International studies suggest that methadone treatment can reduce the spread of HIV among heroin-dependent drug users[1-9]. The type of treatment programmes in these studies, however, were predominantly detoxification clinics or methadone maintenance programmes with strict entry criteria. Because of these strict regimes, only a small (and probably selected) group of drug users could benefit from these methadone programmes.
In Amsterdam, The Netherlands, methadone programmes are implemented according to the harm-reduction approach. The main goal is to keep in contact with as many drug users as possible, therefore methadone is prescribed in different types of programmes in combination with social medical care and needle-exchange facilities[11,12]. With this approach, an estimated 60-70% of all drug users in Amsterdam are reached and the impact on the spread of HIV infection could be high. The aim of the present study was to evaluate methadone maintenance treatment in Amsterdam in relation to the incidence of HIV.
In December 1985, an open and ongoing cohort study among drug users was started in Amsterdam[13-15]. Participants were mainly recruited from local methadone outposts and a sexually transmitted disease clinic for addicted prostitutes and they were asked to return every 4 months for follow-up visits. At all visits, blood samples were taken and a standardized questionnaire was administered. Questions about current behaviour referred to the period between the current and previous visit. Blood specimens were tested for HIV by enzyme-linked immunosorbent assay; positive specimens were confirmed by immunoblotting.
To ensure detailed and valid assessment of methadone treatment, data from the Central Methadone Register [CMR, which covers all methadone prescriptions in Amsterdam on a daily basis (until 1989 weekly)] were linked to data of the cohort. Methadone data were available for 1058 participants, between 1985 and 1996.
Of these 1058 participants, 795 were HIV-negative at cohort entry and 621 had at least one follow-up visit. To increase the precision of the methadone variables and minimize recall bias for behavioural variables only visits with time intervals between the present and preceding cohort visit of ≤?6 months were included in the analysis, leaving 6234 visits of 582 cohort participants (total number of person-years, 1906).
The methadone prescriptions were summarized over the time intervals between the cohort visits in the following variables (for categories see Table 1): mean methadone dosage on the weeks it was prescribed, frequency of programme attendance (percentage of weeks with methadone prescriptions) and main site of prescription. Individual changes in methadone dispensing between two visits were also studied, but these were only recorded when participants had received methadone treatment on two subsequent visits. The variables were change in mean methadone dosage, change in frequency of programme attendance and change in main site of prescription.
General characteristics and (drug use/sexual) risk behaviour variables were used to adjust for possible confounding. General characteristics included sex, calendar year, nationality, ethnic group, homelessness, age and in-patient hospital care in the period preceding the time interval over which the other variables were measured. Drug use variables included current injecting, duration of injecting drug use, current barbiturate and/or tranquilliser use and units of alcohol per day. Sexual behaviour included type of steady partner, HIV status of steady partner and current prostitution. Among current injectors, the frequency of injecting, main type of drug injected, injecting mainly at home, main injecting partner, borrowing needles used by somebody else and the percentage of new needles obtained via needle-exchange programme were also included.
Due to the time window between the moment of infection and the occurrence of antibodies, the most probable period of infection was defined using the results of polymerase chain reaction (PCR), p24 antigen tests  and HIV-RNA data. The time interval between the last negative and first positive PCR, p24 antigen or HIV-RNA visit was defined as the period of infection. For nine seroconverters there were no PCR, p24 antigen or HIV-RNA data and so the time interval between the last HIV-antibody negative and first HIV-antibody positive visit was defined as the period of infection for this group.
All variables were treated as time-dependent. Poisson regression analysis was used to model incidence of HIV and to calculate the 95% confidence intervals (CI). Multivariate models were built using stepwise techniques, variables with a P-value <0.10 were considered as potential independent determinants. A P-value <0.05 was considered statistically significant.
Sixty per cent of the study group (n=582) were male and 76% was Dutch. At intake the mean age was 31 years (SD, 7; range 16-52). Seventy-five per cent had a history of injecting (mean duration 9.6 years, SD, 6.3; range <1 to 28). The total follow-up time was 1906 years, 58 study participants seroconverted during follow-up. The incidence of HIV was 6.0 per 100 person-years among current injectors and 0.2 per 100 person-years among non-injectors (overall 3.0 per 100 person-years).
For drug users who received methadone dosages up to 80mg/day the incidence of HIV increased with increasing methadone dosage (Table 1). A mean dosage greater than 80mg/day was associated with a two-times lower incidence of HIV compared with dosages of 40-60mg/day, but this difference was not statistically significant. For change in methadone dosage a dose-effect relationship was observed: the trend test was statistically significant [relative risk (RR), 0.8; 95% CI, 0.6-1.0; P likelihood ratio statistic=0.04]. Compared with drug users who did not interrupt the programme, those with a low frequency of programme attendance were twice as likely to become infected with HIV (borderline significant). An increase in the frequency of programme attendance was significantly associated with a higher incidence of HIV. The incidence of HIV for a decrease in the frequency of programme attendance was also elevated, but this difference was not statistically significant. Drug users who received methadone mainly via the outpatient clinic for prostitutes and foreigners and drug users who did not have a main site of methadone prescription or who received methadone mainly in police station or prison (category ‚other‚), had a significantly higher incidence of HIV (compared with clients of the methadone outposts). Independent predictors were change in frequency of programme attendance (RR increase, 2.4; 95% CI, 1.2-4.6; RR decrease, 1.5; 95% CI, 0.7-3.2; compared with no change) and change in mean methadone dosage (RR, 0.8; 95% CI, 0.6-1.0; per category). Adjustment for the other methadone variables did not change these RR values.
As current injecting is the most important risk factor for HIV infection and related to the different methadone treatment modalities, the univariate results of Table 1 were adjusted for current injecting (data not shown). For methadone dosages up to 40 mg/day the RRs approached unity, indicating that current injectors more often receive higher dosages. The RR for receiving methadone via the outpatient clinic for prostitutes and foreigners decreased from 2.1 to 1.7, indicating a higher number of current injectors among clients of this outpatient clinic. The RR for a low frequency of programme attendance increased somewhat, from 2.1 to 2.6, and became statistically significant. This indicates a relatively high number of non-injectors among those with low frequency of programme attendance. The other RRs did not change substantially.
Because of the low number of seroconverters among the non-injectors, the possible interaction between current injecting and the methadone treatment modalities could not be studied. Therefore we repeated the analyses selected on current injecting. The results were comparable with the models adjusted for current injecting.
Of the general characteristics and drug use/sexual risk behaviour variables, homelessness, current injecting and in-patient hospital care were independent risk factors for HIV infection (Table 2).
To examine potential confounding of general characteristics and current risk behaviours on the association between methadone maintenance treatment and incidence of HIV, all significant univariate risk factors were added to the multivariate methadone treatment model containing change in frequency of programme attendance and change in methadone dosage. For an increasing frequency of programme attendance the RR increased from 2.4 to 3.2 and for a decreasing frequency from 1.5 to 2.5 (P<0.05); the RR of change in methadone dosage did not change. Adjustment for the non-significant variables did not change the results substantially.
Among Amsterdam drug users with a history of methadone treatment, no statistically significant protective effect of high methadone dosages in itself could be demonstrated, although the effect was in the expected direction[4,18-20]. Apart from low statistical power as high dosages were not frequently prescribed, this might be explained by self-selection bias. In the Amsterdam harm reduction approach, the level of the dosage is a personal choice and probably correlated with risk behaviour. For example, a previous study in our cohort indicated that more severely addicted drug users, including active injectors, received higher methadone dosages. Self-selection bias is less present when changes within an individual are examined; we found that an increasing methadone dosage was significantly and independently associated with lower incidence of HIV.
The causal mechanism between methadone and HIV infection is that high methadone dosages may lead to cessation of injecting, resulting in a lower incidence of HIV. In this study, the RR for receiving more than 80mg methadone per day did not change substantially after adjustment for current injecting. This indicates that the low incidence of HIV among non-injectors could probably not be explained by the prescription of high methadone dosages.
Changes in the methadone prescription regime (decrease in methadone dosage and increase in frequency of programme attendance) were associated with an increased risk for HIV. In general, the methadone dosage is reduced to reach detoxification, on request of the client or involuntarily during imprisonment. It may be that, when the attempt to detoxify fails, or when released from prison there is an increased risk for unsafe behaviour. The urge to inject after a period of non-injecting is demonstrated by a study of Van Haastrecht et al:; relapse to injecting on the very day of release from prison occurred in one of every three released intravenous drug users. Since we observed a dose-effect relation for change in methadone dosage, an increase in methadone dosage (which is the active policy) can be seen as protective for incidence of HIV infection. Continuous retention in treatment, which, among other things, can be achieved by relatively high, adequate methadone dosages [20,22] and the prevention of relapse to injecting after a period of non-injecting (e.g. continuation of the methadone maintenance treatment during imprisonment) are therefore important goals in preventing the spread of HIV among drug users.
Receiving methadone via the outpatient clinic for addicted prostitutes and for foreigners and current prostitutes were associated with a higher incidence of HIV, which was explained by a high level of current injecting in these groups. Another study in our cohort indicated that drug-using prostitutes cease injecting less often, compared with injecting drug users without commercial partners (MW Langendam, GHA van Brussel, RA Coutinho et al, manuscript submitted). Since the methadone dosage is already relatively high among the clients of this outpatient clinic, additional prevention measures should be developed for this subgroup of drug users, also because of the potential of spread of HIV to the general population.
In conclusion, among drug users who receive methadone maintenance treatment in a harm-reduction setting, with ancillary services such as needle-exchange programmes and HIV testing and counselling, the prescription of high methadone dosages is not sufficient to stop the spread of HIV. However, the present results indicate that an individual increase of the dosage and measures to achieve high treatment retention could contribute to HIV prevention.
The authors wish to thank H. Reurs for data management and M. Prins for critically reading the manuscript.
1. Drucker E, Lurie P, Wodak A, Alcabes,
P. Measuring harm reduction: the effects of needle and syringe exchange programs and methadone maintenance on the ecology of HIV.AIDS
1998, 12 (suppl A):
2. Yancovitz SR, Des Jarlais DC, Peskoe N, et al
. A randomized trial of an interim methadone maintenance clinic.Am J Public Health
3. Stark K, Müller R, Bienzle U, Guggenmoos-Holzmann I. Methadone maintenance treatment and HIV risk-taking behavior among injecting drug users in Berlin.J Epidemiol Comm Health
4. Ball JC, Lange WR, Myers CP, Friedman SR. Reducing the risk of AIDS through methadone maintenance treatment.J Health Social Behav
5. Baker A, Kochan N, Dixon J, Wodak A, Heather N. HIV risk-taking behavior among injecting drug users currently, previously and never enrolled in methadone treatment.Addiction
6. Serpelloni G, Carrieri MP, Rezza G, et al.Methadone treatment as a determinant of HIV risk reduction among injecting drug users: a nested case-control study.AIDS Care
7. Williams AB, McNelly EA, Williams AE, et al
. Methadone maintenance treatment and HIV type 1 seroconversion among injecting drug users. AIDS Care
8. 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
9. Metzger DS, Woody GE, McLellan AT, et al
. Human immunodeficiency virus seroconversion among intravenous drug users in- and out-of-treatment: an 18-month prospective follow-up.J Acquir Immune Defic Syndr
10. Buning EC, Coutinho RA, van Brussel GHA, et al
. Preventing AIDS in drug addicts in Amsterdam [letter].Lancet
11. Plomp HN, van der Hek H, Adèr HJ. The Amsterdam Methadone Dispensing Circuit: genesis and effectiveness of a public health model for local drug policy.Addiction
12. Langendam MW, van Haastrecht HJA, van Brussel GHA, et al
. Differentiation in the Amsterdam methadone dispensing circuit: determinants of methadone dosage and site of methadone prescription. Addiction
13. van den Hoek JAR, Coutinho RA, van Haastrecht HJA, et al.Prevalence and risk factors of HIV infections among drug users and drug-using prostitutes in Amsterdam.AIDS
14. van den Hoek JAR, van Haastrecht HJA, Coutinho RA. Risk reduction among intravenous drug users in Amsterdam under the influence of AIDS. Am J Public Health
15. van Haastrecht HJA, van den Hoek JAR, Bardoux C, Leentvaar-Kuijpers A, Coutinho RA. The course of the HIV epidemic among intravenous drug users in Amsterdam, The Netherlands.Am J Public Health
16. Horsburgh CR, Yih OU C, Jason J, et al
. Duration of human immunodeficiency virus infection before detection of antibody. Lancet
17. Bruisten SM, Mientjes GHC, van Delft P, et al
. HIV-1 infection detected by polymerase chain reaction frequently precedes antibody seroconversion in drug users [letter].AIDS
18. Hartel DM, Schoenbaum EE, Selwyn PA, et al
. Heroin use during methadone maintenance treatment: the importance of methadone dose and cocaine use.Am J Public Health
19. Caplehorn JRM, Bell J, Kleinbaum DG, Gebski VJ. Methadone dose and heroin during maintenance treatment. Addiction
20. Bell J, Chan J, Kuk A. Investigation the influence of treatment philosophy on outcome of methadone maintenance. Addiction
21. van Haastrecht HJA, Bax J, van den Hoek JAR. High rates of drug use, but low rates of HIV risk behaviours among injecting drug users during incarceration in Dutch prisons.Addiction
22. Cooper JR. Methadone treatment and acquired immunodeficiency syndrome.JAMA