Exposure to opioid maintenance treatment reduces long-term mortality Amy Gibson 1 , Louisa Degenhardt 1 , Richard P. Mattick 1 , Robert Ali 2 , Jason White 3 & Susannah O’Brien 1 National Drug and Alcohol Research Centre, UNSW, Australia, 1 Drug and Alcohol Services South Australia, Australia 2 and Clinical and Experimental Pharmacology, University of Adelaide, Australia 3 ABSTRACT Aims To (i) examine the predictors of mortality in a randomized study of methadone versus buprenorphine main- tenance treatment; (ii) compare the survival experience of the randomized subject groups; and (iii) describe the causes of death. Design Ten-year longitudinal follow-up of mortality among participants in a randomized trial of metha- done versus buprenorphine maintenance treatment. Setting Recruitment through three clinics for a randomized trial of buprenorphine versus methadone maintenance. Participants A total of 405 heroin-dependent (DSM-IV) partici- pants aged 18years and above who consented to participate in original study. Measurements Baseline data from original randomized study; dates and causes of death through data linkage with Births, Deaths and Marriages regis- tries; and longitudinal treatment exposure via State health departments. Predictors of mortality examined through survival analysis. Findings There was an overall mortality rate of 8.84 deaths per 1000 person-years of follow-up and causes of death were comparable with the literature. Increased exposure to episodes of opioid treatment longer than 7 days reduced the risk of mortality; there was no differential mortality among methadone versus buprenorphine participants. More dependent, heavier users of heroin at baseline had a lower risk of death, and also higher exposure to opioid treatment. Older participants randomized to buprenorphine treatment had significantly improved survival. Aboriginal or Torres Strait Islander participants had a higher risk of death. Conclusions Increased exposure to opioid maintenance treatment reduces the risk of death in opioid-dependent people. There was no differential reduction between buprenorphine and methadone. Previous studies suggesting differential effects may have been affected by biases in patient selection. Keywords Buprenorphine, longitudinal, maintenance treatment, methadone, mortality, opioid dependence, RCT. Correspondence to: Amy Gibson, NDARC, UNSW, Sydney, NSW 2052, Australia. E-mail: amy.gibson@med.unsw.edu.au Submitted 31 May 2007; initial review completed 16 August 2007; final version accepted 29 October 2007 INTRODUCTION Opioid dependence is associated with mortality rates approximately 13 times higher than the general popula- tion of the same age and sex [1,2]. Research to date has demonstrated that one of the more effective ways of reducing this increased mortality risk is the provision of opioid replacement therapy which, to date, has been examined for methadone: in one Swedish study, untreated heroin-dependent people had mortality rates 63 times the general population, while the mortality rate was eight times lower in those receiving methadone com- pared to untreated heroin-dependent people [3]. An Aus- tralian study showed that the relative risk of an untreated heroin-dependent person dying was 3.5 times that of a patient receiving methadone maintenance treatment [4]. The diverse predictors of mortality in opioid- dependent subjects have been considered in a number of cohort studies. A London cohort of heroin-dependent participants recruited in 1969 noted that neither the length of heroin use nor the age at study intake predicted survival; however, external factors such as drug market and treatment system changes were associated with mor- tality rate changes [5]. A Glasgow cohort recruiting 69% of its participants with heroin as the principal drug of choice (11% in methadone treatment) noted that treat- ment did not have a significant impact on survival; however, the risk of fatality increased through the drug RESEARCH REPORT doi:10.1111/j.1360-0443.2007.02090.x © 2008 The Authors. Journal compilation © 2008 Society for the Study of Addiction Addiction, 103, 462–468