2 Vertebral Compression Fractures
Summary
Vertebral compression fractures (VCFs) result from either trauma or pathologic weakening of the bone by conditions such as osteoporosis or malignancy. They often cause severe pain, physical limitation, and disability, and lead to increased morbidity and mortality with considerable heath care expenditures. Fracture classification systems have been designed to identify spinal instability and guide treatment. The public health and economic impact is considerable. Regardless of etiology, timely and accurate diagnosis will assist in determining the appropriate treatment.
2.1 Introduction
Vertebral compression fractures (VCF) result from either trauma or pathologic weakening of the bone by conditions such as osteoporosis or malignancy. They often cause severe pain, physical limitation, and disability, and lead to increased morbidity and mortality. 1 – 6 Vertebral fractures, whether they are symptomatic or asymptomatic, are associated with increased morbidity and mortality, and patients with osteoporotic vertebral fragility fractures have an increased risk of mortality compared to age-matched controls. 7 – 9 Annual health care expenditures to diagnose and treat vertebral fractures are considerable and are on the rise. 10
The thoracolumbar region is the most common location for fractures of the vertebral column. 11 – 13 The thoracic spine is functionally rigid due to coronally oriented facet joints, thin intervertebral disks, and stabilization offered by the rib cage, so significant trauma is required for fracture or dislocation unless the vertebrae are pathologically weakened by disease (▶Fig. 2.1). The presence of the spinal cord in this region predisposes the patient to neurological injury if the canal is compromised causing significant impingement on indwelling the neural tissue. The lumbar spine, on the other hand, is relatively flexible due to the thicker intervertebral disks, sagittally orientated facet joints, and absence of the rib cage (▶Fig. 2.2). The relatively lesser incidence of neurological injury in lumbar fractures can be attributed to the large size of the neural canal and the greater resilience of the cauda equina nerve roots. The thoracolumbar junction (T10–L2) is uniquely positioned between the rigid thoracic spine and the mobile lumbar spine. This transition from the less mobile thoracic spine with its associated ribs and sternum to the more dynamic lumbar spine subjects the region to significant biomechanical stress. 12 , 13 Among thoracolumbar injuries, 50 to 60% affect the transitional zone (T11–L2), 25 to 40% affect the thoracic spine, and 10 to 14% involve the lower lumbar spine and sacrum. 14
2.2 Presentation and Diagnosis
Vertebral fractures may have a profound and untoward effect on quality of life. The presentation of patients with VCFs varies depending on etiology, severity, and the level or levels affected. Discrete onset axial back pain after lifting or bending is often reported and may or may not be associated with sciatica. Pain is typically positional, worsened by weight-bearing activities, transition from one position to another, and by lying supine. Patients without significant trauma rarely have neurologic deficits, but elderly osteoporotic patients, in particular, have marked functional impairment with activity reduced by postural changes and pain. Multiple thoracic fractures can result in severe kyphosis, which may lead to restrictive lung disease. 15 Lumbar fractures may alter abdominal anatomy, leading to constipation, abdominal pain and distention, reduced appetite, premature satiety, and urinary incontinence. 16 These restrictions increase the risk of losing independence, social isolation, and clinical depression.
2.3 Epidemiology
2.3.1 Osteoporosis
Osteoporosis is the most common condition associated with VCFs and affects approximately 700,000 individuals each year in the United States and 30 to 50% of people older than 50 years worldwide. 17 – 20 Osteoporosis or “porous bone” is a systemic bone disorder resulting from an imbalance between bone formation and bone loss and is characterized by diminished bone strength. Bone strength is dependent on both the quality of the bone and the bone mineral density (BMD). Bone quality refers to architecture, mineralization, accumulation of damage (microfractures), and turnover. Bone density is determined by peak bone mass, which is usually attained by age 30 and subsequent bone loss. 21
Osteoporosis can occur in both sexes, but is more common in women following menopause. 21 Both men and women have age-related decline in BMD in middle age due to increased bone resorption as compared to bone formation. This is especially common in women as they experience more rapid bone loss in the early years after menopause. 21 Peak bone mass is achieved only after linear bone growth has ceased. Therefore, bone mass attained early in life may be the most important determinant of skeletal strength later in life. 21 , 22 Factors that affect achievement of peak bone mass include lifetime calcium and vitamin D intake, physical activity, smoking, alcohol, eating disorders, and endocrine disorders. 22 Characteristics associated with low bone mass later in life include female sex, increased age, estrogen deficiency, white race, low body mass index (BMI), and a family history of fractures. 21
As a complication of osteoporosis, VCFs are associated with significant morbidity and mortality and are an increasing public health risk as the population continues to age (▶Fig. 2.3). Vertebral fracture incidence increases substantially with age in both males and females. The annual incidence of vertebral fractures increases from 0.9% among middle-aged women in their 50s to 60s to an incidence of 1.7% among those 80 years and older. 10 , 17 , 23 The mere presence of a vertebral fracture is associated with increased risk of future fractures. If the patient presents with one VCF, the risk of having another one within the first year is increased by fivefold. 17 If the patient has two fractures, the risk increases to 12-fold and if three or more fractures are present, the risk is 75 times increased for having another vertebral fracture.
BMD as measured in gram per square centimeter (g/cm2) and its related T-score are used to screen for the objective presence of osteoporosis. The T-score is the number of standard deviations above or below the mean BMD for a healthy 30-year-old reference patient. The U.S. standard is to use data for a 30-year-old of the same sex and ethnicity, but the World Health Organization (WHO) recommends using data for a 30-year-old white female for everyone. Values for 30-year-olds are used in postmenopausal women and men older than 50 years because they better predict risk of future fracture. The criteria of the WHO are as follows: Normal is a T-score of -1.0 or higher. Osteopenia is defined as between -1.0 and -2.5. Osteoporosis is defined as -2.5 or lower, meaning a bone density that is 2.5 standard deviations below the mean of the reference. 24
It should be noted that while bone density scanning and determination of the patient’s BMD and T-score are valuable from a screening and treatment monitoring perspective, it does not determine which vertebral fractures are osteoporotic VCFs as a greater proportion of vertebral fragility fractures occurs in patients with either a normal bone density or with osteopenia as compared to patients whose T-score shows osteoporosis. The clinical scenario and amount of force required to produce that VCF is a far better determination of whether the patient’s fracture is traumatic or the result of osteoporosis than their T-score.
Using data from the National Health and Nutrition Examination Survey (NHANES), an estimated 16% of U.S. men and almost 30% of U.S. women aged 50 and over have osteoporosis as defined by their bone BMD. 25 An estimated 50% of women and 20% of men in the United States will sustain an osteoporotic fracture at some point in their lifetime. 16 The prevalence of vertebral fracture is similar in men and women and increases with age from less than 5% in those younger than 60 years of age to 11% in those age 70 to 79 years to 18% in those aged 80 years or older. 26 The age-correlated prevalence rates in the United States are very similar to those in Europe, as published by the European Vertebral Osteoporosis Study. 18 , 27 – 31
More than 2 million osteoporotic fractures including 700,000 VCFs were diagnosed in 2005 and that number of osteoporotic fractures is projected to increase to over 3 million fractures by 2025, an increase of 52%. 32 Importantly, this calculation may be an underestimate as it has been projected that less than one-third of all vertebral fractures are clinically diagnosed (see Chapter 34: Treatment of Osteoporosis after Vertebral Augmentation). 33
2.3.2 Cancer
Metastases
The spine is the most frequent site of bone metastasis. 34 In patients with cancer, the risk for pathologic VCF results from bone involvement, with fracture incidence estimated to be 24, 14, 6, and 8% among patients with multiple myeloma (MM) and cancers of the breast, prostate, and lung, respectively. 35 , 36 These fractures are most often of the compression or burst type and may result in axial or radicular pain, and sometimes neurologic and/or motor/sensory deficits. The incidence of symptomatic metastatic spinal tumors in the United States is estimated to be approximately 160,000 per year 37 and between 6 and 24% of these patients are predicted to have a VCF at some point over the course of their disease. 35 Unfortunately, pathologic fractures of the spine are often the initial manifestation of a metastatic malignancy (▶Fig. 2.4).
Multiple Myeloma
Although often perceived as a rare disease, MM is in fact the second most commonly diagnosed hematologic malignancy in the Western world. 38 MM comprises 1.6% of all bone malignancies in the United States and occurs primarily in the elderly, with a median age at diagnosis of 69 years. 39 The 5-year survival rate is less than 50%. 40 VCFs are the most common type of fractures in patients with MM, occurring at the onset of the disease in 34 to 64% of patients. 41 , 42
Unlike osteoporosis, which causes weakening of trabecular bone structure and lower bone mass, the mechanisms of MM leading to significant structural weakness are varied and remain incompletely understood. There is increased osteoclast precursor differentiation and, consequently, enhanced bone resorption. 43 Early diagnosis and treatment are critical in slowing the disease progression and the corresponding deterioration of the patient’s quality of life. 44
Most MM patients report back pain at the time of initial diagnosis. 45 Up to 80% of patients with MM are diagnosed during routine imaging. 46 Radiological findings include diffuse bone loss, focal osteolytic bone lesions, bone marrow edema, and fragility fractures of the axial skeleton. 47 Bone marrow edema is a common MR imaging finding of acute VCFs but may be difficult to see due to bone marrow signal abnormalities seen in patients with MM. 48 Diffuse bone loss alone is often misdiagnosed as osteoporosis until more symptoms associated with MM develop. 49 When focal osteolytic lesions or significant diffuse bone loss is evident, risk for vertebral fractures is high. 50
The patient’s osteoporosis status could be an indicator of disease progression to MM. 51 Twenty percent of osteoporotic patients presenting with vertebral fractures may have either monoclonal gammopathy of undetermined significance or MM. 50 Almost 80% of MM patients are diagnosed with osteoporosis and the level of BMD has a major impact on survival. 52 , 53 However, there are several problems with the use of BMD as a diagnostic tool of MM. First, the decline in BMD as a result of aging and/or osteoporosis is well understood; it declines 0.1 to 0.2% per year due to aging while after menopause and onset of osteoporosis, it peaks to 1 to 2% and then slows back to the decline seen in normal aging. 47 The decline in BMD in MM-induced osteoporosis is less well understood and unpredictable. In a recent study, Borggrefe et al found that BMD of fracture cases in MM patients were significantly reduced in men, but not in women. 54 Second, as has been mentioned previously, BMD and T-score as determined by dual X-ray absorptiometry (DXA) has been challenged as a limited tool for the diagnosis of osteoporosis itself as it only partially estimates fracture risk. The use of DXA as a diagnostic evaluation tool in MM is even less valid. 55 Also, routine assessment of BMD in MM patients is not recommended due to methodological difficulties in these patients and the frequent use of bisphosphonates in all symptomatic MM patients. 56 Finally, there are no established and well-defined clinical criteria to differentiate osteoporotic VCFs and MM-induced osteoporotic VCFs, but MR imaging can be helpful in patients who are suspected to have MM (▶Fig. 2.5). 57
2.3.3 Trauma
The annual incidence of traumatic spinal injuries is greater than 160,000 in North America. 58 According to Hu et al, the incidence of traumatic spinal injuries was 64/100,000 population/y in Canada alone. 59 Of the injuries to the thoracolumbar region, around 50 to 60% affect the transitional zone (T11–T12), with another 25 to 40% affecting the thoracic spine, and the remainder affecting the lower lumbar spine and sacrum. 14 The peak incidence of thoracolumbar fractures is observed in patients who are between 20 and 40 years of age with these injuries being more common in men. 14 , 60 Thoracolumbar fractures are associated with neurologic injury in 20 to 36% of cases 61 , 62 and the type of fracture affects both the chances for and the extent of a neurological deficit. In a multicenter study, the incidence of neurological deficit ranged from 22 to 51% depending on the fracture type (AO Classification type A: 22%; type B: 28%; type C: 51%; ▶Table 2.1). 11 , 63
Most thoracolumbar fractures are associated with trauma and 65% occur due to motor vehicle injuries or falls from a height, with a lesser portion of these fractures attributed to sports injuries or violence. Due to the high-velocity nature of these traumatic fractures, they are commonly associated with other injuries such as hemopneumothorax, rib fractures, vessel injuries, and diaphragmatic rupture. 64 , 65 Seatbelt (Chance) fractures and flexion–distraction injuries are other traumatic spinal fractures that are often associated with intra-abdominal visceral injuries. First described in 1948 by G.Q. Chance, the Chance fracture is a purely bony injury extending from posterior to anterior through the spinous process, pedicles, and vertebral body (VB), respectively. Seen most often in the upper lumbar spine in adults, it typically results from excessive flexion, as caused by a lap belt (without shoulder belt support) injury during a motor vehicle accident. Incidence has been reduced dramatically by the required use of shoulder belts in motor vehicles.
History and physical examination help guide the choice of the initial imaging modality. Many times, advanced and/or dynamic imaging will be needed to detect spinal instability. A careful history regarding vertebral fracture injury mechanism, pain, and the presence or absence of neurological symptoms is essential. The patient may report pain associated with movement and/or with prolonged positions and give-way weakness or a feeling of “instability.” Axial, nonradiating back pain of stabbing or aching quality is the most common symptom. Patients with neurological injury may complain of weakness, paresthesia, or anesthesia below the injury level and may have urinary retention. Thorough inspection of the spine should be performed after a careful log roll maneuver to look for abrasions, tenderness, local kyphosis, and a palpable gap in between spinous processes. Neurological assessment should follow the standard American Spinal Injury Association (ASIA) guidelines (▶Fig. 2.6). 66
As the spinal cord normally ends between the L1 and L2 level and the cauda equina fills the distal canal, varied neurological injury patterns can be observed with thoracolumbar fractures. Neurological injuries at or above L1 can injure the spinal cord and injuries below L2 typically affect only the nerve roots of the cauda equina. Conus medullaris syndrome, characterized by exclusive damage to sacral innervations to the bowel and bladder with intact lumbar nerve roots, is a unique feature of T12–L1 injuries.
2.4 Vertebral Fracture Classification Systems
Numerous classification systems have been proposed to describe thoracolumbar spinal fractures resulting from osteoporosis, neoplasia, or acute trauma. Each system was designed to improve upon the last as a reliable way to identify spinal instability and, ultimately, to guide treatment.
In 1938, Watson-Jones first described the concept of spinal instability and developed a system that integrated the posterior ligamentous complex to described seven fracture types in three major patterns: simple wedge fractures, comminuted fractures, and fracture-dislocations. 67
Over a decade later in 1949, Nicoll improved the classification system by formally defining the anatomy that was responsible for spinal stability, including the VB, the disks, the intervertebral joints, and the interspinous ligament. 68 He emphasized the importance of the interspinous ligament as a crucial component of stability.
Holdsworth was the first to describe stability in terms of two columns. 69 , 70 In his two-column theory, compressing the anterior column causes distraction of the posterior column and vice versa. Holdsworth expanded on Nicoll’s assumption that the interspinous ligament was paramount in determining stability. He described a posterior ligamentous complex (the posterior column) and believed that the entire posterior column was crucial for stability. Therefore, loss of integrity to this posterior ligamentous complex would define instability. Holdsworth also described mechanisms of spinal injuries and defined six groups: anterior wedge-compression fractures, dislocations, rotational fracture-dislocations, extension injuries, burst fractures, and shearing fractures. This classification system predates the modern imaging modalities and is now rarely used.
Panjabi 71 and White et al 72 described three conceptually separate but interdependent components responsible for stability: (1) a passive subsystem consisting of the vertebrae, facets, intervertebral disks, spinal ligaments, and joint capsules; (2) an active subsystem of spinal muscles and tendons; and (3) a control subsystem for neural feedback, including the neural control centers and force transducers located in ligaments, tendons, and muscles.
Introduced in 1982, the Allen and Ferguson system, also rarely used, was a “mechanistic” system that classified spinal injuries based on three common mechanical modes of failure of the spine: compression–flexion, distraction–flexion, and compression–extension. 73
Also introduced in 1982, the Denis three-column spinal stability classification system describes fracture types and included possible mechanisms of failure. The anterior column is defined as the anterior longitudinal ligament, the anterior half of the VB, and the anterior half of the annulus fibrosis. The posterior column consists of everything posterior to the posterior longitudinal ligament: the facet joints, pedicles, and supraspinous ligaments. The novel middle column contains a middle osteoligamentous complex formed by the posterior half of the VB, the posterior longitudinal ligament, and posterior annulus fibrosus (▶Fig. 2.7). 74
As this model has an additional column, novel mechanisms of injury are described. There are four major types: compression, burst, seatbelt type, and flexion–distraction. Compression fractures are classified according to whether they are caused by anterior or lateral flexion. These injuries cause anterior column compression sometimes associated with posterior column distraction. Burst fractures are classified according to whether an axial load alone caused the fractures or whether flexion, rotation, and/or lateral flexion are also involved. These injuries cause anterior and middle column compression, and may sometimes also produce associated posterior column distraction. Seatbelt fractures are defined by their distinctive flexion–disruption mechanism of injury. The anterior column may be intact or distracted, but the middle and posterior columns are distracted. Fracture-dislocations are classified according to whether the dislocation involves rotation, shear, or flexion–distraction. Each of these injuries is further subdivided and can produce any pattern of column involvement (▶Table 2.2 and ▶Fig. 2.8). Minor fractures—those of the transverse processes, articular processes, pars interarticularis, and spinous processes involve only a part of the posterior column and are not thought to lead to acute instability. 75
Using this system, only compression fractures without involvement of the posterior wall would be amenable to treatment with vertebral augmentation. While low-velocity compression fractures were currently considered stable, Denis felt that all burst fractures were unstable and required surgical instrumentation for stabilization. As every burst fracture required the same treatment, the five subdivisions were thought to be of little value.
The following year, McAfee et al simplified this system and was the first to describe a stable burst fracture: a fracture having anterior and middle column compression without posterior column disruption. The McAfee system of classifying acute traumatic spinal injuries is based on three forces (axial compression, axial distraction, and translation within the transverse plane) as they act to injure the middle column. Computed tomography can be used as a reliable method of identifying unstable burst fractures by illustrating facet joint subluxation or disruption of the neural arch. 76 With this system, vertebral augmentation would be appropriate for compression fractures, including stable burst fractures.
The AO (Arbeitsgemeinshaft fur Osteosynthesefragen) Classification is a highly detailed system published in 1994 by Magerl et al. 61 This Swiss system returns to a two-column model and remains the basis for modern fracture fixation. It classifies thoracolumbar fractures into three major groups based on the mechanism of injury: compression (type A), distraction (type B), and multidirectional with translation or torsion (type C). It takes into account the morphological appearance, the direction of the force or mechanism of injury, and increasing destruction of the injury. 77 Its comprehensive description includes the nature of injury, the degree of instability, and prognostic aspects that are important for choosing the most appropriate treatment.
The modified AO classification includes morphology of the fracture, neurological status, and description of relevant patient-specific modifiers (▶Fig. 2.9). 78 The fracture morphology is assessed based on three main injury patterns: type A0–A4 (compression injury to the VB without posterolateral ligamentous complex [PLC] involvement), type B1–B3 (tension band disruption—the failure of posterior [PLC] or anterior [anterior longitudinal ligament] constraints), and type C (translation injuries). Neurologic status is classified as follows: no neurologic injury (N0), transient (resolved) neurologic deficit (N1), radicular symptoms (N2), incomplete spinal cord injury or any degree of cauda equina injury (N3), complete spinal cord injury (N4), and unknown neurologic status (NX). Two modifiers were also included: M1 to indicate indeterminate posterolateral complex status and M2 to designate patient-specific comorbidity (that might preclude general anesthesia and surgery, such as ankylosing spondylitis or diffuse burns overlying the surgical approach).
Based on this classification, type A4 fractures (burst fracture or sagittal split involving both end plates) and B1–B3 and C injuries almost always require instrumented surgical stabilization for optimal outcome. 11 Vertebral augmentation may be indicated to treat type A injuries, specifically, A1.1, A1.2, A1.3, A2.1, A2.2, A2.3 and A3.1.
The Thoracolumbar Injury Classification and Severity Scale. (TLICS) was introduced in 2005 by the Spine Trauma Study Group (STSG; ▶Table 2.3). It has been validated and shown to have good reliability. It uses a composite scoring system based on the mechanism of injury (as determined by the morphology on imaging), integrity of the posterior ligamentous complex, and neurological status of the patient. 79 Evaluation of the PLC requires MR imaging, a modality that has limited worldwide availability. 80 The TLICS is designed to assist in clinical management of thoracolumbar spine injuries and uses features important in predicting spinal stability, future deformity, and progressive neurologic compromise, as well as predicting appropriate treatment recommendations. 79 , 81 The three major morphologic descriptors are compression or burst, rotation/translation, and distraction. For compression injuries, the prefixes axial, flexion, or lateral more precisely describe the injury morphology.
A score of 3 or less, as with compression injuries, suggests a stable fracture, while a score of 5 or more suggests that surgical instrumentation for stabilization should be considered. This classification system also provides a guide for determining the optimal approach (anterior, posterior, and combined anteroposterior) for surgically treated patients based on the status of the posterior ligamentous complex and the patient’s neurologic status. A score of 4 might be handled conservatively or surgically, according to the physician’s judgment. A1, A2, and A3.1 fractures as classified by the AO system would be assigned 0 to 3 points using the TLICS system and vertebral augmentation would be included in the treatment options.
The systems described above were all designed to identify spinal instability and ultimately to guide treatment. The Genant semiquantitative method of classifying osteoporotic vertebral fractures was introduced in 1993 by Harry Genant, who noted 82 :
“Vertebral fractures are the most common consequence of osteoporosis, occurring in a substantial portion of the post-menopausal population. Most vertebral fractures, however, are not clinically recognized, and can accumulate silently. It is established that the presence of a vertebral fracture is a strong risk factor for subsequent osteoporotic fractures, and that those with low bone density and vertebral fractures are at highest risk.”
Prior to the introduction of the Genant system, osteoporotic vertebral fractures were frequently underdiagnosed worldwide, with false-negative rates as high as 30% despite a strict radiographic protocol that provided an unambiguous vertebral fracture definition and minimized the influence of inadequate film quality. Failure was a global problem attributable to either the lack of radiographic detection or a use of ambiguous terminology in imaging reports. A standardized method for viewing and describing vertebral fractures was needed. 83
The Genant classification system is based on height loss involving the anterior, posterior, and/or middle VB (▶Fig. 2.10). In this semiquantitative assessment, each vertebra receives a severity grade based upon the degree of vertebral height loss. Unlike the other approaches, the type of the deformity (wedge, biconcavity, or compression) is not linked to grading. Thoracic and lumbar vertebrae from T4 to L4 are graded on visual inspection and without direct vertebral measurement as normal (grade 0), mildly deformed (grade 1: reduction of 20–25% of height over 10–20% of the vertebral area), moderately deformed (grade 2: reduction of 26–40% of height over 21–40% of the vertebral area), or severely deformed (grade 3: reduction of >40% of height and vertebral area; grade 0.5 designates “borderline” vertebrae that show some deformation but cannot be clearly assigned as a grade 1 fracture.)
The Genant semiquantitative method has been tested and applied in a number of clinical drug trials and epidemiological studies. 82 , 84 – 87 The reproducibility of the method for the diagnosis of prevalent and incident vertebral fractures was found to be high, with intraobserver agreement of 93 to 99% and interobserver agreement of 90 to 99%. In experienced hands, the approach is both sensitive and specific. 87 A “spinal fracture index” can be calculated from this semiquantitative assessment as the sum of all grades assigned to the vertebrae divided by the number of the evaluated vertebrae. Loss of individual vertebral height can be assessed on serial imaging for a meaningful interpretation of follow-up radiographs. 7