ORIGINAL ARTICLE
Variables affecting orthodontic tooth movement with clear aligners Justin R. Chisari,a Susan P. McGorray,b Madhu Nair,c and Timothy T. Wheelerd West Palm Beach and Gainesville, Fla
Introduction: In this study, we examined the impacts of age, sex, root length, bone levels, and bone quality on orthodontic tooth movement. Methods: Clear aligners were programmed to move 1 central incisor 1 mm over the course of 8 weeks. Thirty subjects, ages 19 to 64, were enrolled, and measurements were made on digital models (percentage of tooth movement goal achieved). Morphometric features and bone quality were assessed with cone-beam computed tomography. Data from this study were combined with data from 2 similar studies to increase the power for some analyses. Results: The mean percentage of tooth movement goal achieved was 57% overall. Linear regression modeling indicated a cubic relationship between age and tooth movement, with a decreasing rate of movement from ages 18 to 35 years, a slightly increasing rate from ages 35 to 50, and a decreasing rate from ages 50 to 70. The final decreasing trend was not apparent for women. As would be expected, the correlation was significant between the percentage of the goal achieved and the cone-beam computed tomography superimposed linear measures of tooth movement. A significant negative correlation was found between tooth movement and the measurement apex to the center of rotation, but bone quality, as measured by fractal dimension, was not correlated with movement. Conclusions: The relationship between age and tooth movement is complex and might differ for male and female patients. Limited correlations with cone-beam computed tomography morphology and rate of tooth movement were detected. (Am J Orthod Dentofacial Orthop 2014;145:S82-91)
T
he use of clear aligners to produce orthodontic tooth movement (OTM) provides an opportunity to measure incremental movement and investigate factors that might affect the rate of movement. The broad principles of OTM are based largely on bone and tissue remodeling, specifically the resorption and deposition of alveolar bone as force is applied. The biology of OTM has proven to be an extremely complex process involving an array of coordinated biochemical reactions, including critical cell signaling pathways and a wide range of cellular differentiation, leading to bone remodeling.1 As the science of bone biology continues to evolve, several theories of OTM have surfaced. The a
Private practice, West Palm Beach, Fla. Assistant professor, Department of Biostatistics, University of Florida, Gainesville. c Chairman and professor, Department of Oral & Maxillofacial Diagnostic Sciences, University of Florida, Gainesville. d Professor, Department of Orthodontics, University of Florida, Gainesville. All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and none were reported. Supported by Align Technology, San Jose, Calif, and the Southern Association of Orthodontists. Address correspondence to: Susan P. McGorray, Department of Biostatistics, Box 117450, Gainesville, FL 32611; e-mail, spmcg@ufl.edu. Submitted, June 2012; revised and accepted, October 2013. 0889-5406/$36.00 Copyright Ó 2014 by the American Association of Orthodontists. http://dx.doi.org/10.1016/j.ajodo.2013.10.022 b
S82
pressure-tension theory has emerged as the most popular concept behind the movement of teeth. Bone remodeling involves an intricate arrangement of coordinated cellular activity leading to bone resorption performed by osteoclasts, followed by bone formation carried out by osteoblasts.2 Dolce and Holliday3 have reported that although the precise biologic response to orthodontic force has not been identified, several hypotheses regarding the mechanisms by which osteoblasts and osteocytes sense this initial mechanical stimulus have been proposed, including strainsensitive ion channels, shear stress receptors, integrin activation, and cytoskeleton reorganization. Three phases of tooth movement have been described in the literature: initial phase, lag phase, and secondary phase.4,5 The secondary stage accounts for most of the tooth movement, and teeth during this period move at a faster, more continuous pace.6 The magnitude and direction of force placed on teeth during OTM, in addition to the length of time these forces are in place, also play critical roles in how teeth move. Forces applied to teeth cause various types of tooth movement depending on the location of the center of resistance of that tooth and the direction in which the force is applied. It is understood that the center of resistance for a given tooth changes based on tooth size,
Chisari et al
number of roots, and amount of tooth root that is submerged in bone. Variability among patients can affect OTM. Factors including age, sex, root length, bone levels, bone density, medications, and certain systemic conditions can have inhibitory, synergistic, or additive effects on OTM.7 The majority of literature on the effects of age on OTM has been completed using animal models. Bridges et al8 reported that a significantly greater amount and rate of tooth movement occurred in younger rats compared with their older counterparts in all 3 phases of tooth movement. Similar findings of the effects of increasing age on the rate and amount of tooth movement have been reported by MisawaKageyama et al9 and Harris.10 There has also been some indication that whereas there is a delay in the onset of tooth movement in adult rats, once the secondary phase of tooth movement is reached, the movement becomes equally efficient among the 2 age groups.11 The effect of age on OTM clearly exists and is likely due in part to a decreased biologic response. Although there is individual variability from patient to patient, a direct difference in OTM between the sexes has not been shown in the literature. Medications with pharmacologic effects can impact the cells targeted in OTM. Some of these medication classes include bisphosphonates, estrogens, NSAIDS and other analgesics, corticosteroids, calcium regulators, and supplements.7,12,13 There is little human experimental data on the effects of medications on OTM and limited information from animal models. However, knowing the biochemical action of these medications has led to concerns regarding how they can affect orthodontic treatment.13 Any medication that interferes with or alters bone biology might impact the rate of tooth movement. Systemic factors or nutritional deficiencies affecting bone metabolism have also been found to alter OTM. Specifically, diseases of bone can have a significant impact on the rate of tooth movement as well. Reduced or complete lack of osteoclast function can lead to a condition known as osteopetrosis, characterized by sclerosis of the skeleton and inhibited tooth movement and eruption. On the other hand, in Paget's disease, uncontrollable bone turnover occurs because of the overactivity of osteoclasts.14 Since OTM stimulates an inflammatory process in the periodontal ligament and surrounding tissues, it is thought that any chronic inflammatory disease such as thyroiditis, asthma, and even allergies can affect the movement of teeth.15 Other variables that might be of significance in OTM are root length, bone levels, and the density or quality of bone. Age-related decreases in bone turnover as well as a
S83
relatively higher level of bone density have been documented.8,13 Alveolar bone levels, on average, decrease over time, impacting OTM by changing the center of resistance. The center of resistance of a tooth is largely influenced by its surroundings, particularly in regard to root morphology, bone levels, and bone quality.16 Thus, patients with alveolar bone loss or abnormally long roots will have centers of resistance farther from the point of force application (more apically). Alternatively, the more the root tapers, the more the center of resistance moves coronally.17 Clear aligners with sequentially programmed movement provide an excellent model for investigating tooth movement. A single tooth can be isolated, and frequent measurements made with polyvinyl siloxane or digital impressions provide incremental information regarding the pattern of movement. For example, McGorray et al18 characterized the weekly pattern of tooth movement using this model over 8 weeks, along with subsequent relapse. Kravitz et al19 compared actual tooth movement with aligners with predicted movement over the course of treatment. Aligner treatment is now a commonly prescribed modality for OTM in adolescents and adults, and a better understanding of the pattern of movement and factors that influence movement could lead to more efficient treatment. The purposes of this study were to better characterize the pattern of tooth movement with clear aligners with programmed movement over 8 weeks and to examine the influence of age, sex, root length, morphometric measurements, and bone quality on the rate of OTM. MATERIAL AND METHODS
The design for this study was similar to 2 previous studies that investigated specific aspects of tooth movement with clear aligners.18,20 Approval was obtained from the University of Florida Institutional Review Board for the Protection of Human Subjects. This project was a prospective single-center clinical trial involving subjects of 2 age groups with minor incisor malalignments, who were otherwise healthy and would be undergoing orthodontic treatment. The first group included 7 men and 8 women between the ages of 18 and 35 years, inclusive. The second group consisted of 5 men and 10 women 50 years of age and older. Throughout this article, this study will be called Study 3. Study 1 investigated the role of relaxin in tooth movement and relapse, and has been previously described.18 No difference in tooth movement was detected when comparing those who received relaxin injections with those who received placebo injections; thus, data from both groups in this study were combined, yielding a sample size of 37 subjects. Cone-beam computed
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
Chisari et al
S84
tomography (CBCT) imaging was not performed in Study 1. Study 2 was similarly designed but included CBCT imaging and examined the role of aligner material fatigue in tooth movement.20 Subjects in this study received a new aligner every week rather than every 2 weeks as in Study 1. No difference was detected in total tooth movement when comparing the weekly aligner and the biweekly aligner groups. For studies 1 and 2, the biweekly tooth movement goal was 0.50 mm, for a total 8-week goal of 2 mm. Study 3 was designed to broaden the age range and to be used in conjunction with the previous studies to examine the role of age and other factors in tooth movement. In the 3 studies combined, the total number of subjects was 82. All subjects were in good health and had acceptable malocclusions as defined in the inclusion criteria, which have been described in a previous study.18 Once a subject was accepted into the trial, the right or left maxillary central incisor was selected as the target tooth. The selection was based on the target tooth's not being blocked out by the adjacent teeth to allow a total anteroposterior movement of 1 mm. Tooth movement was accomplished using a series of 4 maxillary aligners (Invisalign; Align Technology, San Jose, Calif), each programmed in increments of 0.25 mm of anterior movement of the central incisor being studied, as described above. Aligners were collected every 2 weeks from each subject, and new aligners were dispensed. The study termination visit and final time point for data collection was at week 8. Polyvinyl siloxane impressions were taken weekly and sent to Align Technology for scanning to create 3-dimensional models. Tooth movement measurements from baseline through week 8 were made from each scanned model using ToothMeasure proprietary software (Align Technology). CBCT measurements and fractal analysis were completed using a combination of software including InVivo (Anatomage, San Jose, Calif), ImageJ (National Institutes of Health, Bethesda, MD), and Tact Workbench (Wake Forest University, Winston-Salem, NC). The following are definitions of the measurements used in the CBCT superimposition analysis. D U1 (x) refers to the distance between lines drawn through the midpoint of the incisal edges of the superimposed target tooth perpendicular to the anteroposterior axis (the plane of prescribed tooth movement). D U1 (s) is the length of the line connecting the midpoint of the incisal edges of the superimposed target tooth. D Apex refers to the length of a line connecting the change in apex of the superimposed target tooth. Rotation angle is the angle created by the intersection of lines drawn from the midpoint of the incisal edge to the apex of the target tooth. The apex of this angle is considered the center
April 2014 Vol 145 Issue 4 Supplement 1
of rotation. Tooth length refers to the distance from the midpoint of the incisal edge to the apex of the target tooth from the initial computed tomography image. Crown length is the portion of the tooth length that is coronal to the bone. Bone to C-rot is the section of tooth length between the center of rotation and a line connecting the most coronal aspect of the faciolingual crestal bone. These variables are illustrated in Figure 1. All study subjects were instructed to wear the aligner appliance full time. They were allowed to remove the appliance when eating, drinking, or brushing their teeth. Their medication and medical histories were taken initially. Each subject recorded aligner wear time in a diary format. At the conclusion of the study, the participants were routinely treated orthodontically with clear aligners. To determine subject eligibility, 2 visits were required. The first visit was designed to identify potential subjects with malocclusions needing minor incisor alignment of at least the maxillary incisors. Those with medical conditions or intraoral problems, including significant periodontal disease, chronic daily use of any nonsteroidal or anti-inflammatory medication, current smokers, or history of significant cardiac disease, uncontrolled hypertension, bleeding disorders, or renal disease, were also excluded. Subjects who were determined to be eligible based on these procedures proceeded to the next visit. The screening visit was designed to finalize the subject's eligibility and collect initial records. The following procedures were performed at this visit: impressions were taken with polyvinyl siloxane for preparation of the Invisalign appliances, impressions were sent to Align Technology after confirmation of eligibility, and intraoral and extraoral photographs and CBCT imaging were done. For women, a negative urine pregnancy test immediately before this procedure was required. After the investigator (T.T.W.) reviewed all subject information to confirm eligibility, the subjects were enrolled into the study and assigned a unique number. At the first study visit (week 0), the first aligner was delivered to each subject. The acceptable visit window for weeks 1 through 8 was 6 1 day, and all 30 treatment subjects successfully satisfied this requirement. During the study visits of weeks 1 through 8, the following procedures were performed: intraoral clinical examination, maxillary occlusal and frontal photographs, and polyvinyl siloxane impressions. In addition, during the study visits of weeks 2, 4, and 6, the aligner (used during the previous 2 weeks) and the wear diary were collected, and the next aligner and diary were dispensed. At the study termination visit, week 8, the same procedures were performed, and CBCT imaging of the maxilla took place.
American Journal of Orthodontics and Dentofacial Orthopedics
Chisari et al
S85
Statistical analysis
Fig 1. Superimposed CBCT measurements. Blue is initial tooth position, and red is final tooth position (figure from thesis of Carl Drake; Gainesville: University of Florida; 2010).
Weekly anteroposterior movement of the target tooth was recorded with polyvinyl siloxane impressions. These impressions were sent to Align Technology, and digital models were created so that OTM could be measured using Align Technology's ToothMeasure software. The digitized model fabricated each week was superimposed on the baseline digital model, taken at week 0, according to the best fit of unmoved teeth, particularly the posterior dentition. The most central portion on the facial surface of the clinical crown of the target tooth, referred to as the centroid, was determined, and subsequent tooth movements in all dimensions were measured from this point at each study visit. The same investigator (J.R.C.) measured the digital models for all 30 subjects. CBCT scans of each subject were performed at the screening visit and at the study termination visit (week 8). Using the InVivo software, the images were superimposed on each other and registered at the curvature of the palate in addition to other stable maxillary structures. Measurements recorded from these superimpositions are shown in Figure 1. A fractal dimension score was calculated for each subject, representing the quality of the bone.21 Higher fractal dimensions correspond to greater morphologic complexity of the bone. Weekly wear time was calculated for each subject, with mean, median, and standard deviation of weekly wear time used to characterize each participant's wear patterns.
Orthodontic tooth movement was quantified using descriptive statistics for the digital model analysis. Chi-square tests of equality of proportions and analysis of variance (ANOVA) were used to compare subject characteristics and results over the 3 studies. To standardize the studies, the primary outcome was the percentage of tooth movement goal achieved over 8 weeks. Spearman correlation coefficients were estimated to examine the relationship between that outcome variable and age, morphometric measurements, and compliance. Median weekly hours of wear were used to represent compliance during the study, and this would not be overly influenced by a week of limited wear or excessive wear and would better represent typical weekly compliance. Linear regression modeling was used to examine the relationship between the percentage of tooth movement goal achieved and multiple covariates. Based on the R2 value (percentage of variability in the outcome explained by the model), the best 1, 2, 3, and so on variable models were determined. Model building was concluded when additional variables did not significantly improve the previous model. Potential interactions between covariates and the influence of outliers were also examined. A P value less than 0.05 was considered statistically significant, and analyses were performed using SAS software (version 9.1.3; SAS Institute, Cary, NC) and R software (version 2.15; R Foundation for Statistical Computing, Vienna, Austria). RESULTS
The demographic characteristics of the subjects for the 3 studies can be found in Table I. No significant differences were detected when comparing the studies for sex, race, compliance, and percentage of tooth movement goal achieved. Tooth movement goal and age were not compared because they differed in the designs of the studies. The pattern of tooth movement over 8 weeks is shown in Figure 2. Most tooth movement occurred in the first week of the 2-week wear cycle. Table II presents summary statistics and group comparisons for the demographic variables and planned biweekly tooth movement. The percentage of the goal achieved did not differ significantly by sex or race. Although it was not statistically significant (P 5 0.06), subjects with a smaller goal had a higher mean percentage of goal achieved, 62%, compared with 54% for those with a planned movement of .50 mm. There was no significant correlation with age, according to the Pearson correlation coefficient estimate of 0.021 (P 5 0.90). However, examining this correlation separately for men and women, different patterns were suggested, with
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
Chisari et al
S86
Table I. Demographic characteristics and descriptive statistics Sex, female (%) Race (%) White Black Other Tooth movement 2-week goal 0.25 mm (%) 0.50 mm (%) Age (y), median Mean (SD) Range Compliance, median hours/wk Mean (SD) Range % goal achieved, median Mean (SD) Range
Study 1 (n 5 37) 70
Study 2 (n 5 15) 60
Study 3 (n 5 30) 60
76 13 11
54 13 33
77 10 13
0 100 27.1 26.7 (5.1) 18.6-40.5 (n 5 34) 138 (24) 28-162 59.0 53.2 (16.2) 11.0-77.5
0 100 23.1 25.1 (4.9) 20.0-35.0 (n 5 15) 144 (14) 104-157 58.0 55.4 (15.0) 17.5-73.0
100 0 43.5 40.7 (15.5) 19.0-68.0 (n 5 30) 147 (8) 128-160 68.0 61.6 (20.0) 11.0-93.0
Total (n 5 82) 65 72 12 16 37 64 27.9 31.5 (12.3) 18.6-68.0 (n 5 79) 142 (18) 27-162 61.8 56.9 (17.7) 11.0-93.0
P value 0.63* 0.32* y
y
0.13z 0.17z
*Chi-square test of equality of proportions; ynot tested because of different study designs; zANOVA.
Fig 2. Mean percentage of tooth movement goal achieved for the 3 studies. Solid line indicates tooth movement goal (assuming linear movement) over the 8 weeks.
the women having a positive correlation coefficient estimate, 0.09 (P 5 0.52), whereas the men had a negative correlation coefficient estimate, 0.23 (P 5 0.25). We also did not detect a significant correlation between compliance (measured by median weekly hours worn) and percentage of goal achieved, with the Pearson correlation coefficient estimate of 0.06 (P 5 0.59). This relationship is displayed in Figure 3. Summary statistics for the CBCT measurements and their correlations with percentage of goal achieved are given in Table III. High correlation would be expected
April 2014 Vol 145 Issue 4 Supplement 1
between percentage of goal achieved and DU1(x), DU1(s), and rotation angle, since these should correspond well with the model-based tooth movement percentage of goal achieved measurement. Significant correlations were not noted for most morphometric measures. A negative correlation with percentage of goal achieved was identified for the apex to center of rotation measurement; this is illustrated in Figure 4. Linear regression modeling was used to examine the relationship between percentage of tooth movement goal achieved and covariates. Covariates considered for
American Journal of Orthodontics and Dentofacial Orthopedics
Chisari et al
S87
Table II. Descriptive statistics and comparisons of
demographic variables and planned treatment goals Variable Sex Male Female Race White Black Other Goal 0.25 mm 0.50 mm
n
Mean
SD
Range
P value *
29 53
54.4 58.2
16.1 18.5
17.5-80.0 11.0-93.0
0.35
59 10 13
58.0 58.4 50.5
16.1 19.4 22.8
11.0-93.0 24.0-85.3 15.0-77.5
0.37
30 52
61.6 54.1
20.0 15.8
11.0-93.0 11.0-77.5
0.06
*Two-sample t test or ANOVA.
inclusion in the model included age, age2, race, study biweekly goal (0.25 or 0.50 mm), compliance, and study number (1, 2, or 3). Sequentially, we determined the best models consisting of 1 to 5 variables. When 2 or more variables were selected for a model, a potential interaction between these variables was also considered for inclusion in the model. This is appropriate if the combined impact of the 2 variables differs from what would be expected from their individual contributions (ie, synergy or antagonism). Age and age2 were included to have more flexibility to identify the relationship between age and tooth movement. The best 1-variable model consisted of the biweekly goal (R2 5 0.042). The best 2-variable model included age and age2 (R2 5 0.084). The best 3-variable model included age, age2, and their interaction (ie, age3) (R2 5 0.14). No additional variables significantly improved the model. The fitted models and the actual data points are displayed in Figure 5. Because this model includes intercept, age, age2, and age3, it is considered a cubic age model. The overall cubic age model suggests that tooth movement slows from ages 18 through 35 years, increases slightly until approximately age 50, and then declines. We also fit this model separately for men and women. Model fit statistics and parameter estimates are provided in Table IV, and the predicted tooth movements for men and women are also illustrated in Figure 5. The decline in tooth movement after the age of 50 is not apparent in the model, with only the data from women. DISCUSSION
The combined data indicate that despite having aligners programmed to move 1 central incisor 1 mm labially (0.25 mm per aligner), on average only 57% of that movement was achieved. This discrepancy might be due to several reasons. It has been postulated that a greater percentage of tooth movement would occur if the prescription in each aligner was decreased from 0.5
Fig 3. Correlation between compliance (median weekly hours wear) and percentage of tooth movement goal achieved. Circles indicate observed data, solid line represents best linear fit (Pearson correlation coefficient estimate 5 0.06; P 5 0.59), and dotted line indicates best linear fit with outlier removed (Pearson correlation coefficient estimate 5 0.12; P 5 0.28).
to 0.25 mm. In studies 1 and 2, where each set of aligners was programmed for 0.5 mm of tooth movement, only 54% of the tooth movement was achieved; slightly more, 62%, was achieved with the smaller 0.25-mm goal. As mentioned earlier, the magnitude and direction of force placed on teeth during OTM, in addition to the length of time these forces are in place, can play critical roles in how teeth move. The use of aligners has increased in both the adolescent and young adult populations and among older adults seeking orthodontic treatment in recent years, raising concerns regarding the efficiency of tooth movement, particularly in the older population. Conventional thinking and clinical experience from several studies has led to the belief that the rate of tooth movement decreases with age.8-10 Additional factors, including sex, root length, bone levels, and bone density, can have various affects on tooth movement as well.7 Our results suggest that the relationship between age and tooth movement is complex and can differ depending on sex. Regression modeling supports an overall cubic relationship between OTM and age, represented by an s-shaped curve. Further exploratory analysis showed a quadratic (u-shaped) relationship for women and a more linear relationship for men. These trends might be the consequence of decreased quality of bone (osteoporosis) typically seen in older women. However, data
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
Chisari et al
S88
Table III. Descriptive statistics for CBCT measure-
ments and Spearman correlation coefficient estimates of specified variables with percentages of tooth movement goal achieved (n 5 45)
Variable D U1 (x)
Study 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2
Mean 1.56 0.85 1.09 1.63 1.17 1.33 0.73 0.41 0.52 5.31 4.08 4.49 24.87 22.84 23.52 12.27 11.51 11.76 12.60 11.33 11.76 0.99 1.05 1.03 5.14 5.64 5.47 7.46 5.70 6.29 0.41
SD 0.38 0.37 0.50 0.40 0.49 0.51 0.26 0.14 0.24 1.32 1.63 1.63 2.02 2.09 2.26 0.74 0.93 0.94 1.74 1.87 1.90 0.12 0.22 0.19 1.25 2.24 1.96 2.01 1.57 1.90 0.11
Min 0.80 0.22 0.22 0.80 0.34 0.34 1.32 0.67 1.32 2.70 1.60 1.60 21.67 17.90 17.90 10.84 9.82 9.82 10.56 6.44 6.44 0.71 0.75 0.71 2.89 1.90 1.90 4.59 1.18 1.18 0.25
Max 2.02 1.96 2.02 2.09 2.32 2.32 0.39 0.09 0.09 7.50 8.90 8.90 30.32 26.90 30.32 13.27 13.79 13.79 17.74 14.95 17.74 1.23 1.78 1.78 7.70 10.19 10.18 12.82 9.26 12.82 0.63
3 Total Fractal dimension 2 3 Total
0.49 0.46 1.71 2.11 1.98
0.16 0.14 0.20 0.05 0.23
0.20 0.20 1.37 1.99 1.37
0.87 0.87 2.00 2.21 2.21
D U1 (s) D Apex
Rotation angle
Tooth length
Crown length
Root length
Crown/root ratio
Bone to C-rot
Apex to C-rot
Bone C-rot/ apex C-rot
Spearman correlation with % goal 0.90* 0.69* 0.37* 0.86* 0.56* 0.44* 0.72* 0.21 0.09 0.86* 0.59* 0.47* 0.42 0.08 0.17 0.17 0.07 0.07 0.40 0.12 0.16 0.12 0.13 0.12 0.10 0.13 0.15 0.18 0.27 0.35* 0.08 0.19 0.24 0.25 0.00 0.13
*Significant, P \0.05.
from studies 2 and 3 showed no statistical correlation between age and fractal dimension, a bone complexity and quality indicator. Caution must be used in interpreting the fractal dimension analysis, and additional information such as bone mass or structural properties of the bone might be needed to truly assess bone quality.21 Hormone replacement therapy can also play a role in the differing patterns. Of the many CBCT and digital model measurements examined, significant positive correlations were noted with DU1 (x), DU1(s), and rotation angle. That is, as
April 2014 Vol 145 Issue 4 Supplement 1
tooth movement increased, these measurements increased in both the CBCT superimposition and the digital model analysis. The DU1(s) and rotation angle difference between studies 2 and 3 implies more uncontrolled tipping in Study 2, likely because of the higher programmed tooth movement in that study. There was a significant negative correlation for the measurement apex to the center of rotation, indicating that as tooth movement increased, the apex to the center of rotation measurement decreased. This could be the result of root resorption, but it was more likely due to decreased alveolar bone levels.22,23 Other authors have examined the efficiency of tooth movement with aligners. Kravitz et al19 examined pretreatment and posttreatment locations for 37 subjects, with 401 teeth. They considered all types of movement and the complete treatment period (means of 10 maxillary and 12 mandibular aligners per treatment). Overall, the mean accuracy of tooth movement was 41%. It was expected that this would be lower than our observed 57% because the study was significantly longer and involved multiple types of movement and multiple teeth per subject. Bollen et al24 quantified tooth movement with aligners by assessing the ability to complete treatment (25-35 sequential aligners). Of 51 subjects, only 15 (29%) completed their initial series. This suggests that tooth movement is less than optimal and illustrates the potential usefulness of understanding factors that affect tooth movement. Factors related to tooth movement via archwires and brackets were examined by Dudic et al.25 Thirty subjects, ranging in age from 11 to 43 years, participated; younger subjects (age 15 and younger) appeared to have more tooth movement that those 16 years and older. Multiple teeth per subject were studied, and correlations within subjects were not adequately accounted for, making the results difficult to interpret. No relationship was detected between sex or tooth location and amount of movement. Compliance with the treatment plan is a key element in achieving tooth movement with aligners, and we were surprised that our measure of compliance was not related to the percentage of tooth movement goal achieved. The use of diaries to obtain time of wear requires accurate and truthful recording by the participants. The subjects in these studies appeared to be highly motivated, and their compliance level was uniformly high, with the exception of 1 outlier. Similar compliance rates have been reported in a similar study.19 Compliance might play a role in tooth movement in less motivated patients, but because of the uniformity of our sample, we were unable to detect this.
American Journal of Orthodontics and Dentofacial Orthopedics
Chisari et al
S89
Fig 4. Correlation between tooth movement and apex to the center of rotation measurement. Dots represent observed data, and line indicates best linear fit (Pearson correlation coefficient estimate 5 0.35; P 5 0.0184).
Fig 5. Cubic age model representing age vs the amount of tooth movement for the 3 data sets. Solid line, overall cubic model; dotted line, model using only data from women; dashed line, model using only data from men; solid squares, observed data for men; open circles, observed data for women.
We were fortunate to be able to combine data from 3 studies to investigate factors associated with tooth movement. Each of the component studies was designed to address a specific question regarding aligners, and the
similar designs, patient populations, and outcome measures allowed the combined data to provide a sufficient sample size to address our larger question. Care was taken in combining data, with separate comparisons of subject characteristics in the 3 studies and acknowledgment and testing (in our regression modeling) for the impact of unidentified study-specific components that could have affected tooth movement. No unidentified study-specific components were detected. The advantages of using the aligner model for tooth movement include ease of patient wear and patient recruitment, reliable outcome measures, and the benefit of a human model. Of particular interest is the ability to measure tooth movement on an incremental time scale. In all 3 substudies, stair-step movement was observed, with most movement occurring during the first week of the 2-week aligner time period. We have previously investigated whether this was due to a lack of continuous force caused by aligner fatigue, but we did not find this to be the case.20 Wear time appeared to be uniform throughout the 8-week study. It would be of interest to further characterize the movement over the 2-week period with more frequent measurements. As a consequence of not achieving biweekly tooth movement goals, accuracy decreased, resulting in poorer aligner fit over time. This can result in revised treatment planning and longer treatment times. There are several limiting factors associated with studying tooth movement with this model. Patient
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
Chisari et al
S90
Table IV. Linear regression model fit and parameter estimates (parameter estimates and P values) Variables in the model Model Overall Women Men
2
R 0.20 0.20 0.24
Overall fit (P value) 0.0006 0.0116 0.07
Intercept 252.9 (\0.0001) 145.7 (0.18) 235.7 (0.0041)
compliance is the single most important factor contributing to the amount of tooth movement seen. Unfortunately, even recording wear time on a daily basis has limited value. Clinical experience has suggested that a more continuous force enhances tooth movement. Removing the aligners for eating and brushing results in an interrupted force. Another factor, addressed by other studies, is the loss of anchorage of adjacent teeth during tooth movement. Whereas only 1 target tooth was programmed for tooth movement, movement of adjacent teeth is expected and can falsely minimize the amount of tooth movement of the target tooth. Polyvinyl siloxane impressions can have inaccuracies because of operator error or subject movement. Lastly, the use of aligners to produce tooth movement might be considered a limitation because it is not a perfect system for producing tooth movement. The ability to make an accurate prediction regarding how teeth will move when a force is applied has long been a challenge for orthodontists. Contributing to this factor is the large degree of variability of OTM seen among patients of the same age group, not to mention those across age groups. Different types of tooth movement, combined with varying biologic responses, make prediction difficult. Although there will always be variability in patients, our results suggest that there are general trends in tooth movement for different age groups, and that these might differ for male and female subjects. CONCLUSIONS
In this single-center prospective clinical trial, we examined the variables affecting tooth movement using a clear aligner model. When all 3 data sets were combined, the mean percentage of tooth movement achieved compared with the tooth movement goal was 57%. The results of regression modeling suggest an overall cubic relationship between OTM and age, represented by an s-shaped curve. Further exploratory analysis revealed a quadratic (u-shaped) relationship for women, indicating an increase in tooth movement in younger and older women, and a more linear
April 2014 Vol 145 Issue 4 Supplement 1
Age 16.23 (\0.0001) 15.49 (0.59) 9.92 (0.0229)
Age2 0.41 (\0.0001) 0.079 (0.80) 0.37 (0.0234)
Age3 0.0033 (0.0001) 0.0001 (0.98) 0.0030 (0.0232)
relationship for men, with decreased movement at older ages. Significant positive correlations were found between CBCT superimposition measurements and digital model measurements of the percentage of goal achieved for the DU1 (x), DU1(s), and rotation angle variables, whereas a significant negative correlation was seen with the percentage of goal achieved and the apex to center of rotation measurement. REFERENCES 1. Masella RS, Meister M. Current concepts in the biology of orthodontic tooth movement. Am J Orthod Dentofacial Orthop 2006; 129:458-68. 2. Proff P, Romer P. The molecular mechanism behind bone remodeling: a review. Clin Oral Investig 2009;13:355-62. 3. Dolce C, Holliday LS. Toward molecular orthodontics. In: Riolo ML, Avery JK, editors. Essentials for orthodontic practice. Ann Arbor and Grand Haven, Mich: EFOP Press; 2003. 4. Kohno, Matsumoto Y, Kanno Z, Warita H, Soma K. Experimental tooth movement under light orthodontic forces: rates of tooth movement and changes of the periodontium. J Orthod 2002;29: 129-35. 5. Krishnan V, Davidovich Z. Cellular, molecular, and tissue-level reactions to orthodontic force. Am J Orthod Dentofacial Orthop 2006;129:469.e1-32. 6. Iwasaki LR, Haack JE, Nickel JC, Morton J. Human tooth movement in response to continuous stress of low magnitude. Am J Orthod Dentofacial Orthop 2000;117:175-83. 7. Krishnan V, Davidovitch Z. The effect of drugs on orthodontic tooth movement. Orthod Craniofac Res 2006;9:163-71. 8. Bridges T, King GJ, Mohammed A. The effect of age on tooth movement and mineral density in the alveolar tissues of the rat. Am J Orthod Dentofacial Orthop 1988;93:245-8. 9. Misawa-Kageyama Y, Kageyama T, Moriyama K, Kurihara S, Yagasaki H, Deguchi T, et al. Histomorphometric study on the effects of age on orthodontic tooth movement and alveolar bone turnover in rats. Eur J Oral Sci 2007;115:124-30. 10. Harris EF. Effects of patient age and sex on treatment: correction of Class II malocclusion with the Begg technique. Angle Orthod 2001;71:433-41. 11. Ren Y, Maltha J, Van't Hof MA, Kuijpers-Jagtman A. Age effect on orthodontic tooth movement in rats. J Dent Res 2003;82:38-42. 12. Rinchuse DJ, Rinchuse DJ, Sosovicka MF, Robison JM, Pendleton R. Orthodontic treatment of patients using bisphosphonates: a report of 2 cases. Am J Orthod Dentofacial Orthop 2007;131:321-6. 13. Bartzela T, T€ urp JC, Motschall E, Maltha JC. Medication effects on the rate of orthodontic tooth movement: a systematic literature review. Am J Orthod Dentofacial Orthop 2009;135:16-26.
American Journal of Orthodontics and Dentofacial Orthopedics
Chisari et al
14. Helfrich M. Osteoclast diseases and dental abnormalities. Arch Oral Biol 2005;50:115-22. 15. Davidovitch Z, Krishnan V. Role of basic biological sciences in clinical orthodontics: a case series. Am J Orthod Dentofacial Orthop 2009;135:222-31. 16. Proffit WR. Biologic basis of orthodontic therapy; mechanical principles of orthodontic force control. In: Proffit WR, Fields HW, Sarver DM, editors. Contemporary orthodontics. 4th ed. St Louis: Mosby; 2007. 17. Choy K, Pae E, Park Y, Kim K, Burstone CJ. Effect of root and bone morphology on the stress distribution in the periodontal ligament. Am J Orthod Dentofacial Orthop 2000; 117:98-105. 18. McGorray SP, Dolce C, Kramer S, Stewart D, Wheeler TT. A randomized, placebo-controlled clinical trial on the effects of recombinant human relaxin on tooth movement and shortterm stability. Am J Orthod Dentofacial Orthop 2012;141: 196-203. 19. Kravitz ND, Kusnoto B, BeGole E, Obrez A, Agran B. How well does Invisalign work? A prospective clinical study evaluating the
S91
20.
21.
22. 23.
24.
25.
efficacy of tooth movement with Invisalign. Am J Orthod Dentofacial Orthop 2009;125:27-35. Drake CT, McGorray SP, Dolce C, Nair M, Wheeler TT. Orthodontic tooth movement with clear aligners. ISRN Dent 2012 [Epub 2012 Aug 14]. Hua Y, Nackaerts O, Duyck J, Maes F, Jacobs R. Bone quality assessment based on cone beam computed tomography imaging. Clin Oral Implants Res 2009;20:767-71. Liang S, Hosur KB, Domon H, Hajishengallis G. Periodontal inflammation and bone loss in aged mice. J Periodontal Res 2010;45:574-80. Renvert S, Persson RE, Persson GR. A history of frequent dental care reduces the risk of tooth loss but not periodontitis in older subjects. Swed Dent J 2011;35:69-75. Bollen AM, Huang G, King G, Hujoel P, Ma T. Activation time and material stiffness of sequential removable orthodontic appliances. Part 1: ability to complete treatment. Am J Orthod Dentofacial Orthop 2003;124:496-501. Dudic A, Giannopoulou C, Kiliaridis S. Factors related to the rate of orthodontically induced tooth movement. Am J Orthod Dentofacial Orthop 2013;143:616-21.
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
Variables affecting orthodontic tooth movement with clear aligners Justin R. Chisari,a Susan P. McGorray,b Madhu Nair,c and Timothy T. Wheelerd West Palm Beach and Gainesville, Fla
Introduction: In this study, we examined the impacts of age, sex, root length, bone levels, and bone quality on orthodontic tooth movement. Methods: Clear aligners were programmed to move 1 central incisor 1 mm over the course of 8 weeks. Thirty subjects, ages 19 to 64, were enrolled, and measurements were made on digital models (percentage of tooth movement goal achieved). Morphometric features and bone quality were assessed with cone-beam computed tomography. Data from this study were combined with data from 2 similar studies to increase the power for some analyses. Results: The mean percentage of tooth movement goal achieved was 57% overall. Linear regression modeling indicated a cubic relationship between age and tooth movement, with a decreasing rate of movement from ages 18 to 35 years, a slightly increasing rate from ages 35 to 50, and a decreasing rate from ages 50 to 70. The final decreasing trend was not apparent for women. As would be expected, the correlation was significant between the percentage of the goal achieved and the cone-beam computed tomography superimposed linear measures of tooth movement. A significant negative correlation was found between tooth movement and the measurement apex to the center of rotation, but bone quality, as measured by fractal dimension, was not correlated with movement. Conclusions: The relationship between age and tooth movement is complex and might differ for male and female patients. Limited correlations with cone-beam computed tomography morphology and rate of tooth movement were detected. (Am J Orthod Dentofacial Orthop 2014;145:S82-91)
T
he use of clear aligners to produce orthodontic tooth movement (OTM) provides an opportunity to measure incremental movement and investigate factors that might affect the rate of movement. The broad principles of OTM are based largely on bone and tissue remodeling, specifically the resorption and deposition of alveolar bone as force is applied. The biology of OTM has proven to be an extremely complex process involving an array of coordinated biochemical reactions, including critical cell signaling pathways and a wide range of cellular differentiation, leading to bone remodeling.1 As the science of bone biology continues to evolve, several theories of OTM have surfaced. The a
Private practice, West Palm Beach, Fla. Assistant professor, Department of Biostatistics, University of Florida, Gainesville. c Chairman and professor, Department of Oral & Maxillofacial Diagnostic Sciences, University of Florida, Gainesville. d Professor, Department of Orthodontics, University of Florida, Gainesville. All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and none were reported. Supported by Align Technology, San Jose, Calif, and the Southern Association of Orthodontists. Address correspondence to: Susan P. McGorray, Department of Biostatistics, Box 117450, Gainesville, FL 32611; e-mail, spmcg@ufl.edu. Submitted, June 2012; revised and accepted, October 2013. 0889-5406/$36.00 Copyright Ó 2014 by the American Association of Orthodontists. http://dx.doi.org/10.1016/j.ajodo.2013.10.022 b
S82
pressure-tension theory has emerged as the most popular concept behind the movement of teeth. Bone remodeling involves an intricate arrangement of coordinated cellular activity leading to bone resorption performed by osteoclasts, followed by bone formation carried out by osteoblasts.2 Dolce and Holliday3 have reported that although the precise biologic response to orthodontic force has not been identified, several hypotheses regarding the mechanisms by which osteoblasts and osteocytes sense this initial mechanical stimulus have been proposed, including strainsensitive ion channels, shear stress receptors, integrin activation, and cytoskeleton reorganization. Three phases of tooth movement have been described in the literature: initial phase, lag phase, and secondary phase.4,5 The secondary stage accounts for most of the tooth movement, and teeth during this period move at a faster, more continuous pace.6 The magnitude and direction of force placed on teeth during OTM, in addition to the length of time these forces are in place, also play critical roles in how teeth move. Forces applied to teeth cause various types of tooth movement depending on the location of the center of resistance of that tooth and the direction in which the force is applied. It is understood that the center of resistance for a given tooth changes based on tooth size,
Chisari et al
number of roots, and amount of tooth root that is submerged in bone. Variability among patients can affect OTM. Factors including age, sex, root length, bone levels, bone density, medications, and certain systemic conditions can have inhibitory, synergistic, or additive effects on OTM.7 The majority of literature on the effects of age on OTM has been completed using animal models. Bridges et al8 reported that a significantly greater amount and rate of tooth movement occurred in younger rats compared with their older counterparts in all 3 phases of tooth movement. Similar findings of the effects of increasing age on the rate and amount of tooth movement have been reported by MisawaKageyama et al9 and Harris.10 There has also been some indication that whereas there is a delay in the onset of tooth movement in adult rats, once the secondary phase of tooth movement is reached, the movement becomes equally efficient among the 2 age groups.11 The effect of age on OTM clearly exists and is likely due in part to a decreased biologic response. Although there is individual variability from patient to patient, a direct difference in OTM between the sexes has not been shown in the literature. Medications with pharmacologic effects can impact the cells targeted in OTM. Some of these medication classes include bisphosphonates, estrogens, NSAIDS and other analgesics, corticosteroids, calcium regulators, and supplements.7,12,13 There is little human experimental data on the effects of medications on OTM and limited information from animal models. However, knowing the biochemical action of these medications has led to concerns regarding how they can affect orthodontic treatment.13 Any medication that interferes with or alters bone biology might impact the rate of tooth movement. Systemic factors or nutritional deficiencies affecting bone metabolism have also been found to alter OTM. Specifically, diseases of bone can have a significant impact on the rate of tooth movement as well. Reduced or complete lack of osteoclast function can lead to a condition known as osteopetrosis, characterized by sclerosis of the skeleton and inhibited tooth movement and eruption. On the other hand, in Paget's disease, uncontrollable bone turnover occurs because of the overactivity of osteoclasts.14 Since OTM stimulates an inflammatory process in the periodontal ligament and surrounding tissues, it is thought that any chronic inflammatory disease such as thyroiditis, asthma, and even allergies can affect the movement of teeth.15 Other variables that might be of significance in OTM are root length, bone levels, and the density or quality of bone. Age-related decreases in bone turnover as well as a
S83
relatively higher level of bone density have been documented.8,13 Alveolar bone levels, on average, decrease over time, impacting OTM by changing the center of resistance. The center of resistance of a tooth is largely influenced by its surroundings, particularly in regard to root morphology, bone levels, and bone quality.16 Thus, patients with alveolar bone loss or abnormally long roots will have centers of resistance farther from the point of force application (more apically). Alternatively, the more the root tapers, the more the center of resistance moves coronally.17 Clear aligners with sequentially programmed movement provide an excellent model for investigating tooth movement. A single tooth can be isolated, and frequent measurements made with polyvinyl siloxane or digital impressions provide incremental information regarding the pattern of movement. For example, McGorray et al18 characterized the weekly pattern of tooth movement using this model over 8 weeks, along with subsequent relapse. Kravitz et al19 compared actual tooth movement with aligners with predicted movement over the course of treatment. Aligner treatment is now a commonly prescribed modality for OTM in adolescents and adults, and a better understanding of the pattern of movement and factors that influence movement could lead to more efficient treatment. The purposes of this study were to better characterize the pattern of tooth movement with clear aligners with programmed movement over 8 weeks and to examine the influence of age, sex, root length, morphometric measurements, and bone quality on the rate of OTM. MATERIAL AND METHODS
The design for this study was similar to 2 previous studies that investigated specific aspects of tooth movement with clear aligners.18,20 Approval was obtained from the University of Florida Institutional Review Board for the Protection of Human Subjects. This project was a prospective single-center clinical trial involving subjects of 2 age groups with minor incisor malalignments, who were otherwise healthy and would be undergoing orthodontic treatment. The first group included 7 men and 8 women between the ages of 18 and 35 years, inclusive. The second group consisted of 5 men and 10 women 50 years of age and older. Throughout this article, this study will be called Study 3. Study 1 investigated the role of relaxin in tooth movement and relapse, and has been previously described.18 No difference in tooth movement was detected when comparing those who received relaxin injections with those who received placebo injections; thus, data from both groups in this study were combined, yielding a sample size of 37 subjects. Cone-beam computed
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
Chisari et al
S84
tomography (CBCT) imaging was not performed in Study 1. Study 2 was similarly designed but included CBCT imaging and examined the role of aligner material fatigue in tooth movement.20 Subjects in this study received a new aligner every week rather than every 2 weeks as in Study 1. No difference was detected in total tooth movement when comparing the weekly aligner and the biweekly aligner groups. For studies 1 and 2, the biweekly tooth movement goal was 0.50 mm, for a total 8-week goal of 2 mm. Study 3 was designed to broaden the age range and to be used in conjunction with the previous studies to examine the role of age and other factors in tooth movement. In the 3 studies combined, the total number of subjects was 82. All subjects were in good health and had acceptable malocclusions as defined in the inclusion criteria, which have been described in a previous study.18 Once a subject was accepted into the trial, the right or left maxillary central incisor was selected as the target tooth. The selection was based on the target tooth's not being blocked out by the adjacent teeth to allow a total anteroposterior movement of 1 mm. Tooth movement was accomplished using a series of 4 maxillary aligners (Invisalign; Align Technology, San Jose, Calif), each programmed in increments of 0.25 mm of anterior movement of the central incisor being studied, as described above. Aligners were collected every 2 weeks from each subject, and new aligners were dispensed. The study termination visit and final time point for data collection was at week 8. Polyvinyl siloxane impressions were taken weekly and sent to Align Technology for scanning to create 3-dimensional models. Tooth movement measurements from baseline through week 8 were made from each scanned model using ToothMeasure proprietary software (Align Technology). CBCT measurements and fractal analysis were completed using a combination of software including InVivo (Anatomage, San Jose, Calif), ImageJ (National Institutes of Health, Bethesda, MD), and Tact Workbench (Wake Forest University, Winston-Salem, NC). The following are definitions of the measurements used in the CBCT superimposition analysis. D U1 (x) refers to the distance between lines drawn through the midpoint of the incisal edges of the superimposed target tooth perpendicular to the anteroposterior axis (the plane of prescribed tooth movement). D U1 (s) is the length of the line connecting the midpoint of the incisal edges of the superimposed target tooth. D Apex refers to the length of a line connecting the change in apex of the superimposed target tooth. Rotation angle is the angle created by the intersection of lines drawn from the midpoint of the incisal edge to the apex of the target tooth. The apex of this angle is considered the center
April 2014 Vol 145 Issue 4 Supplement 1
of rotation. Tooth length refers to the distance from the midpoint of the incisal edge to the apex of the target tooth from the initial computed tomography image. Crown length is the portion of the tooth length that is coronal to the bone. Bone to C-rot is the section of tooth length between the center of rotation and a line connecting the most coronal aspect of the faciolingual crestal bone. These variables are illustrated in Figure 1. All study subjects were instructed to wear the aligner appliance full time. They were allowed to remove the appliance when eating, drinking, or brushing their teeth. Their medication and medical histories were taken initially. Each subject recorded aligner wear time in a diary format. At the conclusion of the study, the participants were routinely treated orthodontically with clear aligners. To determine subject eligibility, 2 visits were required. The first visit was designed to identify potential subjects with malocclusions needing minor incisor alignment of at least the maxillary incisors. Those with medical conditions or intraoral problems, including significant periodontal disease, chronic daily use of any nonsteroidal or anti-inflammatory medication, current smokers, or history of significant cardiac disease, uncontrolled hypertension, bleeding disorders, or renal disease, were also excluded. Subjects who were determined to be eligible based on these procedures proceeded to the next visit. The screening visit was designed to finalize the subject's eligibility and collect initial records. The following procedures were performed at this visit: impressions were taken with polyvinyl siloxane for preparation of the Invisalign appliances, impressions were sent to Align Technology after confirmation of eligibility, and intraoral and extraoral photographs and CBCT imaging were done. For women, a negative urine pregnancy test immediately before this procedure was required. After the investigator (T.T.W.) reviewed all subject information to confirm eligibility, the subjects were enrolled into the study and assigned a unique number. At the first study visit (week 0), the first aligner was delivered to each subject. The acceptable visit window for weeks 1 through 8 was 6 1 day, and all 30 treatment subjects successfully satisfied this requirement. During the study visits of weeks 1 through 8, the following procedures were performed: intraoral clinical examination, maxillary occlusal and frontal photographs, and polyvinyl siloxane impressions. In addition, during the study visits of weeks 2, 4, and 6, the aligner (used during the previous 2 weeks) and the wear diary were collected, and the next aligner and diary were dispensed. At the study termination visit, week 8, the same procedures were performed, and CBCT imaging of the maxilla took place.
American Journal of Orthodontics and Dentofacial Orthopedics
Chisari et al
S85
Statistical analysis
Fig 1. Superimposed CBCT measurements. Blue is initial tooth position, and red is final tooth position (figure from thesis of Carl Drake; Gainesville: University of Florida; 2010).
Weekly anteroposterior movement of the target tooth was recorded with polyvinyl siloxane impressions. These impressions were sent to Align Technology, and digital models were created so that OTM could be measured using Align Technology's ToothMeasure software. The digitized model fabricated each week was superimposed on the baseline digital model, taken at week 0, according to the best fit of unmoved teeth, particularly the posterior dentition. The most central portion on the facial surface of the clinical crown of the target tooth, referred to as the centroid, was determined, and subsequent tooth movements in all dimensions were measured from this point at each study visit. The same investigator (J.R.C.) measured the digital models for all 30 subjects. CBCT scans of each subject were performed at the screening visit and at the study termination visit (week 8). Using the InVivo software, the images were superimposed on each other and registered at the curvature of the palate in addition to other stable maxillary structures. Measurements recorded from these superimpositions are shown in Figure 1. A fractal dimension score was calculated for each subject, representing the quality of the bone.21 Higher fractal dimensions correspond to greater morphologic complexity of the bone. Weekly wear time was calculated for each subject, with mean, median, and standard deviation of weekly wear time used to characterize each participant's wear patterns.
Orthodontic tooth movement was quantified using descriptive statistics for the digital model analysis. Chi-square tests of equality of proportions and analysis of variance (ANOVA) were used to compare subject characteristics and results over the 3 studies. To standardize the studies, the primary outcome was the percentage of tooth movement goal achieved over 8 weeks. Spearman correlation coefficients were estimated to examine the relationship between that outcome variable and age, morphometric measurements, and compliance. Median weekly hours of wear were used to represent compliance during the study, and this would not be overly influenced by a week of limited wear or excessive wear and would better represent typical weekly compliance. Linear regression modeling was used to examine the relationship between the percentage of tooth movement goal achieved and multiple covariates. Based on the R2 value (percentage of variability in the outcome explained by the model), the best 1, 2, 3, and so on variable models were determined. Model building was concluded when additional variables did not significantly improve the previous model. Potential interactions between covariates and the influence of outliers were also examined. A P value less than 0.05 was considered statistically significant, and analyses were performed using SAS software (version 9.1.3; SAS Institute, Cary, NC) and R software (version 2.15; R Foundation for Statistical Computing, Vienna, Austria). RESULTS
The demographic characteristics of the subjects for the 3 studies can be found in Table I. No significant differences were detected when comparing the studies for sex, race, compliance, and percentage of tooth movement goal achieved. Tooth movement goal and age were not compared because they differed in the designs of the studies. The pattern of tooth movement over 8 weeks is shown in Figure 2. Most tooth movement occurred in the first week of the 2-week wear cycle. Table II presents summary statistics and group comparisons for the demographic variables and planned biweekly tooth movement. The percentage of the goal achieved did not differ significantly by sex or race. Although it was not statistically significant (P 5 0.06), subjects with a smaller goal had a higher mean percentage of goal achieved, 62%, compared with 54% for those with a planned movement of .50 mm. There was no significant correlation with age, according to the Pearson correlation coefficient estimate of 0.021 (P 5 0.90). However, examining this correlation separately for men and women, different patterns were suggested, with
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
Chisari et al
S86
Table I. Demographic characteristics and descriptive statistics Sex, female (%) Race (%) White Black Other Tooth movement 2-week goal 0.25 mm (%) 0.50 mm (%) Age (y), median Mean (SD) Range Compliance, median hours/wk Mean (SD) Range % goal achieved, median Mean (SD) Range
Study 1 (n 5 37) 70
Study 2 (n 5 15) 60
Study 3 (n 5 30) 60
76 13 11
54 13 33
77 10 13
0 100 27.1 26.7 (5.1) 18.6-40.5 (n 5 34) 138 (24) 28-162 59.0 53.2 (16.2) 11.0-77.5
0 100 23.1 25.1 (4.9) 20.0-35.0 (n 5 15) 144 (14) 104-157 58.0 55.4 (15.0) 17.5-73.0
100 0 43.5 40.7 (15.5) 19.0-68.0 (n 5 30) 147 (8) 128-160 68.0 61.6 (20.0) 11.0-93.0
Total (n 5 82) 65 72 12 16 37 64 27.9 31.5 (12.3) 18.6-68.0 (n 5 79) 142 (18) 27-162 61.8 56.9 (17.7) 11.0-93.0
P value 0.63* 0.32* y
y
0.13z 0.17z
*Chi-square test of equality of proportions; ynot tested because of different study designs; zANOVA.
Fig 2. Mean percentage of tooth movement goal achieved for the 3 studies. Solid line indicates tooth movement goal (assuming linear movement) over the 8 weeks.
the women having a positive correlation coefficient estimate, 0.09 (P 5 0.52), whereas the men had a negative correlation coefficient estimate, 0.23 (P 5 0.25). We also did not detect a significant correlation between compliance (measured by median weekly hours worn) and percentage of goal achieved, with the Pearson correlation coefficient estimate of 0.06 (P 5 0.59). This relationship is displayed in Figure 3. Summary statistics for the CBCT measurements and their correlations with percentage of goal achieved are given in Table III. High correlation would be expected
April 2014 Vol 145 Issue 4 Supplement 1
between percentage of goal achieved and DU1(x), DU1(s), and rotation angle, since these should correspond well with the model-based tooth movement percentage of goal achieved measurement. Significant correlations were not noted for most morphometric measures. A negative correlation with percentage of goal achieved was identified for the apex to center of rotation measurement; this is illustrated in Figure 4. Linear regression modeling was used to examine the relationship between percentage of tooth movement goal achieved and covariates. Covariates considered for
American Journal of Orthodontics and Dentofacial Orthopedics
Chisari et al
S87
Table II. Descriptive statistics and comparisons of
demographic variables and planned treatment goals Variable Sex Male Female Race White Black Other Goal 0.25 mm 0.50 mm
n
Mean
SD
Range
P value *
29 53
54.4 58.2
16.1 18.5
17.5-80.0 11.0-93.0
0.35
59 10 13
58.0 58.4 50.5
16.1 19.4 22.8
11.0-93.0 24.0-85.3 15.0-77.5
0.37
30 52
61.6 54.1
20.0 15.8
11.0-93.0 11.0-77.5
0.06
*Two-sample t test or ANOVA.
inclusion in the model included age, age2, race, study biweekly goal (0.25 or 0.50 mm), compliance, and study number (1, 2, or 3). Sequentially, we determined the best models consisting of 1 to 5 variables. When 2 or more variables were selected for a model, a potential interaction between these variables was also considered for inclusion in the model. This is appropriate if the combined impact of the 2 variables differs from what would be expected from their individual contributions (ie, synergy or antagonism). Age and age2 were included to have more flexibility to identify the relationship between age and tooth movement. The best 1-variable model consisted of the biweekly goal (R2 5 0.042). The best 2-variable model included age and age2 (R2 5 0.084). The best 3-variable model included age, age2, and their interaction (ie, age3) (R2 5 0.14). No additional variables significantly improved the model. The fitted models and the actual data points are displayed in Figure 5. Because this model includes intercept, age, age2, and age3, it is considered a cubic age model. The overall cubic age model suggests that tooth movement slows from ages 18 through 35 years, increases slightly until approximately age 50, and then declines. We also fit this model separately for men and women. Model fit statistics and parameter estimates are provided in Table IV, and the predicted tooth movements for men and women are also illustrated in Figure 5. The decline in tooth movement after the age of 50 is not apparent in the model, with only the data from women. DISCUSSION
The combined data indicate that despite having aligners programmed to move 1 central incisor 1 mm labially (0.25 mm per aligner), on average only 57% of that movement was achieved. This discrepancy might be due to several reasons. It has been postulated that a greater percentage of tooth movement would occur if the prescription in each aligner was decreased from 0.5
Fig 3. Correlation between compliance (median weekly hours wear) and percentage of tooth movement goal achieved. Circles indicate observed data, solid line represents best linear fit (Pearson correlation coefficient estimate 5 0.06; P 5 0.59), and dotted line indicates best linear fit with outlier removed (Pearson correlation coefficient estimate 5 0.12; P 5 0.28).
to 0.25 mm. In studies 1 and 2, where each set of aligners was programmed for 0.5 mm of tooth movement, only 54% of the tooth movement was achieved; slightly more, 62%, was achieved with the smaller 0.25-mm goal. As mentioned earlier, the magnitude and direction of force placed on teeth during OTM, in addition to the length of time these forces are in place, can play critical roles in how teeth move. The use of aligners has increased in both the adolescent and young adult populations and among older adults seeking orthodontic treatment in recent years, raising concerns regarding the efficiency of tooth movement, particularly in the older population. Conventional thinking and clinical experience from several studies has led to the belief that the rate of tooth movement decreases with age.8-10 Additional factors, including sex, root length, bone levels, and bone density, can have various affects on tooth movement as well.7 Our results suggest that the relationship between age and tooth movement is complex and can differ depending on sex. Regression modeling supports an overall cubic relationship between OTM and age, represented by an s-shaped curve. Further exploratory analysis showed a quadratic (u-shaped) relationship for women and a more linear relationship for men. These trends might be the consequence of decreased quality of bone (osteoporosis) typically seen in older women. However, data
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
Chisari et al
S88
Table III. Descriptive statistics for CBCT measure-
ments and Spearman correlation coefficient estimates of specified variables with percentages of tooth movement goal achieved (n 5 45)
Variable D U1 (x)
Study 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2 3 Total 2
Mean 1.56 0.85 1.09 1.63 1.17 1.33 0.73 0.41 0.52 5.31 4.08 4.49 24.87 22.84 23.52 12.27 11.51 11.76 12.60 11.33 11.76 0.99 1.05 1.03 5.14 5.64 5.47 7.46 5.70 6.29 0.41
SD 0.38 0.37 0.50 0.40 0.49 0.51 0.26 0.14 0.24 1.32 1.63 1.63 2.02 2.09 2.26 0.74 0.93 0.94 1.74 1.87 1.90 0.12 0.22 0.19 1.25 2.24 1.96 2.01 1.57 1.90 0.11
Min 0.80 0.22 0.22 0.80 0.34 0.34 1.32 0.67 1.32 2.70 1.60 1.60 21.67 17.90 17.90 10.84 9.82 9.82 10.56 6.44 6.44 0.71 0.75 0.71 2.89 1.90 1.90 4.59 1.18 1.18 0.25
Max 2.02 1.96 2.02 2.09 2.32 2.32 0.39 0.09 0.09 7.50 8.90 8.90 30.32 26.90 30.32 13.27 13.79 13.79 17.74 14.95 17.74 1.23 1.78 1.78 7.70 10.19 10.18 12.82 9.26 12.82 0.63
3 Total Fractal dimension 2 3 Total
0.49 0.46 1.71 2.11 1.98
0.16 0.14 0.20 0.05 0.23
0.20 0.20 1.37 1.99 1.37
0.87 0.87 2.00 2.21 2.21
D U1 (s) D Apex
Rotation angle
Tooth length
Crown length
Root length
Crown/root ratio
Bone to C-rot
Apex to C-rot
Bone C-rot/ apex C-rot
Spearman correlation with % goal 0.90* 0.69* 0.37* 0.86* 0.56* 0.44* 0.72* 0.21 0.09 0.86* 0.59* 0.47* 0.42 0.08 0.17 0.17 0.07 0.07 0.40 0.12 0.16 0.12 0.13 0.12 0.10 0.13 0.15 0.18 0.27 0.35* 0.08 0.19 0.24 0.25 0.00 0.13
*Significant, P \0.05.
from studies 2 and 3 showed no statistical correlation between age and fractal dimension, a bone complexity and quality indicator. Caution must be used in interpreting the fractal dimension analysis, and additional information such as bone mass or structural properties of the bone might be needed to truly assess bone quality.21 Hormone replacement therapy can also play a role in the differing patterns. Of the many CBCT and digital model measurements examined, significant positive correlations were noted with DU1 (x), DU1(s), and rotation angle. That is, as
April 2014 Vol 145 Issue 4 Supplement 1
tooth movement increased, these measurements increased in both the CBCT superimposition and the digital model analysis. The DU1(s) and rotation angle difference between studies 2 and 3 implies more uncontrolled tipping in Study 2, likely because of the higher programmed tooth movement in that study. There was a significant negative correlation for the measurement apex to the center of rotation, indicating that as tooth movement increased, the apex to the center of rotation measurement decreased. This could be the result of root resorption, but it was more likely due to decreased alveolar bone levels.22,23 Other authors have examined the efficiency of tooth movement with aligners. Kravitz et al19 examined pretreatment and posttreatment locations for 37 subjects, with 401 teeth. They considered all types of movement and the complete treatment period (means of 10 maxillary and 12 mandibular aligners per treatment). Overall, the mean accuracy of tooth movement was 41%. It was expected that this would be lower than our observed 57% because the study was significantly longer and involved multiple types of movement and multiple teeth per subject. Bollen et al24 quantified tooth movement with aligners by assessing the ability to complete treatment (25-35 sequential aligners). Of 51 subjects, only 15 (29%) completed their initial series. This suggests that tooth movement is less than optimal and illustrates the potential usefulness of understanding factors that affect tooth movement. Factors related to tooth movement via archwires and brackets were examined by Dudic et al.25 Thirty subjects, ranging in age from 11 to 43 years, participated; younger subjects (age 15 and younger) appeared to have more tooth movement that those 16 years and older. Multiple teeth per subject were studied, and correlations within subjects were not adequately accounted for, making the results difficult to interpret. No relationship was detected between sex or tooth location and amount of movement. Compliance with the treatment plan is a key element in achieving tooth movement with aligners, and we were surprised that our measure of compliance was not related to the percentage of tooth movement goal achieved. The use of diaries to obtain time of wear requires accurate and truthful recording by the participants. The subjects in these studies appeared to be highly motivated, and their compliance level was uniformly high, with the exception of 1 outlier. Similar compliance rates have been reported in a similar study.19 Compliance might play a role in tooth movement in less motivated patients, but because of the uniformity of our sample, we were unable to detect this.
American Journal of Orthodontics and Dentofacial Orthopedics
Chisari et al
S89
Fig 4. Correlation between tooth movement and apex to the center of rotation measurement. Dots represent observed data, and line indicates best linear fit (Pearson correlation coefficient estimate 5 0.35; P 5 0.0184).
Fig 5. Cubic age model representing age vs the amount of tooth movement for the 3 data sets. Solid line, overall cubic model; dotted line, model using only data from women; dashed line, model using only data from men; solid squares, observed data for men; open circles, observed data for women.
We were fortunate to be able to combine data from 3 studies to investigate factors associated with tooth movement. Each of the component studies was designed to address a specific question regarding aligners, and the
similar designs, patient populations, and outcome measures allowed the combined data to provide a sufficient sample size to address our larger question. Care was taken in combining data, with separate comparisons of subject characteristics in the 3 studies and acknowledgment and testing (in our regression modeling) for the impact of unidentified study-specific components that could have affected tooth movement. No unidentified study-specific components were detected. The advantages of using the aligner model for tooth movement include ease of patient wear and patient recruitment, reliable outcome measures, and the benefit of a human model. Of particular interest is the ability to measure tooth movement on an incremental time scale. In all 3 substudies, stair-step movement was observed, with most movement occurring during the first week of the 2-week aligner time period. We have previously investigated whether this was due to a lack of continuous force caused by aligner fatigue, but we did not find this to be the case.20 Wear time appeared to be uniform throughout the 8-week study. It would be of interest to further characterize the movement over the 2-week period with more frequent measurements. As a consequence of not achieving biweekly tooth movement goals, accuracy decreased, resulting in poorer aligner fit over time. This can result in revised treatment planning and longer treatment times. There are several limiting factors associated with studying tooth movement with this model. Patient
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
Chisari et al
S90
Table IV. Linear regression model fit and parameter estimates (parameter estimates and P values) Variables in the model Model Overall Women Men
2
R 0.20 0.20 0.24
Overall fit (P value) 0.0006 0.0116 0.07
Intercept 252.9 (\0.0001) 145.7 (0.18) 235.7 (0.0041)
compliance is the single most important factor contributing to the amount of tooth movement seen. Unfortunately, even recording wear time on a daily basis has limited value. Clinical experience has suggested that a more continuous force enhances tooth movement. Removing the aligners for eating and brushing results in an interrupted force. Another factor, addressed by other studies, is the loss of anchorage of adjacent teeth during tooth movement. Whereas only 1 target tooth was programmed for tooth movement, movement of adjacent teeth is expected and can falsely minimize the amount of tooth movement of the target tooth. Polyvinyl siloxane impressions can have inaccuracies because of operator error or subject movement. Lastly, the use of aligners to produce tooth movement might be considered a limitation because it is not a perfect system for producing tooth movement. The ability to make an accurate prediction regarding how teeth will move when a force is applied has long been a challenge for orthodontists. Contributing to this factor is the large degree of variability of OTM seen among patients of the same age group, not to mention those across age groups. Different types of tooth movement, combined with varying biologic responses, make prediction difficult. Although there will always be variability in patients, our results suggest that there are general trends in tooth movement for different age groups, and that these might differ for male and female subjects. CONCLUSIONS
In this single-center prospective clinical trial, we examined the variables affecting tooth movement using a clear aligner model. When all 3 data sets were combined, the mean percentage of tooth movement achieved compared with the tooth movement goal was 57%. The results of regression modeling suggest an overall cubic relationship between OTM and age, represented by an s-shaped curve. Further exploratory analysis revealed a quadratic (u-shaped) relationship for women, indicating an increase in tooth movement in younger and older women, and a more linear
April 2014 Vol 145 Issue 4 Supplement 1
Age 16.23 (\0.0001) 15.49 (0.59) 9.92 (0.0229)
Age2 0.41 (\0.0001) 0.079 (0.80) 0.37 (0.0234)
Age3 0.0033 (0.0001) 0.0001 (0.98) 0.0030 (0.0232)
relationship for men, with decreased movement at older ages. Significant positive correlations were found between CBCT superimposition measurements and digital model measurements of the percentage of goal achieved for the DU1 (x), DU1(s), and rotation angle variables, whereas a significant negative correlation was seen with the percentage of goal achieved and the apex to center of rotation measurement. REFERENCES 1. Masella RS, Meister M. Current concepts in the biology of orthodontic tooth movement. Am J Orthod Dentofacial Orthop 2006; 129:458-68. 2. Proff P, Romer P. The molecular mechanism behind bone remodeling: a review. Clin Oral Investig 2009;13:355-62. 3. Dolce C, Holliday LS. Toward molecular orthodontics. In: Riolo ML, Avery JK, editors. Essentials for orthodontic practice. Ann Arbor and Grand Haven, Mich: EFOP Press; 2003. 4. Kohno, Matsumoto Y, Kanno Z, Warita H, Soma K. Experimental tooth movement under light orthodontic forces: rates of tooth movement and changes of the periodontium. J Orthod 2002;29: 129-35. 5. Krishnan V, Davidovich Z. Cellular, molecular, and tissue-level reactions to orthodontic force. Am J Orthod Dentofacial Orthop 2006;129:469.e1-32. 6. Iwasaki LR, Haack JE, Nickel JC, Morton J. Human tooth movement in response to continuous stress of low magnitude. Am J Orthod Dentofacial Orthop 2000;117:175-83. 7. Krishnan V, Davidovitch Z. The effect of drugs on orthodontic tooth movement. Orthod Craniofac Res 2006;9:163-71. 8. Bridges T, King GJ, Mohammed A. The effect of age on tooth movement and mineral density in the alveolar tissues of the rat. Am J Orthod Dentofacial Orthop 1988;93:245-8. 9. Misawa-Kageyama Y, Kageyama T, Moriyama K, Kurihara S, Yagasaki H, Deguchi T, et al. Histomorphometric study on the effects of age on orthodontic tooth movement and alveolar bone turnover in rats. Eur J Oral Sci 2007;115:124-30. 10. Harris EF. Effects of patient age and sex on treatment: correction of Class II malocclusion with the Begg technique. Angle Orthod 2001;71:433-41. 11. Ren Y, Maltha J, Van't Hof MA, Kuijpers-Jagtman A. Age effect on orthodontic tooth movement in rats. J Dent Res 2003;82:38-42. 12. Rinchuse DJ, Rinchuse DJ, Sosovicka MF, Robison JM, Pendleton R. Orthodontic treatment of patients using bisphosphonates: a report of 2 cases. Am J Orthod Dentofacial Orthop 2007;131:321-6. 13. Bartzela T, T€ urp JC, Motschall E, Maltha JC. Medication effects on the rate of orthodontic tooth movement: a systematic literature review. Am J Orthod Dentofacial Orthop 2009;135:16-26.
American Journal of Orthodontics and Dentofacial Orthopedics
Chisari et al
14. Helfrich M. Osteoclast diseases and dental abnormalities. Arch Oral Biol 2005;50:115-22. 15. Davidovitch Z, Krishnan V. Role of basic biological sciences in clinical orthodontics: a case series. Am J Orthod Dentofacial Orthop 2009;135:222-31. 16. Proffit WR. Biologic basis of orthodontic therapy; mechanical principles of orthodontic force control. In: Proffit WR, Fields HW, Sarver DM, editors. Contemporary orthodontics. 4th ed. St Louis: Mosby; 2007. 17. Choy K, Pae E, Park Y, Kim K, Burstone CJ. Effect of root and bone morphology on the stress distribution in the periodontal ligament. Am J Orthod Dentofacial Orthop 2000; 117:98-105. 18. McGorray SP, Dolce C, Kramer S, Stewart D, Wheeler TT. A randomized, placebo-controlled clinical trial on the effects of recombinant human relaxin on tooth movement and shortterm stability. Am J Orthod Dentofacial Orthop 2012;141: 196-203. 19. Kravitz ND, Kusnoto B, BeGole E, Obrez A, Agran B. How well does Invisalign work? A prospective clinical study evaluating the
S91
20.
21.
22. 23.
24.
25.
efficacy of tooth movement with Invisalign. Am J Orthod Dentofacial Orthop 2009;125:27-35. Drake CT, McGorray SP, Dolce C, Nair M, Wheeler TT. Orthodontic tooth movement with clear aligners. ISRN Dent 2012 [Epub 2012 Aug 14]. Hua Y, Nackaerts O, Duyck J, Maes F, Jacobs R. Bone quality assessment based on cone beam computed tomography imaging. Clin Oral Implants Res 2009;20:767-71. Liang S, Hosur KB, Domon H, Hajishengallis G. Periodontal inflammation and bone loss in aged mice. J Periodontal Res 2010;45:574-80. Renvert S, Persson RE, Persson GR. A history of frequent dental care reduces the risk of tooth loss but not periodontitis in older subjects. Swed Dent J 2011;35:69-75. Bollen AM, Huang G, King G, Hujoel P, Ma T. Activation time and material stiffness of sequential removable orthodontic appliances. Part 1: ability to complete treatment. Am J Orthod Dentofacial Orthop 2003;124:496-501. Dudic A, Giannopoulou C, Kiliaridis S. Factors related to the rate of orthodontically induced tooth movement. Am J Orthod Dentofacial Orthop 2013;143:616-21.
American Journal of Orthodontics and Dentofacial Orthopedics
April 2014 Vol 145 Issue 4 Supplement 1
