clinical investigations in critical care High-Resolution CT Findings in Mild Pulmonary Fat Embolism*

Transcription

1 clinical investigations in critical care High-Resolution CT Findings in Mild Pulmonary Fat Embolism* Katerina Malagari, MD; Nikos Economopoulos, MD; Christophoros Stoupis, MD; Zoe Daniil, MD; Spyros Papiris, MD, FCCP; Nestor L. Müller, MD; and Dimitrios Kelekis, MD Objective: The aim of this article is to describe the high-resolution CT (HRCT) findings in mild cases of fat embolism syndrome (FES). Material and methods: Nine patients with FES were examined with HRCT of the lungs (collimation, 1 mm/edge-enhancement algorithm). The median age of the patients was 26 years (range, 17 to 35 years). Five cases were included prospectively, and four cases were reviewed retrospectively. Of the major clinical criteria for FES, respiratory signs were present in six patients, CNS signs were present in two patients, and petechiae was present in six patients. HRCT patterns were recorded and analyzed. The type of injury and FES-associated clinical findings were also recorded. Results: HRCT findings included ground-glass opacities in seven patients, associated with thickened interlobular septa in five patients and a patchy distribution resulting in a geographic appearance in four patients. A nodular pattern was observed in two patients. Resolution of the abnormalities occurred within 16.4 days (range, 7 to 25 days). Conclusion: The HRCT findings of mild fat embolism consist of bilateral ground-glass opacities and thickening of the interlobular septa. Centrilobular nodular opacities are present in some patients. (CHEST 2003; 123: ) Key words: fat embolism; high-resolution CT; pulmonary fat embolism; trauma Abbreviations: FES fat embolism syndrome; FIS fracture index score; HRCT high-resolution CT; PFES pulmonary fat embolism syndrome Fat embolism occurs in nearly all patients ( 90%) with bone fractures during orthopedic prosthesis procedures and rarely in other pathologic conditions. 1,2 Although fat emboli can virtually reach any organ in the body, the results of the embolic shower are most often evident in the lungs, brain, *From the 2nd Department of Radiology (Drs. Malagari, Economopoulos, Stoupis, and Kelekis), National and Kapodistrian University of Athens; Clinic of Pulmonology and Intensive Care (Drs. Daniil and Papiris), Evangelismos General Hospital, National and Kapodistrian University of Athens, Athens, Greece; and Department of Radiology (Dr. Müller), Vancouver General Hospital, University of British Columbia, Vancouver, BC, Canada. Manuscript received January 24, 2002; revision accepted June 13, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( permissions@chestnet.org). Correspondence to: Katerina Malagari, MD, 19 Monis Kyccou, Athens, Greece; kmalag@acn.gr and skin. Approximately 3 to 4% will acquire the classical triad of the fat embolism syndrome (FES), which consists of respiratory distress, cerebral abnormalities, and petechial hemorrhages, while the rest of the patients will remain asymptomatic. 1,3 Clinically severe cases make up 10 to 20% of all patients in whom the diagnosis can be made, 1,3 and have been associated with mortality rates as high as 20% in some series. 3 Diagnosis is crucial; it is mainly based on clinical criteria and may be masked by associated pathology especially in the multitrauma victim. Imaging plays a key role in the differential diagnosis and assists in the recognition of FES. For editorial comment see page 982 In this article, we describe the high-resolution CT (HRCT) findings in nine patients with mild pulmo Clinical Investigations in Critical Care

2 Table 1 Patient Demographic Data, Type and Mechanism of Injury, and Timing of FES Signs* Patient No. Age, yr/gender Type of Injury Bone Fracture FIS Day After Injury, No. 1 21/Male MVA Femoral shaft /Male MVA Femoral shaft /Female MVA Femoral shaft /Male MVA Femoral shaft /Male MVA Femoral shaft /Male Fall from height Femoral neck, pelvis, calcaneus /Male Industrial crush injury Femoral neck, pelvis, radius, ulna /Male MVA Femoral neck, pelvis /Male MVA Femoral shaft, tibia, fibula 7 3 Mean values 2.3 Median values 26 *MVA motor vehicle accident. From Shier et al. 7 nary FES (PFES), and discuss the correlation withthe pathophysiology of the syndrome. To our knowledge, there are only three articles describing CT findings of pulmonary fat embolism 4 6 : two articles were single case reports, and one article described the findings in six patients with severe fat embolism. Materials and Methods A total of nine patients (eight males and one female; median age, 26 years; range, 17 to 35 years) with documented FES were included in this study. The first five cases were prospectively selected because of the mild severity of imaging findings and indolent clinical course, and the following four cases were retrospectively collected after review of the medical records of patients treated in our hospitals for FES in the last 3 years with the same criteria. No postoperative cases are included in this patient group. All patients reached the hospital within 12 h from the injury and had sustained multiple comminuted long-bone and pelvis fractures that are listed in detail in Table 1. To evaluate the severity of bone trauma related to the development of FES, the fracture index score (FIS) introduced by Shier et al 7 was recorded (Table 1). There was no clinical evidence of chest trauma, and the hospital admission chest radiographic findings were normal. None of the patients were in shock prior to the development of pulmonary dysfunction. No prophylactic treatment with corticosteroids was administered prior to chest CT. Diagnosis of FES was subsequently established clinically based on the criteria of Gurd and Wilson. 8 The diagnosis of FES requires the presence of at least one major and four minor criteria. Major clinical criteria include respiratory involvement, cerebral involvement, and petechial rash. Minor criteria include retinal changes, pyrexia ( 39 C), tachycardia ( 110 beats/min), fat present in the urine, and jaundice. Laboratory features are also included in the list of minor criteria, notably thrombocytopenia, high sedimentation rate, fat macroglobulinemia, a sudden unexplainable drop of hematocrit or platelet levels, and fat globules in sputum. In all patients, FES-associated symptoms developed within 1 to 3 days from the time of injury (day 2 to day 3 after trauma) [Table 1]. CT of the chest was performed the same day or the day following the onset of symptoms. Respiratory symptoms included dyspnea (n 5) and/or a drop in Pao 2 below 80 mm Hg (n 5) [Table 2]. Signs of mild confusion in the absence of CT evidence of brain trauma were an indication of FES in two patients, while an unexplained tachycardia combined with other relevant symptoms raised clinical suspicion for FES in eight patients (Table 2). Delayed appearance of petechiae was observed in six of nine patients recognized after a thorough search at the buccal mucosa and conjunctiva after the third day after injury. Skin petechiae Table 2 Physical Signs of FES* Patient No. Pao 2,mmHg Dyspnea Tachycardia Petechiae CNS Signs Retinal Signs Pyrexia Renal Signs 1 85 A P A P P P A 2 68 (room air) P P P A NE A P 3 92 P P P A NE P A 4 80 A A A P P P A 5 65 (room air) P P P P P A A 6 88 A P P P P P P 7 70 (room air) P P P P NE A P 8 55 (room air) A P A P P P A 9 58 (room air) P P P P A P P Total Mean *According to Gurd and Wilson. 8 P present; A absent; NE not examined. Of a total of nine patients. CHEST / 123 / 4/ APRIL,

3 were observed in only four of the patients and were located at the skin over the sternum and neck. Retinal changes at fundoscopy were revealed in five patients. Other signs of FES in our patients included pyrexia (n 6) and fat droplets in the urine (n 4) [Table 2]. All patients responded well to supportive pulmonary care and hydration, and none required intubation. Oxygen (3 to 4 L/min; fraction of inspired oxygen, 24% and 26%, respectively) was administered via Venturi mask in five of nine patients. Definitive orthopedic surgery for internal fixation of fractures was postponed in six patients until respiratory improvement. CT scans of the chest were performed on a Tomoscan SR 7000 Philips scanner (Philips Medical Systems; Eindhoven, Netherlands) using 1-mm collimation, and the images were reconstructed using an edge-enhancement algorithm. The scans were obtained during breathhold at end-inspiratory phase with 1.5-s scanning time and milliampere adjustments according to body weight. All patients were examined in supine position. Window width and levels appropriate for visualization of pulmonary parenchyma were used: window level 500 to 700; window width, 1,000 to 1,600 Hounsfield units (Table 3). Chest CT findings were interpreted by three observers (K.M., N.E., C.S.), who reached a decision by consensus on the pattern and distribution of abnormalities. Nodular opacities were evaluated according to size and location of the nodules in relation to the secondary pulmonary lobule. Distribution of lung findings was evaluated in the craniocaudal axis (apical/basal) and the axial plane (peripheral/central/patchy). Follow-up imaging was performed with daily chest radiographs to monitor potential deterioration. Follow-up CT was performed 7 to 25 days after the initial examination. Results HRCT demonstrated ground-glass opacities in seven patients (Figs 1, 2) and nodular opacities in two patients (Figs 3, 4). In five patients, the groundglass opacities were associated with interlobular septal thickening (Figs 1, 2). No airspace consolidation was observed in any of our cases. A peripheralsubpleural, nongravity-dependent distribution was noted in three patients with ground-glass opacities (Fig 2). In four patients, the ground-glass opacities had a patchy distribution with sharp margination between areas of involved and noninvolved lung Patient No. Table 3 HRCT Findings* Ground Glass Septal Thickening Nodular Opacities 1 A P P 2 P P A 3 P P A 4 P P A 5 P A A 6 P A A 7 P P A 8 P A A 9 A A P Total *See Table 2 for expansion of abbreviations. Of a total of nine patients. Figure 1. A 17-year-old man with a comminuted femur fracture. Top, A: HRCT scan obtained the second day after injury shows ground-glass opacities. Bottom, B: HRCT at a lower level demonstrates ground-glass opacities confined to some lobules with a sharp margination between areas of involved and noninvolved lung, resulting in a geographic appearance. Also noted is smooth and nodular interlobular septal thickening. Artery/bronchus diameter ratios are normal throughout the lung, and no bronchial cuffing is observed. (Fig 1, bottom, B), resulting in geographic appearance. Septal thickening was predominately nodular, confined to the areas of ground glass and was not associated with bronchial cuffing. In cases with a nodular pattern, nodules were 1 to 2 mm in diameter, and tended to be situated in the centrilobular regions, along the interlobular septa and along the interlobar fissures (Figs 3, 4). No zone predominance or gravity dependence was noted in the nodular pattern. No quantification of the extent of the abnormalities on HRCT was attempted due to the small number of patients. Follow-up chest radiographs did not show deterioration or development of consolidation in any of the patients. Follow-up HRCT was performed after normalization of arterial oxygen tension and complete resolution of the abnormalities on the chest radiographs. In two patients, follow-up HRCT revealed persistence of the nodules, which resolved on 1198 Clinical Investigations in Critical Care

4 Figure 2. A 19-year-old man at 2 days after femur shaft fractures. Top, A: HRCT scan obtained at the lower lung zones reveals a predominantly peripheral distribution of ground-glass opacities associated with smooth and nodular septal thickening. Bottom, B: HRCT obtained at a lower level shows relative sparing of some secondary lobules. Findings in this patient are more pronounced than those observed in the patient in Figure 1. the repeat CT a week later. Follow-up HRCT confirmed resolution of the ground-glass opacities after 7 to 15 days and resolution of the nodular opacities within 15 to 25 days. Mean resolution time was 16.4 days (range, 7 to 25 days). Discussion Trauma-related fat embolism is defined by the blockage of blood vessels from fat globules that are usually released from traumatic or iatrogenic injury to the long bones. 1 3 The lung is the most frequently affected organ, but reported incidence of pulmonary fat embolism varies considerably between different series; Heitzman 9 reported that some degree of pulmonary fat embolism regularly follows long-bone injury. The clinical syndrome, however, is reported to develop between 0.5% and 2% following fractures of the long bones (particularly the shafts of femurs and the tibias), reaching 5 to 10% after multiple Figure 3. A 21-year-old man with tachycardia and confusion the second day after injury. Top, A: HRCT obtained at the level of right upper lobe bronchus demonstrates multiple nodules 1 to 2 mm in diameter. The positive vessel sign (arrowhead) seen in some of the nodules indicates a centrilobular distribution, while other nodules are distributed along the interlobular septa and interlobar fissures. Bottom, B: HRCT obtained at the level of bronchus intermedius shows similar pattern and distribution findings. A positive vessel sign (arrowheads) and nodular thickening of interlobular septa and interlobar fissures are again noted. fractures including the pelvis with occasional visualization of fat droplets in veins by imaging techniques. 2,10,11 All our patients had sustained comminuted long-bone and pelvis fractures (Table 1). The probability of a fracture being associated with FES is well reflected by the FIS introduced by Shier et al. 7 Mild PFES may remain undetected; chest radiographic findings may be entirely normal as in the majority of our patients, while in more severe cases diffuse alveolar opacities may be seen. With the widespread availability of CT, more subtle findings may be recognized that may contribute to an earlier recognition of PFES in the proper clinical setting. The main finding was the presence of ground-glass opacities, which reflects less severe parenchymal involvement. This differs from severe fat embolism, which characteristically results in extensive bilateral CHEST / 123 / 4/ APRIL,

5 Figure 4. A 27-year-old man with hypoxemia and buccal petechiae 2 days after injury. Top, A: HRCT obtained just below the tracheal bifurcation reveals a predominantly nodular pattern. Note that bronchovascular bundles are thin and smooth. Bottom, B: HRCT obtained at the level of pulmonary veins shows similar pattern and severity of findings. airspace consolidation on the radiograph and CT. 4 In our series, airspace consolidation developed in none of the patients. The radiologic differential diagnosis includes lung contusion, pulmonary edema, and aspiration. Since our CT findings appeared at least 24 h after the injury, traumatic contusion was a less likely diagnosis. Pulmonary signs are usually the first manifestations of PFES and typically develop within 24 to 48 h from the time of injury. 1,3 This latent period facilitates the differentiation from pulmonary contusion. In addition, in our patients no bronchial cuffing or pleural effusion were present and the artery/bronchus diameter ratio was normal, assisting in the differentiation from increased pressure/fluid overload pulmonary edema. The characterization mild embolism refers to the indolent course, the mild physiologic impairment, and the absence of consolidations or ARDScompatible findings at HRCT. We believe that if CT had not been performed, these mild severity cases might have been missed. Moreover, CT has contributed in early recognition of FES, since in our series as well as in other studies, petechiae, a rather specific finding, seem to be rather subtle and have to be carefully looked for. 12,13 Nodular opacities were demonstrated in two of our patients. Nodules at HRCT in early fat embolism have been also reported by Heyneman and Müller, 5 while Arakawa et al 4 described small nodules of various sizes ( 10 mm) in five of six of their patients (83.3%) Although we lack pathologic correlation, these nodules most probably represent areas of edema, hemorrhage, and atelectasis. Complete resolution was observed in all our patients. However, development of microcalcifications is also reported in the literature. 6 The centrilobular distribution of some of the nodules observed in our series may indicate the vasculogenic origin of the initial insult suggested by the mechanical theory of the pathophysiology of PFES. Mechanical obstruction of lung capillaries is one of the two major pathways of PFES 1,9,12,13 ; fat globules released from intracellular fat from traumatized marrow enter injured venous structures at the fracture site. These globules act like embolic showers at the pulmonary capillaries or bypass the lungs to the systemic circulation embolizing many organs, notably the brain and skin. The biochemical pathway also plays an important role in FES. Increased vessel wall permeability is the final result of either pathway leading to extravasation, interstitial hemorrhage, alveolar wall damage with passage of blood and fat microglobules into the alveoli, and alveolar collapse. 1,9,12,13 Imaging findings therefore resemble increased permeability edema. 4,14 16 The geographic pattern of ground-glass opacities confined to a number of secondary pulmonary nodules while sparing neighboring ones could be related to the variation of lobular perfusion over time. 9 According to the latter physiologic phenomenon, more severe alterations occur at the lobules perfused at the time of the embolic event while the underperfused are relatively protected. We infer though, that with time, redistribution associated with gravity may subsequently distort this initially pathophysiologyrelated distribution. Although not observed in our patients possibly due to early detection and mild form of PFES, gravity dependence of the parenchymal abnormalities is reported by most authors. 1,2,4 6 Arakawa et al 4 noted that gravity dependence appears later in the course of the syndrome, and we can infer that it may be better related to the fate of extravascular fluid in the lung rather than to its site of origin. In conclusion, HRCT may reveal early and subtle cases of PFES. The HRCT findings of mild fat 1200 Clinical Investigations in Critical Care

6 embolism consist of bilateral ground-glass opacities and thickening of the interlobular septa. Centrilobular nodular opacities are present in some of the patients. HRCT findings of PFES reflect the pathophysiology of the syndrome and the ischemic and toxic effects on lung parenchyma. In mild cases, HRCT may suggest the diagnosis prior to the development of overt clinical manifestations. References 1 Levy DL. The fat embolism syndrome: a review. Clin Orthop 1990; 261: Aoki N, Soma K, Shinedo M, et al. Evaluation of potential fat emboli during placement of intramedullary nails after orthopedic fractures. Chest 1998; 3: Mudd KL, Hunt A, Matherly RC, et al. Analysis of pulmonary fat embolism in blunt force fatalities. J Trauma 2000; 48: Arakawa H, Kurihara Y, Nakajima Y. Pulmonary fat embolism syndrome: CT findings in six patients. J Comput Assist Tomogr 2000; 24: Heyneman LE, Müller NL. Pulmonary nodules in early fat embolism syndrome: a case report. J Thorac Imaging 2000; 15: Hamrick-Turner J, Abbitt PL, Harrison RB, et al. Diffuse lung calcifications following fat emboli and adult respiratory distress syndromes: CT findings. J Thorac Imaging 1994; 9: Shier MR, Wilson RF, James RE, et al. Fat embolism prophylaxis: a study of four treatment modalities. J Trauma 1977; 17: Gurd AR, Wilson RI. The fat embolism syndrome. J Bone Joint Surg Br 1974; 56B: Heitzman RE. Pulmonary fat embolism. In: Heitzman RE, ed. The lung: radiologic-pathologic correlations. St. Louis, MO: Mosby, 1983; Harris AC, Torreggiani WC, Lyburn ID, et al. CT and sonography of traumatic fat embolism in the common femoral vein. AJR Am J Roentgenol 2000; 175: Liu P, Armstrong P, Skippen P. Post-traumatic fat embolism in the inferior vena cava. Can Assoc Radiol J 1990; 41: Robert JH, Hoffmeyer P, Broquet PE, et al. Fat embolism syndrome. Orthop Rev 1993; 22: D Heere MS, Houghton D, Ginzburg E. Fat embolism syndrome. J Trauma Nurs 1999; 6: Feldman F, Ellis K, Green WM. The fat embolism syndrome. Radiology 1975; 114: Berrigan TJJ, Carsky EW, Heitzman ER. Fat embolism: roentgenographic-pathologic correlations in 3 cases. AJR Am J Roentgenol 1966; 96: Maruama Y, Little JB. Roentgen manifestations of traumatic pulmonary fat embolism. Radiology 1962; 79: CHEST / 123 / 4/ APRIL,