Self-reported side effects in buprenorphine and methadone patients receiving antiretroviral therapy: results from the MANIF 2000 cohort study M. Patrizia Carrieri 1,2,3 , Perrine Roux 1,2,3 , Julien Cohen 1,2,3 , Isabelle Ravaux 4 , Pierre Dellamonica 5 , Camelia Protopopescu 1,2,3 , Bruno Spire 1,2,3 & Andrew Rosenblum 6 NSERM, U912 (SE4S), Marseille, France, 1 Université Aix Marseille, IRD, UMR-S912, Marseille, France, 2 ORS PACA, Observatoire Régional de la Santé Provence Alpes Côte d’Azur, Marseille, France, 3 Hôpital La Conception, Service des Maladies Infectieuses, Marseille, France, 4 Hôpital L’Archet, Service des Maladies Infectieuses, Nice, France 5 and National Development and Research Institutes, New York, NY 10010, USA 6 ABSTRACT Aims The aim of the study was to investigate the relationship between methadone and buprenorphine treatment and self-reported symptoms in HIV-infected opioid dependent individuals receiving antiretroviral therapy (ART). Design Longitudinal study. Setting The French MANIF2000 cohort was used to compare self-reported symptoms in buprenorphine and methadone patients also receiving ART. Participants We selected individuals receiving ART and OAT (342 visits among 106 patients). Measurements Symptoms were self-reported using a list of 14 symptoms (e.g. nausea, fatigue, fever) perceived during the previous 4 weeks, including three painful symptoms (abdominal or mus- cular pain, headaches). A two-step Heckman approach enabled us to account for the non-random assignment of OAT: a probit model identified predictors of starting either buprenorphine or methadone. A Poisson regression based on generalized estimating equations (GEE) was then used to identify predictors of the number of symptoms while adjusting for the non-random assignment of OAT. Findings The median (interquartile range) number of symptoms was 4 (1–6) and 2 (1–6) among buprenorphine and methadone patients, respectively. After adjustment for non-random assign- ment of OAT type, depressive and opioid withdrawal symptoms, anxiolytics consumption and daily cannabis use, methadone patients were more likely to report a lower number of symptoms than those receiving buprenorphine. Conclusions Methadone patients on ART report fewer symptoms than buprenorphine patients on ART under current treatment conditions in France. Further experimental research is still needed to identify an OAT–ART strategy which would minimize the burden of self-reported symptoms and potential interactions, while assuring sustainability and response to both treatments. Keywords HAART, opioid agonist treatment, pain, selection models, side effects. Correspondence to: M. Patrizia Carrieri, INSERM U912/ORS PACA–23, rue S.Torrents 13006, Marseille, France. E-mail pmcarrieri@aol.com Submitted 16 December 2009; initial review completed 25 February 2010; final version accepted 3 June 2010 INTRODUCTION The introduction of antiretroviral therapy (ART) has highlighted the problem of maintaining adherence (a critical component of ART effectiveness) and managing patient’s ART-related symptoms, which are known to be a major determinant of non-adherence [1]. HIV-infected injecting drug users (IDUs) are at higher risk of non- adherence [2], due probably to a stronger perception of side effects related to ART [3] or to more chaotic life-styles [4]. One factor which contributes significantly to ART discontinuation is medication side effects, or toxicity [5,6]. Methadone and buprenorphine have been shown to be both beneficial for ART adherence [7] and retention in OAT associated with long-term virological response [8]. While a directly observed therapy (DOT) model—used for providing care for both drug dependence and human immunodeficiency virus (HIV)—has been shown to be effective in assuring adherence and good virological out- comes [9], other models based mainly on take-home agonist medications have been shown to ensure good RESEARCH REPORT doi:10.1111/j.1360-0443.2010.03108.x © 2010 The Authors, Addiction © 2010 Society for the Study of Addiction Addiction, 105, 2160–2168