Skip to main content Accessibility help


  • Access
  • Open access
  • Cited by 45


      • Send article to Kindle

        To send this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the or variations. ‘’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Heroin on trial: Systematic review and meta-analysis of randomised trials of diamorphine-prescribing as treatment for refractory heroin addiction
        Available formats

        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        Heroin on trial: Systematic review and meta-analysis of randomised trials of diamorphine-prescribing as treatment for refractory heroin addiction
        Available formats

        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        Heroin on trial: Systematic review and meta-analysis of randomised trials of diamorphine-prescribing as treatment for refractory heroin addiction
        Available formats
Export citation



Supervised injectable heroin (SIH) treatment has emerged over the past 15 years as an intensive treatment for entrenched heroin users who have not responded to standard treatments such as oral methadone maintenance treatment (MMT) or residential rehabilitation.


To synthesise published findings for treatment with SIH for refractory heroin-dependence through systematic review and meta-analysis, and to examine the political and scientific response to these findings.


Randomised controlled trials (RCTs) of SIH treatment were identified through database searching, and random effects pooled efficacy was estimated for SIH treatment. Methodological quality was assessed according to criteria set out by the Cochrane Collaboration.


Six RCTs met the inclusion criteria for analysis. Across the trials, SIH treatment improved treatment outcome, i.e. greater reduction in the use of illicit ‘street’ heroin in patients receiving SIH treatment compared with control groups (most often receiving MMT).


SIH is found to be an effective way of treating heroin dependence refractory to standard treatment. SIH may be less safe than MMT and therefore requires more clinical attention to manage greater safety issues. This intensive intervention is for a patient population previously considered unresponsive to treatment. Inclusion of this low-volume, high-intensity treatment can now improve the impact of comprehensive healthcare provision.


Joint first authors.

See editorial, pp. 3–4, this issue.

Declaration of interest

J.S. and N.L. have contributed to National Treatment Agency/Department of Health English Guidelines on the role of injectable prescribing in the management of opiate addiction (2003; chaired by J.S.), and J.S. also chaired the broader-scope pan-UK working group when preparing the 2007 ‘Orange Guidelines' for the UK Department of Health, providing guidance on management and treatment of drug dependence and misuse and is chairing the new expert group updating the Department of Health Guidelines (2014). J.S. and his employing organisation have provided consultancy advice and received honoraria, travel and conference support, and consultancy fees from pharmaceutical companies including current and potential future suppliers of diacetylmorphine and methadone (ViroPharma, Martindale, TEVA, Reckitt Benckiser) and have conducted research involving collaboration with the pharmaceutical industry to investigate possible new treatment medications (Martindale, Mundipharma, iGen). J.S., N.M. and N.L. have previously undertaken research study of British heroin policy and have given varied commentaries and contributed to professional and public debate. A.U. has been mandated to document and evaluate the Swiss cohort study on heroin-assisted treatment by the Federal Office of Public Health, resulting in (unpaid) scientific publications and (unpaid) presentations at conferences (expenses reimbursed); expert consultation and project participation for the World Health Organization and United Nations Office on Drugs and Crime on substitution treatment for opiate addiction. W.vdB. is chair of the working group that is currently preparing the Netherlands Interdisciplinary Guideline on Opioid Addiction Treatment. He also was the scientific director of the Central Committee on the Treatment of Heroin Addiction (CCBH), which was responsible for the planning, execution and reporting on the Dutch trial on heroin-assisted treatment. W.vdB. has separately provided consultancy advice and received honoraria, travel and conference support, and consultancy fees from various pharmaceutical companies including current and potential future suppliers of buprenorphine (Reckitt Benckiser), extended-release naltrexone (Alkermes) and nalmefene (Lundbeck). C.H. has contributed to the German guidelines on opioid substitution treatment of the German Medical Association and has provided consultancy advice to the German Ministry of Health on the development of the revisions of the Narcotics Law. C.H. has undertaken research evaluation of the German and European drug policy and given commentaries and contributed to professional and public debate. He has also provided consultancy advice and received honoraria, travel and conference support, and consultancy fees from pharmaceutical companies including current and potential future suppliers of diacetylmorphine and other opioids.

The prescribing of diamorphine (pharmaceutical heroin) as part of treatment for heroin addiction initially appears counter-intuitive. Addiction is increasingly recognised as a chronic relapsing disorder 1 for which effective treatments exist. 2 However, public responses to addiction tend to polarise around addiction as ‘badness’ or addiction as ‘illness’, and counter-intuitive treatments often generate strongly felt passions associated with these differing frames of reference. When debate gets heated, it is particularly important for science to contribute cool-headed analysis.

Fifteen years ago, Bammer et al 3 considered this issue at which time there was only one randomised trial (from the UK during the 1970s) investigating heroin-prescribing in an unsupervised clinical situation 4 and one small randomised trial from Switzerland of a new approach of fully supervised self-administration of the prescribed heroin. 5 This latter approach of entirely supervised administration of every injected dose has become standard clinical practice in a new generation of well-designed and executed trials – a further five randomised clinical trials internationally over the past 15 years (see below).

Diamorphine has been prescribed at different times in treatment of heroin addiction for more than a century, in countries that originally included the USA 6 and has continued throughout the century, to a variable extent, in the UK, 7 but it was not until the work of Uchtenhagen and his colleagues in Switzerland in the 1990s that the approach of supervised injectable heroin (SIH) treatment was properly established. 810

Two features characterise the new approach. First, SIH treatment is not a first-line treatment, but an option for patients who have not responded to standard treatments such as oral methadone maintenance treatment (MMT) or residential rehabilitation. Second, all injectable doses (typically 150–250 mg diamorphine per injection) are taken under direct medical or nursing supervision, thereby ensuring compliance, monitoring, safety and prevention of possible diversion of prescribed diamorphine to the illicit market; this requires the clinics to be open several sessions per day, every day of the year. This model of treatment involves screening and appropriate patient selection, structured induction and monitoring, and a high level of support and interaction with staff – thus significantly different from the ‘public health’ approach of open-access supervised injecting rooms. 11 In contrast, SIH treatment is a high-cost, low-volume specialist intervention.

The Cochrane Collaboration has conducted a systematic review of all heroin-prescribing trials. 12 While the Cochrane review compared ‘SIH treatment plus methadone v. oral MMT’, based on trials of SIH treatment, a considerable portion of the reported comparisons have drawn on the findings from a wider group of heroin prescribing randomised controlled trials (RCTs), including ‘heroin provision of various modality and route of administration’, i.e. supervised and unsupervised prescribing practices and prescribing of both injectable and inhalable heroin.

The aims of this paper are: (i) to undertake a systematic review and meta-analysis of a defined narrow group of randomised trials of SIH prescribing and (ii) to examine the political and scientific response to the published findings.


Search strategy

The review was conducted according to the PRISMA guidelines ( The search strategy targeted studies that reported on the effect of SIH treatment in a range of outcome domains among individuals with heroin-dependence unresponsive to standard treatments. Computer-based internet databases used for this search included MEDLINE (PubMed database), Web of Science and Scopus. There were no language or publication year restrictions. The combinations of keywords used in the database search included ‘addiction’, ‘assisted’, ‘supervised’, ‘dependence’, ‘diacetylmorphine’, ‘diamorphine’, ‘heroin’, ‘maintenance’, ‘prescription’ and ‘treatment’. The initial data searches and screening of irrelevant abstracts were conducted by T.G. Subsequent data checking and searches were overseen by N.M. and J.S. Lead clinicians and/or researchers who have been at the forefront of testing and trialling SIH trials co-authored this paper.

Inclusion criteria and selection of studies

The methodology was designed to collect evidence in a sequential and logical manner. The review has a clear focus on evidence of SIH treatment efficacy as well as allowing a broad scope for learning about the scientific and political response to the published findings. Only studies that had the key search terms in the abstract and also had opiate use, retention in treatment, mortality and side-effects as outcome variables were considered. Thus, methodological papers were excluded. Papers were also excluded if they were assessing the pre-existing unsupervised heroin treatment provision, which focused on policy aspects, which were only reporting profile of trial participants or which were separately reporting on measures of cost-effectiveness, community perspectives and patient satisfaction or longer-term (beyond the trial follow-up period) effects.

Data extraction

Information extracted from each study included the location of the study, author names, year of publication, sample size, groups studied, time to follow-up, outcome measures and effect-size estimates.

Statistical analysis

Mantel–Haenszel random effects pooled risk ratios and corresponding 95% CI for SIH treatment patients v. comparison groups were calculated using Review Manager 5.2 for Windows 7 with fuller (compared with the latest Cochrane review of 2011) outcome data. Heterogeneity between studies was assessed through the I 2 statistic. Lastly, funnel plots were used to assess potential publication bias for the meta-analyses.


A total of 2599 records were identified using the search terms (Fig. 1).

Fig. 1 Selection of studies for meta-analysis.

All papers were in English language. Table 1 summarises the six trials included in the review.

Table 1 Six randomised trials of supervised injectable heroin (SIH) (plus flexible supplementary doses of oral methadone): key features and outcomes

Main paper Country Sample size;

groups studied


Cochrane risk of bias 12 using

five criteria

recommended by the

Cochrane Handbook 18
Perneger et al 5 Switzerland n = 51

SIH (+OM): n = 27

OM, detox, rehab:

n = 24
6 months Random sequence generation

Allocation concealment

Incomplete outcome data

Selective reporting

Blinding (objective outcomes)

Blinding (subjective outcomes)





Retention: SIH: 93% v. OM 92%

Self-reported illicit heroin use: SIH: 22%,

OM: 67% (P = 0.002)

SAEs data not reported
van den Brink

et al 13
The Netherlands Injectable trial:

n = 174

SIH (+OM): n=76

OM: n = 98 (also

SInhH trial, n = 75)
12 months Random sequence generation

Allocation concealment

Incomplete outcome data

Selective reporting

Blinding (objective outcomes)

Blinding (subjective outcomes)





Retention: SIH 72% v. OM 85%

Self-reported 40% improvement in at

least one domain (physical, mental,

social): SIH 56% v. OM 31% (P = 0.002)

SAEs: reported data limited to 11 SAEs

(two definitely or probably and nine

possibly related to injectable heroin)
March et al 14 Spain n = 62

SIH (+OM): n = 31

OM: n = 31
9 months Random sequence generation

Allocation concealment

Incomplete outcome data

Selective reporting

Blinding (objective outcomes)

Blinding (subjective outcomes)





Retention: SIH 74% v. OM 68%

Self-reported illicit heroin use in past

30 days (mean days): SIH = 8.3 v.

OM = 16.9 (P = 0.02)

SAEs: SIH = 7 (two unrelated and five

probably or definitely related to study

drug) v. OM = 7
Haasen et al 15 Germany n = 1015

SIH (+OM): n = 515

OM: n = 500
12 months Random sequence generation

Allocation concealment

Incomplete outcome data

Selective reporting

Blinding (objective outcomes)

Blinding (subjective outcomes)





Retention: SIH 67% v. OM 40%

Improvement in drug use (measured

by either UDS and self-report): SIH 69%,

OM 55% (P<0.001)

Improvement in physical/mental

health: SIH 80%, OM 74% (P = 0.023)

Combined reduced drug use and

improved physical/mental health

(responder): SIH 57% v. OM 45% (P<0.001)

SAEs: SIH = 177 (58 possibly, probably or

definitely related to study drug) v. OM = 15

et al 16
Canada n = 251

SIH (+OM): n = 115

OM: n = 111 (also

SIHM+OM, n = 25)
12 months Random sequence generation

Allocation concealment

Incomplete outcome data

Selective reporting

Blinding (objective outcomes)

Blinding (subjective outcomes)





Retention: SIH 88% v. OM 54% (P<0.001)

Self-reported reduction in illicit drug use

or other illegal activities (improvement of

20% for either domain): SIH = 67%, OM = 48%

(P = 0.004)

SAEs: SIH = 51 v. OM = 18
Strang et al 17 England n = 127

SIH (+OM): n = 43

OOM: n = 42 (also

SIM+OM, n = 42)
6 months Random sequence generation

Allocation concealment

Incomplete outcome data

Selective reporting

Blinding (objective outcomes)

Blinding (subjective outcomes)





Retention: SIH (or other treatment) 88% v.

OOM 69%

Reduction in ‘street’ heroin – 50% or

more negative UDS during weeks 14–26

(responder): SIH 66% v. OOM 19%


SAEs: SIH = 7 (two probably related

to study drug) v. OOM = 9

SAE, serious adverse event; OM, oral methadone; OOM, optimised oral methadone; SIM, supervised injectable methadone; SinhH, supervised inhalable heroin; SIHM, supervised injectable hydromorphone. L, low risk of bias; U, unclear; H, high risk of bias.

In addition to the six main papers from the individual trials, a broader set of papers is available, reporting other data such as secondary SIH treatment outcomes, observational longer-term outcomes, health economic data, family perspectives, community perspectives and patient satisfaction. Alongside the results of a meta-analysis of the effects of SIH treatment, this broader set of papers is outlined, although not integrated in our formal analysis for the reasons listed in Table 2.

Table 2 Thirty papers excluded from this review presented in chronological order (from the most recent to the oldest), country and reason for exclusion

Paper Country Reason for exclusion
1 Byford et al 19 England Outcomes not in the scope of this review (health economics)
2 Groshkova et al 20 England Patients' perspective
3 Verthein et al 21 Germany Study not RCT (longer-term outcomes)
4 Vogel et al 22 Germany Outcomes not in the scope of this review (secondary outcomes)
5 Nosyk et al 23 Canada Outcomes not in the scope of this review (health economics)
6 Marchand et al 24 Canada Patients' perspective
7 Verthein et al 25 Germany Study not RCT (longer-term outcomes)
8 Blanken et al 26 The Netherlands Study not RCT (longer-term outcomes)
9 Blanken et al 27 The Netherlands Patients' perspective
10 Eiroa-Orosa et al 28 Germany Outcomes not in the scope of this review (secondary outcomes)
11 Haasen et al 29 Germany Outcomes not in the scope of this review (secondary outcomes)
12 Karow et al 30 Germany Outcomes not in the scope of this review (secondary outcomes)
13 Lasnier et al 31 Canada Community perspectives
14 Miller et al 32 England Community perspectives
15 Oviedo-Joekes et al 33 Spain Study not RCT (longer-term outcomes)
16 Oviedo-Joekes et al 34 Canada Outcomes not in the scope of this review (secondary outcomes)
17 Oviedo-Joekes et al 35 Canada Outcomes not in the scope of this review (secondary outcomes)
18 Scafer et al 36 Germany Outcomes not in the scope of this review (secondary outcomes)
19 Haasen 37 Germany Outcomes not in the scope of this review (health economics)
20 Perea-Milla et al 38 Spain Outcomes not in the scope of this review (secondary outcomes)
21 Haasen et al 39 Germany Outcomes not in the scope of this review (secondary outcomes)
22 Romo et al 40 Spain Patients' perspective
23 Miller et al 41 England Patients' perspective
24 Verthein et al 42 Germany Study not RCT (longer-term outcomes)
25 Dursteler-Macfarland et al 43 Switzerland Patients' perspective
26 Dijkgraaf et al 44 The Netherlands Outcomes not in the scope of this review (health economics)
27 Rehm et al 45 Switzerland Study not RCT (longer-term outcomes)
28 Guttinger et al 46 Switzerland Study not RCT (longer-term outcomes)
29 Rehm et al 8 Switzerland Study not RCT (longer-term outcomes)
30 Hartnoll et al 4 England Unsupervised heroin treatment provision

Six randomised trials in six countries over 15 years: synthesis of findings

In this section, we present the trials in historical sequence. The early heroin trial from the 1970s 4 was not included since this was not based on the new approach of supervised injecting. The series of SIH treatment trials commences with the 1998 Perneger trial in Switzerland, the crucible of the new supervised injecting clinic approach. All of the new randomised trials summarised in this article have taken as their study participants chronic heroin-dependent individuals who have repeatedly failed in orthodox treatment (either currently still failing in treatment as evidenced by continued regular heroin injecting, or alternatively currently no longer engaged in treatment), apart from a subsample of the German study, and they have included randomised comparison with the standard treatment of oral MMT. Generally, the results were consistent and each trial has progressively strengthened the evidence-base for this new treatment approach.

(a) Switzerland, 1998

This small study (n = 51) was important as the first randomised trial of this new supervised treatment approach. 5 Participants were studied over a 6-month period of injected diamorphine or oral MMT. The two groups had equivalent retention, but the diamorphine-prescribed group had significantly greater reductions in illicit heroin use and in crime after 6 months of treatment. Continued illicit heroin use was self-reported by only 22% of the heroin-prescribed group v. 67% of the control group. This early trial contributed to establishing the feasibility of the randomised trial study design, the potential acceptability of the supervised injectable clinic modality, the high doses of diamorphine maintenance that could safely be administered, as well as the good short-term outcome at 6 months.

(b) The Netherlands, 2003

The two Dutch multi-site randomised trials 13 constituted a significant step-change in the evidence-base, bringing sufficient sample size (n = 594) and study rigour to reach more robust conclusions. One of the trials studied the efficacy and safety of injectable diacetylmorphine (n = 174), the other the efficacy and safety of inhalable diacetylmorphine (n = 375) and will not be considered further in this article. Retention rate for MMT at 12 months was higher (85%) than for SIH (72%), but a much larger proportion of the heroin- prescribed group were ‘responders’ on the pre-determined composite scale of response (57% v. 32%). In addition, the Dutch trials showed that SIH was cost-effective for this target population. 44 The study method and the results from the Dutch trial guided the construction of the later trials reported in this article.

(c) Spain, 2006

This small (n = 62) randomised trial was undertaken in Andalucia 14 and found equivalent retention, and significantly greater reduction in self-reported illicit heroin use in the diamorphine group at their selected 9-month follow-up point. Despite the small sample size and the continued reliance on self-report, these findings provided further evidence of benefit to previous studies and also contributed the perspectives of the families of the heroin addicts in SIH treatment. 40

(d) Germany, 2007

This multi-site trial 15 is the largest conducted to date (n = 1015), and found slightly higher retention in the heroin compared with the methadone group. It found greater proportions of the heroin-prescribed group reporting reduced heroin use and being ‘responders’ on the multidimensional outcome measure. An advance in this trial was the attention to ensuring good dosage for participants randomised to oral MMT (thus addressing concern that the apparent advantage of heroin-prescribing may be an artefact of suboptimal treatment in the control group). This trial also incorporated various other study elements (two different recruitment strands; and two styles of counselling therapy – neither of which was associated with meaningful differences in outcome). This trial was the first to include objective laboratory test results for illicit heroin, but these were not available across all participants, were only incorporated into the composite score and were not reported separately.

(e) Canada, 2009

The Canadian NAOMI (North American Opiate Medication Initiative) trial (n = 226), a two-site randomised trial, 16 was the first of the randomised trials to be conducted outside Europe and was carried out in severely affected participants not currently in treatment but with multiple previous treatment attempts. Significantly higher rates of retention (in SIH or other treatment) and clinical response scores occurred in those randomised to diamorphine. This trial also included a small subsidiary arm (n = 25) that was an exploratory double-masked evaluation of injectable hydromorphone and which included objective laboratory urinanalysis, and the results showed broadly equivalent benefits. 47

(f) England, 2010

The UK three-site RIOTT (Randomised Injectable Opioid Treatment Trial) 17 was important as the first trial to be conducted with laboratory illicit opioid test results as the pre-declared primary outcome measure. This three-way randomised trial compared two forms of supervised injectable maintenance (SIH and supervised injectable methadone maintenance) against an optimised version of oral MMT. Although the sample size was modest (n = 127 across the three groups), the investigators had the benefit of the previous trials to guide calculations of sample size and power, as well as improved laboratory analytical methods involving assay for papaverine and other components of illicit heroin. 48 Good retention was achieved in all groups. At months 4–6, the heroin-treated group was significantly more likely to provide urine specimens negative for markers of illicit heroin than the optimised MMT group. This trial also reported on the speed of onset of the benefit observed in the heroin-treated group (as had the Dutch trial), and again benefits were evident within 2 months of treatment.

Effects of SIH treatment

Opiate use outcome data

Across the trials, different opiate use reduction (or abstinence) outcome measures were used, which prevents exploration of the pooled results in relation to this outcome. Nonetheless, there was a positive effect of SIH on illicit heroin use reported by each individual study. 5,1317

Retention in treatment outcome data

Utilising available data from four studies, 5,15,16,17 our meta-analysis identified a significant advantage of SIH over oral MMT treatment in retention in treatment: overall RR = 1.37 (95% CI 1.03–1.83), heterogeneity (P<0.00001), I 2 = 91% (Fig. 2). The Dutch 13 and the Spanish 14 studies were excluded from the analysis of retention because of the specific construction of the two study conditions, i.e. as per trial designs, the participants in the MMT groups had an automatic right to be offered SIH at the end of the randomised trial period. The possibility of exclusion of the RIOTT for the same reason was considered; however, this was not required as there was no automatic right to be offered injectable maintenance at the end of the 6-month randomised trial period, even though there was, in practice, a sympathetic consideration of this request if it was made.

Fig. 2 Supervised injectable heroin (SIH) + flexible doses of oral methadone v. oral methadone: retention in treatment.

Mortality outcome data

The six trials collectively identified 16 events of death (SIH: n = 6; oral MMT: n = 10) resulting in a numerical advantage of SIH over oral MMT, but crossing the mid-line: RR = 0.65 (95% CI 0.25–1.69), heterogeneity (P = 0.89), I = 0% (Fig. 3).

Fig. 3 Supervised injectable heroin (SIH) + flexible doses of oral methadone v. oral methadone: mortality.

Side-effects data

Taking all side-effects together (serious adverse events probably or definitely related to study medication), the five trials (the Swiss study did not report side-effects data) showed a significant higher risk of side-effects in the SIH compared with the oral MMT treatment groups: RR = 4.99 (95% CI 1.66–14.99), heterogeneity (P = 0.25), I 2 = 26% (Fig. 4).

Fig. 4 Supervised injectable heroin (SIH) + flexible doses of oral methadone v. oral methadone maintenance treatment (MMT): side-effects (serious adverse events probably/definitely related to study medication).

Publication bias

Figure 5 presents the funnel plots to assess potential publication bias for the meta-analyses. We have restricted this to a visual inspection of the plots in line with recommendations not to perform statistical tests of asymmetry where there are a small number of trials. 49 The first two funnel plots (Fig. 5a and 5b) relate to the outcomes of retention (as reported earlier and in Fig. 2) and of mortality (as reported earlier and in Fig. 3), and they indicate that the studies had, respectively, very small and small standard errors and RR estimates spanning from below 1 to approximately 2. However, with the outcome of side-effects, the funnel plot in Fig. 5c indicates that the studies included much more variable standard error estimates, with RR above 10, which may reflect small sample size or other limitations.

Fig. 5 Funnel plot of comparison: supervised injectable heroin (SIH) + flexible doses of oral methadone v. oral methadone – outcomes: (a) retention in treatment; (b) mortality; and (c) side-effects.

National and international impact on clinical practice and policy

At the international level, the 1961 and 1971 UN conventions 50,51 contain no explicit regulations concerning the prescribing of diamorphine (heroin) in the context of substitution treatment provision, leaving it to the competence of national governments to regulate in this area. National legislation differs greatly between countries. With the exception of the UK, the development of regulation by means of law and guidelines around heroin prescribing for opioid treatment is a very recent matter.

(a) Countries in which diamorphine exists as a medicinal product

(i) Full approval of diamorphine as a medicinal product (UK). The medical use of heroin is, and always has been, recognised in the UK as a legitimate medicine which a doctor may prescribe for the relief of pain and suffering, as well as for the treatment of opioid dependence. 52,53 However, since the late 1960s, the authority to prescribe diamorphine for addiction treatment has been restricted to doctors with a special licence (essentially being addiction specialists), while all medical practitioners continue to have the authority to prescribe diamorphine for other conditions (e.g. severe pain relief, acute management of coronary infarction).

(ii) Approval of diamorphine as a medicinal product for the specific indication of treatment-refractory heroin dependence (Switzerland, Germany, The Netherlands and Denmark). In Switzerland, Germany and The Netherlands, heroin has been given approved medication status as a legitimate (albeit reserved for severe cases) opioid substitution treatment. In 2001, injectable heroin was registered in Switzerland as a medication for maintenance treatment in opioid dependence, followed by its inclusion on the list of provisions to be fully paid by health insurance in 2002; and, finally, a legal basis was obtained through revision of the narcotic law in 2008. 54 A similar process has been followed and completed over the last decade in The Netherlands 13 and in Germany. 55 In Denmark in 2008, amendment to the Controlled Substances Act was adopted, which allowed the provision of supervised heroin-prescribing and its integration into the existing therapeutic network as an additional treatment for long-term heroin addicts. 56

(b) Countries which have approved diamorphine for research trials

There are also other countries where, in recent years, approval has been given for diamorphine to be prescribed within the context of a randomised trial (e.g. Canada and Spain) – for example, through approval by the Office of Controlled Substances of Health Canada and a Section 56 exemption from Canada's Narcotics Control Act 57 and, in Spain, through Royal Decrees 75/1990 58 and 5/1996. 59 Approval was granted in 2007 for a similar trial in Belgium that was recently finalised (no report available yet).

(c) Countries in which diamorphine is totally prohibited and hence not available as a medicinal product nor as a research medication

Finally, there are all other countries in the rest of the world where either (i) such treatment appears never to have been seriously proposed or (ii) heroin-prescribing trials have been proposed but have then either been blocked or approval has not been granted (Australia, USA and France).


Main findings

A total of six randomised trials from six countries have been included in this review. Based on the evidence that has been accumulated through these clinical trials, heroin-prescribing, as a part of highly regulated regimen, is a feasible and effective treatment for a particularly difficult-to-treat group of heroin-dependent patients. Diamorphine hydrochloride (pharmaceutical heroin) is now registered as a medicinal product for this indication in five European countries (Switzerland, The Netherlands, Germany, UK and Denmark). New research is now testing whether further improvements could be achieved with combination of SIH and incentive reinforcement (termed contingency management, CM) or other specific rehabilitation strategies. Following the conduct of this series of rigorous randomised trials, several countries have altered research restrictions and there has also been new regulatory approval and politically supported changes in narcotics laws of these countries, so that this potentially effective treatment is now becoming available for at least some of the patients whose addiction was previously considered untreatable (and still is, in most countries). An additional option has been added to the clinical algorithm, which can improve personalisation of individually relevant treatment provision, to the benefit of individuals as well as society at large.

Comparison with Cochrane

It is appropriate to compare and contrast the conclusions from the above analyses with the conclusions from earlier and more recent Cochrane Reviews. 12,60 The original 2005 Cochrane Review 60 examined studies published up to 2002 (and with only two of the studies included in our analysis above) and concluded that, even though there were some results in favour of heroin treatment, ‘no definitive conclusions about the overall effectiveness of heroin prescription (was) possible’. By the time of the later Cochrane Review in 2011, all six of the above-randomised trials were included in the new Cochrane analysis, 12 and the Cochrane group concluded that, on the basis of the expanded current evidence, ‘heroin prescription should be indicated to people who (are) currently or have previously failed maintenance treatment, and it should be provided in clinical settings where proper follow-up is ensured’, while also noting that adverse events were consistently more frequent in the heroin groups.

However, a major difference exists in the approach taken by our analyses v. the main approach taken by the Cochrane Review: the Cochrane group have included all trials of heroin prescribing, regardless of whether the administration was supervised or for take-home administration (although with additional analyses later included along the lines of the above analyses), whereas we have regarded the SIH approach as a distinct treatment necessitating its own specific scrutiny and analysis. We consider this distinction important because we wish to avoid any possible contamination of analyses, which could result from inclusion of findings from earlier trials in which supplies of heroin were given to addicts on a take-home basis. We thus consider it more appropriate to analyse solely the trials of the new clinical approach of SIH, and this is the basis of our analyses above. The overall conclusions are similar, but a clearer and stronger signal emerges from the more specific narrower approach we have taken.

Obstacles to fuller impact

The introduction of effective interventions, even when demonstrably effective, can sometimes, at first, be viewed as controversial. SIH treatment is often viewed thus. A number of concerns have been raised and we address these in turn.

(a) Concerns about the adequacy of the scientific evidence

This was previously a major obstacle, but has now largely been addressed by the series of trials described above. All of the trials have broadly shown similar benefits and in the same direction – with regard to ‘street’ heroin and other drug use as well as in secondary outcome domains such as physical, mental health and social functioning where these have been studied (Spain, 38 Germany 22,28,29,30,36,41 and Canada 34,35 ). Also, the latest 2011 Cochrane review 12 reaches a more positive conclusion on SIH than the original 2005 Cochrane review. 60 However, scientific questions still remain. The new empirical evidence from randomised trials on heroin treatment has mostly focused on short-term outcome, with the randomisation phase of treatment being a maximum of 12 months. Nevertheless, longer-term data are also available from eight extended follow-up studies in four countries (Switzerland, 8,45,56 The Netherlands, 26 Spain 33 and Germany 21,25,42 ) with a consistent finding of additional sustained benefit across a range of different outcome categories. We also need to learn more about the process and influences on remission of illicit drug use and elimination of related problems, and, more importantly, enhanced quality of life and social functioning of these patients.

(b) Concerns about security, public safety, and potential for diversion and abuse

Much concern has been expressed over security, public safety and potential for diversion of prescribed heroin. Three of the randomised trials have evaluated the impact of newly established injectable clinics on crime in trial localities: The Netherlands, 27 Canada 31 and the UK. 32 Findings to date suggest no negative effects of the new supervised injecting clinics on public safety, and actual reports of growing local public support.

(c) Concern about rebound damage to other treatments such as oral MMT and rehabilitation

Concern that prescribed diamorphine would preferentially attract heroin users and would undermine other treatments has not been borne out. Most of the six trials actually experienced difficulty in recruiting participants, either failing to reach target recruitment 14,16,17 or needing to extend the planned recruitment time. 15,17 It appears that for many marginalised heroin users, the attraction of prescribed diamorphine is rarely sufficient to promote engagement in highly structured treatment. Recent documented experience 20,24,27,40,41,61,62 suggests that many patients attending the new injecting clinics aim at sobriety in the longer term or return to healthier stability in existing MMT programmes. However, this still needs to be studied further. A suitable response to the needs and aspirations of this patient group will involve investment of collective effort to developing recovery-oriented heroin maintenance – an approach that will combine heroin pharmacotherapy and a sustained menu of recovery support services to assist patients and families in achieving long-term addiction recovery.

(d) Financial costs

In a context of ever-increasing health costs and competing health priorities, heroin prescribing might be difficult for governments to embrace. Findings of international research 19,23,37,44 have consistently demonstrated a considerable economic benefit of SIH because of the reduction in the costs of criminal procedures, imprisonment and healthcare. Different models of possible service provision of heroin treatment may identify variants of SIH treatment which are more affordable, and this was being explored in England 63,64 up until 2015 when the central funding for this new treatment was not renewed

(e) Hijack by campaigning groups

The encouraging findings from the randomised trials has been picked up by groups campaigning for major changes in the law and the trials have been described as if they were trials of legalisation (which they were not). These misrepresentations are not only misleading but also risk damaging the robustness of the conclusions and the integrity of the clinical procedures. This difficulty is not unique to the heroin trials, and it similarly interferes with objective discussion of harm reduction policies and practices; 6567 however, careful attention to accurate secondary reporting of the findings of the heroin trials is important so that they are properly understood and the potential for advancement properly identified.

(f) Diamorphophobia

A critical concern relates to public and political anxiety about the acceptability of the idea of heroin being a medicinal product. While diamorphine has existed as a pharmaceutically manufactured medicinal product in the UK for more than a century, the situation is very different in most other countries where heroin is usually regarded as always an illicitly manufactured drug of abuse and addiction. This has contributed to an inability to establish clinical research trials (e.g. Australia 68 ) and to the refusal to provide continuity of diamorphine treatment for individuals beyond the end of trial treatment (e.g. Spain). It is possible that the Canadian identification of similar benefits with injectable hydromorphone 19 may point to an avenue which might circumvent more severe expressions of such diamorphophobia.

(g) Safety

Several of the trials have reported instances of sudden-onset respiratory depression in people receiving injectable diamorphine, at a rate of about 1 in every 6000 injections, 16,17 hence well below the hazard from injecting street heroin but nevertheless producing clinically critical events. These have all been safely managed with resuscitation measures, but, as noted in the 2011 Cochrane review, this necessitates specific attention and emphasises the importance of supervision of injection by appropriately trained staff. 43 This repeated finding warrants fuller study, and future research will clarify whether it relates to the medicinal product (diamorphine/heroin) itself or to some other aspect of drug-taking behaviour or drug treatment provision. Some such work is ongoing.

Next steps

A trial of SIH treatment has been conducted (2011–2013) in Belgium and future versions of the analysis will be likely to include data from this trial also, once the findings from this further trial have been peer-reviewed and published.


The key limitation of this review is that the analysis synthesised the interpretation of the primary data in each paper rather than considering the primary data directly. Future research could compare SIH treatment outcomes across these trials for a number of outcomes by analysing individual patient data generated by the different research groups.


Funding was received from the EMCDDA (European Monitoring Centre for Drugs and Drug Addiction) to support review of the international randomised trials of supervised injectable diamorphine (heroin) prescribing and the preparation of a 2012 European report on the evidence-base for this approach for the treatment of entrenched heroin addiction and associated evolving practice.


We thank Marina Davoli, Marica Ferri and the Cochrane Drugs and Alcohol Group for their contribution. J.S. is supported by the National Institute for Health Research (NIHR) Biomedical Research Centre for Mental Health at South London and Maudsley NHS Foundation Trust and King's College London.


1 McLellan, AT, Lewis, DC, O'Brien, CP, Kleber, HD. Drug dependence, a chronic medical illness: implications for treatment, insurance, and outcomes evaluation. JAMA 2000; 284: 1689–95.
2 Strang, J, Babor, T, Caulkins, J, Fischer, B, Foxcroft, D, Humphreys, K. Drug policy and the public good: evidence for effective interventions. Lancet 2012; 379: 7183.
3 Bammer, G, Dobler-Mikola, A, Fleming, PM, Strang, J, Uchtenhagen, A. The heroin prescribing debate: integrating science and politics. Science 1999; 284: 1277–8.
4 Hartnoll, RL, Mitcheson, MC, Battersby, A, Brown, G, Ellis, M, Fleming, P, et al. Evaluation of heroin maintenance in controlled trial. Arch Gen Psychiatry 1980; 37: 877–84.
5 Perneger, TV, Giner, F, del, RM, Mino, A. Randomised trial of heroin maintenance programme for addicts who fail in conventional drug treatments. BMJ 1998; 317: 1318.
6 Musto, D. The American Disease: Origins of Narcotic Control. Oxford University Press, 1999.
7 Strang, J, Gossop, M. Heroin Addiction and the British System: Treatment and Other Responses, vol. 2. Routledge, 2004.
8 Rehm, J, Gschwend, P, Steffen, T, Gutzwiller, F, Dobler-Mikola, A, Uchtenhagen, A. Feasibility, safety, and efficacy of injectable heroin prescription for refractory opioid addicts: a follow-up study. Lancet 2001; 358: 1417–23.
9 Uchtenhagen, A. Narcotics Prescription for Heroin Addicts: Main Results of the Swiss National Cohort Study (Medical Prescription of Narcotics). Karger, 1998.
10 Uchtenhagen, A. Heroin-assisted treatment in Switzerland: a case study in policy change. Addiction 2010; 105: 2937.
11 Wright, NM, Tompkins, CN. Supervised injecting centres. BMJ 2004; 328: 100–2.
12 Ferri, M, Davoli, M, Perucci, CA. Heroin maintenance for chronic heroin-dependent individuals. Cochrane Database Syst Rev 2011; 12: CD003410.
13 van den Brink, W, Hendriks, VM, Blanken, P, Koeter, MW, van Zwieten, BJ, van Ree, JM. Medical prescription of heroin to treatment resistant heroin addicts: two randomised controlled trials. BMJ 2003; 327: 310.
14 March, JC, Oviedo-Joekes, E, Perea-Milla, E, Carrasco, F. Controlled trial of prescribed heroin in the treatment of opioid addiction. J Subst Abuse Treat 2006; 31: 203–11.
15 Haasen, C, Verthein, U, Degkwitz, P, Berger, J, Krausz, M, Naber, D. Heroin-assisted treatment for opioid dependence: randomised controlled trial. Br J Psychiatry 2007; 191: 5562.
16 Oviedo-Joekes, E, Brissette, S, Marsh, DC, Lauzon, P, Guh, D, Anis, A, et al. Diacetylmorphine versus methadone for the treatment of opioid addiction. N Engl J Med 2009; 361: 777–86.
17 Strang, J, Metrebian, N, Lintzeris, N, Potts, L, Carnwath, T, Mayet, S, et al. Supervised injectable heroin or injectable methadone versus optimised oral methadone as treatment for chronic heroin addicts in England after persistent failure in orthodox treatment (RIOTT): a randomised trial. Lancet 2010; 375: 1885–95.
18 Higgins, JPT, Green, S. Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.1 (updated September 2009). The Cochrane Collaboration, 2008.
19 Byford, S, Barrett, B, Metrebian, N, Groshkova, T, Cary, M, Charles, V, et al. Cost-effectiveness of injectable opioid treatment v. oral methadone for chronic heroin addiction. Br J Psychiatry 2013; 203: 341–9.
20 Groshkova, T, Metrebian, N, Hallam, C, Charles, V, Martin, A, Forzisi, L, et al. Treatment expectations and satisfaction of treatment-refractory opioid-dependent patients in RIOTT, the Randomised Injectable Opiate Treatment Trial, the UK's first supervised injectable maintenance clinics. Drug Alcohol Rev 2013; 32: 566–73.
21 Verthein, U, Schafer, I, Degkwitz, P. Social integration after 4 years of heroin-assisted treatment. Rehabilitation 2013; 52: 243–50.
22 Vogel, M, Knopfli, B, Schmid, O, Prica, M, Strasser, J, Prieto, L, et al. Treatment or “high”: benzodiazepine use in patients on injectable heroin or oral opioids. Addict Behav 2013; 38: 2477–84.
23 Nosyk, B, Guh, DP, Bansback, NJ, Oviedo-Joekes, E, Brissette, S, Marsh, DC, et al. Cost-effectiveness of diacetylmorphine versus methadone for chronic opioid dependence refractory to treatment. Can Med Assoc J 2012; 184: E31728.
24 Marchand, KI, Oviedo-Joekes, E, Guh, D, Brissette, S, Marsh, DC, Schechter, MT. Client satisfaction among participants in a randomized trial comparing oral methadone and injectable diacetylmorphine for long-term opioid-dependency. BMC Health Serv Res 2011; 11: 174–84.
25 Verthein, U, Haasen, C, Reimer, J. Switching from methadone to diamorphine: 2-year results of the German heroin-assisted treatment trial. Subst Use Misuse 2011; 46: 980–91.
26 Blanken, P, Hendriks, VM, van Ree, JM, van den Brink, W. Outcome of long-term heroin-assisted treatment offered to chronic, treatment-resistant heroin addicts in the Netherlands. Addiction 2010; 105: 300–8.
27 Blanken, P, van den Brink, W, Hendriks, VM, Huijsman, IA, Klous, MG, Rook, EJ, et al. Heroin-assisted treatment in the Netherlands: History, findings, and international context. Eur Neuropsychopharmacol 2010; 20 (suppl 2): S105158.
28 Eiroa-Orosa Haasen, C, Verthein, U, Dilg, C, Schafer, I, Reimer, J. Benzodiazepine use among patients in heroin-assisted vs methadone maintenance treatment. Findings of the German randomised controlled trial. Drug Alcohol Depend 2010; 112: 226–33.
29 Haasen, C, Verthein, U, Eiroa-Orosa Schafer, I, Reimer, J. Is heroin-assisted treatment effective for patients with no previous maintenance treatment? Results from a German randomised controlled trial. Eur Addict Res 2010; 16: 124–30.
30 Karow, A, Reimer, J, Schafer, I, Krausz, M, Haasen, C, Verthein, U. Quality of life under maintenance treatment with heroin versus methadone in patients with opioid dependence. Drug Alcohol Depend 2010; 112: 209–15.
31 Lasnier, B, Brochu, S, Boyd, N, Fischer, B. A heroin prescription trial: case studies from Montreal and Vancouver on crime and disorder in the surrounding neighbourhoods. Int J Drug Policy 2010; 21: 2835.
32 Miller, P, McKenzie, S, Lintzeris, N, Martin, A, Strang, J. The community impact of RIOTT, a medically supervised injectable maintenance clinic in South London. Mental Health Subst Use Dual Diagn 2010; 3: 248–59.
33 Oviedo-Joekes, E, March, JC, Romero, M, Perea-Milla, E. The Andalusian trial on heroin-assisted treatment: a 2 year follow-up. Drug Alcohol Rev 2010; 29: 7580.
34 Oviedo-Joekes, E, Guh, D, Brissette, S, Marchand, K, Marsh, D, Chettiar, J, et al. Effectiveness of diacetylmorphine versus methadone for the treatment of opioid dependence in women. Drug Alcohol Depend 2010; 111: 50–7.
35 Oviedo-Joekes, E, Guh, D, Marsh, DC, Brissette, S, Nosyk, B, Krausz, M, et al. Characteristics and response to treatment among aboriginal people receiving heroin-assisted treatment. Can J Public Health 2010; 101: 210–2.
36 Scafer, I, Eiroa-Oroa, FJ, Verthein, U, Dilg, C, Haasen, C, Reimer, J. Effects of psychiatric comorbidity on treatment outcomes in patients undergoing diamorphine or methadone maintenance. Psychopathology 2010; 43: 8895.
37 Haasen, C. Gesundheitsökonomische Begleitforschung (Unpublished Economic Evaluation Report, German Trial). 2009.
38 Perea-Milla, E, Aycaguer, LCS, Cerda, JCM, Saiz, FG, Rivas-Ruiz, F, Danet, A, et al. Efficacy of prescribed injectable diacetylmorphine in the Andalusian trial: Bayesian analysis of responders and non-responders according to a multi domain outcome index. Trials 2009; 10: 70–6.
39 Haasen, C, Eiroa-Orosa Verthein, U, Soyka, M, Dilg, C, Schafer, I, et al. Effects of heroin assisted treatment on alcohol consumption. Findings of the German randomised controlled trial. Alcohol 2009; 43: 259–64.
40 Romo, N, Poo, M, Ballesta, R. From illegal poison to legal medicine: a qualitative research in a heroin-prescription trial in Spain. Drug Alcohol Rev 2009; 28: 186–95.
41 Miller, P, Forzisi, L, Lintzeris, N, Zador, D, Metrebian, N, Strang, J. Groin injecting in injectable opioid treatment service users in South London. Addict Res Theory 2009; 17: 381–9.
42 Verthein, U, Bonorden-Kleij, K, Degkwitz, P, Dilg, C, Kohler, WK, Passie, T, et al. Long-term effects of heroin-assisted treatment in Germany. Addiction 2008; 103: 960–6.
43 Dursteler-Macfarland, KM, Stohler, R, Moldovanyi, A, Rey, S, Basdekis, R, Gschwend, P, et al. Complaints of heroin-maintained patients – a survey of symptoms ascribed to diacetylmorphine. Drug Alcohol Depend 2006; 81: 231–9.
44 Dijkgraaf, MG, van der Zanden, BP, de Borgie, CA, Blanken, P, van Ree, JM, van den Brink, W. Cost utility analysis of co-prescribed heroin compared with methadone maintenance treatment in heroin addicts in two randomised trials. BMJ 2005; 330: 1297.
45 Rehm, J, Frick, U, Hartwig, C, Gutzwiller, P, Uchtenhagen, A. Mortality in heroin-assisted treatment in Switzerland 1994–2000. Drug Alcohol Depend 2005; 79: 137–43.
46 Guttinger, F, Gschwend, P, Schulte, B, Rehm, J, Uchtenhagen, A. Evaluating long-term effects of heroin-assisted treatment: the results of a 6-year follow-up. Eur Addict Res 2003; 9: 73–9.
47 Oviedo-Joekes, E, Guh, D, Brissette, S, Oviedo-Joekes, E, Guh, D, Brissette, S, et al. Double-blind injectable hydromorphone versus diacetylmorphine for the treatment of opioid dependence: a pilot study. J Subst Abuse Treat 2010; 38: 408–11.
48 Paterson, S, Lintzeris, N, Mitchell, TB, Cordero, R, Nestor, L, Strang, J. Validation of techniques to detect illicit heroin use in patients prescribed pharmaceutical heroin for the management of opioid dependence. Addiction 2005; 100: 1832–39.
49 Sterne, JAC, Harbord, RM. Funnel plots in meta-analysis. Stata J 2004; 4: 127–41.
50 United Nations Single Convention on Narcotic Drugs. UN, 1961.
51 United Nations Convention on Psychotropic Substances. UN, 1971.
52 Rolleston Report: Report of the Departmental Committee on Morphine and Heroin Addiction. Ministry of Health, 1926.
53 Interdepartmental Committee on Drug Addiction (Brain Committee). Drug Addiction, Second Report. Her Majesty's Stationery Office, 1965.
54 Federal Law on Narcotics and Psychotropic Substances. Revision 30.11.2008. Department of Health, 2008.
55 Pfeiffer-Gerschel, T, Kipke, I, Floter, S, Lieb, C, Raiser, P. German National Report on Drugs to the European Monitoring Centre for Drugs and Drug Addiction. Munich, Germany, 2009.
56 Order 29.5.2008 of the Ministrz of Health and Prevention. Denmark, 2008.
57 Gartry, CC, Oviedo-Joekes, E, Laliberte, N, Schechter, MT. NAOMI: the trials and tribulations of implementing a heroin assisted treatment study in North America. Harm Reduct J 2009; 6: 2.
58 Ministerio De La Presidencia. Real Decreto 75/1990 De 19 De Enero, Por El Que Se Regulan Los Tratamientos Con Opiáceos De Las Personas Dependientes De Los Mismos. Boletín Oficial Del Estado Del 11 De Abril 1990. Ministerio de la Presidencia, 1990.
59 Ministerio De La Presidencia. Real Decreto 5/1996 De 15 De Enero, Sobre Modificación Del Real Decreto 75/1990, De 19 De Enero, Por El Que Se Regulan Los Tratamientos Con Opiáceos De Personas Dependientes De Los Mismos, Y De Ampliación De Su Anexo. Boletín Oficial Del Estado Del 20 De Febrero 1996. Minesterio de la Presidencia, 1996.
60 Ferri, M, Davoli, M, Perucci, CA. Heroin maintenance for chronic heroin dependents. Cochrane Database Syst Rev 2005; 2: CD003410.
61 Clark, S. RIOTT, Recovery and Me: A Personal Story from Our New Members' Council Recruit. SLaM news, 2426, 2010.
62 Hannah, X. I predict a RIOTT. Black Poppy 2012: 14: 27–8.
63 Strang, J, Groshkova, T, Metrebian, N. New heroin-assisted treatment: recent evidence and current practices of supervised injectable heroin treatment in Europe and beyond (EMCDDA Insight): 143–144. European Monitoring Centre for Drugs and Drug Addiction, Lisbon, Portugal, 2012 (
64 Department of Health (DoH). Invitation to tender: the piloting of supervised injectable Opioid Treatment. DoH, 2012 (
65 MacCoun, R, Reuter, P. Drug War Heresies: Learning from Other Vices, Times and Places. Cambridge University Press, 2001.
66 Schechter, M. Science, ideology, and needle exchange programs. Ann Am Acad Political Social Sci 2002; 582: 94101.
67 Strang, J. Harm reduction for drug users: exploring the dimensions of harm, their measurement and the strategies for reduction. AIDS Public Policy J 1992; 7: 145–52.
68 Bammer, G. The act heroin trial: intellectual, practical and political challenges. Drug Alcohol Rev 1997; 16: 287–96.