• Heroin-assisted treatment;
  • opiate dependence;
  • patient-treatment-matching;
  • pharmacotherapy;
  • RCT


Aims  To investigate which baseline patient characteristics of treatment-resistant heroin addicts differentially predicted treatment response to medical heroin prescription compared to standard methadone maintenance treatment.

Design  Two open-label randomized controlled trials; pooled data.

Setting  Methadone maintenance programmes and heroin treatment centres in six cities in the Netherlands.

Participants  Four hundred and thirty heroin addicts.

Intervention  Methadone plus injectable heroin or methadone plus inhalable heroin compared to methadone alone prescribed over 12 months: heroin maximum 1000 mg per day, methadone maximum 150 mg per day.

Main outcome measure  Dichotomous, multi-domain response index, including validated indicators of physical health, mental status and social functioning.

Findings  Data of the inhalable and injectable heroin trials were pooled. Intention-to-treat analysis showed that treatment with medically prescribed heroin plus methadone was significantly more effective (51.8% response) than standard methadone maintenance treatment (28.7%) (95% CI of response difference: 14.1–32.2%). Multivariate logistic regression analyses showed that only one of all baseline characteristics was predictive of a differential treatment effect: patients who had previously participated in abstinence-orientated treatment responded significantly better to heroin-assisted treatment than to methadone treatment (61% versus 24%), while patients without experience in abstinence-orientated treatment did equally well in heroin-assisted or methadone maintenance treatment (39% and 38%, respectively).

Conclusions  The effect of heroin-assisted treatment is not dependent on clinical characteristics, with the exception of previous abstinence-orientated treatment: medical prescription of heroin is most effective for those patients who have previously participated in abstinence-orientated treatment.