Favorable mortality profile of naltrexone implants for opiate addiction.
Reece A.S. Journal of Addictive Diseases: 2010, 29, p. 30–50.

Few treatments for opiate addiction arouse as much controversy as naltrexone implants. Inserted under the skin, these block the effects of heroin for up to several months - for some, a magic bullet, for others, an unsafe and ethically dubious experiment. More evidence from Australia that the overdose death risk is less than with oral forms of the drug.

Summary:
At an Australian addiction treatment clinic the death rate of opiate-dependent patients treated with the heroin substitute buprenorphine between 2000 and 2007 was compared with those treated with implants of naltrexone which block the effects of heroin and allied drugs for up to several months. During this time 2518 patients were prescribed buprenorphine free of charge (though there were pharmacy charges) and 255 were implanted, mostly at their own expense. Records were matched with those of deaths recorded by the registrar of the relevant Australian state.

Main findings:
Four naltrexone and 43 buprenorphine patients were known to have died. There were no deaths while a naltrexone implant was active but four among patients at times when the implant would no longer have been effective, resulting in an overall death rate of 3 deaths per 1000 years the patients had been tracked. For buprenorphine, the corresponding figure was 5.35, largely due to 40 deaths during times when the patients were not being prescribed the drug. Not only was the overall death rate lower among naltrexone patients, but this was also generally the case in each age band. All but one of the four naltrexone deaths occurred several months after the implant would no longer have been effective.