3
Epidemiology of the HIV/AIDS Epidemic

As noted in Chapter 1, the global HIV/AIDS epidemic consists of many separate, individual epidemics spread unevenly through sub-Saharan Africa, each with its own distinct characteristics that depend on geography, the specific population affected, the frequencies of risk behaviors and practices, and the temporal introduction of the virus. In addition, biological factors may influence the spread of the epidemic by increasing or decreasing susceptibility to the virus, altering the infectiousness of those with HIV, and hastening the progression of infection to disease and death. Such biological factors may include the presence of classical STDs, male circumcision, and the viral characteristics of both HIV-1 and HIV-2 and their multiple genetic strains.

In sub-Saharan Africa, many of the behavioral patterns and biological conditions that can precipitate rapid HIV transmission were present at the time HIV was introduced into selected populations. Within a relatively brief period of time, massive HIV epidemics were ignited in some areas, affecting over 11 million African adults and resulting in 3 million AIDS-related deaths to date, with many more expected in the next few years (World Health Organization, 1995a). These estimates represent over two-thirds of the worldwide total of all HIV infections and AIDS cases. By the year 2000, as many as 20 million individuals on the continent of Africa will be HIV infected, and at least 8 million people will have died of AIDS (World Health Organization, 1993). It is within the African region that HIV will clearly have its greatest impact on morbidity and mortality, in addition to profound economic, demographic, and social consequences.

This chapter gives an overview of the epidemiology of the HIV/AIDS epidemic



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences 3 Epidemiology of the HIV/AIDS Epidemic As noted in Chapter 1, the global HIV/AIDS epidemic consists of many separate, individual epidemics spread unevenly through sub-Saharan Africa, each with its own distinct characteristics that depend on geography, the specific population affected, the frequencies of risk behaviors and practices, and the temporal introduction of the virus. In addition, biological factors may influence the spread of the epidemic by increasing or decreasing susceptibility to the virus, altering the infectiousness of those with HIV, and hastening the progression of infection to disease and death. Such biological factors may include the presence of classical STDs, male circumcision, and the viral characteristics of both HIV-1 and HIV-2 and their multiple genetic strains. In sub-Saharan Africa, many of the behavioral patterns and biological conditions that can precipitate rapid HIV transmission were present at the time HIV was introduced into selected populations. Within a relatively brief period of time, massive HIV epidemics were ignited in some areas, affecting over 11 million African adults and resulting in 3 million AIDS-related deaths to date, with many more expected in the next few years (World Health Organization, 1995a). These estimates represent over two-thirds of the worldwide total of all HIV infections and AIDS cases. By the year 2000, as many as 20 million individuals on the continent of Africa will be HIV infected, and at least 8 million people will have died of AIDS (World Health Organization, 1993). It is within the African region that HIV will clearly have its greatest impact on morbidity and mortality, in addition to profound economic, demographic, and social consequences. This chapter gives an overview of the epidemiology of the HIV/AIDS epidemic

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences in sub-Saharan Africa, including its status and modes of transmission. The chapter ends with a discussion of remaining gaps in knowledge and a set of recommendations for future research. STATUS OF THE EPIDEMIC The origin of HIV continues to be an enigma, and the timing of the first human infection remains unknown. Attempts to determine the origins of the disease led to early speculation that AIDS originated in Africa. Not surprisingly, this speculation led many African leaders to resent the implication that Africans were to blame for AIDS. The controversy about the origin of AIDS resulted in a "backlash" and denial that HIV even existed within high-risk populations in sub-Saharan African countries and proved very unhelpful for designing effective prevention programs. Efforts to acknowledge that the problem existed and to initiate efforts to control its spread were delayed in some countries for several years. Because theories about where and when AIDS originated have become so entangled in politics, and because the epidemic is now too far advanced for the question to really matter, attempts to find definitive answers to these questions have been given a low priority. There are a few isolated reports in the literature in the 1970s and even earlier of people dying of opportunistic infections that have now become known as the trademarks of AIDS (Henig, 1993). However, AIDS was not recognized as a clinical entity until 1981. In Africa the first reports of AIDS-like syndromes and "slim" appeared in the literature between 1983 and 1985 (Van de Perre et al., 1984; Piot et al., 1984; Serwadda et al., 1985). Since the early 1980s, the prevalence of HIV infection among certain populations in Africa has increased dramatically, and it is expected to grow even more rapidly in the future. Factors in the spread of HIV are discussed in Chapters 2 and 4 with regard to the larger societal context and individual attitudes and behavior, respectively. HIV/AIDS Statistics As of December 1994, nearly 350,000 AIDS cases had been reported from the African region (World Health Organization, 1995a). As noted earlier, this sum represents one-third of the global number (1,025,073) of AIDS cases reported since the start of the epidemic. Allowing for under-diagnosis, incomplete reporting, and reporting delays, WHO estimates that more than 3 million cases of AIDS have occurred in Africa, comprising 70 percent of the global total of 4.5 million. In sub-Saharan Africa, 11 million adults are estimated to have been infected with HIV. This number represents nearly two-thirds of the estimated 18 million cumulative HIV infections that have occurred worldwide. More than half of these 11 million infected adults are women, and as many as 1 million African

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences children are estimated to have been infected as a result of mother-to-child transmission (World Health Organization, 1994). As discussed previously, sub-Saharan Africa is geographically, demographically, socially, and culturally heterogeneous, and the extent and spread of HIV infection and AIDS have accordingly been heterogeneous in the region. Thus, it is difficult if not impossible to generalize about the AIDS epidemic within the region. Yet some overall characteristics and trends can been seen. Wherever possible, we provide specific examples, with the proviso that the quoted rates of infection are pertinent only to the specific population and geographical area for which they are cited. There have been only a few nationally or regionally representative seroprevalence studies conducted to date in sub-Saharan Africa, and information is available predominantly on the groups with the highest risk of HIV infection. In addition, sentinel surveillance systems have been developed to monitor changes in the levels of HIV infection among specific segments of the population, including those with high-risk behaviors, such as women engaged in commercial sex activities or patients receiving care for STDs, as well as groups more representative of the general population (i.e., at lower risk), such as blood donors and women seeking prenatal care. HIV prevalence is not uniformly distributed even among all countries of sub-Saharan Africa. As described in Chapter 1, to date the epidemic has disproportionately affected East and Southern Africa. Among certain urban populations in the worst-afflicted parts of the region, such as those in Kigali, Rwanda, and Kampala, Uganda, up to one in every three adults is infected with HIV (Rwandan HIV Seroprevalence Study Group, 1989; Ministry of Health [Uganda], 1989). Overall HIV-1 infection patterns for lower-risk urban populations in Africa are shown in Figure 3-1. High levels of infection (in excess of 10 percent) are found among these populations in many urban areas throughout East and Southern Africa and in Abidjan, Côte d'Ivoire, in West Africa (U.S. Bureau of the Census, 1994a). Recent data suggest that HIV seroprevalence is still low in most rural as compared with urban settings (Figure 3-2). However, HIV infection appears to be increasing in rural areas as well. In Tanzania, the Bukoba district probably always had a higher HIV seroprevalence than Dar es Salaam. However, even within the Bukoba district, urban centers exhibit higher rates of infection than do rural areas: 24 percent versus 5 percent, respectively (Mhalu et al., 1987; Schmutzhard et al., 1989). Even within a particular geographic area, some population groups are disproportionately affected by the epidemic. The highest infection rates are usually found among men and women between 20 and 40 years old; people with STDs and tuberculosis; and, as discussed in Chapter 2, certain occupational groups, such as long-distance truck drivers, military personnel, and women employed in the commercial sex and entertainment industries (including those who work in

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences FIGURE 3-1 African HIV-1 Seroprevalence for Lower-Risk Urban Populations. SOURCE: U.S. Bureau of the Census (1994b). bars and hotels). HIV infection rates of well over 80 percent have been reported for commercial sex workers in East and Central Africa (Piot et al., 1987, 1988; Quinn, 1991). Figure 3-3 shows the levels of HIV seroprevalence among populations of commercial sex workers in selected countries. Indeed, the AIDS epidemic in each country can be seen as a series of epidemics among subpopulations with varying levels of risk. For example, in Nairobi, Kenya, available data clearly show HIV infection spreading first and most extensively to commercial sex workers, followed by STD clinic patients—no doubt including many clients of those commercial sex workers (Piot et al., 1987). Finally, infection can be seen to be spreading among the general population, as evidenced by the initially slow but accelerating spread among pregnant women (Figure 3-4). In serologic surveys, pregnant women are often used as surrogates for the general population. Such surveys are convenient because in many countries, pregnant women attend government clinics to receive prenatal care and may be readily tested there. To some extent, pregnant women can be considered as being at slightly higher risk than the general population because they are demonstrably sexually active. Moreover, they are drawn from a limited age range and tend to

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences FIGURE 3-2 Urban/Rural Differentials in HIV Infection. SOURCE: U.S. Bureau of the Census (1994a). be younger than adult women in general, given typical patterns of age-specific fertility rates. The population of pregnant women may also be biased toward those in marital (formal or informal) unions. Nevertheless, in many countries data on pregnant women provide the most representative picture of HIV infection among the general population. Seroprevalence data from a number of studies of pregnant women conducted since the mid-1980s demonstrate the heterogeneity mentioned above (Figure 3-4) (U.S. Bureau of the Census, 1994c). There has been a consistent and rapid increase in HIV infection levels among pregnant women in Francistown, Botswana; Blantyre, Malawi; and Kampala, Uganda. By 1992, between 25 and 35 percent of pregnant women in these cities were infected. Infection rates among pregnant women rose at a much more moderate pace in Nairobi, Kenya; Bangui, the Central African Republic; and Dar es Salaam, Tanzania. However, they still reached 15 percent by 1993. Infection levels among pregnant women in Abidjan, Côte d'Ivoire, increased rapidly to about 10 percent by 1987, appeared to have reached a plateau below 15 percent through 1991, but have increased recently above 15 percent. Meanwhile, infection levels in Kinshasa, Zaire, have

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences FIGURE 3-3 HIV Seroprevalence for Commercial Sex Workers in Sub-Saharan Africa: Circa 1992. NOTE: Includes infection from HIV-1 and/or HIV-2. SOURCE: U.S. Bureau of the Census (1994a). FIGURE 3-4 HIV Seroprevalence for Pregnant Women in Selected Urban Areas of Africa: 1985-1994. NOTE: Includes infection from HIV-1 and/or HIV-2. SOURCE: U.S. Bureau of the Census (1994a).

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences been relatively stable at 5 to 6 percent since the mid-1980s (Piot et al., 1990; Piot and Tezzo, 1990). Although the epidemic in Africa was first recognized in Central and East Africa, and these regions continue to have the highest infection levels, there is increasing evidence that the epidemic is spreading into West and Southern Africa. In Abidjan, Cô d'Ivoire, in West Africa, HIV-1 prevalence among adults increased from 1 percent in 1986 to more than 15 percent in 1992. In Nigeria, a country with more than 105 million inhabitants, the largest population in the region, studies indicate that the prevalence of HIV-1 and HIV-2 among prostitutes in Lagos may be rising rapidly (Dada et al., 1993; Olaleye et al., 1993; see also below). Within the last few years, investigators have noted the introduction of HIV among high-risk populations in Nigeria. Although relatively rare during the late 1980s, HIV has been increasing since 1990 throughout Nigeria, according to several surveys. This trend is of critical importance since the population of Nigeria, estimated at over 105 million, represents more than one-sixth of the total population of sub-Saharan Africa (World Bank, 1995). In one recent study, 12.3 percent and 2.1 percent of 885 female prostitutes in Lagos State were infected with HIV-1 and HIV-2, respectively, a rise from a combined prevalence of only 1.7 percent 2 years previously (Dada et al., 1993). Women in the youngest age group, ages 12 to 19, had the highest prevalence (20 percent). In addition, prostitutes residing in the port area of Lagos, which serves as a major convergence of overland and sea routes within and outside Nigeria, had the highest prevalence of HIV-1 infection. A highway region that is traversed by the overland interstate highway also had high rates. Because Lagos is the largest cosmopolitan city in Africa, the constant migratory movement of people into and out of this major trade center provides further opportunity for HIV dissemination. The virus may be spreading even more rapidly in Southern Africa than in West Africa. For example, in Botswana, HIV prevalence among pregnant women increased from 10 percent in 1991 to 34 percent in 1993 in Francistown, and from 6 percent in 1990 to 19 percent in 1993 in Gaborone (U.S. Bureau of the Census, 1994c). Similar disturbing data are emerging from South Africa, suggesting a three-fold increase in HIV prevalence between 1990 and 1993 among women attending prenatal clinics in most regions of the country. Aggregated data collected in prenatal clinics across South africa show a rapid increase in overall prevalence from under 1 percent in 1991 to 1.7 percent in 1992, 2.8 percent in 1993, and 6.4 percent in 1994 (U.S. Bureau of the Census, 1994a). Thus, although HIV infection rates are high among many populations and subgroups in sub-Saharan Africa, there remains much variation in incidence and prevalence rates recorded to date, both geographically and by population subgroups. The probable causes of this heterogeneity in seroprevalence are multiple, and include behavioral, biological, and societal factors. Trying to explain the phenomenon by a single factor such as civil war, male circumcision, STDs, or

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences rate of partner change is simplistic. Instead, it appears that the simultaneous occurrence of several risk factors for HIV transmission determines how rapidly and to what level HIV spreads among the population and who will become infected. In the absence of facilitating factors, HIV infection could remain endemic at low levels over long periods of time until a critical prevalence of infection is reached, and the spread of HIV-1 accelerates (Nzilambi et al., 1988). This epidemiologic diversity not only reflects differences in sexual and other behaviors, but also suggests that the epidemic has not reached an equilibrium in most areas. The epidemiological evidence suggests that HIV prevalence may be stabilizing in some large urban centers (see Figure 3-4) and potentially in some rural areas (Wawer et al., 1994b; U.S. Bureau of the Census, 1994a). However, it must be recognized that a stable prevalence can conceal a significant level of new HIV infection replacing those who die (Wawer et al., 1994b). For example, in the absence of migration, a stable adult seroprevalence of 20 percent suggests that up to 2 percent of adults become newly infected each year, replacing the approximately 10 percent of the infected who are expected to die annually in African settings. Demographics of HIV Infection The HIV epidemic and the demographic structure of sub-Saharan populations will have complex interactions over time. The population of sub-Saharan Africa is predominantly young, in sharp contrast with the age structure in developed countries; 45 percent of the population of the region is under the age of 15, compared with one-third or less for the other major geographic regions (Decosas and Pedneault, 1992; Quinn, 1994). Among persons aged 15 and over, those in the 15-39 age group represent over two-thirds of all sub-Saharan adults; only in Latin America do young adults so predominate, whereas in Asia and the developed regions, young adults represent at most half of all adults (United Nations, 1993). In urban areas, one finds a prominent one-sided bulge caused by the migration of young males into the cities for employment (with some rural areas reporting a proportional ''deficit" of young males who have migrated away) (Serwadda et al., 1992). For example, the prevalence of HIV infection among both urban and rural populations in Uganda is highest in the 25-to-44-year-old bracket among males and in the 15-to-34-year-old bracket among females (Figure 3-5) (Wagner et al., 1993; Serwadda et al., 1992). A significant contributor to the elevated prevalence of infection in sub-Saharan Africa is the fact that behavioral factors associated with HIV transmission—including multiple partners and impermanent relationships—are generally more common among the young, this coupled with the high proportion of young adults found in sub-Saharan African countries (Anderson et al., 1991). Accordingly, the large number of young persons under age 15, who will soon enter their

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences FIGURE 3-5 HIV Seroprevalence of Adult Population of Rural Rakai District, Uganda, by Age and Sex, 1992. SOURCE: Maria Wawer (personal communication, 1995). sexual and reproductive lives, must represent a priority group for AIDS and STD prevention. The difference in the age distribution of peak HIV prevalence between men and women occurs in the region because, on average, sexual partnerships are formed between older men and younger women (see also Chapter 2). The distortion of the urban population profile caused by male migration initially resulted in equal numbers of infected men and women (Quinn, 1994). However, male-to-female transmission of the virus is more efficient than female-to-male transmission in the absence of other cofactors (Haverkos and Quinn, 1995), so that as the epidemic has spread into the larger rural population, the absolute number of infections has become higher among women than men (Rowley et al., 1990; Anderson et al., 1991). HIV-2 Infection One unique feature of the AIDS epidemic in Africa is the remarkable viral heterogeneity of HIV infection. Within HIV-1 there are now nine recognized

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences subtypes, labeled A through I, as well as the more recently identified subtype O. In addition to HIV-1, which is common predominantly in Central, Southern, and East Africa, a more distinct variant, labeled HIV-2, has been identified, predominantly in West African countries. Originally identified in 1986, HIV-2 was recognized among high-risk populations such as commercial sex workers in urban centers in West Africa. While having some genetic relationship to HIV-1, in evolution HIV-2 may be more closely related to a simian immunodeficiency virus (SIV) (Kanki, 1994). Although similar in terms of morphology, cell tropism, and overall genetic organization, HIV-1 and HIV-2 differ significantly in terms of nucleotide sequences, with only 42 percent homology (Clavel et al., 1986; Guyader et al., 1987). Genomic studies further demonstrate that HIV-2 has 70 percent or more homology with SIV. Genetic sequencing of HIV-2 isolates also shows a wide divergence among individual strains of HIV-2, similar to that observed with HIV-1. Thus, it is highly probable that the divergence of HIV-1 and HIV-2 occurred earlier than the beginning of the current epidemic (Myers, 1994). A common ancestor with similar properties and pathogenic potential may have existed a long time ago (Myers, 1994); the spread of HIV in Africa was most likely a result of simultaneous modifications of epidemiologic parameters in West and Central Africa, such as rapid urbanization and increased mobility, leading to infection of larger populations with HIV-1 and HIV-2 (Rowley et al., 1990; Anderson et al., 1991; Decosas et al., 1995) (see also Chapter 2). Although HIV-2 can cause AIDS, it is increasingly clear that its pathogenic potential is lower than that of HIV-1. In cross-sectional studies, individuals infected with HIV-2 were found to have immunologic abnormalities similar to although less marked than those associated with HIV-1. In a prospective study among prostitutes in Senegal who were HIV-2 positive, no reduction in CD4 lymphocyte levels and no clinical abnormalities were found (Marlink, 1994). A less aggressive course of HIV-2 infection is also suggested by other observations in Dakar, Senegal; whereas HIV-2 is predominant among asymptomatic people, HIV-1 is more frequent among hospitalized patients with AIDS (Poulsen et al., 1993; Kanki et al., 1994; Marlink, 1994). The routes of transmission and risk factors for HIV-1 and HIV-2 are similar. Like HIV-1, HIV-2 is transmitted primarily sexually (Kanki et al., 1994). However, the latency period for HIV-2 appears to be longer, and vertical transmission of HIV-2 from mother to infant is rare (Matheron et al., 1990; Poulsen et al., 1992; Adjorlolo-Johnson et al., 1994). HIV-2 infection rates have risen steadily over the past two decades in countries of West Africa, including Côte d'Ivoire, Senegal, Guinea-Bissau, Burkina Faso, The Gambia, and Cape Verde (Kanki, 1991; Naucler et al., 1991; Markovitz, 1993; Poulsen et al., 1993). In several urban centers of West Africa, 15 to 64 percent of female prostitutes are infected. In Guinea-Bissau and The Gambia, HIV-2 is the prevalent infection, and HIV-1 is rare. In Côte d'Ivoire and Burkina

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences FIGURE 3-6 African HIV-2 Seroprevalence for Lower-Risk Urban Populations. SOURCE: U.S. Bureau of the Census (1994b). Faso, HIV-2 and HIV-1 are both present among appreciable proportions of the population. The geographic pattern of HIV-2 shows a higher prevalence in West Africa and in other African countries with a Portuguese colonial history (Kanki, 1991). Troop movements among these former Portuguese colonies and travel facilitated by cultural ties surely contributed to the spread of HIV-2 in these select countries. Conversely, several countries bordering those with substantial HIV-2 infection have as yet shown little evidence of an HIV-2 epidemic. The highest prevalence of HIV-2 infection among high-risk urban adults is found in Côte d'Ivoire and The Gambia, where infection rates are 37 percent and 27 percent, respectively. HIV-2 seroprevalence among low-risk urban adults is far lower (see Figure 3-6); only in Guinea-Bissau does seroprevalence exceed 10 percent (U.S. Bureau of the Census, 1994b). The highest HIV-2 infection rates are found among populations with high HIV-1 prevalence, including people with tuberculosis or STDs and female prostitutes. As noted above, in contrast to HIV-1, which shows a distinct peak at ages 25 to 40, the age-specific prevalence of

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences seeking for STDs, and patterns of alcohol and illicit drug use (which influence sexual behaviors). Of particular interest is the potential effect of a vaccine against HIV on the course of the epidemic. The efficacy of a vaccine can be expressed as the product of three factors (Blower and McLean, 1994): the vaccine take (the fraction of recipients in whom the vaccine induces any immunological effect), the degree of the vaccine (the reduction in susceptibility per sexual partnership among those in whom the vaccine takes), and the fraction of those ceasing sexual activity before the vaccine-induced protection wanes. Whether a vaccine could ultimately lead to eradication of HIV and if so, the extent of coverage necessary to achieve eradication depend on R0, the reproductive rate of HIV. Modeling results show that vaccines with moderate efficacy or those administered to a population with a severe HIV epidemic (as measured by R0) could not achieve eradication; for example, 100 percent of the population at risk would need to be vaccinated if R0 were 2.0 and the efficacy of the vaccine were 50 percent (Blower and McLean, 1994). These calculations assume no change in risk behavior. Perinatal Transmission The second major mode of HIV transmission in Africa is perinatal, which accounts for approximately 15 to 20 percent of all AIDS cases in sub-Saharan Africa, in contrast with 5 to 10 percent worldwide (Quinn et al., 1994). The large numbers of infected children in Africa are explained by the high proportion of women infected with the HIV virus and the large number of children each women bears. Serologic surveys of pregnant women in Africa find that between 6 and 30 percent are HIV-positive (U.S. Bureau of the Census, 1994c). Sub-Saharan Africa accounts for three of every four women who have been infected with HIV worldwide (World Health Organization, 1995b). Perinatal transmission may occur in utero through transplacental infection, at the time of delivery, or through breastfeeding or other routes. The probability of mother-to-child transmission varies according to different studies: 27 percent in Kampala, Uganda; 30 percent in Kigali, Rwanda; 39 percent in Lusaka, Zambia; 39 percent in Nairobi, Kenya; 39 percent in Kinshasa, Zaire; and 42 percent in Brazzaville, Congo (Ryder et al., 1989; Hira et al., 1989; Lallemant et al., 1989; Miotti et al., 1990). In comparison, transmission rates have been lower in North America and Europe, ranging from 7 to 30 percent (Blanche et al., 1989; Rogers et al., 1989; Oxtoby, 1990; European Collaborative Study, 1991). Unfortunately, the results of various studies published to date are not strictly comparable because of differences in recruiting strategies for prospective studies and the criteria used to determine HIV infection among children. Nevertheless, risks of perinatal transmission reported in African studies appear to be generally higher than those reported in North American and European studies, probably because of large differences between the duration and intensity of breastfeeding by seropositive

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences women in Europe and North America compared with women in sub-Saharan Africa. To resolve the issue of lack of standardization of study protocols, a consensus meeting was held in Ghent, Belgium, in 1992. It resulted in a definition of HIV infection in children and a clinical classification of HIV-infected mothers. Factors that affect the probability of perinatal transmission include the disease stage and immune status during pregnancy, as measured by CD4 and CD8 cell counts; the conditions of pregnancy and delivery; the particular viral strain involved; the infectious, parasitic, and nutritional environment in which the mother and child live; the presence of chorioamnionitis and funisitis; and whether the infant is breastfed (Ryder et al., 1989; St. Louis et al., 1993; Semba et al., 1994). Of these factors, the rate of transmission associated with breastfeeding is most difficult to evaluate because infants are often exposed to HIV infection during pregnancy and at birth, as well as postnatally. Although the majority of transmission occurs pre-or intrapartum, there have been several documented cases of postnatal transmission to the infant in which the mothers were infected after delivery (Van de Perre et al., 1991; Lepage et al., 1987). Transmission in these cases occurred from mother to infant during the first year of life while the infant was being breastfed. Postnatal transmission in these instances was probably facilitated because the mothers were seroconverting and therefore had a high level of viremia, and because no immunity was passively transferred to the infants transplacentally or through breast milk. The rate of postnatal transmission estimated from these studies might be higher than that estimated for asymptomatic seropositive mothers who breastfed their infants, but is still likely to be low. The majority of infants who are infected with HIV-1 acquire the infection in utero or during childbirth. When the mother is infected prenatally, the additional risk of HIV-1 transmission via breastfeeding is estimated to be 14 percent (Dunn et al., 1992); when the mother is infected postnatally, the risk of HIV-1 transmission is 29 percent (Dunn et al., 1992). As noted earlier, the risk of perinatal transmission is much higher for HIV-1 than for HIV-2. All babies born to an HIV-infected mother carry passively acquired maternal antibodies to HIV. Those infants who are not infected will gradually lose those antibodies, which may nevertheless persist in some cases beyond a year. Since standard tests for HIV can detect only HIV antibodies and not the virus itself, they cannot be used reliably to determine which infants born to HIV-positive mothers have been infected until the maternal antibodies have been lost (Hardy, 1991). The problem, therefore, is that the HIV status of infants born to HIV-infected mothers cannot be ascertained until well after birth. It is possible that a new inexpensive HIV test will be developed that can reliably yield positive results only if the infant is HIV-positive when cord blood is tested, although a negative result would not mean conclusively that the infant was HIV-negative (Miles et al., 1993). If such a test were developed, HIV-positive mothers might be advised that an infant who tested positive could be breastfed.

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences There has been much discussion concerning whether breastfeeding should be discouraged in areas where HIV is very prevalent. In 1992, a special group representing WHO/UNICEF concluded that breastfeeding should be promoted in all developing countries, regardless of HIV infection rates (World Health Organization, 1992). Breastfeeding provides a mechanism for increased spacing of births, as well as better nutrition and protection against diarrheal diseases, pneumonia, and other infections. Where the primary causes of infant deaths are infectious diseases and malnutrition, the benefits of breastfeeding outweigh the risk of HIV transmission via breastfeeding, even for women known to be infected with HIV. However, in areas with low infant mortality rates from infectious diseases, women known to be infected with HIV should be advised to use a safe feeding alternative to breastfeeding; women whose HIV status is unknown should be advised to breastfeed (World Health Organization, 1992). Several studies support the WHO/UNICEF recommendations (Choto, 1990; Kennedy et al., 1990; Nicoll et al., 1990; Ryder et al., 1991; Dunn et al., 1992; Hu et al., 1992). Mortality among HIV-infected infants is much higher in Africa than in North America or Europe. In examining mortality rates, it is necessary to distinguish between African studies, in which survival is reported globally for infants born to seropositive mothers because of difficulty in diagnosing HIV infection during the first year, and American studies, in which mortality is often reported among infected infants only (Quinn et al., 1994). In Africa, the ultimate cause of death is not easily determined because the diagnostic tools are lacking, and because infants whose clinical deterioration is rapid do not always reach the hospital before they die. In Kinshasa, Zaire, and Brazzaville, Congo, mortality at 12 months was found to be 21 and 37 percent, respectively, for children born to HIV-seropositive mothers and 4 percent for controls (Ryder and Hassig, 1988). In a study in Malawi, mortality rates were 32 percent for the first 24 months for infants born to seropositive mothers and 11 percent for controls (Taha et al., forthcoming). In rural Rakai, Uganda, infant mortality rates were 210 per 1,000 live births for children born to HIV-seropositive mothers and 111 per 1,000 for those born to HIV-seronegative mothers (Sewankambo et al., 1994). High mortality rates among HIV-infected children are due not only to the direct effects of HIV, but also to the profound disruption of the family unit associated with the infection (Preble, 1990). In many cases, the parents themselves are incapacitated by HIV infection, becoming progressively less capable of caring for their families. Parental loss and worsening socioeconomic status affect the survival of children in the family regardless of HIV serologic status. In countries with large numbers of HIV-seropositive women, the impact of AIDS on overall childhood survival is already being felt. In Zimbabwe, approximately half of pediatric hospital admissions were from HIV-associated illness. However, the extent of illness among hospitalized children that is due to HIV infection and AIDS is unknown because of the difficulty involved in making a diagnosis in young children. Using available HIV seroprevalence data for women

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences living in several African countries, Valleroy et al. (1990) estimated the percentage increase in infant-child mortality rates that is due to HIV-attributable mortality. They estimated that infant mortality rates would increase by 6 to 38 percent in Kampala and by 1 to 6 percent in Nairobi as a result of HIV infection (Valleroy et al., 1990). WHO also estimated that within the previous 10 years, nearly 500,000 infants in Africa had been born with HIV infection (Chin, 1990). By the year 2000, there will be an additional 10 million HIV-infected children in Africa. In addition, 5 to 10 million children under age 10 are expected to become orphans during the 1990s because of the death of one or both parents from AIDS. During the decade, it is estimated that infant and child mortality rates in some African countries will increase by 50 percent as a result of AIDS. Parenteral Transmission The third mode of HIV infection is parenteral transmission, which includes blood transfusions, injections, and scarification. This mode represents less than 10 percent of all HIV cases in sub-Saharan Africa (Piot et al., 1988). Blood transfusion with HIV-infected blood is known to be a very efficient means of transmission, with over 90 percent of recipients becoming infected. Unfortunately, blood screening, which is universal in industrialized countries, is not widely available in many developing areas of the world. The impact of not screening is substantial. In Central Africa, HIV seroprevalence is between 2 and 18 percent among blood donors (Mhalu and Ryder, 1988). The public health impact of exposing African populations to unscreened blood units has been documented in several countries. In one survey of 2,384 health care workers in Kinshasa, 9 percent of HIV-seropositive individuals had received blood transfusions as compared with 5 percent of HIV-seronegative individuals (Mann et al., 1986c). In a study of children who were admitted to general pediatrics or measles wards and whose mothers were HIV-seronegative, 31 percent of seropositive children had received blood transfusions, as compared with only 7 percent of seronegative children (Mann et al., 1986a). In another study of older children admitted to a pediatrics ward but not diagnosed with AIDS, 60 percent of children infected with HIV had received blood transfusions, whereas only 33 percent of the HIV-seronegative children had received transfusions (Mann et al., 1986b). HIV transmission via blood transfusion is often associated with preventable endemic tropical diseases. In pediatric populations, malaria-associated anemia is highly prevalent, and patients are often given multiple transfusions for treatment. In one study in Kinshasa, 87 percent of blood transfusions at one hospital had been given to children for malaria-induced anemia. In that study, it was estimated that as many as 561 new pediatric cases of HIV infection would occur each year in one hospital if donated blood were not screened for HIV (Greenberg et al., 1988). With increasing awareness of the transmission of HIV through blood transfusion,

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences transmission through infected blood and blood products is being reduced as appropriate screening of donated blood is introduced. Other measures being taken to protect the blood supply include recruiting donors from among low-risk population groups on a voluntary and unpaid basis. Health-care workers are also being encouraged to revise their guidelines on transfusion to ensure that the procedure is carried out only when absolutely necessary, and that saline solutions are used as blood substitutes whenever possible. Needles and other sharp instruments used by traditional African healers are an unproved but possible important mode of HIV transmission (N'Galy et al., 1988; Berkley et al., 1989). Traditional healers normally establish patient practices in a village, or have a mobile practice in which they visit a circuit of different towns and villages. Many practitioners give their patients injectable antibiotics; injection equipment used is sterilized poorly or not at all. Ironically, the practice of giving prophylactic antibiotics may facilitate HIV transmission because the patient receives an additional injection with blood-contaminated needles and syringes. Because needle use is so ubiquitous in Africa, it is difficult to determine a true causal relationship between needle use and needle transmission. Cosmetic scarification, with its custom of using communally shared cutting utensils, is another possible mode of HIV transmission, as is tattooing with unsterilized needles. HIV transmission through the use of unsterilized paraphernalia by injecting drug users has not been documented as a major mode of HIV transmission in Africa. In contrast with the situation in developed countries, injectable drugs such as heroin or cocaine are not commonly found or used in Africa, although they are becoming more popular in certain port areas and among the more affluent population (World Health Organization, 1994). The low incidence of injectable drug use in Africa has been attributed to the expense of the drugs and associated paraphernalia. REMAINING GAPS IN KNOWLEDGE Limitations of Existing African Behavioral Data With a more complex set of data, it would be possible to understand better the social and behavioral processes that underlie the HIV/AIDS epidemic. To develop more effective AIDS-and STD-prevention strategies, additional information is needed on sexual behaviors (including sexual networking), particularly their determinants (as discussed in Chapter 4), and on barriers to the effective and rapid adoption of preventive measures, including reduced numbers of partners, safe and nonpenetrative sex, condom use, and care seeking for symptoms of STDs (as discussed in Chapter 5). Specifically, data on timing of entrance into a sexual network, dominant sexual practices, and timing of permanent or transitory exit from and re-entry into sexual networks would provide a better understanding

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences of the social and behavioral aspects of HIV transmission. More complex data would also provide information on the conditional probability of infection and its association with particular practices, the presence of other diseases, and possible duration dependencies. Unfortunately, social science research related to sexuality, AIDS, and STDs in Africa has most frequently been conducted in urban areas and among groups that fit the Western concept of high-risk behavior, such as commercial sex workers and the military; other groups in the general population that may also be at high risk because of elevated underlying HIV prevalence are less likely to be contacted (Udvardy, 1990). Logistical problems associated with community-based research have further resulted in a preponderance of urban clinic, hospital, and high-risk group studies, and more recently, community-based urban and rural serosurveys involving little or no behavioral research (Kaheru, 1989; Rwandan HIV Seroprevalence Study Group, 1989). The number of studies combining HIV serologic and behavioral research among representative community-based populations, particularly among rural dwellers, remains small (Konde-Lule et al., 1989; Killewo et al., 1990; Serwadda et al., 1992; Mulder et al., 1994b). The need for such research is underlined by the fact that the vast majority of Africans reside in rural areas, and may account for the bulk of the region's HIV infection. Only recently have researchers started to collect data on actual African sexual practices and patterns. It has been noted that ''anthropologists have devoted relatively little attention to the systematic study of sexual behavior. There are many studies of marriage and divorce, of the social, economic and ritual roles of women, of changes in male-female relationships and in other institutionalized forms of gender behavior. Where sexual practices are mentioned, however, they are often of a generalized nature" (Brokensha, 1988:167-168). Similarly, data on sexual networks and the determinants of partner selection remain limited in quantity and scope (Orubuloye et al., 1990; Obbo, 1993), so that empirical findings are incomplete at best. We still have much to learn about the acceptance of and barriers to other aspects of HIV/AIDS prevention, including condom use. As yet, there are inadequate data on care seeking for STDs and AIDS, and on the determinants of acceptance of HIV serological testing. When considering available data, as well as data that will become available in the next decade, we must be careful not to overgeneralize findings from one African setting to another (Ntozi and Lubega, 1990). As emphasized throughout this report, Africa is culturally diverse, and neighboring groups of closely related peoples can have very different cultural expectations. Given evidence of variability in sexual beliefs and practices, the utility of overarching models of African sexuality has been questioned (Schoepf, 1990). To summarize, available social and behavioral data have limitations from the viewpoint of supporting the development of more effective AIDS-prevention strategies or projecting future transmission. Little is known about the range of

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences sexual options open to individuals; the types of sexual practices conducted with different partners; and changes brought about by migration, various degrees of urbanization, and AIDS itself. The next chapter surveys the available knowledge on these issues. Chapter 7 examines means of building an indigenous capacity for HIV/AIDS-related research in Africa. The Need to Combine Epidemiological and Social/Behavioral Research To date, the interpretation and utility of much epidemiological and social/behavioral research have been limited by the lack of a multidisciplinary approach. Data on reported behavior change may be difficult to assess in the absence of biological validation that such change is sufficient to reduce STD/HIV acquisition. Efforts to model the demographic effects of the HIV/AIDS epidemic are hindered by a paucity of data sets that combine fertility, mortality, migration, and other sociodemographic information with HIV serology. Conversely, serological studies that fail to collect adequate behavioral data miss an important opportunity to assess the effects of factors such as sexual practices, sexual networks, and injecting drug use practices within given populations. There is also a growing realization that the design, execution, and analysis of clinical trials for HIV vaccines, STD control, antiretroviral drugs, and genital barrier methods/viricides all depend on appropriate behavioral research to guide enrollment, ensure adherence to trial protocols, and permit adequate interpretation of epidemiological results (including the very basic need to control for potential differential behavioral change among study groups). The disjunction between epidemiological and social/behavioral research has been due in part to a perception by behavioral scientists that biological specimen collection is difficult, intrusive, and unacceptable to subjects and to caution among clinical/epidemiological researchers in posing questions about potentially intimate behaviors within the context of studies that collect substantial biological samples. However, such reluctance to implement multidisciplinary research is rapidly losing its rationale as the STD/HIV epidemic continues to intensify and as techniques for and experience in the application of combined epidemiological/social science studies improve. On the biological front, assessment of HIV prevalence/incidence has been greatly facilitated by the development of serological collection methods that do not call for venous blood collection. Finger prick/filter paper and saliva tests are now well established as tools in HIV epidemiology (Behets et al., 1992; Belec et al., 1994; Nyambi et al., 1994; Pappaioanou et al., 1993; Frerichs et al., 1994), and HIV assessment from urine samples is under development (Cao et al., 1988; Berrios et al., 1995). Such samples can be collected by lay personnel in nonclinic settings. Urine samples can also be used to quantify gonorrhea and chlamydia prevalence (Chernesky et al., 1994; Lee et al., 1995; Smith et al., 1995), and there is positive experience with home-based collection of self-administered vaginal

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences swabs for trichomonas culture (Wawer et al., 1995a) and the determination of bacterial vaginosis (Nugent et al., 1991; Speigel et al., 1983). From the behavioral viewpoint, detailed data on sexual networks and practices have been collected on subjects in diverse cultural settings (Dyson, 1992; Caldwell et al., 1993; Orubuloye et al., 1994). A few studies have integrated biological specimen and detailed behavioral information collection in community surveys and have achieved high participation rates (Mulder et al., 1994a; Wawer et al., 1995a). The Demographic and Health Surveys (DHS) project is currently planning a pilot survey that would combine biological sample collection with standard DHS sociodemographic and contraceptive information (Cynthia Stanton, personal communication, 1995). Although experience is still limited, data suggest that a combined epidemiological/social science research approach will prove acceptable in many clinic and population-based settings. Ethical Issues in STD/HIV Research According to guidelines for research involving human subjects developed jointly by the Council for International Organizations of Medical Sciences (CIOMS) and WHO, when research is conducted by investigators of one country on subjects of another, the ethical standards applied should be no less exacting than if the research were carried out in the initiating country (Council for International Organizations of Medical Sciences and World Health Organization, 1992). It has also been noted, however, that "the great complexity, varied presentation, and wide distribution of HIV infection challenge this stance" (Christakis, 1988:31) and that the stated purpose of the CIOMS/WHO guidelines is, in part, to anticipate such issues and suggest they can be applied to the special circumstances of developing countries (Christakis, 1988). Ethical standards related to biological and behavioral data collection, and to intervention trials of medical and behavioral prevention modalities, include voluntary informed consent, confidentiality, randomization, avoidance of physical/psychological risk, and provision of STD/HIV counseling and preventive services (Christakis, 1988; Barry, 1988). The concept of justice is also relevant, particularly in the case of international research: the burdens of research should be justly distributed, and disadvantaged communities should be assured of reaping an equal share of potential benefits—such as access to effective vaccines that have been tested in part in developing countries (Beauchamp and Childress, 1983; Christakis, 1988; Garner et al., 1994). Each of these issues is complex, and only a few salient points can be summarized here. Voluntary informed consent represents an ethical imperative in behavioral and medical research. In studies conducted in Africa (or other regions, for that matter), informed consent procedures must take cultural practices into consideration. Thus, community-based research may require the consent of the head of

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences the household prior to the enrollment of other household members, in addition to confidential individual consent. It is obvious that consent forms require careful translation into the local language (or prevalent European language or dialect, if appropriate). In addition, the legalistic language required by U.S.-based institutional committees is frequently incomprehensible and inappropriate for African contexts. Some flexibility is required to develop appropriate formulations, while safeguarding the inviolate principal of voluntary and informed participation in research activities. A recent review of HIV/AIDS intervention research concludes that the majority of behavioral intervention studies reported in the literature were methodologically inadequate for assessing intervention effectiveness (Oakley et al., 1995). The authors consider randomized controlled trials most appropriate for evaluating the effectiveness of behavioral interventions, but note that "it is commonly argued by behavioral researchers that random allocation to experimental groups is ethically more dubious than the uncontrolled experimentation resulting from less robust designs or from the implementation of unevaluated programs." The authors conclude that the resulting methodological weaknesses have led to a situation in which there is "a troubling lack of … soundly based preventive interventions," and they call for more randomized controlled trials to provide adequate guidance for investment in HIV/AIDS behavioral interventions (Oakley et al., 1995:484). A similar call for controlled trials has been sounded by Aral and Peterman (1993). Involving subjects in inadequately controlled trials is itself ethically questionable, as it requires commitment (and potentially some inconvenience and risk) on the part of study subjects with no assurance that this commitment will result in useful data or effective programs. Assessment of reasonable research risk must take into consideration existing services, prevention strategies, and medical care in a given setting. It is obvious that regardless of where it is conducted, research must never take advantage of a lack of alternative services to test strategies that are of dubious benefit or are associated with inappropriate risk. However, in places where diagnostic and service delivery alternatives are very limited, testing of STD/HIV prevention, diagnostic, and treatment strategies that would not be applicable in the North American or European context may be appropriate, provided such research offers distinct potential for the development of locally appropriate and sustainable approaches that would otherwise not exist. Examples include innovative STD interventions based on limited diagnostics. In any case, the involvement of host country researchers, care givers, and policy makers in ethical decision making is essential in weighing the local risks and benefits of particular research efforts. A potential barrier to the integration of behavioral and epidemiological research has been the perception that serological testing for HIV must always include mandatory HIV counseling (i.e., that for ethical reasons and in keeping with U.S. domestic federal regulations, subjects cannot be enrolled unless they agree, a priori, to receive their HIV results). Researchers (both behavioral and

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences other) have thus shied away from collecting blood specimens or at best have collected unlinked, anonymous specimens, thus severely constraining the analysis and interpretation of results. In reality, the desirability of voluntary HIV counseling may be dictated by both the value of enrolling truly representative population samples in some intervention trials (and not only those persons who agree to receive their results) and host country regulations and standards. Indeed, large population-based HIV studies in countries such as Uganda and Tanzania have adopted voluntary testing strategies, with the proviso that HIV results be made readily available and that the programs provide information and motivation for subjects to receive their results. In Nairobi, Kenya, women were tested in perinatal HIV transmission studies after giving voluntary informed consent and were given an appointment one week later to collect their results. Among the 243 women who were told that they were infected, three-quarters did not report their HIV-positive status to their partner, 1 committed suicide, 7 were beaten, and 11 were replaced by another wife or expelled from their home. The investigators subsequently adopted a policy respecting women's right not to know their HIV test results (Temmerman et al., 1995). Finally, although it is ethically important to provide an appropriate level of HIV education and prevention services to all study participants, in a randomized trial it is also important to ensure sufficient difference between treatment and control groups to allow interpretation of data. Studies that "overtreat" control-group subjects and thus do not meet this basic criterion are themselves of questionable ethical standing, as they may lead to false conclusions and subsequent ineffective programs or the dismissal of a useful approach to HIV prevention. Services provided to control-group subjects must at a minimum meet local standards of care; it is not necessarily appropriate for them to meet U.S. standards of care. To borrow an example from medical interventions, WHO recently concluded that comparing simplified intrapartum Zidovudine (AZT) regimens with untreated controls can be ethically appropriate in settings where AZT is not otherwise available, even if such a control group would no longer be appropriate in the United States (World Health Organization, forthcoming). RECOMMENDATIONS KEY RECOMMENDATION 1. Basic surveillance systems for monitoring the prevalence and incidence of STDs and HIV must be strengthened and expanded. Good social science research is as dependent as public health and medical research on reliable and valid HIV/AIDS surveillance data. With the implementation of various interventions aimed at controlling HIV transmission, periodic

OCR for page 69
Preventing and Mitigating AIDS in Sub-Saharan Africa: Research and Data Priorities for the Social and Behavioral Sciences monitoring of STD and HIV prevalence and incidence among selected populations is essential both for assessment of the impact of these programs and for decision making on program design and implementation. Recommendation 3-1. More emphasis must be placed on HIV incidence studies for monitoring trends in HIV infection rates. Although seroprevalence provides important information regarding currently infected individuals in an area, measuring incidence is also critically important for estimating the rate of change in the spread of HIV infection in a given population. In particular, data on current incidence provide the most direct and immediate information regarding the potential effects of a given intervention. Together, prevalence and incidence studies can provide information regarding the current status of the epidemic in terms of numbers of infected individuals and the rate of spread within a given population on an annual basis. Recommendation 3-2. STD and HIV prevalence and incidence data should be combined with behavioral and demographic information. Current surveillance systems are often limited, incomplete, and inconsistent, and they rarely measure behavioral or demographic variables. Given new, noninvasive techniques for the collection and analysis of biological specimens (including blood, urine, vaginal secretions, and saliva), accurate assessment of STD and HIV prevalence and incidence can readily be combined with behavioral and demographic information. In conjunction with periodic serosurveys, demographic information is needed to elucidate the differential spread of STD and HIV infection in rural and urban settings and variations in seroprevalence and incidence by gender, educational level, profession, income level, age, and other demographic factors. This type of information is critical for targeting prevention messages to selected groups at risk of acquiring and transmitting HIV and for projecting the effects of HIV and other STDs on a population over time.