Article Text

PDF

Prolonged coma due to cerebral fat embolism: report of two cases
  1. Leonidas Gregorakos,
  2. Katerina Sakayianni,
  3. Despina Hroni,
  4. Victoria Harizopoulou,
  5. Nikos Markou,
  6. Fani Georgiadou,
  7. Maria Adamidou
  1. Intensive Care Unit, Athens Trauma Hospital, Athens, Greece
  1. Correspondence to: Dr L Gregorakos, 51 Agias Sofias Str, 10444 Athens, Greece

Abstract

Fat embolism syndrome remains a rare, but potentially life threatening complication of long bone fractures. The true incidence is difficult to assess as many cases remain undiagnosed. Cerebral involvement varies from confusion to encephalopathy with coma and seizures. Clinical symptoms and computed tomography are not always diagnostic, while magnetic resonance imaging is more sensitive in the detection of a suspected brain embolism. Two cases of post-traumatic cerebral fat embolism, manifested by prolonged coma and diffuse cerebral oedema, are presented. The clinical course of the disease as well as the intensive care unit management are discussed.

Statistics from Altmetric.com

Fat embolism syndrome (FES) remains a rare, but potentially life threatening complication of long bone fractures, characterised by pulmonary insufficiency, neurological dysfunction, fever and petechial rash, usually occurring during the first 48 hours after injury.1 The true incidence of this syndrome cannot be accurately assessed as many subclinical forms remain unrecognised. It varies from 0.5% to 30% of fractured patients2 with higher rates in multiply injured patients and presents a mean mortality rate of 10%.3 Cerebral involvement has been frequently reported and seems to aggravate the prognosis of FES.4 Clinical manifestations vary widely, from a simple alteration of vigilance (70%–85% of the cases), to seizures and coma.2 We present two cases of post-traumatic cerebral fat embolism, manifested by prolonged coma and diffuse cerebral oedema.

Case reports

CASE 1

An 18 year old man who sustained an open fracture of the left tibia but no head injury, after a motorbike accident, was admitted to the emergency room of our hospital. Upon admission the patient was alert, oriented, normotensive, and eupnoeic. Neurological examination showed no abnormalities. Two hours later, external fixation of his fracture was performed under general anaesthesia. After complete recovery from anaesthesia he was transferred to the orthopaedic department; he was haemodynamically stable and had a normal respiratory pattern. Twelve hours postoperatively he developed signs of FES including fever (>39°C), tachycardia, petechiae over the shoulder area, hypoxaemia, oliguria, and thrombocytopenia. Six hours later he was unconscious, responding only to painful stimuli. Neurological examination demonstrated a Glasgow coma score of 5 (eye opening 1, motor response 3, verbal response 1), and a normal reaction of both pupils to light. An immediate endotracheal intubation was performed because of unconsciousness and respiratory insufficiency. The patient was transferred to the intensive care unit (ICU) and mechanical ventilation initiated. Chest radiography on ICU admission demonstrated diffuse pulmonary infiltrates. Cerebral computed tomography, performed shortly after mental status deterioration, revealed diffuse brain oedema. Antioedema treatment was instituted, and mild hypocapnia was induced. Retinal examination performed on day 2 demonstrated characteristic cottonwool spots across the vascular beds, which are indicative of fat embolism. Magnetic resonance imaging (MRI), performed on day 3 showed multiple areas of increased intensity in the cerebral white matter, which confirmed the diagnosis of cerebral fat embolism. The patient remained under mechanical ventilation, general supportive care in the ICU, and regular neurosurgical evaluation. High doses of methylprednisolone (30 mg/kg/8 hours intravenously) were administered for three doses, followed by prednisolone 50 mg/day, intravenously, tapering off after one week. A tracheostomy was performed and mechanical ventilation continued for 22 days. Thereafter, he was breathing spontaneously through the tracheostomy. One month after initial admission he started opening his eyes to intense stimuli, and computed tomography showed resolution of cerebral oedema. The tracheostomy tube was removed on the 35th day of hospitalisation, by which time he had a progressive resumption of consciousness and improving responsiveness to external stimuli. He was in the ICU for six weeks, and his entire hospital stay after the accident was two months. He was discharged from the hospital without any neurological sequelae. MRI performed before discharge detected no abnormalities.

CASE 2

A 20 year old man sustained bilateral closed tibia fractures and abdominal trauma after a traffic accident. Upon his admission to the emergency room no signs of craniocerebral injury were present. Neurosurgical examination detected no neurological abnormalities, with a Glasgow coma score of 15. Blood pressure was 140/90 mm Hg and heart rate 120 beats/min. As indicated by intraperitoneal lavage, an urgent laparotomy was performed, which showed rupture of the mesenterium. Limited small intestine resection and enteroentero anastomosis were performed. At the same time the patient underwent closed reduction and cast immobilisation of his fractures. After surgical intervention and complete recovery from anaesthesia, the patient was transferred extubated to the ICU. Six hours after admission the patient had a fever (39oC), became cyanotic, comatose, and had seizures. His Glasgow coma score was 4 (eye opening 1, motor response 2, verbal response 1), while pupils were equal in size and reactive to light. He was immediately intubated and mechanically ventilated. In addition, signs of systemic fat embolism, such as tachycardia, conjunctival petechiae, hypoxaemia, and diffuse pulmonary infiltrates on chest radiography, were present. Laboratory analysis revealed thrombocytopenia and anaemia. Arterial blood gas analysis showed a partial oxygen tension of 7.73 kPa (58 mm Hg) and partial carbon dioxide tension of 4.27 kPa (32 mm Hg).

On day 2, computed tomography of the brain showed diffuse cerebral oedema, without any change in neurological status. He remained under mechanical ventilation for 15 days, receiving the same treatment as case 1. MRI was performed on day 5, which showed diffuse high intensity lesions in the subcortical and periventricular white matter. During his stay at the ICU he suffered two episodes of nosocomial pneumonia. Serial blood culture and protected brush specimen cultures were obtained. Pseudomonas aeruginosa was isolated from bronchoalveolar lavage fluid and an appropriate antibiotic treatment was given, according to sensitivity. He was extubated on day 15 and was transferred to the neurosurgical department five days later with moderate disorientation. MRI of the brain on day 25 demonstrated a substantial resolution of the previous hyperintense lesions. He was finally discharged from the hospital 35 days after the accident without any neurological deficit.

Discussion

FES with fat droplets larger than 8 μm in diameter occurs in more than the 90% of patients with long bone fractures. This syndrome denotes clinical or subclinical respiratory insufficiency, and usually runs a mild course and responds well to measures for ventilatory support.1 In our patients the diagnosis was established by Gurd's criteria,5 the characteristic fat globules on retinal examination, and the findings on brain computed tomography and MRI. There are two theories explaining the pathogenesis of FES. According to the “mechanical” theory, free fat particles from the bone marrow enter torn vein sinusoids at the site of the fracture and embolise the pulmonary arterioles. The “chemical” theory suggests that fat emboli arise from plasma fat when, through some type of systemic stimulus associated with trauma and other medical conditions, chylomicrons coalesce and fuse to form larger fat globules.5 Victims of the fulminant form of the syndrome on postmortem examination present occlusion of small blood vessels by fat emboli with areas of brain microinfarction and haemorrhage.6, 7 Due to cerebral fat emboli, the brain often appears oedematous and shows an inflammatory reaction while the numerous petechiae can cover the surface of the brain.8 Endothelial damage results from toxic free fat and capillary obstruction by fat globules with associated platelet aggregation, release of vasoactive substances, and development of coagulopathy.9

The presence or the reopening of a patent foramen ovale and a right to left shunt due to pulmonary hypertension is associated with an increased risk for systemic manifestations of FES.10 However, several studies have failed to demonstrate any intracardiac shunts in patients suffering from this syndrome.8, 11 Published studies remain undecided whether a patent foramen ovale or other intracardiac defects are prerequisites for the development of cerebral symptoms.

Our patients are characteristic cases of cerebral fat embolism, which was followed by a severe course of disease in the ICU. They responded to conventional treatment and administration of corticosteroids. Among several pharmacological treatments, only steroids have proved to be beneficial in the prophylaxis and treatment of FES, both in high and low doses.1214 The mode of action has not been elucidated, but seems to be related to their anti-inflammatory and antiadhesive effects.

The long stay in the ICU was due to the increased intracranial pressure, following cerebral oedema. These cases suggest that cerebral oedema plays a major part in the neurological deterioration in the fulminant type of FES and that cerebral computed tomography is indicated in such patients.15 Computed tomography is also useful in excluding traumatic cerebral involvement in multiple trauma patients presenting with deterioration in their mental state, but is not specific for the diagnosis of cerebral fat embolism.16 MRI is a better diagnostic tool for the confirmation of brain embolism, as it is more sensitive in the detection of degenerative and vascular injuries and non-haemorrhagic contusion. In addition, it seems to correlate well with the clinical neurological course.17, 18

The patients described above, despite the initial severe neurological syndrome and prolonged coma, had complete cerebral recovery. Several studies report that cerebral dysfunction associated with FES appears to be reversible in the majority of the cases, so physicians should not give up hope even in the most severe forms of the syndrome.8, 10, 11, 19 It is likely that fat emboli of very small diameter cause reversible ischaemia and transient perivascular oedema, allowing recovery from cerebral dysfunction.17

Cerebral fat embolism should always be suspected when severe neurological dysfunction appears in a patient with long bone fractures, in the absence of an initial major head injury. Although the diagnosis may be straightforward in patients with isolated long bone fractures, it may be difficult in the multiple trauma patient with head injury or under sedation. In addition, hypoxia may be associated with pulmonary contusion and infection and petechiae may occur after blood transfusion.

Many authors have suggested that early open reduction and surgical stabilisation of long bone fractures may reduce the incidence of FES.2, 10 Repeated manipulation of the fractured fragments may further stimulate the release of bone marrow fat into the circulation.10 Similarly, the type of surgical fixation seems to influence the incidence of fat embolisation. External fixation and plate osteosynthesis have several advantages compared with intramedullary nailing techniques, which further increase intramedullary pressure and promote fat emboli release.10, 20 However, the syndrome is frequently observed despite prompt surgical intervention.

As pathophysiology of the syndrome still remains obscure, attention should be paid to prevention and early recognition of this entity. Optimal immobilisation, adequate fluid and blood replacement, meticulous monitoring, and use of steroids are principles commonly accepted in current treatment.2, 10

Acknowledgments

Contributors

All of the authors listed above have contributed substantially to the collection of data and writing of the paper.

Leonidas Gregorakos is the guarantor for the paper.

References

View Abstract

Footnotes

  • Conflict of interest: none.

  • Funding: none.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.