HIV Prevalence and Risk Behavior among Injecting Drug Users Who Participate in "Low-Threshold" Methadone Programs in Amsterdam
Christina Hartgers, MS, Anneke (JAR) van den Hoek, MD, PhD, Pieta Krijnen MS, and Roel A. Coutinho, MD, PhD
American Journal of Public Health, April 1992, vol 82 no 4
The authors are with the Municipal Health Ser-vice, Department of Public Health and Environment, Amsterdam, the Netherlands
Methadone treatment for intravenous heroin users can help reduce the risk of infection with the human immunodeficiency virus (HIV). The magnitude of this reduction depends on the ability of meth adone programs to reduce illicit opiate injecting.1-5 The Amsterdam "low-threshold" methadone programs endeavor to take down the barriers that treatment programs and/or methadone maintenance pro grarns6, 7 put up for drug users (DUs) by not prohibiting illicit use of opiates or other drugs. This enables these programs to contact many active DUs who do not want to quit drug use. Low-threshold pro grams do not aim at treating the addiction; instead, they aim primarily at contacting heroin users and regulating or stabilizing their use by preventing withdrawal symp toms. Within the "harm reduction" approach, 8,9 the programs' function is to signal and take care of sociomedical problems in an early phase. Together with social support, this approach should provide a foundation on which initiatives to quit drug use will have a better chance of succeeding. 10 After the onset of the HIV epidemic, the hope arose that low-threshold methadone programs would have a function in controlling the spread of HIV, 9,11 either by informing active DUs about the HIV risk or through the prescription of methadone, which would stabilize addiction and lifestyle and thereby reduce injecting. The present study aims at evaluating the effect of the Amsterdam low-threshold pro grams on HIV risk behavior and on the spread of HIV among injecting drug users (IDUs) participating in an epidemiological study of HIV infection
An estimated 5000 to 6000 DUs were in Amsterdam in each quarter of 1985 to 1989.12 , 13 Among them, approximately 40% had injected in the previous month. The Drugs Department of the Municipal Health Service operates seven low threshold methadone clinics, 1 2 -14 which daily provide on-site primary medical and social care and supply methadone to a yearly population of about 3500 DUs. Nonresidents get methadone only when it is medically indicated or when it is needed to enable the user to return to the city or country of origin. Illicit opiate or cocaine use is not a reason for dismissal from the program. Neither is discontinuation of the program for several days, although a reintake (which entails seeing a doctor) is then re quited. If clients want to quit drug use, they either enter a methadone reduction scheme with urine screening for opiates or are directed to a treatment clinic. Regu lated DUs get methadone prescribed by their general practitioner. Between 1981 and 1984, the flow to general practitioners was approximately as large as the opposite flow. 14 In 1987, the average daily methadone dose in the low-threshold programs was approximately 35 mg (range: 5 to 120 mg). Doses between 60 and 120 mg were pre scribed in about 3% of cases. The average yearly duration of methadone supply for Dutch DUs was 21 weeks.
Study Population and Variables
In 1985 an epidemiological study of HIV infection among DUs15-17 was initiated. This ongoing study involves voluntary, confidential HIV antibody testing and counseling for DUs, combined with an interview by specially trained nurses using a standard demographic/behavioral questionnaire. DUs enroll in the study through a low threshold methadone clinic or through a sexually transmitted diseases clinic for addicted prostitutes. Most DUs entering the study (83%) have at some time injected; 55% have injected in the month before entering. The present study concerns all 386 DUs who had injected drugs at some time (hereafter called "injecting drug users," or IDUs) and who enrolled in the study through low-threshold methadone clinics from December 1985 until March 1989. Interviews and serological tests were completed for all 386 IDUs; 110 (28%) were IUV seropositive. The mean age of the 386 IDUs at entrance was 29.7 years (standard deviation = 5.9, range: 16 to 57); 260 were male (67%) and 126 female (33%); and 311 (81%) were Dutch. Limited data are available on DUs who get methadone on prescription in Amsterdam12,13 but are not available specifically on IDUs. With regard to age, there was no significant difference between the 386 study participants and all DUs who got methadone on prescription in Amsterdam in 1987. The male/female ratio in the study was 2:1, as compared with 3:1 among DUs attending lowthreshold methadone clinics in 1987. Behavior reported over the 6 months preceding the interview is called "current" behavior. Behavior in the period prior to these 6 months as far back as 5 years before the interview is indicated as occurring "in the previous 5 years." Current IDUs who started injecting at least 5 years ago are called "long-term IDUs." The variable frequency of borrowing used needles concerns the full 5 years preceding the interview. Long-term regular methadone users (LTM users) were defined as IDUs who started using methadone at least 5 years preceding the interview and who reported daily methadone use in the 5 years preceding the interview. The non-LTM users were defined as IDUs who started using methadone less than 5 years preceding the interview and IDUs who reported irregular use of methadone during these 5 years. Time of the interview was divided into four consecutive periods of 10 months each: (1) December 1985 to September 1986, (2) October 1986 to July 1987, (3)
August 1987 to May 1988, and (4) June 1988 to March 1989.
Statistical Analysis and Serology Statistics used were X 2 , Student's t, Spearman rank correlation coefficient, Mann-Whitney test, odds ratios (ORs), and 95% confidence intervals (CIs). 1 8 P values (two sided) of less than .05 were taken as significant. Logistic regression models were used to determine the independent contribution of different characteristics in predicting HIV serostatus. All demographic and drug use variables that had a univariate statistically significant association with HIV serostatus were considered for entry into the model. Three injecting variables (first and last occurrence of injecting and frequency of borrowing) were considered crucial and were entered at the first step. At each further step, the variable that contributed most to the prediction of HIV serostatus was selected. To provide estimates of the statistical association between LTM use and HIV infection while controlling for possible confounders, adjusted ORs and CIs were derived from logistic regression coefficients. Serological tests used for HIV testing were enzyme-linked immuno sorbent assays and/or immunoblotting.l6
Predictors of HIV Infection
A model was built (see Methods) that optimally could predict HIV serostatus. First occurrence of injecting, injecting mainly heroin and cocaine in the previous 5 years, nationality, and time living in Amsterdam were significantly associated with HIV infection (see Table 1).
Differences between LTM Users and Non-LTM Users
One hundred and ninety four IDUs (50%) met the criteria for LTM use and 189 (49%) fell in the non-LTM group (3 IDUs had missing data). Table 2 shows the differences between the two groups with regard to duration and regularity of methadone use. Compared with non LTM users on demographic and drug use variables, LTM users are older, more often male, and more often Dutch; they have lived longer in Amsterdam, started injecting longer ago, and inject mainly heroin less often and cocaine, either by itself or together with heroin ("speedball"), more often (see Table 3). LTM users inject daily as often as non-LTM users. Among the 189 nonLTM users, 135 persons (71%) currently use methadone daily and 54 persons (29%) do not (see Table 2). The proportion of current daily injectors is also similar in these subgroups: 49 of 135 (36%) versus 19 of 54 (35%). Frequency of borrowing and reuse of own needles and syringes are similar among LTM and non-LTM users (see Table 3). LTM users have a higher HIV prevalence than non-LTM users: 37% versus 20% (OR = 2.42, CI = 1.54-3.83). There is a positive relation between number of years since the first methadone prescription and being HIV positive (P = .005). The higher seroprevalence of LTM users is found in each of the four consecutive time periods: for LTM users it is 35%, 49%,28%, and 32%, respectively, as compared with 21%, 16%, 24%, and 18% among non-LTM users.
Controlling for Possible Confounding
To control for possible confounding, the variable of LTM use was entered in a logistic regression model together with the variables of age, sex, German or South European nationality, time living in Amsterdam, first occurrence of injection, and last occurrence of injection, with HIV serostatus as a dependent variable. The adjusted OR was 1.60 (CI = 0.93-2.74), indicating a slightly-but not statistically significant-increased risk of being HIV seropositive for LTM users. As we considered duration of injecting an important possible confounder, we explored the relation between LTM use, HIV risk behavior, and HIV serostatus separately in the subgroup of 223 longterm IDUs. Among these IDUs, duration and regularity of methadone use in the 133 LTM users and 90 non-MM users were similar to what was found in the total sample (see Table 2). With regard to demographic and drug use variables, the differ ences (and similarities) between the 133 LTM users and 90 non-LTM users were the same as in the total group, with one exception: the proportion of males among MM users and non-LTM users are the same. As in the total sample, the 133 LTM users are more often HIV positive than the 90 non-LTM users (43% vs 28%; OR = 1.97, CI = 1.10-3.52). Adjusted for age, sex, nationality, and time living in Amsterdam, the OR for LTM use was 1.34 (CI = 0.69-2.63). To exclude possible confounding due to exposure to HIV before coming to Amsterdam, we did a separate analysis among Dutch IDUs who were Amsterdam residents for at least 5 years. In this subgroup of 230IDUs, the 147 LTM users differ from the 83 non-LTM users on four demographic and drug use variables: LTM users are older, started injecting longer ago, currently inject more often, and inject mainly cocaine more often. With regard to current daily injecting, having borrowed at least 10 times, and current borrowing, there are no significant differences between LTM users and non-LTM users. LTM users are again more often HIV positive than non-LTM users (38% vs 21%; OR = 2.39, CI = 1.22-4.71). Adjusted for age, sex, and first and last occurrence of injecting, the OR was 2.06 (CI = 1.02-4.16). HIV infection may have been the reason for long-term participation in lowthreshold methadone programs. To exclude this potential bias, we added the variable daily methadone use to a separate multivariate logistic regression analysis of incident HIV infections.19 The behavior of 31 IDUs, who were HIV-seronegative at entrance into the study but who seroconverted for HIV during follow-up, was compared with that of 202 randomly selected controls, who remained seronegative throughout their follow-up period. In
univariate analysis, daily methadone use, reported over a mean period of 8 months preceding the first visit after seroconversion or the selected control visit, was not "protective" (OR = 0.67, CI = 0.311.47). When the variable daily methadone use was adjusted for the independent risk factors for HIV seroconversion, its OR became 0.81 (CI = 0.35-1.83).
An important finding of the present study is that IDUs participating in lowthreshold programs who report long-term daily use of methadone engage in similar or even more HIV risk behavior than IDUs who use methadone for a shorter time and/or irregularly. Univariately, being HIV seropositive is associated with LTM use and with the number of years since the first methadone prescription. In contrast, studies from other countries have reported that DUs in long-term methadone maintenance have a lower HIV seroprevalence than DUs not in treatment; 20-24 which suggests that these maintenance programs are successful in protecting IDUs from HIV infection. In the present study, the HIV prevalence of LTM users who entered the study before October 1986 (i.e., in the first time period) is 35%. These IDUs started using daily methadone before 1981. Because HIV prevalence among 145 DUs entering lowthreshold programs in 1983 and 1984 was only 34%,~ presumably HIV hardly cir culated among DUs in 1981. This indicates that these LTM users got infected mainly while enrolled in the low-threshold programs.
The present study has some limitstions. First, the short average duration of yearly methadone supply in low-threshold methadone programs (21 weeks) suggests that, in reality, LTM users did not use methadone continuously in the 5 years preceding intake. Due to privacy regulations, it is impossible to validate our data with data from the methadone program records. Thus, it is unknown whether the actualyearly duration of methadone use in the total sample significantly differs from our data. Nevertheless, we hold it possible that the self-reports overestimate the actual duration of methadone use. For the present study, it is assumed that LTM users were supplied with methadone more continuously-and during a much longer time-than non-LTM users.
Second, because no data on IDUs in low-threshold programs are available, the representativeness of the sample is un
known. However, although our sample is not atypical for low-threshold program participants with regard to number of years of methadone user13 .26 or age, males are underrepresented in the present study.
Third, comparing methadone users with non-users, preferably randomly assigned, would have been more appropri ate. However, to find a group of comparable IDUs who have never used methadone is-within the Amsterdam context-impossible: only young or newly starting injectors have never used methadone, and so random assignment was not feasible.
Finally, HIV-seropositive IDUs may have entered the study preferentially-for example, because of health problems. However, all IDUs in the present study entered before March 1989, when zidovu dine or prophylaxis for Pneumocystis car inii pneumonia or tuberculosis was not yet given to DUs in the Netherlands. Thus, DUs seemed to enter the study because of HIV risk through injecting, not because of health problems.
With regard to demographics, LTM users were found to be more often Dutch and to have lived a longer time in Amsterdam than non-LTM users. This is in agreement with the intake rules of lowthreshold programs, which discourage nonresident drug users from participating.
With regard to drug use, current daily injecting is similar for LTM and non-LTM users, as is borrowing of needles and sy ringes. LTM users inject mainly heroin less often and inject cocaine, either by it self or together with heroin, more often. A similar relation was found previously; Chaisson et al.27 reports that methadone treatment reduces heroin injecting more strongly than it does cocaine injecting. Speedball injecting is associated with an increased HIV risk, both in the present sample and in New York City. 28
To summarize, there were no indications that LTM users engaged in less HIV risk behavior than non-LTM users. Furthermore, when we looked at the effect of LTM use on HIV risk, while controlling for confounders, we found no indication of a protective effect of LTM use. The same was true in a separate analysis of incident HIV infectionst19: a protective effect of daily methadone use could not be identified.
These findings suggest that the lowthreshold programs do not have an effect on HIV risk because risk behavior is not reduced. Another possibility is that a risk reduction takes place shortly after the start of methadone prescription but does
not grow over the years. In contrast, in a prospective study among IDUs in US methadone maintenance programs; the prevalence of injecting declined from 81% at the beginning of admission to 63% at the end of the period. After this prevalence declined gradually to 42% in the third year and 29% in the fourth year of participation. However, no such decline is apparent in our sample.
To conclude, our results do not support the view that long-term regular participation-as compared with shortterm and/or irregular participation-in low-threshold methadone programs in Amsterdam is associated with less HIV risk-injecting behavior or with a decreased risk of HIV. The apparent failure of the programs to reduce such behavior may be due to the low average methadone dose level or to the (permitted) irregular attendance. 1 . 2,29 However, raising methadone doses to a level at which little or no illicit opiate use occurs, combined with measures to increase attendance, may lower the number of clients, given that not all DUs wish to stop illicit use. Such measures thus would counteract the possibil ity of providing medical and social care to a large group of DUs, many of whom are HIV seropositives.
It therefore seems necessary to judge carefully the relative merits of the lowthreshold approach versus other kinds of approaches. In any case, we suggest that contacts with IDUs in the low-threshold programs should be used more intensively for HIV prevention efforts and that an in crease in methadone dose level, together with measures to enforce regular attendance, may be required if these programs want to play a role in HIV prevention.
This study was supported by the Netherlands Foundation for Preventive Medicine (grant no. 28-1258). The authors thank the nurses B. Frolich, B. Scheringa-Troost, G. Stienstra, and J. Teeuwissen for interviewing and collecting blood samples; Dr. J. Goudsmit and M. Bakker for performing the laboratory tests; G.H.A. van Brussel and P. Drotman for critical comments on earlier drafts; H.J.A. van Haastrecht for critical continents and data management; and M. ter Pelle and T. Mamanaya for preparing the manuscript.
1. Cooper JR. Methadone treatment and acquired immunodeficiency syndrome. JAMA. 1989;262:1664-1668.
2. Dole VP. Methadone treatment and the acquired inununodeficienry syndrome epidemic. JAMA. 1989;262:1681-1682.
3. Novick DM, Joseph H, Croxson TS, et al Absence of antibody to human immunodeficiency virus in long-term, socially rehabilitated methadone maintenance patients. Arch Intern Med. 1990;150:97-99.
4. Zweben JE, Payte IT. Methadone maintenance in the treatment of opioid dependence, a current perspective. WestJMed 1990;152(specialissue):588-599.
5. Zweben JE, Sorenson 7L. Misunderstandings about methadone. J Psychoactive Drugs. 1988;20:275-281.
6. Dole VP. Methadone maintenance treatment for 25 000 heroin addicts. JAMA. 1971;215:1131-1134.
7. Ball JC, Lunge WR, Myers CP, Friedann SR. Reducing the risk of AIDS through methadone maintenance treatment. J Health Soc Behav. 1988;29:214-226.
8. van Bruasel GHA. Openbare gezondheidszorg en drugproblematiek. T Soc Gezondheidszorg. 1986;64:817-819.
9. Buning EC, Coutinho RA, van Brussel GHA, van Samen GW, van Zadelhoff AW. Preventing AIDS in drug addicts in Amsterdam. Lancet. 1986;1:1435. Letterto the Editor.
10. van Brussel GHA, Butting EC. Methadon per bus in Amsterdam. T Alc Drugs. 1979;5:135-137.
11. Butting EC. AIDS-related interventions among drug users in the Netherlands. Presented at the European Symposium on AIDS and Drug Abuse; September 25-28, 1989; Stockholm, Sweden.
12. BuningEC.DeGG&GDenhetdrugprobleem in cijfers,11, III, N. Amsterdam, The Netherlands: GG&GD; 1987, 1988, 1990.
13. Plump HN, Reijneveld SA. Methadon very strekking aan druggebruikers in Amsterdam, analyse van gegevens van de Centmle Methadon Registmtie over de jaren 1984 tot 1987 Amsterdam, The Netherlands: VU utigeverij; 1991.
14. Driessen FMHM. De methadonverstrekking in Amsterdam, 1981-1984, resultaten van de analyse van de gegevens van de Centrale Methadon Registrotie. Amsterdam,The Netherlands: Vrije Universiteit en GG&GD; 1987.
15. van den Hock JAR, van Haastrecht HJA, Coutinho RA. Risk reduction among intravenous drug users in Amsterdam under the influence of AIDS. Am J Public Health. 1989;79:1355-1357.
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 drugusing prostitutes in Amsterdam. AIDS. 1988;2:55-M.
17. van den Hoek JAR, van Haastrecht HJA, Scheeringa-Troost B, Goudsmit J, Coutinho RA. HIV infection and STD in drug-addicted prostitutes in Amsterdam: potential for heterosexual HIV transmission. Genitourin Med. 1989;65:146-150. Goudsmit J, Couthinho RA.
18. Miettinen O. Estimability and estimation in case-referent studies. Am J Epidemiol. 1976;103:226-235.
19. van Ameijden EIC, van den Hoek JAR, van Haastrecht HIA, Coutinho RA. The hams reduction approach and risk factors for HIV-seroconversion in injecting drug users in Amsterdam, the Netherlands. Am J Epidemiol In press.
20. Novick D, Kreek MJ, Des Jarlais D, et al. Antibody to LAV, the putative agent of AIDS, in parenteral drug abusers and methadone-maintained patients: therapeutic, historical, and ethical aspects. National Institute on Drug Abuse Research Monograph 1986;67:318-320.
21. WattersJK,NewmeyerJA,FeldmanHW, Biemacki P. Street-based AIDS prevention for intravenous drug users in San Francisco: prospects, options, and obstacles. In: Community Epidemiology Work Group Proceedings, Vol II. Rockville, Md: National Institute on Drug Abuse; 1986.
22, Hartel D, Selwyn PA, Schmnbaum EE, Mein RS, Friedland GIL Methadone maintenance treatment and reduced risk of AIDS and AIDS-specific mortality in intravenous drug users. Presented at the Fourth International Conference on AIDS; June 12-16, 1988; Stockholm, Sweden.
23. Blix O, Gronbladh BA. AIDS and I V heroin addicts: the preventive effect of meth- , adone maintenance in Sweden. Presented at the Fourth International Conference on AIDS; June 12-16,1988; Stockholm, Sweden.
24. Williams AB, Vranisan K, Gorter R, Brodie B, Meakin R, Moss A. Methadone maintenance, HIV-1 serostatus and race in injection drug users in San Francisco, California. Presented at the Sixth International Conference on AIDS; June 20-24, 1990; San Francisco, Calif.
25. CoutinhoRA,BusJM,RuitenbergFJ.The epidemiology of LAV/HTLV-III infections in Europe. In: Staquet M, Hemmer R, Baert A, eds. ClfnicalAspectsofAIDSand AIDS-Related Complex. Oxford, England: Oxford University Press; 1986:7-14.
26. Reijneveld SA: Methadonverstrekking in Amsterdam in 1989, jaaroverzicht van de Centrale Methadon Registratie. Amsterdam, the Netherlands: GG&GD; 1990.
27. Chaisson RE, Bacchetti P, Osmond D, et al. Cocaine use and HIV infection in intravenous drug users in San Francisco. JAMA. 1989; 261:561-565. .
28. Novick DM, Trigg HL, Des Jarlais DC, et al. Cocaine injection and ethnicity in paren teral drug users during the early years of the human immunodeficiency virus (HIV) epidemic in New York City. J Med Vbol 1989;29:181-185.
29. Dole VP. Implications of methadone maintenance for theories of narcotic addiction. JAMA. 1988;260:3025-3029.