Sir-Peretti-Watel et al .'s paper (Peretti-Watel et al. 2002) was quite a surprise for the investigators of the European School Survey on Alcohol and other Drugs (ESPAD) 1999, who introduced the additional sport variables (Choquet & Arvers 2002). The main focus of this letter is a complementary analysis. Peretti-Watel et al. (2002) concluded that 'results stress the need to control for age and gender'. Yet since sport practice decreases by age while substance use increases, it appeared more reasonable to perform a separate analysis on each age group rather than adjustments on age. The authors also mentioned that 'the relationship between drug use and sport activity depends on the type of sport'. Since the average number of sports practised was about 3.4 (CI 95% 3.3-3.5) in boys and 2.5 (CI 95% 2.4-2.6) in girls, analysis according to the type of sport should be performed cautiously. But the authors did not mention participation in sport competitions, one of the sport questions added to the ESPAD questionnaire. We propose a complementary analysis of the ESPAD 1999 sample described by Peretti-Watel et al. (2002). We analysed the age group 16-18 years (n = 5983; 96.4% replied to the questions concerning sport activities), which is the most representative of the scholar population (the under-16s and over-18s may be following academic programmes not selected in the sample). We performed logistic regressions separately by gender and by each substance, considering not only alcohol, tobacco and cannabis (see ESPAD 1999 criteria, Hibell et al. 2000) but also illicit drugs other than cannabis (7.9%). As explanatory variables, we took into account the exact age, the duration of sports practice