Portal de Periódicos da CAPES

Alcoholism, Hypercortisonism, Fat Embolism and Osseous Avascular Necrosis

2001; Lippincott Williams & Wilkins; Volume: 393; Linguagem: Inglês

10.1097/00003086-200112000-00002

1528-1132

John Paul Jones, Leonard F. Peltier,

Diabetic Foot Ulcer Assessment and Management

John Paul Jones, Jr. (Fig 1) was born in Oakland, CA in October 1937. After his preliminary basic education, he attended the School of Public Health of the University of California in Berkeley. While in college Jones spent several summers working in the Division of Research of the California State Health Department. This work led to his coauthorship of several scientific papers. After receiving his Bachelor’s degree in 1959, he attended the School of Medicine of the University of California in San Francisco from which he was granted his medical degree in 1963. In his senior year he was elected to membership into Alpha Omega Alpha, the medical honor society. His internship at the San Francisco General Hospital was followed by a 2-year fellowship in Rheumatology in the Department of Medicine. From 1966 to 1969, Jones was a resident in the Department of Orthopaedic Surgery at the University of California, San Francisco. He then entered private practice in Lake County, CA.Fig 1.: Dr. John Paul Jones, Jr.Jones became a member of the Orthopaedic Research Society, the Osteoarthritis Research Society, and the Undersea and Hyperbaric Medical Society to all of which he was an active contributor. He spent many years working with others to perfect a total hip arthroplasty for dogs. The major work of Jones focused on the problems of osteonecrosis. His first papers on the subject were published while he was still a resident. Much of the research has been done in the Diagnostic Osteonecrosis Center and Research Foundation of which he has been president since 1974. He has been active in the National Osteonecrosis Foundation of which he currently is vice president, and the Association Research Circulation Osseous of which he also has been president. The classic paper 1 published in 1971 is a good account of Jones’s concepts and findings regarding the etiology of osteonecrosis; his ideas and concepts have stood the test of time. Avascular necrosis results from an interruption of blood supply to bone. Mechanisms responsible for osseous ischemia may include changes in functional requirements (vasospasm, hypotension, hypoxia or a sudden increase in tissue metabolic demands), or, more probably, mechanical disturbances (sudden and complete interruption or occlusion of vessels by traumatic severance, compression, obliterative endarteritis, thrombosis or embolism). Evidence demonstrates that the extent and location of such circulatory disturbance determines the magnitude of the resulting osseous necrosis. Clinically, about 10% of posterior hip dislocations are complicated by avascular necrosis of the femoral head; the mechanism of physical interference, principally with the extraosseous blood supply, is fairly obvious. However, in approximately 35% of displaced fractures of the femoral neck, the proximal fragment eventually becomes necrotic, presumably because the intraosseous circulation, in addition, has been disrupted. In recent experiments, the nutrient artery of the rabbit femur has been found to contribute more than one-third of the total blood flow to the upper femoral epiphysis and metaphysis; in addition, it is known that interruption of both extraosseous and intraosseous vessels to the femoral head of rabbits is more likely to cause necrosis than severance or ligation of extraosseous vessels alone. But in nontraumatic avascular necrosis the precise mechanism(s) and location of vascular interruption (intraosseous and/or extraosseous) remain a mystery, even though over 30 years ago Kistler wrote: “The destruction of an artery to a bone does not duplicate embolism of its branches or end-capillaries within the bone. The large vessels of the medullary tissues may become channels for collateral circulation from another external source. Therefore, a few small emboli may produce greater injury to osteoid tissue than severe trauma. ” HYPOTHESIS In order to construct a working hypothesis, various physiological and metabolic abnormalities, not previously considered in this light, were investigated to determine whether a common etiologic denominator existed in nontraumatic avascular necrosis. Sufficient circumstantial evidence was accumulated to form a hypothesis implicating continuous or intermittent fat embolism in the bones of adults as a possible initiating event in the evolution and pathogenesis of nontraumatic avascular necrosis. Fat embolism has been observed in alcoholism, in hypercortisonism, during crises in the sickle-cell hemoglobinopathies (especially SC disease), and in decompression syndromes and pancreatitis, each of which may be complicated by avascular necrosis. The origin of these fat emboli is generally considered to be the fatty bone marrow or subcutaneous or deep adipose tissue depots. It is not generally recognized that under certain circumstances the liver may serve as a major fat depot. In patients with severe hepatic steatosis or acute fatty infiltration or metamorphosis the liver may be composed principally of neutral fat and may weigh more than 5000 gm (which is about 20 times the average total amount of fat capable of being extracted from the adult human tibia and femur combined). Blunt trauma to the liver, whether or not ribs have been fractured, has been known to cause fat embolism since the clinical report by Zenker in 1861. Obviously this hazard is increased if the liver is enlarged. Owens and Northington experimentally confirmed in vivo that liver lipid is an excellent source of embolic fat following blunt trauma to the right upper quadrant of the abdomen. The number of emboli was increased when frank hepatic necrosis was present. If for any reason the oxygen supply to hepatic cells becomes inadequate, those parenchymal, and potentially fat-laden, cells adjacent to central veins will be the first to become necrotic, since they are farthest away from the source of oxygenated blood. Also, certain medications have been associated with massive hepatic necrosis, including halothane (Fluothane) anesthesia and isonicotinic acid hydrazide (Isoniazid), and it is speculated that concurrent, prolonged, or excessive administration of these agents may contribute to the release of fat emboli. Both clinical and experimental studies have shown that the fatty liver is capable of spontaneously releasing large numbers of embolic-sized fat globules into hepatic venous channels after rupture of fatty cysts into adjacent sinusoids and central veins. Precise quantitation of the amount of this potentially embolic fat per unit time has not been established either clinically or experimentally in vivo, although hepatic perfusion studies in vitro indicate that the magnitude of embolism is directly related to the degree of intracellular fat present at the beginning of the perfusion. Furthermore, the liver is more heavily perfused by circulating blood volume than is any extremity, which should facilitate more extensive drainage of fat globules through the hepatic venous outflow tracts. Moreover, pancreatic enzymes, upon entering portal venous radicles in patients with pancreatitis, could conceivably cause further release of intravascular fat from fat-laden hepatic cells. After penetrating the pulmonary vascular bed, systemic fat emboli enter vessels supplying the brain, kidneys and all other organs and tissues of the body, including those parts of the skeleton which are most vulnerable to avascular necrosis because in general they derive their nourishment from only a few primary vessels–principally terminal arteries with relatively poor collateral circulation. CLINICAL OBSERVATIONS Alcoholism Axhausen probably was the first to report avascular necrosis in an alcoholic patient. More recently, Vignon et al reported that five of their nine patients with idiopathic necrosis had significant alcoholism. Malka noted that three of his six patients with necrosis of the femoral head were alcoholics. Serre and Simon and Patterson et al, found that 19% and 17% of their respective series of patients with avascular necroses had a significant alcoholic history. Since 1963 we have explored the concept that osseous avascular necrosis may be specifically associated with alcoholism, since the alcohol-induced fatty liver is probably the commonest source of continuous, low-grade and relatively asymptomatic (unless delirium tremens is related) showers of systemic fat emboli. Thirty patients who had idiopathic avascular necrosis, all of whom experienced prolonged, excessive alcohol consumption, were studied. Avascular necrosis was diagnosed on the basis of roentgenograms in all 30 patients and confirmed histologically in 19. There was no evidence of major antecedent trauma or any known systemic condition usually associated with nontraumatic avascular necrosis. Eighteen patients were Caucasians, 11 were Negro and one was Chinese; 24 were male and six were female; age range was 24–68 years (mean 48.6 years). This is the same age and sex distribution as those alcoholics with fatty nutritional cirrhosis in a larger series. One other patient (Case 1), who did not show evidence of osseous avascular necrosis, was studied. Several of these patients experienced delirium tremens, unexplained pneumonitis and hypertension. Twenty-seven of 31 patients studied had significant hepatomegaly and 22 of 24 patients showed abnormal bromsulphalein retention. Six of a sample of eight had hypertriglyceridemia, and, in seven of nine livers examined, fatty metamorphosis was demonstrated histologically. Evidence of systemic fat embolism was found in nine of 19 patients: Lipuria was detected in eight cases, probable intravascular fat globules in resected femoral heads were seen in two instances, and emboli were seen at autopsy in one case. Seventy-seven lesions were found in these 30 patients. There was marked bilateral symmetry. Fifty-six percent of the necrotic lesions were present in the femoral head, the most frequent epiphyseal region to be involved, and 11% of the lesions were located in the humeral head. Metadiaphyseal (shaft and neck) infarctions within the distal femur and proximal tibia accounted for nearly one-third of the lesions. Patients may have symptoms for six months or more before developing the first roentgenographic manifestations of necrosis. Although lesions of epiphyseal regions often produce pain and stiffness, metadiaphyseal lesions are not disabling since no cortical or periosteal involvement is present. There is no evidence that extension of pre-existing intramedullary infarctions will cause arthritis. The insidious initial symptom is usually pain in the groin with intermittent radiation down the anteromedial thigh which usually increases during weight-bearing. Progressive stiffness subsequently develops which limits hip motion, particularly flexion, abduction and internal rotation. Although early lesions in the humeral head are often asymptomatic, aching shoulder pain later supervenes, often radiating to the deltoid tuberosity of the humerus. Virchow believed that investigation of a disease should be concentrated on the early dynamic process rather than on the end result. If loss of osseous vascularity is sufficient to cause necrosis, the roentgenographic and histological appearances of the reparative processes are similar regardless of the etiological mechanism(s), which has probably disappeared in the interval. Therefore, those patients manifesting the earliest evidence of necrosis are presented. The following clinical observations are summarized and, we believe, represent the sequence of early pathophysiological events. Case 1. A 32-year-old Caucasian male alcoholic, while hospitalized died suddenly and unexpectedly. No resuscitative attempt was made. At autopsy the fatty liver weighed 4,600 gm, approximately three times the normal weight. Sections, stained with Oil Red O fat stain, demonstrated liquid fat globules escaping from centrilobular fatty hepatic cysts. Multiple fat globules were found entering the hepatic sinusoidal and central venous circulation. Embolic-sized fat globules were also lying within hepatic veins en route to the pulmonary circulatory system. Multiple deformed fat emboli occluded pulmonary arteries and alveolar capillaries with resulting edema and microhemorrhages. Use of our bone fat-staining technique isolated fat globules, probably intravascular, in subchondral capillaries of the right femoral head. However, in this case there was no gross, roentgenographic or microscopic evidence of avascular necrosis. Case 2. A 56-year-old, Caucasian male alcoholic also died suddenly and unexpectedly while hospitalized. No resuscitative attempt was made. Sections taken at autopsy showed multiple fat globules within the central and hepatic veins of the 2600 gm fatty liver. There was also evidence of renal and pulmonary fat embolism. Although he had reported no symptoms, a wedge-shaped zone with sclerotic demarcation was found in the anterosuperior quadrant of the right femoral head; there was no gross disruption of structural integrity. This segmental subchondral lesion was more obvious on hematoxylin and eosin-stained sections. There was definite avascular necrosis with trabecular lacunae devoid of osteocytes and with partially calcified fibrous tissue within the interstices of the dead bone. Deformed fat globules were found in subchondral haversian canals adjacent to the infarcted region. In addition, serpentine metadiaphyseal infarctions were found in the right proximal tibia and in the right distal femur, localized entirely within metaphyseal cancellous bone. Linear strands of calcification were present at the fibrous tissue margins. Necrotic trabeculae, amorphous débris and replacement of the marrow by fibrous tissue were also present. Case 3. A 42-year-old Negro male alcoholic, without evidence of sickle-cell hemoglobinopathy, had hepatomegaly, a biopsy-proven fatty liver, hypertriglyceridemia and lipuria. He developed pain and progressive stiffness of the right hip. Although the liver was grossly enlarged, extending below the right iliac crest, roentgenograms of the femoral heads appeared normal. Twenty months later, tomography delineated a segmental, funnel-shaped lesion in the right femoral head. Intraosseous phlebography, differential radioactive phosphorus-32 uptake, differential manometry and differential oximetry studies suggested segmental avascularity, which was confirmed by biopsy. Necrotic marrow trabeculae and ischemic-appearing appositional bone were also present. In this patient we attempted to revascularize the femoral head early, before it had undergone irreversible fragmentation and architectural collapse, by excising the segmental necrotic zone and performing a musculoosseous pedicle transfer. This procedure was suggested by R. M. Jameson, M.D., and is a modification of that described by Judet for osteosynthesis of femoral neck fractures: Viable cancellous bone from the intertrochanteric crest is nourished by vessels supplying the quadratus femoris muscle. Insufficient time has elapsed to determine whether this operation has been successful. Hypercortisonism Avascular necrosis associated with prolonged corticosteroid therapy was first described in 1957 by Pietrogrande and Mastromarino, and was confirmed by Block-Michel et al. Necrosis occurring after renal homotransplantation was first reported in 1965, and since then additional cases have been reported. Bravo et al reported on five patients who developed avascular necrosis after renal homotransplantation and treatment with corticosteroids. Autopsy findings on one of these patients included fatty infiltration of hepatic parenchymal cells and fat embolism of the lungs and kidneys. These investigators also consistently found intra-and extracellular lipid-staining particles in the synovial fluid and postulated that phagocytosis of these lipid components may facilitate lysosomal enzyme release with resultant synovitis. Since 1963, 32 patients in our series, all of whom had been treated with corticoids for various diseases, developed osseous avascular necrosis which was diagnosed roentgenographically in all cases and confirmed histologically in 12. Nine individuals had received high doses of corticoids for immunosuppression after renal homotransplantation. Thirteen patients were male and 19 were female. The average (mean) age of the patients when necrosis was diagnosed was, for the transplant group 21.3 years (14–30 years) and for the non-transplant group, 43.1 years (17–68 years). In no patient was there evidence of major antecedent trauma or any known systemic condition usually associated with nontraumatic avascular necrosis. Efforts were made to obtain presumptive evidence of systemic fat embolism in 11 patients who had avascular necrosis; such evidence was detected in six of these. Fat globules were consistently found in the urine in five patients, in two of whom renal glomerular fat was demonstrated by biopsy, and two patients appeared to have intravascular fat globules in their necrotic femoral heads. Seventy-one lesions were found in these 32 patients. The distribution of lesions was similar to that in the alcoholic group and multiple bones were involved with symmetrical bilaterality. One patient, who underwent two renal homotransplantations, developed 11 lesions. Approximately one-fourth of the lesions were in non-weight-bearing bones and about 80% affected proximal epiphyseal regions of bones. The distal femur (particularly the lateral femoral condyle), the body of the talus, the carpal scaphoid and the humeral capitellum were less commonly affected. In this series, no tibial lesions were seen and there were no metadiaphyseal bone infarctions. The following clinical observations are summarized and, we believe, represent a spectrum of early pathophysiological events. Case 4. A 25-year-old Caucasian female, who had Cushing’s syndrome, developed severe pain bilaterally in the medial thighs and in the knees three months after renal homotransplantation. Eight months after operation, corticoids were discontinued; she experienced right upper quadrant pain and within a week developed clinical manifestations suggesting systemic fat embolism. The patient died nine months after transplantation; routine autopsy did not reveal the cause of death. Specimens taken at autopsy showed large numbers of fat globules within alveolar capillaries. Fat emboli were detected in multiple glomeruli as well as within myocardial and cerebral vessels. The liver parenchyma contained large quantities of fat, especially concentrated within the centrilobular zones adjacent to central veins, in the form of large fatty cysts, some of which measured 75 to 100 micra in diameter. Despite hypercortisonism and systemic fat embolism there was no radiographic evidence of avascular necrosis in this patient; therefore this patient was not included in the 32 cases analyzed above. Case 5. A 24-year-old Negro female, with no evidence of hemoglobinopathy, had received high-dosage corticosteroids since 1961 to prevent the bullous eruption of lichen planus. She had experienced multiple episodes of probable systemic fat embolism, characterized by episodic, unexplained fever, tachycardia, cough, dyspnea, chest pain, visual blurring and microscopic hematuria. Hypertriglyceridemia, elevated serum lipase and urinary fat globules had been found consistently since September 1964 during episodes of right upper abdominal pain, hepatomegaly and bromsulphalein retention. However, systemic fat embolism was not confirmed histologically on examination of a biopsy specimen from the kidney. She first experienced pain in the right groin in October 1965, but routine roentgenograms were negative. The pain continued and two months later tomograms of the right femoral head revealed a definite early cone-shaped translucent lesion surrounded by an irregular sclerotic margin with its apex pointing toward the center of the obliterated epiphyseal line. Its base abutted the anterosuperior articular cartilage and involved the subchondral bone. Ancillary studies were then performed to assess the viability of the tissue. A differential phosphorus-32 uptake study showed hypervascularity near the apex of the lesion, and this finding was confirmed by normal manometry, oximetry and intraosseous phlebography. A trocar biopsy specimen obtained near the apex of the wedge showed definite avascular necrosis with lacunae devoid of osteocytes. However, there was considerable evidence of regeneration and reossification with appositional new bone deposited on the surfaces of dead bone trabeculae as entombing cocoons. The hypervascularity was probably due to granulation tissue invading the interstices of the dead bone, generating osteoblasts to resurface the dead trabeculae and resorbing necrotic bone and marrow débris with the aid of osteoclasts and histiocytes. Case 6. A 20-year-old Caucasian female, treated with high dosage corticosteroide (sic) following renal homotransplantation, sustained two episodes of probable systemic fat embolism, characterized by unexplained focal glomerular microinfarctions four months after transplantation, and by fever, tachycardia, cough, hypotension, and electrocardiographic abnormalities six months after transplantation. A roentgenogram of her left hip had been normal four months after transplantation. She first experienced pain in the left medial thigh and groin eight months after transplantation. Two months later there was definite roentgenographic evidence of avascular necrosis involving the left femoral head, with sclerosis and flattening. Tomograms revealed fragmentation, cystic rarefaction and segmental collapse of the femoral head. Lipuria was present and a biopsy of the transplanted kidney revealed fat globules, within glomeruli, which were considered evidence of systemic fat embolism. In summary, the last two patients manifested concurrently evidence of hypercortisonism, systemic fat embolism and avascular necrosis. EXPERIMENTAL OBSERVATIONS To investigate the significance of these clinical findings and to provide further verification of the existence and potential etiologic significance of presumed intravascular fat emboli in human surgical and autopsy specimens, a technique using an Oil Red O fat stain and a reticular fiber counterstain has been developed. While prelabeled exogenous or endogenous fat can be easily identified within intraosseous vessels and differentiated from marrow fat, it is difficult to distinguish unlabeled autogenous intravascular fat from extravascular fat. The ultrastructure and precise location of collagen (reticular) fibers surrounding terminal arterioles and capillaries have been demonstrated by electron microscopy. These fibers are usually found as a narrow sheath lying between the thin basement membrane of the endothelial lining cells and the perivascular cells (pericytes). The walls of intraosseous vessels may be more easily identified by reticular fiber staining. When a fat globule is found encased by a reticulin-stained vessel wall, the globule is probably intravascular, though it may or may not be an embolus. If thin sectioning (15–25 micra) is performed, there is less “intravascular” superimposition of perivascular adipose cells or lipids normally occurring in compact bone. To prevent artifactitious marrow fat flow-over into adjacent haversian canals, it is necessary to float frozen sections onto ice-cold slides. Moreover, adipose cellular disruption, pressure oozing and overflow into intertrabecular spaces and haversian canals must be prevented by placing cover slips gently on slides without the use of weights. Evidence that a fat globule is both intravascular and embolic is supported by deformation of such a globule into an oval, cylindrical or bullet-shaped configuration, which indicates intravascular penetration and terminal impaction. Small spherical fat particles lying within haversian vessels are either nonembolic-sized globules (usually located proximal to an occluding embolus), or superimposed extravascular adipose cells or artifactitious free-floating fat. Those deformed globules found lying in the geometric plane of the section within reticulin-stained capillaries of smaller haversian canals in regions of trabecular or cortical bone removed from marrow spaces are considered to be fat emboli. Animal Experiments In previous experiments, fat embolism of bone was demonstrated for the first time. A preparation of iodized poppyseed oil (Lipiodol, Fougera & Co.) was infused into the distal aorta of rabbits. Intraosseous fat emboli were demonstrable histologically during a five-week period after the infusion, by which time they had especially obstructed subchondral arterioles and capillaries of the femoral head. Preliminary studies indicated that at the end of five weeks the initiating event (intraosseous fat embolism) was vanishing, and histologic and radioautographic evidence of avascular necrosis had appeared, although antecedent ischemic manifestations had been apparent in animals killed earlier. Focal anemic infarctions were found only in metaphyseal and epiphyseal regions of the femoral heads. These lesions were characterized by disorganization of all cellular elements (hematopoietic, adipose and osteogenetic), liquefaction fat necrosis, segmental loss of tissue pattern with coagulation necrosis, and the presence of amorphous débris. Separation along cementing lines resulted in trabecular microcracks and sequestration into marrow spaces of many small fragments of dead bone; it is interesting that the necrosis in our animals appeared similar histologically to a lesion of epiphyseal ischemic necrosis reported by Aegerter and Kirkpatrick. Although bone marrow components had completely regenerated by 17 weeks, focal osteocytic death persisted. Discussion The extent to which these preliminary observations on rabbits may apply to man is not known, but certainly the underlying principles, although highly speculative, are important since they provide a basis for postulation of a theoretical mechanism whereby intraosseous fat embolism may result in necrosis of a secondary haversian system in the human. It is emphasized that experimental division of an extraosseous vessel does not, except in isolated instances, simulate occlusion of the intraosseous endings of that vessel by emboli. Osteocytic lacunae are distributed in a cylindrical pattern around parent central vessels and they often extend to the outermost cementing line. The enveloping cementing line apparently represents a structural barrier between the outermost ring of osteocytes and adjacent cells located in the interstitial bone, and the canaliculi of these two areas usually do not communicate. The passage of nutrition and oxygen from the haversian vessel into the surrounding interhaversian bone is prevented, and therefore it is normal for osteocytes to die in locations between haversian cylinders. However, if circulation through a central and terminal haversian capillary is interrupted by embolic occlusion of its lumen (Fig. 2b), then abnormal osteocytic necrosis occurs within range of the cylinder itself, since there is no collateral circulation beyond the point of occlusion. Frost found that microscopic cracks appear predominantly in areas of occluded canaliculi, or micropetrosis, and suggested that this bone is more brittle. With anoxia of the osteon it is conceivable that there is disruption of intralamellar bonds along cementing lines between parallel layers of bone tissue, due to microfatigue in the absence of repair, which may account for the microscopic cracks, which extend to trabecular surfaces as cleavage planes and result in trabecular thinning, with sequestration into the intertrabecular spaces of small spicules of dead bone.Fig 2B.: Diagrammatic representation of hypothetical mechanism whereby fat embolism of a terminal intraosseous capillary may result in avascular necrosis of osteocytes in a haversian system, causing microfractures of superficial lamellae, and, inevitably, thin, fragile and dead trabeculae.Embolism was first proposed as a mechanism producing avascular necrosis of bone by Axhausen in 1909, whereas Rieger first suggested that traumatic fat embolism was the cause of osteochondritis dissecans. Kahlstrom et al observed two patients who had previously sustained multiple fractures and who subsequently experienced idiopathic bone infarctions. These investigators also considered that unrecognized traumatic fat embolism was a possible mechanism in the pathogenesis of osseous avascular necrosis. It is interesting that one patient in our overall series, a 51-year-old Caucasian man, was involved in an automobile accident in August 1946, and was hospitalized with shock, rupture of liver, fractured second and third ribs on the right with perforation of the right lung, and lacerations of the forehead and right calf. Eight months later, he was again hospitalized after a second automobile accident. He had a fracture of the left tibial plateau, laceration of the right heel and a ruptured spleen, which required a splenectomy. Abdominal roentgenograms revealed enlargement of the liver; however, tests to determine possible systemic fat embolism were apparently not performed during either hospitalization. The patient denied chronic alcoholism and liver function tests were normal. He could remember no prior trauma to his right hip, he had no history of any disease associated with nontraumatic osseous avascular necrosis, and he has had no pain or stiffness in the right hip. In March 1967, roentgenograms, taken for another purpose, revealed idiopathic segmental avascular necrosis of the right femoral head. Tomograms demonstrated no structural collapse. An intraosseous phlebogram revealed immediate normal drainage from the periphery of the head and neck; 15-minute films suggested residual Hypaque ® sequestered in the avascular-appearing segment. Avascular necrosis was confirmed histologically by inspection of an open needle biopsy specimen. However, to my knowledge, avascular necrosis has not yet been reported in humans after a single episode of documented traumatic fat embolism, although it is likely that, if the patient survives, most small infarctions, if present, are entirely asymptomatic and are completely resorbed and replaced by healthy cancellous tissue. Experimentally, one episode of fat emboli, even with temporary systemic recycling of the fat globules, does not result in gross or roentgenographic evidence of necrosis. But if vulnerable bones were bombarded with multiple or continuous showers of fat emboli over periods of several months or years, as is potentially the case in alcoholism and hypercortisonism, as well as in chronic, recurrent pancreatitis, episodic sickle-cell crises and periodic decompression sickness, it is conceivable that healing would be prevented by further focal infarctions which may result in obvious irreversible osseous avascular necrosis.