Percutaneous Vertebroplasty: Treatment of Painful Vertebral Compression Fractures with Intraosseous Vacuum Phenomena

Transcription

1 Wilfred C. G. Peh 1,2 Michael S. Gelbart 2 Louis. Gilula 2 Dallas D. Peck 2,3 Received January 17, 2002; accepted after revision October 22, Department of Diagnostic Radiology, Singapore General Hospital, Outram Rd., Singapore ddress correspondence to L.. Gilula. 2 Mallinckrodt Institute of Radiology, Washington University Medical Center, St. Louis, MO Present address: Gem State Radiology, 877 W. Main St., Ste. 603, oise, ID JR 2003;180: X/03/ merican Roentgen Ray Society Percutaneous Vertebroplasty: Treatment of Painful Vertebral Compression Fractures with Intraosseous Vacuum Phenomena OJECTIVE. This study was undertaken to determine the efficacy of percutaneous vertebroplasty in treating patients with painful compression fractures of the thoracic and lumbar vertebrae that contain intraosseous vacuum phenomena. MTERILS ND METHODS. Nineteen cases of painful vertebral compression fractures with intraosseous phenomena occurring in 18 patients (six men, 12 women; age range, years; mean age, 75.5 years) were identified from 393 percutaneous vertebroplasties performed in 199 patients during 32 and a half months. ll patients had osteoporosis, with severe vertebral compression to less than one third of the vertebral body height in 13 of 19 cases. ffected levels were T6 (n = 1), T8 (n = 2), T9 (n = 1), T11 (n = 1), T12 (n = 4), L1 (n = 5), L2 (n = 2), L3 (n = 1), L4 (n = 1), and L5 (n = 1). ll cases had the typical intravertebral body vacuum cleft appearance on radiographs. Imaging and clinical features were analyzed. RESULTS. The mean volume of polymethyl methacrylate injected was 7.43 ml (range, ml). Typically, the polymethyl methacrylate filled the intravertebral vacuum cleft. Complications during radiography consisted of minimal polymethyl methacrylate leakage into the adjacent disk (15/19 cases) and the paravertebral soft tissues (8/19 cases). No complications required surgical intervention. t clinical follow-up, pain relief was complete in eight patients (44.4%), partial in six patients (33.3%), and unchanged in four patients (22.2%). CONCLUSION. Percutaneous vertebroplasty is effective in the treatment of patients with painful vertebral compression fractures with intraosseous vacuum phenomena. I n 1987, percutaneous vertebroplasty was described for treatment of the aggressive type of vertebral hemangioma [1]. Currently, the indications for vertebroplasty encompass compression fractures due to osteoporosis, vertebral myeloma, and metastases [2 13]. Percutaneous vertebroplasty provides pain relief and bone strengthening of collapsed vertebrae, and it is increasingly being accepted as a major treatment option in management of the resultant intractable midline back pain caused by collapsed vertebrae. Kümmell s disease describes vertebral collapse that occurs some time after an injury [14]. This collapse may potentially be a consequence of a vascular insult that leads to secondary bone necrosis [15 17]. It has been reported that finding a vacuum phenomenon or cleft in a vertebra is diagnostic of Kümmell s disease [18 25]. The aim of our study was to determine the efficacy of percutaneous vertebroplasty in treating patients who have painful vertebral compression fractures containing intraosseous vacuum phenomena. Materials and Methods Three hundred ninety-three percutaneous vertebroplasties were performed in 199 consecutive patients over 32 and a half months (June 9, 1998, to February 20, 2001) at the Mallinckrodt Institute of Radiology. Of these, 18 patients (six men, 12 women; mean age, 75.5 years) had 19 vertebroplasties performed for painful compression fractures with intraosseous vacuum phenomena. The indication for vertebroplasty was focal, intense, and intractable spinal pain related to that compression fracture without evidence of definite radicular signs and symptoms. Routine exclusion criteria were bleeding disorders, unstable fractures due to posterior element involvement, and no evidence of vertebral compression on radiographs. Relative contraindications included the inability of the patient to lie prone for the 1 2 hr required for the procedure, lack of surgical backup or patient monitoring facilities, and neurologic symptoms or signs resulting from neurologic compression by 1411

2 Peh et al. vertebral body collapse. Severe vertebral body compression fractures (or vertebra plana) were not regarded as a contraindication to vertebroplasty [13]. Percutaneous vertebroplasty was performed under stringent sterile conditions and with fluoroscopic guidance using a C-arm angiographic unit (ngioskop 33; Siemens, Erlangen, Germany). ll patients were in a prone position during the procedure, and their vital signs were monitored continually. IV fentanyl citrate (Sublimaze; bbott Laboratories, North Chicago, IL) and midazolam hydrochloride (Versed; Roche Pharmaceuticals, Manati, Puerto Rico) were administered for analgesia and sedation, respectively. fter a small skin incision was made, an 11- or 13-gauge bone biopsy trochar needle was advanced through the pedicle into the vertebral body. Ideally, the needle tip was placed in the anterior one third to one fourth of the vertebral body close to the midline. In our study, the major exception to this needle-tip placement was the situation in which the center of the vertebral body was prominently more depressed than the sides of the vertebral body. In that situation, the needle was placed more laterally to decrease the incidence of polymethyl methacrylate leakage into the disk. When the needle tip was optimally positioned, the stylet was removed from the trochar and intraosseous epidural venography using ml of iohexol (Omnipaque 180; Nycomed, Princeton, NJ) was performed to determine whether the needle was positioned within a direct venous anastomosis. Intraosseous venography also enabled observation of the venous structures that filled first and indicated the most likely place to observe first for potential polymethyl methacrylate venous filling. Contrast material injection was stopped immediately after filling the fracture cleft so that the cleft would not be fully opacified. Methyl methacrylate powder (Osteobond copolymer bone cement; Zimmer, Warsaw, IN) was mixed with sterilized barium sulfate powder (E-Z-EM; Westbury, NY) and 1.2 g of tobramycin (Nebcin; Eli Lilly, Indianapolis, IN). Later in our experience, we discontinued the use of tobramycin because it was unavailable; we then began to use IV cefazolin (ncef; eecham, Philadelphia, P) unless a patient was allergic to penicillin. Liquid methyl methacrylate monomer was then added to the powder mixture and mixed into a toothpastelike consistency. The polymethyl methacrylate mixture was placed in the back of a 20-mL syringe and then backfilled into a screw-type 10-mL syringe (LeVeen; oston Scientific, Glen Falls, NY). Under lateral fluoroscopic control, the polymethyl methacrylate mixture was injected until it reached the posterior quarter of the vertebral body, until the mixture started to pass into the disk space or paravertebral tissues, or until the fracture cleft was completely filled. If too much resistance for the 10-mL syringe was encountered, special strongwalled 1-mL syringes (Medallion [Merit Medical Systems, South Jordan, UT]) were backfilled quickly and used to fill the vertebral body more completely. If leakage occurred outside the vertebra, the injection was halted for 1 2 min to allow the polymethyl methacrylate to harden, in an attempt to plug the leak, or the needle was repositioned. If the polymethyl methacrylate mixture did not pass the midline to the opposite side of the vertebral body, the opposite pedicle was then punctured. In some cases, only the vacuum cleft was filled with polymethyl methacrylate and a large amount of bone without polymethyl methacrylate was present below the fracture cleft. When this occurred, a needle (typically directed through the opposite pedicle) was inserted in the vertebral body inferior to the vacuum cleft to try to get polymethyl methacrylate to fill the remaining part of the vertebral body. Two musculoskeletal radiologists analyzed the imaging and clinical features in consensus. Particular attention was paid to the presence of vacuum phenomenon, the volume of polymethyl methacrylate injected, the location of involved vertebra, radiographic complications of polymethyl methacrylate leakage, and clinical outcome. The MR images of all patients were reviewed for the presence of a cleftlike lesion in the compressed vertebral body. Radiographs obtained before the vertebroplasty were also reviewed to determine whether vacuum phenomenon was present in the disks adjacent to the compressed vertebral body. Radiographs obtained after the procedure were reviewed to specifically check whether an increase occurred in the vertebral body height after the polymethyl methacrylate injection. Patients were evaluated for their pain score before vertebroplasty using a visual analog (0 10) scale. fter vertebroplasty, assessment was performed at regular intervals: immediately after vertebroplasty and at 24 hr, 2 weeks, 1 month, 3 months, 1 year, and 2 years. Institutional review board approval was obtained to allow a person trained in patient research to perform telephone follow-up with each patient after vertebroplasty. During this interview, a questionnaire was completed to determine whether the patient s pain was completely relieved, partially relieved, unchanged, or worse. Results ll cases had the typical intravertebral body vacuum cleft phenomenon appearance on radiographs (Fig. 1). ll patients had osteoporosis, and one patient had coexisting breast carcinoma. The painful collapsed vertebral body in this patient was thought to be due to osteoporosis or osteonecrosis because no imaging features of malignancy were found in the affected vertebra. Involved vertebrae were located from the T6 to L5 levels. ffected levels were T6 (n = 1), T8 (n = 2), T9 (n = 1), T11 (n = 1), T12 (n = 4), L1 (n = 5), L2 (n = 2), L3 (n = 1), L4 (n = 1), and L5 (n = 1). Severe vertebral collapse to less than one third of the vertebral body height occurred in 13 of 19 cases. Painful vertebral compression fractures with intraosseous vacuum phenomena requiring vertebroplasty were present at a single level in 17 of 19 cases; one patient had two levels injected. In this study, adequate MR images were available for evaluation of 14 of the 19 affected vertebrae. On review of MR images, a linear band that was hypointense on both T1- and T2-weighted images was detected in the collapsed vertebral body in all 14 cases. In three of the 14 cases with available MR images, a component of a linear band was present that was hypointense on T1- and hyperintense on T2-weighted images (Fig. 2). This intravertebral-body linear band that was present on MR images, whether hypointense or hyperintense, corresponded to the intravertebral-body vacuum cleft seen on radiographs (Figs. 2 and 3). In 18 of the 19 cases, the height of the vertebra after placement of polymethyl methacrylate was seen to be increased on radiographs obtained for verification after vertebroplasty (Figs. 1, 2, and 4). We found that, occasionally, placing body supports of various heights under the chest and pelvis to hyperextend the spine at the level of the involved vertebral body while these patients were in a prone position helped restore the vertebral body height (Fig. 1). The mean volume of polymethyl methacrylate injected was 7.43 ml (range, ml). The polymethyl methacrylate typically filled the intravertebral vacuum cleft (Figs. 1, 2, and 4). ilateral pedicle injection was performed in two of 19 cases; 17 of 19 cases underwent unilateral pedicle injection. Preferential filling of the vacuum cleft occurred. Radiographic complications included polymethyl methacrylate leakage into the adjacent disk (15/19 cases) and the paravertebral soft tissues (8/19 cases). The mean volume of polymethyl methacrylate injected in cases with disk leakage was 7.2 ml (15/19 cases), whereas the mean volume injected in cases without disk leakage was 8.1 ml (4/19 cases). Preexisting vacuum phenomenon was present in a disk adjacent to the vertebral body with intraosseous vacuum phenomenon in nine of 19 cases. Eight of these nine occurrences were in the disk immediately adjacent to the vertebral body surface with the intraosseous vacuum phenomenon. No radiographic complications were encountered that required follow-up surgery. The mean clinical follow-up period for all patients was 9.9 months. Eleven of the 18 patients survived; seven patients died but were followed up until the time of death. The follow-up period for the 11 surviving patients ranged from 6 to 24 months (mean, 13.6 months). In the patients who died, follow-up ranged from 2 days to 14 months (mean,

3 Percutaneous Vertebroplasty months). Pain relief was complete in eight (44.4%), partial in six (33.3%), and unchanged in four (22.2%) of the 18 patients. The patient who died of cancer-related problems within 2 days of the procedure was counted as having no change in pain for purposes of this study. Discussion Kümmell s disease was originally described in 1895 as a posttraumatic osteitis in which painful kyphosis developed in patients after a symptom-free period lasting from months to years after an injury [14]. The cause of this entity is still poorly understood, but most investigators agree that the vertebral collapse with an intraosseous vacuum phenomenon occurs as a consequence of a vascular insult leading to secondary bone necrosis [15 18]. lthough it is not an absolutely specific finding, the radiographic presence of an intraosseous vacuum phenomenon in the collapsed vertebral body is C considered by some to be diagnostic for Kümmell s disease [18 25]. However, Kümmell [14] did not describe such a cleft or vacuum phenomenon occurring in malignancy. t least two reports of a vacuum phenomenon have been published: one in metastatic disease and one in myeloma of a vertebra [26, 27]. Patients with Kümmell s disease, or, more accurately, vertebrae with vacuum phenomena, are generally middle-aged or elderly. The interval between the acute traumatic episode D Fig year-old woman with compression fracture of T8 vertebral body., Lateral radiograph shows horizontal vacuum cleft (arrows) in inferior aspect of T8 vertebral body. Note T10 compression fracture., xial CT scan obtained through lower T8 vertebra confirms intraosseous vacuum phenomenon. C, Lateral radiograph obtained during vertebroplasty shows needle positioned posteriorly in T8 vertebral body. D, Lateral radiograph after polymethyl methacrylate injection shows filled cleft (arrowheads) and opacification of posterior two thirds of T8 vertebral body. Note increased height of vertebral body compared with. 1413

4 Peh et al. Fig year-old man with fracture of L1 vertebral body., Lateral radiograph shows gas-filled fracture cleft in L1 vertebral body. Note intradiscal gas in T12 L1 disk., Sagittal T1-weighted spin-echo MR image shows low-signal cleft in L1 vertebral body located anteriorly near superior endplate, with low-signal fluid at posterior aspect of cleft (arrowheads). C, Sagittal T2-weighted spin-echo MR image at same level as shows low-signal cleft (arrowhead) in L1 vertebral body near superior endplate anteriorly with high signal fluid (arrow) at posterior aspect of cleft. D, Lateral radiograph shows vertebroplasty needle communicating with vacuum cleft. Small amount of contrast material is seen in layers (arrows) in anterior part of cleft. E, Lateral radiograph shows polymethyl methacrylate in vertebral body with small amount passing into T12 L1 and L1 L2 disks. Vertebra has increased height compared with. C D 1414 E

5 Percutaneous Vertebroplasty and onset of painful symptoms varies from days to years. The lower thoracic and upper lumbar vertebral bodies are principally involved. Single rather than multiple vertebral levels are usually affected [18]. This pattern of development of the intraosseous vacuum phenomenon suggesting Kümmell s disease was found in our patient population. However, we found that the time interval between trauma and development of the intraosseous vacuum phenomenon was difficult to evaluate accurately. This information could probably be determined only in a large-scale prospective series of patients with acute traumatic vertebral compression fractures who were initially radiographed and who were subjected to regular clinical and radiographic follow-up. The cardinal radiographic sign of Kümmell s disease, as described in the literature, is the presence of a bandlike radiolucent area in a collapsed vertebral body. This feature is also known as an intraosseous vacuum cleft [15 25]. In rare circumstances, intraosseous gas may accompany a Schmorl s node, but in these cases, the pattern and distribution of the gas differ from that found in Kümmell s disease [18, 28 30]. In Schmorl s node, a complexly shaped radiolucent area is encountered, with a horizontal component in the disk and a vertical component in the osseous defect [30]. halla and Reinus [24] stated that intraosseous gas had not been described in vertebral osteomyelitis. However, other researchers have reported rare occurrence of intraosseous gas in patients with infection [29]. On MR imaging, the vacuum cleft is seen as a linear area of hypointense signal on T1weighted images. On T2-weighted images, this area may be either hypointense or hyperintense, depending on positioning of the patient s spine and whether gas or fluid occupies the cleft [18, 25, 31 34]. Similar MR imaging observations were made in our series. Percutaneous vertebroplasty can be of great help to patients with acute osteoporotic compression fractures in which pain persists despite correct medical treatment [2, 3, 6 12]. Typically, this procedure is performed in patients who are not candidates for surgery. n additional advantage of percutaneous vertebroplasty is that it can be performed in multiple vertebrae, either at the same time or with subsequent procedures [4 6, 8, 9]. Vertebroplasty can also be performed in patients with severe vertebral body collapse [13]. The use of percutaneous vertebroplasty for the treatment of painful fractured vertebrae with intraosseous vacuum phenomena suggestive of Kümmell s disease has not been previously reported in the peer-reviewed literature listed in Index Medicus. To our knowledge, very little histologic data are available in the cases called Kümmell s disease to support the claim that osteonecrosis is always present when a vacuum phenomenon is radiographically present in a vertebral body. Hasegawa et al. [25], in a histologic study of the vertebrae of five patients with an intravertebral cleft, found the cleft to be lined with smooth fibrocartilaginous tissue, consistent with pseudarthrosis. These researchers did not find necrotic bone in their histologic material, although one patient had what they described as necrotic granulation at the posterior portion of the pseudarthrosis. We found that some vertebrae are very soft on needle placement during vertebroplasty, whereas in other cases, the bone of the vertebral body felt distinctly hard. We believe that a fractured vertebral body with an intraosseous vacuum phenomenon represents a poorly healed vertebral body fracture with impending or established nonunion. The adjacent vertebral body may or may not be ischemic; individual cases probably lie within a spectrum that ranges from normal to osteonecrotic. It is likely that these patients have severe pain due to motion between the fracture fragments on either side of the fracture cleft containing the vacuum phenomena. In some cases, we repositioned the same needle or placed an additional needle so that parts of the vertebral body that did not initially fill with polymethyl methacrylate could be filled. We hoped that this maneuver would prevent the vertebral body from undergoing further collapse. None of this patient cohort had evidence of progression of collapse. Perivertebral venous, paravertebral soft-tissue, and intradiscal leakages are usually of no Fig year-old woman with T12 vertebral body fracture., Sagittal T1-weighted spin-echo MR image shows low-signal cleft (arrow) in vertebral body near superior T12 endplate., Sagittal T2-weighted spin-echo MR image shows low-signal cleft (arrows) in vertebral body near superior T12 endplate. 1415

6 Peh et al. Fig year-old woman with L2 vertebral body compression fracture showing increased height of body after vertebroplasty., Lateral radiograph obtained at start of vertebroplasty shows horizontal vacuum cleft (arrows) in mid and superior parts of L2 vertebral body. Needle tip is placed near posterior aspect of vertebral body., Lateral radiograph obtained after early polymethyl methacrylate injection shows opacification of vertebral body inferior to vacuum cleft. C, Lateral radiograph obtained after additional polymethyl methacrylate injection shows injected material has entered vacuum cleft. Note increase in height of compressed vertebral body and more parallel orientation of endplates compared with. short-term or midterm clinical significance [5, 6, 8 10, 13]. It is also our experience that polymethyl methacrylate leakage into the disk is not rare. Our study appeared to show a relationship between the incidence of polymethyl methacrylate disk leakage and the presence of an intraosseous vacuum cleft leading to the adjacent disk. When leakage into the disk was present, it was almost always at the location of the cleft. Lafforgue et al. [26], in a retrospective study of 310 consecutive patients with at least one vertebral collapse, found that intradiscal vacuum phenomenon adjacent to vertebral collapse was common. Collapse-related vacuum phenomena were seen on radiographs in 15% of their patients and in 21% when all the imaging modalities were considered. In a subsequent article, Lafforgue et al. [35] showed a high association between continuous intervertebral and intravertebral vacuum phenomena. In that study, a significantly higher incidence of intradiscal 1416 vacuum phenomenon was present in cases of intraosseous vacuum phenomenon compared with the control group, and in six cases, communication between the intervertebral and intravertebral gaseous collections occurred through a fractured endplate. Our findings support this postulation. In nine of 19 of our cases, gas was present in the disk. Of these nine cases, eight occurred in proximity to the vacuum cleft; that is, the vacuum disk existed at the superior disk space with an intraosseous vacuum cleft in the superior aspect of the vertebral body (Fig. 2). Therefore, to minimize polymethyl methacrylate leakage into the adjacent disk when performing vertebroplasty in patients with intraosseous vacuum phenomena, close observation of the fluoroscopic image during injection is required. Pain relief is commonly seen after a mean period of 24 hr after the procedure [8]. Complete pain relief or decreased pain has been re- C ported in as many as 95% of patients with osteoporotic compression fractures [7, 10, 12]. Our study population had similar results: Most of these patients (77.7%) obtained complete or partial pain relief from the procedure. It is not clear why the percentage of patients with complete pain relief in our series is not as high as that reported by other researchers. One possible explanation may be that all our follow-ups were performed by a third party who was not at all related to performance of the procedure. We took this approach to ensure that the results were as objective as possible, with the thought that some patients may say they feel better than they actually do to avoid disappointing the treating doctor. Typically, our patients had improved mobility within 24 hr, and most could bear weight shortly after the procedure. Our typical approach was to have the patient stand and walk 1 hr after the end of the procedure. fter vertebroplasty, the amount and

7 Percutaneous Vertebroplasty type of pain medication could be either greatly reduced or stopped in most patients. In some of our patients, particularly those who were bedridden for longer periods of time, it took a few days for pain relief to be recognized because these patients could not tell the difference between the soreness resulting from the procedure and that of their presenting pain. The true relationship between the presence of vacuum phenomena and Kümmell s disease remains unclear. ecause Lafforgue et al. [26] showed intradiscal vacuum phenomenon adjacent to vertebral collapse to be common, the question arises whether the presence of vacuum phenomenon in the adjacent vertebral body always represents osteonecrosis. compressed vertebral body commonly fractures through the endplate; therefore, it is possible that the intraosseous gas came from gas in the disk rather than occurring as an isolated development [35]. Some vertebra in our patients felt hard during needle placement, whereas others were soft. lthough some evidence of necrotic bone was found by Hasegawa et al. [25], we agree with those researchers that the chief emphasis should be that these intraosseous vacuum phenomena represent vertebral body nonunions or pseudarthroses. We also question the necessity for osteonecrosis to be present before the term Kümmell s disease is used because Kümmell [14] did not have histologic proof in any of his cases and his description of Kümmell s disease occurred before the development of radiography. In summary, percutaneous vertebroplasty is a useful technique for management of painful compression fractures with intraosseous vacuum phenomena, all of which probably represent vertebral body pseudarthroses. On the basis of our experience, percutaneous vertebroplasty is a safe procedure that provides pain relief and vertebral stabilization in most patients with painful fractured vertebra with intraosseous vacuum phenomena. References 1. Galibert P, Deramond H, Rosat P, Le Gars D. Preliminary note on the treatment of vertebral angioma by percutaneous acrylic vertebroplasty [in French]. Neurochirurgie 1987;33: Debussche-Depriester C, Deramond H, et al. Percutaneous vertebroplasty with acrylic cement in the treatment of osteoporotic vertebral crush fracture syndrome. Neuroradiology 1991;33[suppl]: Gangi, Kastler, Dietemann JL. Percutaneous vertebroplasty guided by a combination of CT and fluoroscopy. JNR 1994;15: Weill, Chiras J, Simon JM, Rose M, Sola-Martinez T, Enkaoua E. Spinal metastases: indications for and results of percutaneous injection of acrylic surgical cement. Radiology 1996;199: Cotten, Dewatre F, Cortet, et al. Percutaneous vertebroplasty for osteolytic metastases and myeloma: effects of the percentage of lesion filling and the leakage of methyl methacrylate at clinical follow-up. Radiology 1996;200: Deramond H, Depriester C, Toussaint P, Galibert P. Percutaneous vertebroplasty. Semin Musculoskelet Radiol 1997;1: Jensen ME, Evans J, Mathis JM, Kallmes DF, Cloft HJ, Dion JE. Percutaneous polymethylmethacrylate vertebroplasty in the treatment of osteoporotic vertebral body compression fractures: technical aspects. JNR 1997;18: Cotten, outry N, Cortet, et al. Percutaneous vertebroplasty: state of the art. RadioGraphics 1998;18: Deramond H, Depriester C, Galibert P, Le Gars D. Percutaneous vertebroplasty with polymethylmethacrylate: technique, indications, and results. Radiol Clin North m 1998;36: Cyteval C, Sarrabere MP, Roux JO, et al. cute osteoporotic vertebral collapse: open study on percutaneous injection of acrylic surgical cement in 20 patients. JR 1999;173: Cortet, Cotten, outry N, et al. Percutaneous vertebroplasty in the treatment of osteoporotic vertebral compression fractures: an open prospective study. J Rheumatol 1999;26: arr JD, arr MS, Lemley TJ, McCann RM. Percutaneous vertebroplasty for pain relief and spinal stabilization. Spine 2000;25: Peh WCG, Gilula L, Peck DD. Percutaneous vertebroplasty for severe osteoporotic vertebral body compression fractures. Radiology 2002;223: Kümmell H. Die rarefizierende ostitis der wirbelkrper. Deutsche Med 1895;21: rower C, Downey EF Jr. Kümmell disease: report of a case with serial radiographs. Radiology 1981;141: Nicholas JJ, enedek TG, Reece GJ. Delayed traumatic vertebral body compression fracture. I. Clinical features. Semin rthritis Rheum 1981;10: enedek TG, Nicholas JJ. Delayed traumatic vertebral body compression fracture. II. Pathological features. Semin rthritis Rheum 1981;10: Resnick D, Niwayama G. Osteonecrosis: diagnostic techniques, specific situations, and complications. In: Resnick D, ed. Diagnosis of bone and joint disorders, 3rd ed. Philadelphia: Saunders, 1995: Maldague E, Noel HM, Malghem JJ. The intravertebral vacuum cleft: a sign of ischemic vertebral collapse. Radiology 1978;129: Resnick D, Niwayama G, Guerra JJ Jr, Vint V, Usselman J. Spinal vacuum phenomenon: anatomical study and review. Radiology 1981;139: Kumpan W, Salomonowitz E, Seidl G, Wittich GR. The intravertebral vacuum phenomenon. Skeletal Radiol 1986;15: Harverson G. Intravertebral vacuum phenomenon. Clin Radiol 1988;39: Van Eenenaam DP, el-khoury GY. Delayed post-traumatic vertebral collapse (Kümmell s disease): case report with serial radiographs, computed tomographic scans, and bone scans. Spine 1993;18: halla S, Reinus WR. The linear intravertebral vacuum: a sign of benign vertebral collapse. JR 1998;170: Hasegawa K, Homma T, Uchiyama S, Takahashi H. Vertebral pseudoarthrosis in the osteoporotic spine. Spine 1998;23: Lafforgue PF, Chagnaud CJ, Daver LMH, et al. Intravertebral disk vacuum phenomenon secondary to vertebral collapse: prevalence and significance. Radiology 1994;193: Karantanas H. CT and MR imaging of intravertebral vacuum resulting from a malignancy. JR 2001;177: Larsen JL, Smievoll I. Gas within a cervical vertebral body: a case report with CT confirmation. Eur J Radiol 1988;8: ielecki DK, Sartoris D, Resnick D, Van Lom K, Fierer J, Haghighi P. Intraosseous and intradiscal gas in association with spinal infection: report of three cases. JR 1986;147: Sartoris DJ. Significance of the vertebral vacuum phenomenon. (answer to question) JR 1994;163: Naul LG, Peet GJ, Maupin W. vascular necrosis of the vertebral body: MR imaging. Radiology 1989;172: Chevalier X, Wrona N, vouac, Larget-Piet. Thigh pain and multiple vertebral osteonecroses: value of magnetic resonance imaging. J Rheumatol 1991;18: Malghem J, Maldague, Labaisse M, et al. Intravertebral vacuum cleft: changes in content after supine positioning. Radiology 1993;187: Dupuy DE, Palmer WE, Rosenthal DI. Vertebral fluid collection associated with vertebral collapse. JR 1996;167: Lafforgue PF, Chagnaud CJ, Daumen-Legre VMS, et al. Intravertebral vacuum phenomenon ( vertebral osteonecrosis ): migration of intradiscal gas in a fractured vertebral body. Spine 1997;22: