Co-prescribing benzodiazepines to patients in opiate substitution treatment is controversial and often alleged to increase mortality. In an inner-London general practice, patients with problematic benzodiazepine co-dependence were allowed benzodiazepine maintenance treatment (BMT) since 1994, providing an opportunity for analysis. 1) Case-note review of all 278 opiate substitution treatment patients, accruing 1289 patient treatment years; 46% had concurrent BMT. 2) National Health Service database search for patients who died after leaving accrued a further 883 years of information; only patients who left the UK were unaccounted for (4%). Three groups were studied: 1) never obtained benzodiazepine prescription (NOB): n=80); 2) briefly/occasionally prescribed benzodiazepines (BOP): n=71; 3) BMT: n=127. Treatment retention (months); deaths/100 patient treatment years; deaths after leaving the service/100 years of information. Treatment retention: NOB: 34 months; BOP: 51 months; BMT: 72 months. In-treatment mortality: NOB: 1.79/100 patient treatment years; BOP: 0.33/100 patient treatment years; BMT: 1.31/100 patient treatment years. Deaths after leaving service: NOB: 2.24/100 years of information, BOP: 0.63/100 years of information. However, mortality for previously BMT-patients increased by 450% to 5.90/100 years of information. BMT patients had longer treatment retention than NOB or BOP and lower mortality than NOB patients. It is unlikely that patients had access to prescribed benzodiazepines on leaving the service because of restrictions in the national guidelines but co-dependent patients are a high-risk group who may stand to gain most benefit from opiate substitution treatment if combined with benzodiazepine-maintenance.