In the first randomised controlled trial of naltrexone implants, Kunøe et al ^{Reference Kunøe, Lobmaier, Kåre Vederhus, Hjerkinn, Hegstad and Gossop1} identify two inclusion criteria in their methodology: being an in-patient and being 18 years or above. Exclusion criteria are given as psychosis, pregnancy and serious hepatic disease. Of 667 possible participants, 480 are excluded. In the results, the term ‘ineligibility’ is used to describe not completing treatment, starting maintenance and transfer to other clinics. Could the authors clarify when these additional criteria were decided upon and how many were excluded for each reason? Given that all 667 patients were receiving ‘abstinence-oriented’ in-patient treatment, it is notable that only a small proportion of patients was eligible for or wanted such treatment. The characteristics of the ineligible or refusal group could provide important information about which group of opiate-dependent patients are likely to benefit from naltrexone.

Data on opioid use throughout the period of treatment would be of value. In the non-abstainers we would expect both groups to use in the first few days, but behavioural extinction to occur in the naltrexone group.

Participants who had their implants removed were included in the analysis using their last response carried forward. If these patients could not be contacted, would it not be a more conservative assumption that they would have relapsed?

The patient group that was living in a controlled environment (prison or clinic) at follow-up was dealt with by using pre-admission data. This group is missing from the flowchart.