Intravertebral Clefts in Osteoporotic Vertebral Compression Fractures

ARTHRITIS & RHEUMATISM Vol. 48, No. 5, May 2003, pp 1414 1419 DOI 10.1002/art.10984 2003, American College of Rheumatology Intravertebral Clefts in Osteoporotic Vertebral Compression Fractures Fergus McKiernan and Tom Faciszewski Objective. To describe the characteristics, radiographic appearance, and significance of intravertebral clefts in vertebral compression fractures (VCFs) of patients with osteoporosis presenting for vertebroplasty. Methods. This was a prospective radiographic study of 50 consecutive patients with 82 VCFs who underwent vertebroplasty at a tertiary referral center. Patients underwent imaging preoperatively with standing lateral and supine cross-table lateral radiographs and magnetic resonance imaging (MRI). Standing lateral radiographs were also obtained postoperatively. Clefts were defined at the time of vertebroplasty as confluent reservoirs for polymethylmethacrylate (PMMA). Postoperatively, all images were reexamined for the presence and characterization of intravertebral clefts. Results. Twenty-four of 50 patients (48%) had clefted VCFs, and 30 of 82 VCFs (37%) contained clefts. Clefted VCFs were severe, dynamically mobile, and occurred primarily in the thoracolumbar junction. Clefts were detectable by standing lateral radiography in 14%, by supine cross-table radiography in 64%, and by MRI in 96% of fractured vertebrae. Clefts occurred primarily in the anterosuperior portion of the vertebral body, and cleft margins appeared increasingly sclerotic in persistently mobile VCFs. PMMA fill patterns of clefted and unclefted VCFs were distinct. Conclusion. Intravertebral clefts occur frequently in osteoporotic VCFs of patients who present for vertebroplasty. The radiographic characteristics of clefts evolve over time and can be indistinguishable from Fergus McKiernan, MD, Tom Faciszewski, MD: Marshfield Clinic, Marshfield, Wisconsin. Address correspondence and reprint requests to Fergus McKiernan, MD, Center for Bone Diseases, Marshfield Clinic, 1000 North Oak Avenue, Marshfield, WI 54449. E-mail: mckiernan.fergus@ marshfieldclinic.org. Submitted for publication November 11, 2002; accepted in revised form February 6, 2003. Kümmell s disease in severe, persistently mobile, clefted osteoporotic VCFs. The intravertebral cleft is generally considered a radiographic sign of avascular necrosis of the vertebral body (Kümmell s sign) and is not thought to be associated with acute osteoporotic vertebral compression fracture (VCF) (1). In a retrospective radiographic survey, intravertebral clefts were identified in 17 of 2,000 adult patients with benign and malignant musculoskeletal disorders and have thus been considered rare (2). Recent appreciation and characterization of the dynamic mobility of many osteoporotic VCFs has led to reconsideration of these assertions (3). Dynamic mobility refers to a change in the fractured vertebral body height in different body postures and implies the gross disruption of cortical and cancellous vertebral bone (Figure 1). Dynamic fracture mobility occurs frequently in patients presenting for vertebroplasty, and its magnitude can be substantial (4). The radiographic correlate of gross corticocancellous disruption in these mobile fractures is the intravertebral cleft (Figure 2). Here we describe the frequency, radiographic appearance, and significance of intravertebral clefts within VCFs of patients with osteoporosis presenting for vertebroplasty. The distinction between severe, persistently mobile, clefted osteoporotic VCFs and the clinical radiographic diagnosis of Kümmell s disease is challenged. PATIENTS AND METHODS The Institutional Review Board of Marshfield Medical Research Foundation approved this prospective radiographic analysis of 50 consecutive patients with 82 VCFs who underwent vertebroplasty at a tertiary referral center. Selection criteria for vertebroplasty, surgical outcome, and results of height restoration in a subset of this cohort have been reported (4). Dynamic fracture mobility was determined preoperatively by comparing standing lateral radiographs with supine crosstable lateral radiographs centered on the fractured (index) vertebra. Dynamic mobility was present when anterior, middle, 1414

INTRAVERTEBRAL CLEFTS IN OSTEOPOROTIC VCFs 1415 Figure 1. A, Standing lateral and B, supine cross-table lateral radiographs of a severe T12 vertebral compression fracture of 12 days duration, demonstrating immediate, complete postural height restoration in the supine position (dynamic mobility). T11 is fixed (immobile). The anatomic cleft within T12 is not apparent radiographically. C, Postoperative standing lateral (upper panel) and anteroposterior radiographs (lower panel), demonstrating polymethylmethacrylate fill. or posterior index vertebral height, manually measured to the nearest millimeter with unaided vision, changed between standing and supine radiographs. When no measurable change occurred, the VCF was considered immobile or fixed (Figure 1). Vertebral radiographs were obtained in accordance with the recommendations of the National Osteoporosis Foundation Working Group on Vertebral Fractures (5), and interradiographic measurement error was eliminated as previously described (4). All patients had preoperative short tau inversion recovery (STIR) sequence magnetic resonance imaging (MRI) of the index vertebra, and some underwent computed tomography or bone nuclear scintigraphy as clinically indicated. Intravertebral clefts were defined at the time of vertebroplasty as low-resistance, confluent reservoirs for polymethylmethacrylate (PMMA) (Figure 3D). Once an intravertebral cleft was identified, the preoperative radiographs, MRIs, and scintigraphic images were reexamined to confirm by which imaging modalities the cleft could be appreciated. Intravertebral location of the cleft, intraspinal location of clefted vertebrae, fracture severity, presence of dynamic mobility, radiographic appearance, and PMMA fill pattern were characterized for each VCF. Fracture severity was determined semiquantitatively and characterized by convention as mild Figure 2. A, Standing lateral and B, supine cross-table lateral radiographs of a severe and debilitating T12 vertebral compression fracture of 36 days duration (arrow in A), demonstrating dynamic mobility and a large intravertebral cleft seen as air density on plain radiograph (arrow in B). C, Short tau inversion recovery sequence magnetic resonance imaging, showing water density (arrow). Figure 3. A, Standing lateral and B, supine cross-table lateral radiographs of a severe and debilitating T12 vertebral compression fracture of 28 days duration, demonstrating an intravertebral cleft bounded by linear sclerosis (arrow in A) and a vacuum cleft sign (arrow in B). C, Short tau inversion recovery sequence magnetic resonance imaging, showing signal void (arrow). D, Postoperative standing lateral radiograph, showing confluent reservoir for polymethylmethacrylate (arrow).

1416 MCKIERNAN AND FACISZEWSKI Figure 4. A, Standing lateral and B, supine cross-table lateral radiographs of a severe and debilitating L1 vertebral compression fracture of 1 year s duration (arrow in A), demonstrating dynamic mobility and double intravertebral clefts seen as air densities on plain radiograph (arrows in B). C, Short tau inversion recovery sequence magnetic resonance imaging showing water density (arrow). (20 25%), moderate (25 40%), or severe ( 40% vertebral height reduction at any posterior, middle, or anterior point) (6). Fracture age was the best estimate of time from onset of fracture pain to vertebroplasty. Statistical significance was determined by Fisher s exact tests and Kruskal-Wallis nonparametric tests. Figure 6. A, Standing lateral and B, standing anteroposterior radiographs, and C, sagittal reconstruction computed tomography of a severe and debilitating clefted T12 vertebral compression fracture (VCF) of 95 days duration (arrows). D, Postoperative standing lateral radiograph, showing contrast of the cleft fill pattern of the T12 VCF (arrow) with the prior trabecular fill of T11 and L1 seen in A. Figure 5. A, Standing lateral and B, supine cross-table lateral radiographs of an intravertebral cleft within a T11 vertebral compression fracture of 139 days duration, visible as mature osseous sclerosis bordering a vacuum cleft sign (arrow in A) and dynamic mobility (arrow in B). C, Postoperative standing lateral radiograph, showing a cleft fill pattern (arrow). RESULTS Fifty consecutive patients underwent 56 vertebroplasty procedures to treat 82 VCFs. Thirty-three of 50 patients (66%) were women and 17 (34%) were men. The average patient age was 78.8 years. Average fracture age was 125 days (4.2 months). Twenty-four of 50 patients (48%) had at least 1 clefted VCF, and 30 of 82 VCFs (37%) contained clefts. Every clefted VCF was dynamically mobile, and no fixed VCF was clefted (P 0.001). Twenty-two of 30 clefted fractures (73%) occurred at the thoracolumbar junction (T11 L1), whereas only 15 of 52 unclefted fractures (29%) occurred at that site (P 0.001). Seventy-nine percent of clefted VCFs were severe, compared with 39% of fixed VCFs (P 0.001). Clefts were appreciated preoperatively by standing lateral radiography in 4 of 29 cases (14%) (Figures 3A, 5A, and 6A), by supine cross-table radiography in 18 of 28 cases (64%) (Figures 2B, 3B, 4B, and 5B), and by STIR sequence MRI in 25 of 26 cases (96%) (Figures 2C and 4C). The anatomic cleft (i.e., the gross corticocancellous disruption that permitted dynamic mobility) was sometimes not apparent radiographically (Figures 1A

INTRAVERTEBRAL CLEFTS IN OSTEOPOROTIC VCFs 1417 Persistent PMMA filling of unclefted vertebrae could result in a similar fill pattern (i.e., central opacification with spiculated margins), but the kinetics of cleft filling and the lack of sharp margins distinguished this from a mature cleft fill pattern. Both cleft and trabecular fill patterns could coexist in a single vertebra (Figures 7A and B). Figure 7. Postoperative standing lateral radiographs of A, T12 and L1 vertebral compression fractures (VCFs) of 53 days duration and B, T12 VCF of 28 days duration and L1 VCF of 12 days duration, demonstrating trabecular (small arrows) and cleft (large arrows) fill patterns. Both fill patterns are visible within each L1 VCF. and B). Persistently mobile fractures were characterized by increasing degrees of sclerosis at the cleft margins (Figures 3A and B, 5A and B, and 6A and B). Bone scintigraphy of 12 clefted VCFs demonstrated intense radiotracer uptake in 10, moderate uptake in 1, and mild uptake in 1. Nuclear scintigraphy did not appear to discriminate clefted from unclefted VCFs. In one instance, an intravertebral cleft was appreciated only at the time of PMMA injection. There was no instance where reexamination of the preoperative imaging studies led to recognition of a cleft that had been overlooked. Consistent with previous observations, the cleft content (air or water) varied with body position and was time dependent (7). Clefts were typically characterized by air density upon assumption of the supine position (Figure 6C) and water density after sustained maintenance of that posture. Clefts occurred just beneath the superior end-plate or occupied the anterosuperior portion of the vertebral body 80% of the time (Figures 2C, 3C, 5C, and 7A and B). Postoperatively, the PMMA fill patterns of clefted and unclefted VCFs were distinct. Persistently mobile, clefted vertebrae filled as a confluent reservoir for PMMA, with uniform opacity and sharp radiographic margins (Figures 3D, 6D, and 7A and B). Fixed, unclefted vertebrae filled with a trabeculated pattern that was spiculated at the margins (Figures 6A and B, and 7A and B). The mobile, unclefted VCF seen in Figure 1 filled with a hybrid pattern. DISCUSSION This prospective radiographic study documents the frequent occurrence of intravertebral clefts within VCFs of patients with osteoporosis presenting for vertebroplasty. Clefts represent gross corticocancellous disruption and predict dynamic fracture mobility. In recently fractured, mobile osteoporotic vertebrae, clefts may be radiographically inapparent (Figure 1). In the absence of fracture union and with persistent mobility, cleft margins become increasingly sclerotic and more apparent radiographically. Intravertebral clefts are often appreciated on the preoperative, supine cross-table lateral radiographs of patients with osteoporotic VCFs who present for vertebroplasty and are almost always appreciated by STIR sequence MRI as signal void. The void content (air or water) varies with posture and over time. Mobile, clefted VCFs cluster at the thoracolumbar junction and are the dominant fracture type at that location in this patient population. The thoracolumbar junction is the spinal region of greatest dynamic load, which might, therefore, predispose to fracture nonunion (8) and possibly pseudarthrosis (9). Every VCF that contained a radiographic cleft has also been mobile thus far. Within the vertebral body, clefts occur primarily in the anterior region just beneath the superior endplate, where vascular supply is most tenuous (10) and age-related microarchitectural deterioration is greatest (11). Clefted mobile VCFs were significantly more severe than unclefted, fixed VCFs that underwent vertebroplasty. Thus, the presence of an intravertebral cleft, particularly within vertebrae at the thoracolumbar junction, may predict a poorer anatomic fracture outcome. Ironically, the cleft also appears to be the intrinsic fracture property that permits dynamic mobility and presents the opportunity for significant anatomic restoration during vertebroplasty (4). Kümmell s disease is a radiographic diagnosis thought to represent delayed, posttraumatic avascular necrosis of the vertebral body (1,2). Kümmell s disease is considered uncommon and unrelated to postmenopausal or age-related osteoporosis. The diagnostic hallmark of Kümmell s disease is the intravertebral vacuum

1418 MCKIERNAN AND FACISZEWSKI cleft sign (Kümmell s sign) (12). Evidence indicates, however, that Kümmell s disease occurs primarily in elderly patients with osteoporosis, occurs in women more often than in men, appears primarily in the thoracolumbar junction (approximately three-fourths of reported cases occur between T11 and L1), and is associated with a greater degree of fracture severity (1,2,7,12 17). Kümmell s sign also favors the anterosuperior portion of the vertebral body (15,16). Postural accentuation of the intravertebral cleft in spinal extension (i.e., dynamic fracture mobility) has been noted repeatedly in Kümmell s disease (2,7,14,16,17). These similarities between Kümmell s disease and the severe, persistently mobile, clefted osteoporotic fractures that present at vertebroplasty suggest that Kümmell s disease is neither a distinct nor a rare pathophysiologic entity but rather the most catastrophic outcome of an injury response common to all osteoporotic vertebrae. Furthermore, since it is not unusual for the radiographic detection of a vertebral fracture to lag behind the clinical presentation of fracture pain, the characteristic posttraumatic delay of Kümmell s disease does little to distinguish this diagnosis from that of an ordinary osteoporotic vertebral fracture (18). Avascular necrosis is often, although not consistently, documented in Kümmell s disease (17), but trabecular microcracks, microfractures, and avascular necrosis of bone are increasingly recognized histopathologic features of radiographically fractured and even unfractured osteoporotic vertebrae (19,20). We and others (Lieberman I: personal communication) have observed osseous necrosis in bone obtained from the region of the intravertebral cleft at the time of vertebroplasty (results not shown). Careful histologic examination of the cleft surface in a small number of older, persistently mobile Kümmell s disease fractures showed areas of osseous necrosis bounded by a fibrocartilaginous membrane, suggesting pseudarthrosis (9). We speculate that cumulative vertebral microdamage, particularly within the anterosuperior vertebral body, predisposes the vertebral body to structural failure, a cycle of increasing compressive and dynamic stress, and increased intraosseous pressure that results in further vascular compromise and osseous necrosis. These biomechanical factors propagate a cleavage plane of gross corticocancellous disruption, which, in time, and particularly at the thoracolumbar junction, could result in fracture nonunion and lead to pseudarthrosis. Thus, avascular necrosis in ordinary osteoporotic vertebrae could simultaneously predispose to and result from vertebral collapse. In summary, the intravertebral cleft is a radiographic correlate of gross corticocancellous disruption and is common in VCFs of patients with osteoporosis who present for vertebroplasty. The anatomic vertebral cleft may not be detectable radiographically until cleft margins become sclerotic or cleft contents can be appreciated as air or water densities. When present, clefts are almost always visible on STIR sequence MRIs and are usually visible on supine cross-table lateral radiographs. Clefts cluster at the thoracolumbar junction and occur primarily in the zone beneath the anterosuperior endplate. Kümmell s disease may not be a distinct pathophysiologic entity, but rather may be the most catastrophic outcome of an injury response shared by all osteoporotic vertebrae. These observations cannot be generalized to all osteoporotic VCFs because this population sample is subject to substantial referral and selection bias. Clinical and radiographic outcomes differ between clefted and unclefted VCFs. Timing and technical approaches to vertebroplasty may differ between clefted and unclefted VCFs and require further investigation. REFERENCES 1. Theodorou DJ. The intravertebral vacuum cleft sign. Radiology 2001;221:787 8. 2. Kumpan W, Salomonowitz E, Seidl G, Wittich GR. The intravertebral vacuum phenomenon. Skeletal Radiol 1986;15:444 7. 3. Faciszewski T, McKiernan F. Calling all vertebral fractures classification of vertebral compression fractures: a consensus for comparison of treatment and outcome. J Bone Miner Res 2002;17: 185 91. 4. McKiernan F, Jensen R, Faciszewski T. The dynamic mobility of vertebral compression fractures. J Bone Miner Res 2003;18:24 9. 5. Kiel D, the National Osteoporosis Foundation Working Group on Vertebral Fractures. Assessing vertebral fractures. J Bone Miner Res 1995;10:518 23. 6. Black DM, Palermo L, Nevitt MC, Genant HK, Epstein R, San Valentin R, et al. Comparison of methods for defining prevalent vertebral deformities: the Study of Osteoporotic Fractures. J Bone Miner Res 1995;10:890 902. 7. Malghem J, Maldague B, Labaisse MA, Dooms G, Duprez T, Devogelaer JP, et al. Intravertebral vacuum cleft: changes in content after supine positioning. Radiology 1993;187:483 7. 8. Ismail AA, Cooper C, Felsenberg D, Varlow J, Kanis JA, Silman AJ, et al, the European Vertebral Osteoporosis Study Group. Number and type of vertebral deformities: epidemiological characteristics and relation to back pain and height loss. Osteoporos Int 1999;9:206 15. 9. Hasegawa K, Homma T, Uchiyama S, Takahashi H. Vertebral pseudarthrosis in the osteoporotic spine. Spine 1998;23:2201 6. 10. Ratcliffe JF. The arterial anatomy of the adult human lumbar vertebral body: a microarteriographic study. J Anat 1980;131:57 79. 11. Thomsen JS, Ebbesen EN, Mosekilde L. Zone-dependent changes in human vertebral trabecular bone: clinical implications. Bone 2002;30:664 9. 12. Resnick D, Niwayama G. Diagnosis of bone and joint disorders. 3rd ed. Philadelphia: WB Saunders; 1995. p. 2973. 13. Bhalla S, Reinus WR. The linear intravertebral vacuum: a sign of

INTRAVERTEBRAL CLEFTS IN OSTEOPOROTIC VCFs 1419 benign vertebral collapse. AJR Am J Roentgenol 1998;170: 1563 9. 14. Modena V, Maiocco I, Bosio C, Bianchi A, DeFilippi PG, Daneo V. Intravertebral vacuum cleft: notes on five cases. Clin Exp Rheumatol 1985;3:23 7. 15. Chou LH, Knight RQ. Idiopathic avascular necrosis of a vertebral body: case report and literature review. Spine 1997;22: 1928 32. 16. Maldague BE, Noel HM, Malghem JJ. The intravertebral vacuum cleft: a sign of ischemic vertebral collapse. Radiology 1978;129: 23 9. 17. Lafforgue P, Chagnaud C, Daumen-Legre V, Daver L, Kasbarian M, Acquaviva PC. The intravertebral vacuum phenomenon ( vertebral osteonecrosis ): migration of intradiscal gas in a fractured vertebral body? Spine 1997;22:1885 91. 18. Fogelman I, Cook GJR. Scintigraphy in metabolic bone disease. In: Favus MJ, Goldring SR, Christakos S, editors. Primer on the metabolic bone diseases and disorders of mineral metabolism. 4th ed. Philadelphia: Lippincott; 1999. p. 146 52. 19. Wenzel TE, Schaffler MB, Fyhrie DP. In vivo trabecular microcracks in human vertebral bone. Bone 1996;19:89 95. 20. Antonacci MD, Mody DR, Rutz K, Weilbaecher D, Heggeness MH. A histologic study of fractured human vertebral bodies. J Spinal Disord Tech 2002;15:118 26.