ENGLISH :
Aims : The Cocaine Rapid Efficacy Screening Trials (CREST) were designed by the National Institute on Drug Abuse Division of Treatment Research and Development (NIDA, DT R&D) to rapidly screen a number of medications potentially useful for the treatment of cocaine dependence. Design : Each CREST trial was designed to compare several medications in a single trial against an unmatched placebo. The placebo group was included in each trial to avoid the nearly universal positive response to medications seen in open-label trials. In addition, a common set of procedures and outcome measures were employed throughout to increase comparability of results obtained from different trials and from different times. Participants : In all, 18 medications were screened in seven different trials, conducted in four different sites throughout the United States involving 398 cocaine-dependent patients. Findings : Three medications were found to be promising enough to include in subsequent larger trials. Common statistical procedures for evaluating medications were developed to facilitate comparisons across sites and across time. A portion of the data were pooled and analyzed, which yielded some useful insights into cocaine dependence and its treatment. Finally, a review of individual trials together with the pooled analysis revealed several potential improvements for future screening trials. Conclusions : Overall, the CREST trials proved to be useful for rapidly screening medications for treatment of cocaine dependence, but several modifications in design should be made before this framework is applied further. (Review' s abstract)