Objective: To assess recent developments in the HIV epidemic in injecting drug users (IDUs) in New York City. With >50,000 cases of AIDS in IDUs, New York has experienced the largest HIV/AIDS epidemic in IDUs of any city in the world.
Methods: Serial cross-sectional surveys conducted continuously from 1990 to 2001 of IDUs entering the Beth Israel Medical Center (BIMC) drug detoxification program in New York City. HIV serostatus, use of prevention services, and risk behaviors were measured. Individuals were permitted to participate multiple times in the surveys but not more than once in any year.
Results: Two thousand eight hundred eighty-seven individuals contributed 3100 observations from 1990 to 2001. There was a substantial and consistent decline in the prevalence of HIV infection among IDUs entering the BIMC detoxification program, from 54% (165/304) in 1991 to 13% (39/303) in 2001 (P < 0.0001). The decline was highly linear, with r2 = 0.92 and a slope of −3.7% in seroprevalence per year. The decline occurred for both males and females, both short and long-term IDUs, and the three largest racial/ethnic subgroups (all P < 0.001 by Cochran–Armitage testing). Use of HIV prevention services increased substantially, particularly syringe exchange and voluntary HIV counseling and testing. General reductions in injection risk behaviors occurred, but substantial numbers of IDUs continued to engage in both receptive and distributive syringe sharing. Two conditional types of risk reduction not currently recommended by health authorities were reported: “informed altruism,” in which persons who knew that they were HIV seropositive reduced transmission behavior, and “partner restriction,” in which persons who shared needles and syringes primarily confined this sharing within small social networks.
Conclusions: HIV infection continues to decline in this population of IDUs in New York City, suggesting the possibility of bringing very high prevalence epidemics under control. Risk elimination may not be required; rather, multiple forms of risk reduction may be effective in reducing HIV transmission within a local population of IDUs.
HIV infection has been described in injecting drug users (IDUs) in >100 countries throughout the world. 1 In parts of eastern Europe, Russia, central Asia, and China, HIV is spreading quite rapidly among IDUs, with an incidence of ≥20% per year and a prevalence ≥70%. 2 IDUs form the leading edge of the HIV epidemic in many of these areas, with the possibility for further HIV spread from IDUs to the larger non–drug-injecting population.
Effective interventions to reduce HIV transmission among IDUs include increased access to sterile injection equipment (through syringe exchange and/or pharmacy sales), community-based outreach, and long-term drug abuse treatment. 3 When these interventions are applied early and on a large scale, it is possible to contain HIV transmission at very low levels (<1% per year) and avert HIV epidemics in local populations of IDUs. 4–6
If the prevalence of HIV infection reaches a high level (≥30%) among a large population of IDUs, however, reducing continuing transmission becomes much more difficult. With a large number of HIV-infected IDUs who can transmit the virus, even modest levels of continuing risk behavior can perpetuate unacceptably high incidence. For example, the incidence of HIV infection among IDUs is ∼3% per year in Amsterdam, 7 6% per year in Bangkok, 8 and 3–5% per year in Vancouver, British Columbia, Canada. 9 Reducing HIV transmission after high seroprevalence levels have been reached among large IDU populations remains a major challenge in HIV epidemiology and prevention research.
We report trends in the prevalence of HIV infection, use of prevention services, and risk behavior among IDUs in New York City for the last decade (1990–2001). New York City experienced the first, and to date the largest, HIV epidemic in IDUs of any city in the world. 10 New York City also served as the epicenter for the spread of HIV along the East Coast 11,12 and throughout the United States as a whole. 13 The virus spread rapidly among IDUs during the late 1970s, and the prevalence of HIV infection then stabilized at ∼50% from the middle 1980s through the early 1990s. The incidence of HIV infection was estimated to be ∼13 cases per 100 personyears during the period of rapid spread 10 and 4–5 cases per 100 person-years during the period of stable high prevalence. 14 From 1990 to 1996, there were declines in the prevalence, incidence, and injection risk behavior of HIV infection. 15,16 As reported here, from 1997 to 2001, the decline in the prevalence of HIV infection continued but with a different pattern of risk reduction. The recent reductions in risk behavior do not reflect only a simple risk reduction (less sharing of needles and syringes) but may also reflect forms of risk reduction that are not currently recommended by health authorities.
From *Beth Israel Medical Center and †National Development and Research Institutes, Inc., New York, NY.
Received for publication January 13, 2003; accepted October 9, 2003.
Supported by grant AI46370 from the National Institute of Allergy and Infectious Diseases, the National Institutes of Health.
Reprints: Don C. Des Jarlais, Beth Israel Medical Center, 1st Avenue at 16th Street, New York, NY 10003 (e-mail: firstname.lastname@example.org).
The data reported here are part of an ongoing series of studies of IDUs entering the Beth Israel Medical Center (BIMC) drug detoxification program in New York City (methods previously described 10,15–18). The detoxification program serves the city as a whole: approximately one half of its patients live in Manhattan; one quarter, in Brooklyn; one fifth, in the Bronx, and the rest (5%), elsewhere. The program is quite large, with 5000 to 7000 admissions per year, depending on the number of beds allocated to the service. The only criteria for admission are age of 18 years or older and presenting a drug abuse problem sufficient to indicate inpatient detoxification. (The specific indications for inpatient detoxification vary by drugs being used.) These criteria and the admissions policies did not change over the period of 1990–2001.
Patients in the detoxification program were selected for potential participation using techniques designed to produce an unbiased sample of IDUs entering the program. Research staff visited the general admission wards of the program in a set order and examined the intake records to identify patients admitted within the past 3 days who had reported injecting illicit drugs within the previous 2 months. All of these newly admitted IDUs on the specific ward were then asked to participate in the study. The study was fully described to each potential subject in turn, and signed informed consent was obtained from those who agreed to participate. Some eligible patients were unavailable because of appointments scheduled by hospital staff (for radiography, appointments with social workers, physicians, etc). Over 95% of the patients who were approached by our research staff agreed to participate. After all of the available newly admitted IDUs on the ward had been asked to participate, the staff rotated to the next ward. Data collection was continuous over the entire study period.
A trained interviewer administered a structured questionnaire, covering demographics, drug use, injection, sexual risk behavior, and use of HIV prevention services. Risk behaviors were assessed for the 6 months before the interview. Subjects who reported “injecting with needles or syringes used by others” (“receptive sharing”) were then asked from how many different people they had obtained used needles and syringes and their relationships with those people. Pretest counseling was provided, and separate informed consent was obtained for HIV testing. The New York City Department of Health Laboratories, using replicate enzyme-linked immunosorbent assay with Western blot confirmation, conducted the testing for antibody to HIV.
Many drug users undergo multiple detoxification treatments over their drug use careers. Within the sampling procedures described above, we permitted participation multiple times in the study but no more than once in any 12-month period. Note, if we had not permitted subjects to provide data in >1 calendar year, samples from later years would have become biased toward persons with shorter drug use histories and lower prevalence of HIV infection.
We were primarily interested in trends over time in dichotomous variables (eg, the proportion of subjects interviewed in each year who were HIV positive and the proportion of subjects interviewed in each year who reported various risk behaviors in the 6 months before the interview). The Cochran–Armitage test for trend 19 was the primary statistical test used to assess statistical significance of changes in these proportions. χ2 tests were used to assess changes in demographic characteristics and prevalence with demographic subgroups of subjects seen across different years. Linear regression was used to provide an easily interpretable estimate of change per year. In the linear regression models, N equaled 12 years of data collection.
We also used multiple logistic regression to identify subject characteristics associated with HIV-seropositive status. In this analysis, we used only 1 interview per subject. For subjects who had participated more than once in the study, we randomly selected 1 interview for use in the logistic regression analysis. (Because only 5% [153/2887] of the individuals contributed >1 interview, inclusion of all interviews would not have substantively affected the results of these regression analyses. Rerunning the analyses with all interviews confirmed this.)
We also obtained data on the number of syringes exchanged by syringe exchange programs in New York City to help interpret the data on use of these programs by the subjects.
Data analyses were conducted using SAS software. 20
A total of 2887 individuals contributed 3100 observations (both interview and HIV testing) from 1990 through 2001. (As noted above, individual subjects were permitted to participate in the study more than once but only once per year.) Table 1 presents the number of observations per year and sociodemographic characteristics of the subjects seen in each year. The mean age did not differ significantly over the period. There was variation in the sex distribution by year, but no trend toward an increase or decrease in the percentage of males was found. There was significant variation in the proportions of white, black, and Hispanic subjects over the period. Inspection of the proportions showed a decrease in the proportion of black subjects and an increase in the proportion of white subjects.
Figure 1 shows the prevalence of HIV infection per year from 1990 to 2001. There was a very substantial and statistically significant decline in the prevalence of HIV infection during the period (P < 0.0001 by Cochran–Armitage testing), from a high of 54% among subjects interviewed in 1991 to a low of 13% among subjects interviewed in 2001. There was a strong linear association between year of interview and prevalence of HIV infection (r2 = 0.92; P < 0.001). The fitted regression model showed a reduction of 3.7% points per year in the prevalence of HIV infection.
There were also significant declines in the prevalence of HIV infection among both male and female subjects, the three largest racial/ethnic groups, and both short-term injectors (persons who had been injecting for <10 years at the time of interview) and long-term injectors (persons who had been injecting for ≥10 years at the time of interview) (Fig. 2) (all P < 0.001 by Cochran–Armitage tests). Given the long history of the HIV epidemic in IDUs in New York City, 21 loss of HIV-seropositive subjects to disability or death is likely to be a relatively important component of declining seroprevalence among persons who have injected for long periods, while reduced incidence of HIV infection is likely to be a more important component of declining seroprevalence among persons who have injected for short periods. The significant declines among both long-term and short-term injectors suggest that both loss of seropositive subjects and reduced incidence of HIV infection were important factors in the overall decline in the prevalence of HIV infection. (We also examined prevalence among “new” injectors—persons injecting for ≤5 years. Even with the reduced sample sizes, there was a significant decline in prevalence over time among this group [data not presented; available from D.C. Des Jarlais].)
We next examined trends in the utilization of HIV prevention services and in injection risk behaviors as possible contributing factors to the declines in prevalence. Table 2 shows the percentage of subjects reporting having received HIV counseling and testing (at any time before the interview) and the percentage of subjects reporting use of syringe exchange (in the 6 months before the interview) who were interviewed in each year. Use of each of these HIV prevention services increased substantially during the early 1990s and then reached a plateau around 1994, with ∼90% of the subjects reporting that they had been tested for HIV (at some time before the interview) and 40% to 50% of subjects reporting that they had used the syringe exchange programs (in the 6 months before the interview).
The increase in the percentage of subjects using the syringe exchange programs from 1992 to 1994 corresponds to the legalization, public funding, and expansion of the programs during that period. Although the percentage of subjects reporting use of the syringe exchange programs in the previous 6 months stabilized from 1994 on, the programs expanded during this period in terms of the volume of syringes exchanged. The numbers of syringes exchanged by New York City syringe exchange programs increased from ∼250,000 in 1991 to 1.3 million in 1994 to 2.1 million in 1995 to 2.2 million in 1996 to 2.4 million in 1997 to 3.0 million in 1998 and then stabilized at ∼3 million in 2000. These increases occurred through larger numbers of syringes exchanged by persons visiting programs, with some of the sterile syringes then being transferred to IDUs who were not personally using the exchange programs (often referred to as “secondary exchanging”). The percentage of subjects who personally used the syringe exchanges is thus likely to be an underestimate of the extent of syringe exchange activity in the city. Of subjects participating in this study, over one half of those using the syringe exchange programs reported that they “usually” gave some of their syringes to other IDUs (D.C. Des Jarlais et al, unpublished data).
As shown in Figure 3A, distributive sharing (passing on used needles and syringes to others) declined significantly among both HIV-seropositive and HIV-seronegative subjects (both P < 0.001 by Cochran–Armitage tests). HIV-seropositive subjects were less likely to report distributive sharing in almost all years; distributive sharing is presented separately for HIV-seropositive and HIV-seronegative subjects in Figure 3A.
Trends in receptive sharing (injecting with a needle or syringe that had been used by someone else) are presented in Figure 3B. (Because there were no differences in receptive sharing between HIV-seropositive and HIV-seronegative subjects, data are not presented separately by serostatus.) There was no consistent pattern in receptive sharing with any other person over the entire period. There was a statistically significant decline in reporting any receptive sharing from 51% in 1990 to 25% in 1996, the period in which the syringe exchange programs expanded. After 1996, however, there was a leveling or even a possible increase over time (to 39% in 2001). The trend over the entire period did not reach statistical significance (P = 0.07 by Cochran–Armitage testing). The declining trend in receptive sharing with >1 other person, however, was highly statistically significant (P < 0.0001). Receptive sharing outside of one's immediate social network (close friends, sexual partners, and relatives) was low throughout the period, although the declining trend over time was not statistically significant (P = 0.08 by Cochran–Armitage testing).
There were no trends in safe versus unsafe sexual behavior with persons of the opposite sex over this period (data not presented; available from D.C. Des Jarlais). (A separate article on sexual risk behavior is in preparation.) Data on HIV infection in IDUs reporting male-with-male sexual behavior from 1990 to 1999 were presented by Maslow et al. 18
The data in Figures 2B, C and Figure 3A suggest differences in HIV serostatus by race/ethnicity, length of time injecting drugs, and distributive sharing. We used logistic regression to examine the relationships among individual demographic and drug use characteristics (sex, race/ethnicity, age, age at first injection, and length of injecting drug use), individual behaviors (any receptive sharing, any distributive sharing, and use of syringe exchange programs, all in the 6 months before interview), and year of interview (as a 12-point ordinal scale for the years 1990 to 2001) with being HIV positive. Table 3 presents the results of the bivariate analyses and a multiple logistic regression model. In the bivariate analyses, year of interview, Hispanic and black race/ethnicity, distributive syringe sharing, age, age at first injection, and years injecting were all associated with HIV serostatus. All variables that were associated with HIV serostatus at P < 0.15 were then entered into a multiple logistic regression model, with the exception of age, which was highly correlated with years injecting. All these predictors were statistically significant independent predictors of HIV serostatus in the multivariate model.
The odds ratio for year of interview was 0.84 in both the univariate logistic regression and the multiple logistic regression. Thus, statistical control for the other significant predictors did not affect the relationship between year of interview and HIV serostatus.
Being black or Hispanic and having injected for a longer time were associated with being seropositive. These results are consistent with findings of earlier studies of HIV infection in IDUs in New York. 10
Only 1 of the 3 individual risk behaviors was associated with HIV serostatus. Distributive sharing (passing on used needles and syringes) in the 6 months before the interview was inversely associated with being HIV seropositive. As discussed below, this is likely to be the result of the subjects knowing their HIV status from previous HIV testing and changing their behavior to reduce the risk of transmitting HIV to others.
Before considering the implications of the declining trend in the seroprevalence of HIV infection among IDUs entering the BIMC detoxification program in New York City, 4 methodological issues should be discussed.
First, loss of HIV-seropositive IDUs to the subject population of IDUs entering the drug detoxification treatment program at BIMC is undoubtedly an important factor in the declining prevalence, and we were not able to measure directly the loss of HIV-seropositive subjects in this population. Loss of HIV-seropositive subjects could occur through cessation of injecting drug use (or reduction to levels not requiring detoxification treatment) or could be due to disability and death. In a previous study 10 of death rates among HIV-seropositive IDUs in New York, we found a mortality rate of ∼5% per year. The New York City Department of Health collects data on deaths of IDUs diagnosed with AIDS by year. 22 Deaths of IDUs with AIDS rose from 2843 in 1990 to a peak of 5935 in 1993 and then fell rapidly to 1716 in 1996 and to only 103 in 1999. (The steep decline corresponds largely with the introduction of highly active antiretroviral treatment in the city.) It is thus very unlikely that the declining seroprevalence over the 1990–2001 period (Fig. 1) was attributable to an increase in deaths due to AIDS in IDUs in the subject population.
Second, the population of IDUs entering the BIMC detoxification program is clearly not a random sample of IDUs in New York City. A number of other recent studies of IDUs in New York City, however, also showed declining and/or low prevalence of HIV infection. For example, between 1991 and 1996, 16 prevalence fell by 3% per year among IDUs entering methadone maintenance, 4% per year among IDUs seen at a lower east side research storefront, 6% per year among IDUs seen at a Harlem storefront, and 2% per year among IDUs seen at city sexually transmitted disease clinics. More recent data also showed low prevalence among IDUs—3% among young IDUs seen at our lower east side storefront from 1997 to 1999 and 9% among young IDUs seen from 1997 to 1999 at a Harlem storefront, 23 19% among New York City syringe exchange participants seen between 2000 and 2001 (N. Braine, unpublished data), and 0 among young IDUs seen at our lower east side storefront from 2002 to 2002. 24 Prevalence among methadone patients in the Bronx declined from 43% in the late 1980s 25 to “approximately 25% currently” (E. Schoenbaum, personal communication). Prevalence among IDUs in the Bushwick section of Brooklyn was 17% in 2002–2003 (S.R. Friedman, unpublished data). The current prevalence among street-recruited IDUs in Harlem is ∼20% (D. Vlahov, personal communication). In a city as large and diverse as New York, there undoubtedly will be some variation in HIV infection in different groups of IDUs. 26 It is likely, however, that the decline seen in the detoxification patients over the 1990–2001 period reflects a trend occurring in the population of IDUs in the city as a whole. Note, the last time the prevalence of HIV infection was in the 10% to 30% range among IDUs entering the BIMC detoxification program was in 1978–1979. 27
Third, the declining seroprevalence strongly suggests that the incidence of HIV infection was low among this population of IDUs during the 1990–2001 period. From 1995 to 1997, we conducted a cohort incidence study of BIMC detoxification patients in which the incidence of HIV infection was 1.35 cases per 100 person-years (95% confidence interval, 0.65–2.47). 15 From 1990 to 1997, we also conducted a “record-based” incidence study of BIMC detoxification patients who sought voluntary HIV counseling and testing. The incidence in this study was 1.62 cases per 100 person-years (95% confidence interval, 0.20–5.85). 15 We also conducted serologic algorithm for recent HIV seroconversion testing for the incidence of HIV infection among BIMC detoxification subjects; the preliminary analyses showed a reduction of >50% in incidence during the 1990–2001 period (D.C. Des Jarlais, unpublished data).
Other recent studies of IDUs in New York also found a low incidence of HIV infection. In a meta-analysis of 10 studies on the incidence of HIV infection among IDUs in New York City from 1992 to 1997 (including the 2 studies of detoxification patients noted above), the weighted average incidence was 0.7 case 100 person-years. There were >6000 person-years of observation in this meta-analysis. Two more recent incidence studies of young IDUs found rates of 0.0 per 100 person-years and 0.7 case per 100 person-years. 23 Thus, the available data indicate that the incidence was low both among the detoxification patients and among IDUs in New York City as a whole during the period.
Fourth, the behavioral data in this study were obtained through self-reports, and it is likely that social desirability led to some underreporting of risk behaviors. 28 In general, however, studies of self-reported drug use and HIV risk behavior among IDUs showed reasonable reliability and validity. 29,30
Multiple Forms of Drug Injection Risk Reduction
The “official” message for HIV prevention programs for drug users in the United States 22 involves a hierarchy of methods for reducing risk due to injecting drug use: stopping use of illicit drugs; stopping injection of illicit drugs; using a new sterile needle for every injection; and using one's own needle and syringe, without sharing with others. The drug user at risk for exposure to or transmitting HIV is encouraged to choose the highest possible level of this hierarchy that he or she is capable of achieving; however, all levels are considered to have some effectiveness in reducing HIV transmission, and no 1 level is a realistic first step for all IDUs.
The IDUs entering the detoxification program clearly had not been able to achieve either the first level or the second level of this hierarchy in the period before entering the program. Approximately 60% to 70% of them reported that they had achieved the fourth level—not sharing needles and syringes.
The substantial minority who continue to share needles and syringes appear to have adopted 2 additional forms of HIV prevention that are not included in the hierarchy noted above. The HIV-seropositive subjects reported adopting an “informed altruism” method of reducing HIV transmission. Most subjects (∼90% in recent years) reported having been tested for HIV before recruitment into the study. Those who were HIV seropositive were also significantly less likely to report distributive sharing of needles and syringes than HIV-seronegative subjects (Fig. 3A and Table 3).
Our ethnographic work in another study of IDUs in New York provides an example of this type of risk reduction. When IDUs inject in a group setting without sufficient needles and syringes for everyone, they discuss their serostatus. (In 1 case, they used the HIV documentation obtained from the research project.) Members of the group then inject with the HIV-seronegative subjects, injecting before HIV-seropositive subjects (C. Torrico, unpublished data). This “ordered sharing/informed altruism” with respect to transmitting HIV through injection risk behavior is similar to the informed altruism among IDUs who know that they are HIV seropositive and are more likely to consistently use condoms with their regular sexual partners. 31,32 Informed altruism is also a primary factor in the current Centers for Disease Control and Prevention Serostatus Approach to Fighting the Epidemic HIV prevention program. 33
It is important to note the different pattern in risk behaviors between the early and the late 1990s among these subjects. Receptive syringe sharing declined with the legalization and expansion of syringe exchange programs that began in late 1992 but then stabilized around 1996. The legalization and expansion of the syringe exchange programs have not led to elimination of receptive sharing of needles and syringes. This residual receptive sharing, however, appears to be restricted with respect to the number of sharing partners.
The combination of reduced rates of transmission risk behavior by HIV-seropositive subjects (informed altruism) (Figs. 3A, B) and exposure risk behavior that is primarily confined to dyads or small groups (partner restriction) would be expected to produce relatively low rates of new HIV infections among the population. The combined effect of these 2 additional forms of risk reduction would reduce the likelihood that an HIV-seropositive IDU would pass on a needle or syringe to an HIV-seronegative IDU. With low rates of injecting risk partner change, the likelihood of new infections would further decrease over time, because the partner-restricted dyads or small groups in which receptive sharing was occurring would increasingly involve persons of the same HIV status.
On the basis of these New York City data, we would suggest consideration and research concerning adding 2 new levels to the hierarchy of messages for reducing HIV transmission among IDUs: know your HIV status, and if you are positive, do not pass on your needles and syringes (and do not engage in unprotected sexual intercourse); and if you feel you need to share needles and syringes (or to engage in unprotected sex), try to do this with as few other persons as possible. It may seem dangerous to “recommend” continuation of any form of risk behavior to persons at high risk for HIV infection. These additional forms of risk reduction may also not be effective in reducing hepatitis B or C transmission among IDUs. However, these additional forms of risk reduction do conform to the idea of a hierarchy of risk reduction practices, each level of which has some effectiveness in reducing HIV transmission, with an individual IDU choosing the maximum level of risk reduction that he or she is likely to achieve.
We would caution that the effectiveness of these additional forms of risk reduction may require an environment in which there is reasonably good access to sterile injection equipment and where the stigmatization of HIV is not so great that HIV-seropositive subjects cannot disclose their status to potential risk partners.
Whether the declines in the prevalence of HIV infection, distributive sharing among HIV-positive subjects, and “higher risk” receptive sharing among HIV-negative subjects continue over time can only be answered through continued research. Given the recent resurgence of HIV risk behavior among young men who have sex with men in New York and other cities in the United States, 34 the possibility of a similar resurgence in risk behavior and HIV infection in IDUs must also be considered.
Relevance to Other Cities
Declines in the prevalence of HIV infection among IDUs were reported in several other areas where the prevalence of HIV infection reached high levels among IDUs, including Amsterdam, 35,36 Edinburgh, 37 Geneva, 38 northern Italy, 39 and several northeastern and southern cities in the United States. 40 The declines in prevalence in these cities appear to be combinations of loss of HIV-seropositive subjects to the active injecting population and risk reduction among the active injectors. The New York City situation is unique, however, in terms of the sheer size of the IDU population (estimated to be between 160,000 41 and 200,000 42), in the magnitude of the decline (from ∼50% to <15% among our detoxification subjects), and in the consistency of the decline over a long period. As noted above, the last time that the prevalence of HIV infection was in the 10% to 30% range among our detoxification subjects was in 1978–1979. 27
There are, of course, other areas where the prevalence of HIV infection appears to be remaining high indefinitely (eg, Bangkok 43) and areas in which the prevalence has reached extremely high rates of ≥80% (eg, in parts of India 44 and China 45).
There was a substantial and consistent decline in the prevalence of HIV infection among IDUs entering the BIMC detoxification program in New York from 1990 to 2001. Other recent studies also showed declining and/or relatively low prevalence of HIV infection among IDUs in the city. This has occurred in the context of expanded syringe exchange and increased voluntary HIV counseling and testing. In addition to general reductions in injection risk behavior, 2 conditional forms of risk reduction were noted: informed altruism, in which persons who knew that they were HIV seropositive reduced transmission behavior, and partner restriction, in which persons who shared needles and syringes primarily confined this sharing within small social networks. The trends in the prevalence of HIV infection from 1990 to 2001 suggest that, over long periods, it may be possible to bring very large, high seroprevalence HIV epidemics among IDUs under control.
1. Ball AL, Rana S, Dehne K. HIV prevention among injecting drug users: responses in developing and transitional countries. Public Health Rep. 1998; 113(suppl 1):170–181.
2. UNAIDS/WHO. AIDS Epidemic Update: December 2002
[UNAIDS/02.46E]. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS); 2002.
3. National Institutes of Health US. Interventions to Prevent HIV Risk Behaviors.
Bethesda, MD: National Institutes of Health; February, 1997:104.
4. Des Jarlais DC, Friedman SR. Strategies for preventing HIV infection among injecting drug users: taking interventions to the people. In: Schneiderman N, Speers MA, Silva JM, et al, eds. Integrating Behavioral and Social Sciences with Public Health. Washington: American Psychological Association; 2001:141–158.
5. Stimson GV, Des Jarlais DC, Ball A, eds. Drug Injecting and HIV Infection: Global Dimensions and Local Responses.
London: UCL Press; 1998.
6. Wodak A. HIV Epidemiology and Prevention Research in Drug-Using Populations in Australia.
Office of AIDS Research, National Institutes of Health, World Health Organization Substance Abuse Department; June 25–26, 1998.
7. Coutinho R. Interventions among injecting drug users: the Amsterdam experience. 4th Science Forum: Research Synthesis Symposium on the Prevention of HIV in Drug Abusers, 1997, Flagstaff, AZ.
8. Vanichseni S, Kitayaporn D, Mastro TD, et al. Continued high HIV-1 incidence in a vaccine trial preparatory cohort of injection drug users in Bangkok, Thailand. AIDS. 2001; 15:397–405.
9. Spittal PM, Craib KJ, Wood E, et al. Risk factors for elevated HIV incidence rates among female injection drug users in Vancouver. CMAJ. 2002; 166:894–899.
10. Des Jarlais DC, Friedman SR, Sotheran JL, et al. Continuity and change within an HIV epidemic: injecting drug users in New York City, 1984 through 1992. JAMA. 1994; 271:121–127.
11. D'Aquila RT, Williams AB, Petersen LR, et al. HIV seroprevalence among Connecticut intravenous drug users in 1986-87: race/ethnicity as a risk factor for HIV seropositivity. 3rd International Conference on AIDS, 1987, Washington, DC.
12. Weiss SH, Ginzburg HM, Goeddert JJ, et al. Risk for HTLV-III exposure and AIDS among parenteral drug abusers in New Jersey. 1st International Conference on AIDS, 1985, Atlanta, GA.
13. Friedman SR, Perlis T, Des Jarlais DC. Laws prohibiting over-the-counter syringe sales to injection drug users: relations to population density, HIV prevalence, and HIV incidence. Am J Public Health. 2001; 91:791–793.
14. Holmberg S. The estimated prevalence and incidence of HIV in 96 large US metropolitan areas. Public Health. 1996; 86:642–654.
15. Des Jarlais DC, Marmor M, Friedmann P, et al. HIV incidence among injecting drug users in New York City, 1992-1997: evidence for a declining epidemic. Am J Public Health. 2000; 90:352–359.
16. Des Jarlais DC, Perlis T, Friedman SR, et al. Behavioral risk reduction in a declining HIV epidemic: injection drug users in New York City, 1990-1997. Am J Public Health. 2000; 90:1112–1116.
17. Des Jarlais DC, Friedman SR, Novick D, et al. HIV-l infection among intravenous drug users in Manhattan, New York City, from 1977 through 1987. JAMA. 1989; 261:1008–1012.
18. Maslow CB, Friedman SR, Perlis TE, et al. Changes in HIV seroprevalence and related behaviors among male injection drug users who do and do not have sex with men: New York City, 1990-1999. Am J Public Health. 2002; 92:382–384.
19. Agresti A. Categorical Data Analysis.
New York: John Wiley & Sons; 1990.
20. SAS. SAS/STAT™ User's Guide, Version 8.
Cary, NC: SAS Institute, Inc.; 1999.
21. Des Jarlais DC, Hagan HH, Friedman SR, et al. Maintaining low HIV seroprevalence in populations of injecting drug users. JAMA. 1995; 274:1226–1231.
22. Centers for Disease Control and Prevention. CDC HIV/AIDS Prevention Website.
National Center for HIV, STD and TB Prevention/Divisions of HIV/AIDS Prevention.
23. Des Jarlais DC, Diaz R, Perlis T, et al. Variability in the incidence of HIV, HBV, and HCV infection among young injecting drug users in New York City. Am J Epidemiol. 2003; 157:467–471.
24. Perlis TE, Des Jarlais DC, Torrico CR, et al. Homelessness among young geographically mobile drug users. 103rd annual meeting of the American Public Health Association, 2002, Philadelphia.
25. Hartel DM, Schoenbaum EE. Methadone treatment protects against HIV infection: two decades of experience in the Bronx, New York City. Public Health Rep. 1998; 113(suppl 1):107–115.
26. Rockwell R, Deren S, Goldstein MF, et al. Trends in the AIDS epidemic among New York City's injecting drug users: localized or citywide? J Urban Health. 2002; 79:136–146.
27. Des Jarlais DC, Friedman SR, Novick DM, et al. HIV-l infection among intravenous drug users in Manhattan. JAMA. 1989; 261:1008–1012.
28. Des Jarlais D, Paone D, Milliken J, et al. Audio-computer interviewing to measure HIV-risk behaviour among injecting drug users: a quasi-randomised trial. Lancet. 1999; 353:1657–1661.
29. Des Jarlais DC, Friedmann P, Hagan H, et al. The protective effect of AIDS-related behavioral change among injection drug users: a cross-national study. Am J Public Health. 1996; 86:1780–1785.
30. Weatherby NL, Needle R, Cesari H, et al. Validity of self-reported drug use among injection drug users and crack cocaine users recruited through street outreach. Eval Program Plann. 1994; 17:347–355.
31. Friedman S, Curtis R, Neaigus A, et al. Social Networks, Drug Injectors' Lives and HIV/AIDS.
New York: Plenum; 1999.
32. Vanichseni S, Des Jarlais DC, Telles P, et al. Condom use with casual partners among injecting drug users. 9th International Conference on AIDS, 1993, Berlin, Germany.
33. Janssen RS, Holtgrave DR, Valdiserri RO, et al. The serostatus approach to fighting the HIV epidemic: prevention strategies for infected individuals. Am J Public Health. 2001; 91:1019–1024.
34. Stall R, Pollack L, Mills T, et al. Use of antiretroviral therapies among HIV-infected men who have sex with men: a household-based sample of 4 major American cities. Am J Public Health. 2001; 91:767–773.
35. Coutinho RA. The Amsterdam Cohort Studies on HIV infection and AIDS. J Acquir Immune Defic Syndr. 1998; 17(suppl 1):S4–S8.
36. Perez K, Rodes A, Merona M, et al. Behavioural surveillance among intravenous drug users (IDUs) out-of-treatment in Barcelona (Spain), 1993-2000. 14th International AIDS Conference, 2002, Barcelona, Spain.
37. Robertson E, Sloboda Z, Boyd G, et al. eds. Rural Substance Abuse: State of Knowledge and Issues
[National Institute on Drug Abuse, Research Monograph Series no. 168]. Rockville, MD: US Department of Health and Human Services, National Institutes of Health; 1997.
38. Broers B, Junet C, Bourquin M, et al. Prevalence and incidence of HIV, hepatitis B and C among drug users on methadone maintenance treatment in Geneva between 1988 and 1995. AIDS. 1998; 12:2059–2066.
39. Nicolosi A, Correa L, Chiesa A, et al. Incidence and prevalence trends of HIV infection among IDUs in northern Italy, 1990-1994. Int Conf AIDS. 1996; 11:359.
40. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report.
Atlanta, GA: Centers for Disease Control and Prevention; 2001: 13(2).
41. New York City Department of Health and Mental Hygiene. AIDS/HIV Semi Annual Report Surveillance Update
[4th quarter]. New York: NYC DOH; 2002.
42. Eisenhandler J, Drucker E. Opiate dependency among the subscribers of a New York area private insurance plan. JAMA. 1993; 269:2890–2891.
43. Vanichseni S, Choopanya K, Des Jarlais DC, et al. HIV among injecting drug users in Bangkok: the first decade. Int J Drug Policy. 2002; 13:39–44.
44. (MACS) MSACS. Epidemiological Analysis of HIV/AIDS in Manipur, September 1986 to February 2002.
Imphal: Manipur State AIDS Control Society; 2002.
45. Zheng X, Zhang J, Tian C, et al. Cohort study of HIV infection among IDUs in Ruili, Longchuan, and Luxi of Yunnan Province, China. Biomed Environ Sci. 1993; 6:348–351.