Measuring dental drift and orthodontic tooth movement in response to various initial forces in adult rats Gregory J. King, DMD, DMSc, Stephen D. Keeling, DDS, MS, Elizabeth A. McCoy, AA, and Thomas H. Ward, DMD Gahzesville, Fla.
A method for the quantification of orthodontic tooth movement in the rat is presented. Reliability, sensitivity, and validity were assessed and tooth movement kinetics were determined for initial forces of 20, 40, and 60 gin. The appliance consisted of a 9 mm length of closed coil spring suspended between a cleat and bonded to the occlusal surface of the maxillary first molars and the maxillary incisors. Initial tipping forces were placed by suspending known weights from the anterior end of these coils before fixation to the incisors. Tooth movement was quantified from enlarged cepnalograms by measuring the position of a reproducible landmark on the molar cleat with respect to either zygomatic amalgam implants or a barbed broach placed submucosally on the palate. All measurements were made along the molar-incisor vector by projecting at 90 ° to this line. Validity and sensitivity were assessed by quantifying molar distal drift and comparing these results with reports of bone turnover rates adjacent to distally drifting adult rat molars. Reliability was obtained by estimating the error of a single measurement in a longitudinal study of 12 adult male Sprague-Dawley rats (180 to 200 days) receiving both amalgam and broach implants and a cross-sectional study of 72 animals divided equally into six groups to be killed at 1, 3, 5, 7, 10, and 14 days. No orthodontic forces were used in this portion of the study. Implant stability within the craniofacial complex was assessed by measuring bilateral broaches as a function of time with respect to each other. There were no systematic errors between replicate films for either the amalgam or the broach method. The 95% confidence limit for a single determination of molar position was 62 ixm using the amalgams and 47 p.m for the broach (p < 0.001). The latter could be reduced to 23 rtm when the average of four independent determinations was used. Homologous implants did not differ with respect to each other in the sagittal plane but did in the transverse plane (p < 0.01), migrating laterally 9 ixm/day. Linear regression analysis of molar distal movement over time predicted 7.7 ttm/day distal drift (p < 0.01), which compared favorably with reports of 6.7 p.m/day of alveolar bone turnover during this drifting process. Characteristic three-part cumulative tooth movement kinetics were obtained for the 40 and 60 gm initial force groups. No individual time point at 60 gm differed from its counterpart at 40 gm. All time points on the 20 gm curve differed significantly from those on the higher force curves (p < 0.001). Comparisons within the highest force group indicated that mesial molar movement differed from zero at all times and that 1, 3, 5, and 7 days did not differ from each other, but 10 and 14 days differed from several other days. In the 20 gm group, day 10 differed from day 7. In the 20 gm group, distal molar drift resumed shortly after initial deformation at 14 i.tm/day. This method is sufficiently reliable, precise, and valid for the quantification of orthodontic tooth movement under experimental conditions. Tooth movement in response to varying initial forces indicated that (1) orthodontic appliances can be overloaded, resulting in no further enhancement of tooth movement and (2) under, conditions of overload, all three processes represented in the tooth movement curve are affected equally. (AMJ ORTHOO OENTOFACORTHOP 1991 ;99:456-65.)
A
reliable, sensitive, and valid in vivo model for orthodontic tooth movement would facilitate studies on biologic mechanisms and studies on how From the Department of Orthodontics, College of Dentistry, Unive~ity of Hofida. Supported by NIDR Grant DE-08659. 8/1122216
456
various physical and pharamcologic manipulations affect clinical outcomes. The interpretation of most available tooth movement data is unclear because of several deficiencies in the models currently being used: (1) tooth movement often is not quantified in descriptive histologic studies, ~4 (2) continuously erupting teeth are used to represent human tooth movement, 2 (3) over-
Voh~me 99 Number 5
Dental drift and tooth movement
m
457
!
Fig. 1. Rat positioned in the restrainer for appliance activation. Note bilateral mouth props/suction devices and tongue retraction with a suture tied back to one of the posterior pillars.
lapping orthopedic and orthodontic changes are not factored out,'- (4) various levels of tissue trauma caused by the appliances confuse similar changes that are specific to orthodontic tooth movement,' (5) extraneous occlusal forces are not eliminated, j4 (6) the reliability, sensitivity, and validity of the methods are unevaluated, ~4 and (7) appliance force magnitudes are imprecise or unknown."3 The purposes of this study were (1) to test the reliability, sensitivity, and validity of a proposed rodent model for orthodontic tooth movement that addresses most of these problems and (2) then to apply it to a study of the time course of tooth movement changes in response to precise initial forces. MATERIALS AND METHODS Animals Adult male Sprague-Dawley strain rats (180 to 200 days old) were used because of availability, cost, genetic homogeneity, and, at this age, a relatively slow rate of growth. These methods can be adapted to younger animals to address questions requiring a more rapidly growing animal. Head-holding device This instrument was designed to serve as a surgical restrainer. It was fabricated from Lucite and consisted of a base (6.5 x 9 inch) with two vertical pillars supporting adjustable ear rods that were coplanar and parallel to the base. Two vertical pillars also were placed on the base caudal to the animal's head to serve as convenient points for tying tongue-retraction sutures. Two removable pulleys were placed rostral to the an-
imal's head to accommodate appliance activation procedures. Two more pillars were positioned on either side of the pulleys. These pillars were equipped with adjustable suction devices, which were fabricated from aluminum tubing (3/,6 inch diameter). One end of the tubing was flattened, sealed with solder, and modified by placing six slits in it. These could be positioned to provide both cheek retraction and a dry intraoral field. The animal was positioned with the head up. The mouth was propped with the suction devices in place bilaterally. The tongue was retracted by placing a suture (6-0 silk) through it and tying this to one of the pillars located caudal to the head. This format permitted unassisted placement and activation of all appliances and surgical manipulations (Fig. l ) . Appliances and animal manipulations Rats were weighed and anesthetized with intramuscular ketamine (87 mg/kg) and xylazine (13 mg/kg). After mounting the animal in the head holder, the occlusal surfaces of both maxillary first molars were prepared by roughing the surface with a green stone, washing with acetone, etching with a phosphoric acid gel for 3 minutes (Healthco International, Boston, Mass.) and sealing with an orthodontic composite resin (System I, Ormco, Glendora, Calif.). Direct-bonded orthodontic cleats (American Orthodontics, Sheboygan, Wis., No. 593-852-165) were modified by removing one of the cleat arms and reducing and reshaping the base to a trapezoid with the apex oriented opposite the remaining cleat arm. This attachment was rinsed in acetone, air-dried, and treated frst with sealant followed by bonding material (System
458
K h l g et al.
Am. J. Orthod. Dentofac. Orthop. May 1991
r
l
Fig. 2. Rat positioned in the restrainer for appliance activation.
I). The attachment then was placed on the occlusal surface of the prepared maxillary first molar with the trapezoidal apex on the mesial aspect and the cleat arm pointing distally. A 6 to 7 mm length of coarse endodontic barbed broach with a 0.67 mm interbarb distance (Maillefor, Switzerland) then was inserted through the palatal mucosa along the palatal bone lingual to the maxillary first molars. The broach was inserted using a hemostat and the end was pushed below the tissue with an explorer, leaving a pinpoint insertion wound that healed readily. Before release of the rat, mandibular first and second molars were extracted with scissors used to remove the crowns and an explorer to elevate the major roots. The rat then was allowed to recover for approximately 1 week by monitoring wound healing and weight gain. For appliance activations the animal again was placed head up in the restrainer as described previously. One end of a 9-mm length of closed coil (0.006 inch Hi T; arbor diameter: 0.022 inch, Unitek, Monrovia, Calif.) was attached to a steel ligature wire loop (0.009 inch). This then was slipped over the molar cleat. The anterior end of the coil was spread apart and attached with suture material to a premeasured weight (Fig. 2). The weight was suspended over the pulley and allowed to hang freely, exerting the prescribed intial force through the coil to the molar. The anterior end of the coil then was bonded to the lateral surface of the maxillary incisor by means of the same procedure as described for attaching the cleat to the maxillary molar. The excess coil and suspended weight were removed and the mandibualt incisal edges were trimmed with scissors.
Measurement of tooth movement
Radiographic exposures were made using a second head-holding device (Fig. 3). This instrument was a simplified version of the restrainer and consisted of ear rods for positioning the rat face down on the base. The base was equipped with a spirit level and four height adjustment knobs to facilitate precise leveling of the instrument. A film packet was placed on the base under the animal's head. The device was placed on the floor and radiographed at a 127 cm object-to-film distance (10 mA, 60 kVp, 4 seconds) with the central ray perpendicular to the base. These cephalograms routinely displayed the implants, molar cleats, and orthodontic springs (Fig. 4, A). These were mounted in slide holders and viewed at approximately × 8 magnification using a rear-screen projection device (Ektagraphic 200 Audioviewer; Kodak, Rochester, N.Y.). Corrections were not made for lens distortion. However, all cephalometric exposures were made with uniform positioning; cephalograms were mounted uniformly and these were all projected with the same lens in an effort to control for this variable across cephalograms, animals, and experimental groups. With acetate tracing paper, a template was made for the following cephalometric structures: the broaches with their adjacent barbs, the molar cleats, and the incisors. By means of these templates, landmark points were selected (1) at the intersection of a barb with its broach body, (2) at the trapezoidal apex of the molar cleat, and (3) on the distal surface of the incisor (at the location of the spring attachment in the animals with force applications). These templates were oriented by best-fit visual inspection on subsequent radiographs and
Vohtme 99
Dental drift and tooth movement
459
Number 5
X-R~ ...............
•
i
127cm
ilm
Fig. 3. Rat positioned for a cephalometric exposure.
the landmarks for each structure registered on a second, overlying acetate paper (Fig. 4, B). The choice of a cephalometric landmark for the tooth is problematical because the amount of tooth movement measured will vary as a function of the landmark's distance from the center of rotation of the tipping tooth. Since the issue of greatest importance in the choice of a dental landmark for this study was reproducibility, the mesiooccclusal surface was chosen for its accessibility and anatomic uniformity. This location also offers the advantage of not traumatizing the periodontal ligament. These tracings were digitized to record the X and Y coordinates of these cephalometric landmarks using a Hipad digitizing tablet (Houston Instruments, Austin, Texas) interfaced with a microcomputer (IBM-PC/AT, IBM Corp., Atlanta, Ga.). A computer program measured distances, adjusting for magnification, by using the known, digitized interbarb distances of the broach. The molar and implant landmarks were projected at 90 ° to the molar-incisor vector and the distance from the tooth to the implant measured. The difference between these measurements made on appliance placement and radiographs taken at death was calculated and reported as tooth movement. The mean of right and left measurements represented molar movement in each animal. Error assessments
The reliability of two methods of measuring molar position (one for molar-to-broach implant; the second, for molar-to-amalgam implant) was determined by estimating the error of a single measurement. This was accomplished in a longitudinal study of 12 animals that received both barbed broach implants as described pre-
Adult JIG Force 0 gins Osy Appllonce 3"emplate
°O
{o Fig. 4. A, Representative cephalogram demonstrating broach and amalgam implants and molar cleats. B, Tracing of the landmarks from a cephalogram.
viously and intraosseous amalgam implants as described by Tuncay and Killiany. 5 The latter were placed in a triangular array by making a single incision medial to the dental arch, extending from 1 mm mesial to the maxillary first molar to the zygomatic maxillary suture and reflecting the tissue laterally to expose the lateral wall of the maxilla. An 18-gauge needle with an 0.16
460
King et al.
Am. J. Orthod. Oentofac. Orthop. May 1991
Table I. Weight and percentage weight change of the 14-day animals by group*
Weight at death (gin)
Initial weight (gm) Force group
Mean
0 20 40 60
495.83 500.77 589.00 448.00
gm gm gm gm
[
SE
Mean
4.64 10.08 19.89 7.16
407.08 375.77 425.50 333.50
[
Change? (%)
SE 10.97 9.84 19.91 12.32
Mean 17.9 - 24.9 - 27.9 -25.7
-
[
SE 2.0 1.4 1.7 2.1
*Percentage weight change equals (weight at death-initial weight)/initial weight. tAnalysis of variance indicated significant differences existed among the groups in amount of weight loss (p = 0.001). Scheffe pairwise comparisons indicated these differen.ces were manifested between the 0 force group and each of the other groups (p < 0.05); no significant differences in weight loss were detected between the 20, 40, and 60 gm groups (p > 0.05).
inch stainless steel wire inserted in the lumen was adapted to serve as an amalgam carrier. Holes to receive the amalgam were prepared with a Y2 round burr. The tissue was sutured with 6-0 silk. Molar cleats were used to facilitate tooth localization but springs were not placed. Templates of the amalgam images were constructed and landmarks selected. Measurements of molar positions in relation to both broach and amalgam implants were obtained from the tracings of 132 radiographs representing two independent radiographs taken on each of I l rats on days 1,3, 5, 7, 10, and 14. (Note: one of the original rats died after day 1.) The radiographs were obtained after rcpositioning the rat between exposures: Implant stability within craniofacial complex
The relative stability of the barbed broaches was examined by comparing movement of right and left homologous implants in the sagittal and transverse dimensions at each time point on a group of 72 rats prepared with bilateral molar cleats and broaches but not orthodontic springs. These rats were divided equally into six subgroups to be killed at 1, 3, 5, 7, I0, and 14 days. Movement of the broaches was referenced to the right molar-incisor vector. Validation of method for measurement of tooth movement
Rodent molars are known to drift distally. 6 Histologic measurement of alveolar bone turnover has been reported at 6.7 I.tm/day of new bone formation in the adult rat. Since the width of the periodontal ligament does not change during this drifting process, total bone formation should be balanced with resorption. Therefore an estimate of 6.7 I.tm/day of distal dental drift can be predicted. This was chosen as the "gold standard" for specificity and sensitivity assessment. The group of 72 rats previously described was used to make
this assessment and molar movement was determined as described previously. Tooth movement time courses
Three groups of 72 rats each were fitted with appliances delivering 20, 40, and 60 gm of initial tipping force to the maxillary first molars. Each force group of 72 rats was divided equally into six subgroups to be killed at 1, 3, 5, 7, 10, and 14 days. These animals were radiographed at appliance.placement and prior to death. Tooth movement was calculated for each animal as described previously. Statistical methods
The error estimate for the single measurement of molar-to-broach distance or molar-to-amalgam distance was determined by the square root of one half of the variance of the absolute difference between two replicate measures for each method. Confidence limits of a single determination were determined at the 95% level by doubling the standard deviation of the random error. The difference in amount of random error between these two methods was examined by a one-sample t test. The means and standard errors of molar movement were calculated for each time point and each force level. In addition, the means and standard errors of the movement of homologous broach implants to one another were determined. An analysis of variance (ANOVA) was performed to examine differences among force groups at each time point and across times within each force group. Paii~vise (Scheffe) comparisons were performed between individual force groups/days when ANOVA indicated that significant differences existed (p < 0.05). Linear regression analyses, using the mean movement data (molar-to-broach and broach-to-broach) of each time point obtained from pooled longitudinal
Volume 99 Number 5
(N = 11) and cross-sectional (N = 72, divided into six equal groups) zero force animals, were performed to predict daily molar drift and daily movement of homologous broaches to one another in the sagittal and transverse dimensions. Finally, linear regression analyses, using the mean molar movement data for days 1, 3, 5, and 7, were performed to examine the slope of the delay period at each force level (20, 40, and 60 gm). RESULTS Of the 300 original animals used, 14 died, 26 lost molar cleats from both sides, and 67 lost one molar cleat. Thus the success rate for the appliance in the surviving animals was 79.2%, Death and bilateral appliance loss necessitated adding six replacement animals each to the 20 and 60 gm 14-day groups. The animals showed a tendency to lose weight as a function of time in appliances. Table I demonstrates the average initial weight and weight at death along with percentage change for each of the force groups over 14 days. The addition of coils bilaterally resulted in a significant increase in weight loss. However, the amount of initial force placed on the springs had no effect on animal weight loss. There were no systematic errors between replicate films for either the molar-to-amalgam or the molar-tobroach method as indicated by the failure to demonstrate statistically significant differences between their means using Student's t test for paired data (p > 0.05). The 95% confidence limits for a single determination of molar position with the amalgam implants was estimated to be 62 p.m; with the broach implant, 47 Ixm. The difference between these two methods was significant statistically (p < 0.001). Because molar position could be determined more reliably with the broach implant and because the method was easier to perform with substantially less trauma to the animals, the use of amalgam implants as a marker in all future studies was abandoned. The 95% confidence limit for the molar-to-broach method could be reduced to 23 p.m when the average of four independent determinations was used. Thus movement data for each animal in this study represented the average of right and left sides (when available) of four independent films at each time point with the broach implant method. In regard to relative broach implant stability within the craniofacial complex as a function of time, ANOVA indicated there were differences in the transverse plane (p < 0.01) but not in the sagittal plane (p > 0.05). Simple regression equations predicted that the broach implants drifted apart 9 p.m/day laterally (R z = 0.81; p = 0.01) (Fig. 5).
D e n t a l drift a n d tooth m o v e m e n t
461
Movement of broach implants to one another over time in the craniofacial complex
0.10A
E E E ¢1) > O
0.00'
y=0.009x-0.028, -0.10
R^2=0.81, p=0.01
,
i
,
,
,
,
,
2
4
6
8
10
12
14
~ 8
, 10
J 12
J 14
Day
0.15"
E E
Sagi~al
0.05
E >
o
-0.05
-0.15 0
i 2
i 4
~ 6
Day Fig. 5. Linear regression analyses of mean implant movement in the transverse and sagittal planes with respect to its contralateral homologue vs. time (days).
The results of the ANOVA procedure performed on molar movement over time in the control (0 force) animals indicated that this movement was significantly different from zero (p < 0.001), with significant differences occurring between multiple pairs of individual days (p < 0.05). The results of the linear regression analysis on molar movement over time predicted that adult "rat molars drifted distally 7.7 I.tm/day (R 2 = 0.86, p = 0.007) (Fig. 6). The molar movement for each animal in the 20, 40, and 60 gm force groups was adjusted by subtracting this predicted distal drift. Three-part cumulative tooth movement kinetics were obtained for the 20, 40, and 60 gm initial force groups (Fig. 7). Analysis of variance indicated that all
462
King et al.
Am. J. Orthod. Dentofac. Orthop. May 1991
Molar movement in the control animals and prediction of rate of physiologic drift
Molar movement over time as a function of initial force level, adjusted for predicted distal drift. 0.5'
0.02 ']
. . . . ,,---
y = - 0.0077x + 0.003, R^2 = 0.86, p = 0.007
60 gram 40 gram
0.00'
---e---
0.4"
20 gram ~S
E
-0.02 '
E E
""
I= 0.3" t""
-0.04 •
E O
.;
:>
o 0.2"
=i
E > o
-0.06 •
3~ 0.1 -0.08 '
0
. 0
-0.10
0 2
4
~ 6
8
10
12
14
Day -0.12 0
'
2
4
6
8
10
12
14
Day Fig. 6. Cumulative maxillary molar m o v e m e n t in untreated animals and linear regression analysis of m o v e m e n t vs. time (N = 12; vertical bars represent standard errors of the mean).
time points on the 20 gm plot differed significantly (p < 0.001) from the comparable points on the 40 and 60 gm curves. However, no individual time point at the highest force level differed from its counterpart at 40 gm (p > 0.05). Analysis of variance indicated that significant differences in movement occurred among days in all three force groups (p < 0.005). Pairwise comparisons within the 40 and 60 gm groups showed that movements at 1, 3, 5, and 7 days did not differ from each other; movements at days 10 and 14 differed from that at all other days (p < 0.05). In the 20 gm initial force group, movements at days 7 and 10 differed (p < 0.05). Results of the linear regression analyses performed on the mean movement data at days 1, 3, 5, and 7 at each force level indicated no significant movement (i.e., during the delay period) in the 40 and 60 gm groups
Fig. 7. Cumulative tooth m o v e m e n t adjusted for distal drift in r e s p o n s e to 20, 40, and 60 gm inital forces tipping maxillary molars mesially. Each point represents the mean of 12 observations. Vertical bars represent standard errors of the mean.
(Fig. 8). However, there was a significant distal movement predicted at the rate of 14 I.tm/day during this period in the 20 gm animals (R'- = 0.94, p = 0.001).
DISCUSSION This experimental model for tooth movement had several positive features. The tooth movement curves obtained by initial forces between 40 and 60 gm were comparable to reports in the literature. 7 However, the marked distal drift of rodent molars did require correcting the data for expected drift. This presented no problem because the distal molar drift was linear and correlated highly with time. The method did eliminate extraneous forces from occlusion and tissue impingement from the appliances. Each of these are critically important when sensitive histologie and biochemical methods are being applied wherein molecules associated with trauma and inflammation are being studied." Extraneous environmental factors could raise background levels of such molecules,
Volume 99 Number 5
D e n t a l drift a n d tooth m o v e m e n t
463
Slope of the delay period as a function of initial force level 0.4R^2 = 0.28, NS
=
60 gram
y = - 0.003x + 0.25, R^2 = 0.08, NS
s
40 gram
a----
20gram
y = O.006x + 0.20,
0.3'
0.2' O
=E
y=-0.014x+0.14,
R^2=0.94, p=0.001
0.1
0.0 0
i
i
I
i
2
4
6
8
Day Fig. 8. Linear regression analyses of days 1, 3, 5, and 7 vs. cumulative tooth movement adjusted for distal drift (N = 12; vertical bars represent standard errors of the mean).
making the demonstration of statistically significant differences difficult. The use of bonding to place these appliances was quite feasible and easy with the headrestraining device described. Modifications also should readily allow the construction of other intraoral appliances to achieve different types of force. Removal of opposing teeth did present the problem of supereruption of the teeth being treatedS; however, this was not critical in the context of this study because movement was assessed only in the anteroposterior plane. Although this method allowed the precise loading of appliances with virtually any force level of biologic relevance, problems still remain with knowing the actual average force on the treated tissues or the force at death. The load-deflection characteristics of the springs used in this study can be determined readily and these are linear. 9 Force at death could be determined by measuring spring deflection on the cephalograms. With force at death and the initial force, an average could be calculated. However, a light force that remains active throughout the 10 to 14 days required to achieve a tooth movement curve is difficult to achieve. Tire lightest commercially
available closed coil in 9 mm lengths deflects approximately 0.2 mm with a 20 gm force. Half of that activation was lost at initial deformation and, since the maxillary incisors erupt both inferiorly and posteriorly, the point of anterior attachment of the springs moved toward deactivation. Although this did not apparently affect the tooth movement curves at the higher initial forces, the data at 20 gm initial force suggested that the appliances were deactivated shortly after initial deformation. This is the only reasonable interpretation for the reassertion of molar distal drift in this group after the initial deformation.. Distal molar movement highly correlated with time during this period (p < 0.001, R ~ = 0.94) and occurred at a rate of 14 IJ.m/day. Since distal drift in the untreated controls was measured at 7.7 p.m/day, this likely represented both relapse from the initial deformation and resumed distal drift. It is well known that when orthodontic forces are removed, the initial rate of relapse is very high. ~ It is unlikely that the springs were actually exerting a distal force since they were secured to the molars with a ligature loop and the cleat was open to the distal aspect.
464
King et al.
Although nonspecific changes in animal weight have not been demonstrated to influence orthodontic tooth movement, such a correlation could influence data comparisons among groups under certain experimental conditions. The weight losses reported here are interpreted to be the result of the extensive oral manipulations and anesthesias required ih the method. Since the magnitude of orthodontic force applied had no statistically significant effect on weight change, comparisons among force groups can be validly made. However, comparisons between the 0 force and force application groups should be approached with caution. Our findings clearly demonstrated that with this method, tooth movement was quantifiable and could readily distinguish the effects of 20 and 40 gm initial forces. Assuming that four independent cephalograms are used at both appliance placement and at death, the 95% confidence limit is 23 p.m. This was well within the initial deformation at all three force levels and would predict that significant distal molar drift could be detected by 3 days. The validity and sensitivity of this method were assessed by measuring distal molar drift. By means of histomorphometric methods, new bone formation was shown to accrue at a rate of 6.7 I.tm/day in adult rodent alveolar bone during molar drift. 6 Since the periodontal ligaments did not change in width during drift, it also was assumed that new resorption balanced new formation. It therefore seemed logical to set a distal tooth movement rate at 6.7 p.m/day. Our methods demonstrated 7.7 p.m/day of distal molar drift. This level of agreement supported the validity of this method and also suggested an acceptable precision. Orthopedic effects will not confound these tooth movement measurements if the broaches are placed on the palatal bone on the same side of the midpalatal suture as the teeth being moved. However, interbroach measurements did confirm a significant degree of transverse growth at this suture, indicating that use of a broach as a measurement landmark on the opposite side of this suture should be approached with caution. These methods proved to be sufficiently simple to perform as to enable an experienced technician to carry them out without supervision after a brief training period. Of the two methods evaluated to assess molar position, the subperiosteal broach proved to be significantly more reliable than the previously reported zygomatic amalgam approach.5 Apparently, images of the multiple barbs on the broach permitted more reproducible visual orientation of the tracing templates than the rounded outlines of the amalgam images. In addition, the broach implants could be inserted easily in seconds;
Am. J. Orthod. Dentofac. Orthop. May 1991
amalgam placement required a 45-minute surgical procedure with concomitant surgical trauma to the experimental animals. The classical tooth movement curve has three parts that represent distinctly different processes. 3'7 The initial movement begins almost instantaneously and is a reflection of tissue deformation. The second phase represents a delay in movement, which reflects recruitment of cells and the establishment of a microenvironment that will allow the appropriate tissue modeling and remodeling. The final phase represents tissue turnover to allow reduction of the applied strain terminating in appliance deactivation. Characteristic tooth movement kinetics were obtained for the 40 and 60 gm initial force groups, whereas the 20 gm group had a reduced deformation phase, distal drift during the times when the delay in tooth movement occurred in the higher force groups, and a statistically significant but abortive mesial movement at 10 days. Interpretation of these findings should be approached with some caution because neither the decay rates of the springs nor the level of force being delivered at death were determined. However, several interpretations seem reasonable'. First, a 50% increase in initial force above 40 gm did not result in a greater amount or rate of tooth movement. This finding supported the observations of other investigators, which indicated that orthodontic appliances can be overloaded, a condition that results in no further increases in the rate or amount of tooth movement achieved. ~o-~2 These findings also demonstrated that overload conditions affect all three processes represented by the clhssical tooth movement kinetics. Our findings indicated that a maximum deformation of the molar alveolar tissues was achieved at a force level between 20 and 40 gm since an increased deformation with increased force above 40 gm was not observed. The equality of the delay periods in the 40 and 60 gm data suggested that a finite amount of time is required to recruit and activate cells and stimulate enough tissue remodeling so that further tooth movement can be detected. Tran Van et al. 8 have reported that activation for bone remodeling in the alveolar process after extraction of opposing teeth required 2 to 3 days; resorption proceeded for 2 to 3 days and reversal required 4 more days. Therefore approximately 8 to 9 days are required for one bone remodeling cycle exclusive of the time required for formation. Although it is currently unknown whether orthodontic forces have any effect on this process, there are reports that peak osteoclast numbers appear in orthodontically treated tissues seveal days after appliance activation. ~The 7- to 10-day delays
Volume 99 Number 5
reported here are only slightly longer than that scale. This could represent a difference between the appearance of the cells histologically and the actual turnover of enough tissue to.detect a significant tooth movement. The most intriguing finding of the study was represented by the statistically significant, but abortive, mesial molar movement seen at 10 days in the lightest force group. If the appliances were in fact no longer active shortly after deformation, this can be interpreted only if we postulate that the remodeling signal that was initiated at deformation took from 7 to 10 days to be manifested as tooth movement and this did not require constant stress. There is only one study in the orthodontic literature in which the long-term effects of forces of limited duration were assessed. In this autoradiographic study, Kvam ~3 reported two significant waves of labeled cells in the periodontal ligament after appliance removal. In the orthopedic literature, there is considerable evidence that deformations of short duration do result in substantial bone remodeling. Tennis players have been reported to have more and denser bone in their humeri on the playing side, with cortical thickness 28% to 34% increased. ~4Loads within the physiologic range applied for 1 h o u r / d a y in sheep radii and ulnae resulted in significant bone remodeling. ~5With an avian model, Rubin and Lanyon t6-~s reported that removal of load bearing resulted in loss of bone in the wing and that constant loads within the physiologic range were not effective in its prevention. However, intermittent compressive forces of 100 seconds/day for 8 weeks resulted in a 24% increase in cross-sectional diameter of these bones. Although it remains premature to propose such an effect in orthodontic tooth movement on the basis of this experiment, this possibility will be explored in future studies. Such information could be significant for determining the precise nature and timing of the most effective orthodontic signals for clinical problems. CONCLUSIONS 1. The method presented herein is a reliable, sensitive, and valid technique for quantifying mesial orthodontic tooth movement using tipping forces on the rat molar. 2. Maxillary molars drift distally at a rate of 7.7 p.m/day in adult male Sprague-Dawley rats. 3. Palatally placed barbed broaches represent a more reliable, less traumatic, and more easily executed superpositional landmark than zygomatic amalgams. 4. Bilaterally placed implants do not migrate with respect to each other in the sagittal plane but do move laterally at a rate of 9 I.tm/day, necessitating their place-
D e n t a l drift a n d tooth m o v e m e n t
465
ment on same side of the midpalatal suture as the molar being tested. 5. Orthodontic appliances can be overloaded, resuiting in no further increases in tooth movement. .6. Overload conditions equally affect all three processes represented by the classical tooth movement kinetics. REFERENCES 1. YamasakiK. The role of cystic AMP,calciumand prostaglandins in boneresorptionassociatedwithexperimentaltooth movement. J Dent Res 1983;62:877-81. 2. Storey E. The nature of tooth movement. AM J OR'rHOD 1973;63:292-314. 3. Roberts WE, Chase DC. Kinetics of cell proliferationand migration associated with orthodontieally-inducedosteogenesls. J Dent Res 1981;60:!74-81. 4. DavidovitchZ, Nicolay O, Alley K, Zwilling B, Lanese R, Shanfeld J. First and second messengerinteractionsin stressed connectivetissues in vivo. In: Norton L, BurstoneC, eds. The biology of tooth movement. Boca Raton, Florida: CRC Press, i 989:97-129. 5. TuncayOC, KillianyDM. The effect of gingival fiberotomyon the rate of tooth movement. AM J ORTHOD1986;89:212-5. 6. Vignery A, Baron R. Dynamic histomorphometryof alveolar bone remodelingin the adult rat. Anat Ree 1980;196:191-200. 7. Reitan K. Clinical and histologicalobservationson tooth movement during and after orthodontic treatment. AM J OR'H-IOD 1967;53:721-45. 8. TranVan P, VigneryA, BaronR. Cellularkienticsof the alveolar bone remodeling sequence in the rat. Anat Rec 1982;202:44551. 9. Boshart BF, Currier GF, Nanda RS, Duncanson MG. Loaddeflection rate measurementsof activated open and closed coil springs. Angle Orthod 1990;60:27-34. 10. Boester CH, Johnston LE. A clinical investigationof the concepts of differentialand optimalforce on canineretraction.Angle Orthod 1974;44:113-9. 11. BurstoneCJ, Groves MA. Threshold and optimumforce values for maxillarytooth movement. J Dent Res 1961;39:695. 12. Hixon EH, Aasen TO, Arango J, et al. On force and tooth movement. AM J ORTItOn1970;57:476-89. 13. Kvam E. Cellular dynamicson the pressure side of the rat periodontiumfollowingexperimentaltooth movement.ScandJ Dent Res 1972;80:369-83. 14. JonesJ J, Prist JP, Hayes WC, TichenorCC, Nagel DA. Humeral hypertrophy in response to exercise. J Bone Joint Surg 1977; 61A:539-46. 15. O'Connor JA, Lanyon LE..The influenceof strain rate on adaptive bone remodeling. J Biomech 1982;15:727. 16. LanyonLE, Rubin CT. Static vs. dynamicloads as an influence on bone remodeling. J Biomech1984;17:897-905. 17. RubinCT, LanyonLE. Regulationof bone formationby applied dynamic loads. J Bone Joint Surg 1984;66A:397-402. 18. RubinCT, Lanyon LE. Regulationof bone mass by mechanical strain magnitude. Calcif Tissue 1985;37:411-7. Reprhlt requests to:
Dr. Gregory J. King Department of Orthodontics Box J-444, JHMHC Universityof Florida Gainesville, FL 32610
A method for the quantification of orthodontic tooth movement in the rat is presented. Reliability, sensitivity, and validity were assessed and tooth movement kinetics were determined for initial forces of 20, 40, and 60 gin. The appliance consisted of a 9 mm length of closed coil spring suspended between a cleat and bonded to the occlusal surface of the maxillary first molars and the maxillary incisors. Initial tipping forces were placed by suspending known weights from the anterior end of these coils before fixation to the incisors. Tooth movement was quantified from enlarged cepnalograms by measuring the position of a reproducible landmark on the molar cleat with respect to either zygomatic amalgam implants or a barbed broach placed submucosally on the palate. All measurements were made along the molar-incisor vector by projecting at 90 ° to this line. Validity and sensitivity were assessed by quantifying molar distal drift and comparing these results with reports of bone turnover rates adjacent to distally drifting adult rat molars. Reliability was obtained by estimating the error of a single measurement in a longitudinal study of 12 adult male Sprague-Dawley rats (180 to 200 days) receiving both amalgam and broach implants and a cross-sectional study of 72 animals divided equally into six groups to be killed at 1, 3, 5, 7, 10, and 14 days. No orthodontic forces were used in this portion of the study. Implant stability within the craniofacial complex was assessed by measuring bilateral broaches as a function of time with respect to each other. There were no systematic errors between replicate films for either the amalgam or the broach method. The 95% confidence limit for a single determination of molar position was 62 ixm using the amalgams and 47 p.m for the broach (p < 0.001). The latter could be reduced to 23 rtm when the average of four independent determinations was used. Homologous implants did not differ with respect to each other in the sagittal plane but did in the transverse plane (p < 0.01), migrating laterally 9 ixm/day. Linear regression analysis of molar distal movement over time predicted 7.7 ttm/day distal drift (p < 0.01), which compared favorably with reports of 6.7 p.m/day of alveolar bone turnover during this drifting process. Characteristic three-part cumulative tooth movement kinetics were obtained for the 40 and 60 gm initial force groups. No individual time point at 60 gm differed from its counterpart at 40 gm. All time points on the 20 gm curve differed significantly from those on the higher force curves (p < 0.001). Comparisons within the highest force group indicated that mesial molar movement differed from zero at all times and that 1, 3, 5, and 7 days did not differ from each other, but 10 and 14 days differed from several other days. In the 20 gm group, day 10 differed from day 7. In the 20 gm group, distal molar drift resumed shortly after initial deformation at 14 i.tm/day. This method is sufficiently reliable, precise, and valid for the quantification of orthodontic tooth movement under experimental conditions. Tooth movement in response to varying initial forces indicated that (1) orthodontic appliances can be overloaded, resulting in no further enhancement of tooth movement and (2) under, conditions of overload, all three processes represented in the tooth movement curve are affected equally. (AMJ ORTHOO OENTOFACORTHOP 1991 ;99:456-65.)
A
reliable, sensitive, and valid in vivo model for orthodontic tooth movement would facilitate studies on biologic mechanisms and studies on how From the Department of Orthodontics, College of Dentistry, Unive~ity of Hofida. Supported by NIDR Grant DE-08659. 8/1122216
456
various physical and pharamcologic manipulations affect clinical outcomes. The interpretation of most available tooth movement data is unclear because of several deficiencies in the models currently being used: (1) tooth movement often is not quantified in descriptive histologic studies, ~4 (2) continuously erupting teeth are used to represent human tooth movement, 2 (3) over-
Voh~me 99 Number 5
Dental drift and tooth movement
m
457
!
Fig. 1. Rat positioned in the restrainer for appliance activation. Note bilateral mouth props/suction devices and tongue retraction with a suture tied back to one of the posterior pillars.
lapping orthopedic and orthodontic changes are not factored out,'- (4) various levels of tissue trauma caused by the appliances confuse similar changes that are specific to orthodontic tooth movement,' (5) extraneous occlusal forces are not eliminated, j4 (6) the reliability, sensitivity, and validity of the methods are unevaluated, ~4 and (7) appliance force magnitudes are imprecise or unknown."3 The purposes of this study were (1) to test the reliability, sensitivity, and validity of a proposed rodent model for orthodontic tooth movement that addresses most of these problems and (2) then to apply it to a study of the time course of tooth movement changes in response to precise initial forces. MATERIALS AND METHODS Animals Adult male Sprague-Dawley strain rats (180 to 200 days old) were used because of availability, cost, genetic homogeneity, and, at this age, a relatively slow rate of growth. These methods can be adapted to younger animals to address questions requiring a more rapidly growing animal. Head-holding device This instrument was designed to serve as a surgical restrainer. It was fabricated from Lucite and consisted of a base (6.5 x 9 inch) with two vertical pillars supporting adjustable ear rods that were coplanar and parallel to the base. Two vertical pillars also were placed on the base caudal to the animal's head to serve as convenient points for tying tongue-retraction sutures. Two removable pulleys were placed rostral to the an-
imal's head to accommodate appliance activation procedures. Two more pillars were positioned on either side of the pulleys. These pillars were equipped with adjustable suction devices, which were fabricated from aluminum tubing (3/,6 inch diameter). One end of the tubing was flattened, sealed with solder, and modified by placing six slits in it. These could be positioned to provide both cheek retraction and a dry intraoral field. The animal was positioned with the head up. The mouth was propped with the suction devices in place bilaterally. The tongue was retracted by placing a suture (6-0 silk) through it and tying this to one of the pillars located caudal to the head. This format permitted unassisted placement and activation of all appliances and surgical manipulations (Fig. l ) . Appliances and animal manipulations Rats were weighed and anesthetized with intramuscular ketamine (87 mg/kg) and xylazine (13 mg/kg). After mounting the animal in the head holder, the occlusal surfaces of both maxillary first molars were prepared by roughing the surface with a green stone, washing with acetone, etching with a phosphoric acid gel for 3 minutes (Healthco International, Boston, Mass.) and sealing with an orthodontic composite resin (System I, Ormco, Glendora, Calif.). Direct-bonded orthodontic cleats (American Orthodontics, Sheboygan, Wis., No. 593-852-165) were modified by removing one of the cleat arms and reducing and reshaping the base to a trapezoid with the apex oriented opposite the remaining cleat arm. This attachment was rinsed in acetone, air-dried, and treated frst with sealant followed by bonding material (System
458
K h l g et al.
Am. J. Orthod. Dentofac. Orthop. May 1991
r
l
Fig. 2. Rat positioned in the restrainer for appliance activation.
I). The attachment then was placed on the occlusal surface of the prepared maxillary first molar with the trapezoidal apex on the mesial aspect and the cleat arm pointing distally. A 6 to 7 mm length of coarse endodontic barbed broach with a 0.67 mm interbarb distance (Maillefor, Switzerland) then was inserted through the palatal mucosa along the palatal bone lingual to the maxillary first molars. The broach was inserted using a hemostat and the end was pushed below the tissue with an explorer, leaving a pinpoint insertion wound that healed readily. Before release of the rat, mandibular first and second molars were extracted with scissors used to remove the crowns and an explorer to elevate the major roots. The rat then was allowed to recover for approximately 1 week by monitoring wound healing and weight gain. For appliance activations the animal again was placed head up in the restrainer as described previously. One end of a 9-mm length of closed coil (0.006 inch Hi T; arbor diameter: 0.022 inch, Unitek, Monrovia, Calif.) was attached to a steel ligature wire loop (0.009 inch). This then was slipped over the molar cleat. The anterior end of the coil was spread apart and attached with suture material to a premeasured weight (Fig. 2). The weight was suspended over the pulley and allowed to hang freely, exerting the prescribed intial force through the coil to the molar. The anterior end of the coil then was bonded to the lateral surface of the maxillary incisor by means of the same procedure as described for attaching the cleat to the maxillary molar. The excess coil and suspended weight were removed and the mandibualt incisal edges were trimmed with scissors.
Measurement of tooth movement
Radiographic exposures were made using a second head-holding device (Fig. 3). This instrument was a simplified version of the restrainer and consisted of ear rods for positioning the rat face down on the base. The base was equipped with a spirit level and four height adjustment knobs to facilitate precise leveling of the instrument. A film packet was placed on the base under the animal's head. The device was placed on the floor and radiographed at a 127 cm object-to-film distance (10 mA, 60 kVp, 4 seconds) with the central ray perpendicular to the base. These cephalograms routinely displayed the implants, molar cleats, and orthodontic springs (Fig. 4, A). These were mounted in slide holders and viewed at approximately × 8 magnification using a rear-screen projection device (Ektagraphic 200 Audioviewer; Kodak, Rochester, N.Y.). Corrections were not made for lens distortion. However, all cephalometric exposures were made with uniform positioning; cephalograms were mounted uniformly and these were all projected with the same lens in an effort to control for this variable across cephalograms, animals, and experimental groups. With acetate tracing paper, a template was made for the following cephalometric structures: the broaches with their adjacent barbs, the molar cleats, and the incisors. By means of these templates, landmark points were selected (1) at the intersection of a barb with its broach body, (2) at the trapezoidal apex of the molar cleat, and (3) on the distal surface of the incisor (at the location of the spring attachment in the animals with force applications). These templates were oriented by best-fit visual inspection on subsequent radiographs and
Vohtme 99
Dental drift and tooth movement
459
Number 5
X-R~ ...............
•
i
127cm
ilm
Fig. 3. Rat positioned for a cephalometric exposure.
the landmarks for each structure registered on a second, overlying acetate paper (Fig. 4, B). The choice of a cephalometric landmark for the tooth is problematical because the amount of tooth movement measured will vary as a function of the landmark's distance from the center of rotation of the tipping tooth. Since the issue of greatest importance in the choice of a dental landmark for this study was reproducibility, the mesiooccclusal surface was chosen for its accessibility and anatomic uniformity. This location also offers the advantage of not traumatizing the periodontal ligament. These tracings were digitized to record the X and Y coordinates of these cephalometric landmarks using a Hipad digitizing tablet (Houston Instruments, Austin, Texas) interfaced with a microcomputer (IBM-PC/AT, IBM Corp., Atlanta, Ga.). A computer program measured distances, adjusting for magnification, by using the known, digitized interbarb distances of the broach. The molar and implant landmarks were projected at 90 ° to the molar-incisor vector and the distance from the tooth to the implant measured. The difference between these measurements made on appliance placement and radiographs taken at death was calculated and reported as tooth movement. The mean of right and left measurements represented molar movement in each animal. Error assessments
The reliability of two methods of measuring molar position (one for molar-to-broach implant; the second, for molar-to-amalgam implant) was determined by estimating the error of a single measurement. This was accomplished in a longitudinal study of 12 animals that received both barbed broach implants as described pre-
Adult JIG Force 0 gins Osy Appllonce 3"emplate
°O
{o Fig. 4. A, Representative cephalogram demonstrating broach and amalgam implants and molar cleats. B, Tracing of the landmarks from a cephalogram.
viously and intraosseous amalgam implants as described by Tuncay and Killiany. 5 The latter were placed in a triangular array by making a single incision medial to the dental arch, extending from 1 mm mesial to the maxillary first molar to the zygomatic maxillary suture and reflecting the tissue laterally to expose the lateral wall of the maxilla. An 18-gauge needle with an 0.16
460
King et al.
Am. J. Orthod. Oentofac. Orthop. May 1991
Table I. Weight and percentage weight change of the 14-day animals by group*
Weight at death (gin)
Initial weight (gm) Force group
Mean
0 20 40 60
495.83 500.77 589.00 448.00
gm gm gm gm
[
SE
Mean
4.64 10.08 19.89 7.16
407.08 375.77 425.50 333.50
[
Change? (%)
SE 10.97 9.84 19.91 12.32
Mean 17.9 - 24.9 - 27.9 -25.7
-
[
SE 2.0 1.4 1.7 2.1
*Percentage weight change equals (weight at death-initial weight)/initial weight. tAnalysis of variance indicated significant differences existed among the groups in amount of weight loss (p = 0.001). Scheffe pairwise comparisons indicated these differen.ces were manifested between the 0 force group and each of the other groups (p < 0.05); no significant differences in weight loss were detected between the 20, 40, and 60 gm groups (p > 0.05).
inch stainless steel wire inserted in the lumen was adapted to serve as an amalgam carrier. Holes to receive the amalgam were prepared with a Y2 round burr. The tissue was sutured with 6-0 silk. Molar cleats were used to facilitate tooth localization but springs were not placed. Templates of the amalgam images were constructed and landmarks selected. Measurements of molar positions in relation to both broach and amalgam implants were obtained from the tracings of 132 radiographs representing two independent radiographs taken on each of I l rats on days 1,3, 5, 7, 10, and 14. (Note: one of the original rats died after day 1.) The radiographs were obtained after rcpositioning the rat between exposures: Implant stability within craniofacial complex
The relative stability of the barbed broaches was examined by comparing movement of right and left homologous implants in the sagittal and transverse dimensions at each time point on a group of 72 rats prepared with bilateral molar cleats and broaches but not orthodontic springs. These rats were divided equally into six subgroups to be killed at 1, 3, 5, 7, I0, and 14 days. Movement of the broaches was referenced to the right molar-incisor vector. Validation of method for measurement of tooth movement
Rodent molars are known to drift distally. 6 Histologic measurement of alveolar bone turnover has been reported at 6.7 I.tm/day of new bone formation in the adult rat. Since the width of the periodontal ligament does not change during this drifting process, total bone formation should be balanced with resorption. Therefore an estimate of 6.7 I.tm/day of distal dental drift can be predicted. This was chosen as the "gold standard" for specificity and sensitivity assessment. The group of 72 rats previously described was used to make
this assessment and molar movement was determined as described previously. Tooth movement time courses
Three groups of 72 rats each were fitted with appliances delivering 20, 40, and 60 gm of initial tipping force to the maxillary first molars. Each force group of 72 rats was divided equally into six subgroups to be killed at 1, 3, 5, 7, 10, and 14 days. These animals were radiographed at appliance.placement and prior to death. Tooth movement was calculated for each animal as described previously. Statistical methods
The error estimate for the single measurement of molar-to-broach distance or molar-to-amalgam distance was determined by the square root of one half of the variance of the absolute difference between two replicate measures for each method. Confidence limits of a single determination were determined at the 95% level by doubling the standard deviation of the random error. The difference in amount of random error between these two methods was examined by a one-sample t test. The means and standard errors of molar movement were calculated for each time point and each force level. In addition, the means and standard errors of the movement of homologous broach implants to one another were determined. An analysis of variance (ANOVA) was performed to examine differences among force groups at each time point and across times within each force group. Paii~vise (Scheffe) comparisons were performed between individual force groups/days when ANOVA indicated that significant differences existed (p < 0.05). Linear regression analyses, using the mean movement data (molar-to-broach and broach-to-broach) of each time point obtained from pooled longitudinal
Volume 99 Number 5
(N = 11) and cross-sectional (N = 72, divided into six equal groups) zero force animals, were performed to predict daily molar drift and daily movement of homologous broaches to one another in the sagittal and transverse dimensions. Finally, linear regression analyses, using the mean molar movement data for days 1, 3, 5, and 7, were performed to examine the slope of the delay period at each force level (20, 40, and 60 gm). RESULTS Of the 300 original animals used, 14 died, 26 lost molar cleats from both sides, and 67 lost one molar cleat. Thus the success rate for the appliance in the surviving animals was 79.2%, Death and bilateral appliance loss necessitated adding six replacement animals each to the 20 and 60 gm 14-day groups. The animals showed a tendency to lose weight as a function of time in appliances. Table I demonstrates the average initial weight and weight at death along with percentage change for each of the force groups over 14 days. The addition of coils bilaterally resulted in a significant increase in weight loss. However, the amount of initial force placed on the springs had no effect on animal weight loss. There were no systematic errors between replicate films for either the molar-to-amalgam or the molar-tobroach method as indicated by the failure to demonstrate statistically significant differences between their means using Student's t test for paired data (p > 0.05). The 95% confidence limits for a single determination of molar position with the amalgam implants was estimated to be 62 p.m; with the broach implant, 47 Ixm. The difference between these two methods was significant statistically (p < 0.001). Because molar position could be determined more reliably with the broach implant and because the method was easier to perform with substantially less trauma to the animals, the use of amalgam implants as a marker in all future studies was abandoned. The 95% confidence limit for the molar-to-broach method could be reduced to 23 p.m when the average of four independent determinations was used. Thus movement data for each animal in this study represented the average of right and left sides (when available) of four independent films at each time point with the broach implant method. In regard to relative broach implant stability within the craniofacial complex as a function of time, ANOVA indicated there were differences in the transverse plane (p < 0.01) but not in the sagittal plane (p > 0.05). Simple regression equations predicted that the broach implants drifted apart 9 p.m/day laterally (R z = 0.81; p = 0.01) (Fig. 5).
D e n t a l drift a n d tooth m o v e m e n t
461
Movement of broach implants to one another over time in the craniofacial complex
0.10A
E E E ¢1) > O
0.00'
y=0.009x-0.028, -0.10
R^2=0.81, p=0.01
,
i
,
,
,
,
,
2
4
6
8
10
12
14
~ 8
, 10
J 12
J 14
Day
0.15"
E E
Sagi~al
0.05
E >
o
-0.05
-0.15 0
i 2
i 4
~ 6
Day Fig. 5. Linear regression analyses of mean implant movement in the transverse and sagittal planes with respect to its contralateral homologue vs. time (days).
The results of the ANOVA procedure performed on molar movement over time in the control (0 force) animals indicated that this movement was significantly different from zero (p < 0.001), with significant differences occurring between multiple pairs of individual days (p < 0.05). The results of the linear regression analysis on molar movement over time predicted that adult "rat molars drifted distally 7.7 I.tm/day (R 2 = 0.86, p = 0.007) (Fig. 6). The molar movement for each animal in the 20, 40, and 60 gm force groups was adjusted by subtracting this predicted distal drift. Three-part cumulative tooth movement kinetics were obtained for the 20, 40, and 60 gm initial force groups (Fig. 7). Analysis of variance indicated that all
462
King et al.
Am. J. Orthod. Dentofac. Orthop. May 1991
Molar movement in the control animals and prediction of rate of physiologic drift
Molar movement over time as a function of initial force level, adjusted for predicted distal drift. 0.5'
0.02 ']
. . . . ,,---
y = - 0.0077x + 0.003, R^2 = 0.86, p = 0.007
60 gram 40 gram
0.00'
---e---
0.4"
20 gram ~S
E
-0.02 '
E E
""
I= 0.3" t""
-0.04 •
E O
.;
:>
o 0.2"
=i
E > o
-0.06 •
3~ 0.1 -0.08 '
0
. 0
-0.10
0 2
4
~ 6
8
10
12
14
Day -0.12 0
'
2
4
6
8
10
12
14
Day Fig. 6. Cumulative maxillary molar m o v e m e n t in untreated animals and linear regression analysis of m o v e m e n t vs. time (N = 12; vertical bars represent standard errors of the mean).
time points on the 20 gm plot differed significantly (p < 0.001) from the comparable points on the 40 and 60 gm curves. However, no individual time point at the highest force level differed from its counterpart at 40 gm (p > 0.05). Analysis of variance indicated that significant differences in movement occurred among days in all three force groups (p < 0.005). Pairwise comparisons within the 40 and 60 gm groups showed that movements at 1, 3, 5, and 7 days did not differ from each other; movements at days 10 and 14 differed from that at all other days (p < 0.05). In the 20 gm initial force group, movements at days 7 and 10 differed (p < 0.05). Results of the linear regression analyses performed on the mean movement data at days 1, 3, 5, and 7 at each force level indicated no significant movement (i.e., during the delay period) in the 40 and 60 gm groups
Fig. 7. Cumulative tooth m o v e m e n t adjusted for distal drift in r e s p o n s e to 20, 40, and 60 gm inital forces tipping maxillary molars mesially. Each point represents the mean of 12 observations. Vertical bars represent standard errors of the mean.
(Fig. 8). However, there was a significant distal movement predicted at the rate of 14 I.tm/day during this period in the 20 gm animals (R'- = 0.94, p = 0.001).
DISCUSSION This experimental model for tooth movement had several positive features. The tooth movement curves obtained by initial forces between 40 and 60 gm were comparable to reports in the literature. 7 However, the marked distal drift of rodent molars did require correcting the data for expected drift. This presented no problem because the distal molar drift was linear and correlated highly with time. The method did eliminate extraneous forces from occlusion and tissue impingement from the appliances. Each of these are critically important when sensitive histologie and biochemical methods are being applied wherein molecules associated with trauma and inflammation are being studied." Extraneous environmental factors could raise background levels of such molecules,
Volume 99 Number 5
D e n t a l drift a n d tooth m o v e m e n t
463
Slope of the delay period as a function of initial force level 0.4R^2 = 0.28, NS
=
60 gram
y = - 0.003x + 0.25, R^2 = 0.08, NS
s
40 gram
a----
20gram
y = O.006x + 0.20,
0.3'
0.2' O
=E
y=-0.014x+0.14,
R^2=0.94, p=0.001
0.1
0.0 0
i
i
I
i
2
4
6
8
Day Fig. 8. Linear regression analyses of days 1, 3, 5, and 7 vs. cumulative tooth movement adjusted for distal drift (N = 12; vertical bars represent standard errors of the mean).
making the demonstration of statistically significant differences difficult. The use of bonding to place these appliances was quite feasible and easy with the headrestraining device described. Modifications also should readily allow the construction of other intraoral appliances to achieve different types of force. Removal of opposing teeth did present the problem of supereruption of the teeth being treatedS; however, this was not critical in the context of this study because movement was assessed only in the anteroposterior plane. Although this method allowed the precise loading of appliances with virtually any force level of biologic relevance, problems still remain with knowing the actual average force on the treated tissues or the force at death. The load-deflection characteristics of the springs used in this study can be determined readily and these are linear. 9 Force at death could be determined by measuring spring deflection on the cephalograms. With force at death and the initial force, an average could be calculated. However, a light force that remains active throughout the 10 to 14 days required to achieve a tooth movement curve is difficult to achieve. Tire lightest commercially
available closed coil in 9 mm lengths deflects approximately 0.2 mm with a 20 gm force. Half of that activation was lost at initial deformation and, since the maxillary incisors erupt both inferiorly and posteriorly, the point of anterior attachment of the springs moved toward deactivation. Although this did not apparently affect the tooth movement curves at the higher initial forces, the data at 20 gm initial force suggested that the appliances were deactivated shortly after initial deformation. This is the only reasonable interpretation for the reassertion of molar distal drift in this group after the initial deformation.. Distal molar movement highly correlated with time during this period (p < 0.001, R ~ = 0.94) and occurred at a rate of 14 IJ.m/day. Since distal drift in the untreated controls was measured at 7.7 p.m/day, this likely represented both relapse from the initial deformation and resumed distal drift. It is well known that when orthodontic forces are removed, the initial rate of relapse is very high. ~ It is unlikely that the springs were actually exerting a distal force since they were secured to the molars with a ligature loop and the cleat was open to the distal aspect.
464
King et al.
Although nonspecific changes in animal weight have not been demonstrated to influence orthodontic tooth movement, such a correlation could influence data comparisons among groups under certain experimental conditions. The weight losses reported here are interpreted to be the result of the extensive oral manipulations and anesthesias required ih the method. Since the magnitude of orthodontic force applied had no statistically significant effect on weight change, comparisons among force groups can be validly made. However, comparisons between the 0 force and force application groups should be approached with caution. Our findings clearly demonstrated that with this method, tooth movement was quantifiable and could readily distinguish the effects of 20 and 40 gm initial forces. Assuming that four independent cephalograms are used at both appliance placement and at death, the 95% confidence limit is 23 p.m. This was well within the initial deformation at all three force levels and would predict that significant distal molar drift could be detected by 3 days. The validity and sensitivity of this method were assessed by measuring distal molar drift. By means of histomorphometric methods, new bone formation was shown to accrue at a rate of 6.7 I.tm/day in adult rodent alveolar bone during molar drift. 6 Since the periodontal ligaments did not change in width during drift, it also was assumed that new resorption balanced new formation. It therefore seemed logical to set a distal tooth movement rate at 6.7 p.m/day. Our methods demonstrated 7.7 p.m/day of distal molar drift. This level of agreement supported the validity of this method and also suggested an acceptable precision. Orthopedic effects will not confound these tooth movement measurements if the broaches are placed on the palatal bone on the same side of the midpalatal suture as the teeth being moved. However, interbroach measurements did confirm a significant degree of transverse growth at this suture, indicating that use of a broach as a measurement landmark on the opposite side of this suture should be approached with caution. These methods proved to be sufficiently simple to perform as to enable an experienced technician to carry them out without supervision after a brief training period. Of the two methods evaluated to assess molar position, the subperiosteal broach proved to be significantly more reliable than the previously reported zygomatic amalgam approach.5 Apparently, images of the multiple barbs on the broach permitted more reproducible visual orientation of the tracing templates than the rounded outlines of the amalgam images. In addition, the broach implants could be inserted easily in seconds;
Am. J. Orthod. Dentofac. Orthop. May 1991
amalgam placement required a 45-minute surgical procedure with concomitant surgical trauma to the experimental animals. The classical tooth movement curve has three parts that represent distinctly different processes. 3'7 The initial movement begins almost instantaneously and is a reflection of tissue deformation. The second phase represents a delay in movement, which reflects recruitment of cells and the establishment of a microenvironment that will allow the appropriate tissue modeling and remodeling. The final phase represents tissue turnover to allow reduction of the applied strain terminating in appliance deactivation. Characteristic tooth movement kinetics were obtained for the 40 and 60 gm initial force groups, whereas the 20 gm group had a reduced deformation phase, distal drift during the times when the delay in tooth movement occurred in the higher force groups, and a statistically significant but abortive mesial movement at 10 days. Interpretation of these findings should be approached with some caution because neither the decay rates of the springs nor the level of force being delivered at death were determined. However, several interpretations seem reasonable'. First, a 50% increase in initial force above 40 gm did not result in a greater amount or rate of tooth movement. This finding supported the observations of other investigators, which indicated that orthodontic appliances can be overloaded, a condition that results in no further increases in the rate or amount of tooth movement achieved. ~o-~2 These findings also demonstrated that overload conditions affect all three processes represented by the clhssical tooth movement kinetics. Our findings indicated that a maximum deformation of the molar alveolar tissues was achieved at a force level between 20 and 40 gm since an increased deformation with increased force above 40 gm was not observed. The equality of the delay periods in the 40 and 60 gm data suggested that a finite amount of time is required to recruit and activate cells and stimulate enough tissue remodeling so that further tooth movement can be detected. Tran Van et al. 8 have reported that activation for bone remodeling in the alveolar process after extraction of opposing teeth required 2 to 3 days; resorption proceeded for 2 to 3 days and reversal required 4 more days. Therefore approximately 8 to 9 days are required for one bone remodeling cycle exclusive of the time required for formation. Although it is currently unknown whether orthodontic forces have any effect on this process, there are reports that peak osteoclast numbers appear in orthodontically treated tissues seveal days after appliance activation. ~The 7- to 10-day delays
Volume 99 Number 5
reported here are only slightly longer than that scale. This could represent a difference between the appearance of the cells histologically and the actual turnover of enough tissue to.detect a significant tooth movement. The most intriguing finding of the study was represented by the statistically significant, but abortive, mesial molar movement seen at 10 days in the lightest force group. If the appliances were in fact no longer active shortly after deformation, this can be interpreted only if we postulate that the remodeling signal that was initiated at deformation took from 7 to 10 days to be manifested as tooth movement and this did not require constant stress. There is only one study in the orthodontic literature in which the long-term effects of forces of limited duration were assessed. In this autoradiographic study, Kvam ~3 reported two significant waves of labeled cells in the periodontal ligament after appliance removal. In the orthopedic literature, there is considerable evidence that deformations of short duration do result in substantial bone remodeling. Tennis players have been reported to have more and denser bone in their humeri on the playing side, with cortical thickness 28% to 34% increased. ~4Loads within the physiologic range applied for 1 h o u r / d a y in sheep radii and ulnae resulted in significant bone remodeling. ~5With an avian model, Rubin and Lanyon t6-~s reported that removal of load bearing resulted in loss of bone in the wing and that constant loads within the physiologic range were not effective in its prevention. However, intermittent compressive forces of 100 seconds/day for 8 weeks resulted in a 24% increase in cross-sectional diameter of these bones. Although it remains premature to propose such an effect in orthodontic tooth movement on the basis of this experiment, this possibility will be explored in future studies. Such information could be significant for determining the precise nature and timing of the most effective orthodontic signals for clinical problems. CONCLUSIONS 1. The method presented herein is a reliable, sensitive, and valid technique for quantifying mesial orthodontic tooth movement using tipping forces on the rat molar. 2. Maxillary molars drift distally at a rate of 7.7 p.m/day in adult male Sprague-Dawley rats. 3. Palatally placed barbed broaches represent a more reliable, less traumatic, and more easily executed superpositional landmark than zygomatic amalgams. 4. Bilaterally placed implants do not migrate with respect to each other in the sagittal plane but do move laterally at a rate of 9 I.tm/day, necessitating their place-
D e n t a l drift a n d tooth m o v e m e n t
465
ment on same side of the midpalatal suture as the molar being tested. 5. Orthodontic appliances can be overloaded, resuiting in no further increases in tooth movement. .6. Overload conditions equally affect all three processes represented by the classical tooth movement kinetics. REFERENCES 1. YamasakiK. The role of cystic AMP,calciumand prostaglandins in boneresorptionassociatedwithexperimentaltooth movement. J Dent Res 1983;62:877-81. 2. Storey E. The nature of tooth movement. AM J OR'rHOD 1973;63:292-314. 3. Roberts WE, Chase DC. Kinetics of cell proliferationand migration associated with orthodontieally-inducedosteogenesls. J Dent Res 1981;60:!74-81. 4. DavidovitchZ, Nicolay O, Alley K, Zwilling B, Lanese R, Shanfeld J. First and second messengerinteractionsin stressed connectivetissues in vivo. In: Norton L, BurstoneC, eds. The biology of tooth movement. Boca Raton, Florida: CRC Press, i 989:97-129. 5. TuncayOC, KillianyDM. The effect of gingival fiberotomyon the rate of tooth movement. AM J ORTHOD1986;89:212-5. 6. Vignery A, Baron R. Dynamic histomorphometryof alveolar bone remodelingin the adult rat. Anat Ree 1980;196:191-200. 7. Reitan K. Clinical and histologicalobservationson tooth movement during and after orthodontic treatment. AM J OR'H-IOD 1967;53:721-45. 8. TranVan P, VigneryA, BaronR. Cellularkienticsof the alveolar bone remodeling sequence in the rat. Anat Rec 1982;202:44551. 9. Boshart BF, Currier GF, Nanda RS, Duncanson MG. Loaddeflection rate measurementsof activated open and closed coil springs. Angle Orthod 1990;60:27-34. 10. Boester CH, Johnston LE. A clinical investigationof the concepts of differentialand optimalforce on canineretraction.Angle Orthod 1974;44:113-9. 11. BurstoneCJ, Groves MA. Threshold and optimumforce values for maxillarytooth movement. J Dent Res 1961;39:695. 12. Hixon EH, Aasen TO, Arango J, et al. On force and tooth movement. AM J ORTItOn1970;57:476-89. 13. Kvam E. Cellular dynamicson the pressure side of the rat periodontiumfollowingexperimentaltooth movement.ScandJ Dent Res 1972;80:369-83. 14. JonesJ J, Prist JP, Hayes WC, TichenorCC, Nagel DA. Humeral hypertrophy in response to exercise. J Bone Joint Surg 1977; 61A:539-46. 15. O'Connor JA, Lanyon LE..The influenceof strain rate on adaptive bone remodeling. J Biomech 1982;15:727. 16. LanyonLE, Rubin CT. Static vs. dynamicloads as an influence on bone remodeling. J Biomech1984;17:897-905. 17. RubinCT, LanyonLE. Regulationof bone formationby applied dynamic loads. J Bone Joint Surg 1984;66A:397-402. 18. RubinCT, Lanyon LE. Regulationof bone mass by mechanical strain magnitude. Calcif Tissue 1985;37:411-7. Reprhlt requests to:
Dr. Gregory J. King Department of Orthodontics Box J-444, JHMHC Universityof Florida Gainesville, FL 32610
