RESUMO
To develop a model for estimating the time delay between last heroin consumption and blood sampling in chronic drug users. Eleven patients, all heroin inhalers undergoing detoxification, were included in the study. Several plasma samples were collected during the detoxification procedure and analyzed for the heroin metabolites 6-acetylmorphine (6AM), morphine (MOR), morphine-6-glucuronide (M6G) and morphine-3-glucuronide (M3G), according to a UHPLC/MSMS method. The general linear mixed model was applied to time-related concentrations and a pragmatic four-step delay estimation approach was proposed based on the simultaneous presence of metabolites in plasma. Validation of the model was carried out using the jackknife technique on the 11 patients, and on a group of 7 test patients. Quadratic equations were derived for all metabolites except 6AM. The interval delay estimation was 2-4 days when only M3G present in plasma, 1-2 days when M6G and M3G were both present, 0-1 day when MOR, M6G and M3G were present and <2 h for all metabolites present. The 'jackknife' correlation between declared and actual estimated delays was 0.90. The overall precision of the delay estimates was 8-9 h. The delay between last heroin consumption and blood sampling in chronic drug users can be satisfactorily predicted from plasma heroin metabolites.