In France, high-dosage buprenorphine (HDB) is the main substitution treatment for narcotic addiction. Few data have been published concerning clinical factors predicting a good response to this treatment in a daily practice. A hospital-based multicenter clinical research program (PHRC) was undertaken in heroin-addicted patients, diagnosed according to DSM-III-R, to detect clinical criteria susceptible of predicting a good response to HDB administered during a 3-month treatment period. At the inclusion time in the study, a diagnostic structured interview (DIGS) was performed, and the Addiction Severity Index (ASI), Zuckerman scale, depression scale from Jouvent, and CGI were scored. MMPI was also administered. Good response was defined as an ongoing participation in the study, with absence of opiate detected in 75% of urine collected during the last month of treatment. Only subjects treated for at least 1 month were eligible for analyses. One hundred fifteen patients were recruited and 73 were analyzed. Patients received 8.5+/-2.6 mg (m+/-S.D.) of buprenorphine for 1 to 3 months. A forward stepwise logistic regression showed that six clinical parameters may predict a good response to treatment: probability to respond to buprenorphine was higher in subjects having a high psychopathology (ASI) subscore, low disinhibition and boredom susceptibility factor scores (Zuckerman scale), no alcohol dependence, no family history of addiction or mood disorder, and duration of opiate dependence less than 10 years. Only the MMPI D subscale was a psychological pattern correlated to a good response to substitution treatment. These findings are important to consider when making the decision to prescribe HDB substitution treatment in opiate addiction.