Osteoporosis is a major health problem to the elderly population as well as to the society. This is characterized by low bone mass, deterioration of the osseous tissue microarchitecture, increased bone fragility and susceptibility to fracture.1 With complications associated with osteoporotic fracture including deteriorated health conditions and increased mortality, the focus of modern fracture management is not only on normal fracture but especially on osteoporotic condition to boot.
Fracture healing is a very complicated process, with its full mechanism being far from thoroughly understood. Nonetheless, both normal and osteoporotic fracture healings employ the same regenerative process: hematoma formation, soft callus formation, callus mineralization, and callus remodeling. With study suggesting a reduced capacity of bone healing and bone remodeling in osteoporotic fracture,2 osteoporosis was found to decrease the responsiveness of bone to mechanical loading, as well as impairing both angiogenesis3 and osteoprogenitor cell recruitment.4 Moreover, the decreased osteogenic capacity with aging5 and reduced osteogenic stimulations from physical inactivities6 could lower the osteoinductive capacity as shown in demineralized bone matrix.7
With advances in the fields of medical sciences and biophysical engineering, the enhancement on bone fracture healing under non-invasive mechanical stimulation has become possible. The use of low intensity pulsed ultrasound (LIPUS) was demonstrated with a significant enhancement in the proliferation and differentiation of human periosteal cells,8 as well as an increase in bone regeneration9 and distraction osteogenesis.10 Together with clinical trials showing the accelerated healing in complex fracture,11 tibial fracture,12 and non-union,13 the use of LIPUS was shown to have an impressive effect on normal fracture healing. By the same token, the efficacy of using LIPUS on the osteoporotic counterpart was, therefore, expected, we hypothesized that LIPUS could accelerate the osteoporotic fracture healing.
The objective of this study was to investigate the effects of LIPUS on normal and osteoporotic bones fracture healings, with multi-modality analyses including radiography, micro-computed tomography (micro-CT), histomorphometry, and biomechanical tests.
A total of 120 female Sprague–Dawley (SD) rats were used. At 6 months old, 60 rats were randomly chosen for bilateral ovariectomies (OVX) according to our established protocol14 (FDA-verified animal model of osteoporosis15). Sham OVX surgical procedures were performed on the remaining 60 rats and part of the visceral fat, instead of the ovaries, was resected as control. With standard chow provided, rats were housed for 3 months.16 The reduction in bone mineral density (BMD) was confirmed by peripheral quantitative computed tomography (pQCT, Densiscan 2000; Scanco Medical, Brüttisellen, Switzerland), through which the 5th lumbar vertebra, the right femoral head and the right femoral shaft were measured.
Three months after OVX and Sham procedures, closed femoral shaft fractures were performed according to our established protocol,17 which was modified from the Einhorn protocol.18 A sterilized Kirschner wire (K-wire, ϕ1.2 mm; Sanatmetal Ltd., Eger, Hungary) was inserted into the medullary canal retrogradely, following drilling and reaming with an 18 G needle (ϕ 1.27 mm, Fig. 2.1.1C). The K-wire then perforated the proximal femur through the piriformis fossa, and the tip was bended to leave a 3 mm length to prevent distal migration. The distal end of the K-wire was cut at the level of the articular surface to allow free joint movement. A custom-made 3-point-bending apparatus, with a metal blade (weighted 500 g) dropping from a height of 35 cm, was used to create fracture on the midshaft of the femur. The quality of fracture with fracture gap smaller than 0.5 mm and displacement <0.5 mm was confirmed by anteroposterior (A-P) and lateral radiographies. All procedures were conducted by one experienced orthopedic surgeon. For analgesia, single dose of buprenorphine (0.03 mg/kg, s.c.; Temgesic, Schering-Plough, NJ) was given 15 min before the surgery as well as on a daily basis for the next 3 days. Rats were allowed unrestricted cage activities at all time.
Within the Sham and the OVX groups, rats were sub-divided randomly into the control group and the treatment group. There were, as a result, four groups: (1) sham control (Sham-C), (2) sham treatment (Sham-T), (3) ovariectomized control (OVX-C), and (4) ovariectomized treatment (OVX-T). With six rats in each group, LIPUS treatments (Exogen 3000+; Smith & Nephew Inc, Memphis, TN) were given to the Sham-T and OVX-T groups at 20 min a day, 5 days a week starting on the next day post fracture. During treatment, rats were anesthetized by using the appropriate dosage of ketamine–xylazine–saline mixture. They were laid on the ventral side and the probe from the LIPUS machine was placed over the lateral side with respect to the fracture site. The skin was shaved and a thin layer of coupling gel was applied between the probe and the contact surface. For the control groups, the same treatment routine was followed only that the LIPUS machines were left turned off (sham treatment) for the same duration of time as the treatment groups. Rats were euthanized by overdosed sodium pentobarbital at 2, 4, and 8 weeks. Radiological analysis was performed using weekly radiography.17 Micro-CT,17 histomorphometric assessment17 and mechanical test were carried out using freshly harvested femora.17 (Different patches of harvested samples were used for the micro-CT/histimorphometry assessments and those for the mechanical test. And there were only 4 and 8 weeks time points for the mechanical test.) All the 120 rats (n = 6 each for four groups and three-time points on radiography/micro-CT/histomorphometry; n = 6 each for four groups and two-time points for mechanical test) were obtained from the Laboratory Animal Services Center of the Chinese University of Hong Kong. The Animal Experimentation Ethics Committee of the authors' institution approved the care and experimental protocol of this study (Ref. No. 07/010/ERG).
Fracture healing was monitored weekly on A-P and lateral radiography (Faxitron X-ray Corporation, model 43855C, Wheeling, IL), based from our established protocol.17 Qualitatively, both the A-P and lateral radiographs were assessed by two surgeons and the frequencies on callus bridging were scored. Bone sample was counted as bridged if: (i) bone bridging of the four cortices was observed in both AP and the lateral radiographs, and (ii) it must receive a consensus from both the surgeons. Quantitatively, the temporal changes of the callus morphology were measured using the lateral radiograph with Metamorph Image Analysis System (Universal Imaging Corporation, Downingtown, PA)17. The callus width (CW) and the callus area (CA) were measured. CW was defined as the maximal outer diameter of the mineralized callus minus the outer diameter of the femur, whereas CA was the total sum of the external mineralized CA.
Based on our established protocol,17 the micro-CT (µCT40, Scanco Medical, Brüttisellen, Switzerland) was used to scan region of interest (ROI) 5 mm proximal and distal to the fracture line. Three-dimensional (3-D) image of mineralized tissue was created using a low pass Gaussian filter (sigma = 1.2, support = 2). The contoured ROI was selected under two-dimensional (2-D) CT images. The low-density (threshold = 165–350) and high-density (threshold = 350–1000) tissues were reconstructed separately to differentiate the newly formed mineralized callus (low-density) from those old cortices (high-density) using our established protocol,19 which was modified from the Gabet's one.20 Normalized percentage of the tissue volumes (bone volume, BV/total bone volume, TV) was calculated.21
Freshly harvested femora were first fixed in 10% neutral buffered formalin and then in 8% formic acid for decalcification. The femora were cut into halves along the mid-sagittal plane. The specimens were embedded into paraffin. Sectioned samples 7 µm in thickness were mounted on the silane-coated glass slides and were stained with safranin-O/fast green (Saf-O).22 The region covering 1.5 mm proximal and distal to the fracture line was evaluated. The cartilage area (Cg.Ar) was measured with Metamorph Image Analysis System.17
Four-point bending test was performed at 4 and 8 weeks post-treatment.17 A material test machine (H25KS Hounsfield Test Equipment Ltd. Redhill, Surrey, UK) was used to test the femora to failure with a constant displacement rate at 5 mm/min. The femora were loaded in the anterior–posterior direction, with the inner and outer span of the blades at 8 and 20 mm apart, respectively.17 The values of ultimate load, stiffness, and the energy to failure were calculated and analyzed using the QMAT Professional Material testing software.
All quantitative data were expressed as mean ± standard deviation (SD) and analyzed with SPSS 13.0 (SPSS Inc, Chicago. IL). One-way analysis of variance (ANOVA) with Bonferroni post hoc tests were used to compare the differences among groups at the corresponding time points, unless otherwise specified. Statistical significance was set at p < 0.05.
During the study, the body weights of all the rats were remained constant with no difference from baseline values (At Start: 308 ± 8.4 g; End Point: 311 ± 8.9 g; Paired t-test: p = 0.468).
Peripheral Quantitative-Computed Tomography
Three months after the ovariectomzied procedures, the changes in BMD values at the 5th lumbar vertebra, the right femoral head and the right femoral shaft were measured and the reductions by 9.6%, 4.6%, and 2.3% were found, respectively.
For callus bridging assessment on radiography, the time for Sham-C, Sham-T, OVX-C, and OVX-T were found at weeks 5, 4, 6 and 3, respectively (Fig. 1A). Quantitative comparison of the earliest bridging time among four groups showed 5.08 weeks for Sham-C, 4.75 weeks for Sham-T, 5.5 weeks for OVX-C, and 4.17 weeks for OVX-T, while significant difference was found between OVX-T and OVX-C (p = 0.001; Fig. 1B). Table 1 also summarizes the bridging rate at weeks 4 and 8, which showed both Sham-T and OVX-T contained more samples with completed bridging than their controls. Non-parametric analyses (Kruskal–Wallis) indicated significant difference among four groups at week 4 (p = 0.012) and week 8 (p = 0.001), while further pairwise comparison (Mann–Whitney) with Bonferonni correction showed significant difference between OVX-T and OVX-C at week 4 (p = 0.0004) and week 8 (p < 0.001) but not between Sham groups (week 4: p = 0.187; week 8: p = 0.09). Quantitatively, both the CW and CA increased from weeks 2 to 4, at where they both reached their peaks (Fig. 2A and B, respectively). The OVX-T was significantly higher than the OVX-C at week 2 (CW was higher by 30.1%, CA was higher by 39.1%; CW, CA: p < 0.001), week 3 (CW was higher by 18.8%, CA was higher by 38.5%; CW, CA: p = 0.001) and week 4 (CW was higher by 21.7%, CA was higher by 33.6%; CW: p = 0.002, CA: p = 0.001). The Sham-T was significantly higher than the Sham-C at week 2 (CW was higher by 15.2%, CA was higher by 28.9%; CW: p = 0.015, CA: p = 0.009). Moreover, the CW of the OVX-T was significantly higher by 15.0% than the Sham-T at week 4 (p = 0.015).
Table 1. Comparison of Callus-Bridging Rate of Femoral Fracture Healing at Weeks 4 and 8
There was significant difference among four groups at week 4 (p = 0.012) and week 8 (p = 0.001) by Kruskal–Wallis test, while pairwise comparison (Mann–Whitney test) with Bonferroni correction indicated significance between OVX-T and OVX-C at week 4 (p = 0.004) and week 8 (p < 0.001) but not in Sham groups at both time points (week 4: p = 0.187; week 8: p = 0.09).
At week 4, Sham-T and OVX-T were shown to have better callus-bridging responses than their corresponding control groups (Fig. 3). At week 8, similar observation with better-remodeling responses were observed in both the Sham-T and OVX-T groups. At both weeks 4 and 8, OVX-T was shown to have the best callus-bridging and bone-remodeling processes. Table 2 showed the quantitative micro-CT analysis on BV/TV. Although no significant difference was observed between any groups at any time points, two noticeable trends were observed. First, an overall increasing trend of BV/TV value was observed. Second, from weeks 2 to 4, the increased in BV/TV values from Sham-T (increased by 18.7%) and OVX-T (increased by 26.1%) were higher than the Sham-C (increased by 2.96%) and OVX-C groups (increased by 16.3%).
Table 2. Quantitative Micro-CT Analysis on BV/TV Values
0.4889 ± 0.0840
0.5066 ± 0.0473
0.4852 ± 0.0467
0.4483 ± 0.0331
0.5803 ± 0.0502
0.5216 ± 0.0586
0.6119 ± 0.0497
0.5215 ± 0.0606
0.6985 ± 0.0370
0.7206 ± 0.0432
0.6620 ± 0.0236
0.6612 ± 0.0160
Safranin O/Fast Green staining was performed which indicated the presence of proteoglycan with a red color. Figure 4 showed Sham-T was significantly higher than the Sham-C at week 2 (p < 0.001) and week 4 (p < 0.001), and OVX-T was significantly higher than the OVX-C at week 2 (p = 0.004) and week 4 (p = 0.007). On the comparison between Sham-T and OVX-T, Sham-T was significantly higher at week 4 (p = 0.007). In addition, OVX-T was found to decrease the most (decreased by 11.6%) from weeks 2 to 4.
Figure 5 showed the results from mechanical test in ultimate load (A), stiffness (B) and energy-to-failure (C). Under the ultimate load assessment, at week 4 the Sham-T and OVX-T were significantly higher than the Sham-C (p = 0.019) and OVX-C (p = 0.015), respectively. This same trend of better treatment groups than the control groups was as well observed in stiffness measurement. At week 8, the stiffness value from the OVX-T group was significantly higher by 119.0% than the OVX-C group (p = 0.018). Moreover, between OVX-T and Sham-T groups, the stiffness of OVX-T was higher, without significance, than that of the Sham-T by 37.4% and 36.9% at weeks 4 and 8, respectively. On energy-to-failure (C), the OVX-T was significantly higher than the OVX-C at weeks 4 (p = 0.017) and 8 (p = 0.006). The Sham-T was significantly higher than the Sham-C at week 8 (p = 0.023). At week 4, the Sham-T was higher than the Sham-C without significant by 45.6%.
This study investigated the possible effect of LIPUS on fracture healing per se, as well as particularly on osteoporotic fracture healings. From the corresponding perspectives of all the results, the use of LIPUS was shown to have an enhancing effect on both normal and osteoporotic fracture healings. Both the treatment groups were shown to have significantly enhanced callus formation, faster mineralization and better remodeling. Moreover, on the comparison between the Sham-T and the OVX-T groups, the OVX-T group was shown to have a significantly faster early phase callus formation than the Sham-T, faster callus bridging (earlier endochondral ossification), and comparable mechanical properties (stiffness, ultimate load and energy-to-failure) as Sham-T group. Since the osteoporotic bone was shown to have slower bone-healing response,2–7 with most results comparable to its normal counterpart, this non-invasive biophysical intervention of LIPUS was confirmed to enhance osteoporotic fracture healing.
LIPUS enhanced the fracture healings on both normal and osteoporotic bones. In accordance to parameters reflected on callus formation, the CW and CA measurements on Sham-T and OVX-T groups were shown to be significantly higher than their control groups during the period from weeks 2 to 4, representing a better intramembranous ossification. This was consistent with a previous study which showed the callus formed rapidly during the first 12 days of LIPUS treatment.23 From the weekly radiographs, the callus bridging of both the treatments were observed to be around 4.17 weeks (OVX-T) and 4.75 weeks (Sham-T), which were faster than both the control groups where bridging were around 5.08 weeks (Sham-C) and 5.5 weeks (OVX-C). This result was further consolidated through the histology analysis in which faster decreases in values of Cg.Ar from both the treatment groups were observed, indicating a faster early phase endochondral ossification. This enhanced endochondral ossification results in a greater callus formation through augmenting mineral deposition (mineralization),24 leading to a narrowing fracture gap and increased cortical bone mass. This was similar to the study of Wang et al.25 who showed more advanced endochondral ossification and a smaller fracture gap after 14 days of ultrasound treatment. This faster rate of callus bridging could also be observed in quantitative micro-CT analysis in which greater increases in BV/TV values from the treatment groups than the control groups from weeks 2 to 4 were observed. These enhanced callus formation, increased bone volume, cortical bone mass, and mineral apposition all indicated that LIPUS might enhance the anabolic activity of osteoblasts, especially early in the differentiation as a lineage.26 In addition to enhanced callus formation (confirmed by increased CW, CA) and mineralization [confirmed by micro-CT, faster decline of CA (after week 4) and Cg.Ar], LIPUS treatment was, as well, shown to accelerate the restoration of mechanical properties including the ultimate load, energy to failure and stiffness (significant differences were observed from the first two parameters). Other studies also showed similar improved mechanical results.19 The efficacies of LIPUS on both the normal and osteoporotic fracture healings were demonstrated, from the aspects of callus formation, mineralization, and remodeling.
Under the intervention of LIPUS, the OVX-T showed comparable healing responses than the Sham-T group in most parameters, while OVX groups indicted relatively more significant differences in various assessments than Sham groups. The results showed OVX-T with a significantly higher CW measurement (week 4), earlier and higher percentages of completed callus bridging (Fig. 1B, Table 1), higher ratio of increment in BV/TV value, faster response of endochondral ossification and a higher stiffness measurement. This was consistent with our previous study, in which the use of low-magnitude high-frequency vibration (LMHFV) demonstrated relative better effects on osteoporotic fracture healing than on the age-matched non-osteoporotic one.21 With altered bone metabolism in osteoporosis, the mechanical and biological factors that are involved in the bone fracture healing process are certainly affected.27 This alteration in healing led to a delay in callus maturation and consequently decelerated the fracture healing process.27 This was in agreement with our study, in which the OVX-C was shown to have the lowest healing response. An in vitro study using LIPUS on cells from osteoporotic rat showed acceleration of osteoblast cell differentiation by increasing mineralization of the fracture-healing process.28 This was concurred in this study with the OVX-T showed the highest CW and CA (from weeks 2 to 4), the fastest drop in CW (by 35.8%) from weeks 4 to 5 and the fastest fracture gap bridging (micro-CT images, weekly radiographs and the greatest decreased in Cg.Ar). Moreover, significantly higher CW was found in the OVX-T than the Sham-T. Another study from Rubinacci et al.29 also showed osteoporotic condition could sensitize cortical bone with cyclic mechanical stimulation. This might provide an explanation on our findings that OVX groups showed more significant enhancements in response to LIPUS treatment in various assessments, as compared with Sham groups.
As mentioned before, our previous study reported beneficial effects of LMHFV on both normal and osteoporotic fracture healings.21 Together with the findings of the current LIPUS study, it was implied that osteoporotic condition could sensitize osseous tissue under external intervention. LMHFV is a physical vibration that was applied systemically. This vibratory stimulation might not exceed the minimum effective strain30 (MES) in order to stimulate response at certain point during the treatment as rat was in constant movement and changing body posture all along. For LIPUS, acoustic signal was applied directly on the region of interest consistently. Whether this might lead to an efficacy difference between the two methods remained unknown, further investigation would be necessary. Nonetheless, both showed enhanced osteoporotic fracture healing responses. As the action of estrogen was purposed a negative modulator of the mechanotransduction process through estrogen receptor-beta (ER-beta) signaling pathway,19 further studies of ER-beta in response to LIPUS shall be performed. On limitations, due to different samples with heterogeneous tissue compositions, we were not able to measure the energy difference that will be reaching the fracture sites. Although the rat model used in this study is FDA-verified animal model of osteoporosis,15 the osteoporosis status of the rats could not be absolutely confirmed due to the lack of BMD norm for SD rats. Furthermore, the low BMD drop at femoral shaft due to its cortical properties may not be a good indication of osteoporosis, although this is a well-accepted site for closed fracture model. On micro-CT assessment, a single threshold was used on the measurements of all the groups. However, since bone mineralization was not constant for all the groups, this approach might not reflect the true values of the mineralization at the fracture site.31
In conclusion, the efficacies of LIPUS on both the normal and osteoporotic fracture healings were demonstrated. These were reflected through enhanced callus formation, faster mineralization and remodeling, and improved mechanical strength. Furthermore, although with decreased repair potential, osteoporotic bone was shown with comparable healing responses as the normal bone under LIPUS intervention, implying the good clinical application potential of LIPUS for elderly fracture patients.
This project was supported by General Research Fund (Ref: 462707) and AO Research Fund (Ref: S-07-1C).