Our aim is to identify predictors of first-time ecstasy use in a prospective study among young people at risk. As part of the multidisciplinary Netherlands XTC Toxicity Study (NeXT), we monitored 188 subjects aged up to 18 years who were ecstasy-naive at baseline but seemed likely to start taking ecstasy in the near future. After an 11- to 26-month follow-up period, 160 respondents remained (85.1%; mean age 21.0 years, 58.1% females): 65 who took ecstasy at least once (ecstasy users) and 95 non-users. At baseline and four times during follow-up, respondents completed self-report questionnaires. Cox regression analysis was used to examine the effects of baseline respondent characteristics on incident ecstasy use. Development of peer group ecstasy use was analyzed by logistic regression. Intention to use ecstasy, low education, and current weekly cannabis use independently increased the hazard rate for first ecstasy use. Although ecstasy use among peers at baseline was not a predictor, the proportion of ecstasy users with ecstasy-using peers increased markedly during the study. Our results suggest that targeted prevention activities should focus in particular on young people who have strong intentions to take ecstasy, especially if they are also regular smokers of cannabis. (Contains 2 tables and 1 figure.)