ENGLISH :
Objective: Assess the validity of medical products reporting program (MEDWatch) reports of abuse/dependence and withdrawal associated with Ultram (tramadol). Methods: Reports of possible abuse/dependence or withdrawal associated with Ultram during 13 quarters following launch were spontaneously reported to the manufacturer Ortho-McNeil Pharmaceutical (OMP) and also solicited from 255 NIDA grantees and addiction treatment professionals by an Independent Steering Committee (ISC). Reports were classified by the ISC using DSM-IV criteria, by the Drug Safety and Surveillance (DSS) units of Robert Wood Johnson Pharmaceutical Research Institute (PRI) using World Health Organization Adverse Reaction Terms (WHOART) terms, and reported to the food and drug administration (FDA) via MEDWatch. Rates of abuse/dependence and withdrawal per 100 000 persons exposed were calculated separately for classifications made by the PRI and the ISC, and confidence intervals calculated to determine the degree to which they agreed. Results: For 681 reports submitted to PRI, confidence intervals of ISC ratings contained PRI ratings 12 of 13 times for abuse/dependence, and 12 of 13 times for withdrawal. For 242 reports submitted to the ISC, confidence intervals of ISC ratings contained PRI ratings 10 of 13 times for abuse/dependence, and 12 of 13 times for withdrawal. Proactive surveillance increased the total number of cases of abuse/dependence but not withdrawal. Many cases of withdrawal without signs or symptoms of abuse/dependence were identified. Conclusions: There was good/excellent concordance between MEDWatch and ISC classifications. Proactive surveillance increased cases of abuse/dependence but not withdrawal. Withdrawal with no signs or symptoms of dependence was common. More use of proactive surveillance is likely to improve assessments of public health risks associated with adverse events. (Review's abstract.)