Cerebral Fat Embolism Syndrome

Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

1301NEUROLOGIC/HEAD AND NECK IMAGING

Wende N. Gibbs, MD, MA Michael J. Opatowsky, MD, MBA Elizabeth C. Burton, MD

HistoryA 57-year-old African American woman with a history of sickle cell b-thalas-semia presented to the hospital with a 3-day history of chest and back pain unresponsive to oxygen and analgesic therapy. Her medical history included multiple painful crises due to sickle cellinduced vaso-occlusion, multiple blood transfusions, hypertension, and diabetes mellitus. The evening of her admission, she developed a fever and hypotension, and a sepsis evaluation was initiated. She was treated with empirical antimicrobial agents and received oxygen, a red blood cell transfusion, and medications for pain. Two nights later, she became mildly confused. Unenhanced computed tomography (CT) of the head was performed. The following morning, magnetic resonance (MR) imaging of the brain and MR angiography of the head and neck were performed. Later that day, the patients hypotension worsened, and she became difficult to arouse. Intubation was performed to protect her airway, and she was transferred to the intensive care unit. Thrombocytopenia was diagnosed, and the prophylactic an-tithrombotic therapy initiated at admission was temporarily withheld. Multiple studies for infectious agents returned negative results. The patients status con-tinued to deteriorate with the development of respiratory failure, renal failure, and ischemic hepatitis (shock liver). Sedation was discontinued, and 48 hours later, with the patient still unresponsive, comfort measures were instituted. The patient died on the 10th day of her hospitalization.

Imaging FindingsUnenhanced CT of the head performed at the onset of the patients mental status changes showed subtle sulcal effacement, abnormally small ventricles, and mildly narrowed basal cisterns. These findings were suggestive of early-stage cerebral edema (Fig 1). There was no evidence of acute hemorrhage or infarction at CT.

MR imaging of the brain and MR angiography of the head and neck per-formed the following morning showed innumerable punctate foci of restricted diffusion throughout the brain, including the cortex and subcortical white matter of the cerebral hemispheres and cerebellum, corona radiata, internal capsules, caudate nuclei, thalamus, middle cerebellar peduncles, and body and splenium of the corpus callosum (Fig 2a2c). Susceptibility-weighted images

AIRP Best Cases in Radiologic-Pathologic CorrelationCerebral Fat Embolism Syndrome in Sickle Cell -Thalassemia1

RadioGraphics 2012; 32:13011306 Published online 10.1148/rg.325115055 Content Codes: 1From the Department of Radiology, Baylor University Medical Center, 3500 Gaston Ave, Dallas, TX 75246 (W.N.G., M.J.O.); and Department of Pathology, Johns Hopkins Hospital, Baltimore, Md (E.C.B.). Received March 16, 2011; revision requested May 20 and received June 28; accepted July 29. All authors have no financial relationships to disclose. Address correspondence to W.N.G., Barrow Neurological Institute, 350 W Thomas Rd, Phoenix, AZ 85013 (e-mail: [email protected]).

RSNA, 2012

EDITORS NOTEEveryone who has taken the course in radiologic pathology at the Armed Forces Institute of Path- ology (AFIP) remembers bringing beautifully illus- trated cases for accession to the Institute. The long-standing and ex-cellent AFIP course in radiologic pathology has transitioned under the auspices of the American College of Radiology to a new home in Silver Spring, Md, entitled the American Institute for Radiologic Pathology (AIRP). In recent years, the staff of the Institute has judged the courses best cases by organ system, and recognition is given to the winners on the last day of the class. With each issue of RadioGraphics, one or more of these cases are published, written by the winning resident. Radio- logic-pathologic corre-lation is emphasized, and the causes of the imaging signs of various diseases are illustrated.

1302 September-October 2012 radiographics.rsna.org

Figures 1, 2. (1) Axial unenhanced CT image demon-strates ventricles and sulci that are smaller than normal for a patient of this age, suggesting mild cerebral edema. There was no evidence of infarction or hemorrhage in this study. (2) Axial diffusion-weighted MR images (b at a higher level than a) and ADC map (c; at the same level as a) show innumerable punctate foci representing restricted diffusion due to cytotoxic edema and emboli throughout the brain. This is the starfield pattern that characterizes cerebral fat embolism syndrome, so called because of its resemblance to a starry sky at night (d). (Fig 2d courtesy of George Wells.)

RG Volume32 Number5 Gibbsetal 1303

demonstrated small focal regions of susceptibility artifact representing hemorrhage or hemosiderin deposition throughout the gray- and white-matter structures of the brain, brainstem, and cerebel-lum (Fig 3). T2-weighted images revealed abnor-mal foci of signal hyperintensity in corresponding regions. No vascular abnormalities were detected at MR angiography of the head and neck.

Pathologic EvaluationAutopsy findings included vertebral bone marrow infarcts and necrosis with associated fat emboli within multiple organs, including the brain, lungs, kidneys, and liver. Macroscopic examination of the

Figure 3. Axial susceptibility-weighted MR images obtained at the levels of the lateral ventricles (a), cerebellum (b), and corona radiata (c) show multiple foci of susceptibility artifact due to microhemorrhages in both gray- and white-matter structures. In a, note the involvement of the internal capsules (ar-row) and the splenium of the corpus callosum (arrowhead). These are unusual sites of mi-crohemorrhage in the absence of trauma.

brain showed cerebral edema without evidence of herniation, and diffuse microhemorrhages involv-ing both gray- and white-matter structures, includ-ing the cerebral and cerebellar cortex and subcor-tical white matter, corpus callosum, internal cap-sules, thalamus, basal ganglia, and brainstem (Fig 4a, 4b). Microscopic sections of the brain, brain-stem, and cerebellum showed diffuse perivascular hemorrhages, as well as multiple microinfarcts consisting of pallor, neuronal loss, apoptosis, and axonal spheroids. Fat globules and sickle-shaped red blood cells were present within the microvas-culature (Fig 4c). The diffuse microhemorrhages corresponded to the numerous foci of susceptibil-ity artifact and restricted diffusion seen at MR imaging (Figs 2, 3). Evidence of end organ dam-age seen at autopsy included acute renal tubular necrosis, resolving ischemic hepatitis, and vascular thromboses involving multiple organs.

1304 September-October 2012 radiographics.rsna.org

Figure 4. (a) Autopsy photograph of a coronal brain section shows diffuse microhemorrhages throughout the gray and white matter. Microhemorrhages in the corpus callosum (black arrowhead), basal ganglia (arrow), and cortex (white arrowhead) correspond to foci of susceptibility artifact depicted in Figure 3a. (Scale is in centi-meters.) (b) Autopsy photograph of axial sections through the cerebellum (top), pons (middle), and midbrain (bottom) show additional diffuse microhemorrhages. The cerebellar microhemorrhages correspond to foci of susceptibility artifact in Figure 3b. (Scale is in centimeters.) (c) High-power photomicrograph (original magnifi-cation, 100; osmium tetroxide stain) of a histologic slice from the corpus callosum shows black osmium staining of multiple fat emboli within microvessels (arrow).

DiscussionSickle cell disease is a hereditary hemoglobin-opathy caused by mutations in the b-globin gene. The homozygous sickle cell hemoglobin (Hb S) mutation is the most common genetic abnormal-ity found in patients with sickle cell disease. Sickle cell b-thalassemia results from a heterozygous Hb S mutation that causes a reduction in, or absence of, the synthesis of b-globin chains. This condition may be phenotypically indistinguishable from sick-le cell anemia. These hemoglobin gene mutations cause abnormal hemoglobin polymerization at low oxygen tension levels, which leads to increased density of red blood cells and cell membrane dam-age with resultant rigidity and deformation. These changes result in the premature destruction of red blood cells (hemolysis) and vaso-occlusion. Intra-vascular hemolysis is responsible for endothelial injury, which may lead to coagulopathy, vasomo-

tor instability, and proliferative vasculopathy, with subsequent pulmonary hypertension. Vaso-occlu-sion often results in tissue ischemia and infarction. Patients with sickle cell disease commonly experi-ence crises of acute pain due to vaso-occlusion, which may occur in any part of the body. These vaso-occlusive crises can lead to localized or gen-eralized bone marrow necrosis, bone infarction, and avascular necrosis (1).

Fat embolism syndrome is a rare but poten-tially lethal complication of sickle cell disease that is not widely recognized (2). Fat embolism syndrome more commonly occurs as a complica-tion of trauma, especially in fractures of the long bones (3). In the setting of sickle cell disease, the syndrome is caused by bone marrow infarcts and necrosis, with subsequent embolization of fat to multiple organs. Notably, not all patients with bone marrow necrosis and fat embolism progress to fat embolism syndrome. The diagnosis is based on clinical manifestations, including progressive

RG Volume32 Number5 Gibbsetal 1305

respiratory distress, cerebral involvement, and cutaneous petechiae (2,4). Secondary diagnostic criteria include tachycardia, fever, anemia, and thrombocytopenia (5). The initial symptoms are pulmonary, ranging from dyspnea and severe hy-poxemia to acute respiratory distress syndrome. Neurologic symptoms follow, including altera-tion in the level of consciousness, seizures, focal neurologic deficits, and coma (5). A recent review of 24 cases of fat embolism syndrome in patients with sickle cell disease included four patients with sickle cell b-thalassemia. All 24 patients (100%) experienced respiratory distress, 50% had cuta-neous petechiae, 75% experienced lethargy, and 75% had central nervous system involvement (2).

The pathogenesis of fat embolism syndrome remains controversial. Bone marrow infarction is common in patients with sickle cell disease and is associated with acute chest syndrome (6). Fat emboli from bone marrow necrosis are thought to enter osseous venous channels and then the ve-nous circulation. Ultimately the fat emboli occlude capillaries and small arteries of the end organs. Biochemical mechanisms of injury also have been proposed by which the fat emboli produce local ischemia and inflammation, with release of inflam-matory mediators and vasoactive amines, and platelet aggregation. In addition, hydrolysis of free fatty acids produces toxic intermediates that dam-age capillary endothelium. The cascade of endo-thelial damage, alveolar injury, increased capillary permeability, and damaged lung surfactant can lead to acute respiratory distress syndrome (7,8).

Neurologic dysfunction in the setting of fat embolism syndrome may result directly from ves-sel occlusion by fat emboli, from disruption of the blood-brain barrier by toxic free fatty acids, or both (2,9). Additional factors, such as hypoxia, hypotension, and a systemic inflammatory re-sponse, likely contribute to neurologic manifesta-tions of the syndrome (10). Fat emboli may reach the brain by traversing the pulmonary capillary bed or via a right-to-left cardiac shunt such as occurs in the presence of a patent foramen ovale. Reported central nervous system findings at au-topsy include multiple cerebral petechiae, anemic lesions, and fat globules in the microvessels of the brain and spinal cord. Petechiae represent micro-scopic hemorrhagic infarcts that are produced either by vessel wall rupture due to embolism or by extravasation of blood into healthy tissues sur-rounding an area of necrosis (7).

Characteristic findings on diffusion-weighted and susceptibility-weighted MR images provide valuable supporting evidence for the diagnosis of cerebral fat embolism syndrome. The starfield pat-tern, which consists of innumerable bright punc-

tate foci of restricted diffusion against the dark background of brain parenchyma, has a limited differential diagnosis including diffuse axonal in-jury; cardiogenic, septic, or fat emboli; vasculitis; and minute hemorrhagic metastases (1114). In cerebral fat embolism syndrome, this pattern is thought to represent numerous sites of cytotoxic edema related to hemorrhage and infarction due to cerebral vessel occlusion by fat emboli. In patients with sickle cell disease and clinical manifestations of cerebral fat embolism syndrome, the starfield pattern is pathognomonic. Gradient-echo images and susceptibility-weighted images frequently demonstrate corresponding foci of susceptibility artifact representing microhemorrhages (7,15). Susceptibility-weighted imaging, which exploits differences in magnetic susceptibility between substances such as blood, iron, and calcification, is more sensitive than gradient-echo imaging in depicting microhemorrhages such as those oc-curring in cerebral fat embolism syndrome (16). Other MR imaging findings commonly seen in ce-rebral fat embolism syndrome include diffuse hy-perintense foci representing small areas of edema throughout the brain on fluid-attenuated inversion-recovery images and T2-weighted images (5). The appearance of the brain on T1-weighted MR im-ages obtained both before and after the administra-tion of an intravenous contrast medium is often normal. Findings at CT are typically negative, even in symptomatic individuals; however, mild cerebral edema may be seen, as occurred in our case (9).

Few pathologic processes produce widespread abnormalities of the gray- and white-matter struc-tures on diffusion-weighted and susceptibility-weighted MR images. Diffuse axonal injury due to brain trauma can produce abnormal foci of increased signal intensity on T2-weighted images and susceptibility artifact on gradient-echo images and susceptibility-weighted images, representing edema and microhemorrhages in the subcortical white matter, corpus callosum, internal capsule, and brainstem. Small peripheral microhemor-rhages can be found in patients with cerebral amyloid angiopathy; however, these foci develop more gradually and do not demonstrate restricted diffusion. Multiple cardiogenic emboli may have a similar appearance, but they typically occlude the terminal cortical branches, producing wedge-shaped infarctions. Septic emboli and hemorrhagic metastases can be found at the gray matterwhite matter junction, but they do not typically involve the cortex, they vary in size, and they enhance af-ter the administration of an intravenous contrast medium. Small-vessel vasculitis could produce

1306 September-October 2012 radiographics.rsna.org

multiple foci of microhemorrhage and infarction with restricted diffusion. The clinical history may be helpful for differentiating between these entities. In our patient, autopsy findings of cerebral edema and diffuse microhemorrhages and histologic find-ings of extensive intravascular fat emboli associated with petechiae and microinfarcts correlated well with the findings in antemortem imaging studies.

The treatment of patients with fat embolism syndrome is primarily supportive. Therefore, pre-vention, early diagnosis, and symptom manage-ment are paramount (17). The diagnosis is based primarily on clinical evidence and supported by findings in radiologic and pathologic investiga-tions, including the identification of fat globules in urinary sediment, peripheral blood, or sputum; demonstration of necrosis at bone marrow aspira-tion biopsy; quantification of fat-laden macro-phages in secretions obtained with bronchoalveo-lar lavage; and serum measurement of secretory phospholipase A2. Transfusion therapy, particularly exchange transfusion, has shown benefit in the treatment of patients with fat embolism syndrome in the setting of sickle cell disease. Although corti-costeroids have shown some benefit for preventing fat embolism in trauma patients, they do not ap-pear to benefit patients with sickle cell disease (2).

In a number of reported cases, patients with cerebral fat embolism due to trauma experienced neurologic changes that were transient and even-tually followed by a full recovery (1113,15). In patients with sickle cell disease, the clinical course of fat embolism is more often fulminant. Two key factors differentiate the two populations: In patients with sickle cell disease, vaso-occlusion instead of fracture is the causal mechanism for embolization, and the hypoxia and hypoxemia of vaso-occlusive crisis produce continuous sickling of red blood cells (2). In addition, it has been shown that patients with sickle cell disease and multiple vaso-occlusive crises experience chronic pulmonary, neurologic, and renal sequelae, which complicate recovery from new insults (1820). These findings support the role of a tenuous baseline state in patients with sickle cell disease, which differentiates them from a population of previously healthy patients who experience trau-ma, and may partially explain the difference in outcome between the two groups. The fulminant course of fat embolism syndrome in patients with sickle cell disease and the response to transfusion therapy in some of these patients underscore the need for rapid recognition of the disease process so that appropriate treatment can be initiated. Characteristic MR imaging findings, especially

the starfield pattern on diffusion-weighted im-ages, add valuable support to a clinical diagnosis of cerebral fat embolism syndrome.

References 1. Ataga KI, Orringer EP. Bone marrow necrosis in

sickle cell disease: a description of three cases and a review of the literature. Am J Med Sci 2000;320(5): 342347.

2. Dang NC, Johnson C, Eslami-Farsani M, Haywood LJ. Bone marrow embolism in sickle cell disease: a review. Am J Hematol 2005;79(1):6167.

3. Akhtar S. Fat embolism. Anesthesiol Clin 2009;27 (3):533550.

4. Gurd AR, Wilson RI. The fat embolism syndrome. J Bone Joint Surg Br 1974;56B(3):408416.

5. Chen JJ, Ha JC, Mirvis SE. MR imaging of the brain in fat embolism syndrome. Emerg Radiol 2008;15(3):187192.

6. Vichinsky EP, Neumayr LD, Earles AN, et al. Causes and outcomes of the acute chest syndrome in sickle cell disease. National Acute Chest Syn-drome Study Group. N Engl J Med 2000;342(25): 18551865.

7. Zaitsu Y, Terae S, Kudo K, et al. Susceptibility-weighted imaging of cerebral fat embolism. J Com-put Assist Tomogr 2010;34(1):107112.

8. Baker PL, Pazell JA, Peltier LF. Free fatty acids, catecholamines, and arterial hypoxia in patients with fat embolism. J Trauma 1971;11(12):10261030.

9. Simon AD, Ulmer JL, Strottmann JM. Contrast-enhanced MR imaging of cerebral fat embolism: case report and review of the literature. AJNR Am J Neuroradiol 2003;24(1):97101.

10. Kamenar E, Burger PC. Cerebral fat embolism: a neuropathological study of a microembolic state. Stroke 1980;11(5):477484.

11. Parizel PM, Demey HE, Veeckmans G, et al. Early diagnosis of cerebral fat embolism syndrome by diffusion-weighted MRI (starfield pattern). Stroke 2001;32(12):29422944.

12. Ryu CW, Lee DH, Kim TK, et al. Cerebral fat em-bolism: diffusion-weighted magnetic resonance im-aging findings. Acta Radiol 2005;46(5):528533.

13. Aravapalli A, Fox J, Lazaridis C. Cerebral fat embo-lism and the starfield pattern: a case report. Cases J 2009;2:212.

14. Marshall GB, Heale VR, Herx L, et al. Magnetic resonance diffusion weighted imaging in cerebral fat embolism. Can J Neurol Sci 2004;31(3):417421.

15. Suh SI, Seol HY, Seo WK, Koh SB. Cerebral fat embolism: susceptibility-weighted magnetic reso-nance imaging. Arch Neurol 2009;66(9):1170.

16. Haacke EM, Mittal S, Wu Z, Neelavalli J, Cheng YC. Susceptibility-weighted imaging: technical as-pects and clinical applications, part 1. AJNR Am J Neuroradiol 2009;30(1):1930.

17. Habashi NM, Andrews PL, Scalea TM. Therapeutic aspects of fat embolism syndrome. Injury 2006;37 (Suppl 4):S68S73.

18. Adams RJ. Stroke prevention and treatment in sickle cell disease. Arch Neurol 2001;58(4):565568.

19. Moser FG, Miller ST, Bello JA, et al. The spectrum of brain MR abnormalities in sickle-cell disease: a report from the Cooperative Study of Sickle Cell Dis-ease. AJNR Am J Neuroradiol 1996;17(5):965972.

20. Ataga KI, Orringer EP. Renal abnormalities in sickle cell disease. Am J Hematol 2000;63(4):205211.