In the past decades, several countries have gained experience with the prescription of heroin to heroin addicts. This experience goes back to as far as the 1920's, when physicians in the United States and the United Kingdom started to prescribe heroin to morphine addicts for both detoxification and maintenance purposes (Drucker and Vlahov, 1999; Zador, 2001). With the 'heroin epidemic' in the late 1960's, the National Health Service (NHS) drug clinics in the United Kingdom took over the responsibility for prescribing injectable heroin from individual physicians, who were no longer allowed to prescribe heroin to heroin addicts. During the following years, the NHS clinics began to switch the prescribed substance from injectable heroin to injectable methadone and since the mid-1970's to oral methadone (Mitcheson, 1994). During this period of transition (1972-1975), a randomized clinical trial was conducted, comparing the effectiveness of medical prescription of injectable heroin (the former standard treatment) with a medical prescription of oral methadone (the new treatment) (Hartnoll et al., 1980).
The study of Hartnoll, Mitcheson and their colleagues represents the only controlled trial that has been conducted with heroin, prior to the current Dutch trials (Hartnoll et al., 1980). The study population consisted of 96 clients, with a mean age of 23.9 years (range 18-32 years), an average duration of opiate addiction of 5.9 years, and a history of daily heroin injection for at least three months. All participants were strongly committed to continue their drug use. After an extensive intake procedure (during which the patients were not informed about their participation in a trial), the patients were randomly assigned to either treatment with injectable heroin (44 persons), or treatment with oral methadone (52 persons), both prescribed on an outpatient maintenance basis. The prescribed daily heroin doses ranged from 30 to 120 mg, with most prescriptions ranging from 40 to 80 mg, and the daily methadone doses ranged from 10 to 120 mg.
The study produced equivocal results, in that no clear indications were obtained that one treatment produced better overall results than the other. Both treatments showed positive outcomes in certain areas, and negative ones in others. For example, treatment compliance was considerably higher among the patients in the heroin condition than among those in the methadone condition, and they also demonstrated a greater reduction in their illegal heroin use and criminal activities. On the other hand, more patients in the methadone condition had reached abstinence at the one-year follow-up assessment, compared to their heroin condition counterparts. No differences were found with respect to the other areas under investigation, including work performance, housing status, diet or physical complications arising from the patients' drug use, neither from pre- to post-treatment status, nor between the two treatment groups. In addition, no pre-treatment patient factors could be detected, which were clearly indicative for the patients' status after treatment. In addition to these results, most patients in the heroin prescription group could be successfully persuaded after completion of the trial to convert from prescribed injectable heroin to oral methadone, which led the investigators to conclude that a maintenance dose of injectable heroin for a limited period of time does not have adverse consequences on the long-term status of the patients.
Based on the results of this early study, the investigators argued that the (differential) results suggest that the choice for one or the other treatment approach depends on the priorities assigned to the various outcome domains: either helping a minority of heroin addicts to terminate their drug use (by means of methadone maintenance), or helping as many addicts as possible to improve their situation (by means of injectable heroin maintenance). Nevertheless, the findings were used by many to infer that injectable heroin maintenance treatment should be dropped in favor of oral methadone maintenance. Following a heated debate, many of the clinics shifted towards a more interventionist therapeutic approach, and new patients were refused prescribed injectable heroin (Mitcheson, 1994). In 1994, only 1-2% of the estimated 75,000-150,000 heroin users in the United Kingdom had actually received a prescribed supply of any injectable drug, and only a small proportion had received injectable heroin (Strang et al., 1994). Based on a postal survey of community pharmacies in the United Kingdom, heroin accounted for less than 1.7% of all prescriptions in 1995 for treatment of opiate dependence (Strang et al., 1996). More recently, the number of heroin addicts who had received injectable heroin on medical prescription in the year 2000 had dropped to an estimated 400 persons, the majority of whom received their prescriptions from a total of 20 licensed physicians (Metrebian, 2000).
Despite this trend towards decreasing injectable opiate prescriptions, Metrebian and her colleagues (1998), when investigating the prescription of injectable opiates to opiate dependent drug users in the mid 1990's, concluded that the prescription of injectable heroin and methadone is a feasible treatment option for long-term injecting drug users, who had previously tried and failed oral methadone and who were unable or unwilling to give up injecting (Metrebian et al., 1998).

The most recent experience with the prescription of opioids other than oral methadone derives from Switzerland. In 1994, the Swiss started a large-scale study into the effects of prescribing injectable and smokeable heroin, injectable and oral morphine, and injectable and oral methadone, as part of a comprehensive national treatment strategy. Based on the positive results of this study, involving a total of 1,969 opiate dependent drug users who began medical treatment with heroin between January 1994 and December 2000 (Rehm et al., 2001), as well as the support which the prescription practice received in a national referendum on the issue, the medical prescription of heroin continues to be an established treatment at the time of writing this report.
According to the reports produced by the investigators (Uchtenhagen et al., 1996a, 1996b, 1997, 1999; Rehm et al., 2001), and an independent process evaluation conducted by the World Health Organization (WHO, 1999), the results of the Swiss heroin experiment have indeed been positive, with high treatment retention rates (86% after three months, 70% after 12 months, and 50% after 30 months), considerable reductions in the use of illegal drugs (heroin and cocaine) and in the level of the participants' criminal activities, substantial and stable improvements in the domains of physical health, psychological well-being, housing and employment, and a substantial reduction in the number of contacts with drug users and the drug scene in general (Rehm et al., 2001) (see Table 2).

Table 2. Changes in the patients' status in the Swiss heroin study (n = 237)

In addition to these findings, it is important to note that no public order problems were encountered, and no fatalities could be attributed to the study medications used. Finally, the cost-benefit analysis revealed an overall net economic benefit of EURO 35 (approx. 45 Swiss Francs) per patient per day, compared to treatment as usual (mainly methadone maintenance).

Although the Swiss experiment has clearly shown that medical prescription of heroin is feasible, and that methadone with co-prescribed heroin, combined with adequate psychosocial support, can lead to considerable improvements in the medical, psychiatric, and social condition of highly dysfunctional opiate addicts, the study has some important limitations from a scientific point of view, which prevent a conclusive answer to the question whether the medical prescription of heroin itself caused the observed improvements.
First, the results are based on a single group pre-post design, i.e. without a control condition, and hence, without reference to the development in similar patients in non-experimental treatment conditions. Within that design, it should be noted that the presented follow-up data were obtained from only 237 of the 385 patients (62%) who entered the project until March 1995 and remained in the program for at least 18 months (Uchtenhagen et al., 1999; Rehm et al., 2001). Although the Swiss collected data at a later stage on patients in various standard methadone programs, and conducted ad hoc comparisons of the treatment results of these patients with those of the experimental group, these comparisons - which involved statistical matching procedures to control for differences between the experimental and comparison groups - lack the strengths of a randomized design. In addition, it should be noted that the Swiss study did include four rather small randomized trials, in some of which a double-blind procedure was used. Three of these studies were, however, not intended to evaluate the long-term effects of heroin prescription, but to investigate the short-term effects associated with the different experimental substances (intravenous heroin, intravenous methadone, and intravenous morphine) on patient recruitment, treatment retention, treatment compliance, and side-effects. The only randomized trial examining the long-term effects of heroin prescription was a small study conducted in Geneva (Perneger et al., 1998). In this study, 51 treatment-resistant heroin addicts were randomly assigned to either intravenous heroin, combined with additional health and psychosocial services or to some other type of conventional drug treatment, usually oral methadone maintenance. After six months, almost all of the patients using prescribed heroin had stopped their illegal heroin use, and their psychological and social functioning (as measured by the frequency of suicide attempts and the level of criminal activities) was much better than that of the patients in the control group.
Second, it is important to note that no attempts were made in the Swiss study on the long-term effects of medically prescribed heroin to investigate the effects of the prescription of heroin alone, i.e. without mandatory counseling and other psychosocial interventions. Hence, the Swiss experiments do not provide information about the effect of the prescription of heroin per se, but about the effect of a combined package of pharmacological (heroin) and psychosocial interventions, compared to more usual treatment (methadone maintenance without obligatory counseling and less psychosocial support) or no treatment at all. Consequently, it is possible that the benefits of the Swiss heroin maintenance programs could at least be partially attributable to the additional and sometimes mandatory psychosocial interventions (Perneger et al., 1998).
Third, it should be noted that the Swiss experiment with smokeable heroin, in the form of heroin reefers (surrogate cigarettes), failed due to the low bio-availability (10-15%) of the heroin from these cigarettes: due to pyrolytic degradation, an average of 85-90% of the heroin in the cigarette did not enter the body of the patient (Seidenberg and Honegger, 1998).

In summary, the Swiss experiences suggest that the medical prescription of injectable heroin is both feasible, medically safe, and potentially effective in chronic, treatment-resistant heroin addicts, at least under the investigated circumstances, i.e. when the prescription is medically controlled, no take-home heroin is provided, and if psychosocial services are provided (e.g. Wodak, 1998). Important questions, however, remain unanswered, particularly pertaining to the causal attribution of the observed improvements to the experimental treatment, the relative contribution of the prescription of heroin in a combined pharmacological and psychosocial treatment program, and the effects and effectiveness of medically prescribed inhalable heroin. In line with this point of view, a special committee of the World Health Organization (WHO, 1999) concluded in its independent process evaluation, that the Swiss studies "( … ) have shown that it is medically feasible to prescribe intravenous heroin as a maintenance drug, at least under the conditions that prevailed during the studies." (p. 10), but also that "The Swiss studies were not able to examine whether improvements in health status or social functioning in the individuals treated were causally related to heroin prescription per se or a result of the impact of the overall treatment programme." (p. 1).