2
DESCRIBING THE EPIDEMIC OF HIV INFECTION AND AIDS AMONG WOMEN AND CHILDREN IN THE UNITED STATES
As the epidemic of HIV infection and AIDS in the United States continues, the recent shifts in demographic trends and geographic patterns of HIV disease become more pronounced. The dimensions of HIV infection among women and children in particular now appear in bold relief. Increasingly, policymakers are coming to appreciate the magnitude of the problem of HIV disease in these populations and the disproportionate impact that the epidemic is having on women and children of color. An understanding of the epidemiology of HIV infection and AIDS among women and children and sensitivity regarding the character and needs of the communities most affected are necessary to guide the development of sound perinatal HIV screening policy.
Epidemiology of HIV Infection and AIDS Among Women
Data on AIDS cases and HIV seroprevalence (i.e., prevalence of infection) delineate the epidemiology and scope of HIV disease among women.1 AIDS case surveillance provides useful information on gross trends in the epidemic, but these data do not fully indicate the current burden of infection among women or possible future epidemic trends. AIDS is the final clinical stage of HIV infection—a disease that has a lengthy incubation period (the estimated median is at least 8 to 9 years) from initial infection to the development of overt illness. HIV seroprevalence data, therefore, more accurately reflect the extent of infection among
women because they capture not only those women who have recently been infected and remain asymptomatic but also those who have symptoms that do not meet the AIDS case definition. AIDS and HIV prevalence data are generally presented by demographic category. Yet it is important to recognize that membership in a particular racial or ethnic group does not in itself constitute a risk factor for HIV infection. Rather, behavioral (e.g., sexual and drug-using practices) and social (e.g., poverty, lack of education) factors contribute to the differential distributions of infection.
In 1989, eight years after the first report of a woman with AIDS, the number of cases of AIDS among women exceeded 10,000, accounting for roughly 9 percent of all reported cases in the United States. In addition, women constituted a growing proportion of all adult AIDS cases during this time period. Prior to 1984, women accounted for only about 7 percent of such cases; in 1989 the percentage had increased to slightly more than 10 percent. Through September 1990, 14,452 cases of AIDS had been reported among women, and approximately one-third of these cases were reported in the previous year (CDC, 1990).
Unlike the majority of cases among men, AIDS among women is inextricably linked to intravenous (IV) drug use. Of the 71 percent of female adult cases associated with IV drug use, 51 percent of these women were IV drug users themselves and 20 percent were sexual partners of IV drug users. Overall, 32 percent of women with AIDS reported sexual contact with high-risk male partners (including IV drug users). The impact of IV drug use can also be seen in the geographic, racial, and age distributions of women with AIDS. In fact, IV drug use may be the principal determinant of the demographic patterns that have been observed thus far. Although states along the Atlantic coast (as well as the District of Columbia and Puerto Rico) still have the highest concentration of AIDS cases among women, current data indicate that such cases are now appearing beyond these boundaries. (In fact, all states except North Dakota have reported cases among women.) Women with AIDS are still found predominantly in large metropolitan areas (populations greater than 1 million), largely because of the connection to drug use; however, small to medium-sized cities (populations between 50,000 and 1 million) and rural or nonmetropolitan areas (populations less than 50,000) are reporting growing proportions of cases.
Women of color have been disproportionately affected by the disease—approximately 52 percent of women with AIDS are black, 20 percent are Hispanic, and 27 percent are white. These proportions have not changed substantially since 1984. As a result, black and Hispanic women have cumulative AIDS incidence rates, respectively, that are 13 and 9 times that of white women.
In contrast to AIDS, specific trends in the prevalence of HIV infection among women are difficult to establish because of limited data. Several cross-sectional, blinded (i.e., anonymous) seroprevalence studies, however, have described general distribution patterns. Surveys of women attending family planning, prenatal, and abortion clinics indicate a median prevalence rate of 0.2 percent (or about 2 infected women per 1,000 women), with a range of 0 to 2.3 percent.2 Seroprevalence among women seeking clinic services varied markedly by age and race or ethnic group, as well as by geographic location (Sweeney et al., 1990).
Recent surveys of women attending sexually transmitted disease (STD) clinics have found substantially higher seroprevalence rates than were found among family planning, prenatal, and abortion clinics. 3 The median prevalence rate among STD clinics was 0.8 percent (or about 8 infected women per 1,000), with a range of 0 to 13 percent; nearly one-fifth of the clinics had rates of 5 percent or greater. The elevated prevalence of infection among STD clinics is not surprising because many of the individuals attending such clinics are likely to have multiple sexual partners and therefore to be at increased risk for infection (Shapiro et al., 1989).
The highest HIV prevalence rates have been found among women who are IV drug users. Surveys of women who enter drug treatment centers reveal a median prevalence rate of 4.7 percent (i.e., nearly 50 infected women per 1,000), with a range of 0 to 47 percent.4 Most striking, however, was the wide geographic variation in prevalence observed among these centers. For example, among treatment centers in the northeastern United States, the median prevalence rate was 24 percent; in other parts of the country, the median rate was only 3 percent. Since 1988 CDC has collaborated with state health departments and NICHD to conduct a population-based seroprevalence survey of childbearing women. The survey involves collecting filter-paper blood samples from newborn infants for routine metabolic screening and anonymously testing them for HIV antibodies (Pappaioanou et al., 1990). (Because all infants are born with passively acquired maternal antibodies, anonymous newborn screening is a surrogate measure of maternal infection.) These surveys are unique among large-scale seroprevalence surveys because they provide relatively
unbiased estimates of HIV prevalence in a well-defined population of women.
These studies have highlighted the concentration of HIV infection among childbearing women along the Atlantic coast. Yet HIV infection among childbearing women has been detected outside the major metropolitan areas and in all regions of the country. In addition, considerable variation in seroprevalence has been observed both among states and within individual states. For instance, the seroprevalence rate among childbearing women in San Francisco during 1989 was 1.3 infected women per 1,000 childbearing women, nearly twice that of the state of California. In Chicago, the seroprevalence rate was 1.6 per 1,000, considerably higher than the state of Illinois' rate of 0.3 per 1,000. In several metropolitan areas along the East Coast, including New York City and Newark, nearly 1 percent of all women delivering infants were HIV positive. This geographic variation in seroprevalence reflects in part the differential distribution of behaviors that place persons at risk for infection—for example, the concentration of IV drug use and related risk behaviors (e.g., sexual contact with IV drug users) in urban centers in the Northeast (Shapiro et al., 1989).
Age-specific data for HIV seroprevalence among women are limited Several studies in the New York City area and studies of military applicants and blood donors have shown that prevalence rates are highest for young adult and middle-aged women (Shapiro et al., 1989). HIV infection and AIDS among women appear to affect predominantly those of reproductive age. In fact, about 85 percent of women with AIDS have been between the ages of 15 and 44 at the time of diagnosis. Consequently, HIV infection and AIDS (grouped collectively for this purpose as HIV disease) have become an important cause of mortality for this group. Although the death rate for most leading causes of mortality among women aged 15 to 44 remained stable over the past decade, the death rate from HIV disease continued to increase and in fact quadrupled between 1985 and 1988. In 1987, HIV disease became the eighth leading cause of death among women of reproductive age. If current trends continue, it is expected to become one of the five leading causes of mortality by 1991 in women of reproductive age (Chu et al., 1990).
The number of AIDS cases among women is likely to continue to increase rapidly, at least for the next few years. Extrapolating from AIDS case surveillance data, it is estimated that 5,000 AIDS cases were diagnosed among women in 1989. In 1991, health officials estimate that 7,000 to 9,000 cases will be diagnosed among women; in 1993, they anticipate between 9,000 and 15,000 cases.
Epidemiology of HIV Infection and AIDS Among Children
As of 1990, CDC estimated that between 5,000 and 10,000 children in the United States were infected with HIV. Through September 1990, 2,628 cases of AIDS had been reported among children under 13 (roughly 2 percent of all reported cases); about 30 percent of them were reported in the previous year. Of the remaining infected children in the CDC estimate (approximately 2,500 to 7,500), some have continued to be asymptomatic or have died from other causes. Still others have developed HIV-related illness that does not meet the AIDS case definition. Some may have developed AIDS but not yet been reported.
The epidemiology of AIDS among children tends to mirror the larger epidemic in women because most children with AIDS have been infected perinatally. (The risk of exposure to HIV through infected blood or blood products has been largely eliminated.) As of September 1990, more than 80 percent of children with AIDS were reported to have acquired HIV infection perinatally.
Perinatal AIDS cases are still heavily concentrated along the East Coast and in Puerto Rico. To date, however, all but six states have reported at least one case. In 1982, New York, New Jersey, and Florida were the only three states reporting AIDS cases attributable to perinatal transmission. Yet in 1989, these three states accounted for just over half of all perinatal AIDS cases, indicating that the epidemic is clearly spreading beyond these areas.
The AIDS epidemic has affected predominantly children of racial and ethnic minorities. Black children accounted for 58 percent of all perinatally acquired AIDS cases reported in 1989; children of Hispanic origin accounted for 26 percent. Compared with white children, the rate of reported AIDS cases was 16 times higher for black children and 8 times higher for Hispanic children. In addition, children with AIDS are often members of families in which IV drug use is prevalent. Approximately 60 percent of pediatric AIDS cases reported through September 1990 were associated with IV drug use—42 percent of the cases with maternal drug use and 17 percent with maternal sexual contact with an IV drug user.
Although long-term trends in perinatally acquired HIV infection are difficult to predict given current data, some estimates can be made over the short term of the number of infants newly infected with HIV and the number expected to develop AIDS. These estimates are derived from information on HIV seroprevalence among childbearing women, birth
statistics, and the estimated rate of perinatal transmission.5 Based on seroprevalence data from 25 states,6 CDC estimates that there were about 1.5 HIV-infected women per 1,000 women delivering live infants in the United States in 1989, which translates to approximately 5,900 births to seropositive women during a 12-month period. Assuming that 25 to 35 percent of these infants were actually infected, an estimated 1,500 to 2,100 new cases of perinatally acquired infection occurred in one year, with the total number probably closer to 1,750 (Gwinn et al., 1990). This figure exceeds the total number of all cases of perinatally acquired AIDS reported from 1981 through 1989. Even if the incidence of perinatal infection remains stable over the next several years, each birth cohort of HIV-infected children will continue to contribute to the number of newly diagnosed AIDS cases for the next several years. Thus, one could expect increasing numbers of children to develop AIDS each year, at least over the short term.