Catastrophic Cerebral Fat Embolism Syndrome Lacks Hyperacute MRI Findings

PRACTICAL IMPLICATIONS

In patients who experience stroke due to cerebral fat embolism, hyperacute MRI may lack the characteristic diffusion-weighted imaging/fluid-attenuated inversion recovery changes expected in acute ischemic stroke.

A 64-year-old woman with sickle cell (hemoglobin SC) disease and previous embolic strokes was admitted for vaso-occlusive pain crisis and developed sudden onset coma. Hyperacute brain MRI showed only small areas of mismatch between diffusion-weighted imaging (DWI) and T2 fluid-attenuated inversion recovery (FLAIR) sequences that were insufficient to explain her profoundly altered level of consciousness. Repeat MRI finding 24 hours later showed innumerable foci of restricted diffusion and petechial microhemorrhage, confirming a diagnosis of cerebral fat embolism (CFE) due to sickle cell crisis.

Case

A 64-year-old woman with a history of SC disease and previous right posterior cerebral artery stroke presented with 24 hours of back and hip pain radiating down both legs. She was admitted for vaso-occlusive pain crisis and treated with opiate therapy and half normal saline infusion. Within 36 hours of symptoms, she developed rapid onset comatose state with absent brainstem reflexes. The CT scan of the head was unremarkable (eFigure 1, links.lww.com/CPJ/A307). The CT angiogram showed no large vessel occlusion. MRI performed 4.5 hours after coma onset showed a small linear acute infarct with diffusion restriction (b = 1,000 s/mm²) in the left posterior parietal region with no corresponding FLAIR signal. There were also subtle punctate foci of diffusion restriction in the left thalamus and bilateral high frontoparietal lobes (Figure 1). These findings, however, were not significant enough to explain her comatose state. Recombinant tissue plasminogen activator and emergency red blood cell exchange transfusion were administered with no clinical improvement. Continuous video EEG demonstrated severe generalized slowing. Repeat MRI finding 24 hours after coma onset showed extensive multifocal punctate ischemic infarcts with numerous microhemorrhages (the starfield pattern), compatible with cerebral fat embolism syndrome (FES) (Figures 1 and 2, eFigure 1).¹ The patient had no clinical recovery over several days. A Tc-99m radionucleotide angiogram of the head performed on day 6 showed the absence of intracranial blood flow compatible with brain death. Autopsy revealed innumerable punctate acute ischemic infarcts in the bilateral cerebral and cerebellar hemispheres and brainstem with associated petechial hemorrhages and intravascular lipid droplets (Figure 2), confirming the diagnosis of catastrophic cerebral FES. Fat emboli were also present in the lungs and kidneys. No patent foramen ovale or intracardiac shunt was present. A single section of vertebral bone marrow was sampled with no evidence of bone marrow necrosis.

Figure 1. MRI Comparison at 4.5 and 24 Hours.

Diffusion-weighted imaging (DWI), ADC map, and fluid-attenuated inversion recovery (FLAIR) images at 4.5 hours after coma onset (top) and then at 24 hours after coma onset (bottom). (A) At the level of the third ventricle, there is initially 1 small linear area of diffusion restriction in the left posterior parietal region (thick arrow) at 4.5 hours. (B) At the level of the corona radiata, there are very subtle punctate foci of diffusion restriction in the bilateral high frontoparietal lobes (circled). There is otherwise minimal DWI/FLAIR mismatch on MRI at 4.5 hours. (C) At the level of the third ventricle, MRI finding at 24 hours demonstrates extensive multifocal punctate ischemic infarcts (thin arrows). (D) At the level of the corona radiata, MRI finding at 24 hours demonstrates extensive multifocal punctate areas of restricted diffusion appearing bright on a dark background in the characteristic starfield pattern (thin arrows).

Figure 2. Microvascular Pathology and Diffuse Cerebral Microhemorrhages.

(A) Hematoxylin and eosin–stained section of cortex from autopsy demonstrates fat droplets (thick black arrows) in a cerebral microvessel. (B and C) Susceptibility-weighted imaging at 24 hours after coma onset demonstrates diffuse microhemorrhages predominantly in the white matter and the cerebellum (thick white arrows).

Discussion

FES is most commonly seen after long bone fractures and orthopedic procedures, although several nontraumatic etiologies, including SC disease and diabetes mellitus, may rarely trigger FES.² The pathophysiology of FES is debated; the “mechanical theory” posits that bone marrow injury allows fat particles to pass into the marrow venous sinusoids and, thus, enter systemic circulation. Alternatively, the “chemical theory” hypothesizes that systemic stress triggers the release of plasma mediators that provoke lipolysis in various tissues. Free fatty acids leak into the venous circulation with resultant microvascular inflammation and coalescence into fat emboli in end-organ microvasculature.^2,3

The classic symptomatic triad of FES includes signs and symptoms involving the lungs, skin, and brain, with 33%–86% of patients developing neurologic symptoms.² When neurologic symptoms predominate, the clinical syndrome may be referred to as “cerebral fat embolism”.³ Although most cases are a self-limited syndrome with complete neurologic recovery, CFE carries an overall mortality rate up to 10%.²

Brain MRI is critical in the evaluation of CFE. Scattered spot lesions can be seen on DWI, with correlating isointense or hyperintense lesions on T2-weighted imaging.⁴ This appearance of scattered bright lesions on a dark background on DWI has been termed the starfield pattern.¹ Alternatively, confluent symmetric restricted diffusion of the cortical white matter with faintly hyperintense T2 lesions may be seen. Scattered petechial white matter hemorrhages are common on susceptibility sequences. These can be distinguished from the petechial hemorrhage of diffuse axonal injury (DAI) primarily by location; CFE occurs in the periventricular and subcortical white matter bilaterally, whereas DAI occurs in the gray-white interface of the frontotemporal lobes and corpus callosum.⁴

Our case demonstrates that minimal MRI findings may be present in the hyperacute phase of CFE but may emerge as the disease progresses. Although this is not unique to CFE (DWI-negative strokes comprise 6.8% of all strokes), it may cause the diagnosis to be overlooked unless a high level of clinical suspicion is maintained.⁵ Because MRI is increasingly incorporated into the hyperacute evaluation of stroke, the neurologist must be aware that ischemic stroke due to CFE may not be initially detected and a repeat MRI should be obtained to verify the clinical diagnosis.

Appendix. Authors

Study Funding

The authors report no targeted funding.

Disclosure

M.T. Gusler, A. Vagal, S.D. Gilday, and M.L. Flaherty report no disclosures relevant to the manuscript. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.