Background Drug-induced aphthous ulcers have been the subject of several isolated and heterogeneous case reports for the last three decades. Objectives To perform a case–control study to evaluate the risks linked to drug exposure in aphthous ulcers. Methods Eighty patients with typical clinical patterns of aphthous ulcers and 152 control patients who had had consultations for skin tumours were studied. A standardized questionnaire, concerning clinical features, life-style and medications taken during the last month, was completed for each patient. Results Case patients had a much higher intake of medications than control patients, respectively, 5·1 and 2·8 medications per patient (P < 0·0001). Multivariate paired analysis showed an association between aphthous ulcers and two classes of drugs: non-steroidal anti-inflammatory drugs (P < 0·001) and beta-blockers (P = 0·002). Smoking could have a protective effect against aphthous ulcers (P < 0·001). Conclusions Previous case reports and the results of this study suggest a real link between beta-blockers and aphthous ulcers. Our study did not confirm a role of other drugs but a few interesting case reports with positive reintroduction have to be considered. These results could be beneficial for patients, as healing may occur when the incriminated drug is discontinued.