CONTINUING EDUCATION ARTICLE Directions of orthodontic tooth movements associated with external apical root resorption of the maxillary

central incisor
Robert J. Parker, DDS, MS,a and Edward F. Harris, PhDb Conway, Ark, and Memphis, Tenn External apical root resorption is a multifactorial problem encountered in all disciplines of dentistry, but it is most commonly seen in cases treated orthodontically. Specific tooth movements that are most likely to exacerbate external apical root resorption are poorly understood. Purpose of the present investigation was twofold: (1) to quantify apical and incisal movements of the maxillary central incisor in the sagittal and vertical planes from cephalograms and (2) to use stepwise multivariate linear regression analyses to see which tooth movements and skeletodental relationships are most predictive of external apical root resorption. The sample consisted of 110 adolescents with similar pretreatment malocclusions (Class I crowded or bimaxillary protrusive) and treatment planned similarly (extraction of four first premolars) by experienced private practitioners. Each of three practitioners used a different orthodontic appliance; the sample was divided proportionately into cases treated with Tweed standard edgewise technique, Begg lightwire technique, and Roth-prescription straightwire technique. Lateral cephalograms were analyzed at the start, middle, and end of treatment. There was no statistical difference in average external apical root resorption between sexes or among techniques. Measures of tooth movement were highly predictive, explaining up to 90% of the variation in root resorption. Apical and incisal vertical movements and increase in incisor proclination were the strong predictors of external apical root resorption for each regression model. Incisor intrusion with increase in lingual root torque together were the strongest predictors of external apical root resorption. In contrast, distal bodily retraction, extrusion, or lingual crown tipping had no discernible effect. (Am J Orthod Dentofacial Orthop 1998;114:677-83)

xternal apical root resorption (EARR) is a common iatrogenic problem associated with orthodontic treatment. Although EARR has long been recognized as a consequence of mechanically induced tooth movement,1-3 its causes are still poorly understood. EARR occurs during treatment when forces at the apex exceed the resistance and reparative ability of the periapical tissues. Other implicated factors include trauma, tooth devitalization, ectopic eruption of adjacent teeth, traumatic occlusion, recurrent heavy forces to the dentition such as bruxism, and habits such as tongue thrusting and finger nail biting.1,2,4-8 The literature suggests that movement of teeth with mature roots, extensive root movement, and intrusive mechanics enhance the risk of EARR.9-15 It is, however, difficult to isolate and evaluate specific tooth movements likely to enhance EARR because combinations of complex
aIn

mechanical tooth movements, such as extrusion, intrusion, translation, tipping, torquing, and rotations, are produced by a wide array of orthodontic appliances. The purpose of this study was to quantify apical and incisal movements of the maxillary central incisor throughout treatment and determine which movements were clinically relevant in predicting the extent of EARR. The maxillary central incisor was used because (1) EARR is most common on this tooth, (2) this tooth is easily visualized on a lateral cephalogram, and (3) maxillary incisors often undergo more displacement than other teeth during extraction treatment.
MATERIAL AND METHODS Sample Selection Orthodontic treatment records of 110 cases were used in this mixed longitudinal retrospective analysis. Forty-one records were from a solo Tweed standard edgewise practice, 29 were obtained from a solo Begg lightwire practice, and 40 were obtained from a solo Roth prescription straightwire practice. Treatment outcome was not considered in sample selection. To reduce sample variability, cases were used that fulfilled the following criteria: 1. Subjects were white and were adolescents at the start of treatment. 2. Subjects had a Class I sagittal molar relationship 677

bProfessor

private practice. in the Department of Orthodontics, The Health Science Center, University of Tennessee. Reprint requests to: Edward F. Harris, Department of Orthodontics, College of Dentistry, 875 Union Avenue, University of Tennessee, Memphis, Tennessee 38163, e-mail: eharris@utmem1.utmem.edu Copyright 1998 by the American Association of Orthodontists. 0889-5406/98/$5.00 + 0 8/1/88547

678

Parker and Harris

American Journal of Orthodontics and Dentofacial Orthopedics December 1998

Fig. 1. Schematic tracing of maxilla and central incisor shows linear and angular measurements for apex and incisal edge of maxillary central incisor. A, Sagittal tooth movements were measured to PNS-perpendicular. Landmark abbreviations are horizontal apical distance (A_HRZ), horizontal incisal distance (I_HRZ), vertical apical distance (A_VRT), and vertical incisal distance (I_VRT). Angular position () was the angle formed by long axis of the incisor with the palatal plane. B, Schematic tracing shows changes in tooth position. Abbreviations are horizontal apical change (_A_HRZ), vertical apical change (_A_VRT), straight-line movement of apex (A_DIS), horizontal incisal edge movement (_I_HRZ), vertical incisal edge movement (_I_VRT), and straight-line movement of incisal edge (I_DIS).

3.

4.

5. 6.

(Angle) at the start of treatment as assessed from the dental study casts. Subjects had undergone extraction of the four first premolars as part of treatment to alleviate anterior crowding or dentoalveolar protrusion. Each case had a diagnostic quality cephalogram before and after treatment. If a case had cephalograms taken during the course of treatment, one of these as close to midtreatment as possible also was assessed. Root formation of the maxillary central incisors was complete before treatment was initiated.16 There was no history of trauma to the permanent maxillary central incisors as ascertained from patient history and radiographic examination. Permanent maxillary central incisors also were intact, caries-free, and without endodontic treatment.

Maxillary Superimposition

Sample Description

The sample consisted of 47 males and 63 females. Mean start age was statistically equivalent in the male and female patients and among the three treatment modalities as judged from a two-way analysis of variance (grand mean, 13.4 yrs; standard deviation [SD], 1.78). Beginning of active treatment was defined by placement of separators and the end of active treatment was denoted by removal of the fixed appliances. Active treatment time ranged from 1.2 years to 4.4 years for the overall sample. Mean time in active treatment was statistically equivalent in the male and female patients and among the three treatment modalities (grand mean, 2.7 yrs; SD, 0.41).

Cephalometric images of the maxillae were superimposed for the pretreatment, intreatment, and posttreatment x-rays of each patient to quantify movements of the most protrusive maxillary central incisor. Fiducial points were located within the bony maxilla because external landmarks drift and displace at different rates than those within the maxilla. The anterior and posterior nasal spine were identified on the pretreatment radiograph, and these landmarks were penetrated on the film with a needle, leaving small but highly visible reference points. The intreatment radiograph then was placed over the pretreatment film. The palatal cortical plate and trabeculation patterns within each maxilla were used to register on the two bony structures.17,18 A definite change in radiographic opaqueness for the palatal cortical plate and common trabeculations was evident with precise superimposition of these anatomic structures. The ANS and PNS reference points were transferred to the intreatment radiograph. The posttreatment radiograph was then superimposed over the pretreatment radiograph and these two reference points were transferred in the same manner.
Positional Changes

Fig. 1 illustrates the measurements used to calculate the seven positional changes of the maxillary central incisor during orthodontic treatment. Two vertical tooth movements were calculated for each film: (1) the

American Journal of Orthodontics and Dentofacial Orthopedics Volume 114, Number 6

Parker and Harris

679

Table I.

Cephalometric variables assessed at each examination

SNA: SNB: ANB: FMA: IMPA: U1-PP: U1-L1: OP-PP:

posterior-inferior angle at junction of the sella-nasion and nasion-point A planes. posterior-inferior angle at junction of the sella-nasion and nasion-point B planes. obtained by subtraction of SNB from SNA. anterior-inferior angle at junction of Frankfort horizontal and the mandibular plane (Go-Me). posterior-superior angle at junction of the mandibular plane and long axis of mandibular central incisor. inferior-posterior angle at junction of the line through long axis of the maxillary central incisor and the palatal plane. posterior angle at junction of the long axis of maxillary central incisor and long axis of mandibular central incisor. angle at intersection of palatal plane and Downs occlusal plane. Angle is positive when the lines intersect distal to the arches so the occlusal plane is tipped down in front. Overjet: distance parallel with occlusal plane between the incisal edges of maxillary and mandibular central incisors. Overbite: distance perpendicular to occlusal plane between incisal edges of maxillary and mandibular central incisors.

Table II.

Tests for sexual dimorphism in tooth length (mm)

Males Females SD 1.8 1.9 1.9 1.0 0.6 0.6 n 63 32 63 63 32 32 x 26.1 25.1 24.6 1.5 0.9 0.9 SD 1.6 1.2 1.8 1.2 0.7 0.8 Sex difference t Test 2.8* 2.8* 3.0* 0.9 1.4 0.9 x 27.0 26.2 25.7 1.3 0.7 0.7

Maxillary central incisor Pretreatment tooth length Intreatment tooth length Posttreatment tooth length Pretreatment to posttreatment length change Pretreatment to intreatment length change Intreatment to posttreatment length change *P < .05.

n 47 38 47 47 38 38

distance from the maxillary central incisal edge to palatal plane (I_VRT), and (2) the distance from the central incisor apex to palatal plane (A_VRT). Sagittal tooth movements were measured to a plane perpendicular to palatal plane intersecting at PNS. Two parasagittal (horizontal) tooth distances were measured on each film: (1) the distance from the incisal edge to PNS-perpendicular (I_HRZ), and (2) the distance from the root apex to PNS-perpendicular (A_HRZ). Two straightline tooth movements also were calculated from pairs of films: (1) the distance the incisal edge traveled between pretreatment, intreatment, and posttreatment assessments (I_DIS) and (2) the distance the root apex traveled between the three treatment assessments (A_DIS). The other incisor tooth measurement was angular. The posterior-inferior angle formed by the intersection of the long axis of the maxillary central incisor and the palatal plane was measured on each film. Ten skeletodental variables also were measured on each film. The focus here was in determining whether the amount and kind of skeletodental changes during treatment were significantly associated with the amount of root resorption. These are defined in Table I.
Measuring Root Resorption

Table III.

Average changes in root length*

n 70 70 110 x 0.8 0.8 1.4 SD 0.7 0.7 1.1 t Test -8.9 9.8 13.7

Treatment interval Pretreatment to Intreatment Intreatment to posttreatment Pretreatment to posttreatment

*Negative values indicate decreases in tooth length. Sexes and treatment methods were pooled. The one-sample t tests determined whether the average change differed significantly from zero. Sample sizes varied because there were no midtreatment radiographs for the straightwire series. P < .0001.

Each cephalogram was assessed to define the degree and severity of EARR for the most procumbent

maxillary central incisor. Total tooth length was measured quantitatively with the Quick Ceph Image program, intermittently checking measurements directly with electronic sliding calipers. Image contrast and magnification were optimized for each measurement. Total tooth length was measured from the incisal edge to the most apical limit of the root. Tooth length was determined to the nearest 0.1 mm parallel to the long axis of the tooth. Because of the capability with Quick Ceph Image of optimizing magnification, brightness, and contrast of the radiographic image, intraobserver repeatability was high. Gauged with a repeated-measures analysis of variance,19 intraobserver error for root length was just 1.7%.

680

Parker and Harris

American Journal of Orthodontics and Dentofacial Orthopedics December 1998

Regression results predict EARR from incisor movements during first half of treatment*
Table IV. Skeletodental variable entered Apical_Vertical_21 Incisal_Vertical_21 Incisal_Angular_21 Standardized regression coefficient 1.26 1.02 0.89 Uniqueness index 0.41 0.32 0.20 Partial R2 0.15 0.12 0.20

Regression results predicting EARR from incisor movements during treatment*

Table VI. Skeletodental variable entered Apical_Vertical_31 Incisal_Vertical_31 Incisal_Angular_31 Overjet_31 FMA_31 Standardized regression Coefficient 1.45 1.45 0.94 0.07 0.08 Uniqueness index 0.75 0.61 0.27 0.00 0.01 Partial R2 0.19 0.37 0.35 0.00 0.01

*Model R2 was 47% (P < .0001). Models were evaluated without an intercept. The 21 suffixes denote changes from pretreatment to intreatment examinations.

*Model R2 was 92% (P < .0001). Models were evaluated without an intercept. The 31 suffixes denote changes from pretreatment to posttreatment examinations.

Regression results predicting EARR from incisor movements during the second half of treatment*
Table V. Skeletodental variable entered Apical_Vertical_32 Incisal_Vertical_32 Incisal_Angular_32 Overjet_32 Standardized regression cofficient 1.23 1.15 0.81 0.22 Uniqueness index 0.30 0.34 0.14 0.03 Partial R2 0.05 0.21 0.14 0.15

course of orthodontic treatment for the total sample (Table III).

Predictors of EARR

*Model R2 was 55% (P < .0001). Models were evaluated without an intercept. The 32 suffixes denote changes from intreatment to posttreatment examinations.

RESULTS Apical Root Resorption

Two-sample t tests confirmed an obvious sexual dimorphism for length of the maxillary central incisor (Table II). Males had a longer tooth, an average of 1 mm longer, at the pretreatment, intreatment, and posttreatment examinations. In contrast, changes in tooth length (EARR) were very similar between males and females; both sexes experienced statistically equivalent decreases in tooth length during each treatment interval. There also was no statistical difference in tooth length among the Begg, Tweed, and straightwire techniques at any of the three examinations; the average amounts of root loss were statistically the same across all three treatment modalities. These three techniques produce the same degree of apical resorption even though they achieve correction by different means. Resorption resulted in highly significant millimetric decreases in tooth length (P < .001). Of note, average loss from the pre- to intreatment examination was the same as that occurring from _ the intreatment to posttreatment examinations (x = .8 mm) in the Tweed and Begg groups (who had midtreatment x-rays). Average root loss of 1.4 mm occurred across the entire

Skeletodental variables (Table I) and measures of incisor movements (Fig. 1) were assessed with stepwise multiple linear regression analyses.20,21 The purpose was to test whether the amount of apical resorption could be predicted from either (1) starting conditions or (2) amounts of intreatment change. Initial runs had the entrance and deletion criteria for variable selection set at = .20; this retains predictors for subsequent analysis that may not meet a more rigorous probability value because of multicollinearity.22 Predictors not meeting the screening criterion (P > .20) were omitted from further analysis. The stepwise procedure then was used with the entrance and deletion criteria set at = 0.05, which yielded a set of predictors that was tested for multicollinearity using variance inflation factors and analyzing the structure of relationships among the variables.20-22 Substantial multicollinearity needs to be removed from a model because it means that two or more of the independent variables are more strongly correlated with one another than any is with the dependent variable. Multicollinearity confounds interpretation of the regression solution, and it can dramatically alter the values of the partial correlations (which also confuses interpretation).23,24 Multicollinearity also decreases the accuracy of the coefficient estimates.24,25 Use of variance inflation factors to measure and localize multicollinearity in a model has been described in various statistical works.20,24,26 The patients skeletodental morphologic characteristics (as assessed from 10 variables, Table I) was tested on the outside chance that there were clinical relevant predictors of EARR discernible before treatment. These would be useful flags for the orthodontist that a patient

American Journal of Orthodontics and Dentofacial Orthopedics Volume 114, Number 6

Parker and Harris

681

was particularly susceptible (or resilient) to root resorption during treatment with fixed appliances. In fact, none of the starting conditions was associated with the amount of resorption occurring during treatment. On the other hand, testing the amount of intreatment change disclosed consistently significant associations. Three regression models were constructed to detect relevant predictors of EARR (Tables IV, V, and VI). In each case the outcome variable was the amount of root resorption, and the three sets of predictor variables were (1) skeletodental changes during the first half of treatment, (2) changes during the second half of treatment, and (3) changes assessed from start to end of treatment. Uniqueness indexes were calculated to determine the percentage of the criterion variance explained by the individual predictors. In conjunction with uniqueness indexes, standardized regression coefficients, and partial coefficients of determination were evaluated to determine the relative importance of each predictors influence in the model.20 A models coefficient of determination (R2) is the variability of EARR explained by that set of independent variables in linear combination. An R2 above 40% was considered to be both clinically as well as statistically significant, although this was just a gauge based on the likelihood of being able to discern a systematic response clinically. Obviously, much weaker associations would achieve statistical significance alone. EARR that occurred throughout treatment was regressed on skeletodental variables measured from pretreatment to intreatment (Table IV) to detect any early treatment changes predictive of EARR. Three variables, incisal vertical change relative to palatal plane (I_VRT), apical vertical change (A_VRT), and change in angulation of the maxillary central incisor relative to the palatal plane (I_ANG) were retained by the stepwise procedure (ie, the 14 other variables did not meet the entrance criteria.) In linear combination they explained 47% of the variability in EARR. EARR that occurred throughout treatment also was regressed on variables measured during the second half of treatment (Table V) to define which late-treatment changes were predictive of EARR. Four variables entered the model (I_VRT, A_VRT, I_ANG, and OVERJET); in linear combination these explained 55% of the EARR variability. The uniqueness indices, however, revealed that overjet was a weak contributor to this regression model, but this also was the only model where a skeletal variable was retained. EARR was regressed to identify overall treatment criteria predictive of patients liability to EARR (Table VI). In linear combination, five predictor variables were found to account for 92% of the criterion variance

(I_VRT, I_ANG, A_VRT, FMA, and OVERJET). The uniqueness indices revealed that FMA and overjet were weak contributors to this regression model.
DISCUSSION

External apical root resorption remains a common iatrogenic problem in orthodontics. There are primarily two impediments to preventing resorption: (1) root apices are prone to resorption when the periodontium is compressed, so teeth cannot be moved through bone without producing some odontoclasia; and (2) no exact criteria have been found that predict which patients will experience overt resorption and which will exhibit little under the same treatment regimens. Individual responses to seemingly similar forces can be quite variable. However, most patients exhibit little resorption, and the esthetic and functional benefits of treatment outweigh the minor iatrogenic sequelae.
Regression Analysis

Standardized multiple linear regression coefficients revealed repeatedly that vertical movement of the root apex (A_VRT), vertical movement of the incisal edge (I_VRT), and change in proclination (I_ANG) were in the optimum subsets of predictor variables for each of the three regression models described. Repeatedly, apical and incisal intrusive movements and increase in incisor proclination were the most powerful predictors of EARR. Orthodontically, this means incisor tooth intrusion in the vertical plane and lingual root torque in the sagittal plane cause the most apical resorption. Ten skeletodental variables were measured on all cases (Table I), but none was strongly associated with EARR. These variables differed greatly among patients as a result of differences in facial morphologic characteristics, but none was retained in any regression equation, which implies that facial morphologic characteristics are effectively unrelated to the variation in EARR.
Prior Studies

Only a few studies have evaluated the extent of apical root resorption based on orthodontically induced tooth movements.10,27,28 There are substantive methodologic differences among these three studies that make it difficult to compare results (Table VII). Differences in research design may explain why the inferred relationship between maxillary incisor movements and EARR remains contentious. Phillips27 concluded that there was no association between EARR and angular or sagittal apical movements of the maxillary central incisor. DeShields,10 in contrast, found significant correlations between EARR and sagittal apical movements of the maxillary central incisor. Mirabella and

682

Parker and Harris

American Journal of Orthodontics and Dentofacial Orthopedics December 1998

Table VII. Criterion

Summary of characteristics of present and prior studies

Present study Mixed longitudinal 11-18 years Class I Extraction* Edgewise and Begg Palatal plane Lateral ceph White Incisal Apical Angular Regression# Student t Chi-square Phillips (1955) Cross sectional 11-19 years Class I, II, III Extraction and nonextraction Edgewise Palatal plane Lateral ceph Unknown Apical Angular Pearson r Student t DeShields (1969) Cross sectional 11-16 years Class II Nonextraction Edgewise Sella-Nasion Bolton-ANS Periapical Unknown Apical Angular Pearson r Mirabella and rtun (1995) Cross sectional 20-70 years Class I, II, III Unknown Edgewise Palatal plane Periapical Unknown Incisal Apical Angular Regression# Student t Chi-square

Study type Age range Pretreatment malocclusion Treatment plan Appliance Reference planes Radiograph to assess EARR Race Movements measured Statistical tests

*Included only four first-premolar extractions. Tweed standard edgewise plus straightwire edgewise. Unknown whether standard or preadjusted appliances. Measured both vertical and sagittal movements. Straight-line movement calculated. Measured only sagittal movement. #Stepwise multiple linear regression.

rtun28 stated, movement of the roots in either an anterior or posterior direction, is associated with apical root resorption. Although Mirabella and rtun found no statistically significant association between vertical movement of the incisor apex and root resorption, they did caution about interpretation because few patients in their study experienced as much as 1 mm of extrusion or intrusion. In the present study, multiple linear regression revealed that vertical apical movement, vertical incisal movement, and incisor proclination changes were consistently predictive of EARR.
Root Torque

Apical root resorption decreases a roots surface area and, thereby, its periodontal support and ability to resist relapse.29,30 EARR averaged 1.4 mm in the present sample. Jacobson31 suggested that a loss of a millimeter at the apex is trivial because the apical end of the root has the smallest diameter, but Kalkwarf et al32 showed that there is a nearly linear relationship between root length and percentage of periodontal attachment, so minor loss in root length may be important. In addition, loss of root length moves the center of resistance coronally, so the same amount of torque on the tooth will have a greater effect than if the root were intact.33 In the present study, increasing the angle between the central incisor and the palatal plane was strongly correlated with increasing EARR. This angular increase occurred by tipping the crown facially and

torquing the apex lingually, but the change was mainly due to lingual root torque rather than labial crown tipping because all cases were treated with first premolar extractions and mechanics were designed to retract the incisors. A major goal in extraction treatment with an edgewise appliance is to retract the incisors while maintaining torque. In the Begg technique, a major focus in stage 3 is to torque the maxillary incisors and remove the axial tipping that occurred early in treatment. The point is that substantial force is required to torque the apex of the maxillary central incisor lingually, and, unfortunately, these torquing forces concentrate at the apex which is the smallest and most resorption sensitive area of the tooth.8
Intrusion

Intrusive forces are damaging to root surfaces because root shape concentrates pressure at the conical apex.34 Studies that have explicitly examined the effects of intrusive mechanics on root resorption concluded that intrusive forces produce apical resorption.14,35,36 In the present study, intrusive movement of the maxillary central incisor had a strong positive association with the amount of EARR. On the other hand, few studies of intrusion have contrasted other vectors of movement on the same teeth.37 In the present study, the strongest predictors of EARR were vertical incisal and apical movements. It should be noted, however, that no regression model showed intrusion alone to be

American Journal of Orthodontics and Dentofacial Orthopedics Volume 114, Number 6

Parker and Harris

683

the cause of EARR. Each clinically significant model showed that a combination of movements best explained the variation in EARR among patients. These movements primarily were intrusion in combination with lingual root torque.
Sagittal Movements

Studies support the contention that the distance a root is moved through bone influences the amount of EARR. DeShields10 found a significant, positive correlation between EARR and how far the root was transposed horizontally. Sharpe et al29 showed that cases requiring premolar extraction experienced more resorption than those requiring less retraction of the maxillary incisors. Harris and Butler38 and Kaley and Phillips39 also reported that the horizontal amount the maxillary incisors needed to be retracted was positively correlated with EARR. In contrast, the present study showed that exclusively horizontal tooth movements were not significantly associated with EARR when other vectors were accounted for. In addition, the straight-line distance the incisal edge or the apex traveled was not strongly correlated with EARR.
CONCLUSIONS The present investigation quantified apical and incisal movements of the maxillary central incisor and used multivariate regression analyses to see which movements and skeletodental variables were most predictive of EARR. Major findings were: 1. Incisor tooth movements taken in combination are strong statistical predictors of the amount of root resorption experienced during treatment. Specific directions of movement differentially enhance the extent of EARR, and the amount of EARR is a function of the amount of movement. 2. In combination, intrusive movement and lingual root torque were the strongest predictors of EARR. In contrast, bodily retraction in a posterior direction, extrusion, or lingual crown tipping had no discernible influence.
REFERENCES 1. Ketcham AH. A preliminary report of an investigation of apical resorption of permanent teeth. Int J Orthod 1927;13:97-127. 2. Ketcham AH. A progress report of an investigation of apical root resorption of vital permanent teeth. Int J Orthod 1929;15:310-28. 3. Stuteville OH. A summary review of tissue changes incident to tooth movement. Angle Orthod 1938;8:1-20. 4. Massler M, Malone AJ. Root resorption in human permanent teeth. Am J Orthod 1954;40:619-33.

5. Massler M, Perreault J. Root resorption in the permanent teeth of young adults. J Dent Child 1954;21:158-64. 6. Ramfjord SP, Ash MM. Occlusion. 2nd ed. Philadelphia: WB Saunders, 1971. p. 158-70. 7. Odenrick L, Brattstrm V. Nailbiting: frequency and association with root resorption during orthodontic treatment. Br J Orthod 1985;12:78-81. 8. Reitan K. Biomechanical principles and reactions. In: Graber TM, Swain BF, eds. Orthodontics: current principles and techniques. St Louis: CV Mosby, 1985. p. 101-92. 9. Rudolph CE. An evaluation of root resorption during orthodontic treatment. J Dent Res 1940;19:367-71. 10. DeShields RW. A study of root resorption in treated Class II Division 1 malocclusions. Angle Orthod 1969;39:231-45. 11. Sjlien T, Zachrisson BU. Periodontal bone support and tooth length in orthodontically treated and untreated persons. Am J Orthod 1973;64:28-37. 12. Linge BO, Linge L. Apical root resorption in upper anterior teeth. Eur J Orthod 1983;5:173-83. 13. Linge BO, Linge L. Patient characteristics and treatment variables associated with apical root resorption during orthodontic treatment. Am J Orthod 1991;99:35-43. 14. Dermaut LR, De Munck A. Apical root resorption of upper incisors caused by intrusive tooth movement: a radiographic study. Am J Orthod 1986;90:321-6. 15. Levander E, Malmgren O. Evaluation of the susceptibility of root resorption during orthodontic treatment: a study of upper incisors. Eur J Orthod 1988;10:30-8 16. Moorrees CFA, Fanning EA, Hunt EE Jr. Age variation of formation stages in ten permanent teeth. J Dent Res 1963;42:1450-502. 17. Bjrk A. Sutural growth of the upper face, studied by the implant method. Eur Orthod Soc Trans 1964;40:49-65. 18. Johnston LE Jr. Balancing the books on orthodontic treatment: an integrated analysis of change. Br J Orthod 1996;23:93-102. 19. Winer BJ. Statistical principles in experimental design. 2nd ed. New York: McGrawHill Book Company, 1991. 20. Freund RJ, Little RC. SAS system for regression. 2nd ed. Cary [NC]: SAS Institute Inc, 1991. 21. SAS Institute Inc. SAS users guide: statistics. 5th ed. Cary: SAS Institute Inc, 1985. 22. Draper NR, Smith H. Applied regression analysis. New York: John Wiley and Sons Inc, 1966. 23. Gordon RA. Issues in multiple regression. Am J Sociology 1968;73:592-616. 24. Cohen J, Cohen P. Applied multiple regression/correlation analysis for the behavioral sciences. New York: John Wiley and Sons, 1975. 25. Kerlinger FN, Pedhazur EJ. Multiple regression in behavioral research. New York: Holt, Rinehart and Winston, Inc, 1973. 26. Myers RH. Classical and modern regression with applications. 2nd ed. Boston: PWS and Kent Publishing Company, Inc, 1990. 27. Phillips JR. Apical root resorption under orthodontic therapy. Angle Orthod 1955;25: 1-12. 28. Mirabella AD, rtun J. Risk factors for apical root resorption of maxillary anterior teeth in adult orthodontic patients. Am J Orthod 1995;108:48-55. 29. Sharpe W, Reed B, Subtelny JD, Polson A. Orthodontic relapse, apical root resorption, and crestal alveolar bone levels. Am J Orthod 1987;91:252-8. 30. Hollender L, Ronnerman A, Thilander B. Root resorption, marginal bone support and clinical crown length in orthodontically treated patients. Eur J Orthod 1980;2:197-205. 31. Jacobson O. Clinical significance of root resorption. Am J Orthod 1952;38:687-96. 32. Kalkwarf KL, Krejci RF, Pao YC. Effect of apical root resorption on periodontal support. J Prosthet Dent 1986;56:317-9. 33. Vaden Bulke MM, Burstone CJ, Sachdeva RC, Dermaut LR. Location of centers of resistance for anterior teeth during retraction using the reflection technique. Am J Orthod Dentofacial Orthop 1987;91:375-84. 34. Beck BW, Harris EF. Apical root resorption in orthodontically treated subjects: analysis of edgewise and lightwire mechanics. Am J Orthod Dentofacial Orthop 1994;105:350-61. 35. Harry MR, Sims MR. Root resorption in bicuspid intrusion: a scanning electron microscopic study. Angle Orthod 1982;52:235-58. 36. Stenvik A, Mjr IA. Pulp and dentin reactions to experimental tooth intrusion: a histologic study of the initial changes. Am J Orthod 1970;57:370-85. 37. Reitan K. Initial tissue behavior during apical root resorption. Angle Orthod 1974;44:68-82. 38. Harris EF, Butler ML. Patterns of incisor root resorption before and after orthodontic correction in cases with anterior open bites. Am J Orthod Dentofacial Orthop 1992;101:112-9. 39. Kaley J, Phillips C. Factors related to root resorption in the edgewise practice. Angle Orthod 1991;61:125-32.