SEARCH

SEARCH BY CITATION

Keywords:

  • stress fracture;
  • bone remodeling;
  • bisphosphonate;
  • risedronate

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Because bisphosphonates (BPs) are potent inhibitors of bone resorption, we hypothesized that they would retard direct remodeling of stress fractures. The aim of this study was to determine the effect of risedronate on direct remodeling and woven bone callus formation following stress fracture formation in the rat ulna. In 135 adult female Wistar rats, cyclic loading of the ulna created stress fractures. Rats were treated daily with oral saline, or risedronate at 0.1 or 1.0 mg/kg. From each bone, histomorphometry was performed on sections stained with toluidine blue at a standard level along the fracture. The high dose of risedronate caused a significant decrease in the percentage of repaired stress fracture and bone resorption along the stress fracture line at 6 and 10 weeks after loading (p < 0.05). At this dose, intracortical resorption was significantly reduced at 10 weeks after loading and intracortical new bone area was significantly reduced at 6 and 10 weeks. Woven bone formation and consolidation phases of stress fracture repair were not affected by low or high doses of risedronate. In conclusion, high dose bisphosphonate treatment impaired healing of a large stress fracture line by reducing the volume of bone resorbed and replaced during remodeling. We also confirmed that periosteal callus formation was not adversely affected by risedronate treatment. © 2011 Orthopaedic Research Society Published by Wiley Periodicals, Inc. J Orthop Res 29:1827–1833, 2011

Bisphosphonates are potent inhibitors of osteoclasts and bone resorption, and are widely used to treat conditions of excessive bone remodeling, including osteoporosis. In osteoporosis, reduced bone mechanical properties can lead to fatigue damage even under normal loading patterns. Many osteoporotic fractures are complete fractures, however some will occur in the form of incomplete fractures similar to stress fractures seen in athletes.1, 2 While bisphosphonates reduce the risk of fractures in osteoporosis, fractures still occur during treatment, so it is necessary to understand the effect of these medications on subsequent fracture healing.2–5

Stress fractures are an important clinical condition commonly affecting athletes, soldiers, and dancers.6 They are a form of fatigue injury associated with repetitive loading. Excessive bone remodeling in response to repetitive loading during training may contribute to the development of stress fractures1, 7 and it is likely that there will be further interest in bisphosphonates for the prevention and treatment of remodeling associated with fatigue injury in bone. The use of bisphosphonates has been proposed as a preventive measure against stress fractures.8, 9 However a trial in army recruits failed to show any preventive efficacy of risedronate therapy.8 Treatment with pamidronate was reported to produce a beneficial effect in clinical stress fractures in young athletes,10 but this study only had small numbers of cases. Of clinical interest, there have been recent case series of displaced transverse diaphyseal femoral fractures with minimal trauma in patients after long-term bisphosphonate treatment.2 These fractures were classified as being typical of stress fractures.4, 11 There is however, very little information available on the effect of bisphosphonate treatment on progression or repair of stress fractures.

There is also a growing focus on the use of bisphosphonates for a variety of orthopedic conditions that can be affected by an imbalance between catabolic and anabolic processes during bone healing. In experimental studies of complete fracture in rats, treatment with bisphosphonates consistently increased the hard callus volume, strength and mineral content.3, 12, 13 In animal models, stress fractures heal by direct remodeling along the fracture line.14, 15 Treatment with bisphosphonates may therefore have a negative effect on stress fracture healing when compared to complete fractures. For example, alendronate reduced the number of resorption spaces observed along stress fracture lines induced by ulnar loading in the rat. The woven bone response was not affected by alendronate treatment initiated before or after loading.9 These effects were, however, only examined at a maximum of 2 weeks following loading. Another study using the rat ulna loading model demonstrated that treatment with aledronate did not affect resorption space area, repaired crack length or woven bone area at 2 and 4 weeks after stress fracture creation but that mineral apposition rate and bone formation rate were decreased at 4 weeks after loading.16 While mechanical properties were not significantly altered by alendronate treatment at 8 weeks after loading, histology was not examined at this later timepoint.16

Repetitive axial loading of the rat ulna has been used to create reliable stress fractures and associated woven bone. The histology and gene expression profile of stress fracture healing in this model has been extensively characterized previously.15 Stress fracture healing in this model consists of two distinct healing components; the rapid woven bone response on the periosteal surface and direct remodeling, in the form of numerous basic multicellular units (BMUs) progressing along the fracture line.14, 15, 17, 18 The aim of this study was to determine the effect of long-term daily risedronate treatment on the healing of stress fractures in the rat ulna. High and low doses were selected to determine if a dose–response existed, and oral treatment was used to represent a clinical protocol. The gastrointestinal tract absorbs bisphosphonates quite poorly. However, most clinical dosing is still by oral administration and effective doses can be administered this way with few adverse effects.19

Previous studies have used thick sections of plastic embedded bone to examine the histomorphometry of stress fracture healing in the rat ulnar loading model.9, 14 This provides a method for measuring flourochrome labels and bone formation indices, but does not allow for detailed examination of resorption and formation indices of healing along the fracture line. We used thin sections of paraffin embedded bone in this study to allow a detailed analysis of the cellular resorption and formation activity within the BMUs, and thus provide valuable information on the effect of bisphosphonates on the progression and activity of cells within a BMU. Stress fracture healing and woven bone response was examined out to 10 weeks after loading to determine the long-term effect of risedronate treatment on stress fracture healing and woven bone consolidation.

We hypothesized that risedronate treatment would impair and delay the resorption and subsequent bone formation within the BMUs during remodeling along the fracture line. We also hypothesized that treatment with risedronate would not prevent the woven bone response but consolidation of the woven bone, as measured by the reduction in woven bone area over time, would be impaired.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Treatment Groups

The right ulna was loaded in 135 rats. Female Wistar rats aged 16–19 weeks of age and weighing 290–320 g at the time of loading were used. The Animal Ethics Committee of The University of Queensland approved all procedures. Drug treatment was commenced immediately after loading. For each timepoint rats were either treated with saline (VEH), or risedronate at a low dose rate 0.1 mg/kg (RisLo) or a high dose rate 1.0 mg/kg (RisHigh). Risedronate or vehicle were administered daily by oral gavage. These dose rates were estimated as equivalent to approximately one fifth of the standard 35 mg per week human clinical dose (low dose) and twice the standard clinical dose (high dose).a Risedronate was made up as a 0.1 or 1.0 mg/ml solution in the saline and given at 1 ml/kg. Thus for a 300 g rat the dose administered was 0.3 ml. Rats were weighed daily to ensure the correct dose. Food and water intake were also monitored. Each rat received VEH, RisLo or RisHigh by oral gavage daily until the end of the experiment. Rats were euthanized at 2, 6, or 10 weeks after loading (n = 15 rats per group).

Loading Model

Loading sessions were conducted by placing the flexed carpus and elbow in cups and applying axial cyclic loading on the ulnae of rats using methods as described previously.15 Loading was performed in a custom designed loading device which has previously been described in detail.20 To determine displacement in the limb during loading, the loading device was attached to a linear vertical displacement transducer (LVDT) (made in the School of Biomedical Sciences electrical workshop) connected to a maclab (AD Instruments, Colorado Springs, Colorado) and monitored using Chart v5.4 (AD Instruments). Isoflourane and oxygen general anesthesia was used. Loading was applied as cyclic compressive loading at 17–24 N load and 2 Hz cycle frequency. The load range was selected based on previous reports in the literature using similar sized rats, and commenced at 17 N. In all rats the displacement value decreased over the first 1,000–2,000 cycles, and then stabilized, until a rapid increase in displacement leading up to fatigue failure. Loading was manually stopped at the point when a 10% increase in displacement was reached, compared to the lowest measured displacement rather than the initial displacement. The total number of cycles to fatigue ranged from 3,800–13,200 cycles. A single injection of an opioid analgesic (Buprenorphine 0.05 mg/Kg subcutaneously) was used following loading sessions. No evidence of lameness or behavioral changes was seen following loading, nor was food or water intake different among groups.

Paraffin Embedding and Sectioning

A 10–15 mm section of the right ulna containing the fracture was removed from the bone and fixed in cold 10% neutral buffered formalin for 2 days and stored in 70% ethanol until processed. Bones were decalcified for 7 weeks in 14% EDTA (pH 7.2) before being embedded in paraffin using standard protocols. Multiple 5 µm thick transverse sections were obtained along the full length of the fracture line. Toluidine blue staining was performed. Examination of multiple serial sections along the entire length of the fracture was undertaken for qualitative assessment of fracture healing.

Histomorphometry

One reader analyzed the sections and was blinded to group affiliation. Histomorphometry was performed on one toluidine blue-stained section from each bone at a standard level along the fracture. This section was the point where the fracture line was half way between the medial cortical margin and the medullary cavity of the bone in transverse section.15 Toluidine blue staining provided excellent cellular and structural detail. Areas of woven bone, areas of osteoclastic resorption and cement lines between the original and newly formed intracortical bone could be clearly distinguished and measured (Fig. 1). Histomorphometry measurements were obtained using a digitizing pad (Wacom Intuo 2, Vancouver, WA) and KSS software (KSS Stereology, Salt Lake City, UT). Measurements were made directly from histology slides using a mouse to trace the identified areas (resorption and new bone formation, or along the crack) on the digitizing pad connected to the viewing microscope. Cortical and woven bone measurements obtained were cortical area (Ct.Ar, mm2) and woven bone area (Wo.B.Ar, mm2). Woven bone area as a percent of cortical bone area (Wo.B.Ar/Ct.Ar %) was calculated to correct for any differences in cortical area between sections.

thumbnail image

Figure 1. Diagram outlining histomorphometric measurements of a typical transverse section. The area outlined in black is the total resorption area (Rs.Ar) and the area outlined in red, within this, is the area of new bone (Ne.B.Ar). The resorption space is progressing along the fracture line from the margin of the periosteal woven bone (Wo.B). [Color figure can be seen in the online version of this article, available at http://wileyonlinelibrary.com/journal/jor]

Download figure to PowerPoint

Measurements obtained from along the fracture line were total length of the stress fracture (crack) (Cr.Le, µm), total length of repair along the crack (Re.Cr.Le, µm) and length of new bone along the crack (Ne.B.Cr.Le, µm). Length of repair included crack that had previously been resorbed by osteoclasts with or without newly formed bone within the BMU. The percentage resorption along the crack (Re.Cr.Le/Cr.Le, %) and percentage of new bone along the crack (Ne.B.Cr.Le/Cr.Le) were calculated to correct for variations in the total length of the fracture line. Intracortical measurements obtained were total area of intracortical resorption (Rs.Ar, µm2), and intracortical area of healed new bone (Ne.B.Ar, µm2). The areas of resorption and new bone were normalized by the cortical area (Rs.Ar/Ct.Ar, % and Ne.B.Ar/Ct.Ar, %).

Statistical Methods

All variables were checked for normal distribution using a Kolmogorov–Smirnov test. Variables that were not normally distributed were Ne.B.Cr.Le/Cr.Le, Ne.B.Ar, and Re.Cr.Le/Cr.Le. For these variables non-parametric tests were performed. A Kruskal–Wallis ANOVA was used to determine if there was a significant treatment effect and differences between groups were assessed using a Mann–Whitney test. Significance was determined if p < 0.05. All remaining variables were normally distributed, and were analyzed using a two-way ANOVA. If there was a significant treatment effect, differences between groups were assessed using a one-way ANOVA and post hoc Fisher's least significant difference.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Stress Fracture Healing

Stress fractures were consistently in the distal half of the diaphysis on the medial (compressive) surface of the bone. There were no significant differences between any of the groups for fracture length or cortical area indicating that there was a consistent level of damage at a standard position within the bone. The configuration and qualitative assessment of the remodeling, healing, and woven bone responses in all groups were very similar to those described previously by us.15

Histomorphometry

Woven Bone Formation

Results for cortical and woven bone histomorphometry are summarized in Table 1. Woven bone area normalized by cortical area (WoB.Ar/Ct.Ar) was rapidly produced in the first 2 weeks after loading, with WoB.Ar/Ct.Ar greatest at the 2 weeks time-point. This was decreased again by 6 weeks but the decrease was not statistically significant. WoB.Ar/Ct.Ar remained stable between 6 and 10 weeks. There were no significant differences in WoB.Ar/Ct.Ar between treated and control groups at any time-point.

Table 1. Cortical and Woven Bone Histomorphometry Results for VEH, RisLo, and RisHigh Treated Groups
 2 Weeks6 Weeks10 Weeks
 VEH (n = 13)RisLo (n = 15)RisHigh (n = 15)VEH (n = 13)RisLo (n = 14)RisHigh (n = 15)VEH (n = 15)RisLo (n = 14)RisHigh (n = 14)
  1. Values are mean (SEM) for cortical area (Ct.Ar), woven bone area (Wo.B.Ar) and woven bone area as a percentage of original cortical area (Wo.B.Ar/Ct.Ar %) for daily treatment with vehicle (VEH), risedronate (0.1 mg/kg, RisLo) or risedronate (1.0 mg/kg, RisHigh) at 2, 6, or 10 weeks after loading. There were no significant differences between groups for any of these values.

Ct.Ar (mm2)1.36 (0.08)1.26 (0.05)1.33 (0.04)1.30 (0.04)1.33 (0.05)1.38 (0.06)1.31 (0.05)1.31 (0.03)1.38 (0.05)
Wo.B.Ar (mm2)0.75 (0.08)0.79 (0.08)0.76 (0.05)0.63 (0.03)0.68 (0.03)0.62 (0.04)0.61 (0.03)0.63 (0.04)0.70 (0.04)
Wo.B.Ar/Ct.Ar (%)57 (7)64 (6)57 (3)49 (2)52 (2)45 (2)48 (4)48 (3)51 (3)
Remodeling along the Fracture Line

Results for histomorphometry are summarized in Figures 2 and 3. At 2 weeks, active remodeling could be observed along the fracture line. This was primarily in the region where the fracture exited the cortex. Remodeling at 2 weeks was in the form of large resorption spaces that were progressing along the fracture line towards the medullary cavity region. These resorption spaces had a crenated edge and were almost completely filled with cellular activity including obvious multinucleated osteoclasts. There was rarely any new bone at this time-point. Six weeks after loading, more individual resorption spaces could be seen and these had progressed further along the fracture line towards the medullary cavity. There was a rapid increase in the area and length of resorption spaces and associated new bone between 2 and 6 weeks after loading. This new bone formation progressed more slowly between 6 and 10 weeks.

thumbnail image

Figure 2. Effect of treatment with vehicle (control), risedronate at 0.1 mg/kg, (RisLo) or at 1.0 mg/kg (RisHigh) on the percentage of resorption length along the crack (Re.Cr.Le/Cr.Le) (A) and the percentage of new bone along the crack (Ne.B.Cr.Le/Cr.Le) (B) at 2, 6, and 10 weeks after loading. Values are mean ± SE for each group. ap < 0.05 compared to control group at same timepoint.

Download figure to PowerPoint

thumbnail image

Figure 3. Effect of treatment with vehicle (control), risedronate at 0.1 mg/kg, (RisLo) or at 1.0 mg/kg (RisHigh) on area of intracortical resorption normalized by the cortical area (Rs.Ar/Ct.Ar) (A) and the areas of intracortical new bone normalized by the cortical area (Ne.B.Ar/Ct.Ar) (B) at 2, 6, and 10 weeks after loading. Values are mean ± SE for each group. ap < 0.05 compared to control group at same timepoint.

Download figure to PowerPoint

High Dose Risedronate Significantly Reduces Bone Resorption and Formation Within the BMUs

In the RisHigh group, the percentage of the crack that had been repaired (Re.Cr.Le/Cr.Le) was significantly less than in the control group at the two week time-point (p < 0.05). At six weeks Re.Cr.Le/Cr.Le, the percentage of new bone along the crack (Ne.B.Cr.Le/Cr.Le) and the areas of intracortical new bone normalized by the cortical area (Ne.B.Ar/Ct.Ar) were all significantly reduced in the RisHigh group compared to the control group (p < 0.05). At 10 weeks, Re.Cr.Le/Cr.Le, Ne.B.Cr.Le/Cr.Le, Ne.B.Ar/Ct.Ar and the areas of intracortical resorption normalized by the cortical area (Rs.Ar/Ct.Ar) were all significantly reduced in the RisHigh group compared to the control group (p < 0.05; Figs. 2 and 3).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Risedronate is one of the nitrogen-containing bisphosphonates, widely used clinically, particularly to treat osteoporosis.21, 22 Bisphosphonates are potent inhibitors of osteoclastic bone resorption and have effects on osteoclast recruitment, differentiation, activity, and apoptosis.23–26 This study examined the effects of bisphosphonate treatment on stress fracture healing over an extended period. In contrast to complete fractures, stress fractures require early resorption during healing to enable new bone formation along the fracture line.15 We observed that daily risedronate treatment, at a dose equivalent to twice that recommended for treatment of osteoporosis, impaired stress fracture healing in the rat ulna up to 10 weeks after loading. Measures of resorption and subsequent new bone formation associated with the fracture line were significantly reduced at 6 and 10 weeks after loading and the length of the repair along the crack was significantly reduced at 2, 4, and 6 weeks after loading.

Most studies elect to use a subcutaneous administration route for risedronate studies in animals. However, we decided to use the oral route as this is the routine method of administration of risedronate in human patients. We selected 2 dose rates to determine if a dose response could be detected and the high dose rate (1.0 mg/kg) was selected to ensure an anti-resorptive effect was likely to occur while remaining within a clinically relevant dose range. Oral dose rates of 0.5 and 2.5 mg/kg risedronate prevented bone loss in an ovariectomy model in rats while 0.1 mg/kg had no effect.27 Reported results of pre-clinical testing for risedronate suggest that the dose rate needed in a rat to be equivalent to the standard 35 mg/kg/week human dose is 0.5 mg/kg.a

Bone resorption along the stress fracture line was clearly depressed by treatment with risedronate. The area of intracortical resorption was significantly reduced in the high-dose group at 10 weeks after loading. The stress fracture in this model heals by remodeling with BMUs progressing along the crack. The length of the crack undergoing healing depends on the progression of both osteoclastic and osteoblastic activity, and that length was significantly reduced in the high-dose treatment groups at all timepoints. New bone formation was also significantly reduced by treatment at the high dose, 6 and 10 weeks after loading. This likely reflects the decreased resorption surface available for new bone formation. The ulnar loading model has provided a method for examining the effects of BP treatment on the progression of focal BMUs over a predictable time-course and direction, at a known anatomical location. These results demonstrate that BPs can delay and reduce the volume of bone resorbed during the initiation and progression of remodeling in response to a large fatigue crack.

The reduction in newly formed bone along the stress fracture line is consistent with previous work demonstrating that alendronate reduced bone formation rate and mineral apposition rate at 4 weeks after loading in this model.16 The reduced bone formation within BMUs along the fracture line in this study was likely to have resulted from the reduced bone resorption during fracture remodeling, rather than a direct effect of risedronate on bone formation. There is some evidence that bisphosphonates can affect bone formation independently of osteoclast activity. Although bisphosphonates promote apoptosis of osteoclasts, they also reduce apoptosis of osteocytes and may inhibit osteoblast activity.28–30 Mineral apposition rate on the periosteal surface of the femur and tibia was reduced with risedronate or alendronate treatment in growing rats. As this periosteal modeling should be independent of osteoclast activity, this may indicate a direct effect of bisphosphonates on osteoblast activity.31 Most bisphosphonate compound is bound to bone mineral and only osteoclasts are able to release and take up bound bisphosphonates.32 The effects of bisphosphonates on osteoclast function are therefore likely to be more prominent than effects of transiently unbound bisphosphonates on other cell types.32, 33

In the current study, treatment with risedronate at high and low doses following stress fracture formation did not change the area or thickness of the woven bone. The woven bone in this model restores the mechanical properties of the ulna by 2 weeks after loading,34, 35 and this woven bone response was unaffected by risedronate treatment even at the high dose. Woven bone formation following rat ulnar loading was also unaffected by alendronate treatment given for 2 weeks prior to loading.9 In complete fractures, proliferative new bone is not affected by treatment with bisphosphonates. In complete fractures in rats, bisphosphonate treatment does not alter the rate of bone union and consistently results in a larger callus.12, 13, 36 Similar results were observed in models using dogs and sheep.37, 38 In these studies, mechanical properties of the callus were improved3, 12, 13, 36 or not significantly changed by treatment with bisphosphonates.38, 39 Woven bone is a prominent feature of clinical stress fractures and provides mechanical protection to a stress fracture line.40, 41

The area of the woven bone was not significantly affected by bisphosphonate treatment in the current study, even at 10 weeks after loading. The woven bone area decreased in the period between 2 and 6 weeks after loading, in all treatment groups. This indicates that the consolidation of this woven bone was not affected by osteoclast inhibition. It is possible that there is such a prolific resorption response that osteoclast recruitment and activity over the large surface area of highly vascular periosteum is able to progress sufficiently to compensate for ongoing inhibition and loss of osteoclasts by bisphosphonates.

Bisphosphonates are not currently used clinically to treat stress fractures in athletes. Nonetheless, there is growing interest in their potential to prevent excessive remodeling that may predispose to stress fracture development. Milgrom et al.8 tested this hypothesis by treating Israeli military recruits with Risedronate during basic training. Recruits were treated with a loading dose of 30 mg of risedronate daily for the first 10 days then the standard dose rate of 30 mg weekly for a further 12 weeks. This did not reduce the incidence of stress fractures compared to placebo treatment. However, the authors suggested that the dose used may have produced an excessive reduction in remodeling.8 The effect of bisphosphonate treatment on the subsequent healing of the stress fractures that did occur was not reported. In the ulnar loading model, 2 weeks of treatment with alendronate, prior to loading, did not protect the bone from fatigue damage.9 However, that effect was tested using a single loading session where remodeling could not have contributed to the development of the resulting stress fractures. The stress fracture produced in the rat ulna model differs in its creation to those observed clinically. In particular, the effects of bisphosphonate treatment on the progressive accumulation of fatigue damage and remodeling over time, prior to fracture creation, cannot easily be examined using this model. Diaphyseal fractures of the femur, in patients on long-term bisphosphonates, have also been described as fatigue fractures.2, 4, 5, 11 Although we have not studied long-term treatment using a clinical dose, our data suggest a possible mechanism whereby inhibition of direct remodeling of the stress fracture line could retard its repair, leading to greater bone fragility.

The results of this study demonstrate that a bisphosphonate can delay and reduce the volume of bone resorbed during the initiation and progression of remodeling in response to a large stress fracture line. However woven bone formation and subsequent consolidation were not affected by high doses of risedronate even over 10 weeks duration. This suggests that the periosteal reaction to stress fracture is not negatively affected by risedronate, even when treatment is continued after the stress fracture was sustained. This woven bone response is able to return the ulna to its original strength within 2 weeks of stress fracture creation in the rat ulna loading model.35, 42 It is therefore possible that woven bone callus is able to protect the site of the fracture and reduce any detrimental effects of a delay in stress fracture remodeling in clinical stress fracture cases.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

This study supported in part by National Health and Medical Research Council Project Grants 401553 and 511187; and the Rebecca L Cooper Medical Research Foundation. We thank Procter and Gamble Pharmaceuticals Inc for the kind donation of Risedronate sodium (NE-58095).

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES