Distinct Frequency Dependent Effects of Whole-Body Vibration on Non-Fractured Bone and Fracture Healing in Mice

Transcription

1 Distinct Frequency Dependent Effects of Whole-Body Vibration on Non-Fractured Bone and Fracture Healing in Mice Esther Wehrle, 1 Tim Wehner, 1 Aline Heilmann, 1 Ronny Bindl, 1 Lutz Claes, 1 Franz Jakob, 2 Michael Amling, 3 Anita Ignatius 1 1 Institute of Orthopaedic Research and Biomechanics, Centre of Musculoskeletal Research, University of Ulm, Helmholtzstr.14, Ulm, Germany, 2 Orthopedic Center for Musculoskeletal Research, Orthopedic Department, University of Würzburg, Würzburg, Germany, 3 Department of Osteology and Biomechanics, University Medical Center Hamburg-Eppendorf, Hamburg, Germany Received 11 September 2013; accepted 14 March 2014 Published online 14 April 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI /jor ABSTRACT: Low-magnitude high-frequency vibration (LMHFV) provokes anabolic effects in non-fractured bone; however, in fracture healing, inconsistent results were reported and optimum vibration conditions remain unidentified. Here, we investigated frequency dependent effects of LMHFV on fracture healing. Twelve-week-old, female C57BL/6 mice received a femur osteotomy stabilized using an external fixator. The mice received whole-body vibrations (20 min/day) with 0.3g peak-to-peak acceleration and a frequency of either 35 or 45 Hz. After 10 and 21 days, the osteotomized femurs and intact bones (contra-lateral femurs, lumbar spine) were evaluated using bending-testing, m-computed tomography, and histomorphometry. In non-fractured trabecular bone, vibration with 35 Hz significantly increased the relative amount of bone (þ28%) and the trabecular number (þ29%), whereas cortical bone was not influenced. LMHFV with 45 Hz failed to provoke anabolic effects in trabecular or cortical bone. Fracture healing was not significantly influenced by whole-body vibration with 35 Hz, whereas 45 Hz significantly reduced bone formation ( 64%) and flexural rigidity ( 34%) of the callus. Although the exact mechanisms remain open, our results suggest that small vibration setting changes could considerably influence LMHFV effects on bone formation in remodeling and repair, and even disrupt fracture healing, implicating caution when treating patients with impaired fracture healing. ß 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32: , Keywords: whole-body vibration; LMHFV; fracture healing; bone formation Delayed or incomplete fracture healing is a major issue in orthopedic surgery, with an incidence of about 10% of patients and a high socioeconomic burden, implicating the need for effective treatment strategies. 1 Because bone metabolism is partly regulated by mechanical factors and suitable mechanical stimuli can improve bone repair, 2,3 non-invasive physical methods, among them low-magnitude high-frequency vibration (LMHFV), 2,4 could provide a readily applicable and cost-effective approach to prevent or treat delayed fracture healing. LMHFV is characterized by a sinusoidal waveform with an acceleration of <1g (g ¼ gravitational acceleration, 1g ¼ 9.81 m/s 2 ) and a frequency range between 20 and 90 Hz, 5 inducing extremely small strains in bone tissue (approximately 10 me), much less than the peak strains generated during locomotion (2,000 3,000 me). 2 LMHFV provoked anabolic effects in both healthy and osteoporotic bone in preclinical 6 9 and clinical studies. 4 However, few authors have investigated the influence of LMHFV on fracture healing, reporting contradictory results. Goodship et al. 10 demonstrated that LMHFV, locally applied to the fractured sheep tibia via a dynamic fixator (30 Hz, 25 mm interfragmentary movement), increased bone mineral content and mechanical properties of the fracture callus. In contrast, vibrations with 20 Hz inducing 20 mm interfragmentary movement did not provoke any significant effect on fracture Conflict of interest: None. Grant sponsor: German Research Foundation; Grant numbers: FOR793, IG18/7-2, SCHI 504/5-2. Correspondence to: Anita Ignatius (T: þ ; F: þ ; anita.ignatius@uni-ulm.de) # 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. healing in sheep. 11 In both healthy 12 and ovariectomized rats, 13 LMHFV (35 Hz/0.3g) was shown to improve bone repair by inducing a large callus with more bone and improved mechanical quality. The same authors found that LMHFV stimulated callus angiogenesis 14 and, in the late healing phase, bone resorption. 15 In a model of metaphyseal bone healing, vibrations with both high frequency (90 Hz) and acceleration (4g) provoked some positive effects in ovariectomized rats, but resulted in delayed callus bridging in healthy animals, and was therefore considered unsuitable to support fracture healing. 16 In a recent study, the same group compared different frequencies (35 90 Hz) but found that apart from some minor positive effects of certain vibration regimes on bone formation, there was no improvement in the mechanical callus quality with any of the applied vibration treatments. 17 Even though these preclinical studies are difficult to compare, because they were performed in different species, the inconsistent results suggest a critical influence of the vibration protocols used. The optimum conditions remain to be identified. Some of the above indicated studies 16,17 even used vibration settings exceeding the recommended range for humans, 18 and thus may entail a potential risk of adverse effects. Furthermore, to our best knowledge, studies on mice are currently missing, although mice are increasingly used in fracture healing studies due to the benefit of specific gene targeting. Therefore, the aim of this study was to establish a mouse model to investigate the effect of whole body vibration on fracture healing. To find effective vibration parameters, we selected two LMHFV regimes (0.3g/45 Hz; 0.3g/35 Hz), which were reported to induce anabolic effects in non-fractured murine bone 6 and in 1006

2 WHOLE-BODY VIBRATION IN FRACTURE HEALING 1007 fracture healing in rats. 12 We hypothesized that both treatments would influence bone formation and that the effect would be different depended on the applied frequency. METHODS Animals The animal experiment was performed according to international and national recommendations for the care and use of laboratory animals and was approved by the local ethical committee (No and 1113, Regierungspräsidium Tübingen, Germany). Female 10-week-old C57BL/6NCrl mice (n ¼ 48) were purchased from Charles River (Sulzfeld, Germany) and housed in groups of four animals, allowing free movement within the cage. Study Design At an age of 12 weeks the mice received a femur osteotomy stabilized using an external fixator, and were randomly assigned to a control (n ¼ 20) and two treatment groups (each n ¼ 14). The treatment groups received whole-body vibrations (20 min/day on 5 days/week) with 0.3g peak-to-peak acceleration (a peak-to-peak ) and a frequency (f) of either 35 or 45 Hz, starting on the third postoperative day. Limb loading and locomotion were monitored for a subset of mice of each group during the healing period. The animals were sacrificed 10 days (n ¼ 6 per group) or 21 days (control group: n ¼ 14; treatment groups: n ¼ 8) after surgery. Intact bones (contralateral femurs, lumbar spine) and the osteotomized femurs were evaluated using biomechanical testing, m-computed tomography (mct), and histomorphometry. Surgery Surgery was performed as described previously. 19 Briefly, a 0.4 mm osteotomy gap was created at the mid-shaft of the right femur and stabilized using an external fixator (axial stiffness 3 N/mm, RISystem, Davos, Switzerland) under general anesthesia (2% isoflurane, Forene 1, Abbott, Wiesbaden, Germany). Preoperatively, all mice received a single dose of antibiotics (clindamycin-2-dihydrogenphosphate, 45 mg/kg, Clindamycin, Ratiopharm, Germany). Analgesia (25 mg/l, Tramal 1, Gruenenthal GmbH, Aachen, Germany) was provided via the drinking water 2 days before surgery until the third postoperative day. Measurement of Body Weight, Locomotion, and Ground Reaction Force Four days prior to surgery and 4, 11, and 18 days postoperatively, body weight and ground reaction force (GRF) of the operated limb were determined. 19 For GRF measurements the mice moved freely in an acrylic glass channel containing a force plate (HE6 6, Watertown, MA). The peak vertical GRF of the operated limb was averaged from five measurements for each mouse. The locomotion was recorded using an infrared beam detection system (ActimotMot-System; TSE Systems GmbH, Bad Homburg, Germany). All postoperative values were related to preoperative measurements. Application of LMHFV To apply vertical vibration we used custom-made platforms. The vibration parameters were adjusted exactly and recorded using integrated accelerometers (Sensor KS95B.100, measurement amplifier Innobeamer L2, Software Vibromatrix; IDS Innomic GmbH, Salzwedel, Germany). The platforms were equipped with compartmented cages, allowing the simultaneous treatment of three mice. Starting on the third postoperative day, the mice were placed on the vibration platforms for 20 min per day for 5 days per week, and received either 0.3g/35 Hz or 0.3g/45 Hz. The control mice were sham vibrated. Biomechanical Testing To investigate the mechanical properties, intact and osteotomized femurs explanted on Day 21 were subjected to a non-destructive three-point bending test as described previously. 19 Briefly, the proximal end of the femur was rigidly fixed to an aluminum cylinder, which was then fixed to a hinge joint of the 3-point-bending setup, which allows free rotation in the sagittal plane (Z10, Zwick Roell, Ulm, Germany). The femur condyles rested unfixed on a bending support. An axial load of 4 N with an anterior to posterior direction was applied to the midshaft of the femur in the sagittal plane. Flexural rigidity was calculated from the slope (k) of the linear region of the force-deflection curve. The distance between the load vector and the proximal (a) and distal (b) supports was considered when the fracture callus was not located exactly in the middle of the supports (l/2). Flexural rigidity (EI) was calculated according to the formula for asymmetrical bending: EI ¼ k a2 b 2 3l. Micro-Computed Tomography (mct) On Day 21, femurs and lumbar vertebrae were scanned in a mct device (Skyscan 1172, Kontich, Belgium) at a resolution of 8 mm using a voltage of 50 kv and 200 ma. To assess cortical bone, a volume of interest (VOI) was chosen distal to the third trochanter in the diaphysis of the intact left femur covering 250 slices each of 8 mm. The trabecular bone was evaluated in the distal femurs and the lumbar spine (L6). In the distal femur the trabecular VOI (90 slices of 8 mm) was manually selected comprising the whole medullary cavity starting 240 mm above the most proximal point of the growth plate. In the lumbar spine (vertebral body of L6) a spherical VOI (d ¼ 0.8 mm) was chosen covering the medullary cavity while excluding the cortical shell. In the osteotomized femurs the former osteotomy gap was defined as the VOI (Fig. 1B). Hydroxyapatite (HA) cylinders (250 and 750 mg/cm 3 )wereused to calibrate bone mineral density (BMD) and global thresholding was performed to distinguish between mineralized and non-mineralized tissue at a set BMD of mg HA/cm 3 for both cortical bone and fracture callus, 20 and mg HA/cm 3 for trabecular bone. 21 Common ASBMR standard parameters were evaluated. 22 According to the standard clinical evaluation Figure 1. ROI/VOI scheme for callus histomorphometry and mct evaluation. (A) ROI including the periosteal callus used at Day 10; (B) ROI/VOI including the callus in the osteotomy gap used at Day 21.

3 1008 WEHRLE ET AL. of X-rays, the number of bridged cortices per callus was evaluated in two perpendicular planes (Data viewer, Skyscan). A healed fracture was considered as having a minimum of at least three bridged cortices per callus. Histomorphometry The bone specimens were embedded in paraffin (Day 10) or methyl methacrylate (Day 21). Safranin O (Day 10; n ¼ 6 per group) or Giemsa (Day 21; n ¼ 14 in control group; n ¼ 8in treatment groups) staining was performed on longitudinal sections of the calli. In the region of interest (ROI), the relative amounts of newly formed bone, cartilage, and fibrous tissue were determined using light microscopy and image analysis (Leica DMI6000 B; Software MMAF Version MetaMorph 1 ; Leica, Heerbrugg, Switzerland). On Day 10 the periosteal callus between the inner pins of the fixator and on Day 21 the former osteotomy gap (endosteal and periosteal) was defined as the ROI (Fig. 1). Statistical Analysis Sample size was calculated on the basis of previous studies 19 for two main outcome parameters, flexural rigidity in fractured femurs and trabecular BV/TV in non-fractured femurs (Power: 80%, alpha ¼ 0.05, control group/experimental group: 1.75). Results are presented as the mean standard deviation (SD). Data were tested for normal distribution (Shapiro Wilk normality test) and homogeneity of variance (Levene test). Biomechanical data (normally distributed, homogenous variance) were analyzed using one-way analysis of variance (ANOVA) and a post-hoc Bonferroni test (IBM SPSS Statistics Version 19, Chicago, IL). Data derived from repeated measurements (activity, GRF) underwent two-way ANOVA and Bonferroni correction (GraphPad Prism 6, La Jolla, CA). Both mct and histomorphometry data were assessed using Kruskal Wallis and Dunn s multiple comparison tests (IBM SPSS Statistics Version 19). The level of significance was set at p RESULTS Influence of LMHFV on Body Weight, Locomotion, and GRF All mice recovered rapidly from surgery. The body weight did not significantly change during the healing period and did not differ between the treatment groups (data not shown). The activity of the mice nonsignificantly decreased in the early postoperative period without significant differences between the groups (Fig. 2A). These data indicate that the treatments had little influence on the physical condition of the mice. By Day 4 the GRF was significantly reduced in the 45 Hz group compared to the other groups, but subsequently increased to similar values (Fig. 2B), indicating that 45 Hz vibration caused a feeling of discomfort in the operated limb, resulting in decreased loading. Figure 2. In-vivo monitoring of the mice pre-operatively (pre OP), and on Days 4, 11, and 18 after surgery. The postoperative values were related to the preoperative data. (A) Activity of the mice measured using an infrared beam detection system; (B) Ground reaction force (GRF) of the operated limb. n ¼ 5 9; data shown as mean standard deviation; p 0.05, 45 Hz vibration versus control. Influence of LMHFV on Non-Fractured Trabecular and Cortical Bone Vibration using 35 Hz induced a significant anabolic effect in the trabecular bone of the intact femurs (Table 1). mct analysis revealed a significant increase in bone volume/total volume (BV/TV) by 28% and trabecular number (Tb.N) by 29% compared to nonvibrated controls. The trabecular bone in the lumbar spine was not significantly affected. In the cortical bone we found only a slight increase in the biomechanical and structural bone parameters in the 35 Hz group without reaching statistical significance. In contrast, vibration using 45 Hz did not provoke any positive effect on bone formation (Table 1). The biomechanical and structural bone parameters did not differ significantly compared to non-vibrated controls. The comparison of the two treatment groups demonstrated a lower BV/TV ( 27%) and Tb.N ( 25%), and a higher trabecular separation (Tb.Sp) (þ23%) in the distal femur as well as a reduced trabecular thickness (Tb.Th) ( 10%) in the lumbar spine in the 45 HZ group. The cortical thickness (C.Th) and bone area (B.Ar) of the femur were also significantly reduced by 17% and 18%, respectively, in the 45 Hz group compared to the 35 Hz group suggesting distinct effects of both vibration settings. Influence of LMHFV on Fracture Healing Vibration treatment using 35 Hz did not significantly alter the flexural rigidity of the healed femurs on

4 WHOLE-BODY VIBRATION IN FRACTURE HEALING 1009 Table 1. Evaluation of Intact Cortical and Trabecular Bone using Biomechanical Testing and mct Control Group 35 Hz Group 45 Hz Group Femur EI (Nmm 2 ) 2, , , Cort. bone BMD 1, , , E.Pm (mm) P.Pm (mm) B.Ar (mm 2 ) b C.Th (mm) b Trab. bone BV/TV (%) a b Tb.N. (mm 1 ) a b Tb.Th (mm) Tb.Sp (mm) b Lumbar spine BV/TV (%) Tb.N. (mm 1 ) Tb.Th (mm) b Tb.Sp (mm) n ¼ 6 14; data shown as mean standard deviation. a p 0.05, vibration versus control. b p 0.05, 35 versus 45 Hz. Day 21 (Fig. 3A). mct analysis (Fig. 3B E) and histomorphometry (Fig. 4, Table 2) on Days 10 and 21 did not indicate any significant differences in callus size or the relative amounts of tissues in the fracture calli of the 35 Hz and control groups, indicating that 35 Hz vibrations did not support bone repair. In contrast, vibrations at 45 Hz significantly reduced flexural rigidity by 34% and 42% compared to the non-vibrated control and 35 Hz groups, respectively (Fig. 3A). mct analysis confirmed these results through a decreased callus volume ( 48/ 37%, Fig. 3C) with less newly formed bone in the 45 Hz group compared to sham vibrated mice ( 64%) and mice vibrated with 35 Hz ( 65%) (Fig. 3D). BV/TV was also diminished in the 45 Hz group by 37% and 44% compared to non-vibrated controls and the 35 Hz group, although not significantly (Fig. 3E). Consistent with these findings, all fracture gaps were completely bridged with bone in the non-vibrated and 35 Hz groups, whereas in the 45 Hz group none of the fracture gaps were completely bridged at the four assessed locations and only five of eight mice were classified as healed (Table 3). Histomorphometry did not reveal statistically significant differences in bone formation between vibrated and control mice on Day 10, but a significantly reduced amount of cartilage in the 45 Hz group compared to the 35 Hz group ( 36%) (Fig. 4C and E, Table 2). After 21 days, in contrast to mct data, histology did not reveal any statistically significant differences between the groups. This could be explained by the greater accuracy of three-dimensional mct analysis compared to two-dimensional histology. Furthermore, in mct-evaluation a global threshold was used to segment mineralized and nonmineralized tissue, whereas in histomorphometry newly formed bone was identified by its characteristic morphology, thus also including less mineralized bone in the evaluation. Taken together, our data indicate that a vibration setting of 45 Hz considerably impaired fracture healing. DISCUSSION Because of the proposed osteoanabolic effects of wholebody vibration, we investigated the influence of LMHFV on fracture healing in mice, thereby comparing two common vibration protocols (35 or 45 Hz, 0.3g). We confirmed the anabolic effect of LMHFV on intact bone using 35 Hz vibrations, whereas treatment with 45 Hz was ineffective. In contrast, fracture healing was not improved by either of the LMHFV protocols. While the treatment with 35 Hz had no significant influence, vibrations using 45 Hz significantly impaired cartilage and bone formation. We conclude that small changes in the vibration settings could significantly influence the effects on both bone remodeling and repair, and even disrupt fracture healing, a finding, which implies caution in clinical applications. The present study confirmed previous data demonstrating an anabolic effect of LMHFV on intact bone. 6 9 Our results showed that a 3-week treatment with LHMFV of 35 Hz/0.3g significantly enhanced trabecular bone formation in the mouse femur. The trabecular bone of the lumbar spine and the cortical bone of the femur were not significantly affected. Site-specific effects were also described previously 6,8,23 and could be explained by an unequal distribution of the vibration stimuli within the skeleton and by spatial differences in the mechano-responsiveness of the bone. In contrast to the 35 Hz treatment, a frequency of 45 Hz did not induce bone formation in trabecular or cortical

5 1010 WEHRLE ET AL. Figure 3. Evaluation of fractured femora of the control and the vibration groups on Day 21 using three-point bending testing and mct. (A) Flexural rigidity (EI) of the fractured femur; (B) moment of inertia (I x ) of the callus; (C) total callus volume (TV); (D) bone volume (BV) in the callus; (E) BV/TV. n ¼ 8 14; data shown as mean standard deviation; p bone. Our findings are in agreement with other mouse studies reporting an anabolic influence or no effect using vibrations of 32 Hz/0.5g 8 or 45 Hz/0.3g 23, respectively. In contrast, others found an enhancement in trabecular and cortical bone formation in mice subjected to 45 Hz/0.3g vibration that was ineffective in our study. 6,9 The strong dependency of the observed effects on vibration settings may partly explain the divergent results of the animal experiments. Judex et al. 24 demonstrated that the strains generated in the bone matrix, which were recorded using strain gages, were frequency dependent and that 90 Hz/0.15g induced bone formation in rats whereas 45 Hz/0.15g did not. Other authors varied the acceleration (0.1 1g) at a constant frequency (45 Hz) and similarly found no (0.3g) or positive effects (0.1 and 1.0g) on bone formation in mice. 23 One reason for the difficulty in identifying universally efficient treatment regimes from the literature may be because the transition of the input signals of LMHFV to the bone matrix depends on the body mass, the anatomy of the treated individual, and the posture during vibration, 25 and therefore varies between species. Furthermore, the bone response is dependent on age, 6,8 gender, 2 and genetic predisposition 26 of the individual, explaining the inconsistent results even between studies using the same species. The effects of LMHFV are also influenced by the endocrine status, because it was reported that ovariectomy could sensitize the bone to whole-body vibration. 27 This implies the need for LMHFV protocols which are specifically designed for different target groups. Several studies reported that LMHFV significantly improved fracture healing in rats using vibration settings of 35 Hz/0.3g 12,13,15 or in sheep using 35 Hz and 20 mm interfragmentary movements (IFM). 10 Other studies did not find major influences on bone repair using vibrations of Hz/0.5 mm amplitude in rats 17 and 20 Hz/20 mm IFM in sheep, 11 or even found an impairment of fracture healing. 16 Again, it is difficult to compare the different LMHFV protocols used in the literature, because the transition of the

6 WHOLE-BODY VIBRATION IN FRACTURE HEALING 1011 Figure 4. Representative histological sections of the fractured femora stained using Safranin-O (Day 10) and Giemsa (Day 21), respectively. (A) Control group, Day 10; (B) control group, Day 21; (C) 35 Hz group, Day 10; (D) 35 Hz group, Day 21; (E) 45 Hz group, Day 10; (F) 45 Hz group, Day 21. Scale bar ¼ 500 mm. signals to the callus also depends on the stability of the fracture fixation. A stiff fixation of the fragments may hamper, whereas a more flexible fixation might facilitate the transfer of the vibration stimuli to the fracture gap. In our mouse model vibrations with 35 Hz did not significantly stimulate bone repair, although cartilage formation in the callus was slightly pronounced in the early healing stage and bone formation and bending stiffness were tendentially enhanced later, but failed significance, which may be due to the relatively small group size. Another reason for such minor effects may be, that we used a semirigid fracture fixation, providing suitable mechanical conditions for proper fracture healing at physiological limb loading, as previously demonstrated by our group. 19 In the present study the locomotion of the mice and the loading of the operated limb were not affected by the 35 Hz treatments. The peak strains generated in the fracture callus during limb loading were potentially much higher than the additional small signals induced by the vibration, which did not induce a significant additional effect on fracture healing in our model. However, the strains on the fracture callus were not measured in this study, because it is technically sophisticated. One study reported that in the intact mouse tibia the strains induced by vibration were in the range of 5 me. 2,6 The authors implanted a strain gage on the surface of the tibia. It is technically even more challenging to fix a strain gage on the very small fracture callus (0.4 mm fracture gap). A second reason for the finding that LMHFV was not able to significantly improve fracture healing in our study may be that we used young adult mice exhibiting a good regenerative capacity. Shi et al. 13 reported that LMHFV supported fracture healing more effectively in osteoporotic compared to healthy rats, and proposed that low estrogen levels may be associated with an Table 2. Relative Amounts of Tissues in the Fracture Callus Evaluated using Histomorphometry Callus Composition Control Group 35 Hz Group 45 Hz Group 10 day Bone (%) Cartilage (%) a Fibrous tissue (%) day Bone (%) Cartilage (%) Fibrous tissue (%) ROI at Day 10: periosteal callus; ROI at Day 21: osteotomy gap. n ¼ 6 14; data shown as mean standard deviation. a p 0.05, 35 versus 45 Hz. Table 3. Number of Bridged Cortices of the Calli Evaluated using mct in Two Perpendicular Planes and Number of Mice with Successful Fracture Healing (3 Bridged Cortices) Number of Bridged Cortices Clinical Fracture Healing Outcome Group Group Not Healed Healed Control Control Hz Hz Hz Hz 3 5

7 1012 WEHRLE ET AL. increased response to LMHFV treatment. Further studies should be performed to investigate whether LMHFV supports fracture healing under suboptimal healing conditions. In contrast to the 35 Hz treatment, vibrations of 45 Hz significantly impaired fracture healing in our study. We found a smaller callus, a reduced amount of both cartilage and bone and poor mechanical properties compared to non-vibrated controls and mice subjected to 35 Hz vibrations. Consistently, the fracture gaps of the 45 Hz group were incompletely bridged whereas all other mice healed perfectly, which may have different explanations. Because the 45 Hz treatment significantly diminished limb loading during the early postoperative phase, one could speculate that the load induced peak strains in the fracture callus were too small to stimulate proper callus healing and that this was not compensated by the additional small vibration stimulus. This hypothesis is supported by the observation of a small callus with less cartilage and bone being characteristic of bone healing under inferior mechanical load. 10 In contrast, one could argue that the 45 Hz vibrations induced muscle contractions 28 at the operated limb thus leading to critically high IFM, which would also result in poor healing. 3 However, in this case a large callus with poor quality would be expected. 29 As we were unable to record IFM or strains in our study, the reasons for the negative effects of the 45 Hz vibrations on bone healing remain open. Besides, it was proposed that the LMHFV induced effects may rather be provoked by oscillatory accelerations than by matrix strains, which are several orders of magnitude smaller than the peak strains generated during normal locomotory activities. 2 Future studies will aim to understand the underlying mechanisms of the different vibration effects by incorporating analyses of bone formation and resorption as well as gene expression of fracture callus. In conclusion, the findings of this study demonstrated that the distinct vibration effects on non-fractured and fractured bone were attributable to minor changes in the vibration frequency. While LMHFV of 35 Hz/ 0.3g induced bone formation in intact bone but not in fracture healing, vibration with 45 Hz/0.3g did not affect intact bone but considerably impaired fracture healing. To transfer the present results to a clinical setting, caution should be paid when considering LMHFV application for fracture healing. ACKNOWLEGMENTS This study was funded by the German Research Foundation (DFG) (Research Unit FOR793, IG18/7-2, and SCHI 504/5-2). The authors appreciate the technical assistance of Sevil Essig, Ursula Maile, and Marion Tomo. Each author in this manuscript has not and will not receive benefits in any form from a commercial party related directly or indirectly to the content of this manuscript. REFERENCES 1. Tzioupis C, Giannoudis PV Prevalence of long-bone non-unions. Injury 38:S3 S9. 2. Judex S, Gupta S, Rubin C Regulation of mechanical signals in bone. Orthod Craniofac Res 12: Claes LE, Heigele CA Magnitudes of local stress and strain along bony surfaces predict the course and type of fracture healing. J Biomech 32: Rubin C, Recker R, Cullen D, et al Prevention of postmenopausal bone loss by a low-magnitude, high-frequency mechanical stimuli: a clinical trial assessing compliance, efficacy, and safety. J Bone Miner Res 19: Lau E, Al-Dujaili S, Guenther A, et al Effect of lowmagnitude, high-frequency vibration on osteocytes in the regulation of osteoclasts. Bone 46: Xie L, Jacobson JM, Choi ES, et al Low-level mechanical vibrations can influence bone resorption and bone formation in the growing skeleton. Bone 39: Tezval M, Biblis M, Sehmisch S, et al Improvement of femoral bone quality after low-magnitude, high-frequency mechanical stimulation in the ovariectomized rat as an osteopenia model. Calcif Tissue Int 88: Wenger KH, Freeman JD, Fulzele S, et al Effect of whole-body vibration on bone properties in aging mice. Bone 47: Xie L, Rubin C, Judex S Enhancement of the adolescent murine musculoskeletal system using low-level mechanical vibrations. J Appl Physiol 104: Goodship AE, Lawes TJ, Rubin CT Low-magnitude high-frequency mechanical signals accelerate and augment endochondral bone repair: preliminary evidence of efficacy. J Orthop Res 27: Wolf S, Augat P, Eckert-Hubner K, et al Effects of high-frequency, low-magnitude mechanical stimulus on bone healing. Clin Orthop Relat Res Leung KS, Shi HF, Cheung WH, et al Low-magnitude high-frequency vibration accelerates callus formation, mineralization, and fracture healing in rats. J Orthop Res 27: Shi HF, Cheung WH, Qin L, et al Low-magnitude high-frequency vibration treatment augments fracture healing in ovariectomy-induced osteoporotic bone. Bone 46: Cheung WH, Sun MH, Zheng YP, et al Stimulated angiogenesis for fracture healing augmented by low-magnitude, high-frequency vibration in a rat model-evaluation of pulsed-wave doppler, 3-D power doppler ultrasonography and micro-ct microangiography. Ultrasound Med Biol 38: Chow DH, Leung KS, Qin L, et al Low-magnitude high-frequency vibration (LMHFV) enhances bone remodeling in osteoporotic rat femoral fracture healing. J Orthop Res 29: Stuermer EK, Komrakova M, Werner C, et al Musculoskeletal response to whole-body vibration during fracture healing in intact and ovariectomized rats. Calcif Tissue Int 87: Komrakova M, Sehmisch S, Tezval M, et al Identification of a vibration regime favorable for bone healing and muscle in estrogen-deficient rats. Calcif Tissue Int 92: Muir J, Kiel DP, Rubin CT Safety and severity of accelerations delivered from whole body vibration exercise devices to standing adults. J Sci Med Sport 16: Röntgen V, Blakytny R, Matthys R, et al Fracture healing in mice under controlled rigid and flexible conditions

8 WHOLE-BODY VIBRATION IN FRACTURE HEALING 1013 using an adjustable external fixator. J Orthop Res 28: Morgan EF, Mason ZD, Chien KB, et al Microcomputed tomography assessment of fracture healing: relationships among callus structure, composition, and mechanical function. Bone 44: O Neill KR, Stutz CM, Mignemi NA, et al Microcomputed tomography assessment of the progression of fracture healing in mice. Bone 50: Parfitt AM, Drezner MK, Glorieux FH, et al Bone histomorphometry - standardization of nomenclature, symbols, and units. J Bone Miner Res 2: Christiansen BA, Silva MJ The effect of varying magnitudes of whole-body vibration on several skeletal sites in mice. Ann Biomed Eng 34: Judex S, Lei X, Han D, et al Low-magnitude mechanical signals that stimulate bone formation in the ovariectomized rat are dependent on the applied frequency but not on the strain magnitude. J Biomech 40: Kiiski J, Heinonen A, Jaervinen TL, et al Transmission of vertical whole body vibration to the human body. J Bone Miner Res 23: Robling AG, Turner CH Mechanotransduction in bone: genetic effects on mechanosensitivity in mice. Bone 31: Rubinacci A, Marenzana M, Cavani F, et al Ovariectomy sensitizes rat cortical bone to whole-body vibration. Calcif Tissue Int 82: McKeehen JN, Novotny SA, Baltgalvis KA, et al Adaptations of mouse skeletal muscle to low-intensity vibration training. Med Sci Sports Exerc 45: Claes L, Augat P, Suger G, et al Influence of size and stability of the osteotomy gap on the success of fracture healing. J Orthop Res 15: