USING transesophageal echocardiography, echogenic material can be shown in the right ventricle during many orthopedic procedures. 1Recently, transcranial Doppler–recorded microembolic signals have been described during femoral fracture fixation, suggesting that medullary pressurization may be associated with systemic microembolism and pulmonary fat and marrow emboli. 2
Although reamed, locked intramedullary nailing is the procedure of choice for most femoral shaft fractures, some investigators advocate the use of small-diameter, locking, unreamed nails to decrease the risk of fat emboli. 3,4We report a case of severe intraoperative fat embolism after an unreamed locked intramedullary nailing.
A previously healthy 17-yr-old woman was injured on her right side by a motor vehicle. At arrival to our hospital, she was fully awake and alert. There was no evidence of head, chest, or abdominal injuries. She reported pain in the right thigh and the pelvis and had an 8-cm laceration on her right hip. Her heart rate was 130 beats/min (electrocardiography showed sinus tachycardia) and her respiratory rate was 24 breaths/min. Blood pressure was 115/75 mmHg. Laboratory investigations showed mild acidosis (pH; 7.32; partial pressure of car-bon dioxide [Pco2]: 42 mmHg; partial pressure of oxygen [Po2]: 183 mmHg; fraction of inspired concentration of oxygen [Fio2]: 0.4) and hemoglobin concentration was 11.7 g/dl, with a platelet count of 357 × 109/l. Radiographs showed a stable lateral-compression–type pelvic fracture and a displaced, transverse, right femoral shaft fracture. Radiographs of cervical, thoracic, and lumbar sections of the spine were normal, as was the chest radiograph.
Fourteen hours after injury, oxygen saturation as measured by pulse oximetry (Spo2) before induction of general anesthesia was 98% breathing oxygen (Fio2= 0.5). The patient was alert, cooperative, and recalled the accident and all events after injury (Glasgow Coma Scale = 15). Vital signs were normal; however, heart rate was 130 beats/min. Anesthesia was induced using midazolam, fentanyl citrate, and propofol. Tracheal intubation was facilitated using succinylcholine. The patient underwent ventilation using oxygen in air. Anesthesia was maintained with isoflurane, intravenous fentanyl, and morphine. Vecuronium was used to provide muscle relaxation.
The patient’s hip laceration was irrigated and debrided. Using manual traction, the femoral fracture was reduced, and fixation wasachieved using a titanium, unreamed AO femoral nail that was statically locked (9 × 380 mm; Synthes, Mississauga, Canada). The nail passed across the fracture site with minimal difficulty, and no guidewire or reaming was used. During passage of the nail across the fracture site, approximately three brief episodes of oxygen desaturation (Spo2of 83%) were noted. Each episode was no longer than 3 min, and Spo2returned to 98% (Fio2= 1.0). The Pao2value after one episode was 94 mmHg (Fio2= 1.0).
After surgery, the patient did not regain consciousness. At examination, she was unresponsive to pain and displayed Babinski reflexes bilaterally and generalized hyperreflexia, with increased tone in all extremities. No lateralizing neurologic findings were detected. Eight hours after surgery, she had a generalized tonic–clonic seizure, which lasted approximately 1 min. Electroencephalography showed slow wave activity that was consistent with moderate to severe encephalopathy. Computerized tomography (CT) was normal. Magnetic resonance imaging (MRI) was performed. The T2-weighted MRI images showed multiple punctate lesions in the subcortical white matter, the centrum semiovale, and the basal ganglia (fig. 1). Transesophageal echocardiography showed good ventricular function and no intracardiac defect, specifically no patent foramen ovale.
Within 24 hours of injury, the patient’s hemoglobin concentration decreased to 78 g/l and platelet count was 131 × 109/l. Chest radiography showed bilateral, ill-defined pulmonary infiltrates. A pulmonary artery catheter was inserted, and the occluded pressure was 12 mmHg and pulmonary artery pressure was 43/25 mmHg. No petechiae were noted. The patient was comatose and required ventilatory support. She had a low-grade fever without evidence of infection. Respiratory function improved over 4 days, with minimal change in neurologic status. On day 6 after surgery, her neurologic state improved (Glasgow Coma Scale = 6). She was extubated 16 days after surgery and was able to answer questions by gesturing (Glasgow Coma Scale = 11). Twenty-six days after injury, the patient was discharged to a rehabilitation hospital.
Five months later, neurologic function was steadily improving (Glasgow Coma Scale = 15), and she had returned to school. The family noted significant personality and cognitive changes and limited conversational skills. Another MRI was performed. The T2-weighted images (fig. 2) showed resolution of the hyperintense focal subcortical lesions; however, a diffuse periventricular white matter abnormality was noted. The femoral fracture healed uneventfully.
Although intraoperative cardiovascular collapse and death has been reported in one case of prophylactic femoral nailing, 5we report a case of intraoperative cerebral microembolism after unreamed intramedullary rod fixation. The classic triad of fat embolism syndrome consists of respiratory distress, neurologic dysfunction, and petechial rash. 6Factors that determine the “intravasation” of marrow products into the venous circulation include disruption of the medullary contents, pressurization of the marrow cavity, and the availability of open vessels. These factors are present during intramedullary rod fixation.
In this case, intraoperative hypoxemia and hypotension were transient and do not explain the profound and lasting neurologic and cognitive dysfunction. Although preoperative desaturation suggested fat embolism syndrome, no neurologic findings were detectable before anesthesia. In fat embolism syndrome, patients may have focal neurologic signs or generalized encephalopathy (confusion, seizures, coma). 7No intracardiac defect was found; however, paradoxical emboli can occur through a patent foramen ovale. 8In this patient, transpulmonary passage of fat occurred. This has been shown in animal models in which cerebral intravascular fat also is found. 9
The advantages of the interlocking of nails in the treatment of diaphyseal femoral fractures are well-established. However, one of the complications of intramedullary nailing includes embolization of marrow contents, and unreamed femoral nails are advocated to minimize this risk because of the resultant reduction in intramedullary pressure. 4Medullary pressurization and embolization depend on surgical factors, such as the velocity of nail insertion or the gap between nail and cortical bone when entering the distal fragment. 3Our case emphasizes that, although there may be a reduced incidence, severe fat embolism can occur when unreamed, locked femoral nails are used.
Neither the degree nor the duration of hypoxemia in this case would be expected to cause severe cerebral injury in a young patient. It is important to emphasize that MRI is necessary to show the characteristic cerebral lesions in the acute stage of fat embolism syndrome. The CT scan in such cases is normal. 10The distribution of lesions shown during MRI (fig. 1) is characteristic and suggests transpulmonary microemboli, with resultant acute perivascular edema. Takahashi et al . 11characterized the lesions seen during MRI after cerebral fat embolism as high intensity signals located in areas of brain that are perfused by perforating arteries. Their disappearance during subsequent MRI, coincident with resolution of the neurologic signs and symptoms, emphasized the clinical importance. 11Six of 11 patients in the study became alert within 10 days; however, one patient had severe, persisting neurologic impairment for 2 months (55 days), with persistent abnormalities seen during MRI. In our patient, the first MRI scan (fig. 1) showed characteristic hyperintense lesions, suggesting focal perivascular edema. Five months after injury, the patient had marked neurocognitive deficits, and MRI (fig. 2) showed lasting signal intensity abnormalities and morphologic lesions. The characteristic MRI pattern had changed to a nonspecific pattern, consistent with diffuse periventricular ischemic injury. This emphasizes that the microvascular lesions in the acute phase can result in long-term neurologic impairment.
The authors thank Dr. Walter Kucharcyk, Professor, Chairman, Department of Diagnostic Imaging, University of Toronto, for advice and review of the MRI scans.