Material and methods

This study was approved by the ethics in research committee of Bauru Dental School at the University of São Paulo in Brazil.

The sample was selected from the files of the Department of Orthodontics at Bauru Dental School, University of São Paulo. Initial and final dental study models of 71 patients who had type 1 Class II subdivision malocclusions (complete Class II on 1 side and Class I on the other side) and were treated with fixed appliances with a 4 or a 3 premolar extraction protocol were selected. The sample was then divided into 2 groups. Group 1 consisted of 40 patients treated with 4 premolar extractions with an initial mean age of 13.37 ± 1.31 years (range, 10.25-15.91 years), and group 2 consisted of 31 patients treated with 3 premolar extractions (2 maxillary premolars and 1 mandibular premolar on the Class I side) with an initial mean age of 14.44 ± 2.83 years (range, 10-24.25 years). The selection criteria were a full Class II molar relationship on 1 side and a Class I molar relationship on the other side, and the presence of all permanent teeth up to the first molars. The additional selection criteria were (1) no previous orthodontic treatment, (2) no history of facial trauma that could have altered growth of the apical bases, and (3) complete orthodontic records. To select the sample, only the initial anteroposterior relationship was considered. No other dentoalveolar or skeletal characteristic was considered.

Orthodontic mechanics included fixed edgewise appliances, with 0.022 × 0.028-in conventional brackets and the usual wire sequence characterized by an initial 0.015-in twist flex or a 0.016-in nickel-titanium alloy archwire, followed by 0.016, 0.018, 0.020, and 0.021 × 0.025 or 0.018 × 0.025-in stainless steel archwires (3M Unitek, Monrovia, Calif). Accentuated and reverse curve of Spee were used to correct the deepbites. Rectangular wires and elastic chains were used for en-masse retraction and to correct overjet and Class II canine relationship. Extraoral headgear was used to correct the Class II anteroposterior relationship on the Class II side and to reinforce anchorage to maintain the Class I relationship on the Class I side, in the 4 premolar extraction protocol. In the 3 premolar extraction protocol, extraoral headgear was used to reinforce anchorage to maintain the molars in Class II and Class I relationships on the respective sides. When necessary, Class II elastics were used in the 4 premolar extraction group to help obtain a Class I molar relationship on the Class II side, whereas in the 3 premolar extraction group, this procedure was used to help maintain the original molar relationship.

To compare the efficiency of the treatment protocol in each group, the pretreatment and posttreatment occlusal results were assessed on dental casts using the Peer Assessment Rating (PAR) occlusal index, according to the American weightings suggested by DeGuzman et al, and the time spent in the treatment of each group was calculated from the clinical charts. A form was used to calculate the PAR index on the pretreatment and posttreatment study models of each patient ( Fig ). The PAR occlusal index evaluates posterior occlusion, overjet, overbite, dental midline, and crowding.

PAR data collection form.

The amounts of initial and final dental midline deviation and improvement of midline deviation correction were also evaluated with a caliper (Absolute IP67 model; Mitutoyo, Kawasaki, Japan).

Treatment efficiency (treatment efficiency index, TEI) was assessed using an index defined as the rate between the percentage of occlusal improvement (PCPAR) by the treatment time (TT) in months, expressed as TEIPAR = PCPAR/TT ( Table I ). Treatment efficiency increases when a greater percentage of occlusal improvement is associated with a shorter treatment time.

Twenty-two pairs of dental study models were randomly selected and remeasured by the same examiner (T.B.) a month later. The random errors were calculated according to Dahlberg’s formula (Se ² = Σd ² /2n), where Se ² is the error variance, and d is the difference between 2 determinations of the same variable. The systematic errors were estimated with dependent t tests at P <0.05.