(1) The aim of replacement therapy for heroin addiction is to suppress craving for other opiates and to prevent opiate withdrawal symptoms. (2) In France, methadone was the first drug to be licensed for this use, in 1995, with very strict prescribing and dispensing conditions. Buprenorphine was approved in 1996, and was subject to less restrictive conditions. (3) In 2003 in France, an estimated 80 000 people were receiving replacement therapy with buprenorphine and 14 000 with methadone. (4) A meta-analysis of 13 comparative trials involving a total of 2544 patients showed that buprenorphine 6 to 12 mg initially reduced both opiate and benzodiazepine use, whereas doses of 2 to 4 mg had no marked impact on heroin use. This meta-analysis concluded that buprenorphine and methadone had similar efficacy in clinical trials in which the dose was adjusted to outcome. There were more dropouts with buprenorphine than with methadone. A daily dose of 16 mg appeared to be roughly equivalent to 60 mg/day methadone. (5) France appears to be the only country to have relied primarily on buprenorphine as replacement therapy for heroin addiction. This has been the case in France since 1996. The frequency of heroin overdose has fallen markedly in France since 1996, possibly due in part to the availability of replacement therapies. Overall mortality among drug users has also declined, but this is largely due to more effective treatment of HIV infection. (6) In France, a two-year cohort study of patients treated with buprenorphine and a survey conducted during the first year of buprenorphine replacement were funded by the manufacturer, Schering-Plough. The results showed that more than two-thirds of patients remained on treatment, and that, overall, the patients' general condition improved. (7) Opioid-like adverse effects are infrequent under normal conditions of use. There are reports of cases of hepatitis in patients taking buprenorphine, with or without a benzodiazepine. Attribution to buprenorphine is unclear, however, due to the lack of appropriate analyses. (8) Some of the key adverse effects occur during misuse: buprenorphine tablets are often injected, especially during the first few months of treatment (sometimes for more than two years). Injection carries a risk of infections; other potential long-term effects are poorly understood. Compared with methadone users, and regardless of the substances involved, buprenorphine users appear more likely to self-inject. (9) The consequences of sniffing crushed buprenorphine tablets have not been studied. (10) Deaths have been reported following buprenorphine overdose, but they appear to be less frequent than with methadone (0.2 and 0.7 deaths per 1000 users, respectively in 1998). (11) Approaches designed to help patients stop self-injecting have not been tested in comparative trials. Prescriptions of methadone syrup or an injection opiate may be worth trying when all other measures fail.