Full text of DOI:10.1007/BF03020414

GENERAL ANESTHESIA
19
Fat embolism syndrome and elective knee
arthroplasty
[Embolie graisseuse et arthroplastie du genou non urgente]
Kathryn Jenkins FRCA,* Frances Chung FRCPC,* Richard Wennberg FRCPC,† Edward E. Etchells FRCPC,†
Rod Davey FRCSC‡
Purpose: To report a case of fat embolism syndrome (FES) follow-
ing elective left knee arthroplasty and review the diagnosis, investi-
gation, prevention and perioperative management of this condition.
suite d’une investigation poussée, un diagnostic d’EG a été fait, 48 h
après l’opération, confirmé par le développement d’encéphalopathie
diffuse, de thrombocytopénie, d’hypoxémie, de pétéchies thoraciques
et de modifications radiographiques pulmonaires. Un examen d’ima-
gerie par résonance magnétique a confirmé le diagnostic cinq jours
après l’opération. L’état de la patiente s’est graduellement amélioré en
faveur d’une lente résolution de l’encéphalopathie.
Dans les cas déjà publiés d’EG associée à l’arthroplastie du genou, on
retrouve soit un collapsus cardiorespiratoire peropératoire, soit un
dérèglement postopératoire respiratoire, cardiovasculaire et/ou
cérébral comme dans le cas présent.
Clinical features: A 76-yr-old lady presented for left total knee
arthroplasty under general anesthesia. After an uneventful anes-
thetic and initial recovery, she developed respiratory and neurolog-
ical complications six hours postoperatively necessitating supportive
care in the intensive care unit. Following extensive investigation, a
clinical diagnosis of FES was made 48 hr postoperatively supported
by the development of diffuse encephalopathy, thrombocytopenia,
hypoxemia, chest petechiae and chest x-ray changes. A magnetic
resonance imaging scan five days postoperatively confirmed this
diagnosis. Her postoperative course showed gradual improvement
consistent with a slowly resolving encephalopathy.
Conclusion : Le diagnostic clinique d’EG est essentiellement un
diagnostic d’exclusion, corroboré par les examens de laboratoire et de
radiographie. L’identification préopératoire des patients à risque,
l’usage d’un monitorage périopératoire effractif approprié et de tech-
niques chirurgicales modifiées peuvent minimiser le développement de
ce syndrome. Le traitement consiste en une thérapie de soutien.
Previous published cases of FES associated with knee arthroplasty
present either with intraoperative cardiorespiratory collapse or, as
with this patient, in the postoperative period with respiratory, car-
diovascular and/or cerebral dysfunction.
Conclusions: The clinical diagnosis of FES is essentially one of
exclusion, supported by laboratory and radiological investigations.
Preoperative identification of at-risk patients, use of appropriate
invasive perioperative monitoring and modified surgical techniques
may minimize the development of the syndrome. Treatment is sup-
portive.
AT embolism syndrome (FES) has been rec-
ognized since the late 1800’s and extensively
described following traumatic, surgical and
F
atraumatic conditions. We present a case of
FES following elective left knee arthroplasty. Previous
published case reports of FES associated with elective
knee arthroplasty are summarized, revealing common
patterns of presentation. Recommendations for peri-
operative management are suggested to minimize the
risk of this syndrome during knee arthroplasty.
Objectif : Décrire un cas d’embolie graisseuse (EG) survenue après
une arthroplastie du genou non urgente et réexaminer le diagnostic, la
recherche, la prévention et le traitement périopératoire entourant
cette situation.
Éléments cliniques : Une femme de 76 ans s’est présentée pour
une arthroplastie totale du genou gauche sous anesthésie générale. La
récupération, d’abord sans incident, s’est compliquée de troubles res-
piratoires et neurologiques six heures après l’intervention, ce qui a
nécessité un traitement de soutien à l’unité des soins intensifs. À la
Case report
A 76-yr-old female, weighing 66 kg, with osteoarthritis
of her left knee presented for left total knee arthroplasty
in September 2000. Her past medical history included
From the Departments of Anesthesia,* Medicine,† and Orthopedic Surgery,‡ University of Toronto, Toronto Western Hospital, Toronto,
Ontario, Canada.
Address correspondence to: Dr. Frances Chung, Department of Anesthesia, Toronto Western Hospital, 399 Bathurst Street, Toronto,
Ontario M5T 2S8, Canada. Phone: 416-603-5118; Fax: 416-603-6494; E-mail: frances.chung@uhn.on.ca
Accepted for publication July 18, 2001.
Revision accepted October 22, 2001.
CAN J ANESTH 2002 / 49: 1 / pp 19–24

20
CANADIAN JOURNAL OF ANESTHESIA
osteotomy of right hip, right knee arthroplasty, cholecys-
tectomy, hysterectomy, appendectomy, bowel resection
for diverticulitis, posterior fossa craniotomy for cerebel-
lo-pontine angle tumour, lower lumbar disc pathology
and a hiatus hernia with reflux. She had a five-year histo-
ry of stable angina. Her medication included atenolol
100 mg once daily and nitroglycerin spray as required.
Preoperative echocardiography revealed mild global left
ventricular dysfunction and infero-basal hypokinesis con-
sistent with regional ischemia. A 1996 thallium scan was
normal, and her electrocardiogram (ECG) revealed
longstanding left bundle branch block (LBBB).
Preoperative vital signs and blood tests were normal.
Preoperative venous thrombo-prophylaxis with oral
coumadin had been commenced.
The patient had a left total knee arthroplasty, using a
thigh tourniquet, under general anesthesia. A cemented
total condylar system was inserted, using femoral and
tibial intramedullary alignment rods. With standard
monitoring, a rapid sequence induction was carried out
with fentanyl 250 µg, propofol 120 mg and suxam-
ethonium 150 mg. Anesthesia was maintained with des-
flurane (end-tidal concentrations 2–3%) in nitrous
oxide (70%) and oxygen. The patient was paralyzed
with rocuronium, and morphine 8 mg was given intra-
operatively for analgesia. Estimated intraoperative
blood loss was 100 mL and the patient received 1600
mL of 0.9% saline intravenously. Vital signs were stable
throughout the intraoperative course (2.5 hr). On
recovery from anesthesia, the patient was awake, alert
and hemodynamically stable. She required supplemen-
tal oxygen via a Venturi mask (FIO2 0.5) to maintain
oxygen saturations of 97% following a transient drop in
saturations to 90% in the postanesthesia care unit on 4
L·min– 1 oxygen via nasal prongs. Morphine patient-
controlled analgesia was provided. She achieved a max-
imal Aldrete postoperative recovery score of 10 prior to
discharge to the ward.
With further deterioration in her neurological sta-
tus 24 hr postoperatively, she was transferred to the
intensive care unit. Neurological examination revealed
the patient to be stuporous with no focal signs and
bilateral extensor plantar responses. No retinal fat or
petechiae were seen. A cerebral computed tomogram
(CT) was normal. Repeat ABG on 50% oxygen had
deteriorated: pH 7.46, pO 80 mmHg, pCO 36
2
mmHg, base excess 2.3, sat2urations 95%. There was
no evidence of deep venous thromboses clinically or
on venous ultrasounds. A lumbar puncture revealed
no abnormality of cerebrospinal fluid. No other
sources of sepsis were identified.
At 48 hr postoperatively, fat embolism was suspect-
ed, supported by the development of petechiae on the
upper trunk and thrombocytopenia (platelets 98 ×
109·L–1). Magnetic resonance imaging (MRI) was car-
ried out five days postoperatively which revealed mul-
tiple small bilateral predominantly white matter
changes preferentially located in the fronto-parietal
“watershed” area between the major hemispheric vas-
cular territories (Figure, top). The abnormalities
appeared as hyperintense lesions on T2-weighted and
fluid-attenuated inversion recovery (FLAIR)
sequences: T1-weighted sequences were normal. A
repeat CT was unchanged. Electroencephalography
showed an abnormal diffuse theta recording with no
focal abnormalities. A repeat echocardiogram five days
postoperatively detected 8 mm left to right shunt flow
across an interatrial septum consistent with an atrial
septal defect or patent foramen ovale.
Her subsequent postoperative course showed grad-
ual improvement of neurological status consistent
with a slowly resolving encephalopathy. A repeat MRI
one month later showed resolution of the fronto-pari-
etal lesions (Figure, bottom).
Discussion
Six hours postoperatively, she was noted to be
increasingly drowsy, opening eyes to voice, withdrawing
to pain but no verbal response. She was hemodynami-
cally stable with a respiratory rate of 16·min–1, and sat-
urations of 97% on oxygen FIO2 0.5. Arterial blood
gases (ABG) revealed pH 7.47, pCO 42 mmHg, pO
92 mmHg, base excess -0.3, saturations 97% on 50%
inspired oxygen; Hb 127 g·L–1, WBC 16.8 × 109·L–1,
platelets 163 × 109·L–1; glucose 7.5 mmoL·L–1;
Troponin C 0.6 µg·L– 1. There was no detectable meta-
bolic derangement. Her ECG was unchanged (LBBB).
The chest x-ray showed bilateral diffuse infiltrates con-
sistent with pulmonary edema. Furosemide 40 mg iv
was administered which resulted in urinary diuresis but
no improvement in ABG.
Fat embolism occurs in almost all lower extremity
trauma and intramedullary surgery. FES, however, is a
severe multisystem manifestation of embolization that
occurs much less commonly. The incidence of FES in
retrospective reviews is less than 1%.1, 2However, there
is a greater incidence reported in prospective studies
of 11–29%.3–6 Overall mortality varies between
7–20%,1,2 and long-term morbidity is usually due to
neurological dysfunction.7
The commonest surgical procedures predisposing
to FES are intramedullary nailing of long bones, hip
arthroplasty and knee arthroplasty.7 Over the last 30
years, FES associated with knee arthroplasty has been
described in at least 16 case reports.8– 21 Of these,
seven proved fatal and one had poor long-term neu-
2
2

Jenkins et al.: FAT EMBOLISM SYNDROME
21
FIGURE Top; Brain magnetic resonance imaging (MRI) five days postoperatively shows multiple small hyperintense lesions on fluid-
attenuated inversion recovery sequences preferentially located in “watershed” area between major hemispheric vascular territories. Bottom;
Repeat MRI one month later shows resolution of abnormalities.
rological outcome. The remaining 50% recovered
completely.
heart disease, relative fluid overload from periopera-
tive iv administration must also be considered.
Clinical presentation in these patients can be divid-
ed into two distinct groups. Eight patients presented
with intraoperative cardiorespiratory collapse follow-
ing femoral reaming,1 insertion of intramedullary
alignment guide1 or cemented prosthesis,2 or after
tourniquet release.4 Hypotension was recorded in
seven to eight of these cases, with three progressing to
electromechanical dissociation. Hypoxia was noted in
five patients. Two patients in this group died.
Six patients presented in the postoperative period,
varying from immediately postoperatively to four
hours later. Only one patient had been unstable intra-
operatively, with transient hypotension following
tourniquet release. Postoperatively, all patients devel-
oped neurological signs, specifically disorientation or
unconsciousness, and hypoxemia. In two patients this
was associated with hypotension, and another devel-
oped signs of cardiac ischemia (segment depression on
ECG). Three patients in this group died, and one had
poor neurological recovery.
In trauma patients, respiratory signs are seen in
75%, which may progress to respiratory failure in
10%.7,22 Neurological changes, seen in up to 86% of
patients, may present as diffuse encephalopathy
and/or with focal abnormalities.2 3 Petechiae, which
represent dermal fat embolization, are seen in 50–60%
of cases usually after 24–48 hr.2 2 They are usually dis-
tributed to the upper body–chest, neck and conjuncti-
vae. Resolution occurs within seven days. Other
non-specific symptoms, such as tachycardia, fever, reti-
nal changes, jaundice and lipuria may also be seen.7,22
There are no universal criteria used for diagnosis.
Those published by Gurd et al. in 19702 2 divided the
symptoms into major and minor features for diagnosis
(Table I). These criteria have been criticized for not
specifically assessing oxygenation with ABG that may
be a useful early indicator of FES.5 Schonfeld’s scoring
system incorporated this6 (Table II) and Lindeque
proposed respiratory parameters as alternative criteria
for diagnosis of FES in trauma patients (Table III).
Our patient had clinical evidence of FES by both
Gurd’s and Schonfeld’s criteria.
The diagnosis of FES is clinical, usually one of
exclusion. Differential diagnoses of cardiorespiratory
and neurological dysfunction include myocardial
infarction, cardiogenic shock, infection or sepsis, pul-
monary embolism, cerebrovascular accident or meta-
bolic disturbance. In this patient with known ischemic
No investigation is 100% specific for the syndrome.
Laboratory tests should include ABG, complete blood
count to identify anemia or thrombocytopenia, and a
coagulation profile to reveal bone marrow thrombo-

22
CANADIAN JOURNAL OF ANESTHESIA
ritories,2 5 as was the case in our patient. This “water-
shed” distribution of the MRI abnormalities presum-
ably represents a tendency for the embolic particles to
lodge in the small terminal branches of the major
hemispheric vessels.
TABLE I Gurd’s criteria for diagnosis of FES 2 2
Major criteria
Petechial rash
(one necessary for diagnosis) Respiratory insufficiency
Cerebral involvement
Minor criteria
(four necessary for diagnosis) Fever >39.4°C
Retinal signs – fat or petechiae
Tachycardia >120 beats·min– 1
Identification of at risk patients preoperatively is
important. Medically, those with pre-existing cardiac
or respiratory disease have decreased physiological
reserve, and are at greater risk of cardiorespiratory col-
lapse following an embolic load. Patients with metasta-
tic disease, myeloblastic disorders, rheumatoid
arthritis, collagen vascular disease or osteopenia are at
risk due to enlargement and increased liquid marrow
content of the medullary canals.1 9 The risk may be
increased in bilateral surgical procedures, presumably
due to increased embolic load. A patent foramen ovale,
present in up to 34% of people,2 6 provides a means of
paradoxical fat embolization in the event of pulmonary
hypertension and increased right atrial pressure.
Intraoperatively, in addition to standard monitor-
ing in high-risk patients or procedures, use of an arte-
rial line for continuous hemodynamic and blood gas
monitoring is recommended.2 1 Intraoperative trans-
esophageal echocardiography may detect fat
embolization and monitor intraoperative volume sta-
tus and cardiac function.2 1 Pulmonary artery pressure
monitoring has also been recommended, both as a
diagnostic and prognostic tool, in bilateral proce-
dures.2 1 Although more invasive monitoring improves
detection of embolization and its sequelae, there is no
evidence to suggest these interventions improve out-
come in FES.
Jaundice
Renal signs – anuria or oliguria
Laboratory findings
Thrombocytopenia
(one necessary for diagnosis) Anemia
High erythrocyte sedimentation
rate
Fat macroglobulinemia
FES=fat embolism syndrome.
TABLE II Schonfeld’s criteria for diagnosis of FES6
Cumulative score >5 required for diagnosis
Petechiae [5]
Chest x-ray changes [4]
Hypoxemia [3]
Fever [1]
Tachycardia [1]
Tachypnea [1]
Confusion [1]
FES=fat embolism syndrome.
TABLE III Lindeque’s criteria for diagnosis of FES5
1
2
3
4
Sustained pO <8 kPa
Sustained pC2O of >7.3 kPa or a pH <7.3
2
Sustained respiratory rate >35·min– 1 despite sedation
Increased work of breathing: dyspnea, accessory muscle use,
tachycardia and anxiety.
Close postoperative monitoring should identify
early signs of respiratory deterioration, using routine
pulse oximetry, regular blood gas analysis and supple-
mental oxygen.2 1
FES=fat embolism syndrome.
Modification of surgical technique may lessen the risk.
Increased intramedullary pressure has been implicated as
the source of fat emboli due to fat intravasation from the
bone marrow cavity.7 The shift towards the use of total
condylar components from long-stemmed prostheses
obviates the need for extensive intramedullary manipula-
tion. However, intramedullary alignment guides have
also been implicated.1 8 Venting of the femoral shaft and
lavage of the medullary canal minimizes increases in
intramedullary pressure and the amount of fatty bone
marrow present.2 1 Methylmethacrylate cement insertion
results in pressurization of the intramedullary fat and
direct vasodilatation leading to transient hypotension.2 7
The presence of a tourniquet does not prevent emboliza-
tion and FES but may modulate the time course of
events and limits the vasodilatory consequences of the
cement insertion.2 1
plastin- or complement-mediated coagulopathies.2 3
Any other treatable causes of neurological dysfunc-
tion, e.g., toxic, metabolic or infectious, must be
excluded.2 4 Urinalysis may reveal fat globules but this
is non-specific. The ECG may be normal, reflect right
heart strain or may reveal ischemia.7 The chest radi-
ograph is non-specific, with diffuse fluffy bilateral
infiltrates and opacities consistent with increased cap-
illary permeability and edema.7 MRI has been shown
to be more sensitive than CT in the identification of
cerebral FES, with the abnormalities typically appear-
ing as hyperintense lesions on T2-weighted, proton
density and FLAIR sequences, characteristically locat-
ed along the boundary zones of the major vascular ter-

Jenkins et al.: FAT EMBOLISM SYNDROME
23
5 Lindeque BGP, Schoeman HS, Dommisse GF, Boeyens
To prevent hypoxemia intraoperatively, 100% oxygen
administration during cementing and prosthesis inser-
tion in patients at risk has been recommended anecdo-
tally although there are no prospective studies
confirming this.2 1Fluid therapy to prevent hypovolemia
is also important, as canine studies have shown that
acute hemodynamic changes during prosthesis insertion
and fat embolism may increase in hypovolemic states.2 7
Recent reports of intraoperative cardiac instability due
to fat embolization have recommended maintenance of
perfusion pressure with vasoconstrictors to preserve
right heart perfusion and function.2 3
Treatment is essentially supportive, consisting of
cardiovascular and respiratory resuscitation and stabi-
lization. No specific drug therapy for FES is currently
recommended. Small prospective randomized con-
trolled studies of steroid prophylaxis in patients with
long bone fractures have indicated both a decrease in
development of FES and a decreased incidence of
hypoxemia in steroid treated groups compared to con-
trols. However, dose regimens were not standardized,
varying from 1.5 mg·kg–1 methylprednisolone 8
hourly to 30 mg·kg– 14 hourly.4, 5In a canine model of
cemented arthroplasties, Byrick et al. demonstrated no
attenuation of hemodynamic or prostanoid response
in acute lung injury following treatment with methyl-
prednisolone 30 mg·kg– 1iv 20 min prior to reaming.2 8
Other drugs, such as heparin, iv alcohol, dextrans and
hypertonic dextrose, have not proved to be of any
benefit and may be detrimental.21
MC, Vlok AL. Fat embolism and the fat embolism syn-
drome. A double-blind therapeutic study. J Bone Joint
Surg Br 1987; 69: 128–31.
6 Schonfeld SA, Ploysongsang Y, DiLisio R, et al. Fat
embolism prophylaxis with corticosteroids. A prospec-
tive study in high-risk patients. Ann Intern Med 1983;
99: 438–43.
7 Johnson MJ, Lucas GL. Fat embolism syndrome.
Orthopedics 1996; 19: 41–50.
8 Weiss SJ, Cheung AT, Stecker MM, Garino JP, Hughes
JE, Murphy FL Jr. Fatal paradoxical cerebral emboliza-
tion during bilateral knee arthroplasty. Anesthesiology
1996; 84: 721–3.
9 Hume Adams J, Graham DI, Mills E, Sprunt TG. Fat
embolism and cerebral infarction after use of methyl-
methacrylate cement. Brit Med J 1972; 3: 740–1.
10 Letournel E, LaGrange J. Total knee replacement with
the “LL” type prosthesis. Clin Orthop 1973; 94:
249–56.
11 Mallory TH, Kolodzik S. Fat embolization after total
knee replacement (Letter). JAMA 1976; 236: 1451.
12 Browne CH. A case of fat embolism following Shiers
arthroplasty. Postgrad Med J 1976; 52: 247–9.
13 Bisla RS, Inglis AE, Lewis RJ. Fat embolism following
bilateral total knee replacement with the Guepar pros-
thesis. A case report. Clin Orthop 1976; 115: 195–8.
14 Lachiewicz PF, Ranawat CS. Fat embolism syndrome
following bilateral total knee replacement with total
condylar prosthesis: report of two cases. Clin Orthop
1981; 160: 106–8.
15 Zimmerman RL, Kroner III LF, Blomberg DJ, Nollet
DJ. Fatal fat embolism following total knee arthroplas-
ty. Minn Med 1983; 66: 213–6.
16 Orsini EC, Richards RR, Mullen JMB. Fatal fat
embolism during cemented total knee arthroplasty: a
case report. Can J Surg 1986; 29: 385–6.
In summary FES is a significant cause of mortality
and morbidity not only in trauma patients but increas-
ingly recognized in association with elective orthope-
dic surgery. Two main presentations are seen in knee
arthroplasties: intraoperative cardiorespiratory col-
lapse, or postoperative respiratory and cerebral dys-
function. A high index of suspicion is required for
early clinical diagnosis in at-risk patients.
17 Byrick RJ, Forbes D, Waddell JP. A monitored cardio-
vascular collapse during cemented total knee replace-
ment. Anesthesiology 1986; 65: 213–6.
References
18 Caillouette JT, Anzel SH. Fat embolism syndrome fol-
lowing the intramedullary alignment guide in total
knee arthroplasty. Clin Orthop 1990; 251: 198–9.
19 Monto RR, Garcia J, Callaghan JJ. Fatal fat embolism
following total condylar knee arthroplasty. J
Arthroplasty 1990; 5: 291–9.
1 Bulger EM, Smith DG, Maier RV, Jurkovich GJ. Fat
embolism syndrome. A 10-year review. Arch Surg
1997; 132: 435–9.
2 Robert JH, Hoffmeyer P, Broquet P-E, Cerutti P, Vasey
H. Fat embolism syndrome. Orthop Rev 1993; 22:
567–71.
20 Hall TM, Callaghan JJ. Fat embolism precipitated by
reaming of the femoral canal during revision of a total
knee replacement. A case report. J Bone Joint Surg Am
1994; 76: 899–903.
3 Fabian TC, Hoots AV, Stanford DS, Patterson CR,
Mangiante EC. Fat embolism syndrome: prospective
evaluation in 92 fracture patients. Crit Care Med 1990;
18: 42–6.
21 Enneking FK. Cardiac arrest during total knee replace-
ment using a long-stem prosthesis. J Clin Anesthesia
1995; 7: 253–63.
4 Kallenbach J, Lewis M, Zaltzman M, Feldman C,
Orford A, Zwi S. ‘Low-dose’ corticosteroid prophylaxis
against fat embolism. J Trauma 1987; 27: 1173–6.

24
CANADIAN JOURNAL OF ANESTHESIA
22 Gurd AR, Wilson RI. The fat embolism syndrome. J
Bone Joint Surg Br 1974; 56 B: 408–16.
23 Byrick RJ. Fat embolism and postoperative coagulopa-
thy (Editorial). Can J Anesth 2001; 48: 618–21.
24 Jacobsen DM, Terrance CF, Reinmuth OM. The neuro-
logic manifestations of fat embolism. Neurology 1986;
36: 847–51.
25 Stoeger A, Daniaux M, Felber S, Stockhammer G,
Aichner F, zur Nedden D. MRI findings in cerebral fat
embolism. Eur Radiol 1998; 8: 1590–3.
26 Hagen PT, Scholz DG, Edwards WD. Incidence and size
of patent foramen ovale during the first ten decades of
life: an autopsy study of 965 normal hearts. Mayo Clin
Proc 1984; 59: 17–20.
27 Berman AT, Price HL, Hahn JF. The cardiovascular
effects of methylmethacrylate in dogs. Clin Orthop
1974; 100: 265–9.
28 Byrick RJ, Mullen JB, Wong PY, Kay JC, Wiggleworth
D, Doran RJ. Prostanoid production and pulmonary
hypertension after fat embolism are not modified by
methylprednisolone. Can J Anaesth 1991; 38: 660–70.
Joshua Tree National Monument – California