Aims To analyse whether changes in maintenance treatment of opiate-dependent subjects in Sweden were related to changes in opiate-related mortality and inpatient care from 1998 to 2006. Design We collected data from surveys of methadone maintenance treatment units, of buprenorphine and methadone sales, and of mortality and inpatient care in Sweden. Setting Sweden. Participants Patients in maintenance treatment. Measurements Survey data of treatment policy to all units in 2003 and 2005. Trend tests and correlation analyses of data on sales, mortality, inpatient care and forensic investigations. Findings The surveys showed a marked change to a less restrictive policy, with increased use of 'take-away doses' and a reduction of discharges due to side misuse. The one-year retention rate stayed high. Sales of buprenorphine and methadone and the number of patients in treatment increased more than threefold from 2000 to 2006, with the greatest increase for buprenoprphine, introduced in year 2000. There was a significant 20-30% reduction in opiate-related mortality and inpatient care between 2000-2002 and 2004-2006 but not of other drug-related mortality and inpatient care. This decline was larger in Stockholm County, which had a less restricted treatment policy. However, a significant increase in buprenorphine- and methadone-related mortality occurred. For the study period 1998-2006, statistically significant declines occurred only in Stockholm County. Conclusions The liberalization of Sweden's drug policy correlated with an increase in maintenance treatment, a decrease in opiate-related mortality and inpatient care and an increase in deaths with methadone and buprenorphine in the tissues.