A 3-year longitudinal prospective study was conducted to compare the incidence of substance use disorders (SUD) and non-substance use disorders (NSUD) among ecstasy users and two control groups: one of cannabis users and the other of non-drug users. The 5-HTTLPR polymorphism related to NSUD was also studied. A total of 94 subjects were included: 37 ecstasy users, 23 cannabis users and 34 non-drug users. SUD and NSUD disorders were diagnosed according to the fourth edition of the Diagnostic and Statistical Manual of Mental Health Disorders criteria using the Psychiatric Research Interview for Substance and Mental Disorders. Incidence Rates (IR) are presented. The 5-HTTLPR polymorphism was analyzed. Hardy-Weinberg equilibrium was studied. The results of the study showed that the highest IR of SUD was cannabis abuse/dependence in both the ecstasy (IR: 48.6/100 person-year) and cannabis (IR: 2.5/100 person-year) groups. There were no new cases of SUD in non-drug users at follow-up. The highest IR of NSUD was primary mood disorder in both the ecstasy (IR: 4.2/100 person-year) and in the non-drug (IR: 1.3/100 person-year) groups (P < 0.282). There were no new cases of NSUD in the cannabis group at follow-up. 5-HTTLPR polymorphism was associated with lifetime of primary mood disorders in ecstasy group (P = 0.018). Ecstasy use was associated with a higher rate of cannabis abuse/dependence disorders and mood disorders than cannabis use. In the ecstasy users, 5-HTTLPR polymorphism may result in a high vulnerability to primary mood disorders.