Background: To compare the cognitive performances of maintenance patients (MAIN), abstinent ex-users (ABST) and healthy non-heroin using controls (CON). Methods: Case control study of 125 MAIN (94 subjects maintained on methadone, 31 on buprenorphine), 50 ABST and 50 CON. Neuropsychological tests measuring executive function, working memory, information processing speed, verbal learning and non-verbal learning were administered. Results: There were no differences between the cognitive profiles of those maintained on methadone or buprenorphine on any administered test. After controlling for confounders, the MAIN group had poorer performance than controls in six of the 13 administered tests, and were poorer than the ABST group in five. The MAIN group exhibited poorer performance in the Haylings Sentence Completion, Matrix Reasoning, Digit Symbol, Logical Memory (immediate and delayed recall), and the Complex Figure Test (immediate recall). There were no differences between the ABST and CON groups on any of the administered tests. Conclusions: Poorer cognitive performance, across a range of test and domains, was seen amongst maintenance patients, regardless of their maintenance drug. This is a group that is likely might benefit from approaches for managing individuals with cognitive and behavioural difficulties arising from brain dysfunction. (C) 2012 Elsevier Ireland Ltd. All rights reserved.