Research evidence suggests that low-intensity pulsed ultrasound (LIPU) produces significant osteoinductive effects, accelerating the healing of bone defects. The authors investigated the effects of LIPU on mandibular bone defects in a rabbit model. Fifty-six adult Dutch rabbits were divided randomly into control, LIPU-1 (1 MHz), and LIPU-3 (3 MHz) groups. A mandibular defect was created in all rabbits. The effect of LIPU on mandibular defects was assessed by frequency (1 or 3 MHz) and timing (2 and 4 weeks). Bone mineral density (BMD) was measured and stereology and histology performed; results were compared at the end of 2 and 4 weeks. LIPU-3 resulted in significantly higher bone formation compared to the control group at the end of week 4 on histological assessment ( P = 0.008). BMD was significantly higher at 4 weeks than at 2 weeks ( P = 0.03). LIPU-3 increased the numerical density of osteoblasts and osteocytes at the end of week 4 ( P = 0.05 and P = 0.001, respectively). The results of this study are in favour of using LIPU 3 MHz to accelerate mandibular bone healing. However, this study suggests that a frequency of 3 MHz and the longer application of LIPU 3 MHz for 4 weeks can only promote 8% mandibular bone healing in rabbits. Therefore, the use of LIPU has no really convincing, consistent clinical effects on maxillofacial bone healing.
The healing of bone defects in the maxillofacial region is always a challenge in cases of congenital or traumatic defects or tumour resection. Rapid soft tissue growth inside the defect may block bone formation at the defect periphery, resulting in impaired defect healing. A relative lack of certain tissue factors in the centre of the defect, which originate from the edge of the defect, is believed to limit the bone healing process.
Several methods for managing bone defects have been introduced following research in the field of maxillofacial bone regeneration, such as the use of bone grafts and/or barrier membranes. Recently, the effect of ultrasound on bone defect healing has been assessed. Well-designed prospective studies have been reported, indicating that low-intensity pulsed ultrasound (LIPU) can accelerate fracture healing of the long bones in animal models and the fracture repair process in the tibia and radius.
Reports on the application of LIPU in the maxillofacial area are conflicting ( Table 1 ). Ustun et al. reported that LIPU may have positive effects on osseointegration and the stability of dental implants. However, Schortinghuis et al., in two separate studies, reported that there was no statistically significant difference in the percentage of defect closure between groups of rats exposed to LIPU and control rats. El-Bialy et al., in two different studies, showed that LIPU 1.5 MHz accelerated bone formation in mandibular bone distraction and suggested that the later stages of healing were enhanced more by pulsed ultrasound. Erdogan et al. found considerable contributions of LIPU (1.5 MHz) to bone healing in mandibular fractures.
|Reference||Species ( n )||Frequency, MHz||Follow-up, weeks||Evaluation methods||Results|
|El-Bialy et al., 2002||Rabbit (21)||1.5||4||Bone photodensity, vibratory coherence, mechanical stiffness, histological studies||Accelerated bone formation evaluated by photodensitometric, vibratory, elastic, and histological techniques|
|Schortinghuis et al., 2004||Rat (72)||1.5||2, 4||Microradiographs||No statistically significant difference in the percentage of defect closure between the groups|
|Schortinghuis et al., 2005||Rat (64)||1.5||2, 4||Microradiographs||No statistically significant difference in the percentage of defect closure between the groups|
|Erdogan et al., 2006||Rabbit (30)||1.5||3||Three-point bending test, digital radiodensitometric analysis, histological and histomorphometric examinations||Considerable contributions to bone healing in mandibular fractures|
|El-Bialy et al., 2008||Rabbit (36)||1.5||1, 2, 3, 4||Quantitative bone density (QBD), mechanical testing, and histological examination||Earlier stages of bone healing were enhanced more by continuous ultrasound, whereas late stages were enhanced more by pulsed ultrasound|
These previously reported studies used 1.5 MHz ultrasound in the maxillofacial area. Although some compared the bone healing process based on weeks post application, the effect of different frequencies has not been evaluated in this region. Taking into account the reported positive influences of ultrasound on the bone healing processes with the disturbed function of impaired bone healing in large maxillofacial defects, it was decided to investigate the potential of LIPU to stimulate mandibular bone defect healing at two different frequencies, over 2 and 4 weeks, in a rabbit model.
Materials and methods
The study was approved by the university ethics committee. Fifty-six mature male Dutch rabbits weighing a mean 2.4 ± 0.2 kg and aged a mean 18 ± 2 months were included ( Table 2 ). Unilateral mandibular defects were created in all animals and they were subsequently divided into three treatment groups, which were subdivided further into six subgroups. The first two subgroups of animals (G1 and G2, n = 11 in each) received 1 MHz ultrasound treatment daily, 10 min/day, for 2 weeks (G1) or 4 weeks (G2). The second two subgroups (G3 and G4, n = 11 in each) received 3 MHz ultrasound with the same protocol for 2 weeks (G3) or 4 weeks (G4). Animals in the two control groups (G5 and G6, n = 6 in each) received sham application for 2 weeks (G5) or 4 weeks (G6).
|Group||LIPU||Time of sampling after LIPU, weeks||Number of rabbits|
|G5 (control 1)||None||2||6|
|G6 (control 2)||None||4||6|
All rabbits underwent general anaesthesia by intramuscular injection of 44 mg/kg ketamine 10% (Alfasan International BV, Woerden, the Netherlands) and 8 mg/kg xylazine 2% (Alfasan International BV). After placing the animal in a supine position, the mandible was exposed with a 3-cm incision in the diastema area. A unilateral extraoral submandibular approach was used and a 5 mm × 2 mm × 1 mm bone defect was created 5 mm medial to the mental foramen on one side using a round bur (ELA Carbide; Emil Lange – Zahnbohrerfabrik e.K., Engelskirchen, Germany); all layers were then sutured ( Fig. 1 ).
An analgesic (flunixin 5%, 0.15 mg/kg; Erfan Daru Pharmaceutical, Tehran, Iran) and an antibiotic (penicillin 22,000 IU/kg; Erfan Daru Pharmaceutical) were administered intramuscularly preoperatively and twice per day for four postoperative days. The rabbits were housed in separate cages and fed soft food for 1 week. After the first week, a normal diet was resumed. Food and water intake and the weights of the animals were monitored and recorded daily. Animals that had a weight loss of more than 20% of their initial body weight were excluded from the study.
On the second postoperative day, the application of LIPU was started for the animals in the four experimental groups. A commercially available therapeutic ultrasound device (new version 215× class 1 type BF; Nvin Medical Engineering Co., Isfahan, Iran) was used for the ultrasound treatment. The device transmits pulsed ultrasound signals with an operating frequency of 1–3 MHz, which produces 0.5 W/cm 2 average temporal and spatial intensity. Ultrasound applications were performed after placing the animal in a box in order to restrict excessive movements. No sedative agent was given during the ultrasound treatments. An investigator observed the whole ultrasound treatment session to ensure continuity of the application.
The 10-min sessions were repeated on a daily basis for 2 or 4 weeks according to the study group (G1–G4). The same procedure was applied to the animals in the two control groups, without activating the ultrasound device (G5 and G6). On the day after the last LIPU application, all animals were killed with an intravenous injection of 100 mg/kg sodium thiopental (Rotexmedica, Trittau, Germany). The whole mandibles were harvested and soft tissues were stripped off. The mandibles were split at the midline using a scalpel. Animals in groups G1, G3, and G5 were killed on postoperative day 15, while animals in G2, G4, and G6 were killed on postoperative day 30. Bone mineral density (BMD) measurements and stereology and histological examinations were performed on the harvested hemimandibles.
The BMD of the prepared regions of the mandible was measured using dual-energy X-ray absorptiometry (DXA) (Hologic, Lunar Corp., Madison, WI, USA); the DXA device was calibrated daily, in accordance with the manufacturer's recommendations. The regional high-resolution mode of the small animal scan protocol was used.
To measure BMD using DXA, the specimens were positioned centrally at the bottom of a square, thin-walled plastic container. The level in the water bath was set at 8 cm. One at a time, a hemimandible was placed into the container and the device was set to start the test. Each test took about 3 min. A specific area of 0.47 cm 2 was selected and focused on. This area contained the defect created in the mandible. The mandibular bone density and bone content of the rabbits in each group was measured. All DXA measurements and analyses were performed by the same investigator.
The specimens were fixed in 4% neutralized formalin for 2 weeks, decalcified in 10% formic acid, dehydrated, and finally embedded in paraffin. Coronal sections approximately 5 μm in thickness were prepared through the centre of the defects, stained with haematoxylin–eosin (H&E), and observed under an optical microscope (Nikon, Tokyo, Japan). A single blinded pathologist scored the stained sections according to the amount of mineralization in the fracture gap, using the grading system described by Perry et al. With this system, grade 1 indicates fibrous union, grade 2 indicates predominantly fibrous tissue with some cartilage, grade 3 indicates equal amounts of fibrous tissue and cartilage, grade 4 indicates all cartilage, grade 5 indicates predominantly cartilage with some woven bone, grade 6 indicates equal amounts of cartilage and woven bone, grade 7 indicates predominantly woven bone with some cartilage, grade 8 indicates woven bone, grade 9 indicates woven bone and some lamellar bone, and grade 10 indicates lamellar bone. The mean score was calculated for each group.
The volume density, V v (bone or cavity), of the trabecular bone and the bone marrow cavity was estimated using the point-counting method and the following formula: V v (bone or cavity) = P (bone or cavity)/ P (ref), where P (bone or cavity) and P (ref) are the total numbers of points hitting the bone, cavity, and the reference space ( Fig. 2 ).
The optical disector was used to estimate the total number of osteoblast, osteocyte, and osteoclast cells in the defect region of the mandible. For this purpose, the authors used software designed at the Histomorphometry and Stereology Research Centre, Shiraz University of Medical Sciences. The area of the defect was demarcated and counts were done at random locations. The area of the counting frame (a/f) was 21 μm × 21 μm. Section thickness was used as the height of the disector ( h ), excluding the 4-μm thick guard zones at the top and bottom of each section. To calculate the suitable guard zone, z -axis distribution of the nuclei was plotted.
The counted cells were scored and grouped in 10 histograms from percentiles 0–100 through the mandible tissue section, from the upper surface (0%) to the lower surface (100%).
Fig. 3 shows the z -axis distribution of the nuclei. The upper and lower 20% of the histogram were considered as the guard zones and the counting box was placed over the remaining 60% ( h ). According to the histogram, under-sampling was balanced out and corrected. Any nucleus coming into maximal focus within the next focal sampling plane was selected if it was located completely or partly inside the counting frame and did not touch the exclusion line ( Fig. 4 ).
Post-shrinkage thickness was measured during cell counting and used to determine an average thickness of 20 μm ( t ). All cells were distinguished based on differences in morphological characteristics and were counted separately within the frame. In order to determine cell densities, osteoblast, osteocyte, and osteoclast cells counted in the defect area were divided by the total volume of the counting frames as follows:
where ∑ Q − is the number of the nuclei coming into focus during scanning of the section thickness, ∑ P is the total counting of the unbiased counting frame in all fields, and h is the height of the disector. On average, 60–80 osteoblasts, 90–120 osteocytes, and 10–20 osteoclasts were counted per area of defect. a/f is the frame area, t is the real section thickness measured using the microcator, and BA is the block advance of the microtome, which was set at 26 μm. The total number of osteoblasts, osteocytes, and osteoclasts was estimated by multiplying the numerical density ( N v) by V (defect area).
Estimation of the coefficient of error (CE)
The CE ( N v) for the estimate of the numerical density of the osteoblast, osteocyte, and osteoclast numbers, was calculated using the following formula : CE ( N v) = [( n / n − 1) × [(∑( Q − ) 2 /∑ Q − ∑ Q − ) + (∑( P ) 2 /∑ P ∑ P ) − (2∑( Q − P )/∑ Q − ∑ P )]] 1/2 .
Differences between the average BMD and stereology data for the left mandible in a period of 2 and 4 weeks between groups of 1 MHz, 3 MHz, and control were analyzed by one-way analysis of variance (ANOVA) and the least significant difference (LSD) post hoc test. Differences between the average BMD and stereology data of groups 1 MHz, 3 MHz, and control in a period of 2 and 4 weeks were compared using the paired t -test. Histology data were analyzed using non-parametric tests including the Kruskal–Wallis test and Mann–Whitney U -test. The plot was generated using the survival option of GraphPad Prism version 5.01 for Windows (GraphPad Software Inc., San Diego, CA, USA). All data are presented as the mean (±standard deviation). Probability values of P < 0.05 were considered significant.
Eight animals were excluded from the study due to pneumonia and excessive weight loss. The final numbers of animals in each of the six groups were as follow: G1: n = 10 (1 MHz/2 weeks); G2: n = 8 (1 MHz/4 weeks); G3: n = 9 (3 MHz/2 weeks); G4: n = 9 (3 MHz/4 weeks); G5: n = 6 (control/2 weeks); and G6: n = 6 (control/4 weeks).
Analysis of BMD showed a significantly higher value at the end of week 4 compared to week 2 in the group that received LIPU 1 MHz (G2 vs. G1, P = 0.03) ( Fig. 5 ). No significant differences were seen in the comparisons of the other groups ( P > 0.05).
The Kruskal–Wallis test revealed a statistically significant difference in the grade of bone healing across the six different groups ( P = 0.03). The LIPU 3 MHz for 4 weeks group (G4) recorded a higher median score (Md = 8) than the other groups. To determine the difference between each group, the Mann–Whitney U -test was performed. Results of this test revealed a statistically significant difference in grade of bone healing in the group that received LIPU 3 MHz (G4) compared to the control group (G6) at the end of week 4 ( P = 0.008) (the Bonferroni correction was done). No significant differences were seen in comparisons of the other groups (G2 vs. G6, P = 0.34; G1 vs. G5, P = 0.26; G3 vs. G5, P = 0.11).