Mandibular changes secondary to serial extractions compared with late premolar extractions and controls


Variations in treatment times for serial extraction and late premolar extraction patients may be due to differences in the time needed to flatten the occlusal curves. In this study, we compared tooth tipping and occlusal curves in patients treated by serial extractions or late premolar extractions with untreated controls.


Mandibular dental casts and cephalometric radiographs were collected from 90 subjects (30 Class I control subjects, 30 patients with serial extractions, and 30 with late premolar extractions) at 3 time points: T0, baseline for the controls and serial extraction patients; T1, after natural drift and preorthodontics for the controls and the serial extraction patients, and pretreatment for the late premolar extraction patients; and T2, after comprehensive orthodontic treatment for the serial extraction and the late premolar extraction groups. The long axes of the central incisor, canine, and first molar to the palatal plane were measured on digitized headfilms to determine the direction and the amount of tipping between the time points. Three occlusal curves were measured by sphere fitting cusp-tip landmarks on digitized mandibular casts.


From T0 to T1, incisors and canines in the patients with serial extractions tipped distally. Molars at T1 in the patients with serial extractions were tipped forward more than in the late premolar extraction patients and the controls. From T1 to T2, canines and molars in the patients with serial extractions were uprighted.


Serial extractions produce steeper occlusal curves and distal tipping of the incisors and canines after drift (T1). Posttreatment (T2) occlusal curves in the patients with serial extractions are steeper than in the late premolar extraction patients and controls (except for the curve of Spee). After the serial extractions, orthodontic treatment included incisor and canine proclination, with molar uprighting and occlusal curve flattening.


  • Teeth and arch form determine occlusal curvature, an important treatment goal.

  • Serial and late premolar extractions produce different occlusal curvatures.

  • Serial extractions (SE) had smaller radii overall and steeper curvatures at T1 and T2.

  • Incisors and canines were tipped distally in the SE group; molars were tipped mesially at T0-T1.

  • Orthodontics tipped incisors and canines mesially; molars were tipped distally at T1-T2 in the SE group.

Malocclusions with tooth size-arch length deficiencies (TSALD) can be treated either with or without extractions. A nonextraction mixed dentition plan may preserve the leeway space to resolve crowding. Alternatively, expansion therapy can be used; this may move teeth into abnormal stress positions, altering muscle pressure and leading to relapse. Serial extractions (SEs) have been used to resolve severe TSALD in the mixed dentition and need systematic planning to limit the disadvantages and to facilitate the best results.

The earliest documented attempts of SEs were in the mid-18th century. Treatment with SEs was repopularized by Tweed, Dewel, and Kjellgren in the 1950s and 1960s. The reported advantages of SEs include (1) physiologic tooth movement, (2) reduced active treatment time, (3) reduced retention time, (4) reduced load on anchorage units, and (5) better maintenance of the alveolar bone and periodontal tissues. Reported disadvantages include (1) deeper overbite, (2) lingual tipping of incisors, (3) potential scar tissue buildup in the extraction space, (4) creation of a diastema and increased spacing, and (4) alteration of tongue function. The increased overbite with SEs implies a deeper curve of Spee, defined as the anteroposterior anatomic curve established by the occlusal alignment of the teeth, projected onto the median plane. An exaggerated curve of Spee is frequently observed with a deep overbite. This deepening of the curve of Spee can be further aggravated by tooth drift after SEs.

Another occlusal curve, the curve of Wilson, results from the lingual inclination of the mandibular posterior teeth, when the lingual cusps are lower than the buccal cusps. The maxillary buccal cusps are correspondingly higher than the lingual cusps because of buccal tipping of the maxillary posterior teeth. This is seen clinically as lingually rolled mandibular molars and occurs in many patients with deepbite.

The 3-dimensional (3D) arrangement of dental cusps and incisal edges in the natural dentition has been described in the literature as spherical, since each tooth’s occlusal surface touches a segment making up a sphere. Monson’s sphere, a nonscientific observation—a combination of the curves of Spee and Wilson in 3 dimensions—depicts the spherical arrangement of dental cusps and incisal edges in natural human dentitions, with the occlusal surfaces of all teeth touching a segment of the spherical surface. Monson believed that to achieve well-balanced geometric proportions of the face along with optimum function, the jaws can be related to a sphere whose radius is approximately 4 in, with the center equidistant from the occlusal surface of the teeth and the center of each condyle.

The significance of these occlusal curves is reflected in the scoring of posttreatment dental casts by the American Board of Orthodontics’ Objective Grading System, renamed the Cast-Radiograph Evaluation in 2000. Specifically, the curves are addressed by the marginal ridge and the buccolingual criteria of the Cast-Radiograph Evaluation.

Ringenberg and O’Shaughnessy et al compared SE patients with late premolar extraction (LPE) patients; they found active treatment time reductions of 6 months for SE patients and between 4 and 6 months for LPE patients. LPE becomes a treatment option when the permanent dentition has severe crowding requiring comprehensive orthodontics to align the teeth and close the extraction spaces. Since flattening the occlusal curves is a goal of orthodontic treatment, it is possible that the extra time needed to achieve this might explain these differences in SE and LPE treatment times. Because we found no evidence on how SEs affect occlusal curves in our literature search, we therefore studied how SEs and LPEs affect these curves in a retrospective sample of patients at various stages of treatment.

Our first aim was to measure serial changes in the angulations of mandibular teeth after SEs and compare them with those in Class I control subjects and LPE subjects. Our second aim was to measure the occlusal curves of mandibular dentitions after SEs and compare these with Class I control subjects and LPE subjects.

Material and methods

This study was approved by the research ethics board at the University of British Columbia (H12-02568). The sample consisted of 90 subjects following the inclusion and exclusion criteria of O’Shaughnessy et al. The following exclusion criteria were imposed: no anteroposterior discrepancies, no decreased vertical facial heights, no anomalies (eg, missing teeth, clefting), and no excessive deepbite. Treatment records for 30 SE and 30 LPE patients treated between 2007 and 2012 in a specialty pediatric-orthodontic practice were screened according to the following inclusion criteria: Angle Class I, extraction of first premolars in each quadrant, and the availability of pretreatment and posttreatment records. The control group of 30 Class I cases was drawn from the longitudinal sample available at Oregon Health and Sciences University (American Association of Orthodontists Foundation Legacy Fund Web site, 2013); the inclusion criteria here were untreated Angle Class I patients with complete records.

Each of the 3 sample groups contained 15 male and 15 female subjects. The 30 SE subjects had their deciduous canines extracted before the initial records were taken. The serial extraction sequence proceeded with premolar extractions or enucleation before permanent canine eruption. There was a minimum delay of 1 year for physiologic drift after the extractions. The orthodontists prescribed SEs if the patients had significant crowding in the mixed dentition. The 30 LPE subjects were selected from permanent dentition patients with severe crowding. Fixed appliances were placed within 3 months of LPEs. The 30 control group subjects were matched to the 30 SE subjects at baseline (T0). The initial records, according to sex and dental age, conformed to the method of Demirjian et al and Demirjian and Goldstein. Subjects in the 2 treatment groups (SE and LPE) were matched only for sex and not for severity of the initial malocclusion, because the timing did not coincide.

Data were collected from both the cephalometric radiographs and the mandibular dental casts.

For the SE group, the data were analyzed at 3 times: T0, initial records, after extraction of the deciduous canines but before extraction of the deciduous first molars and first permanent premolars (to which a control subject was dentally matched); T1, postdrift, at the end of the observation period, before comprehensive treatment; and T2, after the completion of orthodontic treatment (final records).

The LPE group’s casts were analyzed at 2 times: T1, initial records (permanent dentition, with the second molars partially or fully erupted), before 4 premolar extractions; and T2, after the completion of orthodontic treatment (final records).

The control group’s casts were analyzed at 2 times: T0, to which each SE subject was dentally matched ; and T1, 2 to 3 years later to mirror the SE group after the period of drift. In total, 210 mandibular casts were digitized and measured.

For the cephalometric measurements, 210 cephalometric radiographs were digitized. The images were imported into Rhinoceros software (Robert McNeel & Associates, Seattle, Wash) for landmark definition and analysis similar to that reported by Yoshihara et al. The palatal plane served as a reference plane for tooth tip and inclination measurement at T0, T1, and T2. The following angles were calculated to determine tooth tip ( Fig 1 ): (1) palatal plane to the long axis of the mandibular incisor, (2) palatal plane to the long axis of the mandibular canine, and (3) palatal plane to the axis formed by the line connecting the mesiobuccal cusp and mesial root apex of the first molar.

Fig 1
Tipping of incisor, canine, and molar teeth represented by changes in their axial angulation relative to palatal plane (ANS-PNS) at T0, T1, and T2.

The mandibular dental casts were digitized using an Ortho Insight 3D scanner (version 4.0; MotionView Software, Chattanooga, Tenn), and the files were imported into the Rhinoceros software. A 3D analysis was carried out using the following cuspal landmarks: the vertices of the first molar cusps (distobuccal, mesiobuccal, distolingual, and mesiolingual); and the incisal edges of all 4 mandibular incisors.

The radii of the spheres corresponded to the depth of the occlusal curvatures; the steeper the curvature, the smaller the radius. Best fits were obtained for the following point schemes ( Fig 2 ).

  • 1.

    The distobuccal and mesiobuccal cusps of the mandibular first molars and the 4 mandibular incisors to represent the curve of Spee (anteroposterior curve, viewed in the sagittal plane).

  • 2.

    Only the distobuccal, mesiobuccal, distolingual, and mesiolingual cusps of the mandibular first molars to represent the curve of Wilson (mediolateral curvature, viewed in the frontal plane).

  • 3.

    All the digitized points (distobuccal, mesiobuccal, distolingual, and mesiolingual of the mandibular first molars, and the incisal edges of the 4 mandibular incisors) similar to Monson’s sphere.

Fig 2
Fitted spheres representing the curve of Spee, curve of Wilson, and Monson’s sphere.

In each case, the radii of 3 fitted spheres were used as outcome measures; a sphere of best fit is determined via an optimization function that minimizes the sum of ([x−xc] 2 + [y−yc] 2 + [z−zc] 2 − r 2 ) 2 , where x , y , and z are the point coordinates; xc , yc , and zc are the sphere’s center; and r is the sphere’s radius.

Statistical analysis

A power calculation was done for detecting differences in the curve of Spee. Assuming equal variances of 0.77 mm according to Marshall et al and a mean difference of 0.6 mm, a power of 80%, and an α = 0.05, we would require 26 patients in each group. SPSS software (version 21.0; IBM, Armonk, NY) was used for all statistical analyses, and the threshold for the statistical significance for all tests was set at P <0.05. Bivariate analyses were used to compare the means of the different measurements for cast data and cephalometric data between the SE patients and their matched controls at T0; between the SE group, their matched controls, and the LPE group at T1; and between the SE and LPE groups at T2 (independent sample t tests or analysis of variance [ANOVA] with the Bonferroni post hoc adjustment).


To evaluate intraexaminer reliability, 10 mandibular casts and 10 cephalometric radiographs were randomly selected for duplicate measurements. The intraclass correlation coefficients for the duplicate cephalometric angular measurements ranged from 0.808 to 0.956, indicating satisfactory intraexaminer reliability. The intraclass correlation coefficients for double recordings of radii in the 10 casts ranged from 0.711 to 0.834, indicating satisfactory reliability.

For the cephalometric data, the SE group ( Fig 3 ), from T0 to T1, showed statistically significant uprighting of the incisor and canine (+4.29° and +6.70°, respectively). The molar angle was reduced by a not statistically significant −1.53°. From T1 to T2, the SE group demonstrated the opposite trend; incisor angulation decreased by 0.07°. There was a statistically significant decrease of 4.38° in canine angulation, with a statistically significant increase in average molar angulation (+3.07°) from T1 to T2.

Fig 3
Tip in the SE group from T0-T1-T2 obtained by cephalometric analysis.

At T1, the SE and control groups had no statistically significant difference for incisor angulation, whereas significant differences were observed between the controls and LPE group, as well as between the SE and LPE groups ( Fig 4 ). Average canine angulation indicated no significant differences between the control and the SE groups at T1. There were significant differences for the LPE group compared with the other 2 groups.

Apr 6, 2017 | Posted by in Orthodontics | Comments Off on Mandibular changes secondary to serial extractions compared with late premolar extractions and controls
Premium Wordpress Themes by UFO Themes