Intravertebral vacuum phenomenon in osteoporotic compression fracture: report of 67 cases with quantitative evaluation of intravertebral instability

Transcription

1 J Neurosurg (Spine 1) 100:24 31, 2004 Intravertebral vacuum phenomenon in osteoporotic compression fracture: report of 67 cases with quantitative evaluation of intravertebral instability DONG-YUN KIM, M.D., SANG-HO LEE, M.D., PH.D., JEE SOO JANG, M.D., PH.D., SANG KI CHUNG, M.D., AND HO-YEON LEE, M.D., PH.D. Department of Neurosurgery, Wooridul Spine Hospital, Seoul, Korea Object. The objectives of this study were to: 1) describe the incidence and clinical features of intravertebral vacuum phenomenon (IVVP) in a relatively large number of cases; 2) quantitatively evaluate intravertebral instability and determine the factors affecting instability; and 3) evaluate the efficacy of percutaneous vertebroplasty in the treatment of this phenomenon. Methods. A retrospective review was conducted of the records of 67 patients with IVVP among 652 consecutive cases of osteoporotic compression fracture. Comparisons between the IVVP group and a control group, a stable group, and an unstable group were conducted. Percutaneous vertebroplasty was performed in all patients. There were 67 patients (10.3%) in whom there were 70 vacuum phenomena of the intravertebral space. Intravertebral vacuum phenomena occurred predominantly in the thoracolumbar junction (81%) and in patients with a longer duration of symptoms ( months) compared with the control group. Of 59 vertebrae for which flexion extension radiographs were available, 26 vertebrae were categorized as stable and 33 as unstable. Twenty-one vertebrae (64%) had undergone compression fracture in the unstable group compared with nine (35%) compression fractures in the stable group. There were 28 (85%) fractures of the wedged vertebrae in the unstable group compared with 16 (61%) fractures in wedged vertebrae in the stable group. Percutaneous vertebroplasty was performed with successful clinical outcome. Conclusions. Intravertebral vacuum phenomenon is more common than has been previously appreciated. The results of this study indicate that biomechanics, not ischemic or avascular theory, may play an important role in pathogenesis of this phenomenon. Percutaneous vertebroplasty was found to be a minimally invasive and effective procedure for the treatment of IVVP. KEY WORDS intravertebral vacuum phenomenon necrosis vertebral body intravertebral instability T Abbreviations used in this paper: CT = computerized tomography; IVVP = intravertebral vacuum phenomenon; MR = magnetic resonance; PMMA = polymethylmethacrylate; VB = vertebral body. 24 HE IVVP is known as a relatively uncommon sequela of osteoporotic compression fracture. The finding on plain radiography of an intravertebral vacuum cleft in a collapsed vertebra is considered virtually pathognomonic. 23,26,27,31 Signs peculiar to these cases are flexion extension radiography documented abnormal movement at the intravertebral cleft as well as prolonged back pain aggravated by standing or gait, which improves with bed rest. 6,15,27 The macroscopic findings of the cleft and great degree of motion between the fracture ends suggest that IVVP can be a varied type of pseudarthrosis that is different from the well-known pseudarthrosis of long bone. 9,15 Intravertebral instability has been implicated as a cause of back pain during motion in this phenomenon. The authors of several case reports and clinical series of this phenomenon have used various terms such as avascular necrosis of the VB, 6,26,30,31 intravertebral vacuum cleft, 13,14,16,22,23,26,27 intravertebral pseudarthrosis, 15 vertebral osteonecrosis, 9,23,24 vertebral fluid collection associated with vertebral collapse, 11 delayed posttraumatic vertebral collapse, 18 and Kümmell disease. 9,10 It is not known, however, what percentage of osteoporotic compression fractures is associated with IVVP. In addition, intravertebral instability of IVVP is not well defined. An effective treatment for this phenomenon has not been established. The objectives of our study were to 1) describe the incidence and clinical features of IVVP in a relatively large number of cases; 2) quantitatively evaluate the intravertebral instability and determine the factors affecting instability; and 3) evaluate the efficacy of percutaneous vertebroplasty in the treatment of this phenomenon. Clinical Material and Methods Patient Population We retrospectively reviewed the records of 652 patients with osteoporotic compression fracture who were hospitalized at the Wooridul Spine Hospital between March

2 Intravertebral vacuum phenomenon 2001 and August Intravertebral vaccum phenomenon was defined as the presence of a gas collection within a VB on plain radiography or CT scanning. In several cases with an intravertebral cleft in which the presence of a gas collection was doubtful, those in which the localization of the gas within the VB or with the intervertebral disc could not be established were excluded. The study included 67 patients in whom there were 70 vacuum phenomena of the intervertebral space. Data in this group were compared with those acquired in a cohort of remaining 585 patients with 882 osteoporotic compression fractures without an IVVP (control group). Radiographic and Neuroimaging Assessment In addition to spinal radiography, CT scanning was performed in 57 patients and MR imaging in all patients. An independent observer reviewed each VB on sagittal T 1 - and T 2 -weighted MR images and plain radiographs with regard to the pattern of deformity. Seventy vertebrae were classified into three major categories of deformity: wedged, concave, and flat vertebrae 12 (Fig. 1). Wedged vertebrae were subdivided into those that were superiorly and those that were inferiorly wedged. Each VB was also categorized using the Denis classification system 8 into compression or burst fracture. The burst fracture was characterized on sagittal MR images by the following features: fracture of the posterior wall cortex, loss of posterior height of the VB, and tilting and retropulsion of a fragment of bone into the spinal canal. The presence of a gas collection was assessed using lateral flexion extension radiography. The anterior height of the wedged or flat vertebra and middle height of the concave vertebra was measured on the lateral radiograph and flexion extension radiographs if available. The kyphotic angle was measured using the Cobb method (the upper endplate of the vertebra above and the inferior endplate of the vertebra below the fractured vertebra). The vertebral height was defined as the distance between endplates on the lateral radiograph. The vertebral height of the next normal-appearing adjacent VB was also measured to provide an estimate of the prefracture height. Calculations were made as follows: height lost = [height of the reference vertebra height of the fractured vertebra]; degree of collapse (%) = [height lost/height of the reference vertebra] 100. To correct the magnification of the radiograph, the instability index (in millimeters) of the VB was calculated as the difference of the ratio of the VB height to the reference vertebra height between the flexion extension radiographs multiplied by the height of the reference vertebra. Intravertebral instability was defined as an instability index of greater than 3 mm. A fracture was defined as acute when bone marrow edema was apparent on MR imaging examinations. 33 The height of the VB after vertebroplasty was measured on the postoperative radiographs. FIG. 1. Plain radiographs showing the classification of the deformity. Upper Left: Superiorly wedged vertebra. Upper Right: Inferiorly wedged vertebra. Lower Left: Concave vertebra. Lower Right: Flat vertebra. Vertebroplasty and Outcome Assessment Percutaneous vertebroplasty was performed in all cases, except one in which kyphoplasty was conducted. Percutaneous vertebroplasty was performed via the fluoroscopically guided uni- or bilateral transpedicular approach recommended by Jensen, et al. 19 The procedure was performed after injection of a local anesthetic and induction of conscious sedation in extended posture. The PMMA (Simplex P; Stryker Howmedica, Kalamazoo, MI) with barium sulfate powder for opacification was injected manually, using fluoroscopic guidance, into the lateral projection with an 11-gauge needle. Kyphoplasty was performed using the bilateral tranpedicular approach recommended by Lieberman, et al. 25 Initial pain relief was assessed within the first 48 hours. Preoperative and postoperative pain was assessed using a 10-point visual analog scale. Based on clinical notes and telephone interviews, the observer determined the outcome category into which each case was classified using the following criteria: full recovery to premorbid status and no need for analgesic medication, excellent ; residual back pain not severe enough to impede a premorbid lifestyle, with or without occasional use of nonnarcotic analgesic medication, good ; unimproved pain or functional impairment and the need for analgesic medication, fair ; and increased pain or neurological deficit, poor. Statistical Analysis Statistical analyses were conducted using SPSS for Windows (version 10.0; SPSS, Inc., Chicago, IL). De- 25

3 D. Y. Kim, et al. pending on the variables, intergroup differences were analyzed using the Student t-test, Mann Whitney U-test, and chi-square test. Otherwise, Pearson correlation was applied. The result was considered statistically significant if the probability value was less than Results Clinical Features of IVVP and Control Groups Table 1 provides a brief summary of the treatment group and control groups. The treatment group included 67 patients (10.3% of the total): 61 women and six men, who ranged from 55 to 85 years of age (mean years). Three patients suffered two simultaneous IVVP, accounting for a total of 70 affected vertebrae. In all patients with IVVP there were signs of underlying diffuse osteopenia characterized by concomitant or subsequent compression fractures (35 patients) and low bone mineral density at dual photon absometry (all patients). The demineralization was related to senile or postmenopausal osteoporosis (64 cases), corticosteroid treatment (two cases), and alcohol abuse (one case). None had other risk factors associated with IVVP such as malignancy, infection, history of radiotherapy, vasculitides, or pancreatitis. All patients suffered severe motion pain in the back, notwithstanding conservative treatment with medication and corset therapy. Two patients experienced subtle paraparesis but could walk without assistance. In 41 patients pain was present without an association with trauma, and in 26 pain was due to minor trauma (falling on the buttocks or lifting of heavy objects). The mean duration of symptoms was months. Bone marrow edema was apparent on MR imaging examinations in 27 vertebrae. The IVVP was commonly located at the thoracolumbar junction: eight (11%) at T-11, 22 (31%) at T-12, 27 (39%) at L-1, and the remaining 13 lesions (19%) were between T-8 and L-5 (Fig. 2). Vertebral deformities comprised 55 wedged vertebrae (46 superiorly and nine inferiorly wedged), 14 concave vertebrae, and one flat vertebra. Thirty-six vertebrae (51%) were categorized as compression fractures and 34 (49%) as burst fractures. Intravertebral instability was detected in 33 vertebrae (47%). The control group consisted of 585 patients with 882 osteoporotic compression fractures in which IVVP was absent: 515 women and 70 men in whom the mean age was years. The mean symptom duration was months. Osteoporotic compression fractures without IVVP were also commonly located at the thoracolumbar junction: 63 (7%) at T-11, 154 (18%) at T-12, 193 (22%) at L-1, and the remaining 408 lesions (53%) were between T-5 and L-5 (Fig. 2). Vertebral deformities comprised 627 wedged vertebrae (71%), 213 concave vertebrae (24%), and 42 flat vertebrae (5%). Four hundred eighty vertebrae (54%) were categorized as compression fractures and 402 (46%) as burst fractures. Intravertebral instability was noted in 24 vertebrae (3%). The statistical difference was demonstrated in the duration of symptoms, location of VB collapse, and intravertebral instability between the IVVP group and control group (Student t-test and chi-square test, p ). Radiographic Assessment TABLE 1 Summary of clinical features in patients with IVVP and in controls* Patient Group (%) Variable IVVP Control p Value no. of patients no. of collapses male/female ratio 6:61 70:515 NS age SD (yrs) NS spinal region thoracic 2 (3) 101 (11) thoracolumbar junction 57 (81) 410 (47) lumbar 11 (16) 371 (42) type of fracture NS compression 36 (51) 480 (54) burst 34 (49) 402 (46) category of deformity NS wedged 55 (79) 627 (71) concave 14 (20) 213 (24) flat 1 (1) 42 (5) intravertebral instability 33 (47) 24 (3) symptom duration (mos) * NS = not significant; SD = standard deviation. Significance established using chi-square test. Significance established using Student t-test. Of the 11 vertebrae for which flexion extension radiographs were not available, nine cases of IVVP were demonstrated on the neutral radiograph. In the other two patients, CT scanning was the only modality to demonstrate a vacuum cleft in the VB. Of 59 vertebrae in which the flexion extension radiographs were available, 18 (31%) vacuum phenomena were demonstrated on the neutral lateral radiograph and 41 (69%) only on the extension lateral radiograph. Radiologically based measurement of VB height demonstrated that the mean height lost was 15.1 mm (range mm). The mean degree of collapse was 53% (range 4 90%). The mean instability index was 3.2 mm (range mm). Of 59 vertebrae for which the flexion extension radiographs were available, the mean VB heights on flexion extension radiographs were and mm, respectively. Table 2 provides a summary of data obtained in those cases classified as stable and unstable. Twenty-six vertebrae were categorized as stable and 33 as unstable based on the instability index. The mean instability index was mm in the stable group and mm in the unstable group. The incidence of compression fracture was more common in the unstable group. Twenty-one lesions (64%) were compression fractures and 12 (36%) were burst fractures in the unstable group compared with nine compression fractures (35%) and 17 burst fractures (65%) in the stable group (chi-square test, p 0.05). Wedged vertebrae tended to be more common in the unstable group, but this difference did not reach statistical significance. Twenty-eight fractures (85%) were wedged vertebrae and five (15%) were concave vertebrae in the unstable group compared with 16 wedged vertebrae (61%) and nine concave vertebrae (35%) in the stable group (chi-square test, p = 0.098). Several variables (duration of symptoms, instability index, degree of VB collapse, presence of bone marrow edema, location of frac- 26

4 Intravertebral vacuum phenomenon FIG. 2. Bar graph showing the distribution of the locations of vertebral collapses in the IVVP and control groups. ture, and kyphotic angle) were not shown to be associated with any intergroup differences. The mean postvertebroplasty VB height was mm as demonstrated on postoperative radiography. There was no significant difference in the VB height as measured on preoperative extension radiographs and postvertebroplasty radiographs (paired t-test, p 0.05). Outcome Assessment Cement extravasation occurred at 10 vertebral levels (14.3%). The PMMA entered the epidural space in one patient, the disc space in seven patients, and the paraspinal tissues in two cases. No problems were identified clinically as a result of these extravasations. Almost all patients rated their preoperative back pain intensity as a score of 8 to 10 on the visual analog scale. Marked to complete pain relief, defined as a postoperative pain level score of 0 to 3, was achieved in 42 patients (63%). Moderate pain relief, corresponding to a postoperative pain score of 4 to 6, occurred in 22 patients (33%). Three patients (4%) experienced no significant pain relief. The mean follow-up period was 16.4 months (range 7 23 months). At the time of this review, one patient had died because of an unrelated cause (ischemic stroke) and two could not be located despite attempts to contact them. Of the remaining 64 patients, outcome in nine (14%) was classified as excellent and in 43 (67%) as good, whereas in 12 patients (19%) outcome was considered fair or poor. Illustrative Case Examination. This 72-year-old man presented with an 8-month history of progressive back pain. The pain began suddenly while lifting a heavy object and was aggravated while standing and walking, whereas it was improved with bed rest. No neurological symptom was observed. On physical examination, the patient exhibited severe tenderness over the upper lumbar vertebrae. On radiographic examination, we observed L-1 and L-2 vertebral collapse accompanied by intravertebral vacuum cleft. The intravertebral cleft became prominent when the lumbar spine was in extension but was absent when in flexion (Fig. 3 upper left and center). Computerized tomography scanning clearly demonstrated the presence of gas density within the collapsed vertebrae (Fig. 3 upper right). TABLE 2 Summary of data obtained in the stable and unstable groups Patient Group (%) Stable Unstable p Value no. of patients no. of collapses male/female ratio 2:24 3:28 NS mean age SD (yrs) NS instability index <0.0001* spinal region NS thoracic 1 (3) thoracolumbar junction 22 (85) 26 (79) lumbar 4 (15) 6 (18) type of fracture <0.05 compression 9 (35) 21 (64) burst 17 (65) 12 (36) category of deformity <0.1 wedged 16 (61) 28 (85) concave 9 (35) 5 (15) flat 1 (4) symptom duration (mos) NS bone marrow edema 11 (42) 12 (36) NS collapse (%) NS kyphotic angle NS * Established using the Mann Whitney U-test. Established using the chi-square test. 27

5 D. Y. Kim, et al. FIG. 3. Neuroimages obtained in an illustrative case. Upper Left: Lateral flexion radiograph demonstrating compressed VBs without evident vacuum cleft in L-1 and L-2. Upper Center: Lateral extension radiograph showing the opening of the collapsed vertebrae with the appearance of IVVP with a gaslike area of radiolucency. Upper Right: Axial CT scan clearly demonstrating a gaseous collection in the VB. Lower Left: Sagittal T 1 -weighted MR image demonstrating the markedly low signal intensity horizontal band within the cleft surrounded by areas of low signal intensity. Lower Right: Sagittal T 2 -weighted MR image revealing low signal intensity in L-1 and high signal intensity in L-2. Magnetic resonance imaging revealed abnormally low signal intensity on T 1 -weighted sequences and heterogeneous high signal intensity of the cleft with surrounding band of diminished signal intensity on the T 2 -weighted sequences (Fig. 3 lower left and right). Operation. Percutaneous vertebroplasty was performed via unilateral transpedicular approach after injection of a local anesthetic. Postoperative Course. The patient reported marked reduction in pain after the procedure. Follow-up imaging demonstrated that the PMMA was confined to the vacuum cleft, and flexion extension radiography demonstrated no intravertebral instability (Fig. 4). Discussion General Consideration of IVVP Osteoporosis is a common and socioeconomically important health problem and a major cause of spinal fracture in the elderly. 4,7,21,32 Clinical presentation of osteoporotic compression fracture is quite distinct from that of nonosteoporotic fracture in younger patients. The fracture occurs either spontaneously or after minor trauma, and the initial presentation involves a complaint of back pain only. Most osteoporotic compression fractures heal successfully within a few weeks or months, and pain resolution can be expected with conservative therapy. A minority of patients, however, subsequently experience progressive insidious VB collapse followed by prolonged pain and neurological dysfunction. Analysis of the literature suggests that progression of vertebral collapse and angular deformity results from these fractures only in rare cases. 7,9,21 Reports of such cases, however, are appearing with much greater frequency in recent years, and it is likely that this phenomenon is more common than previously appreciated. 2,17,21,28 The cause of continued VB collapse is not fully understood. One proposed hypothesis implicates IVVP. 3,18,20,24,29 Pathogenesis of IVVP Although several theories have been proposed to ex- 28

6 Intravertebral vacuum phenomenon FIG. 4. Neuroimages obtained in an illustrative case. Left: Postvertebroplasty CT scan revealing bone PMMA confined to vacuum cavity with a sclerotic rim. Center and Right: Postvertebroplasty lateral flexion (center) and extension (right) radiographs demonstrating the vacuum cleft filled with bone PMMA, without evidence of intravertebral instability. plain the cause of the IVVP, its pathogenesis remains unclear. Some investigators have postulated the ischemic theory. Maldague, et al., 26 first described IVVP in 10 patients. For those authors, IVVP was indicative of the absence of bleeding in the vertebral fracture (that is, ischemia), and they termed it ischemic vertebral collapse. This hypothesis has been supported by the following indirect findings. 1) Histopathological examination showed signs of bone necrosis in some cases in which biopsy samples were obtained. 6,15,16,24,26 2) A history of corticosteroid treatment or alcohol abuse, well-known risk factors for avascular necrosis of the femoral head, was found in patients with IVVP. 14,23,24,26,27,29,31 3) Some relationship between the crescent sign of femoral head avascular necrosis and the IVVP 26 has been pointed out. In fact, because all these arguments are debated, the suggestion that avascular necrosis is the main pathological event remains hypothetical. Although histopathological signs of osteonecrosis were present in several reports of IVVP in which biopsy samples were evaluated, recently published pathological evidence of osteoporotic compression fracture indicates that avascular necrosis of the VB is a common histological finding in uncomplicated osteoporotic compression fracture and even in individuals with unfractured osteoporotic VB. 1 In addition, association of avascular necrosis of the femur head, systemic corticosteroid treatment, and alcohol abuse is rare, with only a few reported cases. 6,11,15,16,23 For these reasons, ischemic necrosis of bone, even if extant, is presumably an accessory nonspecific process in IVVP. Some authors have postulated that the IVVP represents an ununited vertebral fracture with possible formation of a pseudarthrosis. 3,15,18,22,29 Baba, et al., 3 reported cases of osteoporosis-related delayed vertebral collapse occurring after a minor injury in which segmental instability was accompanied by an absence of fracture healing. In our study, 81% of 70 vacuum phenomena were located in thoracolumbar intervertebral spaces. This frequency is significantly high compared with the 47% prevalence of osteoporotic compression fractures in the thoracolumbar junction in the control group. Another important fact is that compression fractures and wedged vertebrae were predisposed to the intravertebral instability, whereas burst fractures and concave vertebrae were not. These two lines of evidence suggest that biomechanical properties may play an important role in the pathogenesis of this phenomenon. Anterior wedging results from compressive loading on the anterior aspect of the VBs, which is more conspicuous at the thoracolumbar junction because of an anatomical kyphotic curve. Anterior wedging increases the kyphotic curvature, which produces more bending of the spine and an additional anterior compressive load, resulting in further microfracture; thus, a vicious circle may be formed. Vertebroplasty for IVVP Several authors reported various treatment options for IVVP including conservative brace therapy, 5,14,20,29 laminectomy with posterior fixation, 29 vertebrectomy and reconstruction with strut graft and instrumentation, 6,16,18,20,24,29 and simple bone graft placed in the cleft without instrumentation. 15 The major surgery-related problem encountered in these cases is the diminished mechanical properties of the osteoporotic bone and their poor general medical condition. Because of significant risk caused by comorbid conditions common in this elderly population, as well as technical difficulty related to adequate hardware fixation within osteoporotic bone, surgical intervention has been limited to cases involving concurrent spinal instability or neurological compromise. Because intravertebral instability after pseudarthrosis can be a cause of back pain during motion, this phenomenon is amenable to percutaneous vertebroplasty. Do, et al., 10 reported that vertebroplasty-related outcomes in patients with IVVP are similar to those in patients with involutional osteoporotic fractures. Their results also showed that vertebroplasty was an effective and well-tolerated procedure indicated for the treatment of IVVP in cases in which neurological compromise is absent. 29

7 D. Y. Kim, et al. It is our observation that the intravertebral cavity in this phenomenon allows for its low-pressure filling with viscous cement. During vertebroplasty in which the PMMA was manually injected, we observed, under fluoroscopic visualization, that the PMMA usually filled the vacuum cleft first with low pressure. With continuous injection, the PMMA filled the cancellous bone of a collapsed VB beyond the confines of the vacuum cleft. In addition, our results in this series indicated that the vertebroplasty in extended posture could restore the vertebral height at least to the level of that demonstrated on preoperative extension lateral radiography without requiring further maneuvering. Clinical Implications The incidence of osteoporotic compression fractures associated with IVVP is unknown. Although VB fracture and collapse are readily demonstrated on plain radiographs, visualization of the IVVP is often more difficult. Our findings suggest that IVVP is more common than has been previously thought. In many of these patients, flexion extension radiographs are not routinely obtained after static radiographs; thus, IVVP may not be commonly discovered. We believe that flexion extension radiographs, if they can be obtained in patients with osteoporotic compression fractures, would increase the discovery of this phenomenon. Although the natural history of the IVVP is not fully understood, several authors have reported its evolution. Ito, et al., 18 characterized this phenomenon as delayed vertebral collapse. They followed their patients longitudinally and postulated that delayed VB collapse began with the appearance of an intravertebral cleft approximately 3 weeks after injury and its gradual extension anteriorly over time. The VB anterior wall was eventually destroyed, leading to collapse of the entire VB. Although disappearance of the vacuum phenomena and complete osseous union were observed in a minority of reported cases in which surgical intervention was not performed, 26,27 we observed a significantly longer duration of symptoms in the IVVP group compared with the control group. Percutaneous vertebroplasty has been proven to be an effective treatment for two types of patients with osteoporotic compression fracture: those with chronic pain refractory to medical and brace therapies and those with severe disabling pain caused by more acute fractures. Treatment of ambulatory patients with acute fractures remains controversial; 4,25,32 however, some authors believe that not only chronic compression fractures but also acute or subacute fractures associated with IVVP should be included in the indications for percutaneous vertebroplasty because of its evolutional and chronic nature. Conclusions In this study it was shown that IVVP is more common than has been previously appreciated. We believe that flexion extension radiography performed, if possible, in cases of osteoporotic compression fracture would increase the discovery of IVVP. Although there is a wide range of intravertebral instability in this phenomenon, compression fracture and wedged vertebrae were more common in our group of patients with unstable vertebra. Intravertebral vacuum phenomon had predilection for the thoracolumbar spine. These findings suggest that biomechanics, not ischemic or avascular theory, may play an important role in pathogenesis of this phenomenon. Percutaneous vertebroplasty was found to be a minimally invasive and effective procedure providing pain relief in patients with osteoporotic compression fracture and IVVP. Acknowledgments We thank Yoon Nam Kim and Hyun-Ju Lee for her help with data collection and In Sook Jo for preparation of this manuscript. References 1. Antonacci MD, Mody DR, Rutz K, et al: A histologic study of fractured human vertebral bodies. J Spinal Disord Tech 15: , Arciero RA, Leung KY, Pierce JH: Spontaneous unstable burst fracture of the thoracolumbar spine in osteoporosis. A report of two cases. Spine 14: , Baba H, Maezawa Y, Kamitani K, et al: Osteoporotic vertebral collapse with late neurological complications. Paraplegia 33: , Barr JD, Barr MS, Lemley TJ, et al: Percutaneous vertebroplasty for pain relief and spinal stabilization. Spine 25: , Chevalier X, Wrona N, Avouac B, et al: Thigh pain and multiple vertebral osteonecroses: value of magnetic resonance imaging. J Rheumatol 18: , Chou LH, Knight RQ: Idiopathic avascular necrosis of a vertebral body. Case report and literature review. Spine 22: , Cooper C, Atkinson EJ, O Fallon WM, et al: Incidence of clinically diagnosed vertebral fractures: a population-based study in Rochester, Minnesota, J Bone Miner Res 7: , Denis F: The three column spine and its significance in the classification of acute thoracolumbar spinal injuries. Spine 8: , Do HM: Magnetic resonance imaging in the evaluation of patients for percutaneous vertebroplasty. Top Magn Reson Imaging 11: , Do HM, Jensen ME, Marx WF, et al: Percutaneous vertebroplasty in the treatment of patients with vertebral osteonecrosis (Kummell s disease). Neurosurg Focus 7:Article 2, Dupuy DE, Palmer WE, Rosenthal DI: Vertebral fluid collection associated with vertebral collapse. AJR 167: , Eastell R, Cedel SL, Wahner HW, et al: Classification of vertebral fractures. J Bone Miner Res 6: , Golimbu C, Firooznia H, Rafii M: The intravertebral vacuum sign. Spine 11: , Harverson G: Intravertebral vacuum phenomenon. Clin Radiol 39:69 72, Hasegawa K, Homma T, Uchiyama S, et al: Vertebral pseudarthrosis in the osteoporotic spine. Spine 23: , Hashimoto K, Yasui N, Yamagishi M, et al: Intravertebral vacuum cleft in the fifth lumbar vertebra. Spine 14: , Heggeness MH: Spine fracture with neurological deficit in osteoporosis. Osteoporos Int 3: , Ito Y, Hasegawa Y, Toda K, et al: Pathogenesis and diagnosis of delayed vertebral collapse resulting from osteoporotic spinal fracture. Spine J 2: , Jensen ME, Evans AJ, Mathis JM, et al: Percutaneous polymethylmethacrylate vertebroplasty in the treatment of osteoporotic vertebral body compression fractures: technical aspects. AJNR Am J Neuroradiol 18: ,

8 Intravertebral vacuum phenomenon 20. Kaplan PA, Orton DF, Asleson RJ: Osteoporosis with vertebral compression fractures, retropulsed fragments, and neurologic compromise. Radiology 165: , Kostuik JP, Heggeness MH: Surgery of the osteoporotic spine, in Frymoyer JW, Ducker TB, Hadler N, et al (eds): The Adult Spine: Principles and Practice. Philadelphia: Lippincott-Raven, Vol 2, 1997, pp Kumpan W, Salomonowitz E, Seidl G, et al: The intravertebral vacuum phenomenon. Skeletal Radiol 15: , Lafforgue P, Chagnaud C, Daumen-Legre V, et al: The intravertebral vacuum phenomenon ( vertebral osteonecrosis ). Migration of intradiscal gas in a fractured vertebral body? Spine 22: , Laloux P, Lefebvre S, Esselinckx W, et al: Spinal cord compression secondary to vertebral aseptic osteonecrosis. Spine 16: , Lieberman IH, Dudeney S, Reinhardt MK, et al: Initial outcome and efficacy of kyphoplasty in the treatment of painful osteoporotic vertebral compression fractures. Spine 26: , Maldague BE, Noel HM, Malghem JJ: The intravertebral vacuum cleft: a sign of ischemic vertebral collapse. Radiology 129: 23 29, Malghem J, Maldague B, Labaisse MA, et al: Intravertebral vacuum cleft: changes in content after supine positioning. Radiology 187: , Maruo S, Tatekawa F, Nakano K: [Paraplegia caused by vertebral compression fractures in senile osteoporosis.] Z Orthop Ihre Grenzgeb 125: , 1987 (Ger) 29. Mochida J, Toh E, Chiba M, et al: Treatment of osteoporotic late collapse of a vertebral body of thoracic and lumbar spine. J Spinal Disord 14: , Mok MY, Isenberg DA: Avascular necrosis of a single vertebral body, an atypical site of disease in a secondary APLS. Ann Rheum Dis 59: , Naul LG, Peet GJ, Maupin WB: Avascular necrosis of the vertebral body: MR imaging. Radiology 172: , Truumees E: The role of vertebroplasty and kyphoplasty as parts of a treatment strategy for osteoporotic vertebral compression fractures, in Andersson GBJ (ed): Contemporary Spine Surgery. Baltimore: Lippincott Williams & Wilkins, Vol 3, 2002, pp Yamato M, Nishimura G, Kuramochi E, et al: MR appearance at different ages of osteoporotic compression fractures of the vertebrae. Radiat Med 16: , 1998 Manuscript received February 11, Accepted in final form September 18, Address reprint requests to: Dong-Yun Kim, M.D., Department of Neurosurgery, Wooridul Spine Hospital, 47-4 Chungdam-Dong Kangnam-Gu, Seoul , Korea. nstopaz@yahoo.co.kr. 31