Portal de Periódicos da CAPES

Calling All Vertebral Fractures Classification of Vertebral Compression Fractures: A Consensus for Comparison of Treatment and Outcome

2002; Oxford University Press; Volume: 17; Issue: 2 Linguagem: Inglês

10.1359/jbmr.2002.17.2.185

1523-4681

Tom Faciszewski, Fergus McKiernan,

Bone health and osteoporosis research

Osteoporosis is a disease characterized by low bone mass, microarchitectural deterioration of bone tissue, and the propensity to low-energy fracture.1 The most frequent osteoporotic fracture is the vertebral compression fracture (VCF) of which only one in three manifests as an acutely painful event.2-4 Thus, prevalent radiographic VCFs outnumber clinical fractures. Several fracture definition schemes exist but all are essentially binary decisions (i.e., fracture vs. no fracture) based on that degree of vertebral height reduction that results in disease prevalence and incidence most accurately reflecting the problem of osteoporosis in the general population.5-10 Vertebral fractures serve as sentinel markers for the epidemiological study of osteoporosis1, 11 and are the benchmark by which efficacy of new osteoporosis therapies is established.12 Therefore, consensus on vertebral fracture definition is extremely important because small changes in fracture definition might potentially result in profound alterations in fracture epidemiology, diagnostic strategies, and treatment outcomes. On the other hand, vertebral fracture morphology is descriptive (i.e., wedge, crush, and biconcave), unstandardized, and of uncertain clinical relevance.6, 8 Vertebral fractures, whether clinically apparent or not, are associated with an increased risk of recurrent vertebral fracture,13 future nonvertebral osteoporotic fracture,14, 15 and future hospitalization and mortality.16, 17 Prevalent VCFs predict decreased pulmonary capacity and increased risk of pulmonary death.18-20 The number and severity of prevalent compression fractures correlates with frailty and poor functional status by multiple validated quality of life instruments.21-23 Severely painful VCF may initiate the frustrating cascade of impaired mobility, physical deconditioning, accelerated bone loss, and further frailty. Long-term consequences of VCF include postural deformity, chronic back pain, abdominal crowding, altered body image, social withdrawal, fear of future fracture, and depression.24, 25 Thus, fracture-prevention strategies and interventions that promote postural and functional recovery from acute VCFs are welcome. Conventional treatment of painful VCFs has been supportive, rehabilitative, and, heretofore, rarely surgical. Previous surgical interventions for VCF have resulted in significant morbidity and poor surgical outcomes because of the insufficient material properties of osteoporotic bone and the advanced age and multiple medical comorbidities of this patient population.26 Therefore, the emergence of percutaneous vertebral augmentation (PVA) is a hopeful option for managing the pain and sequelae of vertebral fractures in osteoporotic patients. PVA refers primarily to vertebroplasty, the percutaneous fixation of fractured vertebrae with polymethylmethacrylate (PMMA).27 Kyphoplasty, a proprietary derivative of vertebroplasty, is the PMMA fixation of fractured vertebrae after purported vertebral end plate elevation using a percutaneous inflatable balloon tamp.28 Both procedures claim to relieve vertebral fracture pain, quickly restore functional capacity, and abort the cascade of accelerating frailty previously described.27, 28 The accessibility and perceived technical simplicity of vertebral augmentation, the desire to relieve suffering, the recent assignment of current procedural terminology (CPT) codes for vertebroplasty (but not kyphoplasty), and favorable short-term outcomes reported in small uncontrolled series have fueled enthusiasm for these procedures. The number of web-informed patients and "certified operators" has grown so that both procedures are now performed with increasing frequency outside the referral center setting and independent of clinical trials. Unfortunately, the immediate and long-term risks and benefits of PVA remain poorly defined, and the potential for catastrophic outcome is very real and likely underreported. There are no prospective, randomized studies that have shown long-term superiority of PVA over conventional medical therapy and rehabilitation. There are less than a dozen (mostly retrospective) clinical reports of vertebroplasty (consisting of five or more patients) published in peer review English language literature.27, 29-35 Only one kyphoplasty report meets these criteria.28 Many questions remain unanswered not the least of which are at what point in the evolution of vertebral fracture is augmentation indicated, in which patient, by whom, using which technique and filler material? Until carefully designed, prospective clinical studies of PVA are completed these questions will remain unanswered and the relative merits of any one particular method of augmentation over another remain unproved. In the interim, PVA should be performed cautiously and we should analyze our augmentation experiences rigorously, even if retrospectively. Like others, we are accumulating a clinical database from which we hope to better define the risks and benefits of augmentation, improve candidate selection criteria, refine technical aspects, and describe long-term outcomes. In our first year performing PVA, we evaluated over 100 patients and 250 fractures. Our initial experience with vertebral augmentation suggested that existing binary case definition schemes (i.e., fracture vs. no fracture) and descriptive morphometric analyses of shape (wedge, crush, and biconcave) and degree (mild, moderate, and severe) were simply inadequate to characterize the complexity of fracture types we routinely encountered. Other fracture dimensions including fracture age, the state of fracture reparative activity, the presence of gross trabecular disruption, dynamic fracture mobility, the presence or absence of posterior wall cortical disruption, and fracture retropulsion need to be considered when characterizing individual fractured vertebrae and engaging in PVA. The predictive value of any of these dimensions is, in so far as natural history or treatment outcomes is concerned, unvalidated.36 Fracture mobility has been observed, but its frequency and significance are unknown.27 Individual vertebral fractures are unique and any one fracture may not remain the same over time. These fracture dimensions, infrequently discussed in the medical and surgical osteoporosis literature, become practical issues when considering vertebral augmentation. We should anticipate the possibility that individual fracture types might have different natural histories and respond to different therapeutic interventions. Three cases, representative of our experience with PVA, illustrate the need to consider multiple fracture dimensions when engaging in and interpreting the outcome of PVA. We recognize that referral and selection bias might alter the relative frequencies of our observations but do not, in our opinion, diminish their significance. A 70-year-old white female with steroid-dependent rheumatoid arthritis (RA) and severe osteoporosis interrupted radiation therapy of non-small cell carcinoma of the lung because a painful L1 VCF confined her to bed. Fractures of L4 and L5 were old and asymptomatic. Balloon tamp elevation of the L1 vertebral end plates with PMMA fixation was planned but when intraoperative positioning alone resulted in significant vertebral height restoration, a conventional vertebroplasty was performed instead. Fracture pain was relieved and radiation therapy was resumed (Fig. 1). (A) Severe acute wedge fracture of L1. Chronic moderate biconcave fractures of L4 and L5. (B) Intraoperative hyperextension reveals gross intravertebral trabecular disruption and fracture mobility. An increase of vertebral height of over 100% of fractured height is attained. Vertebral body kyphosis is reduced to normal. (C) Acute severe L1 mobile wedge fracture with intravertebral disruption rendered stable with filling of the vertebral void with PMMA. A 90-year-old white female with severe postmenopausal osteoporosis had multiple previous atraumatic VCF. T12 fractured 6 months before evaluation but fracture pain improved and became tolerable. T11 fractured 3 months before evaluation but fracture pain persisted and became intolerable. There was violation of the posterior cortical wall but no fracture retropulsion. An intravertebral magnetic resonance (MR) signal void correlated with intravertebral gas seen on plain radiographs. Vertebral motion was established with dynamic posturing (hyperextension). She underwent PMMA fixation of T11 after intraoperative postural restoration of vertebral height and experienced rapid pain relief and improved stature (Fig. 2). (A) MR imaging (MRI) reveals many fracture types in 1 patient. T11 is an acute severe crush fracture with edema. T12, L2, L4, and L5 represent a variety of fracture morphologies without edema. (B) Standing lateral radiograph of T11 with 19° of vertebral body kyphosis and a gas-filled cleft. (C) Hyperextension lateral radiograph documents mobility of the T11 fracture with an increase of vertebral height of over 100% of fractured height. Vertebral body kyphosis reduced to 1° of lordosis. A gas-filled cleft is seen to increase. (D) Magnified spot radiograph of T11 vertebra in hyperextension. Vertebral end plate margins and gas-filled cleft are outlined by arrows. (E) Vertebral alignment of neutral standing and hyperextended spine. T11 mobility affects sagittal alignment significantly. (F) Intraoperative radiograph with three-fourths of the vertebral body void filled with PMMA. The chronic severe mobile wedge fracture with intravertebral disruption is rendered stable. A 70-year-old white female with severe postmenopausal osteoporosis and unrecognized Parkinson's disease remained nonambulatory 6 months after compression of T12. The wedge fracture, initially mild, progressed since the time of original presentation and became severe. The T12 vertebral body was edematous, demonstrated an intravertebral MR signal void, and was mobile with dynamic posturing. She underwent T12 PMMA fixation after intraoperative postural restoration of vertebral height. She had rapid pain relief and resumed assisted ambulation (Fig. 3). (A) Chronic mild wedge fracture of T12 with 14° of vertebral body kyphosis. (B) T12 7 weeks later has progressed to a severe wedge fracture with 35° of vertebral body kyphosis on standing radiograph. No trabecular disruption apparent on standing radiograph. (C) MRI scan reveals intravertebral gas manifest as the black signal void on T1 imaging. Vertebral edema is represented by decreased signal intensity of T12. (D) Supine hyperextension radiograph shows mobility with an increase of vertebral height of 110% of fractured height. Vertebral body kyphosis reduced to 23°. Intravertebral gas defines the vertebral void. (E) Computed tomography (CT) scan with intravertebral disruption visible in the neutral position. The size of the void is decreased in this position. (F) Lateral and posterior-anterior (PA) intraoperative radiograph of the chronic severe wedge fracture with intravertebral disruption that is rendered stable by filling the intravertebral void with PMMA. These clinical examples illustrate the heterogeneity and surprisingly dynamic aspect of many fractured vertebrae for which there exists no prevailing nosology or consensus of language. Based on these observations, the correction of vertebral body kyphosis and statural deformities in certain dynamic fracture types is, in part, postural and should not be attributed to any vertebral augmentation procedure. To anticipate the degree of postural fracture reduction and angular correction, we now routinely perform preprocedural dynamic (hyperextension) radiographs. In our experience, a significant proportion of painful fractures we encounter appear mobile—a finding the literature had not led us to anticipate. Compared with conventional preoperative standing radiographs, height of mobile fractures increases an average of 68% (data on file) with dynamic positioning alone. We routinely show posturally induced vertebral end plate elevations, vertebral height restorations and reductions of vertebral body angular kyphosis that exceed published values ascribed to tamp procedures (data on file). Finally, we observe the presence of gross intravertebral trabecular disruption with regularity. This is sometimes manifest on plain radiograph as gas, as signal void on MR, or occasionally seen only at the time of cement injection as a low-resistance/high-capacitance reservoir for PMMA. In some of our cases (e.g., case 2), this finding bears resemblance to the radiographic appearance of delayed posttraumatic vertebral osseous necrosis (Kummel's spondylitis) discussed in the radiology literature of spine trauma.37-41 In this condition, visualization of an intravertebral cleft is both time and position dependent, appearing as gas or water under different circumstances.39 This literature and our experience suggest that vertebral osseous necrosis may be a shared final pathway for both high- and low-energy vertebral injury. We speculate that chronic instability in untreated, mobile fractures may lead to osseous necrosis, a plane of microscopic or gross trabecular disruption, an intravertebral cavity (cleft), or pseudarthrosis. It seems intuitive that microscopic or gross intravertebral motion and fracture pain are causally related, but how is it that two-thirds of all VCFs are painless? Without a consensus of language and a nosology that accommodates these and perhaps other dimensions of vertebral fracture, any future discussion of fracture epidemiology, natural history, therapeutics, and, in particular, the value of PVA may be flawed. Only carefully designed and executed clinical trials will determine whether these observations are verifiable or relevant but failure to anticipate and prospectively incorporate these considerations into data collections may prove regrettable. We speculate that these observations will help determine appropriate clinical indications, selection criteria, surgical technique, and outcome of PVA. They may even influence the very definition of vertebral fracture. The multidimensional scheme for the classification of VCFs shown in Table 1 is an initial attempt to accommodate our observations. A consensus discussion on these (and possibly other) dimensions of the fractured vertebra should improve our understanding of vertebral fracture and clarify the therapeutic role of vertebral augmentation. Six dimensions are identified in this classification scheme. The first dimension, fracture morphometry, describes the number, anatomic locations, morphology, and degree of fracture. The anatomic location is regional and specific (e.g., the 12th thoracic vertebra). Morphology is conventionally characterized as wedge, crush, or biconcave, recognizing that these configurations are not mutually exclusive.8 Fracture degree can be calculated semiquantitatively or quantitatively and characterized as mild (3-3.5 SD below reference normative data or vertebral height reduction of 20-30%), moderate (3.5-4 SD below reference normative data or 30-40% vertebral height reduction), and severe (>4 SD below reference normative data or >40% vertebral height reduction).6, 7, 10 The second dimension is chronicity. Fracture onset is dated to the onset of pain ascribed to that fracture. Recognizing problems inherent in symptom recall and pain localization, we arbitrarily consider acute fractures as those 3 months old. It remains uncertain whether time since fracture or the time dependence of certain fracture dimensions (e.g., morphology, degree, and reparative activity) will prove to be important. The third fracture dimension is fracture reparative activity. Fractures that are edematous on T2-weighted chemical fat suppression or short tau inversion recovery (STIR) MR sequences or those that accumulate radiopharmaceutical on bone scintigraphy are considered persistent or active. Fractures without edema on appropriate MR sequences or that do not accumulate radiopharmaceutical on bone scintigraphy are considered healed. Dynamic fracture mobility is the fourth dimension. Dynamic fracture mobility is determined by comparing standing lateral radiographs to hyperextension supine cross-table lateral radiographs preoperatively and intraoperatively. A measurable decrease in vertebral body kyphosis or increase in anterior (Ha), middle (Hm), or, less likely, posterior (Hp) vertebral height is considered a mobile fracture. Fractures with no measurable changes are considered fixed. Fifth, fractures are assessed for the presence or absence of gross intravertebral trabecular disruption or clefts. Clefts may be characterized by intravertebral gas on plain standing or hyperextension radiographs or as signal void on MR depending on patient positioning and position duration. On occasion, clefts are only appreciated at the time of injection as a low-resistance/high-capacitance reservoir for PMMA. Finally, fractures are classified based on the presence or absence of violation of the posterior vertebral cortex and with or without retropulsion of fracture fragment. The literature to date has discouraged augmentation when retropulsion is present because of the risk of PMMA leakage into the spinal canal. PVA is an exciting new intervention that eventually will find its place in our therapeutic armamentarium. The anticipated availability of skeletal growth factors and resorbable biomaterials makes the future of augmentation seem even brighter. However, before vertebral augmentation can become a permanent part of our armamentarium many important questions not only need to be answered but also need to be asked. If the questions we ask fail to recognize the heterogeneity of clinical VCFs, then the answers we obtain may be unhelpful. We need a consensus of nosology and language that describes our collective experience with vertebral fractures and we need it as clinical trials of PVA are being designed and initiated. In our call to evaluate all vertebral fractures, let's be careful that we aren't calling all vertebral fractures the same.