The aim of this study was to evaluate, over an 18-month period, the clinical performance of a self-etch adhesive [Transbond Plus Self Etching Primer (SEP), 3M Unitek] compared with a conventional adhesive that employs the etch-and-rinse approach (Transbond XT, 3M Unitek). One operator, using the straight-wire technique, bonded 567 metallic brackets to the teeth of 30 patients (age range 12–18 years) in a way that patients acted as self-control. The brackets were bonded following the manufacturers’ instructions except for the fact that the self-etch system was brushed for a longer time than recommended (10–15 seconds) since previous investigations have reported that prolonged application times can improve the bonding efficacy of self-etch systems to enamel. The failure modes were visually classified into: adhesive-enamel, adhesive-bracket, and cohesive failure. The survival rates of the brackets were estimated by Kaplan–Meier and log-rank test (P < 0.05). The failure rates of the self-etch and conventional adhesives were 15.6 and 17.6 per cent, respectively. No significant differences in the survival rate were observed between the materials (P > 0.05). Most of the failures were cohesive and at the adhesive–enamel interface. There was no difference in the fracture debonding mode. These findings indicate that Transbond Plus SEP can be safely used for orthodontic brackets since the survival rates are similar to the conventional Transbond XT.