Osteonecrosis of the Femoral Head of Laboratory Animals: The Lessons Learned from a Comparative Study of Osteonecrosis in Man and Experimental Animals

Spontaneous Osteonecrosis in Laboratory Animals

Osteonecrosis of the femoral head has been well studied in 6-week- to 8-month-old, spontaneously hypertensive rats. During early life, defect in expression of insulin-like growth factor-1, which governs growth and differentiation of the cells via an autocrine mode, seems to lead to delayed upregulation of the mineralization-controlling type X collagen in the rats' hypertrophic cartilage. Fibrous tissue invades the cartilage normally. However, the imperfect mineralization of this cartilage does not adequately prepare an environment conducive to the formation of bone. The secondary ossification center of these rats' femoral capital epiphyses does not mature properly. Consequently, everyday loads of weight bearing suffice to deform the mechanically weak femoral head. In addition, maximal weight-bearing loads compress the arteries entering the femoral head at the lateral facets, which are the weakest and most easily deformed segments of the femoral head. The route of the lateral epiphyseal arteries, the main nutrient suppliers, differs in spontaneously hypertensive rats from that in other rat strains in that the vessels abruptly decrease in diameter at their entry sites into the ossific nucleus of the epiphysis. Stenosis or obstruction of the lumen of these arteries develops. In parallel with the most severely affected parts, the ensuing ischemia induces osteonecrosis at the lateral epiphyseal facets of the femoral heads. Not surprisingly, the incidence of osteonecrosis decreases with decreasing loads on the hip and with nutritional restriction, which is probably related to the rats' lower body weight, shorter femurs, and more advanced epiphyseal ossification. ²¹ , ²⁴ , ²⁸ , ⁶²

Avascular necrosis of the femoral head is sporadically encountered in dogs. Perthes disease–like necrosis of the femoral head and neck occurs in some breeds of small dogs. Hypoxemia due to thrombotic venous occlusion and compression or scarcity of arteries appears to lethally injure these dogs' developing epiphyses. ⁵

Surgically Induced Osteonecrosis

The blood supply of the femoral heads derives from vessels emanating from the medial and lateral circumflex femoral arteries and from anastomoses between the epiphyseal and metaphyseal circulations. Because of these dual networks, it is a challenging exercise to surgically cause femoral capital necrosis in animal species with functionally sufficient transphyseal blood circulation. As established in experiments on time-dependent osteonecrosis in bone-containing chambers implanted in the tibiae of rabbits, osteocytes die within 20 minutes of blood deprivation due to arterial occlusion. ⁷⁴

In animals of all species, the capillary-sized vessels that anastomose with both epiphyseal and metaphyseal circulation are functionally ineffective. Stripping, occluding, or destructing the retinacular vessels results in necrosis of the canine femoral head. The flow of blood is maximally reduced by combination of the stripping of the retinacular vessels with the reaming of the bone. Interstitial edema and hemorrhage are widespread in the bone marrow on the third postoperative day. Magnetic resonance imaging provides clues for and then clear-cut evidence of osteonecrosis by the fourth and seventh postoperative days, respectively. The hydrostatic pressure of intra-articular fluids plays a pivotal role in regulating the femoral head's blood supply, the flow decreasing with increasing intra-articular pressure above the arterial blood pressure. A tamponade of the hip joint compresses the blood vessels of the synovial membrane to such a degree that the flow of blood dwindles in the femoral head. As evinced scintimetrically, bone ischemia does not persist for longer than 4 weeks after raising the intra-articular pressure of puppies' hip joints to 50 mm Hg for 6 hours. Intra-articular collection of fluid to lower the arterial pressure does not significantly impair the blood flow in the femoral head. The circulation in the avascular femoral head is reestablished within 3–4 months of the ischemic event. In contrast, circulation in a replanted avascular spongy bone is restored within 3 weeks. A tract drilled from the greater trochanter into the head acts as a conduit, allowing for the ingrowth of fibrous tissue into the avascular epiphysis. Partial to complete collapse of the femoral head correlates with the extent of creeping substitution in cases in which revascularization of the avascular bone proceeds unrestrained. Collapse does not ensue in cases in which revascularization is suppressed. ⁶ , ¹¹ , ²⁵ , ²⁹ , ⁴⁷ , ⁵⁶

The radiographic-graded extent of necrosis of the femoral heads of patients with Legg-Calvé-Perthes disease ranges from mild (less than 15% involvement) to severe (over 30% involvement). In the rat model of vascular deprivation–induced necrosis of the femoral head, the epiphysis consistently undergoes complete necrosis. This difference (one of several dissimilarities between the disorder in man and rat) appears to be incidental to species differences in size, anatomy, and physiology, and to the artificiality of the experimental setup. Indeed, the clinical and the experimental situations are similar, not equivalent. The surgically induced abrupt stoppage of the arterial and venous blood flow to and from the femoral head of otherwise healthy animals does not truly approximate the pathogenesis of osteonecrosis in adults or Perthes disease in children. ⁵⁰ , ⁶⁰

Stripping the cervical periosteum and cutting the ligamentum teres produce necrosis of the rat's femoral capital epiphysis and often the physeal and articular cartilage as well. The first microscopic signs of necrosis are apparent on the second postoperative day. Complete death of the bone is visible on the fifth postoperative day. Within another day or two, vascularized fibrous tissue originating in the inflamed and hyperplastic joint capsule starts to penetrate the necrotic debris. In the second postoperative week, capillaries sprout and branch, fashioning irregular networks. Subsequently, the capillaries transform into arteries and veins, reestablishing effective circulation. The perfusion exceeds physiologic levels by the second postoperative month, but it normalizes after the third month. The amount of the invading fibrous tissue peaks during the second postoperative week so that its penetration and the revascularization are far advanced by the third postoperative week. Concurrent with the macrophagic, osteoclastic, and chondroclastic resorption of the necrotic tissues, maturation of fibrous tissue coincides with appositional and intramembranous osteogenesis during the second to fourth postoperative weeks. By the second postoperative month, all the necrotic debris has been removed. Repopulation of the reconstituted intertrabecular spaces by fat and hematopoietic cells also occurs at this time. Remodeling of the epiphysis follows the processes of fibrogenesis, chondrogenesis, and osteogenesis. Using ³H-proline-, ³H-tetracycline-, and ⁴⁵Ca-labeling techniques, quantification discloses equivalent amounts of bony calcium and collagen during the different phases, suggesting that osteolysis and osteogenesis nearly balance each other. ¹² , ³³ , ⁷⁴

The repair processes, including the ongoing remodeling, are well advanced by the sixth postoperative week in some rats and by the eighth postoperative week in all rats. Everyday strains of load transfer bring about deformation of the femoral heads, ranging from epiphyseal collapse (flattening) to disfiguring epiphyseal enlargement with an excessive osseous framework, reminiscent of the enlarged bone volume in arthritic patients with subchondral osteosclerosis. This overshooting of new bone formation modifies the epiphysis from cancellous to focally compacta-like bone. Stiffness, strength, and toughness of any newly formed, immature bone are inferior to those of mature bones. Reduced yield strength, reduced elastic modulus, and increased strain-to-failure are characteristic of infarcted bones during their early healing stage. This weakening of the epiphysis explains the architectural distortion of the rats' devascularized femoral heads during the remodeling stages. In some cases, epiphyseal-metaphyseal bony bridges bisect the physeal cartilage, reflecting an aberrant pattern of bone formation in the wake of segmental necrosis of the physis. Coincidental regressive changes of the articular cartilage range from loss of glycosaminoglycans and fibrillation to total destruction, resulting in eburnated bone. In extreme instances, bony and cartilaginous shreds that are randomly scattered in an unorganized collagenous framework and no longer display a hemispherical configuration replace the femoral head. Altogether, the remodeling of the epiphysis, cartilage degeneration, and the accompanying synovitis account for the osteoarthritis-like features developing 2 months or so after the ischemic event. In Sokoloff's view, the trio of degenerative, inflammatory, and regenerative processes simultaneously disrupting the osteochondral junction, joint cartilage, and articular capsule fit the osteoarthritic profile. ³² , ⁴² , ⁵⁰ , ⁵⁵ , ⁵⁸

The biomechanic integrity of the subchondral bone plate and cancellous bone is initially preserved. Later, the subchondral bone deteriorates as the process of epiphyseal remodeling unfolds. Stress-to-strength ratios reflect the tendency of structural failure to be higher in the cancellous bone than in the subchondral bone. The stress levels in the postnecrotic subchondral bone are smaller than in the normal subchondral bone, attaining approximately 70% of the physiologic levels. This reduction suffices to deform the loaded, remodeled epiphysis. The onset of the femoral capital collapse is first and foremost influenced by the extent and intensity of the structural bone deterioration in the main areas of the infarcted epiphysis rather than by the degree of structural degradation of the subchondral bone. ⁷ , ³² , ³³ , ⁵⁵

The frequently poor results of core decompression of patients' necrotic femoral heads are due to the transience of drilling-induced normalization of the intraosseous pressure. All the effects of decompression are lost within 3 weeks of drilling a sheep's necrotic head because the tracts are obliterated. With or without inserting a resorbable stent, the channels are initially sealed off by blood clots and later by fibrous tissue. Advancing simultaneously from the periosteum and marrow along the drilled channel, blood vessels sprout, linking the periosteal-diaphyseal-metaphyseal and epiphyseal-metaphyseal vascular networks with one another. However, the too scant anastomoses fail to adequately furnish the bone's demand for oxygen. ⁵⁷

Blood flow abates in a puppy's femoral head after traction application. It falls to approximately one third of the baseline values in animals with tamponade of the hip joint. These experimental data correlate with the iatrogenic femoral capital necrosis complicating traction treatment of children with congenital hip dysplasia and adults with a nondisplaced fracture of the neck. Joint position–dependent variations of the blood flow in the main arteries as well as rupture, stretching, torsion, and compression of the blood vessels play contributory roles. When a pig's hip is held in the frog leg position, the blood flow in the femoral head decreases dramatically after ligation of the deep femoral or lateral femoral circumflex arteries. It is not easy to induce necrosis of a pig's femoral head solely by the obliteration of the physis because nutrient branches arise from the circumflex femoral and cervical medullary arteries. The femoral head can survive either a dislocation or a ligation of the circumflex femoral arteries and veins. Yet, the heads usually become necrotic by the second to fourth postoperative week in pigs subjected to both procedures because the blood flow decreases to about 15% of baseline values. By the same token, an 80% decrease in the amount of the normally supplied blood is prerequisite for necrosis to ensue in the canine femoral head. ¹⁶ , ⁴⁶ , ⁶¹

Animal Models of Legg-Calvé-Perthes Disease

When comparing the experimental models with Perthes disease in children, it should be kept in mind that perfect restoration of the epiphysis does not occur in Perthes disease; the femoral heads usually do not regain their optimal hemispherical form. Revascularization of the epiphysis fails to materialize in some patients, repair falls short, and destruction of the femoral head ensues. This clinical experience correlates with experimental observations in so far as the shape factor and height-to-length ratio of the epiphyses on the 18th and 36th day after vascular deprivation–induced necrosis of the rats' femoral heads differ significantly from those of control animals' femoral heads. Although some heads are flattened, other animals' femoral heads gain in height. ²⁷ , ⁵⁵

The supposition that joint effusions, coexisting with transient synovitis, cause Perthes disease was not based on experiments. The epiphyseal blood flow in femoral heads remained unchanged when the intra-articular pressure in puppies' hip joints was moderately raised (50 mm Hg). Extreme elevations of the intra-articular pressure (150 mm Hg) were necessary to disrupt the intraosseous circulation. The same holds true for pigs. The pressure in the hip joint must be raised to 150 mm Hg for 10 hours to cause osteonecrosis. In immature pigs with talcum-induced synovitis, the epiphyses remained viable after raising the intra-articular pressure to 200 mm Hg by infusing autologous serum. Two hours after the infusion, the intra-articular pressure in these animals decreased to 35 mm Hg because the infusate oozed out through the inflamed and stretched synovial membrane. The results of this experiment cast doubts on the belief that a tamponade by synovial infusion of the hip joint is a significant causative factor in Perthes disease in children. ³ , ⁷⁵

Notwithstanding its tamponade-like effects, distension of the articular cavity does not lead to epiphyseal ischemia unless it is combined with compression of the joint. The results of some “exotically” extreme conditions appear to contradict these principles. For instance, keeping 6-week-old rabbits' internally rotated hips for 1 whole day in a flexed and abducted position obstructed the vessels so that necrosis of the femoral capital epiphysis ensued. To cite one more example, casting the hip joints in hyperextension, full internal rotation, and abduction for 20 hours caused a Perthes disease-like disorder in puppies. ²⁰

In adult animals with a lifelong persisting physis, a cartilaginous stratum separates the epiphyseal and metaphyseal blood circulations. The small amounts of blood passing through the physis do not satisfy the nutritional demand of the epiphyseal tissues. The rat's femoral head is a case in point. In adult dogs, on the other hand, the physis is obliterated at maturation so that the femoral head is nourished by capsular as well as medullary vessels, each network being capable of substituting for the other. Nonetheless, raising the pressure in the joint for 2 hours to a degree that mildly impedes the circulation may lethally injure trabecular osteocytes, although the marrow cells are unharmed. The widespread appositional osteogenesis throughout the whole epiphysis of these dogs attests to an antecedent death of bone cells. ²⁹ , ³⁰ , ⁶⁴ , ⁷³

The epiphysis of the femoral head is already necrotic by the first stage of the naturally occurring canine Perthes disease-like disorder. Disturbed physeal endochondral ossification, in conjunction with the mechanical load on a remodeled asymmetric femoral head, results in its collapse. Manchester Terriers' inherited disorder is characterized by transient disruption of the venous blood flow, inducing lesions, which resemble histopathologic features of Perthes disease. Hindrance of the venous return after instillation of semiliquid silicone into young dogs' femoral necks raises the intraosseous pressure to levels resulting in focal necrosis of the head, suggesting a Perthes disease-like disorder to some authors. Obstruction of venous drainage is associated with elevated cervical intraosseous pressure in children with unilateral Perthes disease. Are orthopedic surgeons' manipulations responsible for Perthes disease in children with congenital hip dislocation? The rate of the blood flow decreases in the femoral heads of puppies when compression or traction is applied at one half of the animals' body weight. Either maneuver, acting side by side with tamponade of the hip joint, reduces the blood flow rate by about 70%. These results were not reproducible in adult dogs. Hence, the blood circulation in femoral heads of young dogs may be especially vulnerable to a variety of insults. ¹ , ³¹ , ³⁴ , ³⁹ , ⁵⁷ , ⁶³ Perhaps the same holds true for the femoral head of children.

Transphyseal bony bridges constitute part of the epiphyseal remodeling in one third of rats with vascular deprivation–induced femoral capital necrosis. Bone develops in the gaps formed by focal necrosis of the physis during the acute ischemic phase. The similarity to Perthes disease is clear. The broken-up physis underlies a remodeled epiphysis and overlies a remodeled metaphysis. Even in children, bony bridges transiently pierce the physis along vessels crossing the cartilage during the revascularization stage. ⁵³

The coexistence of necrosis of the epiphyseal bone and physeal cartilage with viable joint cartilage is a feature of the spontaneously hypertensive rat model that resembles Perthes disease. In both the spontaneously hypertensive rat model and Perthes disease, the growth arrest of the anterolateral portion of the physis and perichondrial ring is concurrent with ongoing or even accelerated growth of the posteromedial portion of the physis. The delayed ossification of the epiphysis contrasts with the premature maturation of the physis. Crawford et al. reason that animal models featuring femoral capital collapse in the wake of vessel severance are not similar to Perthes disease. ⁹ The endeavor of these authors to mimic Perthes disease by removing young goats' physeal cartilage is open to criticism. Physeal ablation led to remodeling of the femoral head and development of transphyseal fibrous, fibrocartilaginous, or osseous bridges. The remodeling triggered by cauterization of the physis of young goats resulted in a growth-retarded, fragmented, and mushroom-shaped femoral head. These changes are unlike those observed in Perthes disease. Instead, the changes reflect an alteration in the bony framework resembling osteoarthritic reaction patterns. Transphyseal bony bridges evolve in the aftermath of necrosis of the physeal cartilage, be it the outcome of a disease, trauma, or otherwise. ⁹ , ²¹ , ²⁴ , ²⁸ , ²⁹ , ⁵³ , ⁶²

Traumatic Osteonecrosis

The pathogenesis of femoral capital osteonecrosis is clear in patients who sustained a fracture of the neck or dislocated hip that interferes with circulation. The blood supply to and drainage from the head is severed because the vessels at the fracture site and in the ligamentum teres are torn. As a rule, the effect of disruption of the extraosseous blood vessels is more severe in young than in adult animals. Effective epiphyseal-metaphyseal anastomotic networks across the effaced physis of adult rabbits' femoral heads minimize an otherwise devastating consequence of a blocked extraosseous circulation. There is an essential difference between posttraumatic osteonecrosis of rabbits and other types of avascular necrosis. Viable osteocytes with well-stained nuclei, persisting within the necrotic foci on either side of the fracture line, span the live bone-dead bone interface. However, viable osteocytes with well-stained nuclei are not present when the fracture line passes through an already necrotic bone. Blood flow may be partially restored by vessels from the cervical medulla, but adequate flow is not reestablished until circulation through the medial femoral circumflex and ligamentum teres arteries is restored. Necrosis of the femoral head not infrequently complicates nonunion of a dog's fractured neck. ² , ⁴⁴

Corticosteroid-Induced Osteonecrosis

Corticosteroid medication is a pivotal risk factor in the development of avascular osteonecrosis. Conditionally, it may cause the disease. The pathogenetic mechanisms involved in steroid-induced osteonecrosis are poorly understood. First recognized after short-term, high-dosed steroid therapy in occasional patients, most cases are diagnosed after long-term treatment. ⁴ By the fifth month of methylprednisolone therapy for lupus erythematosus, the incidence of osteonecrosis is close to 50% of patients screened by sensitive techniques. New osteonecrotic lesions rarely develop after the sixth month of treatment. ⁵¹

Lipid emboli plug the intraosseous vessels, primarily subchondral capillaries and arterioles, of steroid-treated patients who have never manifested clinical signs of osteonecrosis but whose bones, at autopsy, contain scant viable osteocytes. The microvascular anatomy accounts for the preferential embolization of the vessels of the subchondral zone. After intra-arterial fat infusions, lipid-containing emboli block the vessels and result in necrosis of rabbits' femoral heads. Hyperlipemia and bone-marrow necrosis peak within a week of subjecting rabbits to high-dose steroid regimens. The arteries, mainly those of the subchondral zone, are occluded by fat emboli and lipid-loaded fibrin-platelet thrombi, which are partly derived from the steroid-induced fatty liver. The disturbed lipid metabolism of the rabbits is reflected in hypertrophy of bone marrow adipocytes and an elevated ratio of low-density to high-density lipoprotein cholesterol. This lipid-induced hypertrophy of the fat cells cannot expand the marrow cavity within the inflexible osseous cage. Consequently, the intraosseous pressure rises, leading to sinusoidal compression, venous stasis, and, eventually, arterial obstruction, accounting for the ischemic osteonecrosis. The outcome of elevated intraosseous pressure depends to a large extent on additional parameters. Steroid-induced cholesterol deposition reduces the fluidity and permeability of the cell membranes, contributing to the death of the osteocytes. Microfractures, accumulating in the fragile remodeled bone, further compress the subchondral vessels, further compromising the already unstable circulation. The blood flow may be reduced up to one third after continuous steroid treatment for 10 weeks. A failing blood supply of this magnitude would by itself be insignificant but would exacerbate cell death incidental to other conditions. Lastly, intraosseous hypertension inhibits regeneration of the blood vessels. Treating the animals with the platelet aggregation–reducing and fibrinolytic aprotonin decreases the levels of the serum lipids, preventing osteonecrosis. ¹⁵ , ²³ , ²⁶ , ⁵⁹ , ⁶⁵ , ⁶⁶ , ⁶⁸ , ⁶⁹ , ⁷⁷

The uncertain benefit of core decompression has been witnessed in steroid-induced osteonecrosis of rabbits' femoral heads as well. Granulation tissue grows into the surgically created conduit, ushering in revascularization of the avascular head. Myofibroblasts in this granulation tissue secrete an endothelial cell growth factor, stimulating angiogenesis. But by suppressing activities of the myofibroblasts, the steroids inhibit proliferation of the capillaries. Blood flow increases immediately after the core decompression, but it decreases during the second postoperative month. ¹⁹ , ⁶⁷

Steroid medication increases the intraosseous pressure in spite of decreased blood flow. This phenomenon is probably related to the pathogenic role of an obstructed venous drainage in the development of osteonecrosis. The ubiquitous, intimal thickening of the rabbits' veins after an 8-week-long methylprednisolone regimen results from proliferation of myocyte-derived foam cells. To what extent this vasculopathy on the venous side of the circulation contributes to the blood stasis is unknown. ⁵⁴

Mice treated with dexamethasone have bone marrow infiltrated by numerous, lipid-laden, pluripotential stromal cells. Pluripotential marrow stromal cells cultured in a dexamethasone-enriched medium differentiate into adipocytes and partly lose the expression of osteocalcin messenger RNA and type I collagen, paralleling the suppressed maturation of mesenchymal cells into osteoblasts. Addition of the lipid-lowering agent lovastatin to the medium inhibits the steroid-induced expression of adipocytic genes and counteracts the inhibition of expression of osteoblastic genes. Treatment of rabbits with the lipid-clearing clofibrate reduces steroid-induced hepatosteatosis, hyperlipidemia, and accumulation of lipids in osteocytes and protects the femoral head against necrosis. ¹⁰ , ³⁶ , ⁴⁹ , ⁷⁰

Lipopolysaccharide-induced Osteonecrosis

Platelet activation plays a crucial role in the pathogenesis of some patients' osteonecrosis. Yet, endotoxin-induced disseminated intravascular coagulation uncommonly causes osteonecrosis in clinical practice. Low-dose lipopolysaccharide (10 mg/kg of body weight) injected intravenously leads to multifocal osteonecrosis in about 80% of the treated rabbits. Fibrin thrombi clog myriad arterioles and small arteries within 24 hours of the injection. Thrombocytopenia, hyperlipidemia, raised plasma levels of plasminogen activator inhibitor-1 and elevated adipocytic tissue, and macrophage factors are characteristic of the early stage of this disorder. Vessels obliterated by fibrosis surround the osteonecrotic foci a few weeks after the injection. The necrotic areas in steroid-treated rabbits with a Shwartzman reaction are often massive, extending from the epiphysis to the diaphysis of the femur. Corticosteroids reinforce the processes leading to thrombotic vascular occlusion by injuring endothelial cells and amplifying hypercoagulability. Thrombi obstruct the arterioles near regions of hemorrhaging from damaged vessels after an injection of platelet-activating factor and two injections of lipopolysaccharide. Necrosis of the osseous trabeculae develops in one half of the treated rabbits. ³⁷ , ⁷⁶

Immune Reaction–induced Osteonecrosis

The extent of necrosis in femoral epiphyses, metaphyses, and diaphyses increases from the first to the third week after the second of two injections of horse serum in sensitized rabbits. Early on there are significantly more necrotic marrow cells than osteocytes. Thrombi clog arterioles and small arteries near foci of extravasated red blood cells. Antigen-antibody complexes are deposited in the vascular walls close to the osteonecrotic lesions, substantiating the role of an immune reaction in the pathogenesis of this model of osteonecrosis. ³⁸ , ⁴⁸

Physical Injury–induced Osteonecrosis

The low temperature produced by liquid nitrogen cryoprobes inserted into a rabbit's bone precipitates thrombotic occlusion of the microvasculature, bringing about osseous infarction. In the absence of vascularized tissue covering the joint cartilage, reparative processes are limited to the deep compartment of the femoral head. This is clinically relevant because newly formed periosteal bone strengthens, under other circumstances, the involved bone segment at the early healing stages. A comparable strengthening mechanism does not occur at the articular perimeter of the femoral head. Osteogenesis is less active during the first 7 weeks after cold injuries than after other types of damage. The bone volume destroyed by freezing is three times larger than the dimension of the cavity prepared for insertion of the cryoprobe. Osteogenesis gets ahead belatedly within just a thin zone of the cryonecrotic bone. Initially, regeneration is slow, but healing catches up later with complete repair of the defect by the 24th week after the freezing event. ³⁵ , ⁷⁸

Hyperthermia kills osseous tissues, but the threshold at which the thermal stress consistently kills osteocytes is debatable. Raising the ambient temperature of a dog's bone to 42.5 C for 60 minutes causes necrosis up to a distance of 5 mm from a heated stainless steel nail. The bony tissues are only mildly damaged, as evidenced by full recovery of the bone within 12 weeks of heating. But higher temperatures are poorly tolerated. At temperatures over 45 C, each elevation of 1 C prolongs the healing time by a factor of two, after a thermal damage of the rats' distal tail vertebrae. A low pH, such as that which is a result of an impaired blood supply, renders the bone hypersensitive to the injurious effects of high temperatures. ⁴⁵

Necrosis of the femoral head occurred in nearly 40% of rats reared from their 5th to 15th week of life in high cages, which forced the animals to raise themselves up on their hind limbs to reach the food box and drinking aperture. The lateral aspects of these rats' hip joints were rebuilt within the framework of tissue repair. At the entrance sites of the vessels into the epiphysis, the arteries are blocked such that the femoral heads undergo ischemic necrosis. It is provocative that a supposedly inoffensive physical insult, such as lengthy standing on the hind limbs, produces as serious an effect as osteonecrosis of the femoral head. The clinical counterpart of this experimental model is the case of bilateral talar osteonecrosis in an obsessive-compulsive, anorexic woman who stood all day long, never allowing herself to sit down at any activity. The prolonged and excessive biomechanical strain coincided with the patient's osteoporosis to cause osteonecrosis. ⁴⁰ , ⁴¹

The Role of Apoptosis in Experimental Osteonecrosis

Sato and his colleagues advise us to continue using “the term necrosis to describe findings comprising dead cells in histologic sections, regardless of the pathway in which the cells died.” ¹⁷ Apoptosis, i.e., programmed cell death, plays a role in causing osteonecrosis. Oxygen deprivation modifies the expression of the stress protein genes. Apoptotic bodies and DNA fragmentations are observed in the osteocytes and marrow cells of ischemically necrotic femoral heads of rats. The presence of osteocytes with optically empty lacunae in addition to terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL)–positive cells testifies to coexistent bone death of the commonplace and apoptotic types. In humans with idiopathic femoral capital necrosis, apoptotic osteocytes and osteoblasts abound near the subchondral fracture. This contrasts with the absence or scarcity of apoptotic cells in patients with osteonecrosis due to sickle cell disease, trauma, or alcoholism. Apoptotic osteocytes and osteoblasts amass in the bone of prednisolone-treated mice as a result of the cumulative and irreparable disruption of the mechanosensory function of the osteocytic network, starting a chain of events leading to collapse of the femoral head. The bone loss in the methylprednisolone-treated rabbits, histologically obvious before the time at which femoral capital necrosis is first discernible, is more than just a consequence of decreasing synthesis. It is also due to an increased breakdown of the osseous matrix, which correlates with apoptosis of the osteocytes and osteoblasts. In comparison, there are few TUNEL-labeled osteoblasts and no TUNEL-labeled osteocytes in the healthy bone. ¹⁴ , ⁷¹ , ⁷²

Concluding Remarks

Idiopathic osteonecrosis, formerly the most common variant of the disease, is being overtaken by the secondary types. Traumatic and iatrogenic osteonecrosis receives ever more attention with overaging of the aged population, prevalence of traffic accidents, and aggressive therapeutic modalities. Presently, for example, 13% of patients undergoing up-to-date therapy for systemic lupus erythematosus will suffer osteonecrosis sometime during the course of their disease. ¹⁷

Osteonecrosis of patients and animals is akin in some aspects but diverges in others. Femoral heads of small laboratory animals, researchers' favored subjects of experimentation, were totally necrotic after vascular deprivation. Viable tissue replaces the necrotic osseous tissues within a few weeks. A smaller or larger part of, but rarely the entire, femoral head undergoes necrosis in man. The necrotic bony trabeculae, persisting in situ for many years, are partly covered by newly formed appositional bone and separated from each other by a poorly organized fibrous tissue within which intramembranous osteogenesis expresses an attempt at repair.

Analysis of experimental models facilitates the assessment of new treatment modalities. Healing of vascular-deprived necrotic femoral heads is expedited by an exposure of the rats to hyperbaric oxygen. The hyperoxygenation-mediated relief of the ischemia apparently boosts fibroblastic, angioblastic, osteoblastic, and osteoclastic activities. Non–weight bearing further encourages repair processes affected by the hyperbaric oxygenation. ³³ , ⁵² Injections of osteogenic protein-1 augments the healing activities in dogs undergoing the trapdoor procedure, i.e., excavation of all necrotic tissues followed by bone autografting. ⁴³

Cruess described the complex nature of osteonecrosis as microembolic events, vascular impediment, adipocytic hypertrophy, and lipid-incurred osteocytic death, alone or in combination, playing etiopathogenic roles of unequal magnitudes under different circumstances. ¹⁰ Any deduction based on the results in models are burdened by the fact that interventions are conducted in healthy animals having bones and muscles of excellent quality. As comparison, osteonecrosis at the bedside frequently involves patients with diverse musculoskeletal or systemic disease. Hsieh and coauthors rightfully reason that to understand the pathophysiology of any disease, one must reproduce it in an animal model, where it can be monitored throughout its course. ²²