Portal de Periódicos da CAPES

Can Bisphosphonates Be Given to Patients with Fractures?

2001; Oxford University Press; Volume: 16; Issue: 3 Linguagem: Inglês

10.1359/jbmr.2001.16.3.437

1523-4681

H. Fleisch,

Bone fractures and treatments

One of the frequently asked questions concerning patients treated with bisphosphonates is whether individuals who recently have sustained a fracture should take inhibitors of bone resorption such as these. This argument is relevant because many of the patients treated with these drugs do have fractures, such as patients with osteoporosis, Paget's disease, tumor bone disease, and more recently, osteogenesis imperfecta. Fracture healing has always been a major medical concern and it has been the aim of physicians and basic scientists alike to find ways to shorten the healing time and to prevent nonunion. Until recently, the main progress has been in the surgical procedures, which have allowed solid stabilization of the fractured segments. The attempts to develop drugs to stimulate bone formation have not been successful yet, although the discovery of bone-forming growth factors, such as the bone morphogenetic proteins (BMPs), transforming growth factors (TGFs), fibroblast growth factors (FGFs), and others, raises hope that soon we shall make use of their anabolic properties. It is only recently that investigations have addressed the positive or negative influence of bone resorption inhibitors on fracture healing. Rather, emphasis has been largely on the inhibition of fracture incidence. However, with the wide use of the bisphosphonates, more recently, attention has focused on whether these drugs are, in fact, deleterious to fracture healing. Therefore, experiments in various animals are now available, which investigate bisphosphonate effects on the healing of fractures. In this issue of the Journal, Li et al.1 investigated the long-term effect of the bisphosphonate incadronate on fracture healing of the femoral shaft in growing rats. This article is the continuation of another publication by this group,2 based on a similar protocol, but in which the animals were analyzed after 6 weeks and 16 weeks, as opposed to 25 weeks and 49 weeks as in the present work. In this previous study, the authors found that incadronate led to a larger callus, a delay in its turnover, and an increase in ultimate load. In the present study, incadronate was injected three times a week at a dose of 10 μg/kg or 100 μg/kg for 2 weeks before making a transverse osteotomy of the femur by means of a circular blade, which was then fixed by stainless steel wire. The administration was then either discontinued or continued for another 25 weeks or 49 weeks, at which time the animals were killed. After death, the fracture site was investigated by radiography, microradiography, morphometry, and mechanically by the three-point bending method. The authors found that administration for 2 weeks before inducing the fracture had no effect on any of the parameters. In contrast, the continuous treatment until death influenced the process of fracture healing as assessed morphologically. Although all the fracture lines disappeared when analyzed radiographically, changes occurred in the callus. Thus, the amount of callus was increased at both doses after 25 weeks, and with the higher dose after 49 weeks. In contrast, the maturation to a final repair was delayed, because the transformation of woven into lamellar bone and, therefore, the formation of the lamellar cortical shell was slowed. All morphometric parameters of bone turnover were decreased. Last, the biomechanical investigations showed an increase in stiffness and ultimate load but no change in the ultimate stress of the fractured femur in the groups receiving the higher dose. No change was seen with the lower dose. Thus, the slowing of callus turnover was accompanied paradoxically by a higher mechanical strength. These results are in line with earlier studies performed on various animals with alendronate, clodronate (in most studies), etidronate, pamidronate, and tiludronate. Although the results of these are heterogeneous, which is not surprising in view of the many different animal models, bisphosphonates, drug doses, and fracture models used, a certain trend appears. Thus, the size of the callus was either not influenced3 or was increased4-8 but never decreased. This has resulted in the filing of patents for the use of bisphosphonates for the actual treatment of fractures.9 The mechanical parameters were influenced somewhat in a negative sense only in one study and in one of the conditions used10 but usually are not altered4, 7, 8 or improved.5 The article in the current issue of the Journal1 raises several questions: (a) What is the relation between the morphological and biomechanical investigations in fractures, and how can the results be extrapolated to what will occur in vivo? and (b) Can bisphosphonates be administered safely to patients who have sustained a fracture recently? To answer these questions a few considerations on the process of fracture healing are necessary.11 After a fracture, there is loss of bone integrity and continuity as well as a rupture of blood vessels with local avascularity. The spontaneous healing of a mechanically unstabilized fracture proceeds by: (a) spontaneous interfragmentary mechanical stabilization effected by periosteal and endosteal callus formation made of woven bone and interfragmentary fibrocartilage formation, which forms a cushion between the fragments; (b) bone union by intramembranous and endochondral ossification; (c) osteoclastic destruction and substitution of avascular and necrotic areas by Haversian remodeling; and (d) Haversian remodeling of the callus cuff and the fracture site so that the bone resumes as close as possible its previous form. This development is visualized radiologically by relatively extensive callus formation, the more unstable the fracture site the larger, by temporary widening of the fracture gap because of osteoclastic destruction of the fracture ends, by the slow disappearance of the fracture line caused by mineralization of the fibrocartilage and the formation of mineralized bone, and finally by a trend to regain the original form. The process is in many ways similar to the one seen in the embryo and during growth. In case the fragments are fixed rigidly by an internal device such as a compression plate or a leg screw, less or sometimes no bridging by external callus is necessary and the healing proceeds by direct union from one fragment to another, usually by lamellar bone. In the second step, the fractured site is remodeled as it is in nonrigidly fixed fractures, by means of Haversian remodeling. For this purpose, it is perforated by osteoclastic cutter cones, followed by osteoblasts, which fill the canals with bone. This process is favored by the fact that the fracture by itself activates the formation of new bone multicellular units (BMUs) in the surrounding bone. Therefore, in both nonstabilized and stabilized fractures, healing involves more stages of osteoclastic destruction, which explains the changes seen under bisphosphonate therapy. All these processes are modulated by a series of cytokines and their receptors, which either induce bone formation, such as FGFs, platelet-derived growth factor (PDGF), TGF-βs, and the BMPs, or induce bone resorption, such as interleukins 1 and 6 and TGF-α.12 These factors have a large array of effects on chemotaxis, multiplication, differentiation, maturation, and activity of the local cells. The cells producing these various factors either originate from the ruptured vessels, such as thrombocytes, from local tissues, or are transported to the site during angiogenesis. Inflammatory, mesenchymal, cartilage, and bone cells are the main cell units involved. This cascade of events permits the reconstruction of the fracture site so that function is restored as fast as possible. This is warranted only if interference with the mechanical status does not occur. Indeed, any movement or micromovement is deleterious to the reconstruction and stimulates bone destruction. From the studies on the effect of bone forming factors on fracture healing, it appears that it is very difficult to correlate the morphological appearance of the callus with its biomechanical strength. This is no surprise considering the considerable heterogeneity of the callus. Furthermore, mechanical strength itself depends on a great variety of factors. Among these, bone mass probably is the most important. Incidentally, this parameter is not always faithfully reflected by bone mineral density (BMD), because the degree of bone mineralization can change, especially under treatment with bisphosphonates.13 Part of the calcified tissue present may be cartilage, which does not have the same mechanical characteristics as bone. Other parameters are also important for the mechanical strength, such as tridimensional architecture of cortical and trabecular bone; bone quality, including the chemical composition; crystal size, shape, and crystallinity, as well as the chemistry and structure of the matrix, especially collagen. Last, osteocyte viability and bone turnover itself are important. To analyze all these parameters to draw a conclusion on the mechanical properties is not possible at this time, and hasty deductions are dangerous. Therefore, the biomechanical investigations probably are the best means available to assess the effect of a drug on the fracture site. This is similar to what happens in osteoporosis in which the occurrence of fractures is the primary endpoint today. In the article by Li et al.1 the morphology of the fracture site was influenced strongly by the administration of incadronate. The changes observed, such as the larger callus and the decreased callus development, might all be secondary to the known inhibitory activity of the drug on bone resorption. The decrease in bone formation they found is likely secondary to the decrease in bone resorption, because the two processes are linked intimately during remodeling. Although one investigation did show a decrease in the viability of the osteoblasts in the fracture callus,6 currently there is no indication that the decrease in bone formation induced by the bisphosphonates in vivo is direct. On the contrary, the wider mean wall thickness found in some studies14 and the recently described decreased apoptosis of osteoblasts and osteocytes15 would suggest the opposite if anything. Furthermore, in the growth plate, which has a certain similarity with the fracture site, only the resorption but not the formation of either cartilage or bone appears influenced by bisphosphonates. The most important information in this article is the fact that the biomechanical properties are not decreased, but in the case of stiffness and ultimate load are increased and, in the case of ultimate stress, are unchanged. An increase in callus fragility was a concern based on various arguments. One was that the inhibition of bone turnover by a drug might induce a condition resembling osteopetrosis. Large amounts of bisphosphonates can induce an increase of microcracks16 and fractures17 in animals. Because bisphosphonates are deposited preferentially at fracture sites,18 the latter could be a predilected site for such an effect. In contrast, slowing bone resorption and turnover is believed to be one of the mechanisms by which bisphosphonates reduce fractures in osteoporosis. Indeed, a higher number of bone remodeling sites, in which there is excessive osteoclastic destruction of bone, leads to the development of areas of stress concentration, and hence to increased fracture risk. The delay in callus evolution and the massive changes in morphology were another concern. The decrease in ultimate stress (load divided by surface) and hence quality of the callus, seen in the earlier study by the authors,2 which was conducted over a shorter period of time, suggested that at some time point during the fracture healing period there may be an increase in callus fragility. However, the results in the present study do not sustain these fears but actually suggest that the contrary occurs, namely, an improvement of the mechanical properties of the callus. This is likely due, at least in large part, to the increase in callus size and in its outer diameter. It appears that the organism has compensated any negative effect of the drug by modulation of the morphology to obtain the best mechanical function. This phenomenon of adaptation of structure to function is present all over the skeleton and is likely governed by a feedback mechanism called the mechanostat. Therefore, inhibiting bone resorption does not appear harmful with respect to bone strength, at least in this model and at the doses of bisphosphonates administered. The question remains whether this conclusion may be extrapolated to other bisphosphonates. This seems likely because there is no theoretical basis for other bisphosphonates to behave differently. Furthermore, when not given in excess, many bisphosphonates, such as alendronate, clodronate, etidronate, ibandronate, incadronate, minodronate, neridoronate, olpadronate, pamidronate, risedronate, and tiludronate have a positive effect on mechanical characteristics of nonfractured bone. These include parameters such as torsional torque, ultimate bending strength, stiffness, maximum elastic strength, Young's module of elasticity, and others. However, there is an exception if a bisphosphonate with a weak inhibitory activity is administered in a large dose, at which it will induce an inhibition of mineralization. This is the case with etidronate, which has been shown to produce a mechanically weaker bone, both at the fracture site,19 as well as throughout the bone, when given in large amounts.17 Regarding whether it is advisable to administer bisphosphonates to patients with a recent fracture, no specific investigations have been done in humans. However, no human data, despite the very large number of patients treated with these drugs, many of them with fractures, suggest, either in randomized studies or in postmarketing experience, that the administration of bisphosphonates, at least in doses used for osteoporosis, would interfere clinically with fracture repair. Actually, fracture incidence is decreased with both alendronate20 and risedronate.21 The same is true for children with osteogenesis imperfecta, who received relatively larger doses of pamidronate (6.8 ± SD = 1.1 mg/kg per year in one study and 12.4 mg in another one).22, 23 Actually, in one of them radiological fracture healing was given special attention. Furthermore, although the extrapolation from animal to humans is never secure entirely, especially if the experiments are performed on bones of growing rats, which are different from those of adult humans, having no Haversian remodeling, the animal data in the literature strongly suggest that the recurrence of a new fracture is not a reason to suspend or avoid treatment with bisphosphonates, at least at the doses administered in osteoporosis. With alendronate, these were found to decrease bone turnover to premenopausal levels only.24 However, these conclusions are true only for bisphosphonates, which do not inhibit mineralization at the doses given. Guidelines for patients receiving larger doses, such as in Paget's disease and in tumor bone disease, are more difficult to establish. However, there are again no human data indicating that a problem would exist in these patients. Furthermore, the animal results suggest that even larger doses do not increase callus fragility. Thus, the decision will have to be taken on an individual basis, based on the risk-benefit relationship.