To establish the extent to which participants in maintenance treatment programmes (methadone and sublingual buprenorphine) used opioids, benzodiazepines, ethanol, cocaine and amphetamines, and to define the pattern of such use over time. Design: A retrospective analysis of data of 42,610 analyses from urinalysis screening. Setting: Outpatient drug treatment centre, Department of Psychiatry at Innsbruck Medical University, offering maintenance treatment and detoxification programmes in Tyrol, Austria. Measurements: Enzyme-multiplied immunoassay and gas chromatography-mass spectrometry. Results: Urine samples from 23.4% (N = 4998) of those in the methadone treatment group and 16.5% (N = 1110) in the sublingual buprenorphine group tested positive for consumption of additional drugs (p < 0.001). The most frequently detected drugs were benzodiazepines (63% of all methadone, and 40% of all sublingual buprenorphine positive samples; (p < 0.001)) and opioids (26% of all methadone, and 21% of all sublingual buprenorphine positive samples; (p < 0.001)). In both of the treatment programmes morphine abuse (63%) was higher than heroin (7%). There was a significant decrease of benzodiazepine abuse in the study period (2000- 2003) both among clients receiving methadone (p < 0.001) and those receiving sublingual buprenorphine (p = 0.006). Benzodiazepine (Spearman rank correlation coefficient r = 0.281, p < 0.001), cocaine (r = 0.057, p = 0.007), ethanol (r = 0.073, p= 0.028) and morphine (r = -0.179, p < 0.001) abuse were associated with age in sublingual buprenorphine treated patients. Methadone clients showed a lifetime relation for cocaine (r = -0.032, p =0.007), ethanol (Spearman corr. coeff. r = 0.118, p < 0.001), morphine (r = -0.107, p < 0.001) and dihydrocodeine abuse (r = 0.093, p < 0.001). Discussion: The present data indicate that nonprescribed drug use remains a persistent problem for patients in maintenance treatment. Substances misused are mainly opioids and benzodiazepines. Non-prescribed morphine has replaced heroin.