Archival Content: 1999-2005
HIV Prevention Among Drug Users:
Determining the Extent of HIV Infection Among Drug Users
Thus far, PART 2 has given an overview of drug-using practices, the high-risk sexual behaviors of drug users, and the social context of drug use. It also has discussed several specific population groups who are heavily affected by the HIV epidemic. This provides important background for HIV prevention planners and program managers, who will be prioritizing, designing, and implementing HIV prevention programs in their communities.
With this background in mind, one of the first things that prevention planners and program managers will need to do when developing an intervention program is to develop a profile of the extent of HIV infection among the drug users in their community. Some of the data needed to develop such a profile can be obtained from HIV/AIDS surveillance and reporting systems. These data, which are available to local planners and program managers through their health departments, provide information on the number and characteristics of persons diagnosed with AIDS or recently infected with HIV, and about the mode of exposure to HIV.
This information will help to increase the understanding of the extent of drug use among persons infected with HIV as well as the patterns of drug use and HIV risk in certain parts of the U.S. State and local health departments have epidemiologists on staff who can help planning groups and program managers understand these data and their strengths and limitations. Table 2.4 lists some major data sources, a number of which are discussed in "Sources of Data.".
In addition, there are other sources of local data that may be useful to planners. Some of these sources have been designed to supplement national HIV/AIDS surveillance data, and provide information on the scope of the epidemic attributable to injection drug use. For example, several local HIV/AIDS surveillance programs and a number of special studies have begun to report data on the association between HIV exposure and the use of non-injection drugs, such as crack (Chirgwin et al., 1991; Chiasson et al., 1991; Diaz et al., 1993; Diaz et al., 1994; Edlin et al., 1994; Ellerbrock et al., 1995; Sugarman et al., 1995).
Following are brief descriptions of specific sources of national, state, and local data that may be useful to planning groups and program managers as they come to understand the HIV and drug use profile of their communities and as they prepare to select and develop interventions. Definitions of key terminology can be found in Exhibit E.
Almost all persons with AIDS who receive medical attention are reported to the national AIDS surveillance system, making these data representative of persons with AIDS. Of all AIDS cases reported to CDC in 1995, just over 35 percent were associated with injection-drug use (CDC, 1996). Among those with IDU-associated AIDS, just over half were heterosexual males, 20 percent were female, and 13 percent were men who have sex with men. The rest were male and female heterosexual partners of IDUs (11 percent), and children whose mothers were either IDUs or sex partners of IDUs (1 percent). Because of the completeness and representativeness of AIDS surveillance data, emerging trends in characteristics of HIV-infected persons can be detected by analysis of AIDS case surveillance data. However, because AIDS surveillance represents those persons with advanced HIV disease, early detection of trends among subgroups may not be possible. Also, surveillance of AIDS cases may fail to detect cases in some subgroups, such as lesbians, because questions about sexual orientation among females may not be asked or answered.
As of November 1996, confidential HIV infection reporting for adults and adolescents by name is required by 26 states and for children by 29 states. HIV reporting is an adjunct to AIDS surveillance. HIV-reporting data provide a minimum estimate of the number of persons known to be infected with HIV in states with confidential HIV infection reporting. These data may be used to anticipate trends among particular groups, such as adolescents.
These data are representative only of HIV-infected persons who are confidentially tested for HIV in the states where HIV reporting is required. HIV-infection reporting data are not representative of HIV-infected persons who have not been tested, who have been tested anonymously, or who live in states or territories where HIV reporting is not required.
Clinic-based seroprevalence surveys
Clinic-based data compiled by CDC provide important information about HIV seroprevalence among injection drug users. These data also provide a valuable window on the drug-use practices in local areas and how these practices differ from area to area. For example, seroprevalence surveys conducted in local drug treatment centers (DTCs) over the past several years provide information on persons who enter participating DTCs and who report injecting illicit drugs during the previous year (CDC, 1994a). Seroprevalence among IDUs entering DTCs was the second highest of any group surveyed in CDC's National Serosurveillance Program. In 1991-1992, the HIV seroprevalence among IDUs ranged by DTC from less than 1 percent to more than 50 percent.
From 1991 through 1992, 39 STD clinics in which seroprevalence surveys were conducted reported at least 50 injection drug users (the minimum number required for calculating seroprevalence). Median seroprevalence among injection drug users attending these STD clinics was around 6 percent. A number of findings from the surveys of these DTCs (CDC, 1994b) show how data from a national reporting system can be used to plan, design, and implement interventions that are appropriate to local communities:
Geographic differences. HIV seroprevalence among IDUs differs dramatically according to location. DTCs in metropolitan areas along the Atlantic Coast and Puerto Rico showed the highest median HIV seroprevalence, ranging from just under 5 percent to around 40 percent. The median HIV seroprevalence in DTCs in the Midwest ranged from 3 percent to 17 percent, and from less than 1 percent to around 7 percent in the West. Figure 2.3 displays the HIV prevalence reported in 34 metropolitan areas among IDUs entering DTCs (CDC, 1994a).
Although the reasons for these geographic differences are unknown, several plausible hypotheses have been advanced. First, injection drug use (and the potential for HIV transmission) will occur most frequently near channels of drug distribution. Second, like gonorrhea and syphilis, the drug-related transmission of HIV may be a "core group" phenomena whereby the occurrence of "a disease is clustered geographically and sociodemographically in distinct population subgroups. The extent to which those in the core group have sexual [or injecting] partners outside the group determines STD [or HIV] distribution" (Marx 1991, p.93).4
Finally, it has been suggested that the mode of transmission influences geographic distribution of HIV prevalence. The fact that some MSM have traveled widely and live all around the country accounts for the geographic similarity of HIV prevalence in the homosexual population. On the other hand, the fact that IDUs are less mobile and tend to stay in certain parts of the country might account for the geographic differences among HIV prevalence in this population (Huet al., 1994).
Racial and ethnic differences. In all geographic regions, the median HIV seroprevalence by DTC was substantially higher among African Americans than among whites. HIV seroprevalence was generally higher among Hispanics than among whites, a difference largely due to the higher seroprevalence among Hispanics in the Northeast.
Gender differences. Evidence from the DTCs showed that seroprevalence rates among men and women were generally similar. In a different seroprevalence study conducted from 1988 to 1993, however, seroprevalence was similar among men and women except in the South, where the seroprevalence among women was almost twice that among men (Prevots et al., 1996).
HIV seroprevalence among IDUs entering DTCs increased before 1989 and has stabilized, although marked geographic variations remain (CDC, 1994a). Prevots et al. support these findings, and suggest that, based on "estimates of historical infection rates...the peak years of HIV incidence among injection drug users in the United States were 1983-1986" (Prevots et al., 1996, p.739).
Because IDUs entering DTCs may not be representative of all IDUs, these data should be interpreted with caution, and other sources of information used, such as studies conducted among street-recruited IDUs (Deren et al., 1995; Friedman et al., 1995). Nonetheless, DTC seroprevalence data interpreted along with data on HIV seroprevalence among IDUs attending STD clinics may provide a more complete picture of the potential for HIV acquisition or transmission in a given community.
Supplement to HIV/AIDS Surveillance
CDC collaborates with 11 state and local health departments in a project called the Supplement to HIV/AIDS Surveillance (SHAS). The data generated from this project are important because they represent localized drug use that is specifically related to HIV seropositivity in that community. These data also are important because they include information on non-injection drug use. Other state and local health departments may want to explore how they can develop locally relevant data sources in collaboration with the CDC/SHAS process.5
In the SHAS project, interview data are collected from a cross section of persons aged 18 years or older with AIDS (or HIV infection in HIV-reporting states) who are medically able to complete interviews. Among 1,142 persons reported with HIV infection or AIDS who also reported injection drug use, 71 percent reported injecting more than one drug. Overall, 35 percent of IDUs interviewed reported cocaine as their primary drug injected, followed by heroin (28 percent), speedball (17 percent), and amphetamines (16 percent). The primary drug injected varied notably by state or city of residence (see Table 2.5). For example, heroin was reported as the primary drug injected by nearly all IDUs interviewed in Detroit and by almost half of IDUs interviewed in Connecticut, but by one-quarter or fewer of IDUs interviewed in all other areas. (Diaz et al., 1994).SHAS data suggest that, overall, cocaine has been the most popular drug for injection among IDUs with HIV/AIDS, and that cocaine also has been positively associated with the practice of needle sharing (Diaz et al., 1994; Mandell et al., 1994).
SHAS data also suggest that some IDUs may serve as a bridge between the injecting community, in which the prevalence of HIV infection is high, and the crack-using community, in which the prevalence of high-risk sexual behavior is high (Diaz et al., 1994). In Houston and Dayton/Columbus, Ohio, 75 percent and 70 percent of IDUs, respectively, smoked crack in addition to injecting drugs (Williams et al., 1995).
Other sources of data
Additional data about the drug-using population may be available from local drug treatment and mental health professionals and STD clinics, as well as from local surveys of drug use and risk behavior funded by a range of federal agencies, including CDC, the National Institute on Drug Abuse (NIDA), and the National Institute on Mental Health (NIMH).6 Since the drug-using population relies on the mental health and drug treatment system for services, representatives from local agencies providing these services can contribute much to local prevention planning and program development. Drug treatment and STD clinics offer opportunities to collect data as well as to provide HIV and STD prevention services to the drug user. With proper assurances of protection of confidentiality, health professionals can work together to identify populations vulnerable to infection, facilitate contact with, and provide cross referral for those populations.7
It should be noted, however, that the availability of agency data for prevention planners may vary considerably, based on reporting requirements and staff capacity to summarize these data in an aggregate way. Although data may be available, for example, on the percentage of drug users in an agency's funded programs, there may not be data on the types of drugs used by this population. Alternatively, data may be available from a mental health service agency on the percentage of clients/patients who are dual-diagnosed (e.g., mental illness and chemical dependency), but not be available on how many of these clients/patients are homeless. When certain data are not available for planning purposes, it is best to conduct structured interviews with a representative sample of relevant professional providers to retrieve the information.
Surrogate markers of HIV risk behavior can be another valuable indicator of the potential for HIV transmission within a community. One example of a surrogate marker is teenage pregnancy rates, which indicate unprotected sexual intercourse and, therefore, HIV risk behavior. Another marker is rates of sexually transmitted diseases (STDs) acquired through unprotected sexual intercourse. Because sexually transmitted disease rates are a reliable indicator of high-risk behavior, groups with high rates of STDs are at increased risk for the spread of HIV infection once it has been introduced into the group.
Rates of infection with hepatitis B virus (HBV) and hepatitis C virus (HCV) may sometimes be used as surrogate markers because HIV can be transmitted by the same drug injection practices that transmit HBV and HCV. Like HIV, these infections also can be transmitted sexually and perinatally. Because the incidence of HBV, and probably HCV, is commonly higher than that of HIV infection (Bortolotti et al., 1982; Chamot et al., 1992; CDC, 1991), studies of HBV and HCV infections have been suggested as a method of evaluating some HIV prevention programs (Committee on Social and Behavioral Sciences, 1992).
Since syphilis, gonorrhea, chlamydia, hepatitis B, and hepatitis C are reportable diseases, locally reported data on these diseases may be used by HIV prevention planners and program managers as an indirect way to determine the prevalence of high-risk behaviors in the local population and the potential for HIV spread, once it is introduced into the population. However, incidence of these infections is seriously under-reported, especially hepatitis B and C, because the majority of persons infected with HBV and HCV do not have symptoms.
Prevention planners and program managers should be aware, however, that hepatitis B cannot always be used as a reliable surrogate marker for HIV transmission. HIV transmission not only depends on individuals engaging in risk behaviors, but also on the prevalence of HIV in the community. Thus, although the prevalence of HBV among drug users is generally high, HBV would not be considered a reliable surrogate for HIV prevalence in a community where the prevalence of HIV is low, even if HBV prevalence among drug users in that community were high. Current research on the relationship between HBV and HIV seroconversion has demonstrated that trends of incident HBV infection do not parallel trends of incident HIV infection in a population of IDUs followed over time. At an individual level of analysis, however, incident HBV infection is a predictor of incident HIV infection among male but not female IDUs (Levine et al., 1996).
This CDC Web site is no longer being reviewed or updated and thus is no longer kept current. This site remains to assist researchers or others needing historical content.