Drug users are at high risk for human immunodeficiency virus (HIV) infection. Moreover, injecting drug users function as a bridge group through which HIV can be transmitted to non-injecting drug users and the general population, driving the heterosexual spread of HIV [1,2]. Over the years, substantial decreases in HIV incidence or moderately stable HIV incidence rates have been observed among some drug using populations [3–7] as have reductions in borrowing and lending of needles and sexual risk behaviour [5,8–10].
Since highly active antiretroviral therapy (HAART) became generally available in 1996, sexual risk behaviour and sexually transmitted infections (STI) have increased among men who have sex with men (MSM) . Comparable data on drug users in the HAART era are limited, and most studies have evaluated changes in injecting and sexual risk behaviour among HIV-positive drug users using HAART, with conflicting results [12–14]. Recently, we have found that drug users who are treated with HAART are not increasing their risk behaviour .
The present study investigated trends in HIV incidence and both injecting and sexual risk behaviours among HIV-negative drug users of the Amsterdam Cohort Study (ACS). The long follow-up time of this cohort enabled us to evaluate temporal changes in risk behaviour, in HIV incidence, and in HIV transmission routes since 1985. In addition, we evaluated changes in trends since the widespread use of HAART.
Materials and methods
In December 1985, an ongoing, open prospective cohort study among drug users was initiated as part of the Amsterdam Cohort Studies (ACS) on HIV/AIDS . Both HIV-negative and HIV-positive participants were recruited. At follow-up visits every 4–6 months, blood was drawn for HIV testing by enzyme linked immunosorbent assays (ELISA) and confirmation by western blot (since 1995: HIV Blot version 2.2; Genelabs Diagnostic, Singapore). Counselling was conducted and standardized questionnaires were administered by trained nurses to collect data on clinical, epidemiological, sexual, and drug use-related topics.
At ACS entry, data collection on injecting and sexual risk behaviour relates to the 6-month period preceding enrolment. For STI the preceding period was 5 years; after March 1989, it refers back to 1980. At follow-up visits, questionnaires concern the period that has passed since the last visit. The present study included 1315 drug users who were HIV-negative on cohort entry between December 1985 and December 2004.
During HIV-negative follow-up time, we defined five main outcome variables: injecting (yes/no), borrowing needles (yes/no), using needle exchange units (yes/no), having any unprotected sex (yes/no), and having STI (syphilis, gonorrhea, herpes simplex virus or other STI: yes/no). In addition, trends in unprotected sex were studied separately for a steady partner, a casual partner, or exchanging sex for money.
An extra questionnaire was introduced in 1994 by which HIV seroconverters are questioned in detail, as to the putative route of HIV transmission. For those seroconverters who became lost to follow-up until 1994, these questionnaires were filled out by the study physician, based on cohort data and clinical notes. HIV seroconverters were allocated to three hierarchical groups based on transmission mode: (1) injecting; (2) homosexual; or (3) heterosexual.
Yearly HIV incidence per 100 person-years was calculated using person-time techniques. The midpoint between the last negative and first seropositive test was regarded as the date of HIV seroconversion. We studied trends in risk behaviour by calculating the odds ratios (ORs) per calendar year increase in the periods before and after 1996 within a logistic regression model, using the generalised estimating equations (GEE) method with a compound symmetry covariance matrix to adjust for correlations between visits of the same individual . All models were corrected for gender, age (< 28, 28–34, > 34 years, based on tertiles) and calendar year of ACS entry, and we checked for interactions of these variables with calendar time. The most likely route of HIV transmission was studied among those who seroconverted for HIV using logistic regression analysis to calculate the OR for a sexual versus an injecting HIV transmission route before and after 1996 .
Study participants and demographics
Of the 1315 HIV-negative drug users, 63% were male and 77% were of European ethnicity. On ACS entry, the median calendar year was 1991 [interquartile range (IQR), 1987–1998] and the median age was 30 years (IQR, 26–35 years). After 6970 HIV-negative person-years of follow-up, 93 participants seroconverted for HIV and 135 individuals died. Per individual, the median number of visits within 6 months from the preceding visit was 6 (IQR, 1–16). On average, 90% of participants that visited the ACS a given calendar year returned the next year.
HIV incidence and HIV transmission routes
The HIV incidence was 7/100 person-years in 1986 [95% confidence interval (95% CI), 3–16], dropping to 2/100 person-years in 1995 (95% CI, 1–4) (see Fig. 1) and continued to decline thereafter. Since 1999, the HIV incidence has fluctuated between 0 and 0.5/100 person-years.
Due to missing data, 65 out of 93 HIV seroconverters were included in the analysis of HIV transmission routes. Before 1996, 50 were related to injecting risk behaviour and eight were related to unprotected heterosexual contact, whereas after 1996, these figures were two and five, respectively. After 1996, compared with the period before, the unadjusted OR was 15.6 (95% CI, 2.6–94.6) for HIV transmission through unprotected heterosexual contact.
Trends in risk behaviour
The proportion of visits at which injecting was reported declined from 57% in 1986 to 21% in 2004 (Fig. 1). Among injectors only, borrowing needles decreased from 47 to 9% (data not shown). After 1996, the decline in injecting became steeper than before this period (P < 0.05; Table 1). However, among injectors only, the decline in borrowing needles was less pronounced. Interestingly, before 1996, the use of needle exchange units increased significantly, whereas thereafter a significant decrease was found.
Trends in sexual risk behaviour differed in general profile from trends in injecting risk behaviour. Visits at which any unprotected sex was reported declined from 52% in 1990 to 40% in 1996, and remained stable thereafter. Over the period 1990–2004, unprotected sex with steady partners moderately declined from around 35 to 24% (data not shown). Evaluation of time trends before and after 1996 for unprotected sex with casual partners and while exchanging sex for money found them similar to trends for any unprotected sex; that is, no significant decline in risk behaviour occurred after 1996 (Table 1).
The proportion of follow-up visits at which STI were reported decreased from 17% in 1986 to 3% in 2004 (Fig. 1). A significant decline was found before 1996, after which STIs tended to increase slightly over time (Table 1). For STIs reported at study entry, the decline was only significant after 1996.
In order to investigate the role of selective loss to follow-up and changes in inclusion criteria over time, all analyses were repeated with exclusion of those who died during follow-up, those who visited the ACS only once and drug users under 30, leaving 721 participants available for this analysis. In addition, analyses were repeated including 2516 visits more than 6 months from the preceding visit (16% of 16 128 visits). Results were comparable with those above for all outcome variables (data not shown).
We found a steep decline in HIV incidence in the drug user cohort in Amsterdam. After peaking at 7/100 person-years in 1986, the incidence varied between 0 and 0.5 per 100 person-years since 1999. This is comparable with the rate found by others in some West European countries [7,19–21] and The Netherlands , but lower than was recently observed in London . Factors related to drug use that could explain this steep decline in HIV incidence in the ACS include a strong decline in injecting, which became even stronger after the introduction of HAART in 1996. Borrowing needles likewise declined in the whole study period, although this decline became less pronounced in the HAART era.
A declining trend in the use of needle exchange units was observed in our drug user cohort after 1996. This observation was accompanied by a reduction in the absolute number of exchanged needles per calendar year in Amsterdam, which peaked in 1992 with 1 100 000 whereas in 2002 approximately 350 000 needles were exchanged . Since accessibility of needle exchange units remained similar over the years (personal communication, G. van Brussel) these findings confirm that injecting drug use in Amsterdam has become less popular over time. Moreover, in 1986, 54% of ACS participants reported any non-injecting use of hard drugs whereas this proportion was 83% in 2004 (data not shown). This could indicate a shift from injecting to non-injecting drug use. All these observations suggest that harm reduction models such as those implemented in Amsterdam do not have a stimulatory effect on injecting in drug users in Amsterdam. This is also supported by other studies that found acute hepatitis B (HBV) infections reduced among injecting drug users in Amsterdam over the period 1992–2003 . Hepatitis C prevalence, among young drug users in our cohort, has likewise dropped between the periods 1985–1989 (83%) and 2000–2004 (14%) .
Cross-sectional studies from the 1990s found a potentially important role for sexual risk behaviour in HIV transmission among drug users  as did recent prospective studies . Although numbers were small, our study confirms this role. The stable trends in sexual risk behaviour and a possible increase in STI after 1996 raises concern for continued sexual HIV transmission among drug users, underscoring the need for prevention measures that focus on sexual risk reduction in this group.
Other factors might have affected the HIV incidence as well. Even though HIV-positive drug users may have less access to HAART than MSM, a virological response in those who do receive it will leave them less infectious. On the other hand, HAART might increase the pool of HIV-infected drug users because of their prolonged survival. In contrast, now more than 20 years into the HIV epidemic, the majority of those with high-risk behaviour who became infected with HIV early in the epidemic have died.
Selection bias prevents generalizations of our study results to the whole Amsterdam drug-user population and we cannot exclude that a cohort effect might partially explain the observed decrease in HIV incidence and injecting behaviour. Another limitation is that data on risk behaviour were self-reported, and bias toward socially desirable answers could cause underestimation of the proportion engaged in risk behaviour.
Although the decline in injecting and sexual risk behaviour in our cohort started before 1996, that year was chosen as the change point for the trends as we were interested in evaluating risk behaviour since the introduction of HAART. Whether this introduction has a causal relationship to the observed trends in risk behaviour or whether they reflect more general trends could not be determined. Among HIV-positive drug users in the ACS, no causal relationship was found between the use of HAART and the observed decline in injecting risk behaviour . In addition, a small decrease in sexual risk behaviour was demonstrated among those on HAART.
In conclusion, the HIV incidence among drug users in the ACS has declined since 1985. Accompanied by a reduction in injecting drug use and needle sharing, this decline occurred despite continued sexual risk behaviour. At present, new HIV seroconversions are related mainly to unprotected heterosexual contacts. Therefore, HIV prevention programmes for drug users should pay specific attention to the importance of safe sex practices.
The authors would like to thank J. Bax, A. Snuverink and M. Totté for data collection and blood sampling; all participants for study participation; R.B. Geskus for help in statistical analyses, and L. Phillips for the editing of the manuscript.
Sponsorship: This study was performed as part of the Amsterdam Cohort Studies on HIV/AIDS, a collaboration between the Health Service of Amsterdam, the Academic Medical Center, and CLB/Sanquin, Amsterdam, The Netherlands, sponsored by the Netherlands Organisation for Health Research and Development ZonMw, the Ministry of Health, Welfare and Sport, and Aids Fonds Netherlands (grant number 4141).
1. Lukashov VV, Op de Coul ELM, Coutinho RA, Goudsmit J. HIV-1 strains specific for Dutch injecting drug users in heterosexually infected individuals in The Netherlands. AIDS 1998; 12:635–641.
2. Donoghoe MC. Sex, HIV and the injecting drug use [review]. Br J Addict 1992; 87:405–416.
3. Nelson KE, Galai N, Safaeian M, Strathdee SA, Celentano DD, Vlahov D. Temporal trends in the incidence of human immunodeficiency virus infection and risk behavior among injection drug users in Baltimore, Maryland, 1988–1998. Am J Epidemiol 2002; 156:641–653.
4. Kral AH, Lorvick J, Gee L, Bacchetti P, Rawal B, Busch M, et al
. Trends in human immunodeficiency virus seroincidence among street-recruited injection drug users in San Francisco, 1987–1998. Am J Epidemiol 2003; 157:915–922.
5. van Ameijden EJC, Coutinho RA. Maximum impact of HIV prevention measures targeted at injecting drug users. AIDS 1998; 12:625–633.
6. Hernandez-Aguado I, Avino MJ, Perez-Hoyos S, Gonzalez-Aracil J, Ruiz-Perez I, Torrella A, et al
. Human immunodeficiency virus (HIV) infection in parenteral drug users: evolution of the epidemic over 10 years. Valencian Epidemiology and Prevention of HIV Disease Study Group. Int J Epidemiol 1999; 28:335–340.
7. Sabbatini A, Carulli B, Villa M, Correa LM, Nicolosi A. Northern Italian Seronegative Drug Addicts. Recent trends in the HIV epidemic among injecting drug users in Northern Italy, 1993–1999. AIDS 2001; 15:2181–2185.
8. Des Jarlais JC, Perlis T, Friedman SR, Chapman T, Kwok J, Rockwell R, 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.
9. Vlahov D, Junge B, Brookmeyer R, Cohn S, Riley E, Armenian H, et al
. Reductions in high-risk drug use behaviors among participants in the Baltimore needle exchange program. J Acquir Immune Defic Syndr Human Retrovirol 1997; 16:400–406.
10. van Ameijden EJC, van den Hoek JAR, Coutinho RA. Large declines in sexual risk behavior with noncommercial partners among heterosexual injection drug users in Amsterdam, 1989–1995. Am J Epidemiol 1996; 144:772–781.
11. Stolte IG, Coutinho RA. Risk behaviour and sexually transmitted diseases are on the rise in gay men, but what is happening with HIV? Curr Opin Infect Dis 2002; 15:37–41.
12. Tun W, Gange SJ, Vlahov D, Strathdee SA, Celentano DD. Increase in sexual risk behavior associated with immunologic response to highly active antiretroviral therapy among HIV-infected injection drug users. Clin Infect Dis 2004; 38:1167–1174.
13. Vlahov D, Safaien M, Lai S, Strathdee SA, Johnson L, Sterling T, et al
. Sexual and drug risk-related behaviours after initiating highly active antiretroviral therapy among injection drug users. AIDS 2001; 15:2311–2316.
14. Bouhnik AD, Moatti JP, Vlahov D, Gallais H, Dellamonica P, Obadia Y. Highly active antiretroviral treatment does not increase sexual risk behaviour among French HIV infected injecting drug users. J Epidemiol Community Health 2002; 56:349–353.
15. Smit C, Lindenburg K, Geskus RB, Brinkman K, Coutinho RA, Prins M. Highly active antiretroviral therapy (HAART) among HIV-infected drug users: a prospective cohort study of sexual risk and injecting behaviour. Addiction 2006; 101:433–440.
16. van den Hoek JAR, Coutinho RA, van Haastrecht HJA, van Zadelhoff AW, Goudsmit J. Prevalence and risk factors of HIV infections among drug users and drug-using prostitutes in Amsterdam. AIDS 1988; 2:55–60.
17. SAS Institute Inc. Base SAS 9.1 Procedures Guide
. Cary, NC: SAS Institute Inc.; 2005.
18. SPSS Inc. SPSS 13.0 for Windows, Release 13.0
. Chicago, Illinois: SPSS Inc.; 2004.
19. McFarland W, Kellogg TA, Louie B, Murrill C, Katz MH. Low estimates of HIV seroconversions among clients of a drug treatment clinic in San Francisco, 1995 to 1998. J Acquir Immune Defic Syndr 2000; 23:426–429.
20. Des Jarlais JD, Marmor M, Friedmann P, Titus S, Aviles E, Deren S, et al
. HIV incidence among injection drug users in New York City, 1992–1997: evidence for a declining epidemic. Am J Public Health 2000; 90:352–359.
21. Lucidarme D, Bruandet A, Ilef D, Harbonnier J, Jacob C, Decoster A, et al
. Incidence and risk factors of HCV and HIV infections in a cohort of intravenous drug users in the North and East of France. Epidemiol Infect 2004; 132:699–708.
22. Laar MJW van de, Boer IM de, Koedijk FDH, Op de Coul ELM. HIV and sexually transmitted infections in the Netherlands in 2004 An update: November 2005
. Bilthoven: RIVM; 2006.
23. Judd A, Hickman M, Jones S, McDonald T, Parry JV, Stimson GV, et al
. Incidence of hepatitis C virus and HIV among new injecting drug users in London: prospective cohort study. BMJ 1905; 330:24–25.
24. Brussel GHA van, Buster MCA. Monitor of public health care: 2002, 2003, 2004
. Amsterdam: Health Service Amsterdam; 2004.
25. Anon. Abstracts of the 40th Annual Meeting of the European Association for the Study of the Liver. April 13–17, 2005. Paris, France [abstract 415]
. J Hepatol
2005; 42(2 Suppl)
26. van de Laar TJ, Langendam MW, Bruisten SM, Welp EA, Verhaest I, van Ameijden EJ, et al
. Changes in risk behavior and dynamics of hepatitis C virus infections among young drug users in Amsterdam, the Netherlands. J Med Virol 2005; 77:509–518.
27. Kral AH, Bluthenthal RN, Lorvick J, Gee L, Bacchetti P, Edlin BR. Sexual transmission of HIV-1 among injection drug users in San Francisco, USA: risk-factor analysis. Lancet 2001; 357:1397–1401.
28. Strathdee SA, Sherman SG. The role of sexual transmission of HIV infection among injection and non-injection drug users. J Urban Health 2003; 80(4 suppl 3):iii7–iii14.