AIDS

Acquired Immune Deficiency Syndrome (AIDS) was first noticed and described in the United States in 1981, initially in homosexual men. Further cases in hemophiliac patients were reported in 1982. The publication of accounts of this new disease in the newsletter Morbidity and Mortality Weekly Reports triggered responses from physicians in other developed countries who had recently come across similar constellations of symptoms, which indicated a breakdown of the immune system in individuals who had no known exposure to radiation or immunosuppressant drugs.

Origins of HIV

The Human Immunodeficiency Virus (HIV) that is the cause of AIDS was first isolated by Luc Montagnier at the Pasteur Institute in Paris in 1983. HIV is a retrovirus, which means that it stores its genetic information as RNA, stimulating the production of DNA copies of its genome when it enters a host cell. A systematic testing of stored serological samples carried out in the late 1990s in a search for the origins of this disease revealed the earliest documented occurrence of HIV dating from 1959. This was in a blood sample taken from a male subject "L70," one of a number of hospital patients from Western and Central Congo seen in Leopoldville in that year.

HIV is believed to have arisen as a result of ancestral viruses crossing the species barrier, from chimpanzees and monkeys to man. Such crossovers are believed to have occurred at least twice, with the more virulent strain, HIV-1, originating from a simian immunodeficiency virus (SIV) of chimpanzees, and HIV-2 coming from an SIV usually found in sooty mangabey monkeys. Modeling studies based on the genetic diversity of the HIV viruses, reported by B. Korber and colleagues, estimate that the species crossover occurred between 1920 and 1940. It is believed that the virus was present in isolated human populations in rural Central Africa from this time, and began to spread more widely in the 1960s and 1970s, as a result of wars, tourism, and social changes linked to modernization, which all contributed to increased population mobility.

In the 1980s and 1990s HIV/AIDS was identified in every region of the world. UNAIDS, the Joint United Nations Programme on HIV/AIDS, has estimated that 42 million adults and children worldwide were infected with HIV by the end of 2002. Of this number 29 million (70%) were living in sub-Saharan Africa. Table 1 shows the estimated numbers of infected persons in each of the world's major regions at the end of 2002.

Etiology and Disease Progression

HIV spreads by direct contact through body fluids. This may occur during sexual intercourse, or as a result of mother-to-child transmission during pregnancy, delivery, or breastfeeding. The virus may also be transferred in blood used for transfusions, or in blood products, such as the clotting factor supplied to hemophiliacs. Finally, it can be spread by unsterilized hypodermic needles and surgical instruments, so outbreaks among injectable drug users (IDU) are common where injecting equipment is shared. In the 1990s it was speculated that the species transfer may have occurred as a result of live polio vaccine being cultivated in infected primate livers, but this theory has been discounted, as no traces of the virus were found in stored vaccine samples. However, mass vaccination campaigns could have helped to spread the virus in the 1960s, if needles were reused during the campaigns, or left behind and subsequently reused in poorly equipped hospitals. The virus is very fragile: it cannot survive outside of the human host cell and cannot be spread by insect bites, by casual touching, or by sharing of food utensils or clothes.

Both HIV-1 and HIV-2 target specific cells of the immune system: the T-cells, which are the human body's main immunological defense against infection. Primary infection with HIV in an adult usually results in rapid multiplication of the virus in the lymph system, and a rapid decline in T-cells. An immune response is usually provoked within four to six weeks, and after this time antibodies to HIV can be detected in the blood. The disease then enters a latent phase, which has a median duration of around

TABLE 1

nine years in the absence of treatment in developing and developed countries alike. During this latent phase the number of T-cells declines steadily. Eventually, the immune system is so compromised that it is no longer able to respond adequately to a range of infections and cancers, such as Kaposi's sarcoma, that usually pose little threat to healthy individuals. Once such clinical manifestations of AIDS occur, death follows rapidly: in the studies reviewed by J. T. Boerma and colleagues, survival times of five to nine months were reported in developing countries, 9 to 26 months in developed countries.

In infants, the disease usually progresses much more rapidly, because the immature immune system cannot respond adequately in the primary infection phase. Median survival time of infected infants in the absence of treatment is around two years.

HIV mutates rapidly both within an infected individual and across individuals as the infection spreads in a population. The resulting diversity of forms makes it difficult to develop an effective vaccine, and ensures the rapid emergence of drug resistant forms.

Epidemiological Evidence

On a national basis, AIDS epidemics are characterized as generalized if HIV prevalence exceeds 1 percent in the adult population aged 15–49. In developing countries, almost all the evidence for HIV prevalence levels in the general population comes from the anonymous screening of blood samples obtained from pregnant women in antenatal clinics. As a consequence there is very little direct evidence on HIV prevalence among men. Subject to the resulting uncertainties in assessing prevalence levels, UNAIDS estimates that only 5 of the 45 countries of sub-Saharan Africa did not have generalized epidemics by 2000. Adult prevalence had already reached over 20 percent in Southern Africa, and lay between 10 percent and 20 percent in most of Eastern and Central Africa. Western African countries, especially those where HIV-2 is the predominant strain, generally had lower rates, between 2 percent and 5 percent. In the first years of the twenty-first century, HIV prevalence was either stagnant or increasing all over Africa, except for Uganda, where the epidemic appeared to have passed a turning point, with a decline in prevalence from 14 percent in the early 1990s to around 8 percent at the end of the decade.

The other region with a generalized epidemic is the Caribbean, where adult prevalence was over 1 percent by 2001 in nine countries. Southeast Asia as a whole had a prevalence of 0.6 percent in 2001, but in three countries, Cambodia, Myanmar, and Thailand, adult prevalence was over 1 percent, though in Thailand a strong downward trend was evident by this time in the general population. The future course of the epidemic in Asia will be largely determined by trends in the most populous countries, India and China, in both of which HIV infection is concentrated in the high-risk groups (sex workers and IDU) and in particular regions. (Six states in India recorded prevalence levels over 1 percent in 2001; Yunan province in China recorded an increase from 6 percent to 10 percent between 2000 and 2001 in prevalence among sex workers.)

The sub-populations most strongly affected by the epidemic vary widely by region, as shown in Table 1. In high-income industrialized countries by 2002, most of the persons living with HIV were men who had sex with men (MSM) or IDU. Eastern Europe and Central Asia have epidemics that are concentrated among IDU–but these are among the fastest growing epidemics in populations that are difficult to monitor.

There is evidence, summarized by M. Caraël and K. Holmes, that male circumcision is an important risk reduction factor. Consistent condom use has been shown to be effective in reducing sexual transmission in MSM epidemics in industrialized countries. Paradoxically, it does not emerge as a protective factor in community studies in Africa since condoms are mainly used by those who are already infected or believe themselves to be at very high risk. Co-infection with other sexually transmitted diseases has been found to significantly enhance the transmission of HIV.

Population mobility is likely to have been an important factor in the initial spread of HIV, and is implicated in the rapid spread of HIV in Southern African countries in the late 1990s. These countries had a tradition of male labor migration to mines and commercial farming estates that encouraged the formation of temporary partnerships and the growth of commercial sex.

Demographic Impacts of AIDS in Africa

In the epidemics driven by MSM or IDU, the number of males infected exceeds that of females by a factor of two to ten. However, in heterosexually spread epidemics, the number of infected females generally exceeds that of infected males. Several factors account for this: males tend to have more sexual partners than females; HIV transmission from male to female is more efficient than from female to male; and most of all, there are more females than males at risk of infection because of the age difference between sexual partners and the steeply tapering youthful age distribution.

In Africa, where generalized epidemics began in the 1980s, strong evidence of an impact on national mortality trends was detected in the 1990s. Almost all African countries lack national vital registration systems, so cause of death data are not available, and most of the evidence for trends in age-specific mortality rates comes from census and survey data, either from direct questions about household deaths in the year preceding the survey, or indirect enquiries about the survival of relatives. Griffith Feeney has used mortality data from a variety of secondary sources to show that adult mortality in Zimbabwe more than doubled between 1982 and 1997, spanning the time when HIV prevalence rose from virtually zero to almost 30 percent.

In the context of the HIV epidemic, the most widely used indirect estimation techniques, such as the child survival and orphanhood methods, yield mortality estimates that are biased downward, because of the high correlation between the survival of parents and children. Reports of household deaths may also be incomplete as households tend to dissolve upon the death of the head. However, using new analytical techniques based on reported survival of siblings, Ian Timaeus has demonstrated significant rises in adult mortality in the most severely affected countries. For example, five Eastern and Southern African countries that included sibling survival questions in Demographic and Health Survey rounds in the late 1990s recorded large increases in the probability of dying between age 15 and age 60: from an average of 28 percent, five years before the survey, to 45 percent in the survey year. Estimated HIV prevalence at the time of the surveys in these countries averaged 24 percent. By contrast, in five West African countries, with an average HIV prevalence of 2 percent, similar analyses of DHS sibling survival data collected in the 1990s indicated a continuing modest improvement in this index of adult mortality, from 26 percent to 22 percent over a similar 5-year period.

The most compelling evidence about the scale of the impact on mortality comes from longitudinal community based studies, in which repeated serological testing is accompanied by demographic surveillance. Studies of this type were established in East African countries in the early 1990s and subsequently in Southern Africa. They have shown that mortality rates among HIV infected adults are 10 to 15 times the rates observed among uninfected individuals. The relative risk of mortality among HIV-positive individuals has been shown to increase with epidemic maturity, as the balance between individuals who were recently infected and those who have been living with the disease for some time shifts in favor of the latter, who are closer to developing full-blown AIDS. Duration of survival post-infection in adults is strongly related to age at infection, with much quicker progression to AIDS among those infected at a later age.

Community-based studies rarely provide direct comparisons of mortality in HIV-infected and uninfected children, because of the problems of ascertaining the HIV status of a newborn infant. Maternal HIV antibodies cross the placental barrier much more readily than the virus itself, and the commonly available, relatively inexpensive serological tests detect the presence of antibodies rather than virus. However, such studies have also shown that the mortality among children of HIV-infected mothers is between 2.5 and 4 times that observed among children of uninfected mothers. Hospital-based studies (reported by L. Kuhn and Z. A. Stein) have shown that in the absence of antiretroviral therapy, mother-to-child transmission of the virus occurs in about 30 percent of cases, suggesting that mortality rates for HIV infected children could be 5 to 10 times the rates observed among uninfected children.

Although there is strong evidence on the scale of excess mortality among HIV infected persons, considerable uncertainty remains concerning the number and distribution of deaths due to HIV. Official estimates of historical and current deaths in the worst affected countries are based on models fitted to incomplete time series of prevalence data from a limited number of sentinel sites that perform anonymous tests on pregnant women. The predominantly urban character of sentinel clinics means that they tend to be based in communities with relatively high HIV prevalence.

HIV also has an impact on fertility, with HIV-positive women experiencing significantly lower fertility rates than uninfected women. Biological explanations for this effect include increased fetal losses and stillbirths, increased menstrual irregularities and decreased spermatogenesis in male partners. Social mechanisms are probably even more important, as women who are suspected of being HIV-positive will find it harder to remarry following widowhood and divorce, both of which occur more frequently among the HIV positive. Community-based studies have shown that the overall reduction in fertility in HIV-positive women can range from 10 percent to 40 percent, with older HIV positive women experiencing proportionately larger reductions. Since HIV lowers fertility, estimates of HIV prevalence based on anonymous surveillance of pregnant women attending prenatal clinics tend to under-estimate HIV prevalence among women in the community.

Evidence of an impact of HIV on the population age structure is harder to find, because of countervailing demographic tendencies. The increased mortality due to HIV slows population growth, which would tend to make the population structure older. However, since adults are disproportionately affected compared to children, their early death tends to make the age structure younger. At a subnational level, internal migration, which is very high among young adults, often masks or exacerbates the impact on the age structure.

In the worst-affected countries of Southern Africa, where HIV prevalence was estimated to have passed 20 percent by the year 2000 (Botswana, Lesotho, Malawi, Namibia, South Africa, Swaziland, Zambia, and Zimbabwe), life expectancy is projected to fall below 40 years. These countries may well experience periods of negative population growth in the first few decades of the twenty-first century. This is particularly likely to occur in populations that have already experienced significant fertility declines by the time the HIV epidemic took hold. Estimates and projections of the effect on population size in the worst-affected countries, made by the United Nations Population Division, are shown in Figure 1. South Africa is an example of a country in which AIDS is likely to cause negative population growth–as seen in the projections in Figure 2.

Socioeconomic Consequences

It has been estimated (by Alan Whiteside and Clem Sunter) that AIDS has caused annual per capita economic growth to fall by 0.5 to 1.2 percentage points in about half the countries of sub-Saharan Africa. Absenteeism and illness in the work force mean that employers face increasing costs in recruitment, training, insurance, and sick pay.

A 2001 report from the UN Food and Agriculture Organization estimates that 7 million farm workers have died from AIDS-related causes since 1985, and 16 million more are expected to die by 2020, with a consequent decrease in agricultural production. Children of families in which the adults are too sick to work are often taken out of school to work on family farms and to care for their relatives, and orphans tend to have low school attendance rates because the families who care for them cannot afford an extra set of school fees. Teachers are disproportionately affected by AIDS–in 1999, according to UNAIDS, over 15,000 teachers died in sub-Saharan Africa. Because of the age profile of AIDS deaths, it is the most productive part of the labor force that will be lost, and families will lose parents who have not yet finished bringing up their children.

The health sector has come under severe stress in many African countries. Public expenditure surveys in 1997 showed that in 7 out of 16 countries health spending on AIDS exceeded 2 percent of GDP, and accounted for between 25 percent and 70 percent of the total public health expenditure. This generally represents spending on treatment of opportunistic infections, since most African countries cannot afford antiretroviral therapy, except possibly in the one-off doses that are used at delivery to prevent mother-to-child transmission.

Policy Responses

UNAIDS has estimated that US$10 billion per year will be needed in low-and middle-income countries to combat AIDS in the first decade of the twenty-first century.

Official silence and outright denial have paralyzed efforts to control the epidemic in many countries. Uganda, one of the few countries to admit the problem and encourage public discussion early on, is the only African country in which there is clear evidence of a decline in prevalence. Policy responses that have been advocated include education to raise awareness and promote behavioral change; encouragement

FIGURE 1

of voluntary counseling and testing; the promotion and social marketing of condoms and the treatment of other STDs to lower transmission probabilities. Thailand is also widely recognized as a success story in the fight against AIDS; there, the key strategy was the 100 percent condom use campaign in brothels. Thailand has also recorded a marked decrease in the proportion of men having sex with commercial sex workers.

Treatment with antiretrovirals is in great demand by people suffering with HIV, but is very expensive and unlikely to do much to stem the spread of the epidemic in the poorest countries of sub-Saharan Africa. Vaccines and microbicides offer more hope as preventive strategies, but major research efforts are still needed in the development of these medical interventions. Antiretroviral drugs, such as nevirapine, have had proven success when administered in single doses to pregnant women and their newborn infants, but since such treatment does not affect disease progression in mothers, they cannot slow the overall development of the epidemic, or the growth in the number of AIDS orphans.

The demographic consequences of the AIDS epidemic are relatively easy to project if age and sex specific infection rates can be accurately forecast. Unfortunately, there is little agreement among epidemiologists on robust and widely applicable methods

FIGURE 2

for projecting long-term trends in infection, and the data requirements for the more sophisticated projection tools make them unsuitable for most countries in sub-Saharan Africa. Relatively simple models based on fitting parametric curves to observed time series of HIV prevalence in antenatal clinics have been used successfully by UNAIDS for short-term projections. Because these simple models are not structured by sex and age, they are of limited use in making long-term demographic projections. In the absence of effective interventions, most analysts agree that a continuation of the trends observed in the late 1990s could lead to adult prevalence levels of over 25 percent in many African countries in the first decade of the twenty-first century.