The Effects of Masseter Muscle Paralysis on Facial Bone Growth1

Journal of Surgical Research 139, 243–252 (2007) doi:10.1016/j.jss.2006.09.003

Damir B. Matic, M.D.,*,2 Arjang Yazdani, M.D.,* R. Glenn Wells, Ph.D.,† Ting Y. Lee, Ph.D.,† and Bing S. Gan, M.D.* *Division of Plastic and Reconstructive Surgery; and †Department of Medical Biophysics and Department of Diagnostic Radiology and Nuclear Medicine, University of Western Ontario, London, Canada

Submitted for publication July 16, 2006

Key Words: facial growth; masseter; botulin toxin; Background. Understanding the effects of muscle muscle paralysis; rabbit; SPECT; cephalometrics; fa- function on facial bone growth may help us treat chil- cial anomalies. dren with facial anomalies. Facial bone growth is known to be a result of both genetic and epigenetic INTRODUCTION influences. One of the main epigenetic factors controlling growth is thought to be muscle action. The pur- It has been shown that children born with congenital pose of this study was to establish a model of single facial anomalies have disturbed facial growth and de- facial muscle paralysis and to identify the effects mas- velopment [1]. Surgery is done early in these patients seter muscle paralysis has on mandible and zygoma growth. in an attempt to restore normal function and facial Methods. Twenty New Zealand white rabbits were appearance and to allow for normal psychological de- divided into control, paralysis, and sham groups. Mas- velopment [2, 3]. However, surgical correction can seter muscle paralysis was achieved with botulinum cause additional growth disturbances, does not pro- toxin A (BTX). Computed tomographic and single- mote growth, and can worsen the overall outcome for photon emission computed tomography (SPECT) these patients [4–6]. A better understanding of the scans and cephalometric measurements were per- mechanisms involved in controlling and regulating fa- formed. Masseter weights and mandible and zygoma cial growth may help to develop novel approaches, im- volumes, shapes, and metabolism were measured. proving current surgical interventions. Results. Eighteen animals completed the study. Sig- It is known that craniofacial bone growth is influ- nificant decreases in zygoma and mandible volumes enced by both genetic and epigenetic factors [7–14]. with minimal changes in shape were seen on the par- Moss and others have noted that muscle function is one alyzed sides. SPECT showed a decrease in bone pro- of the most important epigenetic factors involved in duction in both zygomas and mandibles on the para- guiding facial bone growth [7, 13, 15]. The notion that lyzed sides. muscle action influences facial growth may have arisen Conclusions. An animal model has been created in from clinical observations of facial characteristics of which the effects of single muscle paralysis on bone people with abnormal occlusion. Patients who are pri- growth can be studied. Masseter muscle function may marily mouth breathers develop long faces; tongue- be responsible in maintaining mandible and zygoma thrust habits result in anterior open bites, and patients volume by controlling bone production. Masseter that have an increased bite force and hypertrophy of function alone has less influence on mandible and zy- their masseter muscles develop a more acute gonial goma shape. © 2007 Elsevier Inc. All rights reserved. angle (square jaw), relatively shorter face, and a larger 1 Presented in June 2004 at the 49th annual Plastic Surgery mandible [16–18]. Research Council meeting in Ann Arbor, MI, and in September 2003 Animal models were created in an attempt to delin- at the 10th meeting of the International Society of Craniofacial eate the influence muscle has on facial bone growth. Surgery in Monterey, CA. Animal research can be divided into scar- and non- 2 To whom correspondence and reprint requests should be ad- dressed at London Health Sciences Center–Westminster Campus, scar-forming studies. Scar-forming studies generally 800 Commissioners Road, London, ON, N6A 4G5, Canada. E-mail: involve either removal of bone or bone sutures, and/or [email protected]. excision, transection, or change in the position of var-

METHODS

Experimental Design

Prior to the start of the experiment, full protocol approval was obtained through the Council of Animal Care at the University of Western Ontario (London, Ontario, Canada). Twenty New Zealand white rabbits (Oryctolagus cuniculus) were used for the study. All animals were female to limit sexual dimorphism. All animals were enrolled after weaning, at 6 weeks of age. Each animal was housed in a separate cage in the same room (Lawson Research Institute, Lon- don, Ontario, Canada) under climate-controlled conditions with 12-h light and dark lighting schedules. All were fed a standard diet of hard pellets (Lab Diet No. 5326; Purina Mills Inc., Richmond, IN) and water ad libitum. The animals were weighed at time 0, 4, 8, and 12 weeks. Animals were divided randomly as follows. Control group (n ϭ 5). All rabbits had facial imaging at time 0, 4, 8, and 12 weeks of the study to document bone growth and metabolism. Imaging instrumentation included computed tomography (CT) FIG. 1. Rabbit skull showing the approximate positions of the and single-photon emission computed tomography (SPECT). superficial and deep masseter muscle bellies as they attach to both the zygoma and the mandible. The dots represent the injection sites Botulinum toxin A (BTX) group (n ϭ 10). All animals had both for both botulinum toxin A in the BTX group and saline in the Sham mandibular regions shaved for clinical land marking, allowing direct group. visualization of the masseter muscle injection sites. All rabbits had an intramuscular injection of 25 mouse units (mu), 2.5 mL of volume, ious facial muscles. Removal of a muscle or bone de- of BTX (Allergan Inc., Irvine, CA) into one randomly chosen superficial masseter muscle belly, at the time of the first scan date (time ϭ creases the blood supply, changes the loading of the 0) (Fig. 1). All animals were injected on the same day. Injection was entire skeleton, introduces another biomechanical performed into three separate points (2.5 mL divided into 0.8 mL, 0.8 force (scar), and often causes the animals to eat a soft mL, and 0.9 mL) within the substance of the masseter. Five days diet, which further alters the biomechanical environ- after injection, paralysis was assessed by clinical palpation of both ment [19]. Given these variables, the results of such masseters at rest and during mastication. The examiner was blinded to the side of injection. Imaging was performed as for the Control studies become difficult to interpret. group. Research to date has been performed on models in Sham group (n ϭ 5). All animals had both mandibular regions which more than one muscle’s function is altered. It is shaved. All rabbits had an intramuscular injection of 2.5 ml of sterile difficult therefore to determine a single muscle’s influ- saline (0.9% sodium chloride) into one randomly chosen superficial ence on adjacent bone growth. A non-scarring type of masseter muscle belly at the time of the first scan date (time ϭ 0). animal model in which only one muscle’s action is Injection was performed into three separate points (2.5 mL volume removed has not yet been created. The purpose of this divided into 0.8 mL, 0.8 mL, and 0.9 mL) within the substance of the masseter (Fig. 1). CT scans were performed at the same time points pilot study, therefore, is to develop a model of individ- as for the other groups. SPECT scans were not performed in this ual muscle paralysis while minimizing the variables group due to limited SPECT scanner availability. known to independently affect facial growth such as scar. This model will then be used to gather pilot data to answer the following questions: What are the effects Anesthesia of paralysis of the masseter muscle on the growth and shape of the bones onto which it attaches (mandible All animals received anesthesia during craniofacial imaging periods and during intramuscular injections. A single dose of a cocktail and zygoma)? What are the effects of paralysis of the of 35 mg/kg of ketamine, 5 mg/kg of xylazine, and 0.75 mg/kg of masseter muscle on the metabolism of the bones onto acepromazine was injected intramuscularly to achieve initial anes- which it attaches? thesia on the imaging days and for accurate injections of BTX and ™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 FIG. 2. (A) Inferior view of rabbit upper skull showing the anatomical landmarks and corresponding points used for cephalometric measurements. Point A, posterior alveolar margin of the posterior maxillary incisor; point B, anterior alveolar margin of the first maxillary molar; point C, most anterior projection of the spina masseterica; point N, most anterior projection of the palatine process of the maxilla; point Z, most posterior projection of zygomatic arch on process of squamosal temporal bone. (B) Lateral view of rabbit upper skull showing the anatomical landmarks and corresponding points used for cephalometric measurements. Point C, most anterior projection of the spina masseterica; point X, most anterior projection of zygomatic process of squamosal temporal bone; point Y, lateral aspect of supraorbital process of frontal bone; point Z, most posterior projection of zygomatic arch on process of squamosal temporal bone. (C) Lateral view of rabbit mandible showing the points used for cephalometric measurements. Point R, most anterior and inferior projection of angulus mandibulae; point S, posterior alveolar margin of posterior mandibular molar; point T, anterior alveolar margin of mandibular first molar; point U, most superior and posterior projection of angulus mandibulae; point V, most anterior projection of processus condyloideus; point W, posterior alveolar margin of mandibular incisor. (D) Inferior view of rabbit mandible showing point used for cephalometric measurements. Point P, volar point of fusion of the bodies of each hemi-mandible. (E) Close-up of rabbit condylar head showing points used for cephalometric measurements. Point L, lateral margin of condylar head at its widest margin; point M, medial margin of condylar head at its widest margin. MATIC ET AL.: MASSETER MUSCLE PARALYSIS ON FACIAL BONE GROWTH 245 246 JOURNAL OF SURGICAL RESEARCH: VOL. 139, NO. 2, MAY 15, 2007 saline into masseter muscles. At the end of the study, a single Cephalometric and Masseter Muscle Measurements intravenous dose of pentobarbital (110 mg/kg) was used to achieve euthanasia. At the end of the study, the superficial and deep masseter muscles on both sides of each animal were harvested and weighed as one unit due to difficulty in accurately separating them during the dissection. Image Acquisition After muscle harvest, the skulls were cleaned by beetle colony and direct cephalometric measurements using digital Vernier calipers (accuracy, Ϯ0.01 mm) were performed. Basic landmarks previously CT mandible and zygoma volume measurements. A coronal heli- described for measurements of cranial, maxillary, and mandibular cal scan using a General Electric (GE) Discovery LS CT scanner (GE structures were used [23, 24]. In addition, extra points were identi- Technologies, Waukesha, WI) was performed on all animals to mea- fied in an attempt to measure small changes in shape and size that sure bone volumes. The scan slice thickness was 1.25 mm with a may be observed (Fig. 2A–E). The following distances were measured 0.5-mm overlap interval at a speed of 3.75 mm/rotation. The X-ray for the upper skull: A-B, A-C, A-Z, B-C, B-N, B-Z, C-N, C-Z, C-Y, Z-Y, tube parameters were 80 kV and 50 mA and were increased accord- Z-N, X-C, X-Z, X-N, and X-Y. The following distances were measured ingly as the animals grew larger. for the mandible: V-U, V-R, V-S, V-T, V-W, U-S, U-T, U-W, U-R, R-S, Once the CT data were acquired, segmentation of the mandible R-T, R-W, S-T, S-W, T-W, P-V, P-U, P-R, and L-M. Qualitative and zygomas was performed using the software program, Analyze observations, noting any consistent patterns of shape change within 4.0 (Mayo Clinic, Rochester, MN). The mandible was divided into two the mandible and maxilla, not measurable with the above linear hemi-mandibles through the midline between the central incisors. distances, were also recorded. The mandibular tooth crowns were not included in the segmentation. The extent of the zygomatic bone is marked by bone sutures, which were not visible on CT scans. Landmarks were therefore created to Statistical Analysis represent the superior, inferior, anterior, posterior, and medial lim- The injected sides of BTX and Sham animals were referred to as its of the zygomatic bone. The same individual performed all of the treated (T) and the contralateral sides were referred to as untreated segmentations. The Analyze 4.0 program counts the number of pixels 3 (U). In Controls, the right side was arbitrarily chosen as the treated in each outline and multiplies by the cubic voxel size (mm )to side to simplify the analysis and reduce confusion. CT volume, calculate the volume of each coronal slice outline. All slice volumes SPECT, masseter muscle weights, animal weights, and cephalomet- are then summed to obtain the total volume of the object of interest. ric data were all analyzed in a similar fashion. The distribution of Volumes for each hemi-mandible and zygoma were calculated in each outcome was examined and, where necessary, the data were each animal. mathematically transformed to approximate a normal distribution. SPECT imaging measurements. SPECT is a modality that gives Descriptive statistics are presented as means Ϯ SD. Graphical functional information regarding a biological process such as bone representation of the data includes error bars denoting standard metabolism in a dynamic in vivo environment. The pharmaceutical error of the mean (SEM). Within each group, treated (T) sides were used for this study was methylene diphosphonate (MDP). Diphos- compared to untreated (U) sides at each time point using paired phonates accumulate during mineralization of bone and do not con- t-tests and, as supportive tests, Wilcoxon signed rank tests. To com- centrate in osteoid (mature bone) or in osteoclast-driven bone resorp- pare groups in terms of the treatment effect within each animal, tion. MDP accumulation, therefore, represents active osteoblast- among group differences were based on the difference between driven bone production. MDP, labeled with the radioisotope treated and untreated sides (T Ϫ U). At each time point, one-way technetium-99m (99mTc), emits gamma-rays of 140 keV that are de- analysis of variance (ANOVA) was used to compare groups where tected by the SPECT scanner’s scintillation camera. These detected three groups were involved and unpaired t-tests where two groups emissions represent MDP accumulation in areas of bone production were involved. Kruskal–Wallis tests and Wilcoxon two-sample tests [20]. A dose of 15 MBq/kg of 99mTc-MDP was injected intravenously provided supportive statistical comparisons for three-group and two- 4 h prior to scanning for equilibration and uptake within the skele- group comparisons, respectively. Since there was some concern that ton. The animals were then scanned using a three-dimensional small the assumptions of the t-tests and ANOVAs may not have been field-of-view SPECT camera (dual-head GE Millennium MG Cam- satisfied, due to small sample sizes, nonparametric tests were used era, GE Technologies). Resolution is 2.26 mm/pixel and total acqui- to support the conclusions made from the parametric tests. sition time was 32 min per scan. Within each group, the effect of time was examined using repeated Once the emission data were collected, the reconstructive algo- measures ANOVA with time as the “within subjects” factor and rithm, “Maximum A Posteriori Ordered Subset Expectation Maximi- post-hoc comparisons were performed using Tukey’s honestly signif- zation” (MAP OSEM) was used to reconstruct the data into a three- icant difference (HSD) t-test and Dunnett’s t-test. Tukey’s HSD was dimensional map [21, 22]. Separate analysis of zygoma and mandible performed to identify differences between pairs of groups once uptake could not be performed accurately due to resolution limita- ANOVA testing showed a difference to exist among groups. Dun- tions (2.26 mm/pixel); therefore, areas of activity corresponding to nett’s test was used to compare subsets of pairs within a comparison. the mandibular angle were measured. Dunnett’s test was used for cephalometric data. Both Tukey’s and

TABLE 1 Mean Changes in Animal Weights (kg) Divided by Group over Time

Control BTX Sham

Variable Mean S.D. Mean S.D. Mean S.D. ANOVA (P value) Kruskal–Wallis (P value) dwt_0-4w 1.0 0.1 1.0 0.2 0.9 0.1 0.329 0.147 dwt_0-8w 1.7 0.2 1.7 0.3 1.9 0.2 0.494 0.465 dwt_0-12w 1.9 0.2 2.0 0.3 2.2 0.2 0.165 0.134

wt ϭ weight; dwt ϭ difference in weight; w ϭ week; S.D. ϭ standard deviation. MATIC ET AL.: MASSETER MUSCLE PARALYSIS ON FACIAL BONE GROWTH 247

TABLE 2 Mean Masseter Weights (g) of Treated and Untreated Sides Divided by Group

masseter_T masseter_U Signed Paired t-test rank Group (n) Mean S.D. Mean S.D. (P value) (P value)

Control (4) 5.92 0.37 5.89 0.40 0.570 0.875 BTX (9) 5.28 0.83 6.10 0.48 <.001 0.008 Sham (5) 4.85 0.47 5.02 0.54 0.237 0.313

Note. P values are reported in bold if statistically signiﬁcant (P Ͻ 0.05). T ϭ treated; U ϭ untreated; S.D. ϭ standard deviation.

Dunnett’s were used to minimize Type 1 error during comparisons, reducing false-positive rates. Two-sided P values Ͻ 0.05 were con- sidered statistically signiﬁcant throughout the analysis.

RESULTS

Eighteen animals completed the study protocol. Two animals died during the study due to anesthetic complications. The first animal was in the Control group (C2) and the second was in the BTX group (B6L). Data FIG. 4. (A) Mean hemi-mandible volumes (mm3) versus time gathered prior to death for both animals were used in (weeks) in the BTX group. Paired statistical comparisons of treated the final analyses. and untreated sides in the BTX group were done at each time period. All animals in the BTX group tolerated BTX injec- Error bars represent SEM. Statistical comparisons were performed after logarithmic transformation of the data. *P Ͻ 0.001. (B) Mean tions without complications. The BTX-injected side differences (T Ϫ U) in hemi-mandible volumes (mm3) among groups was identified 100% of the time by the blinded exam- versus time. Mean differences (T Ϫ U) for all groups were compared iner (D.B.M.) by palpation at rest and during function at each of the four time periods. Error bars represent SEM. Statis- 5 days after injection. Clinically, BTX rabbits did not tical comparisons were performed after logarithmic transformation favor chewing on any particular side and tolerated of the data. *P Ͻ 0.001. eating the standard hard pellet diet. gains across time (week 0–4, week 0–8, and week Animal and Masseter Muscle Weights 0–12) were analyzed among groups, no difference in There were no clinically evident differences in size weights was identified (Table 1). between animals at the end of the study. When weight The BTX-treated masseters were significantly smaller than untreated muscles in the BTX group. No difference existed between treated and untreated sides in the Sham and Control groups (Table 2). In addition, the difference between weights of the two masseters was significant between the BTX and Control groups and BTX and Sham groups but not for the Sham and Control groups (Fig. 3).

CT Volume Measurements Hemi-mandible volumes. Prior to analysis, the data were transformed to a logarithmic distribution to improve normality of data distribution. No differences between sides in either the Control or the Sham group were identified. The hemi-mandibles on the treated FIG. 3. Mean differences (T Ϫ U) in masseter weights (g) among side in the BTX group were smaller in volume than groups. Error bars represent SEM. *P ϭ 0.002. Pairwise comparison untreated sides at weeks 4, 8, and 12 (Fig. 4A). The (Tukey’s HSD) tests indicate that there are statistically significant differences between BTX and Sham groups and BTX and Control mean differences in volumes between treated and un- groups (P Ͻ 0.05) but not between Sham and Control groups (P Ͼ treated sides among groups were significantly different 0.05). at week 4, 8, and 12 (Fig. 4B). Pairwise comparisons 248 JOURNAL OF SURGICAL RESEARCH: VOL. 139, NO. 2, MAY 15, 2007

TABLE 3 Mean Differences (T ؊ U) in Hemi-mandible Volume (mm3) Increases Among Groups Over Time

Control BTX Sham

Variable Mean S.D. Mean S.D. Mean S.D. ANOVA (P value) Kruskal–Wallis (P value) dd_hm_0-4w 74.31 — Ϫ166.30 55.04 3.50 22.08 <0.001* 0.009 dd_hm_0-8w 2.63 — Ϫ178.27 76.43 Ϫ7.63 24.88 0.416 0.228 dd_hm_0-12w 28.91 23.43 Ϫ165.79 78.76 Ϫ5.71 25.72 0.744 0.680

Note. P values are reported in bold if statistically significant (P Ͻ 0.05). * Pairwise comparison (Tukey’s HSD) - BTX versus Sham/Control are significantly different (P Ͻ 0.05). No significant difference exists between Sham and Control groups at any time interval (P Ͼ 0.05, Tukey’s HSD). T ϭ treated; U ϭ untreated; dd ϭ time period; hm ϭ hemi-mandible; w ϭ weeks; S.D. ϭ standard deviation. between groups showed that the BTX group was dif- showed an interaction (P Ͻ 0.001) among groups (P ϭ ferent (P Ͻ 0.05) from both the Sham and the Control 0.015) and assessment (P Ͻ 0.001), confirming that the groups. There were no differences identified between changes in volumes were not consistent with time. the Sham and Control groups. Table 3 shows the dif- Zygoma volumes. Prior to analysis, the data were ference in mean volume increases (T Ϫ U) in all groups transformed to a logarithmic distribution to improve across the three time intervals. The BTX group was normality of data distribution. No differences between different from the Sham/Control groups combined at sides in either the Control or the Sham group were the first time interval. Repeated measures ANOVA identified. The zygomas on the treated side in the BTX group were smaller in volume than untreated sides at week 4, 8, and 12 (Fig. 5A). The mean differences in volumes between treated and untreated sides among groups were different at weeks 4, 8, and 12 (Fig. 5B). Pairwise comparisons between groups showed that the BTX group was different (P Ͻ 0.05) from both the Sham and the Control groups. There were no differences identified between the Sham and Control groups. Table 4 shows the difference in mean volume increases (T Ϫ U) in all groups across the three time intervals (weeks 0-4, 0-8, and 0-12). The BTX group was different from the Sham/Control groups combined at the second time interval. However, repeated measures ANOVA showed no interaction (P ϭ 0.076) among groups (P ϭ 0.033) and assessment (P ϭ 0.087), suggesting that changes in volume were quantitatively similar across time.

Cephalometric Measurements Mandible measurements. Paired t-test analysis showed a signiﬁcant difference in the BTX group for measurements RS (the distance from the most anterior and inferior projection of angulus mandibulae to the posterior alveolar margin of the posterior mandibular molar)

3 and LM (the distance from the lateral margin to the FIG. 5. (A) Mean zygoma volumes (mm ) versus time (weeks) in medial margin of the condylar head at its widest point) the BTX group. Paired statistical comparisons of treated and untreated sides in the BTX group were done at each time period. Error with the treated sides being smaller. Mean differences bars represent SEM. Statistical comparisons were performed after (T Ϫ U), divided into groups, are shown in Table 5. ANOVA logarithmic transformation of the data. *P Ͻ 0.001. (B) Mean differ- analysis identiﬁed a signiﬁcant difference among groups 3 ences (T Ϫ U) in zygoma volumes (mm ) among groups versus time. for measurements RS, RW (the distance from the most Mean differences (T Ϫ U) for all groups were compared at each of the four time periods. Error bars represent SEM. Statistical comparisons anterior and inferior projection of angulus mandibulae to were performed after logarithmic transformation of the data.*P Ͻ the posterior alveolar margin of the mandibular incisor), 0.001. and LM. Pairwise comparison with Dunnett’s t-test iden- MATIC ET AL.: MASSETER MUSCLE PARALYSIS ON FACIAL BONE GROWTH 249

TABLE 4 Mean Differences (T ؊ U) in Zygoma Volume (mm3) Increases Among Groups Over Time

Control BTX Sham

Variable Mean S.D. Mean S.D. Mean S.D. ANOVA (P value) Kruskal–Wallis (P value) dd_zygoma_0-4w 35.69 . Ϫ66.66 48.69 6.09 37.02 0.065 0.079 dd_zygoma_0-8w 22.00 . Ϫ126.25 68.33 Ϫ3.22 26.92 0.011* 0.022 dd_zygoma_0-12w 55.12 6.95 Ϫ110.22 66.15 Ϫ2.59 41.98 0.893 0.684

Note. P values are reported in bold if statistically significant (P Ͻ 0.05). * Pairwise comparison (Tukey’s HSD) - BTX versus Sham/Control are significantly different (P Ͻ 0.05). No significant. T ϭ treated; U ϭ untreated; dd ϭ time period; w ϭ weeks; S.D. ϭ standard deviation. tified a difference between the BTX and Control groups measurement XC (the distance from the most anterior (P Ͻ 0.05) for RS and LM. Both measurements were projection of the zygomatic process of the squamosal smaller on the BTX-treated side. Dunnett’s t-test showed temporal bone to the most anterior projection of the that BTX and Sham groups were significantly different spina masseterica), with the treated side being from the Control group (P Ͻ 0.05) for RW. However, a smaller. Mean differences (T Ϫ U), divided into groups, test for equality of variances (Folded F) identified that are shown in Table 6. ANOVA analysis revealed no the variances between groups were different for measure- differences between groups for any of the measurement RS (P Ͻ 0.05). No difference between variances was ments. For measurement XC, differences (T Ϫ U) ap- identified for measurements RW and LM (P Ͼ 0.05). proached significance. Qualitative analysis of upper There were qualitative changes in mandible shape skulls did not identify any differences between sides or noted on the treated sides in the BTX group. These dif- groups. ferences included an occlusal tilt downward toward the BTX-treated side. At the angle of the mandible, the fossa SPECT Imaging in which the masseter muscle is located was absent. Upper skull measurements. Paired t-test analysis SPECT data were square-root transformed to im- showed a significant difference, in the BTX group, for prove normality of data distribution. No differences

TABLE 5 Mean Differences (T ؊ U) in Cephalometric Measurements (mm) of the Mandible in All Groups at the End of the Study

Control (n ϭ 4) BTX (n ϭ 9) Sham (n ϭ 5)

Variable Mean Diff. S.D. Mean Diff. S.D. Mean Diff. S.D. ANOVA (P value) Kruskal–Wallis (P value)

VU_d Ϫ0.22 0.20 0.20 0.99 Ϫ0.03 0.66 0.684 0.628 VR_d Ϫ0.03 0.42 Ϫ0.19 0.82 Ϫ0.03 0.59 0.884 0.868 VS_d Ϫ0.29 0.34 0.45 0.73 0.18 0.28 0.134 0.075 VT_d 0.29 0.36 0.25 0.44 0.09 0.44 0.728 0.679 VW_d 0.05 0.48 0.04 0.57 0.74 0.72 0.115 0.188 US_d 0.42 0.48 Ϫ0.31 0.83 Ϫ0.08 0.97 0.355 0.320 UT_d 0.38 0.53 0.08 0.75 0.05 0.44 0.704 0.569 UW_d 0.09 0.59 Ϫ0.18 0.60 Ϫ0.45 0.79 0.480 0.279 UR_d 0.12 0.55 0.28 0.59 Ϫ0.32 0.44 0.175 0.126 RS_d Ϫ0.19 0.16 Ϫ1.16 0.89 Ϫ0.19 0.38 0.029 0.034 RT_d Ϫ0.24 0.62 Ϫ0.71 1.49 Ϫ0.41 0.72 0.780 0.707 RW_d 0.81 0.71 Ϫ0.73 0.97 Ϫ0.87 0.87 0.024 0.050 ST_d Ϫ0.04 0.33 0.17 0.30 0.22 0.35 0.460 0.532 SW_d 0.07 0.45 0.07 0.22 0.05 0.32 0.994 0.994 TW_d 0.21 0.41 Ϫ0.26 0.53 0.11 0.33 0.196 0.285 PV_d Ϫ0.15 0.51 Ϫ0.53 0.92 0.52 0.74 0.099 0.087 PU_d 0.49 0.24 Ϫ0.54 1.08 Ϫ0.07 0.16 0.133 0.011 PR_d Ϫ0.30 0.59 Ϫ0.71 1.16 0.04 0.91 0.422 0.401 LM_d Ϫ0.08 0.21 Ϫ0.57 0.28 0.17 0.22 <.001 0.001

Note. P values are reported in bold if statistically signiﬁcant (P Ͻ 0.05). d ϭ difference T Ϫ U; S.D. ϭ standard deviation. 250 JOURNAL OF SURGICAL RESEARCH: VOL. 139, NO. 2, MAY 15, 2007

TABLE 6 Mean Differences (T ؊ U) in Cephalometric Measurements (mm) of Upper Skull in All Groups at the End of the Study

Control (n ϭ 4) BTX (n ϭ 9) Sham (n ϭ 5)

Variable Mean Diff. S.D. Mean Diff. S.D. Mean Diff. S.D. ANOVA (P value) Kruskal–Wallis (P value)

AB_d Ϫ0.27 0.59 0.11 0.44 0 0.35 0.404 0.694 AC_d 0.08 0.37 Ϫ0.03 0.56 Ϫ0.09 0.48 0.885 0.938 AZ_d Ϫ0.15 0.29 Ϫ0.16 0.89 0.23 0.50 0.606 0.574 BC_d 0.13 0.55 0.27 0.38 0.05 0.50 0.687 0.437 BN_d 0.14 0.66 Ϫ0.23 0.33 0.13 0.40 0.234 0.175 BZ_d Ϫ0.06 0.26 Ϫ0.43 0.98 Ϫ0.05 0.38 0.580 0.302 CN_d 0.19 0.44 0.31 0.69 0.10 0.27 0.782 0.863 CZ_d Ϫ0.13 0.43 Ϫ0.51 1.04 0.15 0.36 0.351 0.118 CY_d Ϫ0.48 0.37 0.33 0.77 0.32 0.68 0.150 0.101 ZY_d Ϫ0.01 0.40 Ϫ0.28 0.81 Ϫ0.11 1.14 0.852 0.917 ZN_d Ϫ0.13 0.55 Ϫ0.25 0.91 0.20 0.39 0.553 0.466 XC_d Ϫ0.09 0.78 Ϫ1.32 0.96 Ϫ0.39 0.95 0.075 0.056 XZ_d Ϫ0.33 0.52 0.48 1.03 0.44 0.68 0.299 0.502 XN_d Ϫ0.09 0.71 Ϫ0.91 1.54 0.46 0.69 0.151 0.188 XY_d Ϫ0.21 0.67 0.27 0.69 Ϫ0.36 0.64 0.230 0.253

d ϭ difference T Ϫ U; S.D. ϭ standard deviation. between sides in the Control group were identified. A Bone Volumes and SPECT Imaging significant difference was found in the BTX group at 4 Hemi-mandible and zygoma volumes decreased sig- weeks (P ϭ 0.002). The treated side was smaller (less nificantly on the side of masseter paralysis compared bone production) compared to the untreated side (Fig. to untreated sides in the BTX group. Hemi-mandible 6A). There was also a trend toward less bone produc- volume differences (T Ϫ U) were significantly different tion on the treated sides of BTX animals. Mean differ- between the BTX group and both the Sham and the ences (T Ϫ U) between groups, at each time point, are Control groups between 0 and 4 weeks. This suggests presented in Fig. 6B. At 4 weeks, a significant differ- that paralysis from BTX resulted in less mandible ence between groups (P ϭ 0.001) was identified. growth during the first 4 weeks of the study on the side of paralysis. Reversal of paralysis must have occurred DISCUSSION between 4 and 8 weeks as the volume differences between sides became similar. These changes in volume are supported by the SPECT scan findings, which link BTX was chosen for this study because of its po- a decrease in bone volume with a decrease in bone tency, reliability in achieving paralysis, mainte- production. Further, changes in hemi-mandible vol- nance of nerve continuity, and maintenance of mus- umes across time are also linked with changes in bone cle architecture and function after reversal of production at week 4. However, zygoma volumes in- paralysis [25, 26]. Onset of action in rodents is 24 h creased at similar rates. Masseter function therefore after injection and paralysis lasts 4–6 weeks [27, may not control zygoma growth in the same way as it 28]. Given the larger size and slower metabolism of does for the mandible. the rabbit, paralysis was expected to last 6–8 weeks. Animals were enrolled after weaning, at 6 weeks of Cephalometric Measurements age, to take advantage of the pubertal growth spurt Mandible Measurements that occurs at 10–14 weeks [29, 30]. The size of the masseter muscle in 6-week-old rab- Significant differences among groups were seen for bits approximates the size of adult human corrugator RS, RW, and LM. However, variances were different muscles. The dose for paralysis of these muscles aver- for the RS mean measurements, suggesting that the ages between 20 and 30 mu [31]. Therefore, 25 mu of difference seen may not be true. In addition, other BTX was used for this study. Clinical assessment has distances that use R or W as points were not different been shown to be a reliable method for assessing BTX- between groups. The measured difference therefore induced muscle paralysis [32, 33]. BTX injection may be due to chance. The LM measurement, however, achieved clinical paralysis with masseter atrophy in all may be significant given that the difference was visible BTX group animals. and biologically congruent (Fig. 7). If the bite force that MATIC ET AL.: MASSETER MUSCLE PARALYSIS ON FACIAL BONE GROWTH 251 is distributed through the mandible were less on one side, the condyle on that side would be expected to remodel and decrease in size. Loss of the masseteric fossa on the angle and ramus of the mandible was also seen. Loss of muscle bulk and masseter movement due to paralysis, in the area of the ramus, would have decreased the force on the under- lying bone matrix. This reduction in force would then lead to net bone deposition on the masseter side of the ramus and net resorption on the lingual side of the FIG. 7. Close-up of both condyles of a BTX animal’s mandible. ramus, resulting in loss of the fossa. The condyle on the figure’s right is the treated side. Note that the treated condyle is narrower compared to the untreated side. Upper Skull Measurements Significant differences between sides in the BTX ration. In addition, there was a trend toward a differ- group were seen for measurement XC (length of the ence in XC measurements between groups but this did zygoma in the area of masseter muscle attachment). not reach significance. Increasing animal numbers in Since zygoma lengthening occurs at the ZSS, this find- the future may resolve this issue. ing suggests that masseter muscle function may affect bone growth at this suture by controlling suture sepa- Mandible and Zygoma Bone Growth Summary Clinical masseter muscle paralysis was achieved, di- minishing the forces transmitted to the mandible and zygoma during function. It is likely that the diminished forces caused the remodeling process of bone growth to reduce bone thickness circumferentially to maintain relative bone shape. This would then explain why the volumes of the hemi-mandible and zygoma on the side of paralysis were decreased while shape and relative size were preserved. Masseter muscle function may therefore be involved in maintaining mandibular and zygomatic bone volume through changes in bone metabolism. Masseter muscle function alone is probably not involved in maintaining mandibular and zygomatic bone shape. The findings of this study confirm Her- ring’s observation, that “decreased levels of muscle activity are associated with skeletal elements that are of reasonably normal length but reduced in diameter” [19]. It is conceivable, however, that BTX itself may have been the cause of changes in bone volume, independent of muscle paralysis. Recently, it has been shown that the maturational growth of muscle is influenced by BTX; however, this growth change is partially reversed by exercise [25]. The animals in this study were al- lowed to chew normally, exercising their masseter muscles. In addition, neurotrophic effects on muscle growth and development are preserved after BTX- induced paralysis [26]. Therefore it is less likely that BTX independently influenced bone growth. FIG. 6. (A) Mean masseter SPECT results (counts) versus time This pilot study has taken a step toward understand- (weeks) in the BTX group. Paired statistical comparisons of treated and untreated sides in the BTX group were done at each time period. ing the epigenetic control of facial growth by establish- Error bars represent SEM. Statistical comparisons were performed ing the first model in which a single facial muscle’s after square-root transformation of the data. *P ϭ 0.002. (B) Mean action can be removed to analyze bone growth changes. differences (T Ϫ U) in masseter SPECT results (counts) between It has also identified SPECT as an investigative tool groups versus time (weeks). Mean differences (T Ϫ U) for all groups were compared at each of the four time periods. Error bars represent that can be used to measure bone metabolism changes SEM. Statistical comparisons were performed after square-root secondary to muscle paralysis. These encouraging pre- transformation of the data. *P ϭ 0.001. liminary findings should be confirmed with larger an- 252 JOURNAL OF SURGICAL RESEARCH: VOL. 139, NO. 2, MAY 15, 2007 imal numbers. Electromyography should also be used 14. Enlow DH. Growth and the problem of the local control mech- to confirm both paralysis and its reversal. These anism. Am J Anat 1973;136:403. changes may make it possible to quantify the masseter 15. Petrovic AG, Stutzmann JJ, Oudet CL. Defects in mandibular growth resulting from condylectomy and resection pf the ptery- muscle’s contribution to both mandibular and zygo- goid and masseter muscles. In: McNamara JW, Jr, ed. The matic bone growth. effect of surgical intervention on craniofacial growth, monograph no. 12, Craniofacial growth series. Ann Arbor, MI: Uni- ACKNOWLEDGMENTS versity of Michigan, 1982. 16. Storey AT, Kenny DJ. Growth, development, and aging of oro- facial tissues: neural aspects. Adv Dent Res 1989;3:14. We thank Larry Stitt, Biostatistician, at the University of Western 17. Kiliardis S. The relationship between masticatory function and Ontario for advice and preparation of all of the statistical analyses. craniofacial morphology. Eur J Orthod 1985;7:273. In addition, we thank our granting agencies Plastic Surgery Educa- tion Foundation–Smile Train Award, AO Research Fund of the AO 18. Moss ML, Moss-Salentijn L. The muscle-bone interface: an Foundation for Project Number 03-M73, and the Lawson Health analysis of a morphologic boundary. In: McNamara JW, Jr, ed. Research Institute for generous support of this project. Muscle adaptation in the craniofacial region, monograph no. 8, Funding support was provided by Plastic Surgery Education Craniofacial growth series. Ann Arbor, MI: University of Mich- Foundation–Smile Train Award, AO Research Fund of the AO igan, 1978. foundation–Project no. 03-M73, and Lawson Health Research Insti- 19. Herring SW. Epigenetic and functional influences on skull tute Young Investigators Award. growth. In: Hanken J, Hall BK, ed. The skull. Chicago, IL: University of Chicago Press, 1993. 20. Wagner HN, Szabo Z, Buchanan JW. Principles of nuclear REFERENCES medicine, 2nd ed. Philadelphia, PA: WB Saunders, 1995. 1. Pruzansky S. Craniofacial surgery: the experiment on nature’s 21. Hudson HM, Larkin RS. Accelerated image reconstruction us- experiment II. Historical perspectives. In: McNamara JW, Jr, ing ordered subsets of projection data. IEEE Trans Med Imag- ed. The effect of surgical intervention on craniofacial growth, ing 1994;13:601. monograph no. 12, Craniofacial growth series. Ann Arbor, MI: 22. Bruyant PP. Analytic and iterative reconstruction algorithms .J Nucl Med 2002;43:1343 .ءUniversity of Michigan, 1982. in SPECT 2. Kearns GJ, Padwa BL, Mulliken JB, et al. Progression of facial 23. Hoffman WY, McCarthy JG. The effects of facial nerve ablation asymmetry in hemifacial microsomia. Plast Reconstr Surg on craniofacial skeletal development in neonatal rabbits. Plast 2000;105:492. Reconstr Surg 1994;93:1236. 3. Cousley RR, Calvert ML. Current concepts in the understand- 24. Bardach J, Mooney MMS, Giedrojc-Juraha ZL. A comparative ing and management of hemifacial microsomia. Br J Plast Surg study of facial growth following cleft lip repair with or without 1997;50:536. soft-tissue undermining: an experimental study in rabbits. 4. Bardach J, Eisbach KJ. The influence of primary unilateral cleft Plast Reconstr Surg 1982;69:745. lip repair on facial growth, Part I: lip pressure. Cleft Palate J 25. Chen C, Stott S, Smith H. Effects of botulinum toxin A injection 1977;1:88. and exercise on the growth of juvenile rat gastrocnemius mus- 5. Eisbach K, Bardach J, Klausner E. The influence of primary cle. J Appl Physiol 2002;93:1437. unilateral cleft lip repair on facial growth, Part II: direct ceph- 26. Soha JM, Callaway EM, Van Essen DC. Lack of fiber type alometry of the skull. Cleft Palate J 1978;2:109. selectivity during re-innervation of neonatal rabbit soleus. Dev 6. Kokich VG, Shapiro PA. The effects of lefort I osteotomies on Biol 1989;131:401. the craniofacial growth of juvenile macaca nemestrina. In: Mc- 27. Tarabal O, Caldero J, Ribera J, et al. Regulation of motoneu- Namara JW, Jr, ed. The effect of surgical intervention on ronal calcitonin gene-related peptide (CGRP) during axonal craniofacial growth, monograph no. 12, Craniofacial growth growth and neuromuscular synaptic plasticity induced by bot- series. Ann Arbor, MI: University of Michigan, 1982. ulinum toxin in rats. Eur J Neurosci 1996;8:829. 7. Moss ML. The functional matrix hypothesis revisited. 1. The 28. Filippi GM, Errico P, Santarelli R, et al. Botulinum A toxin role of mechanotransduction. Am J Orthod Dentofacial Orthop effects on rat jaw muscle spindle. Acta Otolaryngol 1993;113: 1997;112:8. 400. 8. Enlow DH. Wolff’s law and the factor of architectonic circum- 29. Glattly AD, McKeown M. The growth of the rabbit skull in stance. Am J Orthod 1968;54:803. 3-dimensions: a radiographic cephalic appraisal. Anat Anz 1982;151:105. 9. Wolff J. Gesetz der Transformation der Knochen. Berlin: Au- gust Hirschwald, 1892. 30. Masoud I, Shapiro F, Moses A. Longitudinal roentgencephalic study of the growth of the New Zealand white rabbit: cumula- 10. Moss ML. The functional matrix hypothesis revisited. 2. The tive and biweekly incremental growth rates for skull and man- role of an osseous connected cellular network. Am J Orthod dible. J Craniofac Genet Dev Biol 1986;6:259. Dentofacial Orthop 1997;112:221. 31. Guyuron B, Tucker T, Kriegler J. Botulinum toxin A and mi- 11. Moss ML. The functional matrix hypothesis revisited. 3. The graine surgery. Plast Reconstr Surg 2003;112:171. genomic thesis. Am J Orthod Dentofacial Orthop 1997;112:338. 32. Borodic GE, Joseph M, Fay L, et al. Botulinum A toxin for the 12. Moss ML. The functional matrix hypothesis revisited. 4. The treatment of spasmodic torticollis: dysphagia and regional toxin epigenetic antithesis and the resolving synthesis. Am J Orthod spread. Head Neck 1990;12:392. Dentofacial Orthop 1997;112:410. 33. Borodic GE, Ferrante R. Effects of repeated botulinum toxin 13. Moss ML, Salentijn L. The primary role of functional matrices injections on orbicularis oculi muscle. J Clin Neuroophthalmol in facial growth. Am J Orthod 1969;55:566. 1992;12:121.