Many international studies show that religion is an important dimension modulating the use of alcohol and drugs by adolescents. to determine which religious variables are associated to frequent or heavy use of alcohol, tobacco and drugs among adolescents in intermediate and high schools in Campinas, Brazil. A cross-sectional study using a self-report anonymous questionnaire was administered to 2,287 students from a convenience sample of seven schools: five from central areas (two public and three private schools) and two public schools from the outskirts of the city, in 1998. The study analyzes data regarding the use of alcohol, tobacco, medicines, solvents, marijuana, cocaine and ecstasy. The religious variables included in the regression analysis were: religious affiliation, church attendance, self-assessed religiousness, and religious education in childhood. For the substances, nicotine, alcohol, marijuana, cocaine, ecstasy and abuse of medicines a logistic regression analysis for dicotomic answer was applied. The heavy use of at least one drug during the last month was more frequent among students that did not have a religious education during childhood. The use in the last month of cocaine, ecstasy and (abuse of) medicines was more frequent among those students that had no religion (cocaine and medicines) and that did not have a religious education during childhood (ecstasy and medicines). this study is consistent with previous investigations demonstrating a strong influence of religious variables over the use of drugs among adolescents. Interesting, it was found that no or weaker religious education during childhood was markedly associated with significant more use of drugs during adolescence.2682-90