Objectives: To study how satisfactory the contents of introductory courses in cardiovascular therapeutics, given to medical students in France, are with respect to the concepts of evidence-based medicine. Methods: Medical school lecturers were asked to provide written course material used as part of medical school courses. Best-available evidence was classified as existent (including two therapeutic subclasses: indicated and contraindicated ), and nonexistent. Four scores (from 0 to 10) were given, according to conformity with best-available evidence, and citation of randomized clinical trials (RCT), meta-analyses and therapeutic objectives. Results: Thirty-four written documents were obtained from 43 faculties. Although the score (mean +/- SEM) of conformity with best-available evidence was 5.43 +/- 0.28 for the existent best-available evidence class, the corresponding scores for the citation of RCT, meta-analyses, and therapeutic objectives were, respectively, 1 +/- 0.2, 0.16 +/- 0.07, and 2.7 +/- 0.3. The four scores were highest when the best-available evidence belonged to the indicated class, intermediate when best-available evidence was nonexistent, and lowest for the contraindicated class (P < .05). These scores were significantly higher when the printed material was intended for specialists. Conclusion: Despite some limitations, the extent of agreement with the best-available evidence is only moderate. Pathophysiologic reasoning is largely preferred to justify the choice of therapeutics recommended to medical students. (Clin Pharmacol Ther 1999;66:173-84.)