The aim of this study was to replicate and extend the findings of O'Connor, Belanger, Marchand, Dupuis, Elie, and Boyer [Addict. Behav. 24 (1999) 537], which had established a psychosocial profile associated with psychological distress in benzodiazepine (BZD) use. Forty-one participants with anxiety or insomnia, receiving maintenance therapy of BZD for at least 8 weeks, participated in a 20-week, tapered discontinuation protocol with physician counselling. Drug type and use was monitored throughout. Questionnaire measures of anxiety, behavioural inhibition, neuroticism, withdrawal complaints, social support, psychological distress, self-efficacy in coping without BZD, quality of life, positive and negative life events, were completed at baseline, postdiscontinuation, and at 3-month follow-up.Measures of baseline psychological distress and anxiety inhibition were consistently associated with both discontinuation and the emergence of withdrawal complaints. Successful withdrawal was characterized by low baseline neuroticism, low behavioural inhibition, higher number of positive events, and higher level of social support satisfaction. Higher dosage (in diazepam equivalent dose) was associated with both poorer outcome and the emergence of withdrawal symptoms. Self-efficacy in coping was negatively associated with relapse but not with outcome. Psychosocial factors play a role at different stages of the BZD withdrawal process and could be targeted in treatment.