- Email: [email protected]
HIV prevalence and risk behaviour in injecting drug users in Berlin
Forensic Science International 62 (1993) 73-81
Frrrensic Science International
HIV prevalence and risk behaviour in injecting drug users in Berlin K. Stark*,
R. Miiller
Landesinstitut fiir Tropenmedizin, Institut ftir Medisinische Siatistik und Informationsverarbeitung. Freie Vniversitiit Berlin, Engeldamm 62-64, 10179 Berlin, Germany
(Accepted I September 1992)
Abstract
In injecting drug users (IDU) newly entering drug treatment centres in Berlin, HIV seroprevalence continuously decreased from 1985 through 1992. A recent cross-sectional study with multi-site sampling (n = 472) showed significantly differing HIV prevalences by site of recruitment (treatment centres 6%, storefront units 20%, infectious disease clinic 56%). In multivariate analysis, the borrowing of syringes in prison was the strongest independent predictor of HIV infection. A majority of the IDU have reduced risks of infection. However, current risk behaviour is common. Of all IDU, 48% reported the borrowing of syringes within the previous 6 months, only 13% had consequently used condoms. Current borrowing of syringes was associated with younger age, shorter history of intravenous drug use, negative HIV serostatus, and non-participation in methadone maintenance programmes. Continued preventive efforts against HIV transmission among IDU are needed. Key words:
HIV infection; Injecting drug users; Risk behaviour; Risk reduction
1. Introduction AIDS and HIV infection are still major problems in injecting drug users (IDU) throughout the world. In Europe, since 1991 IDU represent the largest transmission group among diagnosed AIDS cases. In 1992, 37.4% of the 14 137 AIDS cases were IDU [l]. In Germany, 13.7% of all AIDS cases are IDU (December 1992). From 1983 through 1992, 246 AIDS cases have been diagnosed among IDU in Berlin * Corresponding
author.
0379-0738/93/$06.00 0 1993 SSDI 0379-0738(93)01361-U
Elsevier Scientific Publishers Ireland Ltd. All rights reserved
74
K. Stark, R. Miiller/Forensic Sci. hf. 62 (1993) 73-81
(12.1%). This proportion has been stable in the last years [2,3] To determine the current spread of HIV infection among IDU, seroepidemiological studies are necessary. In many industrialized countries, HIV seroprevalences in IDU have levelled off or even declined in recent years [4-71. IDU have reduced risks of infection mainly by avoiding sharing injection equipment. However, risk behaviour still occurs in substantial proportions of IDU [4,8-lo]. Moreover, the severe HIV epidemics among IDU in Thailand (Bangkok), Myanmar and Poland (Warsaw) again show that under certain circumstances such as lack of information and lack of sterile injection equipment, IDU are very vulnerable to HIV infection [4]. We have been continously monitoring HIV seroprevalence in IDU in drug treatment centres in Berlin since 1985. In an ongoing cross-sectional study with multisite sampling we assess HIV prevalence, determinants of HIV infection, risk behaviour and behaviour changes, sociodemographic variables, health problems and utilization of medical services. In this paper we present data on the HIV epidemiology in IDU in Berlin. 2. Subjects and methods Since 1985 the Institute of Tropical Medicine offers counselling on HIV infection and anonymous testing for HIV antibodies to clients of several drug treatment centres in Berlin. Throughout the study period, about 85% of all IDU entering treatment underwent the test. Sociodemographic and behavioural data were assessed by standardized questionnaires. Screening for HIV antibodies was done by commercially available ELISA. Repeatedly positive results were confirmed by Western blotting according to the guidelines of WHO [l 11. Overall, 1813 IDU have been tested up to 3 1 December 1992. In 1992, we began a cross-sectional study on the epidemiology of HIV infection and other.parenterally transmitted diseases (hepatitis B, C, D, syphilis) in IDU. Participants were anonymously recruited from treatment centres, storefront units, by street recruitment and from an infectious disease clinic. HIV antibodies were determined in serum or saliva. Participants were interviewed by trained personnel using a standardized questionnaire. For statistical analysis, bivariate (N2-test) and multivariate methods (logistic regression analysis) were used. 3. Results 3.1. Time trends of HIV seroprevalence in IDU in treatment centres, 1985-1992 HIV seroprevalence was 24.9% in IDU tested in 1985. Some of these individuals had already stopped i.v. drug use several years before. Thus, HIV seroprevalence increased from 18.4% in IDU having injected until 1983 to 30.9% in those still injecting in 1985. In the following years it declined continously to 3.2% in 1992 (Fig. 1). However, after adjusting for confounding factors (i.v. drug use in prison, history of i.v. drug use, positive seromarkers to hepatitis B virus) the HIV prevalence odds ratio significantly increased until 1988 (Fig. 2). In the following years, the set of confounders changed. Therefore it was not possible to adjust for the whole study period.
K. Stark, R. Miller/Forensic HIV seroprevalence
Sci. ht. 62 (1993)
73-81
75
(%)
5o 1
0
I
I
51983
I
1984
I
,
1985
1986
I
1987
I
I
1988
1989
I
I
I
1990
1991
1992
year of last drug injection Fig. 1. Time trend of the crude
Adjusted
HIV prevalence
HIV seroprevalence
in IDU in treatment
centres
(n = 1813)
odds ratio
5-
4-
3-
2-
~_~.______________.________‘_____-_ .
0
)
‘1983
I
1984
I
1985
I
1986
I
1987
I
1988
year of last injection Fig. 2. Adjusted HIV prevalence odds ratios (with 95”/0 confidence intervals) in IDU from treatment centres. Figures are adjusted for i.v. drug use in prison, history of iv. drug use, and hepatitis B serostatus.
K. Stark, R. Miiller/ Forensic Sri. ht. 62 (1993)
76
73-81
A significant decrease over time was noticed in the prevalence of sexual contacts with HIV-infected partners, imprisonment, i.v. drug use in prison, and prostitution. Other variables, such as lifetime needle sharing and sexual contacts with i.v. drug users, did not show significant changes throughout the study period. 3.2. Results of the cross-sectional study, 1992/93 Data are presented from 472 IDU examined in the ongoing cross-sectional study. The mean age of the participants was 29.8 years (S.D., 5.8; median, 29) and the mean
Table 1 Basic characteristics of the study population (n = 472) Variable
Prevalence (“h)
Sex
Men Women
66.9 33.1
Nationality
German Other
93.2 6.8
Age <25 years 25-29 30-34 235
16.9 34.1 28.8 20.1
History of i.v. drug use <5 years
5-9 210
27.3 27.8 44.9
Site of recruitment
Treatment centres Storefront units Streets Infectious disease clinic
33.1 39.0 14.6 13.3
Injected drugs
Heroin Cocaine Benzodiazepines Barbiturates Amphetamines
98.3 12.5 51.1 23.1 13.8
Previous imprisonment
Men Women
16.3 53.8
i. v. drug use in prison
Men Women
44.9 21.2
Prostitution
Men Women Previous drug treatment
12.7 73.1 71.0
K. Stark, R. Miiller /Forensic
Sci. Int. 62 (1993)
II
73-81
duration of i.v. drug use was 9.6 years (S.D., 6.5; median 8). Table 1 shows basic characteristics of the study population. Sites of recruitment were treatment centres in 33. lo!, storefront units in 39.0%, streets in 14.6%, and an infectious disease clinic in 13.3% of the participants. The main drugs injected were heroin, cocaine, and benzodiazepines (e.g., flunitrazepam, diazepam). The majority of the individuals (79%) injected more than once daily. Of all subjects, 69% have been imprisoned (men, 76.3%; women, 53.8%). Injecting in prison was reported by 44.9% of the men, and 21.2% of the women, respectively.
Table 2 Association of HIV infection and different variables HIV seroprevalence % out of (n)
Variable
n.s.
Sex
17.1 (316) 23.1 (156)
Men Women Age ~25 25-29 30-34 235
CO.05 8.8 (80) 18.6 (161) 17.6 (136) 30.5 (95)
History of i.v. drug use <5
5-9
210 Site of recruitment
<0.001
Treatment centres Storefront units Streets Infectious disease clinic
5.8 20.1 13.0 55.6
(156) (184) (69) (63)
5.6 12.6 17.5 41.8
(89) (182) (91) (110)
Borrowing of syringes/needles
Never l-20 times 21-100 >lOO
Borrowing from:
Good friends/acquaintances Other persons
only
12.2 (147) 27.1 (236)
i.v. drug use in prison
< 0.001
No Yes Borrowing of syringes/needles in prison
10.1 (298) 34.5 (174)
Never t-50 times >50 Sex with HIV+ partners
10.5 (333) 32.6 (95) 54.5 (44)
No Yes
P-value
CO.001 13.7 (387) 43.5 (85)
78
K. Stark, R. Miiller / Forensic Sci. ht. 62 (1993)
73-81
3.3. Determinants of HIV infection
The overall HIV seroprevalence was 19.1% (90/472). In bivariate analysis, HIV infection is associated with age, history of iv. drug use, site of recruitment, frequency of borrowing syringes/needles, frequency of borrowing in prison, person from whom syringes/needles were borrowed, and sex with HIV-infected partners (Table 2). Of those who have borrowed injection equipment in prison more than 50 times, 54.5% were HIV-infected. No association was found between HIV infection and sex, nationality, or previous drug treatment. A multivariate analysis was performed to determine the effect of potential confounders. The variables sex and history of i.v. drug use were found to be important confounders. After adjusting for these confounders, HIV infection was most strongly associated with borrowing in prison (adjusted HIV prevalence odds ratio (POR), 3.9; 95% confidence interval (CI), 1.8-8.5; P < 0.01) and with sex with HIVinfected partners (POR, 3.0; 95% CI, 1.5-9; P < 0.01). 3.4. Risk reduction and current risk behaviour Of all IDU, 67.8% reported having reduced their risk of acquiring HIV infection; 11.2% had not changed their behaviour despite being aware of risks of infection, 8.5% had already avoided risk behaviour before they became aware of the HIV epidemic among IDU, and no further information was available from 12.5% without risk reduction. The patterns of risk reduction are shown in Table 3. Although a majority have reduced risks of HIV infection, continued risk behaviour occurred in significant numbers of IDU in the study. Borrowing of syringes/needles in the 6 months preceding the interview was reported by 47.9% of the participants. Only 12.9% reported to have always used condoms with casual or steady sexual partners (excluding prostitution) in the previous 6 months. For associations of current risk behaviour and different variables, see Table 4. Borrowing of syringes/needles was associated with younger age, a shorter history of i.v. drug use, site of recruitment, previous drug treatment, no current participation in methadone maintenance, and negative HIV serostatus. Table 3 Behaviour changes in IDU under the influence of aids (n = 472) Variable
Prevalence (%)
Reduced frequency of injecting Avoidance of borrowing syringes Reduced borrowing of syringes Borrowing only from well-known persons Borrowing from one partner only Effective cleansing of borrowed syringes always more often Reduced number of sexual partners Use of condoms always more often
2.3 27.3 30.5 23.5 6.4 11.4 6.6 16.7 17.8 14.2
K. Stark, R. Miiller / Forensic Sci. Int. 62 (1993)
Table 4 Association Variable
of current
risk behaviour
(n)
(previous
6 months)
Borrowing of injection equipment (prevalence
79
73-81
in YII)
and different
variables
Condom use in max. SO”% of sexual contacts (prevalence in “/u)
Sex Men (316)
49.4
Women
45.5
83.9 86.5
56.3* 53.4
90.0 81.4
230 (231) History of i. Y. drug use < 5 years (I 29) 5-9 (131)
38.5
82.3
58.9*** 51.9
> IO (212) Site of recruitment Treatment centres (156) Storefront units (184)
36.8
86.0 89.3 80.7
58.3** 41.3 43.5
87.2 85.9 15.4
(63) Previous drug treatment
39.1
84.1
Yes (335) No (137) Methadone maintenance Yes (113)
49.9* 39.4
86.0 81.8
34.5*** 51.0
78.8 81.2
31.1*** 51.3
73.3* 86.6
(156)
Age <25 years (80) 25-29 (161)
Streets (69) Infectious disease clinic
No (359) HIV serostatus Positive (90) Negative (382) P-values:
*<0.05,
**
***
4. Discussion The HIV seroprevalence in IDU in treatment centres is steadily declining since 1985. One reason is risk reduction in the majority of IDU. Preventive measures which have been intensified in Berlin since 1986 (such as improved availability of sterile injection equipment, extended outreach work, AIDS counselling, improved social and medical support for IDU) have probably facilitated such behaviour changes. However, the decreasing HIV seroprevalence is due to selection bias as well. IDU recruited from the treatment centres in the last 5 years were significantly less likely to have been imprisoned, to have injected drugs in prison, or to report prostitution. In different studies these factors were identified as determinants of HIV infection [4,12,13,16]. To reduce effects of selection in the ongoing cross-sectional study, the participants are recruited by multisite sampling. HIV seroprevalence shows significant differ-
K. Stark, R. Miiller / Forensic Sri. Ini. 62 (1993)
80
73-81
ences by site of recruitment. Compared with clients of treatment centres, IDU recruited from the streets or from storefront units are more likely to be HIV-infected. However, HIV seroprevalence in these subpopulations of IDU in Berlin are lower than in similar studies in 1990/9 1 [ 14,151. Sampling bias or a significant increase in numbers of new drug injectors who avoided HIV infection are possible explanations. The same methods of sampling had been applied in the earlier studies. However, in the meantime many IDU have entered methadone maintenance provided by general practitioners. HIV-infected IDU get into such programmes relatively easily, thus being underrepresented in samples of IDU recruited from the streets or from storefront units. In fact, about 50% of the IDU in methadone maintenance programmes are HIV-infected. Multivariate analysis of our results indicates that injecting drugs with borrowed syringes and needles in prison is a strong predictor of HIV infection. Preventive action for imprisoned IDU is urgently needed. At least the means for effective cleansing of used injection equipment should be provided (e.g., disinfectants such as bleach, facilities to boil water). A majority of about 70% of the IDU reported behaviour changes to reduce risks of HIV infection. As reported in other studies, risk reduction in participants in our study in borrowing syringes was much more common (58%) than increased condom use (32%) [4,8,10]. About 30% of the IDU reported that after having changed behaviour, they have borrowed only from well-known persons or partners. These individuals underestimate risks of HIV infection because they may not be aware of HIV risks or actual HIV infection in these persons. Despite significant risk reduction in the participants in our study, current risk behaviour is common. Of all IDU, 48% have borrowed syringes/needles in the 6 months preceding interview, and only 13% have always used condoms in private sexual relations in the previous 6 months. IDU of younger age and those with a shorter history of i.v. drug use were more likely to have borrowed syringes in the previous 6 months. Intervention strategies targeted on these groups should be intensified. Surprisingly, IDU recruited from sites other than treatment centres and IDU who never have been in treatment were less likely to have recently borrowed injection equipment. It seems possible that some of the IDU practice high risk behaviour in the months before they decide to enter treatment. On the other hand, the lower prevalence of recent sharing may be due to prevention measures in storefront units and by outreach workers. This needs further evaluation. References European Centre for the Epidemiological Quarterly Report No. 36 (31.12.92), Paris,
2 3
AIDS/HIV, Bericht des AIDS-Zentrums des Bundesgesundheitsamtes, Monthly Report, Berlin, 12192. 0. Hamouda, B. Schwartlander, M. Koch, et al.: AIDS/HIV 1992. Bericht zur epidemiologischen Situation in der Bundesrepublik Deutschland zum 31.12.1992, Report of the AIDS-Zentrum, Berlin, 1993. D.C. Des Jarlais, S.R. Friedman, D.M. Novick, et al., International epidemiology of HIV and AIDS among injecting drug users. AIDS, 6 (1992) 1053-1068. CF. Robert, J.J. Deglon, J. Wintsch, et al., Behavioural changes in intravenous drug users in Geneva: rise and fall of HIV infection, 1980-1989. AIDS, 4 (1990) 657-660.
4 5
Monitoring 1992.
of AIDS,
AIDS
Surveillance in Europe.
1
K. Stark, R. Miiller / Forensic Sri. Ini. 62 (1993)
6
8 9 10
11 12 13 14 15
16
73-81
81
N. Romano, F. Vitale, D.R. Alesi, et al., The changing pattern of human immunodeficiency virus type 1 infection in intravenous drug users. Am. J. Epidemiol., 135 (1992) 1189-l 196. R. Zangerle, D. Fuchs, H. Rossler et al., Trends in HIV infection among intravenous drug users in Innsbruck, Austria. J. Acquired Immune Defciceny Syndrome, 5 (1992) 865-871. R.P. Brettle, HIV and harm reduction for injecting drug users. AIDS, 5 (1991) 125-136. R.A. Coutinho. Epidemiology and prevention of AIDS among intravenous drug users. J. Acquired Immune Deficiceny Syndrome, 3 (1990) 413-416. J.A.R. Van den Hoek, H.J.A. Van Haastrecht, R.A. Coutinho et al., Risk reduction among intravenous drug users in Amsterdam under the influence of AIDS. Am. J. Public Health, 79 (1989) 1355-1357. WHO, Proposed WHO criteria for interpreting results from Western Blot assays for HIV-l. HIV-2. and HTLV-I/II. wkly. Epidem. Rec., 65 (1990) 281-283. E.E. Schoenbaum, D. Hartel, P.A. Selwyn et al., Risk factors for human immunodeliciency virus infection in intravenous drug users. N. Engl. J. Med., 321 (1989) 874-879. J.A.R. Van den Hoek, R.A. Coutinho, H.J.A. Van Haastrecht et al., Prevalence and risk factors of HIV infections among drug users and drug-using prostitutes in Amsterdam. AIDS, 2 (1988) 55-60. D. Kleiber, HIV - Needle sharing - Sex, Research Reports SPI, Berlin, 1992. K. Stark, R. Milller, I. Guggenmoos-Holzmann and U. Bienzle, Epidemiologie der HIV-lnfektion bei i.v. Drogenkonsumenten in Berlin. In C. Lange (ed.), AIDS: Eine Forschungsbilan,-, Verlag Sigma, Berlin, 1993, pp. 161-173. K. Stark, R. Miiller, 1. Guggenmoos-Holzmann et al., HIV infection in intravenous in Berlin: risk factors and time trends. Min. Wochenschr., 68 (1990) 415-420.
drug abusers
Frrrensic Science International
HIV prevalence and risk behaviour in injecting drug users in Berlin K. Stark*,
R. Miiller
Landesinstitut fiir Tropenmedizin, Institut ftir Medisinische Siatistik und Informationsverarbeitung. Freie Vniversitiit Berlin, Engeldamm 62-64, 10179 Berlin, Germany
(Accepted I September 1992)
Abstract
In injecting drug users (IDU) newly entering drug treatment centres in Berlin, HIV seroprevalence continuously decreased from 1985 through 1992. A recent cross-sectional study with multi-site sampling (n = 472) showed significantly differing HIV prevalences by site of recruitment (treatment centres 6%, storefront units 20%, infectious disease clinic 56%). In multivariate analysis, the borrowing of syringes in prison was the strongest independent predictor of HIV infection. A majority of the IDU have reduced risks of infection. However, current risk behaviour is common. Of all IDU, 48% reported the borrowing of syringes within the previous 6 months, only 13% had consequently used condoms. Current borrowing of syringes was associated with younger age, shorter history of intravenous drug use, negative HIV serostatus, and non-participation in methadone maintenance programmes. Continued preventive efforts against HIV transmission among IDU are needed. Key words:
HIV infection; Injecting drug users; Risk behaviour; Risk reduction
1. Introduction AIDS and HIV infection are still major problems in injecting drug users (IDU) throughout the world. In Europe, since 1991 IDU represent the largest transmission group among diagnosed AIDS cases. In 1992, 37.4% of the 14 137 AIDS cases were IDU [l]. In Germany, 13.7% of all AIDS cases are IDU (December 1992). From 1983 through 1992, 246 AIDS cases have been diagnosed among IDU in Berlin * Corresponding
author.
0379-0738/93/$06.00 0 1993 SSDI 0379-0738(93)01361-U
Elsevier Scientific Publishers Ireland Ltd. All rights reserved
74
K. Stark, R. Miiller/Forensic Sci. hf. 62 (1993) 73-81
(12.1%). This proportion has been stable in the last years [2,3] To determine the current spread of HIV infection among IDU, seroepidemiological studies are necessary. In many industrialized countries, HIV seroprevalences in IDU have levelled off or even declined in recent years [4-71. IDU have reduced risks of infection mainly by avoiding sharing injection equipment. However, risk behaviour still occurs in substantial proportions of IDU [4,8-lo]. Moreover, the severe HIV epidemics among IDU in Thailand (Bangkok), Myanmar and Poland (Warsaw) again show that under certain circumstances such as lack of information and lack of sterile injection equipment, IDU are very vulnerable to HIV infection [4]. We have been continously monitoring HIV seroprevalence in IDU in drug treatment centres in Berlin since 1985. In an ongoing cross-sectional study with multisite sampling we assess HIV prevalence, determinants of HIV infection, risk behaviour and behaviour changes, sociodemographic variables, health problems and utilization of medical services. In this paper we present data on the HIV epidemiology in IDU in Berlin. 2. Subjects and methods Since 1985 the Institute of Tropical Medicine offers counselling on HIV infection and anonymous testing for HIV antibodies to clients of several drug treatment centres in Berlin. Throughout the study period, about 85% of all IDU entering treatment underwent the test. Sociodemographic and behavioural data were assessed by standardized questionnaires. Screening for HIV antibodies was done by commercially available ELISA. Repeatedly positive results were confirmed by Western blotting according to the guidelines of WHO [l 11. Overall, 1813 IDU have been tested up to 3 1 December 1992. In 1992, we began a cross-sectional study on the epidemiology of HIV infection and other.parenterally transmitted diseases (hepatitis B, C, D, syphilis) in IDU. Participants were anonymously recruited from treatment centres, storefront units, by street recruitment and from an infectious disease clinic. HIV antibodies were determined in serum or saliva. Participants were interviewed by trained personnel using a standardized questionnaire. For statistical analysis, bivariate (N2-test) and multivariate methods (logistic regression analysis) were used. 3. Results 3.1. Time trends of HIV seroprevalence in IDU in treatment centres, 1985-1992 HIV seroprevalence was 24.9% in IDU tested in 1985. Some of these individuals had already stopped i.v. drug use several years before. Thus, HIV seroprevalence increased from 18.4% in IDU having injected until 1983 to 30.9% in those still injecting in 1985. In the following years it declined continously to 3.2% in 1992 (Fig. 1). However, after adjusting for confounding factors (i.v. drug use in prison, history of i.v. drug use, positive seromarkers to hepatitis B virus) the HIV prevalence odds ratio significantly increased until 1988 (Fig. 2). In the following years, the set of confounders changed. Therefore it was not possible to adjust for the whole study period.
K. Stark, R. Miller/Forensic HIV seroprevalence
Sci. ht. 62 (1993)
73-81
75
(%)
5o 1
0
I
I
51983
I
1984
I
,
1985
1986
I
1987
I
I
1988
1989
I
I
I
1990
1991
1992
year of last drug injection Fig. 1. Time trend of the crude
Adjusted
HIV prevalence
HIV seroprevalence
in IDU in treatment
centres
(n = 1813)
odds ratio
5-
4-
3-
2-
~_~.______________.________‘_____-_ .
0
)
‘1983
I
1984
I
1985
I
1986
I
1987
I
1988
year of last injection Fig. 2. Adjusted HIV prevalence odds ratios (with 95”/0 confidence intervals) in IDU from treatment centres. Figures are adjusted for i.v. drug use in prison, history of iv. drug use, and hepatitis B serostatus.
K. Stark, R. Miiller/ Forensic Sri. ht. 62 (1993)
76
73-81
A significant decrease over time was noticed in the prevalence of sexual contacts with HIV-infected partners, imprisonment, i.v. drug use in prison, and prostitution. Other variables, such as lifetime needle sharing and sexual contacts with i.v. drug users, did not show significant changes throughout the study period. 3.2. Results of the cross-sectional study, 1992/93 Data are presented from 472 IDU examined in the ongoing cross-sectional study. The mean age of the participants was 29.8 years (S.D., 5.8; median, 29) and the mean
Table 1 Basic characteristics of the study population (n = 472) Variable
Prevalence (“h)
Sex
Men Women
66.9 33.1
Nationality
German Other
93.2 6.8
Age <25 years 25-29 30-34 235
16.9 34.1 28.8 20.1
History of i.v. drug use <5 years
5-9 210
27.3 27.8 44.9
Site of recruitment
Treatment centres Storefront units Streets Infectious disease clinic
33.1 39.0 14.6 13.3
Injected drugs
Heroin Cocaine Benzodiazepines Barbiturates Amphetamines
98.3 12.5 51.1 23.1 13.8
Previous imprisonment
Men Women
16.3 53.8
i. v. drug use in prison
Men Women
44.9 21.2
Prostitution
Men Women Previous drug treatment
12.7 73.1 71.0
K. Stark, R. Miiller /Forensic
Sci. Int. 62 (1993)
II
73-81
duration of i.v. drug use was 9.6 years (S.D., 6.5; median 8). Table 1 shows basic characteristics of the study population. Sites of recruitment were treatment centres in 33. lo!, storefront units in 39.0%, streets in 14.6%, and an infectious disease clinic in 13.3% of the participants. The main drugs injected were heroin, cocaine, and benzodiazepines (e.g., flunitrazepam, diazepam). The majority of the individuals (79%) injected more than once daily. Of all subjects, 69% have been imprisoned (men, 76.3%; women, 53.8%). Injecting in prison was reported by 44.9% of the men, and 21.2% of the women, respectively.
Table 2 Association of HIV infection and different variables HIV seroprevalence % out of (n)
Variable
n.s.
Sex
17.1 (316) 23.1 (156)
Men Women Age ~25 25-29 30-34 235
CO.05 8.8 (80) 18.6 (161) 17.6 (136) 30.5 (95)
History of i.v. drug use <5
5-9
210 Site of recruitment
<0.001
Treatment centres Storefront units Streets Infectious disease clinic
5.8 20.1 13.0 55.6
(156) (184) (69) (63)
5.6 12.6 17.5 41.8
(89) (182) (91) (110)
Borrowing of syringes/needles
Never l-20 times 21-100 >lOO
Borrowing from:
Good friends/acquaintances Other persons
only
12.2 (147) 27.1 (236)
i.v. drug use in prison
< 0.001
No Yes Borrowing of syringes/needles in prison
10.1 (298) 34.5 (174)
Never t-50 times >50 Sex with HIV+ partners
10.5 (333) 32.6 (95) 54.5 (44)
No Yes
P-value
CO.001 13.7 (387) 43.5 (85)
78
K. Stark, R. Miiller / Forensic Sci. ht. 62 (1993)
73-81
3.3. Determinants of HIV infection
The overall HIV seroprevalence was 19.1% (90/472). In bivariate analysis, HIV infection is associated with age, history of iv. drug use, site of recruitment, frequency of borrowing syringes/needles, frequency of borrowing in prison, person from whom syringes/needles were borrowed, and sex with HIV-infected partners (Table 2). Of those who have borrowed injection equipment in prison more than 50 times, 54.5% were HIV-infected. No association was found between HIV infection and sex, nationality, or previous drug treatment. A multivariate analysis was performed to determine the effect of potential confounders. The variables sex and history of i.v. drug use were found to be important confounders. After adjusting for these confounders, HIV infection was most strongly associated with borrowing in prison (adjusted HIV prevalence odds ratio (POR), 3.9; 95% confidence interval (CI), 1.8-8.5; P < 0.01) and with sex with HIVinfected partners (POR, 3.0; 95% CI, 1.5-9; P < 0.01). 3.4. Risk reduction and current risk behaviour Of all IDU, 67.8% reported having reduced their risk of acquiring HIV infection; 11.2% had not changed their behaviour despite being aware of risks of infection, 8.5% had already avoided risk behaviour before they became aware of the HIV epidemic among IDU, and no further information was available from 12.5% without risk reduction. The patterns of risk reduction are shown in Table 3. Although a majority have reduced risks of HIV infection, continued risk behaviour occurred in significant numbers of IDU in the study. Borrowing of syringes/needles in the 6 months preceding the interview was reported by 47.9% of the participants. Only 12.9% reported to have always used condoms with casual or steady sexual partners (excluding prostitution) in the previous 6 months. For associations of current risk behaviour and different variables, see Table 4. Borrowing of syringes/needles was associated with younger age, a shorter history of i.v. drug use, site of recruitment, previous drug treatment, no current participation in methadone maintenance, and negative HIV serostatus. Table 3 Behaviour changes in IDU under the influence of aids (n = 472) Variable
Prevalence (%)
Reduced frequency of injecting Avoidance of borrowing syringes Reduced borrowing of syringes Borrowing only from well-known persons Borrowing from one partner only Effective cleansing of borrowed syringes always more often Reduced number of sexual partners Use of condoms always more often
2.3 27.3 30.5 23.5 6.4 11.4 6.6 16.7 17.8 14.2
K. Stark, R. Miiller / Forensic Sci. Int. 62 (1993)
Table 4 Association Variable
of current
risk behaviour
(n)
(previous
6 months)
Borrowing of injection equipment (prevalence
79
73-81
in YII)
and different
variables
Condom use in max. SO”% of sexual contacts (prevalence in “/u)
Sex Men (316)
49.4
Women
45.5
83.9 86.5
56.3* 53.4
90.0 81.4
230 (231) History of i. Y. drug use < 5 years (I 29) 5-9 (131)
38.5
82.3
58.9*** 51.9
> IO (212) Site of recruitment Treatment centres (156) Storefront units (184)
36.8
86.0 89.3 80.7
58.3** 41.3 43.5
87.2 85.9 15.4
(63) Previous drug treatment
39.1
84.1
Yes (335) No (137) Methadone maintenance Yes (113)
49.9* 39.4
86.0 81.8
34.5*** 51.0
78.8 81.2
31.1*** 51.3
73.3* 86.6
(156)
Age <25 years (80) 25-29 (161)
Streets (69) Infectious disease clinic
No (359) HIV serostatus Positive (90) Negative (382) P-values:
*<0.05,
**
***
4. Discussion The HIV seroprevalence in IDU in treatment centres is steadily declining since 1985. One reason is risk reduction in the majority of IDU. Preventive measures which have been intensified in Berlin since 1986 (such as improved availability of sterile injection equipment, extended outreach work, AIDS counselling, improved social and medical support for IDU) have probably facilitated such behaviour changes. However, the decreasing HIV seroprevalence is due to selection bias as well. IDU recruited from the treatment centres in the last 5 years were significantly less likely to have been imprisoned, to have injected drugs in prison, or to report prostitution. In different studies these factors were identified as determinants of HIV infection [4,12,13,16]. To reduce effects of selection in the ongoing cross-sectional study, the participants are recruited by multisite sampling. HIV seroprevalence shows significant differ-
K. Stark, R. Miiller / Forensic Sri. Ini. 62 (1993)
80
73-81
ences by site of recruitment. Compared with clients of treatment centres, IDU recruited from the streets or from storefront units are more likely to be HIV-infected. However, HIV seroprevalence in these subpopulations of IDU in Berlin are lower than in similar studies in 1990/9 1 [ 14,151. Sampling bias or a significant increase in numbers of new drug injectors who avoided HIV infection are possible explanations. The same methods of sampling had been applied in the earlier studies. However, in the meantime many IDU have entered methadone maintenance provided by general practitioners. HIV-infected IDU get into such programmes relatively easily, thus being underrepresented in samples of IDU recruited from the streets or from storefront units. In fact, about 50% of the IDU in methadone maintenance programmes are HIV-infected. Multivariate analysis of our results indicates that injecting drugs with borrowed syringes and needles in prison is a strong predictor of HIV infection. Preventive action for imprisoned IDU is urgently needed. At least the means for effective cleansing of used injection equipment should be provided (e.g., disinfectants such as bleach, facilities to boil water). A majority of about 70% of the IDU reported behaviour changes to reduce risks of HIV infection. As reported in other studies, risk reduction in participants in our study in borrowing syringes was much more common (58%) than increased condom use (32%) [4,8,10]. About 30% of the IDU reported that after having changed behaviour, they have borrowed only from well-known persons or partners. These individuals underestimate risks of HIV infection because they may not be aware of HIV risks or actual HIV infection in these persons. Despite significant risk reduction in the participants in our study, current risk behaviour is common. Of all IDU, 48% have borrowed syringes/needles in the 6 months preceding interview, and only 13% have always used condoms in private sexual relations in the previous 6 months. IDU of younger age and those with a shorter history of i.v. drug use were more likely to have borrowed syringes in the previous 6 months. Intervention strategies targeted on these groups should be intensified. Surprisingly, IDU recruited from sites other than treatment centres and IDU who never have been in treatment were less likely to have recently borrowed injection equipment. It seems possible that some of the IDU practice high risk behaviour in the months before they decide to enter treatment. On the other hand, the lower prevalence of recent sharing may be due to prevention measures in storefront units and by outreach workers. This needs further evaluation. References European Centre for the Epidemiological Quarterly Report No. 36 (31.12.92), Paris,
2 3
AIDS/HIV, Bericht des AIDS-Zentrums des Bundesgesundheitsamtes, Monthly Report, Berlin, 12192. 0. Hamouda, B. Schwartlander, M. Koch, et al.: AIDS/HIV 1992. Bericht zur epidemiologischen Situation in der Bundesrepublik Deutschland zum 31.12.1992, Report of the AIDS-Zentrum, Berlin, 1993. D.C. Des Jarlais, S.R. Friedman, D.M. Novick, et al., International epidemiology of HIV and AIDS among injecting drug users. AIDS, 6 (1992) 1053-1068. CF. Robert, J.J. Deglon, J. Wintsch, et al., Behavioural changes in intravenous drug users in Geneva: rise and fall of HIV infection, 1980-1989. AIDS, 4 (1990) 657-660.
4 5
Monitoring 1992.
of AIDS,
AIDS
Surveillance in Europe.
1
K. Stark, R. Miiller / Forensic Sri. Ini. 62 (1993)
6
8 9 10
11 12 13 14 15
16
73-81
81
N. Romano, F. Vitale, D.R. Alesi, et al., The changing pattern of human immunodeficiency virus type 1 infection in intravenous drug users. Am. J. Epidemiol., 135 (1992) 1189-l 196. R. Zangerle, D. Fuchs, H. Rossler et al., Trends in HIV infection among intravenous drug users in Innsbruck, Austria. J. Acquired Immune Defciceny Syndrome, 5 (1992) 865-871. R.P. Brettle, HIV and harm reduction for injecting drug users. AIDS, 5 (1991) 125-136. R.A. Coutinho. Epidemiology and prevention of AIDS among intravenous drug users. J. Acquired Immune Deficiceny Syndrome, 3 (1990) 413-416. J.A.R. Van den Hoek, H.J.A. Van Haastrecht, R.A. Coutinho et al., Risk reduction among intravenous drug users in Amsterdam under the influence of AIDS. Am. J. Public Health, 79 (1989) 1355-1357. WHO, Proposed WHO criteria for interpreting results from Western Blot assays for HIV-l. HIV-2. and HTLV-I/II. wkly. Epidem. Rec., 65 (1990) 281-283. E.E. Schoenbaum, D. Hartel, P.A. Selwyn et al., Risk factors for human immunodeliciency virus infection in intravenous drug users. N. Engl. J. Med., 321 (1989) 874-879. J.A.R. Van den Hoek, R.A. Coutinho, H.J.A. Van Haastrecht et al., Prevalence and risk factors of HIV infections among drug users and drug-using prostitutes in Amsterdam. AIDS, 2 (1988) 55-60. D. Kleiber, HIV - Needle sharing - Sex, Research Reports SPI, Berlin, 1992. K. Stark, R. Milller, I. Guggenmoos-Holzmann and U. Bienzle, Epidemiologie der HIV-lnfektion bei i.v. Drogenkonsumenten in Berlin. In C. Lange (ed.), AIDS: Eine Forschungsbilan,-, Verlag Sigma, Berlin, 1993, pp. 161-173. K. Stark, R. Miiller, 1. Guggenmoos-Holzmann et al., HIV infection in intravenous in Berlin: risk factors and time trends. Min. Wochenschr., 68 (1990) 415-420.
drug abusers