To evaluate the incidence of failure of brackets bonded with and without primer. A single-operator, cross-mouth, randomized controlled trial (RCT). The Orthodontic Department at the Postgraduate Dental Education Centre, A-rebro, Sweden. Ethical approval was granted by the Regional Ethical Review Board, Uppsala, Sweden. The protocol was not published before trial commencement. Fifty consecutive patients requiring bimaxillary orthodontic treatment with fixed appliances and with an equal number of teeth on each side of the dental arch, were included in this RCT. A cross-mouth methodology was applied. In each patient, two diagonal quadrants (i.e. upper right and lower left, or vice versa) were randomly assigned to the primer group (control group) and the contralateral diagonal quadrants to the non-primer group (experimental group). The randomization process was as follows: A computer-manufactured block-randomization list was acquired and stored with a research secretary at the Postgraduate Dental Education Centre. Each time a patient gave consent, the secretary was contacted by e-mail, and information about which quadrants were to be bonded with and without primer was obtained. All incidents of bracket failure and debonding noted in patient records during the 2012-14 observation period were compiled by the other co-author, whom was blinded to the study and did not perform any orthodontic treatment on the study patients. Number of bracket failures over 18 months. Failure rate without primer was 5.5 per cent and with primer 3.1 per cent; P = 0.063, odds ratio (OR) 1.89 [95% confidence interval (CI) 0.97-3.68] in the adjusted model. Younger ages (10-13 years), boys, and mandible were significantly associated with higher failure rates. Interaction tests indicated that younger patients had significantly higher failure rates without (12.1 per cent) than with primer (4.1 per cent), P < 0.001, OR 3.51 (95% CI 1.93-6.38) in the adjusted model. No failure rate differences between study settings were found for older patients (14-18 years). The difference between two groups was powered at 5 per cent. Some clinicians may consider a difference less than 5 per cent clinically significant. Bonding Victory Series (TM) brackets with Transbond (TM) XT with or without Transbond (TM) MIP primer seems overall to work equally well in a clinical setting, except in younger children where lower failure rate was found in the primer setting.