Case Report

An autopsy case of non-traumatic fat embolism syndrome

Mai Sakashita,1 Shingo Sakashita ,2 Akiko Sakata,3 Noriko Uesugi,2 Kazunori Ishige,4 Ichinosuke Hyodo4

and Masayuki Noguchi2

1Doctoral Program in Biomedical Science, Graduate School of Comprehensive Human Sciences, University ofTsukuba, Ibaraki, Japan, 2Department of Pathology, Faculty of Medicine, University of Tsukuba, Ibaraki, Japan,3Department of Pathology, Hitachi General Hospital, Ibaraki, Japan, and 4Department of Gastroenterology, Facultyof Medicine, University of Tsukuba, Ibaraki, Japan

Fat embolism syndrome (FES) occurs after long bonefractures and the symptoms appear 24–72h after the initialtrauma. Fat emboli can affect both the pulmonary andsystemic circulation. Apart from the most common type ofFES that originates from bone fracture, non-traumatic FEShas been also reported. We have experienced an autopsycase of non-traumatic FES. An 81-year-old man withhepatocellular carcinoma associated with alcoholic livercirrhosis suddenly lost consciousness before transcath-eter arterial chemoembolization treatment for his diseaseand died 5h after the episode. At autopsy, numerous fatdroplets were detected in the alveolar capillaries of thelung and glomerular capillaries of the kidney. Lipid analy-sis of lung autopsy specimens by thin-layer chromatogra-phy showed that the emboli were composed mainly oftristearin. Free fatty acids (FFA) has been considered to bethe main component of fat emboli and can be a cause ofacute respiratory distress syndrome (ARDS). However, inthe present case, the lung specimen contained tristearinand ARDS did not occur. This is the first report ofnon-traumatic FES in which lipid analysis of humanautopsy specimens has been conducted.

Key words: fat embolism syndrome, non-trauma, lipidanalysis, phosphatidylcholine, tristearin

Fat embolism is defined as mechanical blockage of thelumina of blood vessels by circulating fat particles.1 Fatembolism syndrome (FES) is fat embolism with identifiableclinical pattern of symptoms and the diagnostic triadincludes hypoxemia, neurological symptoms, and a pete-chial rash.2 FES usually occurs after fracture of long bones,and the symptoms appear 24–72 h after the initial trauma.3

On the other hand, non-traumatic FES has also beenreported in association with bone marrow necrosis in sicklecell disease,4 corticosteroid administration, fatty liver dis-ease,5 bone marrow transplantation, pancreatitis, liposuc-tion,6 and diabetes mellitus (DM).7 Although FES ischaracterized by both pulmonary and systemic fat embo-lism, the biological mechanism of FES has not beenclarified in detail.6

We have experienced a case of clinically unexplainedsudden death without any trauma in which autopsy revealednumerous fat emboli in multiple organs, including the lungand kidney. We report this autopsy case of non-traumaticFES and describe the results of lipid analysis by thin-layerchromatography using autopsy materials.

CLINICAL SUMMARY

An 81-year-old Japanese man had been diagnosed threeyears previously as having hepatocellular carcinoma inthe hepatic portal region associated with alcoholic cirrhosis.He had undergone proton radiation therapy and the carci-noma had been well controlled. However, a year before thelast admission, another hepatocellular carcinoma had ap-peared in the S7 region for which transcatheter arterialchemoembolization (TACE) had been performed twice.Because of local re-recurrence of the liver carcinoma in S7,he was admitted to the University Hospital of Tsukuba,Japan, for a third session of TACE. On the day ofadmission, his performance status was good (PS1) and hisvital signs were normal. His height was 161 cm and weightwas 68.6 kg. Other than mild anemia (hemoglobin 11 g/dL),the laboratory data were unremarkable. Serum lipids levelswere within normal limits (triglyceride (TG) 82mg/dL, phos-pholipid 168mg/dL, total cholesterol (TC) 149mg/dL, high-density lipoprotein cholesterol (HDL-C) 38.4mg/dL, andlow-density lipoprotein cholesterol (LDL-C) 84mg/dL). TheC-reactive protein (CRP) level was 0.79mg/dL.

Correspondence: Shingo Sakashita MD, PhD, Department ofPathology, Faculty of Medicine, University of Tsukuba, 1-1-1Tennodai, Tsukuba-shi, Ibaraki 305-8575, Japan.E-mail: s_sakashita@md.tsukuba.ac.jpReceived 20 March 2017. Accepted for publication 31 May 2017.

© 2017 Japanese Society of Pathology and John Wiley & SonsAustralia, Ltd

Pathology International 2017; xx: 1–6 doi:10.1111/pin.12556

In the morning, just after coughing, the patient suddenlylost consciousness on his ward bed and suffered cardiopul-monary arrest (CPA). At that time, TACE was not per-formed. He had no symptom of a petechial rash. After CPA,the levels of all lipids, especially TG, were decreased (TG,82mg/dL on admission and 36mg/dL after CPA; phospho-lipid 168mg/dL and 99mg/dL; TC, 149mg/dL and 83mg/dL;HDL-C, 38.4mg/dL and 21.7mg/dL, LDL-C 84mg/dL and52mg/dL, respectively). Computed tomography revealed nosigns of injury, and the cause of CPA was unknown.Despite immediate cardiopulmonary resuscitation and per-cutaneous cardiopulmonary support, the patient died 5hafter the episode. Autopsy was conducted 3 h after death.

PATHOLOGICAL FINDINGS

Macroscopically, the lungs were reddish and their weightswere increased (left lung, 550 g; right lung, 780 g) (Fig. 1a).Infarction or hemorrhage was not evident. The kidneys werealso reddish and their weights were 140g (left) and 120g(right) (Fig. 2a). Infarction or hemorrhage was not evident.The heart weighed 360g and a patent foramen ovale wasnot evident. The thickness of the left ventricular wall was11mm and that of the right ventricular wall was 4mm. Rightventricular dilation was not evident. The coronary arteries

showed arteriosclerosis. The liver weighed 870g, and awell-circumscribed tumor was detected in the right lobe(S7). It was 3 cm in diameter and yellowish white in color(Fig. 3a). The tumor was classified as the simple nodulartype macroscopically. No tumor was found in the hepaticportal region. The aorta showed severe atherosclerosis.

The autopsy materials were fixed in buffered formalin for7 days, then hematoxylin and eosin (HE) staining wasperformed. For oil red O staining, frozen sections from theformalin-fixed samples were prepared. We also performedfat staining using osmium tetroxide by a modification ofMochizuki et al.'s protocol.8 Briefly, specimens were fixed in10% neutral-buffered formalin and cut to 10�10� 2mm.They were rinsed in tap water for 3 h and stained with amixture of 2% osmium tetroxide and 5% potassium dichro-mate. The specimens were then sealed hermetically andheated at 60°C for 4 h. After rinsing, the samples wereembedded in paraffin, cut into sections 3mm thick, anddeparaffinized. After bleaching with 5% periodic acid, HEstaining was performed.

Microscopically, both of the lungs were congested butthere was no inflammation or diffuse alveolar damage(DAD). Interestingly, numerous alveolar capillaries in bothlungs were significantly dilated and empty in all microscopicfields on HE staining (Fig. 1b). In the kidneys, most of theglomerular capillaries were distended diffusely (Fig. 2b,c).

Figure 1 Lungs. (a) Macroscopic findings, (b) HE staining, �200, (c) Oil red O staining, �200, (d) Osmium staining, �200. (a) The lungsare reddish and their weights are increased. Infarction or hemorrhage is not evident. (b) Alveolar capillaries are dilated diffusely in bothlungs. (c, d) Intravascular fat is confirmed by oil red O staining and osmium staining.

© 2017 Japanese Society of Pathology and John Wiley & Sons Australia, Ltd

2 M. Sakashita et al.

Oil red O staining and osmium staining revealed numerousfat droplets in both the alveolar and glomerular capillaries(Figs. 1c,d, 2d,e). Moreover, luminal fat particles in bloodvessels were detected in the myocardial capillaries of theheart and Glisson capsule in the liver (Fig. 3b). The fatembolism found in the present case was different from bonemarrow embolism, since the fat embolism consisted of onlyadipose cells and there were no accompanying hematopoi-etic cells. In addition, cholesterol embolisms were not foundin any of the organs, although the aorta showed severeatherosclerosis.The liver showed diffuse cirrhosis (Fig. 3c) and the well-

circumscribed liver tumor was histologically diagnosed asmoderately differentiated hepatocellular carcinoma (Fig. 3d).The pathological stage of the hepatocellular carcinoma wasyaT2N0M0, yaStageII. The liver didn’t show steatosis.Although the coronary artery showed arteriosclerosis and

maximum 75% stenosis, no occlusion was evident. Theheart did not show myocardial infarction, and the cardiacconduction system was unremarkable.The bone marrow showed normal cellularity and there

was no evidence of bone marrow necrosis.

As there were no possible causes of death other thansystemic fat emboli, we considered that death had beendue to non-traumatic FES.

We performed thin-layer chromatography (TLC) to analyzethe type of lipid composing the fat droplets found incapillaries. This study was approved by the Ethics Commit-tee of the University of Tsukuba Hospital (H28-162).Although lipid analysis is usually performed using freshtissues, we used formalin-fixed samples of lung becausefresh tissue samples were not available. We were able toextract lipid using the Folch method, and decided to conductTLC, which was performed at Toray Research Center(Tokyo, Japan). Another autopsy case (a 75-year-oldJapanese man who had died due to hemorrhage) was usedas a normal control. We analyzed the levels of phosphatidyl-choline (PC), phosphatidylethanolamine (PE), phosphatidyl-serine (PS), phosphatidylinositol (PI), cholesterol (Ch),tristearin (TG), and stearic acid (FFA). The results are shownin Table 1. The levels of PC, TG, and FFA in the patient'slung samples were 2.47mg/mg, 3.28mg/mg and 1.77mg/mg,respectively, and those in control lung samples were1.15mg/mg, 2.60mg/mg, and 2.83mg/mg, respectively.

Figure 2 Kidneys. (a) Macroscopic findings, (b) HE staining, �400, (c) Periodic acid-methenamine-silver (PAM) staining, �400, (d) Oilred O staining, �400, (e) Osmium staining, �400, (a) The kidneys are reddish and their weights are increased. Infarction or hemorrhageis not evident. (b, c) Glomerular capillaries are dilated in both kidneys. (d, e) Intravascular fat is confirmed by oil red O stain andosmium stain.

© 2017 Japanese Society of Pathology and John Wiley & Sons Australia, Ltd

Non-traumatic FES autopsy 3

DISCUSSION

We have described a rare case of non-traumatic FES.Previously reported cases of non-traumatic FES are sum-marized in Table 2.9–18 Although there have been severalreports of non- traumatic FES associated with fatty liver,5

this is the first report of non-traumatic FES associated withhepatocellular carcinoma. Schulz et al. reported two cases

of fatal fat embolism associated with steatosis,5 andindicated that massive fat embolism could have originatedfrom fatty liver. The liver of the present patient showedcirrhosis, however, there was no evidence of steatosis.

Although many researchers have assumed that FESfinally leads to acute respiratory distress syndrome(ARDS),6 our patient did not show DAD in the lungs. Inaddition, only two out of 16 reported cases with non-traumatic FES had ARDS (Table 2). Therefore, ARDS maynot necessarily occur in non-traumatic FES.

Although the exact pathogenesis of FES remains un-known, several theories have been suggested.15 The twomain theories are the mechanical theory and the biochemi-cal theory. Since the present patient had not suffered anyinjury, the mechanical theory could not fully explain thepathogenesis of the present case.

The biochemical theory includes several possible mecha-nisms. One of them is that release of FFAs is associatedwith FES.6 Pneumocyte lipases can hydrolyze triglycerideinto FFAs, which release toxic substances that damage thecapillary endothelium. FFAs increase vascular permeabilityand inactivate surfactants. Finally, respiratory failure due toARDS may occur. Lipid analysis of the present autopsiedlung sample confirmed that the fat emboli contained

Figure 3 Liver. (a) Macroscopic findings, (b) The histology of the liver without carcinoma, osmium staining, �200, (c) The histologyof liver without carcinoma, Masson trichrome staining, �40, (d) The histology of the tumor in S7, HE staining, �100. (a) Simplenodular-type carcinoma is evident in S7, measuring 3 cm in diameter (arrow). There is no carcinoma in other areas, including thehepatic portal area. (b) Intravascular fat is confirmed in Glisson's capsule. (c) The liver tissue without carcinoma shows cirrhosis.(d) The tumor shows a trabecular pattern and is diagnosed as moderately differentiated hepatocellular carcinoma. There is noevidence of steatosis.

Table 1 Lipid analysis of lung by thin-layer chromatography

Concentration(mg/mg)

Lipid Control Present case

PC 1.15 2.47PS – –

PI 0.29 0.15PE 0.16 0.10Ch 1.72 1.96FFA 2.83 1.77TG 2.60 3.28

–, undetected;PC, phosphatidylcholine; PS, phosphatidylserine; PI, phosphatidylino-

sitol; PE, phosphatidylethanolamine; Ch, cholesterol; TG, tristearin;FFA, stearic acid.

© 2017 Japanese Society of Pathology and John Wiley & Sons Australia, Ltd

4 M. Sakashita et al.

triglyceride (preliminary observation) (Table 1). In addition,Schinaid et al.19 have reported that FFA was only moder-ately higher in patients with long bone fractures andconcluded that mobilization of FFA did not contribute toFES. Moreover, in rat models, Takada et al. have reportedthat an injection of neutral fat caused fat embolism,however, fat embolisms did not occur with an injection ofFFAs. They hypothesized that hyperviscosity of glycerolcaused occlusion of capillaries.3 We thought that ourpresent results might reflect this mechanism, since thepatient's lung samples contained TG (Table 1) and theserum level of TG was decreased significantly after CPA(36mg/dL, as compared with 82mg/dL on admission). Onthe basis of our results and these previous reports, weconsider that TG might have been a contributory factor inthe present case.There were several limitations to this study. To our

knowledge, there have been no previous reports of lipidanalysis using formalin-fixed samples, and thus the varia-tion and reproducibility of this approach has not beenvalidated. In addition, only one control sample was used.Therefore, the present data must be regarded as prelimi-nary, and accumulation of further cases is needed.Nevertheless, this is the first report of lipid analysis being

employed in a case of non-traumatic FES using humanautopsy specimens that contained fat emboli (Table 2).

Although Mendoza et al. employed gas chromatographycoupled with mass spectrometry to examine oily materialbetween muscle fibers in buttock area since the patient gotgluteal injection of vitamin E before death,12 the authors didnot analyze the organ that contained fat embolism. More-over, hitherto, lipid analysis of human FES using organsamples which contain fat embolism was conducted only byusing bronchoalveolar lavage.20 In addition, although sev-eral reports have documented animal models of FEScreated by injecting several kinds of fat, they did notanalyze the fat emboli themselves, even in animal autopsyspecimens.3

In conclusion, we have reported an autopsy case ofnon-traumatic FES with lipid analysis of a lung specimen.This confirmed that the emboli contained TG and therewas no evidence of DAD in the lungs. These resultsindicated that this case of non-traumatic FES was causedby numerous lipid emboli containing TG. The biologicalmechanism of these emboli is still unknown, therefore,accumulation of further cases will help to clarify thepathogenesis of FES.

DISCLOSURE STATEMENT

None declared.

Table 2 Reported cases of non-traumatic fat embolism syndrome

Author Year Age SexUnderlying disease

(or precedent medical practice) Diagnostic toolLipid

analysis PFOARDS(DAD)

Present case 2017 81 M Hepatocellular carcinoma Autopsy Yes No NoCantu CA, et al. 2017 60 F Liposuction Autopsy No Not written Not writtenScarpino M, et al. 2016 64 F Empyema (treated by VATS) MRI No Yes NoMay J, et al. 2016 57 F Sickle-b+ thalassemia,

bone marrow necrosisClinical history, CT,

MRINo Not written No

Mendoza-MoralesRC, et al.

2016 24 F Gluteal injection of vitamin E(for cosmetic enhancement)

Autopsy Yes� No Yes

Kammeyer R, et al. 2016 54 M Hemoglobin SC disease MRI No Not written NoGraff DM, et al. 2015 18 F Hemoglobin SC disease,

bone marrow necrosisAutopsy No No No

Celik SU, et al. 2015 33 F Renal angiomyolipoma CT No Not written NoSchrufer-Poland T,et al.

2015 29 F Arnold-Chiari malformation,Cesarean delivery

Autopsy No No No

Alobeidi F, et al. 2015 8 F Sickle cell disease,bone marrow necrosis

MRI No Yes No

Coronado-Malag�on M,et al.

2010 26 F Soft tissue filter injection CT, BAL No Not written Not written

Bilgrami S, et al. 1999 59 F Diffuse large-cell lymphoma Autopsy No Not written Not writtenBilgrami S, et al. 1999 54 M Diffuse large-cell lymphoma Autopsy No Not written Not writtenSchulz F, et al. 1996 63 F Acute hepatic necrosis with

steatosisAutopsy No Not written No

Schulz F, et al. 1996 46 M Acute hepatic necrosis causedby fulminant viral hepatitis

with steatosis

Autopsy No No No

Rosen JM, et al. 1986 40 M Diffuse mixed cellularitylymphoma

Biopsy, autopsy No Not written Yes

*Gas chromatography coupled with mass spectrometry using oily material between the muscle fibers in buttock area. (tocopherol was detected).

© 2017 Japanese Society of Pathology and John Wiley & Sons Australia, Ltd

Non-traumatic FES autopsy 5

REFERENCES

1 Hulman G. The pathogenesis of fat embolism. J Pathol 1995;176: 3–9.

2 Mellor A, Soni N. Fat embolism. Anaesthesia 2001; 56: 145–54.3 Takada M, Chiba S, Nagai T et al. Inflammatory responses to

neutral fat and fatty acids in multiple organs in a rat model offat embolism syndrome. Forensic Sci Int 2015; 254: 126–32.

4 Kammeyer R, Devnani R, Mehta R. Cerebral fat embolism syn-drome mimicking thrombotic thrombocytopenic purpura in apatient with hemoglobin SC disease. Am J Hematol 2016; 91:539–42.

5 Schulz F, Tr€ubner K, Hildebrand E. Fatal fat embolism in acutehepatic necrosis with associated fatty liver. Am J Forensic MedPathol 1996; 17: 264–8.

6 Habashi NM, Andrews PL, Scalea TM. Therapeutic aspects offat embolism syndrome. Injury 2006; 37: S68–73.

7 Kent SP. Fat embolism in diabetic patients without physicaltrauma. Am J Pathol 1955; 31: 399–403.

8 Mochizuki S, Moritani T, Ishida K, Naganuma H, Saito T,Sasano K. Fat stain for paraffin-embedded specimen by blockstaining osmium tetroxide. Pathol Clin Med 2006; 24: 99–102.[in Japanese].

9 Cantu CA, Pavlisko EN. Liposuction-induced fat embolismsyndrome. BMJ Case Rep 2017; 2017 pii: bcr-2017-219835.https://doi.org/10.1136/bcr-2017-219835.

10 Scarpino M, Lanzo G, Cappelli F et al. Cerebral fat embolismafter video-assisted thoracic surgery. Ann Thorac Surg 2016;102: e409–e11.

11 May J, Sullivan JC, LaVie D, LaVie K, Marques MB. Inside out:Bone marrow necrosis and fat embolism complicating sickle-bþthalassemia. Am J Med 2016; 129: e321–e4.

12 Mendoza-Morales RC, Camberos-Nava EV, Luna-Rosas A,Garc�es-Ramírez L, De la Cruz F, García-Dolores F. A fatalcase of systemic fat embolism resulting from gluteal injectionsof vitamin e for cosmetic enhancement. Forensic Sci Int 2016;259: e1–4.

13 Kammeyer R, Devnani R, Mehta R. Cerebral fat embolismsyndrome mimicking thrombotic thrombocytopenic purpura in apatient with hemoglobin SC disease. Am J Hematol 2016; 91:539–42.

14 Graff DM, Owen E, Bendon R, Bertolone S, Raj A. Distinctiveacellular lipid emboli in hemoglobin SC disease following bonemarrow infarction with parvovirus infection. Case Rep Hematol2015; 2015: 328065.

15 Schrufer-Poland T, Singh P, Jodicke C, Reynolds S, Maulik D.Nontraumatic fat embolism found following maternal death aftercesarean delivery. AJP Rep 2015; 5: e1–5.

16 Alobeidi F, Inusa BP, Singh RR, U-King-Im JM. Cerebralmicrohaemorrhages secondary to fat embolus syndrome insickle cell disease. Postgrad Med J 2015; 91: 55–6.

17 Bilgrami S, Hasson J, Tutschka PJ. Case 23-1998: Fatembolism. N Engl J Med 1999; 340: 393–4.

18 Rosen JM, Braman SS, Hasan FM, Teplitz C. Nontraumatic fatembolization. A rare cause of new pulmonary infiltrates in animmunocompromised patient. Am Rev Respir Dis 1986; 134:805–8.

19 Schnaid E, Lamprey JM, Viljoen MJ, Joffe BI, Seftel HC. Theearly biochemical and hormonal profile of patients with longbone fractures at risk of fat embolism syndrome. J Trauma1987; 27: 309–11.

20 Roger N, Xaubet A, Agustí C et al. Role of bronchoalveolarlavage in the diagnosis of fat embolism syndrome. Eur Respir J1995; 8: 1275–80.

© 2017 Japanese Society of Pathology and John Wiley & Sons Australia, Ltd

6 M. Sakashita et al.