V, Sudha T, et al. ‘A population-based study of human immunodeficiency virus in south India reveals major differences from sentinel surveillance-based estimates’, BMC Med, 2006, 4: 31.
[xvii] International Institute for Population Sciences (IIPS), ORC Macro. National Family Health Survey (NFHS-3), 2005-06: India: Vol. 1. Mumbai: IIPS, 2006. Available at: http://www.nfhsindia.org/volume_1.html (accessed 18 October 2007).
[xviii] Kalyan R. ‘AIDS cases: Dharwad tops list’, Deccan Herald, 7 July 2007.
[xix] Over M, Heywood P, Gold J, et al. HIV/AIDS Treatment and Prevention in India: Modeling the cost and consequences. Washington DC: World Bank, 2004. p. 65.
[xx] Gisselquist D, Correa M. ‘How much does heterosexual commercial sex contribute to India’s HIV epidemic?’, Int J STD AIDS, 2006, 17: 736-42.
[xxi] Siraprapasiri T, Thanprasertsuk S, Rodklay A, et al. ‘Risk factors for HIV among prostitutes in Chiangmai, Thailand’, AIDS, 1991; 5: 579-82.
[xxii] Sawanpanyalert P, Ungchusak K, Thanprasertsuk S, et al. ‘HIV-1 seroconversion rates among female commercial sex workers, Chiang Mai, Thailand: A multi cross-sectional study’, AIDS, 1994; 8: 825-9; Gray JA, Dore GJ, Li Y, et al. ‘HIV-1 infection among female commercial sex workers in rural Thailand’, AIDS, 1997; 11: 89-94.
[xxiii] Sirisopana N, Torugsa K, Mason CJ, et al. ‘Correlates of HIV-1 sereopositivity among young men in Thailand’, J Acquir Immune Defic Syndr, 1996, 11: 492-8.
[xxiv] Punpanich W, Ungchusak K, Detels R. ‘Thailand’s response to the HIV epidemic: Yesterday, today, and tomorrow’, AIDS Educ Prev, 2004, 16 (suppl A): 119-36.
[xxv] US Census Bureau. HIV/AIDS Surveillance Data Base, June 2003 release.
[xxvi] UNDP. ‘Youandaids: The HIV/AIDS portal for Asia.’ Available at:
http://www.youandaids.org/Asia%20Pacific%20at%20a%20Glance/Myanmar/index.asp (accessed 9 September 2007).
[xxvii] Parker R. ‘The global HIV/AIDS pandemic, structural inequalities, and the politics of international health’, Am J Pub Health, 2002, 92: 343-6. p. 344.
[xxviii] Fenton L. ‘Preventing HIV/AIDS through poverty reduction: The only sustainable solution?’ Lancet, 2004, 364: 1186-7. pp. 1186-7.
[xxix] Stillwagon E. AIDS and the Ecology of Poverty. Oxford: Oxford University Press, 2006.
[xxx] Wojcicki JM. ‘Socioeconomic status as a risk factor for HIV infection in women in East Central, and Southern Africa: A systematic review’, J biosoc Sci, 2005, 37: 1-36.
[xxxi] National Intelligence Council. ‘The next wave of HIV/AIDS: Nigeria, Ethiopia, Russia, India, and China.’ Available at: http://www.dni.gov/nic/special_nextwaveHIV.html (accessed 9 September 2007).