Long-term evaluation of Class II subdivision treatment with unilateral maxillary first molar extraction

Correction of Class II subdivision malocclusion has long been a challenge for clinicians. Through the years, a wide variety of treatment modalities have been implemented, such as use of asymmetrical headgear¹ ; unilateral Class II elastics coupled with a coil spring, sliding jigs, or tip-back mechanics on the affected side² ; one, three, or four premolar extractions^3,4 ; bimaxillary surgical procedures⁵ ; TADs-supported unilateral molar distalization⁶ ; and a fixed functional appliance.⁷ 

Despite strong clinical interest, few studies on Class II subdivision treatment have been published. Janson and colleagues observed slightly better treatment success rates in asymmetric extraction of three premolars compared with extraction of four.³  Smile attractiveness and buccal corridors did not differ in Class II subdivision subjects treated with one, three, or four premolar extractions.⁴ 

A retrospective study of varying treatment strategies, ie, intermaxillary elastics, extractions, asymmetrical headgear, fixed functional appliance, and orthognathic surgery, demonstrated comparable occlusal outcomes.⁸  Finally, whereas Herbst treatment was similarly successful in various Class II malocclusions, a Class III tendency was more frequently evident in the subdivision group.⁷ 

Recently, unilateral extraction of a maxillary first molar (M1) followed by fixed appliance treatment has also been advocated in a case report with a favorable result.⁹  However, no case series or long-term follow-up studies have yet been published on the treatment of unilateral M1 extraction in Class II subdivision malocclusion.

Therefore, the objective of this study was to assess long-term treatment changes in a sample of Class II subdivision patients treated with one M1 extraction and fixed appliances.

This retrospective study included 20 Class II subdivision subjects (7 males, 13 females; mean age, 13.0 years; SD, 1.7 years) all consecutively treated by one orthodontist with the Begg light-wire appliance in his private practice (Table 1). The inclusion criteria were white race, Class II subdivision (defined as a unilateral Class II ≥ 1/2 premolar width and Class I on the other side), no missing teeth or tooth agenesis including third molars, permanent dentition, no or mild crowding in the mandibular arch, and unilateral M1 extraction on the Class II side. Clinical records were obtained before treatment (T1), after treatment (T2), and 2.5 years posttreatment on average (T3; range, 1.8 years–4.3 years).

The control subjects were untreated Class II subdivision adolescents (4 males, 11 females; mean age, 12.2 years; SD, 1.3 years at the start of the observation period) selected and matched by age from the archives of the Groningen Longitudinal Growth Study (Table 1).^10–12 

All lateral head films were scanned (Epson Expression 1680 Pro, Suwa, Nagano, Japan) and subsequently digitized by the first author using cephalometric software (Viewbox 3.0; dHAL Software, Kifissia, Greece). The landmarks and reference lines used for the analysis are displayed in Figure 1. The same calibrated examiner scored all study casts using the peer assessment rating (PAR) twice, with a 1-week interval between observations. Twelve tracings and PAR scores were randomly selected and repeated at least 2 weeks after the initial series of measurements to evaluate intraobserver reliability. Joint Photographic Experts Group images of patient smiles were imported into image processing software (Image J version 1.48v, US National Institutes of Health, Bethesda, Md) to assess midline asymmetry. Image J was set to define facial and dental midlines and calculate the linear distance between the midlines.

Descriptive statistics (means, standard deviations) were calculated for all cephalometric and PAR measurements. Intraobserver reliability was assessed using the intracluster correlation coefficient (ICC). The effect of the intervention on the parameters of interest was assessed by fitting a mixed linear model in which each outcome of interest was regressed on treatment, time point, patient age, and outcome baseline value. The mixed model accounts for the correlated nature of data arising from the fact that there are multiple observations within patients; the patient was used as the random effect. The level of statistical significance was set at 5%. Statistical analysis was performed with Stata version 13 software (Stata Corporation, College Station, Texas).

The ICC ranged from 0.75 to 0.99, indicating excellent intraobserver reliability. Demographics and summary values (mean, SD) for the study and control groups are presented in Tables 1 and 2. The results from the adjusted analyses for the effects of therapy on the parameters of interest are shown in Table 3.

Superimposition of the mean tracings at all three time points illustrates the overall treatment and growth effects (Figure 2). Six cephalometric variables (U1 to A-Pog, L1/ML, L1 to A-Pog, Li to Sn-Pog′, N-No, ANS-Me/N-Me) showed a statistical significant association with treatment (Table 3).

The adjusted analysis indicated that during therapy, the maxillary incisors were retracted 2.3 mm more than were the control teeth in relation to A-Pog (β  =  2.31; 95% CI: 0.76, 3.87). At T3, the maxillary incisors relapsed in both groups but remained retracted compared with pretreatment standards in the treated adolescents (mean  =  6.0; SD  =  2.5). Treatment also had a significant effect on the mandibular incisor position relative to A-Pog (β  =  1.34; 95% CI: 0.09, 2.59). In the treatment group, the mandibular incisors were protracted 0.9 mm between T1 and T2 (at T2; mean  =  2.4; SD  =  2.1) and 0.4 mm at T3 (mean  =  2.8; SD  =  2.1). In the growth study sample, the mandibular incisors were slightly retracted at T2 (mean  =  0.4; SD  =  1.7 mm) and moved in the opposite direction at follow-up (mean  =  0.8; SD  =  1.7).

In the extraction group, the mandibular incisor to mandibular plane angle increased significantly from T1 to T3 (β  =  5.92; 95% CI: 1.43, 10.41) compared with control, namely, from 98.5° (SD  =  8.2) to 102.1° (SD  =  6.4). The mandibular incisors in the untreated controls proclined after treatment (mean  =  94.9; SD  =  7.2) and remained stable during the posttreatment period (mean  =  94.9; SD  =  7.2).

Regarding soft tissue measurements, the significant maxillary incisor retraction was not accompanied by equivalent changes either in the upper lip position or the nasolabial angle (Table 3). Following the significant treatment effects on L1/ML and L1 to A-Pog, the lower lip appeared significantly more protrusive relative to Sn-Pog′ throughout the observation period in the treatment group (β  =  1.43; 95% CI: 0.18, 2.67). On the contrary, projection of the labrale inferius was decreased in the matched controls by 0.2 mm from T1 to T2 (at T2; mean  =  1.7, SD  =  2.4) and from T2 to T3 (at T3; mean  =  1.9, SD  =  2.0).

The ratio ANS-Me/N-Me was significantly increased from T1 to T3 in the treatment group (β  =  1.63; 95% CI: 0.26, 3.01) indicating an increase in lower face height that we did not consider clinically significant.

Not related to treatment, the nose became significantly more prominent in the treated subjects (β  =  3.97: 95% CI: 0.62, 7.33).

According to the adjusted model, PAR exhibited a significant decrease with treatment compared with the control group (β  =  −6.73; 95% CI: −10.73, −2.73, Figure 3). The average PAR score in the treatment group at T1 was 22.05 (SD  =  7.2), which was reduced to 2.00 (SD  =  2.5) at the end of treatment. PAR reduction for the unilateral molar extraction group exceeded 90%. All but three cases exhibited PAR scores lower than 6 at the follow-up examination. In contrast, there was a mean absolute increase of 1.3 points in the PAR score of the untreated subjects from T1 to T3 (Figure 2).

Initially, in 13 out of 20 adolescents (65%) from the treatment group, the mandibular midline did not correspond with the facial midline. Both dental midlines deviated in five cases (25%), while the remaining subjects (10%) had a shift of the maxillary midline in relation to the facial midline. After removal of appliances, facial and dental midlines were coincident in nine patients (45%). The maxillary-to-facial midline discrepancy was fully addressed by the therapy in thirteen subjects (65%).

Deviation between maxillary midline to face and between dental midlines ranged between 0.3–2.1 mm and 0.5–1.2 mm, respectively, after treatment. At T2, nine individuals appeared to have midlines perfectly aligned with the face. Midline characteristics of the study group are summarized in Table 4.

This is the first clinical study to evaluate long-term changes in Class II subdivision orthodontic patients undergoing unilateral M1 extraction. During the observation period, the maxillary incisors were significantly retracted in the treatment group, whereas comparable changes in lip projection and nasolabial angle did not take place. In contrast, the only previous study on extraction treatment of asymmetrical Class II malocclusion¹³  that cephalometrically compared three-premolar with four-premolar extraction protocols showed no significant changes in maxillary incisor displacement between groups immediately after treatment. The great variability in the amount of retraction in the abovementioned study, probably resulting from varying premolar extraction patterns within the groups, might have contributed to the lack of significant differences. Nevertheless, retraction of the upper lip was significantly greater in cases wherein four premolars had been extracted. As pointed out in our results, proper axial inclination of maxillary incisors was maintained during an average retraction of 2.1 mm relative to the A-Pog line, while the upper lip followed on average 66% of the maxillary incisor movement. In contrast, Stalpers and colleagues¹⁴  found that the upper lip moved half the distance in the same direction as the maxillary incisors in cases of bilateral M1 extractions.

In Class II therapy with extraction of two maxillary first premolars, patients exhibited significantly more retruded maxillary central incisors after treatment than those with premolar extractions in both jaws or nonextraction therapy.¹⁵  Yet, the distance between upper and lower lips to the esthetic line increased highly significantly in all groups regardless of extraction patterns. These investigators noted slight but insignificant increase in the nasolabial angle between the start and end of treatment in all groups. In another two-maxillary-premolar-extraction study, correction of a mean overjet of 8.6 mm was accompanied by significant retraction of the maxillary incisors and labrale superius and an increase in the nasolabial angle.¹⁶  Nonetheless, these authors concluded that the upper lip did not respond uniformly to the distal movement of the maxillary incisors, and therefore potential decrease of lip projection should not be a matter of concern in less severe Class II division 1 malocclusions. In this context, Katsaros,^17,18  based on relatively small changes in the sagittal position of the lips in both extraction and nonextraction patients, claimed that the influence of growth of the chin and nose on the facial profile might be more important than the extractions themselves.

Leveling of the curve of Spee and tooth alignment in treated subjects were accompanied by a significant proclination and protrusion of the mandibular incisors relative to A-Pog and a similar forward movement of the lower lip as measured by the vertical distance from the subnasale-soft-tissue-Pog line. These findings are consistent with the changes observed in dental and soft tissue parameters after the extraction of two M1s.¹⁴  Moreover, the resulting forward movement of the mandibular incisors reduced the required amount of maxillary incisor retraction and apparently enhanced esthetics. Previous analysis of overjet correction with the same low-friction appliances in bilateral M1 extraction cases showed that approximately one-third of the anteroposterior correction was achieved by protrusion of the mandibular incisors.¹⁹ 

With reference to the skeletal measurements, we found a statistically significant increase in lower-face vertical dimension in the treated subjects. However, the 0.1%–0.5% increase in the ratio of lower anterior facial height to total anterior facial height between time points can be considered clinically irrelevant. Given that such vertical skeletal increase was not apparent in the controls, it can be assumed that it most likely resulted from orthodontic extrusive mechanics during incisor retraction and use of Class II elastics rather than normal craniofacial growth and development. In line with our results, lower face height increased in camouflage therapy of Class II division 1 whites having two maxillary first premolars extracted^16,20  and two M1s extracted in the horizontal- and normal-vertical-face-height patients.¹⁴ 

The statistically significant increase in nose length in the treated subjects may be due to the inclusion of older patients and more males than in the control group. It has been previously demonstrated that essential changes in facial convexity, primarily resulting from an increase in nasal prominence relative to the rest of the soft tissue profile, occur earlier in females (at 10–15 years) than in males (15–25 years).²¹ 

The M1 extraction cases underwent an average reduction of more than 20 PAR points, whereas the malocclusion was slightly increased in untreated controls. According to PAR conventions, a minimum change in the weighted PAR score of 22 points is required for a case to be classified as “greatly improved.”²²  Owing to the asymmetrical Class II malocclusion, our study group initially presented only moderate overjet, which diminished the severity of the malocclusion, and did not a potentially greater PAR reduction after treatment. Nevertheless, as indicated by the improved occlusal outcomes after treatment, the patients benefited substantially from treatment with an M1 extraction.

Similar to past studies on classification of Class II subdivision malocclusion,^13,23  midline asymmetry was most commonly located in the mandibular arch. At T2, maxillary and mandibular midlines were harmonized with the midline of the face in approximately half the subjects. Recent research on smile esthetics has demonstrated maximum acceptable maxillary midline-to-face discrepancies ranging from 2.9 mm to 3.3 mm.^24–26  Additionally, the limit of acceptability for the maxillary-mandibular midline deviation has been estimated to be between 2.1 mm and 3.6 mm.^24–26  In view of these results, it can be postulated that midline esthetics was promoted in the treatment group.

Our investigation presents certain shortcomings, mainly related to the sample characteristics. First, it may be argued that the study group included a relatively small number of subjects, which resulted, in some cases, in imprecise estimates as to the associated confidence intervals range from clinically significant to nonsignificant effects. Second, to allow discrimination of the treatment outcome from normal growth, we used historical control data representative of the general Dutch population; however, use of historical controls can be problematic. As factors such as living standards, lifestyles, and nutrition change across time periods, the comparability between the historical and contemporary samples might be questionable. For example, differences in the general level of nutrition, texture of foods, frequency of eating events,²⁷  and infant feeding methods²⁸  may affect dental arch development. On the other hand, it would have been unethical to recruit controls by deferring treatment until a later time. Prospective comparative studies of M1 extraction with other Class II subdivision treatment approaches may increase our understanding of the management of asymmetrical Class II malocclusion.

• Unilateral M1 extraction in Class II subdivision malocclusion may yield favorable long- term occlusal outcomes. Posttreatment changes in midline esthetics and soft tissue profile are considered acceptable.

The authors would like to thank Dr Demetrios J. Halazonetis, Athens, Greece for the implementation of the cephalometric analysis used in this study.