- Email: [email protected]
Increased cortical remodeling after osteotomy causes posttraumatic osteopenia
Bone 43 (2008) 539–543
Contents lists available at ScienceDirect
Bone j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b o n e
Increased cortical remodeling after osteotomy causes posttraumatic osteopenia Peter Augat a,b,⁎, Lutz Claes c a b c
Biomechanics Research Laboratory, Paracelsus Medical University, 5020 Salzburg, Austria Biomechanics Research Laboratory, Trauma Center Murnau, 82418 Murnau, Germany Institute of Orthopaedic Research and Biomechanics, University of Ulm, 89081 Ulm, Germany
a r t i c l e
i n f o
Article history: Received 4 February 2008 Revised 9 May 2008 Accepted 13 May 2008 Available online 30 June 2008 Edited by: H. Genant Keywords: Fracture healing Osteopenia Bone density Bone histomorphometry Remodeling
a b s t r a c t Following a fracture, substantial bone mineral loss can occur at the affected limb. The aim of this study was to analyze the changes in cortical bone around the site of a fracture. We analyzed bone mineral density by quantitative computed tomography and quantified changes in cortical remodeling by histomorphometry adjacent to an experimental osteotomy in sheep metatarsals. In the cortical bone around the osteotomy, we found a statistically significant 16% reduction in app.BMD within 9 weeks following surgery. This reduction was explained (R = −0.71, P b 0.01) by a more than 6 fold increase in bone remodeling activity within cortical bone at the affected limb. The remodeling activity significantly increased between surgery and week 6, but remained unchanged between week 6 and week 9. We conclude from these findings that posttraumatic bone mineral loss adjacent to a fracture is related to an elevated number of active osteons, indicating a significant increase in bone remodeling activity. Load shielding by the osteosynthesis material and local recruitment of bone mineral are likely causes for this increased remodeling. This post-traumatic bone loss is likely to contribute significantly to frequently observed healing complications like refracture, failure of implant fixation, implant loosening, or cut out. © 2008 Elsevier Inc. All rights reserved.
Introduction Fracture healing occurs through formation of periosteal callus tissue (secondary healing) or increased bone remodeling (primary healing or metaphyseal healing) at the site of the fracture. During the healing process, the amount of bone in the fractured area initially increases. In diaphyseal healing, this is typically perceptible by the formation of hard callus visible on X-rays. However, there is also considerable regional bone loss at sites adjacent to the fracture. Previous studies demonstrated that bone loss around a fracture can amount to almost 30% loss in the same bone [1,2] and more than 10% in neighboring bones [3]. This loss can be observed as early as 6 weeks after fracture [4,5]. Although some recovery is observed after completion of healing [6], the bone mass deficit may persist for many years [7,8]. Although fracture related bone loss occurs to a large extent at sites rich in trabecular bone, there have also been reports of significant bone loss at diaphyseal locations after fracture [2,5,9]. The changes observed at diaphyseal locations were mostly observed with dual X-ray absorptiometry (DXA) which generates a 2D projectional measurement of bone mineral density [2,5,9]. As this measurement is affected by the projected area, any periosteal callus formation that increases the projected area would incorrectly detect a decrease in
⁎ Corresponding author. Paracelsus Medical University Salzburg, and Director Biomechanics Laboratory, Trauma Center Murnau, Prof. Kuentscher Str. 8, 82418 Murnau, Germany. Fax: +49 8841 484573. E-mail address: [email protected] (P. Augat). 8756-3282/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2008.05.017
bone mineral. Measurements with computed tomography, which are not susceptible to this error, were only performed at trabecular rich sites with rather small cortices [2,5,9]. Therefore these previous studies did not entirely reveal to which extent diaphyseal locations are affected by post traumatic bone loss. Because loss of bone mineral is associated with a reduction in mechanical competence [10], the consequence of fracture associated bone loss is weakening of the fractured bone as well as bones adjacent to the fracture. The reduction in bone strength increases the risk of fracture and may also contribute to failure in the maintenance of fracture reduction after osteosynthetic fixation. Therefore, fracture associated bone loss is likely to play a significant role in the occurrence of refractures and loosening of osteosynthetic implants. Although the occurrence of fracture associated bone loss has been well described in the literature, there is limited research on the local distribution, the underlying mechanism and the reason for this phenomenon. It is not known whether the fracture associated bone loss in the diaphysis occurs within the cortex, at the periosteal or at the endosteal surface. Furthermore, there is no knowledge about the specific distribution of bone loss around the fracture, particularly whether bone loss differs proximally or distally from the fracture site. The aim of this study was to quantitatively analyze how a fracture affects the cortical bone around the site of the fracture. We hypothesized that the post-fracture loss of bone mineral is caused by increased cortical remodeling and that variations in cortical remodeling can be observed between proximal and distal locations. Therefore, we assessed cortical remodeling in the diaphysis of ovine
540
P. Augat, L. Claes / Bone 43 (2008) 539–543
Fig. 1. Measurement of bone density with pQCT. Scan locations were identified on an anterior–posterior scout view (right panel). On the left non-osteotomized limb, three equidistant scans were placed in the center of the metatarsus (left panel, top). On the osteotomized limb, three scans were performed: at the center of the osteotomy (left panel, middle) and 15 mm proximal and distal (left panel, bottom) of the osteotomy, respectively.
long bones adjacent to a healing osteotomy by histomorphometric measurement of active osteons. Materials and methods Cortical bone remodeling was studied in the osteotomized and contralateral hind limb of sheep in the course of a bone healing process. Twelve skeletally mature Merino sheep (female, 2–3 years of age, average weight 73 ± 9 kg) received a transverse osteotomy at the mid diaphysis of the right metatarsus under general anesthesia. The osteotomy gap was adjusted to 2 mm and the external ring fixator allowed for interfragmentary movements between 0.2 mm and 1.0 mm. The surgical procedure and the external fixator procedure are described in greater detail elsewhere [11,12]. Following the operation, the animals received metamizol (20 mg per kg bodyweight; Novalgin®, Hoechst, Frankfurt, Germany) for pain relief for a period of
5 days and were allowed full freedom of movement. To monitor remodeling activity within cortical bone, the fluorochrome markers, Calcein (0.33 ml/kg body weight; Calceingreen; Synopharm, Barsbuettel, Germany) and Tetracycline (1.0 ml/kg body weight; Tetracycline-hydrochloride; Caesar and Loretz GmbH, Hilden, Germany), were administered intravenously at 4 and 6 weeks after surgery, respectively. After 9 weeks the animals were killed, the metatarsals were harvested, and the fixator was removed. The protocol was performed in accordance with the Guide for Care and Use of Laboratory Animals and was approved by the animal care and use committee (approval No. 543, Regierungspraesidium Tuebingen). After explantation both metatarsals were analyzed for apparent bone mineral density using a peripheral quantitative computed tomography scanner (pQCT, XCT 960, Stratec, Pforzheim, Germany). For the control limb, three scans 1 cm apart from each other were performed at the center of the metatarsus. For the osteotomized limb,
Fig. 2. Histologic picture (25-fold magnification, bar indicates 400 μm) of osteons under fluorescent light from the untreated control limb (left) and the limb that received an osteotomy (right). Green and yellow labels indicate bone formed at week 4 and week 6, respectively.
P. Augat, L. Claes / Bone 43 (2008) 539–543
541
Table 1 Histomorphometric analysis of density of active osteons in cortical bone Active osteon density in # / mm2 Controla Osteotomy- week 4 Osteotomy week 6
0.06 ± 0.08 0.39 ± 0.33 0.96 ± 0.51
Osteotomy week 9
0.69 ± 0.30
P b 0.01 compared to control P b 0.001 compared to control P b 0.01 compared to week 4 P b 0.001 compared to control P b 0.01 compared to week 4
Average of distal and proximal location (mean ± standard deviation) from measurement in n = 12 sheep. a Density of active osteons in the control group depended neither on location nor time after osteotomy and was therefore averaged from proximal and distal locations and for all three time points.
Fig. 3. app.BMD 9 weeks after surgery of cortical bone proximal and distal from the osteotomy was significantly (P b 0.001) smaller compared to app.BMD in the nonosteotomized control limb. Box and Whisker plot showing median, 25%-, and 75%percentiles, minimum, and maximum of app.BMD measurements in 12 sheep.
the level of the osteotomy was identified on a planar scout view and three pQCT scans were performed (Fig. 1). The first scan was placed at the center of the osteotomy gap, including the periosteal callus formation, followed by scans at 1.5 cm distally and 1.5 cm proximally from the center of the osteotomy gap, respectively. The distal and proximal scans were therefore located between the positions of the two outermost fixator pins, respectively. The axial scans of the diaphysis were analyzed for apparent bone mineral density (app. BMD) of the cortical compartment and total bone mineral content (BMC). Both metatarsals were then processed for undecalcified bone histology and embedded in polymethylmethacrylate. Axial crosssectional slices of 70 μm thickness were obtained 3 cm distally and proximally from the center of the osteotomy. From the nonosteotomized control metatarsus, two slices were obtained 2 cm proximally and distally from the center of the metatarsus. The slices were cut with a histologic precision saw (Exact System, Nordenstadt, Germany) under constant cooling and were smooth grinded. The sections were surface stained with Paragon (Paragon C and C; New York, NY, USA), which labeled newly formed bone and active osteons in blue. The histological slices were examined with conventional light and fluorescence microscopy (Axiophot; Zeiss, Oberkochen, Germany) at 25-fold magnification. Active osteons were counted within the cortical bone compartment, which appear green or yellow under fluorescent light or blue under conventional light (Fig. 2). The
histological slices were then digitized at 12.5-fold magnification and the images were imported into an image analysis system (analySIS; Soft Imaging Systems, Muenster, Germany). By manual segmentation, the periosteal and endosteal bone surfaces were identified and the cortical bone area was measured. Density of active osteons was used as a measure of remodeling activity and was calculated by dividing the number of osteons by the area of the cortical bone cross-section [13]. Repeated analysis of five histological slices by two observers revealed an inter- and intraobserver reproducibility of 9% and 7%, respectively. Shapiro–Wilk test was used to control for normal distribution of the data. Paired sample Student's t-tests were employed to test for significant differences between osteotomized and control limbs, and differences between proximal and distal locations. The density of active osteons at different time points and at different locations was compared using a repeated measures analysis of variance and a Tukey post hoc test. Resulting P-values were adjusted for multiple comparisons using the Bonferroni adjustment. The SPSS statistical software package (SPSS 14.0, Chicago, Illinois) and Excel (Microsoft Corp.) were used for data analysis. Results All animals had uneventful healing of the osteotomy with two animals showing minor signs of pin infection. The app.BMD did not differ between distal and proximal locations (P N 0.2) and was on average 16% lower at the osteotomized limb (859 ± 50 mg/cm3; P b 0.001) compared to the app.BMD at the non-osteotomized control limb (Fig. 3, 1023 ± 58 mg/cm3). The central 1 mm slice at the site of the former osteotomy, which included the periosteal callus apposition in the osteotomized bone, contained 41% (P b 0.02) more mineralized tissue than the central slice in the control limb.
Fig. 4. Density of active osteons 4, 6, and 9 weeks after osteotomy in the control limb and the osteotomized limb, distal and proximal of the osteotomy. Box and Whisker plot showing median, 25%-, and 75%-percentiles, minimum, and maximum. The control limb had significantly (P b 0.001) smaller density of active osteons at all time points.
542
P. Augat, L. Claes / Bone 43 (2008) 539–543
Fig. 5. The reduction in app.BMD adjacent to the osteotomy is partly explained by an increase in density of active osteons at week 9 after surgery (R2 = 0.5, P b 0.01).
Density of active osteons in cortical bone of the control limb on average remained below 0.08 mm− 2. There was neither any local variation between distal and proximal locations (P = 0.8) nor any change over the course of the observation period in the control limb (Fig. 4, P = 0.9). In the osteotomized limb, the median density of active osteons was increased at least 6 fold (Fig. 4, P b 0.001) compared to the control limb. Although the density of active osteons was on average higher at the distal location compared to the proximal location, this difference was not statistically significant (P N 0.2). Therefore, the values for the proximal and distal locations were averaged and demonstrated a significant increase between week 4 and week 6 after osteotomy, but did not change significantly thereafter (Table 1). The reduction in app.BMD adjacent to the osteotomy after 9 weeks of bone healing was associated with an increased remodeling activity in the diaphysis. The app.BMD in the diaphysis showed a significant correlation with the density of active osteons at week 9 (R = −0.71, P b 0.01, Fig. 5) and week 6 (R = −0.42, P b 0.05), but not at week 4 (R = −0.35, P N 0.2). Discussion In our experimental study, we confirmed the clinically observed phenomenon of reduction in bone mineral at the ipsilateral limb following a fracture. By histomorphometric analysis of the cortical bone around the site of fracture, we were able to identify the source of this bone loss. The large number of active osteons in the cortical bone revealed a marked increase in bone remodeling activity following the fracture. This was accountable for the observed post-traumatic loss in bone mineral. The results from our study demonstrate a profound increase in remodeling activity in the cortical bone adjacent to the site of an osteotomy. As early as 4 weeks after osteotomy, we were able to observe a significant increase in bone formation in osteonal remodeling sites. Previous studies on the response of bone markers in humans indicated an increased osteoclastic activity after fracture from as early as 7 days after fracture [14]. Increased osteoclastic activity early after fracture can be found at fracture surfaces, where osteoclasts resorb necrotic or dead bone [15]. During the course of healing, the osteoclastic activity continuously increases within the periosteal and endosteal callus tissue [16]. Our study suggests that also the surrounding cortical bone adjacent to a diaphyseal osteotomy has a markedly increased activity of bone resorbing osteoclasts during the first weeks of healing. At the 9 weeks time point, which fell within the remodeling phase of bone healing, active osteons were still abundant in the adjacent cortex indicating an ongoing remodeling activity. Longer lasting studies in humans demonstrated that osteoclastic activity remained significantly enhanced after bone union, most likely indicating an ongoing remodeling activity in the fracture callus [14]. The large number of fluorescent and Paragon stained osteons in the cortical bone adjacent to an osteotomy indicates an increased bone
formation activity by osteoblasts. Our histologic analysis during 9 weeks of bone healing demonstrated ongoing new bone formation at remodeling sites within the cortex. Previous studies observed that osteoblast activity after fracture is temporarily ceased and recovers within the first week after fracture [17]. Thereafter, fractures with normal healing show increased bone formation throughout the healing period until bony union [14]. Previous studies that have determined serum levels of bone markers were not designed to determine the location of bone turnover activities. Although the fracture site itself is an obvious source for the observed modifications in bone formation and bone resorption markers, the findings of our study suggest that also the bone adjacent to the fracture significantly contributes to enhanced bone turnover during bone repair. One possible explanation of increased bone turnover in the adjacent cortex after a fracture might be the immobilization of the affected limb due to reduced weight bearing activities and the osteosynthetic fixation. Following an osteotomy, sheep in this experimental environment typically show a reduction in load bearing activity for the first 4 weeks after fracture [18,19]. The external fixator further shielded the region between the fixator pins from loading. The samples for bone density measurement and histomorphometric analysis were obtained from within a region covered by the external fixator frame. As a second explanation, changes in the blood supply to the bone have been suggested as a cause for changes in bone turnover [20]. The lack of blood supply after fracture is likely to vary among different locations around the fracture [21]. Distal locations may be more affected than proximal locations because the blood transfer through the site of fracture might have been affected by the injury. The third explanation of increased bone turnover is the recruitment of bone mineral for the augmentation of local bone repair. The changes in bone formation and resorption markers are likely to cause a systemic response, which would trigger modifications in bone turnover throughout all skeletal sites [22], but of course not necessarily distributed homogenously throughout the system. As preoperative measurements were not available, we cannot exclude that the activity at the control site was enhanced from baseline activity in the sheep. Furthermore, the activity at the osteotomized site could also be partly related to the local need of mineral for the bone repair process for which the adjacent bone matrix would be an obvious source. The persistence of the increased remodeling activity suggests that the load shielding effect by the osteosynthesis material or the recruitment of mineral are the causes for the remodeling activity and the subsequent bone loss. Reduced mobility of the animals or changes in blood supply to the fracture are less likely causes of the bone loss. Mobility reached normal levels after around 4 weeks and also blood supply is known to regenerate within a few days after fracture [21,23]. On the other hand, load shielding by the external fixator and the recruitment of mineral for local bone repair may have continued until the time of bony consolidation of the osteotomy. The increased remodeling activity observed at the end of the observation period was associated with lower values of apparent bone mineral density (Fig. 5). The remodeling activity was measured by counting newly formed osteons showing signs of recent osteoid formation either by fluorescence labeling or by Paragon staining. The deficient calcification in these recently deposited osteoid was accountable for the lower value of bone mineral density. The osteon constitutes the basic multicellular unit (BMU) which excavates and refills tunnels through cortical bone [24], and which originates from local osteoclastic activity. If the recruitment frequency and the resorption activity of osteoclasts could be reduced there would be potential for limiting the observed remodeling activity. Because bisphosphonates are potent inhibitors of bone resorption they may also be effective in preventing post fracture loss of bone mineral adjacent to the site of fracture. It therefore appears reasonable to treat patients with osteoporotic fractures not only to alleviate their
P. Augat, L. Claes / Bone 43 (2008) 539–543
underlying osteoporosis but also to prevent their accelerated loss of bone mineral around the site of the fracture [25]. Concerns regarding a possible interaction between bisphosphonate treatment and fracture repair are most likely unsubstantiated. Based on the current literature a recent fracture should not preclude the initiation of therapy, because bisphosphonates have not been shown to interfere with overall fracture strength [26]. Our experimental study also has some limitations. Measurement of app.BMD as well as the histological analysis was performed at a location that was covered by the osteosynthesis material, namely the external fixator. Therefore, bone density and density of active osteons was likely affected by load shielding of the osteosynthesis material and could not be separated from post surgery reduction in weight bearing activities. Moreover, due to the mechanical interaction of the fixator pins with the bone, some periosteal callus reaction was observed at the pin entrance into the bone. This periosteal bone formation was included in the app.BMD analysis at the osteotomized metatarsus. Because this callus tissue was not yet fully mineralized, the measurement of app.BMD might have underestimated the real cortical app.BMD. The fluorescence images of new bone formation in the cortex were not accessible for histomorphometric analysis. Because of incomplete staining of osteons, it was not possible to obtain quantitative measures like osteon size or area covered by osteons. Furthermore, the methods of staining new bone formation differed between weeks 4 and 6, and week 9. While new bone formation in weeks 4 and 6 was labeled by fluorescent stains supplied during the experiment, the staining of newly formed bone at week 9 was performed ex situ on the histologic slices. This may have introduced a systematic error in the counting of active remodeling osteons. Repeated counting of osteons was minimized by switching back and forth between standard and fluorescence illuminations. A further limitation of our study is that bone density measurements were only performed post-mortem and we were not able to provide pre-operative values of bone mineral density at the experimental or the control limb. Thus we cannot exclude whether the control limb experienced any changes in bone mineral density due to the modified loading situation following the surgical intervention. The findings from this study have several implications. (1) Diaphyseal fractures considerably affect bone mineral in the ipsilateral limb. This results in deterioration of the bone mechanical properties and therefore contributes to increased risk of refracture of the affected bone [27,28]. (2) Furthermore, the mechanical quality of the cortex adjacent to a fracture is essential for providing the holding power of osteosynthesis implants such as screws, rods or bone plates [29–31]. Reduced holding power of an implant is likely to impair implant fixation and cause migration, cutout, deformity or shortening [29]. (3) Bone mass measurements after a fracture at the ipsilateral limb are incorrect markers for bone mass before the fracture. In the past, retrospective analyses of bone mineral after the occurrence of a fracture have been used to identify risk factors for fractures. Our study, together with several other studies on changes in bone density after fracture, reveals that bone mass measurement after a fracture considerably underestimates the pre-fracture bone mass as early as 4 weeks and as long as several years after fracture. Special care has therefore to be taken if post-fracture bone mass measurements are interpreted with respect to their ability for fracture prediction [32]. We conclude from these findings that posttraumatic loss of bone mineral adjacent to a fracture is related to an elevated number of active osteons, indicating a significant increase in bone remodeling activity. Load shielding by the osteosynthesis material and local recruitment of bone mineral are likely causes for this increased remodeling. If this bone loss could be moderated by anti-resorptive medication, complications like refracture or failure of implant fixation could potentially be reduced.
543
References [1] Findlay SC, Eastell R, Ingle BM. Measurement of bone adjacent to tibial shaft fracture. Osteoporos Int 2002;13:980–9. [2] Veitch SW, Findlay SC, Hamer AJ, Blumsohn A, Eastell R, Ingle BM. Changes in bone mass and bone turnover following tibial shaft fracture. Osteoporos Int 2006;17:364–72. [3] van der Poest CE, van der WH, Patka P, Roos JC, Lips P. Long-term consequences of fracture of the lower leg: cross-sectional study and long-term longitudinal follow-up of bone mineral density in the hip after fracture of lower leg. Bone 1999;24:131–4. [4] Van der Wiel HE, Lips P, Nauta J, Patka P, Haarman HJ, Teule GJ. Loss of bone in the proximal part of the femur following unstable fractures of the leg. J Bone Joint Surg Am 1994;76:230–6. [5] Karlsson MK, Josefsson PO, Nordkvist A, Akesson K, Seeman E, Obrant KJ. Bone loss following tibial osteotomy: a model for evaluating post-traumatic osteopenia. Osteoporos Int 2000;11:261–4. [6] Emami A, Larsson S, Hellquist E, Mallmin H. Limited bone loss in the hip and heel after reamed intramedullary fixation and early weight-bearing of tibial fractures. J Orthop Trauma 2001;15:560–5. [7] Eyres KS, Kanis JA. Bone loss after tibial fracture. Evaluated by dual-energy X-ray absorptiometry. J Bone Joint Surg Br 1995;77:473–8. [8] Karlsson MK, Nilsson BE, Obrant KJ. Bone mineral loss after lower extremity trauma. 62 cases followed for 15–38 years. Acta Orthop Scand 1993;64:362–4. [9] Westlin NE. Loss of bone mineral after Colles' fracture. Clin Orthop Relat Res 1974:194–9. [10] Hernandez CJ, Keaveny TM. A biomechanical perspective on bone quality. Bone 2006;39:1173–81. [11] Augat P, Margevicius K, Simon J, Wolf S, Suger G, Claes L. Local tissue properties in bone healing: influence of size and stability of the osteotomy gap. J Orthop Res 1998;16:475–81. [12] Claes L, Augat P, Suger G, Wilke HJ. Influence of size and stability of the osteotomy gap on the success of fracture healing. J Orthop Res 1997;15:577–84. [13] Hedgecock NL, Hadi T, Chen AA, Curtiss SB, Martin RB, Hazelwood SJ. Quantitative regional associations between remodeling, modeling, and osteocyte apoptosis and density in rabbit tibial midshafts. Bone 2007;40:627–37. [14] Stoffel K, Engler H, Kuster M, Riesen W. Changes in biochemical markers after lower limb fractures. Clin Chem 2007;53:131–4. [15] Cruess RL, Dumont J. Fracture healing. Can J Surg 1975;18:403–13. [16] Schell H, Lienau J, Epari DR, Seebeck P, Exner C, Muchow S, et al. Osteoclastic activity begins early and increases over the course of bone healing. Bone 2006;38:547–54. [17] Bowles SA, Kurdy N, Davis AM, France MW, Marsh DR. Serum osteocalcin, total and bone-specific alkaline phosphatase following isolated tibial shaft fracture. Ann Clin Biochem 1996;33(Pt 3):196–200. [18] Augat P, Merk J, Ignatius A, Margevicius KJ, Bauer G, Rosenbaum D, et al. Early, full weightbearing with flexible fixation delays the healing of experimental fractures. Clinical Orthopaedics and Related Research 1996;328:194–202. [19] Klein P, Schell H, Streitparth F, Heller M, Kassi JP, Kandziora F, et al. The initial phase of fracture healing is specifically sensitive to mechanical conditions. J Orthop Res 2003;21:662–9. [20] Lucas TS, Bab IA, Lian JB, Stein GS, Jazrawi L, Majeska RJ, et al. Stimulation of systemic bone formation induced by experimental blood loss. Clin Orthop Relat Res 1997:267–75. [21] M. Melnyk, T. Henke, L. Claes, P. Augat, Revascularisation during fracture healing with soft tissue injury. Arch Orthop Trauma Surg 2007. [22] Kaspar D, Neidlinger-Wilke C, Holbein O, Claes L, Ignatius A. Mitogens are increased in the systemic circulation during bone callus healing. J Orthop Res 2003;21:320–5. [23] Bumann M, Henke T, Gerngross H, Claes L, Augat P. Influence of haemorrhagic shock on fracture healing. Langenbecks Arch Surg 2003;388:331–8. [24] Parfitt AM. Skeletal heterogeneity and the purpose of bone remodeling. In: Marcus R, Feldman D, Kelsey J, editors. Osteoporosis. San Diego: Academic Press; 1996. p. 315–29. [25] van der Poest CE, van EM, Ader H, Roos JC, Patka P, Lips P. Alendronate in the prevention of bone loss after a fracture of the lower leg. J Bone Miner Res 2002;17:2247–55. [26] Seebach C, Kurth A, Marzi I. The influence of bisphosphonates on fracture healing. Orthopade 2007;36:136–40. [27] J. Goldhahn, N. Suhm, S. Goldhahn, M. Blauth, B. Hanson, Influence of osteoporosis on fracture fixation — a systematic literature review. Osteoporos Int. 2007. [28] Utrilla AL, Reig JS, Munoz FM, Tufanisco CB. Trochanteric gamma nail and compression hip screw for trochanteric fractures: a randomized, prospective, comparative study in 210 elderly patients with a new design of the gamma nail. J Orthop Trauma 2005;19:229–33. [29] Goldhahn J, Seebeck J, Frei R, Frenz B, Antoniadis I, Schneider E. New implant designs for fracture fixation in osteoporotic bone. Osteoporos Int 2005;16(Suppl 2):S112–9. [30] Koistinen A, Santavirta SS, Kroger H, Lappalainen R. Effect of bone mineral density and amorphous diamond coatings on insertion torque of bone screws. Biomaterials 2005;26:5687–94. [31] Seebeck J, Goldhahn J, Stadele H, Messmer P, Morlock MM, Schneider E. Effect of cortical thickness and cancellous bone density on the holding strength of internal fixator screws. J Orthop Res 2004;22:1237–42. [32] Jones G, Boon P. Which bone mass measures discriminate adolescents who have fractured from those who have not? Osteoporos Int 2008;19:251–5.
Contents lists available at ScienceDirect
Bone j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b o n e
Increased cortical remodeling after osteotomy causes posttraumatic osteopenia Peter Augat a,b,⁎, Lutz Claes c a b c
Biomechanics Research Laboratory, Paracelsus Medical University, 5020 Salzburg, Austria Biomechanics Research Laboratory, Trauma Center Murnau, 82418 Murnau, Germany Institute of Orthopaedic Research and Biomechanics, University of Ulm, 89081 Ulm, Germany
a r t i c l e
i n f o
Article history: Received 4 February 2008 Revised 9 May 2008 Accepted 13 May 2008 Available online 30 June 2008 Edited by: H. Genant Keywords: Fracture healing Osteopenia Bone density Bone histomorphometry Remodeling
a b s t r a c t Following a fracture, substantial bone mineral loss can occur at the affected limb. The aim of this study was to analyze the changes in cortical bone around the site of a fracture. We analyzed bone mineral density by quantitative computed tomography and quantified changes in cortical remodeling by histomorphometry adjacent to an experimental osteotomy in sheep metatarsals. In the cortical bone around the osteotomy, we found a statistically significant 16% reduction in app.BMD within 9 weeks following surgery. This reduction was explained (R = −0.71, P b 0.01) by a more than 6 fold increase in bone remodeling activity within cortical bone at the affected limb. The remodeling activity significantly increased between surgery and week 6, but remained unchanged between week 6 and week 9. We conclude from these findings that posttraumatic bone mineral loss adjacent to a fracture is related to an elevated number of active osteons, indicating a significant increase in bone remodeling activity. Load shielding by the osteosynthesis material and local recruitment of bone mineral are likely causes for this increased remodeling. This post-traumatic bone loss is likely to contribute significantly to frequently observed healing complications like refracture, failure of implant fixation, implant loosening, or cut out. © 2008 Elsevier Inc. All rights reserved.
Introduction Fracture healing occurs through formation of periosteal callus tissue (secondary healing) or increased bone remodeling (primary healing or metaphyseal healing) at the site of the fracture. During the healing process, the amount of bone in the fractured area initially increases. In diaphyseal healing, this is typically perceptible by the formation of hard callus visible on X-rays. However, there is also considerable regional bone loss at sites adjacent to the fracture. Previous studies demonstrated that bone loss around a fracture can amount to almost 30% loss in the same bone [1,2] and more than 10% in neighboring bones [3]. This loss can be observed as early as 6 weeks after fracture [4,5]. Although some recovery is observed after completion of healing [6], the bone mass deficit may persist for many years [7,8]. Although fracture related bone loss occurs to a large extent at sites rich in trabecular bone, there have also been reports of significant bone loss at diaphyseal locations after fracture [2,5,9]. The changes observed at diaphyseal locations were mostly observed with dual X-ray absorptiometry (DXA) which generates a 2D projectional measurement of bone mineral density [2,5,9]. As this measurement is affected by the projected area, any periosteal callus formation that increases the projected area would incorrectly detect a decrease in
⁎ Corresponding author. Paracelsus Medical University Salzburg, and Director Biomechanics Laboratory, Trauma Center Murnau, Prof. Kuentscher Str. 8, 82418 Murnau, Germany. Fax: +49 8841 484573. E-mail address: [email protected] (P. Augat). 8756-3282/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2008.05.017
bone mineral. Measurements with computed tomography, which are not susceptible to this error, were only performed at trabecular rich sites with rather small cortices [2,5,9]. Therefore these previous studies did not entirely reveal to which extent diaphyseal locations are affected by post traumatic bone loss. Because loss of bone mineral is associated with a reduction in mechanical competence [10], the consequence of fracture associated bone loss is weakening of the fractured bone as well as bones adjacent to the fracture. The reduction in bone strength increases the risk of fracture and may also contribute to failure in the maintenance of fracture reduction after osteosynthetic fixation. Therefore, fracture associated bone loss is likely to play a significant role in the occurrence of refractures and loosening of osteosynthetic implants. Although the occurrence of fracture associated bone loss has been well described in the literature, there is limited research on the local distribution, the underlying mechanism and the reason for this phenomenon. It is not known whether the fracture associated bone loss in the diaphysis occurs within the cortex, at the periosteal or at the endosteal surface. Furthermore, there is no knowledge about the specific distribution of bone loss around the fracture, particularly whether bone loss differs proximally or distally from the fracture site. The aim of this study was to quantitatively analyze how a fracture affects the cortical bone around the site of the fracture. We hypothesized that the post-fracture loss of bone mineral is caused by increased cortical remodeling and that variations in cortical remodeling can be observed between proximal and distal locations. Therefore, we assessed cortical remodeling in the diaphysis of ovine
540
P. Augat, L. Claes / Bone 43 (2008) 539–543
Fig. 1. Measurement of bone density with pQCT. Scan locations were identified on an anterior–posterior scout view (right panel). On the left non-osteotomized limb, three equidistant scans were placed in the center of the metatarsus (left panel, top). On the osteotomized limb, three scans were performed: at the center of the osteotomy (left panel, middle) and 15 mm proximal and distal (left panel, bottom) of the osteotomy, respectively.
long bones adjacent to a healing osteotomy by histomorphometric measurement of active osteons. Materials and methods Cortical bone remodeling was studied in the osteotomized and contralateral hind limb of sheep in the course of a bone healing process. Twelve skeletally mature Merino sheep (female, 2–3 years of age, average weight 73 ± 9 kg) received a transverse osteotomy at the mid diaphysis of the right metatarsus under general anesthesia. The osteotomy gap was adjusted to 2 mm and the external ring fixator allowed for interfragmentary movements between 0.2 mm and 1.0 mm. The surgical procedure and the external fixator procedure are described in greater detail elsewhere [11,12]. Following the operation, the animals received metamizol (20 mg per kg bodyweight; Novalgin®, Hoechst, Frankfurt, Germany) for pain relief for a period of
5 days and were allowed full freedom of movement. To monitor remodeling activity within cortical bone, the fluorochrome markers, Calcein (0.33 ml/kg body weight; Calceingreen; Synopharm, Barsbuettel, Germany) and Tetracycline (1.0 ml/kg body weight; Tetracycline-hydrochloride; Caesar and Loretz GmbH, Hilden, Germany), were administered intravenously at 4 and 6 weeks after surgery, respectively. After 9 weeks the animals were killed, the metatarsals were harvested, and the fixator was removed. The protocol was performed in accordance with the Guide for Care and Use of Laboratory Animals and was approved by the animal care and use committee (approval No. 543, Regierungspraesidium Tuebingen). After explantation both metatarsals were analyzed for apparent bone mineral density using a peripheral quantitative computed tomography scanner (pQCT, XCT 960, Stratec, Pforzheim, Germany). For the control limb, three scans 1 cm apart from each other were performed at the center of the metatarsus. For the osteotomized limb,
Fig. 2. Histologic picture (25-fold magnification, bar indicates 400 μm) of osteons under fluorescent light from the untreated control limb (left) and the limb that received an osteotomy (right). Green and yellow labels indicate bone formed at week 4 and week 6, respectively.
P. Augat, L. Claes / Bone 43 (2008) 539–543
541
Table 1 Histomorphometric analysis of density of active osteons in cortical bone Active osteon density in # / mm2 Controla Osteotomy- week 4 Osteotomy week 6
0.06 ± 0.08 0.39 ± 0.33 0.96 ± 0.51
Osteotomy week 9
0.69 ± 0.30
P b 0.01 compared to control P b 0.001 compared to control P b 0.01 compared to week 4 P b 0.001 compared to control P b 0.01 compared to week 4
Average of distal and proximal location (mean ± standard deviation) from measurement in n = 12 sheep. a Density of active osteons in the control group depended neither on location nor time after osteotomy and was therefore averaged from proximal and distal locations and for all three time points.
Fig. 3. app.BMD 9 weeks after surgery of cortical bone proximal and distal from the osteotomy was significantly (P b 0.001) smaller compared to app.BMD in the nonosteotomized control limb. Box and Whisker plot showing median, 25%-, and 75%percentiles, minimum, and maximum of app.BMD measurements in 12 sheep.
the level of the osteotomy was identified on a planar scout view and three pQCT scans were performed (Fig. 1). The first scan was placed at the center of the osteotomy gap, including the periosteal callus formation, followed by scans at 1.5 cm distally and 1.5 cm proximally from the center of the osteotomy gap, respectively. The distal and proximal scans were therefore located between the positions of the two outermost fixator pins, respectively. The axial scans of the diaphysis were analyzed for apparent bone mineral density (app. BMD) of the cortical compartment and total bone mineral content (BMC). Both metatarsals were then processed for undecalcified bone histology and embedded in polymethylmethacrylate. Axial crosssectional slices of 70 μm thickness were obtained 3 cm distally and proximally from the center of the osteotomy. From the nonosteotomized control metatarsus, two slices were obtained 2 cm proximally and distally from the center of the metatarsus. The slices were cut with a histologic precision saw (Exact System, Nordenstadt, Germany) under constant cooling and were smooth grinded. The sections were surface stained with Paragon (Paragon C and C; New York, NY, USA), which labeled newly formed bone and active osteons in blue. The histological slices were examined with conventional light and fluorescence microscopy (Axiophot; Zeiss, Oberkochen, Germany) at 25-fold magnification. Active osteons were counted within the cortical bone compartment, which appear green or yellow under fluorescent light or blue under conventional light (Fig. 2). The
histological slices were then digitized at 12.5-fold magnification and the images were imported into an image analysis system (analySIS; Soft Imaging Systems, Muenster, Germany). By manual segmentation, the periosteal and endosteal bone surfaces were identified and the cortical bone area was measured. Density of active osteons was used as a measure of remodeling activity and was calculated by dividing the number of osteons by the area of the cortical bone cross-section [13]. Repeated analysis of five histological slices by two observers revealed an inter- and intraobserver reproducibility of 9% and 7%, respectively. Shapiro–Wilk test was used to control for normal distribution of the data. Paired sample Student's t-tests were employed to test for significant differences between osteotomized and control limbs, and differences between proximal and distal locations. The density of active osteons at different time points and at different locations was compared using a repeated measures analysis of variance and a Tukey post hoc test. Resulting P-values were adjusted for multiple comparisons using the Bonferroni adjustment. The SPSS statistical software package (SPSS 14.0, Chicago, Illinois) and Excel (Microsoft Corp.) were used for data analysis. Results All animals had uneventful healing of the osteotomy with two animals showing minor signs of pin infection. The app.BMD did not differ between distal and proximal locations (P N 0.2) and was on average 16% lower at the osteotomized limb (859 ± 50 mg/cm3; P b 0.001) compared to the app.BMD at the non-osteotomized control limb (Fig. 3, 1023 ± 58 mg/cm3). The central 1 mm slice at the site of the former osteotomy, which included the periosteal callus apposition in the osteotomized bone, contained 41% (P b 0.02) more mineralized tissue than the central slice in the control limb.
Fig. 4. Density of active osteons 4, 6, and 9 weeks after osteotomy in the control limb and the osteotomized limb, distal and proximal of the osteotomy. Box and Whisker plot showing median, 25%-, and 75%-percentiles, minimum, and maximum. The control limb had significantly (P b 0.001) smaller density of active osteons at all time points.
542
P. Augat, L. Claes / Bone 43 (2008) 539–543
Fig. 5. The reduction in app.BMD adjacent to the osteotomy is partly explained by an increase in density of active osteons at week 9 after surgery (R2 = 0.5, P b 0.01).
Density of active osteons in cortical bone of the control limb on average remained below 0.08 mm− 2. There was neither any local variation between distal and proximal locations (P = 0.8) nor any change over the course of the observation period in the control limb (Fig. 4, P = 0.9). In the osteotomized limb, the median density of active osteons was increased at least 6 fold (Fig. 4, P b 0.001) compared to the control limb. Although the density of active osteons was on average higher at the distal location compared to the proximal location, this difference was not statistically significant (P N 0.2). Therefore, the values for the proximal and distal locations were averaged and demonstrated a significant increase between week 4 and week 6 after osteotomy, but did not change significantly thereafter (Table 1). The reduction in app.BMD adjacent to the osteotomy after 9 weeks of bone healing was associated with an increased remodeling activity in the diaphysis. The app.BMD in the diaphysis showed a significant correlation with the density of active osteons at week 9 (R = −0.71, P b 0.01, Fig. 5) and week 6 (R = −0.42, P b 0.05), but not at week 4 (R = −0.35, P N 0.2). Discussion In our experimental study, we confirmed the clinically observed phenomenon of reduction in bone mineral at the ipsilateral limb following a fracture. By histomorphometric analysis of the cortical bone around the site of fracture, we were able to identify the source of this bone loss. The large number of active osteons in the cortical bone revealed a marked increase in bone remodeling activity following the fracture. This was accountable for the observed post-traumatic loss in bone mineral. The results from our study demonstrate a profound increase in remodeling activity in the cortical bone adjacent to the site of an osteotomy. As early as 4 weeks after osteotomy, we were able to observe a significant increase in bone formation in osteonal remodeling sites. Previous studies on the response of bone markers in humans indicated an increased osteoclastic activity after fracture from as early as 7 days after fracture [14]. Increased osteoclastic activity early after fracture can be found at fracture surfaces, where osteoclasts resorb necrotic or dead bone [15]. During the course of healing, the osteoclastic activity continuously increases within the periosteal and endosteal callus tissue [16]. Our study suggests that also the surrounding cortical bone adjacent to a diaphyseal osteotomy has a markedly increased activity of bone resorbing osteoclasts during the first weeks of healing. At the 9 weeks time point, which fell within the remodeling phase of bone healing, active osteons were still abundant in the adjacent cortex indicating an ongoing remodeling activity. Longer lasting studies in humans demonstrated that osteoclastic activity remained significantly enhanced after bone union, most likely indicating an ongoing remodeling activity in the fracture callus [14]. The large number of fluorescent and Paragon stained osteons in the cortical bone adjacent to an osteotomy indicates an increased bone
formation activity by osteoblasts. Our histologic analysis during 9 weeks of bone healing demonstrated ongoing new bone formation at remodeling sites within the cortex. Previous studies observed that osteoblast activity after fracture is temporarily ceased and recovers within the first week after fracture [17]. Thereafter, fractures with normal healing show increased bone formation throughout the healing period until bony union [14]. Previous studies that have determined serum levels of bone markers were not designed to determine the location of bone turnover activities. Although the fracture site itself is an obvious source for the observed modifications in bone formation and bone resorption markers, the findings of our study suggest that also the bone adjacent to the fracture significantly contributes to enhanced bone turnover during bone repair. One possible explanation of increased bone turnover in the adjacent cortex after a fracture might be the immobilization of the affected limb due to reduced weight bearing activities and the osteosynthetic fixation. Following an osteotomy, sheep in this experimental environment typically show a reduction in load bearing activity for the first 4 weeks after fracture [18,19]. The external fixator further shielded the region between the fixator pins from loading. The samples for bone density measurement and histomorphometric analysis were obtained from within a region covered by the external fixator frame. As a second explanation, changes in the blood supply to the bone have been suggested as a cause for changes in bone turnover [20]. The lack of blood supply after fracture is likely to vary among different locations around the fracture [21]. Distal locations may be more affected than proximal locations because the blood transfer through the site of fracture might have been affected by the injury. The third explanation of increased bone turnover is the recruitment of bone mineral for the augmentation of local bone repair. The changes in bone formation and resorption markers are likely to cause a systemic response, which would trigger modifications in bone turnover throughout all skeletal sites [22], but of course not necessarily distributed homogenously throughout the system. As preoperative measurements were not available, we cannot exclude that the activity at the control site was enhanced from baseline activity in the sheep. Furthermore, the activity at the osteotomized site could also be partly related to the local need of mineral for the bone repair process for which the adjacent bone matrix would be an obvious source. The persistence of the increased remodeling activity suggests that the load shielding effect by the osteosynthesis material or the recruitment of mineral are the causes for the remodeling activity and the subsequent bone loss. Reduced mobility of the animals or changes in blood supply to the fracture are less likely causes of the bone loss. Mobility reached normal levels after around 4 weeks and also blood supply is known to regenerate within a few days after fracture [21,23]. On the other hand, load shielding by the external fixator and the recruitment of mineral for local bone repair may have continued until the time of bony consolidation of the osteotomy. The increased remodeling activity observed at the end of the observation period was associated with lower values of apparent bone mineral density (Fig. 5). The remodeling activity was measured by counting newly formed osteons showing signs of recent osteoid formation either by fluorescence labeling or by Paragon staining. The deficient calcification in these recently deposited osteoid was accountable for the lower value of bone mineral density. The osteon constitutes the basic multicellular unit (BMU) which excavates and refills tunnels through cortical bone [24], and which originates from local osteoclastic activity. If the recruitment frequency and the resorption activity of osteoclasts could be reduced there would be potential for limiting the observed remodeling activity. Because bisphosphonates are potent inhibitors of bone resorption they may also be effective in preventing post fracture loss of bone mineral adjacent to the site of fracture. It therefore appears reasonable to treat patients with osteoporotic fractures not only to alleviate their
P. Augat, L. Claes / Bone 43 (2008) 539–543
underlying osteoporosis but also to prevent their accelerated loss of bone mineral around the site of the fracture [25]. Concerns regarding a possible interaction between bisphosphonate treatment and fracture repair are most likely unsubstantiated. Based on the current literature a recent fracture should not preclude the initiation of therapy, because bisphosphonates have not been shown to interfere with overall fracture strength [26]. Our experimental study also has some limitations. Measurement of app.BMD as well as the histological analysis was performed at a location that was covered by the osteosynthesis material, namely the external fixator. Therefore, bone density and density of active osteons was likely affected by load shielding of the osteosynthesis material and could not be separated from post surgery reduction in weight bearing activities. Moreover, due to the mechanical interaction of the fixator pins with the bone, some periosteal callus reaction was observed at the pin entrance into the bone. This periosteal bone formation was included in the app.BMD analysis at the osteotomized metatarsus. Because this callus tissue was not yet fully mineralized, the measurement of app.BMD might have underestimated the real cortical app.BMD. The fluorescence images of new bone formation in the cortex were not accessible for histomorphometric analysis. Because of incomplete staining of osteons, it was not possible to obtain quantitative measures like osteon size or area covered by osteons. Furthermore, the methods of staining new bone formation differed between weeks 4 and 6, and week 9. While new bone formation in weeks 4 and 6 was labeled by fluorescent stains supplied during the experiment, the staining of newly formed bone at week 9 was performed ex situ on the histologic slices. This may have introduced a systematic error in the counting of active remodeling osteons. Repeated counting of osteons was minimized by switching back and forth between standard and fluorescence illuminations. A further limitation of our study is that bone density measurements were only performed post-mortem and we were not able to provide pre-operative values of bone mineral density at the experimental or the control limb. Thus we cannot exclude whether the control limb experienced any changes in bone mineral density due to the modified loading situation following the surgical intervention. The findings from this study have several implications. (1) Diaphyseal fractures considerably affect bone mineral in the ipsilateral limb. This results in deterioration of the bone mechanical properties and therefore contributes to increased risk of refracture of the affected bone [27,28]. (2) Furthermore, the mechanical quality of the cortex adjacent to a fracture is essential for providing the holding power of osteosynthesis implants such as screws, rods or bone plates [29–31]. Reduced holding power of an implant is likely to impair implant fixation and cause migration, cutout, deformity or shortening [29]. (3) Bone mass measurements after a fracture at the ipsilateral limb are incorrect markers for bone mass before the fracture. In the past, retrospective analyses of bone mineral after the occurrence of a fracture have been used to identify risk factors for fractures. Our study, together with several other studies on changes in bone density after fracture, reveals that bone mass measurement after a fracture considerably underestimates the pre-fracture bone mass as early as 4 weeks and as long as several years after fracture. Special care has therefore to be taken if post-fracture bone mass measurements are interpreted with respect to their ability for fracture prediction [32]. We conclude from these findings that posttraumatic loss of bone mineral adjacent to a fracture is related to an elevated number of active osteons, indicating a significant increase in bone remodeling activity. Load shielding by the osteosynthesis material and local recruitment of bone mineral are likely causes for this increased remodeling. If this bone loss could be moderated by anti-resorptive medication, complications like refracture or failure of implant fixation could potentially be reduced.
543
References [1] Findlay SC, Eastell R, Ingle BM. Measurement of bone adjacent to tibial shaft fracture. Osteoporos Int 2002;13:980–9. [2] Veitch SW, Findlay SC, Hamer AJ, Blumsohn A, Eastell R, Ingle BM. Changes in bone mass and bone turnover following tibial shaft fracture. Osteoporos Int 2006;17:364–72. [3] van der Poest CE, van der WH, Patka P, Roos JC, Lips P. Long-term consequences of fracture of the lower leg: cross-sectional study and long-term longitudinal follow-up of bone mineral density in the hip after fracture of lower leg. Bone 1999;24:131–4. [4] Van der Wiel HE, Lips P, Nauta J, Patka P, Haarman HJ, Teule GJ. Loss of bone in the proximal part of the femur following unstable fractures of the leg. J Bone Joint Surg Am 1994;76:230–6. [5] Karlsson MK, Josefsson PO, Nordkvist A, Akesson K, Seeman E, Obrant KJ. Bone loss following tibial osteotomy: a model for evaluating post-traumatic osteopenia. Osteoporos Int 2000;11:261–4. [6] Emami A, Larsson S, Hellquist E, Mallmin H. Limited bone loss in the hip and heel after reamed intramedullary fixation and early weight-bearing of tibial fractures. J Orthop Trauma 2001;15:560–5. [7] Eyres KS, Kanis JA. Bone loss after tibial fracture. Evaluated by dual-energy X-ray absorptiometry. J Bone Joint Surg Br 1995;77:473–8. [8] Karlsson MK, Nilsson BE, Obrant KJ. Bone mineral loss after lower extremity trauma. 62 cases followed for 15–38 years. Acta Orthop Scand 1993;64:362–4. [9] Westlin NE. Loss of bone mineral after Colles' fracture. Clin Orthop Relat Res 1974:194–9. [10] Hernandez CJ, Keaveny TM. A biomechanical perspective on bone quality. Bone 2006;39:1173–81. [11] Augat P, Margevicius K, Simon J, Wolf S, Suger G, Claes L. Local tissue properties in bone healing: influence of size and stability of the osteotomy gap. J Orthop Res 1998;16:475–81. [12] Claes L, Augat P, Suger G, Wilke HJ. Influence of size and stability of the osteotomy gap on the success of fracture healing. J Orthop Res 1997;15:577–84. [13] Hedgecock NL, Hadi T, Chen AA, Curtiss SB, Martin RB, Hazelwood SJ. Quantitative regional associations between remodeling, modeling, and osteocyte apoptosis and density in rabbit tibial midshafts. Bone 2007;40:627–37. [14] Stoffel K, Engler H, Kuster M, Riesen W. Changes in biochemical markers after lower limb fractures. Clin Chem 2007;53:131–4. [15] Cruess RL, Dumont J. Fracture healing. Can J Surg 1975;18:403–13. [16] Schell H, Lienau J, Epari DR, Seebeck P, Exner C, Muchow S, et al. Osteoclastic activity begins early and increases over the course of bone healing. Bone 2006;38:547–54. [17] Bowles SA, Kurdy N, Davis AM, France MW, Marsh DR. Serum osteocalcin, total and bone-specific alkaline phosphatase following isolated tibial shaft fracture. Ann Clin Biochem 1996;33(Pt 3):196–200. [18] Augat P, Merk J, Ignatius A, Margevicius KJ, Bauer G, Rosenbaum D, et al. Early, full weightbearing with flexible fixation delays the healing of experimental fractures. Clinical Orthopaedics and Related Research 1996;328:194–202. [19] Klein P, Schell H, Streitparth F, Heller M, Kassi JP, Kandziora F, et al. The initial phase of fracture healing is specifically sensitive to mechanical conditions. J Orthop Res 2003;21:662–9. [20] Lucas TS, Bab IA, Lian JB, Stein GS, Jazrawi L, Majeska RJ, et al. Stimulation of systemic bone formation induced by experimental blood loss. Clin Orthop Relat Res 1997:267–75. [21] M. Melnyk, T. Henke, L. Claes, P. Augat, Revascularisation during fracture healing with soft tissue injury. Arch Orthop Trauma Surg 2007. [22] Kaspar D, Neidlinger-Wilke C, Holbein O, Claes L, Ignatius A. Mitogens are increased in the systemic circulation during bone callus healing. J Orthop Res 2003;21:320–5. [23] Bumann M, Henke T, Gerngross H, Claes L, Augat P. Influence of haemorrhagic shock on fracture healing. Langenbecks Arch Surg 2003;388:331–8. [24] Parfitt AM. Skeletal heterogeneity and the purpose of bone remodeling. In: Marcus R, Feldman D, Kelsey J, editors. Osteoporosis. San Diego: Academic Press; 1996. p. 315–29. [25] van der Poest CE, van EM, Ader H, Roos JC, Patka P, Lips P. Alendronate in the prevention of bone loss after a fracture of the lower leg. J Bone Miner Res 2002;17:2247–55. [26] Seebach C, Kurth A, Marzi I. The influence of bisphosphonates on fracture healing. Orthopade 2007;36:136–40. [27] J. Goldhahn, N. Suhm, S. Goldhahn, M. Blauth, B. Hanson, Influence of osteoporosis on fracture fixation — a systematic literature review. Osteoporos Int. 2007. [28] Utrilla AL, Reig JS, Munoz FM, Tufanisco CB. Trochanteric gamma nail and compression hip screw for trochanteric fractures: a randomized, prospective, comparative study in 210 elderly patients with a new design of the gamma nail. J Orthop Trauma 2005;19:229–33. [29] Goldhahn J, Seebeck J, Frei R, Frenz B, Antoniadis I, Schneider E. New implant designs for fracture fixation in osteoporotic bone. Osteoporos Int 2005;16(Suppl 2):S112–9. [30] Koistinen A, Santavirta SS, Kroger H, Lappalainen R. Effect of bone mineral density and amorphous diamond coatings on insertion torque of bone screws. Biomaterials 2005;26:5687–94. [31] Seebeck J, Goldhahn J, Stadele H, Messmer P, Morlock MM, Schneider E. Effect of cortical thickness and cancellous bone density on the holding strength of internal fixator screws. J Orthop Res 2004;22:1237–42. [32] Jones G, Boon P. Which bone mass measures discriminate adolescents who have fractured from those who have not? Osteoporos Int 2008;19:251–5.