The primary concern for patients with osteoporosis is the increased risk for fractures. Osteoporosis-associated fractures most commonly involve the hip, spine, or wrist. More than 1.5 million osteoporotic fractures occur in the United States yearly.³ It is estimated that the annual incidence of hip fractures in the United States will exceed 6.3 million cases by the year 2050.⁴ Osteoporosis-related fractures can result in significant disability. Only a third of patients regain their premorbid level of function after a hip fracture, and a third require placement in a nursing home within 1 year of the fracture.⁵^–⁷ Twenty percent of patients are no longer living 1 year after a hip fracture.^5,⁸ Besides functional disability and chronic pain, osteoporotic fractures can result in significant anxiety, depression, emotional distress, decreased quality of life, and impaired social well-being.

The increased risk for vertebral compression fractures (VCFs) in patients with osteoporosis is of particular concern to spine care providers. Approximately half of all osteoporotic fractures are spine related.⁹ VCFs are responsible for 150,000 hospital admissions, 161,000 doctors visits, and more than 5 million days of restricted activity annually.¹⁰ It is estimated that 25% of women older than 50 years will suffer a symptomatic VCF during their lifetime.⁹ Although many VCFs are essentially asymptomatic or cause limited symptoms, they can also carry significant morbidity, with chronic pain related to the injury developing in up to a third of patients.¹¹ In addition, VCFs can lead to progressive sagittal-plane deformity with concomitant reduced lung capacity. Kyphosis caused by severe or multilevel fractures can also alter the biomechanical stress at other segments and lead to increased risk for additional fractures.¹²^–¹⁴ This disease is associated with 23% higher mortality in women older than 65 years than in age-matched controls, with additional fractures contributing to increasing mortality.¹⁵

Osteoporosis also places a significant burden on national health care expenditures. The estimated health care cost for osteoporosis-associated fractures was $13.8 billion in 1995 and increased to $17 billion in 2001.^2,¹⁶ This figure includes hospital and nursing home expenses, but the majority of the cost is for inpatient medical care. The projected national health care expenditure for osteoporosis is predicted to rise to $50 billion by the year 2040.²

Pathophysiology

Pathophysiology of Osteoporosis

The National Institutes of Health Consensus Conference defined osteoporosis as a skeletal disorder characterized by compromised bone strength, as reflected in the integration of bone density and bone quality, that predisposes to an increased risk for fracture.¹ Bone density is determined primarily by two factors: one’s peak bone mass and the degree of bone loss throughout one’s lifetime. Bone quality is multifactorial and is dependent on one’s bony architecture, mineralization, ratio of bone formation to resorption, and accumulation of damage. Fractures are defined as mechanical failure as a result of a force or load applied to bone. Pathophysiologically, osteoporosis is a disease of decreased bone mass in the absence of a mineralization defect. Therefore, overall bone mass decreases while the remaining bone maintains normal calcification. Bone loss occurs when the rate of bone resorption is greater than that of new bone formation. With aging, osteoclastic resorption exceeds osteoblastic activity. The net effect is progressive loss of skeletal bone mass, which generally begins by the fourth decade.

Osteoporosis that occurs as a result of aging is termed primary osteoporosis. Primary osteoporosis is characterized by a slow phase and an accelerated phase. In the slow phase, the rate of resorption begins to exceed the rate formation, with 0.3% to 0.5% of bone mass being lost per year. The slow phase occurs in both men and women in older age groups and is hypothesized to be due to impaired vitamin D metabolism by the kidneys. The accelerated phase occurs only in postmenopausal women and is probably related to the loss of estrogen, which results in increased bone resorption. In the accelerated phase, bone mass decreases at a rate of 2% to 3% per year and the decrease continues for up to 10 years. Bone loss generally affects cancellous bone earlier and to a greater extent than cortical bone.

Several factors are protective against primary osteoporosis. Higher peak bone mass attained early in life provides a greater reserve to counteract the eventual increase in bone resorption later in life. Peak bone mass demonstrates linear growth until shortly after adolescence, at which time it consolidates for about 5 to 15 years. A balanced diet with sufficient calories, calcium supplementation, and vitamin D also reduces the risk for the development of osteoporosis. Regular impact loading and resistance exercises contribute to greater peak bone mass. Gonadal hormones are also protective against osteoporosis. Estrogen appears to be particularly important in maintaining bone mass in women. Estrogen suppresses resorption of cancellous bone and maintains balanced turnover between osteoblastic and osteoclastic activity. Administration of estrogen to postmenopausal women decreases bone loss, whereas estrogen deficiency results in increased bone resorption. In adolescents, gonadal hormones contribute to the development of maximal peak bone mass.

Several factors increase the risk for the development of osteoporosis. Late menarche or premature menopause, which results in decreased overall lifetime exposure to estrogen, increases one’s likelihood of becoming osteoporotic. Inadequate dietary intake of calcium through either poor nutrition or eating disorders can also lead to increased bone loss. A sedentary lifestyle or the absence of regular exercise likewise increases the risk for osteoporosis. Tobacco and excessive alcohol use also contribute to bone loss.

Ethnicity appears to be related to risk for the development of osteoporosis. White women have the greatest risk for osteoporosis, and this segment of the population also has a greater risk for vertebral and nonvertebral fractures than African American, Native American, and Asian women.^17,¹⁸ African American women have higher overall bone mineral density (BMD) than white women throughout their lifetime. This discrepancy is reflected in a lower lifetime risk for hip fracture in African American women than in white women (6% for African American women versus 14% for white women).¹

Secondary osteoporosis is a pathologic loss of bone density that occurs in the presence of an underlying medical condition. Common endocrinologic causes include hypogonadal states and hyperthyroidism. Gastrointestinal causes include malabsorptive diseases such as celiac sprue and inflammatory bowel disease. Hematologic disease and bone marrow dysplasia can also result in secondary osteoporosis. Chronic nutritional deficiencies can likewise lead to progressive bone loss. Various medications are associated with secondary osteoporosis, of which corticosteroids are the most commonly encountered. Corticosteroids, often used for immune suppression and treatment of asthma and rheumatoid arthritis, suppress bone formation and inhibit calcium absorption.

Pathophysiology of Osteoporotic Vertebral Fractures

Osteoporotic VCFs occur as a result of structural failure of the spinal column under physiologic loading. The ability of the vertebral body to support compressive forces is dependent on the inherent microarchitecture and material properties of the bone. The vertebral body is predominantly cancellous bone with a surrounding thin shell of dense cortical bone. The majority of axially directed loads are therefore resisted by cancellous bone. The microarchitectural properties of cancellous bone that contribute to structural integrity include trabecular thickness, porosity, and interconnectivity. Material properties that affect bone strength are the degree of mineralization, collagen composition, and lifetime accumulation of damage.

Physiologic loading generates axial compressive forces from activities such as lifting, bending, reaching, or falling. As the vertebral body is loaded, force is transmitted from the intervertebral disk to the end plates and then through the predominantly cancellous bone. In osteoporotic conditions, the cancellous bone suffers the greatest loss in bone density. Compressive mechanical strength is related to density squared, and therefore osteoporosis results in an exponential reduction in resistance to stress. Most commonly, fractures are a result of repetitive loading leading to material fatigue. As osteoporotic vertebrae are compressed, cracks form and grow. Ultimately, the vertebral body fails at a lower load than what would be required to induce failure with a single application of force.

Clinical Findings and Diagnosis

Osteoporosis is a progressive disease that generally becomes clinically relevant only when an individual suffers an osteoporosis-related fracture. Normal activity such as lifting, bending, reaching, or falling with low impact may exceed the loading capacity of osteoporotic bone and lead to failure. Mechanical failure results in an acute fracture, which in the spine is usually manifested as back pain, although subclinical VCFs may also occur. Generally, the pain associated with VCFs dissipates once the fracture heals, but in up to a third of patients the pain becomes chronic. VCFs are rarely associated with neurological compromise. However, multiple fractures can lead to progressive kyphosis. With global sagittal imbalance, patients may experience decreased functionality, impaired social well-being, and in severe cases, decreased pulmonary function.

The initial evaluation of a patient suspected of having osteoporosis includes BMD testing, which is repeated serially if pharmacologic treatment is initiated. Younger patients with severe osteopenia may warrant additional hematologic studies or a bone marrow biopsy.

Bone Mineral Density

The current standard for measurement of BMD is dual-energy x-ray absorptiometry (DEXA). DEXA scans evaluate BMD at clinically relevant locations such as the hip and spine, which are particularly prone to osteoporosis-related fracture. DEXA is a favorable screening tool because it exposes the patient to a low dose of radiation (90% less than with a standard chest radiograph). Standard spine DEXA values are obtained by scanning the lumbar spine in the posteroanterior plane from L1 to L4, with a total value for the four sites combined also being reported. BMD is expressed as the bone mineral content (grams) divided by the area (centimeters squared). Spine BMD may be falsely elevated in older patients with excessive osteophytes and calcification of disks or surrounding vasculature. BMD may also appear abnormally low in patients who have previously undergone laminectomy. BMD assessment of the hip includes scanning of the femoral neck, greater trochanter, intertrochanteric region, and total femur.

A patient’s BMD is compared with a Z score (BMD for healthy gender- and age-matched controls) and a T score (BMD for normal healthy young controls at peak bone mass). The WHO defines a BMD of less than 1 SD below the T score as being within normal limits (osteopenia). Osteoporosis is defined as a BMD greater than 2.5 SD below the T score. Severe osteoporosis is a BMD that is greater than 2.5 SD below the T score with at least one osteoporosis-related fracture. Based on the WHO classification, it is estimated that 94% of women older than 75 years meet the criteria for osteopenia, with 38% of women in this age group having osteoporosis.¹⁹ BMD values also correlate with fracture risk. Each 1-SD decrease in age-adjusted BMD measurement equates to a 1.5-fold increase in fracture risk.

Additional methods for assessing BMD include plain film densitometry or simple measurement of the cortex of the metacarpal or other tubular bones on standard radiography. Single-photon absorptiometry measures differences in photon absorption between bone and soft tissue. For ultrasound assessment, two transducers are placed opposite each other at the calcaneus. BMD can be determined by measuring broadband attenuation of the ultrasound beam through the calcaneus. Ultrasonography has proved to be effective in predicting risk for certain osteoporosis-related fractures. Computed tomography (CT) is becoming a useful instrument for assessing BMD. CT scanning at the midpoint of the vertebra can measure the BMD of cancellous and cortical bone, as well as the entire vertebral body. Therefore, both three-dimensional volumetric analysis and selective assessment of cancellous bone are possible with CT imaging. Disadvantages of the use of CT include the necessity for special calibration of the equipment before scanning, as well as variability in results with minor alterations in the location of measurement.

Biochemical Markers

Biochemical markers for bone formation and bone resorption can be quantified as a means of determining the actual rate of bone metabolism. Enzymes and proteins synthesized by osteoblasts and osteoclasts or osteoclast-induced degradation products are measured in serum or urine to assess bone turnover. These markers for bone remodeling can be used to determine risk for the development of osteoporosis. Biochemical markers are also used to monitor disease progression and responsiveness to therapy.

Markers of bone formation are noncollagenous proteins produced by osteoblasts, which are elevated in serum with increased activity. Bone-specific alkaline phosphatase is a product of osteoblasts and osteoblast precursors and is quantified with the use of monoclonal antibodies. Osteocalcin is a 49-residue polypeptide that is associated with 1,25-dihydroxyvitamin D₃ activity and reflects osteoblastic synthesis and deposition of new bone. Additional markers of bone formation include type I collagen propeptides. Type I collagen is the major product of osteoblasts, and immunoassays for type I collagen propeptides provide a nonselective index of total bone formation activity in the body.

Markers of bone resorption reflect increased osteoclastic activity. Such markers include products secreted by osteoclasts, as well as collagen breakdown products. Biochemical markers of collagen degradation are the C- and N-terminal telopeptides of type I collagen cross-links (CTx and NTx). CTx and NTx are measured in urine and are the most sensitive indices of collagen breakdown, with NTx being most commonly referenced. Pyridinoline and deoxypyridinoline are collagen cross-linking amino acids that are excreted in urine when collagen is degraded. Pyridinoline and deoxypyridinoline are quantified with high-performance liquid chromatography. Bone is the predominant reservoir of type I collagen in the body and is therefore the major source of pyridinoline in urine. Collagen cross-links such as deoxypyridinoline are not unique to bone, however, and thus urine concentrations of these markers may not accurately reflect bone turnover. Serum markers for the osteoclast-specific isoform of tartrate-resistant acid phosphatase reflect activity of this lysosomal enzyme in bone resorption. Hydroxyproline is another urinary marker of bone resorption, but its utility is limited by its lack of specificity for bone and variability attributable to differences in diet and degradative losses in the liver.

Biochemical markers are primarily useful for determining risk for the development of osteoporosis and for assessing responsiveness to therapy. Serum markers are generally less variable than urinary markers, which need to be corrected based on creatinine clearance. Studies have demonstrated that levels of bone-specific alkaline phosphatase, osteocalcin, and NTx are higher in postmenopausal women than in premenopausal women.²⁰ With bisphosphonate alendronate therapy, however, alkaline phosphatase, osteocalcin, and CTx levels have been shown to decrease 40% to 50% over a period of 6 to 12 months.²¹ NTx was the most responsive marker for measuring therapeutic response. Additional studies have suggested that NTx and CTx correlate significantly with BMD and risk for fracture.²⁰

Conservative and Medical Management

Preventive measures remain among the most important and effective strategies for managing osteoporosis. Adequate nutrition with an appropriate balance of calcium and vitamin D is essential for optimizing bone quality. Calcium supplementation in the form of calcium carbonate or calcium citrate is primarily effective for postmenopausal women. Vitamin D supplementation is also beneficial, with one study demonstrating that 1200 mg of calcium and 600 to 800 IU of vitamin D result in a 40% decrease in hip fractures and a 16% decrease in mortality.²² Regular weight-bearing, impact exercise increases peak bone mass and thereby reduces the risk for osteoporosis. Wolff’s law states that bone forms by appositional growth in areas of increased stress. With impact loading, differences in electronegative potential occur across compressed surfaces, which subsequently stimulates bone formation. Subjects randomized to aerobics and weight-training programs demonstrate a 5.2% increase in spine density over subjects treated only with calcium supplementation.²³ Strength training also increases bone density in both the spine and hips, whereas immobilization decreases overall bone mass. An active exercise program of jogging and stair climbing in postmenopausal women receiving calcium supplementation resulted in a 5.2% increase in BMD at 9 months.²³ The control group of patients treated with just calcium supplementation experienced a 1.4% loss in BMD. The average BMD in smokers is 1% to 3% lower than that in nonsmokers, with the number of pack-years being inversely correlated with BMD. Although the mechanism is unclear, tobacco use may alter the local acidic environment to facilitate osteoclastic breakdown of hydroxyapatite. Chronic alcohol use also results in increased bone loss. Low vitamin D from malnutrition and decreased activation as a result of chronic liver disease may explain the effect of long-standing alcohol consumption on bone loss.

Bisphosphonates are the first-line pharmacologic agent for osteoporosis. They are analogues of pyrophosphate and function to inhibit osteoclastic activity. Early-generation bisphosphonates, such as etidronate and clodronate, are nonselective and inhibit both bone formation and resorption equally. Second-generation drugs (pamidronate, alendronate) have more selective antiresorptive activity and demonstrate a 50% reduction in spinal and hip fractures.²⁴ Alendronate therapy is associated with a 5% to 7% increase in spinal bone mass at 2 years. Risedronate and zoledronate are third-generation bisphosphonates that preferentially function at sites of active bone resorption.

Estrogen hormonal replacement therapy (HRT) significantly increases BMD in postmenopausal women. However, estrogen HRT is associated with a significant risk for breast cancer, stroke, and deep venous thrombosis, which was found to outweigh its benefit in the treatment of osteoporosis. Estrogen-alone HRT (without progestin) is also known to increase the incidence of endometrial cancer. Selective estrogen receptor modulators (SERMs) are agents that preserve the beneficial effects of estrogen on bone metabolism while having antiestrogenic effects on breast and endometrial tissue. Raloxifene, a SERM, significantly decreases bone resorption. In particular, raloxifene in combination with the bisphosphonate alendronate has proved to be more effective in improving lumbar BMD than either agent alone.²⁵

Several endogenous hormones involved in calcium homeostasis can be administered exogenously to improve bone mass. PTH secreted by the parathyroid gland primarily regulates calcium and phosphate balance. Daily administration of PTH increases bone mass, as well as cortical and cancellous bone thickness. Teriparatide is a PTH agent approved by the Food and Drug Administration (FDA) for use in osteoporotic patients with high fracture risk. Teriparatide, however, is contraindicated in patients with osteosarcoma, Paget’s disease, or previous skeletal radiation therapy. Calcitonin is secreted by thyroid parafollicular cells and increases calcium stores in bone by inhibiting osteoclast activity. Calcitonin is administered intranasally and is an effective agent for painful VCFs.

Surgical Treatment of Osteoporotic Vertebral Compression Fractures

VCFs result from osteoporosis-related loss of bone density with subsequent failure of the anterior column under mechanical loading. Generally, VCFs are stable injuries that respond to conservative treatment consisting of bed rest, pain management, and external immobilization. Prolonged bed rest, however, can increase loss of bone density, result in muscle atrophy, and impair the functional outcome. Use of narcotics to alleviate pain can alter mood and cognitive function, which may further compound any existing medical and age-related conditions. Alternatively, surgical intervention consisting of spinal reconstruction and instrumented stabilization is limited in this population because of poor bone stock and generally high surgical risk stratification.

Percutaneous vertebral cement augmentation has emerged as a minimally invasive technique for stabilization of painful VCFs. Both vertebroplasty and kyphoplasty are procedures with relatively low morbidity that have demonstrated significant beneficial clinical results. The main indication for vertebroplasty or kyphoplasty is chronically painful VCFs in patients who have failed 4 to 6 weeks of conservative therapy. The primary function of vertebroplasty and kyphoplasty is to decrease pain and improve mobility and function. Contraindications to these procedures include fractures with disruption of the posterior vertebral wall, neurological deficit, or complete collapse of the vertebral body.

Vertebroplasty, first described in 1987, involves the forced injection of polymethyl methacrylate (PMMA) into the fractured vertebral body. A Jamshidi needle is inserted percutaneously under biplanar fluoroscopic guidance via either a transpedicular or extrapedicular route into the affected level. Contrast medium is then injected through the needle to ensure that the tip is not within a vessel. Low-viscosity PMMA is forced through the needle at considerable pressure to allow the cement to infiltrate the fissures created by the fracture. The volume of injection is limited by the potential for extravertebral extravasation of cement.

Kyphoplasty is similar to vertebroplasty in that a cannula is introduced percutaneously into the fractured vertebral body under fluoroscopic guidance. An inflatable bone tamp is then placed through the cannula and, on inflation, reduces the fractured fragments, thereby creating a cavity and restoring vertebral body height. The bone tamp is subsequently removed, and PMMA is injected into the cavity. Higher viscosity PMMA can be injected with kyphoplasty than with vertebroplasty because the cement is being introduced into a preformed cavity.

Currently, PMMA is the primary bone cement used for both vertebroplasty and kyphoplasty. Alleviation of pain after PMMA augmentation is probably due to mechanical stabilization of the fractured vertebral body. PMMA also generates heat during exothermic polymerization. This process may cause thermal necrosis of pain receptors within bone and a resultant decrease in sensitivity. PMMA is also known to be chemotoxic and may cause inflammation-induced injury to nerve endings with decreased transmission of pain sensation.

Clinical studies of both vertebroplasty and kyphoplasty demonstrate significant beneficial results. Vertebroplasty is associated with a 70% to 90% success rate in relieving pain.²⁶^–²⁹ Complications from vertebroplasty are reported in less than 10% of patients and include increased pain, radiculopathy, spinal cord compression, pulmonary embolism, infection, and rib fracture. The risk for extravasation of cement after vertebroplasty is widely variable and estimated to be 30% to 67%.³⁰^–³³ Kyphoplasty demonstrates similar positive results, with rates of pain relief ranging from 88% to 100%.^28,³⁴^–³⁷ In an ongoing multicenter study from 1998 to 2000, 603 VCFs in 304 patients were treated by kyphoplasty.²⁸ At an average 18 months’ follow-up, 90% of the patients had symptomatic and functional improvement.

Kyphoplasty is primarily differentiated from vertebroplasty in its potential for restoring vertebral body height and improving overall sagittal alignment. The average restoration of vertebral height with kyphoplasty is reported to be 30% to 35%.^36,³⁸ In the aforementioned multicenter study of 603 VCFs treated by kyphoplasty, of fractures with a 15% or greater estimated loss of vertebral height, treatment improved vertebral body height from an average of 68% of predicted height to 84% of predicted height after treatment.²⁸ Other studies counter that although kyphoplasty may improve local kyphosis, the overall effect on global sagittal alignment is minimal. In a retrospective study of 65 patients undergoing one- to three-level kyphoplasty, Pradhan and colleagues found that kyphoplasty improved the local deformity at the fracture level by an average of 7.3 degrees, or 63% of the preoperative kyphosis.³⁹ They found, however, that the overall angular correction decreased to 2.4 degrees, or 20% of the preoperative kyphosis, when measured one level above and below the fractured level. At three levels above and below the fracture level, the correction decreased to just 1.0 degrees, or 8% of the preoperative kyphosis. The researchers surmised that the majority of the local angular and height correction becomes negated at more distant levels by the relatively softer intervertebral disks. As a result, the radiographic improvement in overall sagittal alignment with kyphoplasty is modest at best. A meta-analysis of the literature revealed that both kyphoplasty and vertebroplasty result in a significant improvement in visual analog pain scores, with an average decrease of 5.68 for vertebroplasty and 4.60 for kyphoplasty.⁴⁰ The risk for extravasation of cement with vertebroplasty was 19.7% but just 7.0% with kyphoplasty. The significant difference in risk for leakage of cement is hypothesized to be due to the injection of higher viscosity cement into the preformed cavity with kyphoplasty.

Patients with osteoporosis-related VCFs are at risk for the development of additional fractures at other spinal levels. Without surgical intervention, additional VCFs will develop in 19.2% of patients within 1 year.^41,⁴² Because both vertebroplasty and kyphoplasty involve relatively rigid cement, the altered biomechanics of the augmented vertebra may create greater stress at weakened adjacent segments. Studies have demonstrated that the relative risk for the development of additional VCFs at a level adjacent to a previous vertebroplasty is 4.62 times greater than that at nonadjacent levels.⁴³ In a meta-analysis of the literature on vertebroplasty and kyphoplasty, the risk for fracture at an adjacent level was 17.9% after vertebroplasty and 14.1% after kyphoplasty.⁴⁰ Although both vertebroplasty and kyphoplasty result in improved clinical outcomes for patients with chronically painful VCFs, the future of vertebral augmentation will probably involve the use of osteobiologic agents and biocompatible cement.

Osteoporosis and Implications for Spinal Instrumentation

With increases in life expectancy, more patients of advanced age are undergoing spine surgery for a variety of indications, including degenerative, deforming, traumatic, rheumatologic, infectious, and oncologic disease. Spine surgery in this population frequently involves instrumented stabilization and reconstruction. However, the use of rigid spinal fixation in the setting of osteoporosis can pose significant technical challenges.⁴⁴

Rigid instrumentation places mechanical demands on the implant-bone interface. As bone quality weakens, the risk of loosening at the implant-bone interface increases under cyclic mechanical loading. Hardware failure in osteoporotic patients may also occur when increased biomechanical demands are placed on the instrumentation, such as when performing spinal corrective maneuvers. Reduction techniques for deformity, spondylolisthesis, and fractures in patients with low BMD increase stress at the screw-bone interface. Ultimately, screw pullout may lead to loss of correction or progression of deformity. Long-segment rigid constructs also alter the biomechanical environment of the segments adjacent to the instrumentation. This may increase the risk for junctional kyphosis at levels proximal to the termination of the construct. Interbody devices for anterior reconstruction may subside through weakened vertebral end plates, which can result in collapse and failure of the anterior column.

Biomechanical studies have demonstrated that screw pullout strength is directly related to BMD.⁴⁵^–⁴⁷ As BMD decreases, so does the force required for axial pullout of pedicle screws. Pedicle screws have been compared with other spinal implants such as laminar hooks and sublaminar wires in the setting of low BMD. Coe and coworkers found that laminar hooks demonstrated better resistance to failure with a posteriorly directed force than did either pedicle screws or wiring. Unlike pedicle screws or wires, the load to failure for laminar hooks did not correlate with the measured BMD.⁴⁸ They observed that pedicle screws failed by stripping the cancellous bone within the pedicle track, consistent with the predominant effect of osteoporosis on cancellous bone. Hooks, however, required failure of the inner cortical bone of the lamina, which is relatively spared by osteoporosis.

Strategies for Spinal Fixation in Osteoporosis

Screw Placement

Several techniques can be used to reduce the risk for screw failure in patients with osteoporotic bone. Longer screws achieve better fixation in low-density bone than shorter screws do. Hackenberg and associates demonstrated that a screw length of 50 mm had significantly greater pullout strength than did 35-mm screws.⁴⁵ Similarly, increasing the screw diameter also improves resistance to failure.⁴⁷ However, the benefit of a larger screw diameter in settings of low BMD was observed only in patients with milder osteoporosis. In severe osteoporosis, the pullout force was low regardless of screw diameter. Pullout resistance for pedicle screws relies largely on the volume of cancellous bone engaged between the threads of the screw. With osteoporosis, however, the cancellous bone is less dense, thereby providing suboptimal fixation. Placing pedicle screws so that they achieve purchase in cortical bone increases their resistance to failure. Bicortical pedicle screws provide up to an additional 30% pullout strength in comparison to unicortical screws by engaging the ventral cortex of the vertebral body.⁴⁹

Triangulation of bilateral pedicle screws is another technique for increasing the load to failure.^50,⁵¹ Ruland and colleagues found that triangulated screws connected by a transverse plate provide better pullout resistance than do laminar hooks or a single pedicle screw.⁵¹ With parallel screw placement, the resistance to pullout for each screw is dependent on the volume of cancellous bone between the threads of the screw. With screw triangulation and a transverse connector, the pullout strength of the construct is contributed by the volume of bone within the trapezoid area in the vertebral body formed by the triangulated screws. With a larger volume of cancellous bone available for resistance to pullout, triangulated screws provide up to twice the pullout strength of a single pedicle screw.

Appropriate pilot hole preparation before screw placement is another technique for improving screw purchase. Tapping or drilling the pilot hole results in removal of bone within the pedicle track. Even screw insertion, removal, and subsequent reinsertion of the same screw decrease the mechanical insertion torque by more than 34%.⁵² Carmouche and coworkers found that tapping the pilot hole with the same-diameter drill bit as the screw significantly decreases pullout resistance in comparison to not tapping or undertapping.⁵³ They observed no significant difference in pullout resistance between undertapping and not tapping the pilot hole. When compared with same-size tapping, undersizing the tap by 0.5 and 1.0 mm increases insertional torque by 47% and 93%, respectively.⁵⁴ Interestingly, undertapping demonstrates less benefit in the thoracic region than in the lumbar spine.⁵³ The authors attributed this observation to the less overall available cancellous bone within the smaller thoracic pedicle in the setting of osteoporosis. Therefore, the size of the tap had little effect, and they surmised that thoracic pedicle screws are probably more dependent on cortical purchase with the pedicle walls.

Screw Augmentation with Hooks

Augmentation of screw-rod constructs with hooks combines the optimal three-dimensional control of pedicle screws with the improved pullout resistance of hooks. In a study of construct stiffness obtained from load deformation curves, pedicle screws with a laminar hook demonstrated significantly more stiffness than a pedicle screw alone.⁵⁵ Halvorson and colleagues found that a pedicle screw with offset hooks at two adjacent levels improves fixation and doubles the expected pullout force.⁴⁶ A hook at the inferior laminar edge of the same level and a second hook one or two levels above, when connected to a rod, function as a three-point load-sharing device, even as the screw begins to displace. This particular construct in osteoporotic bone results in resistance to failure that is equal to or greater than that of a single pedicle screw in normal bone. A pedicle screw with a single laminar hook at the same level, however, does not increase the load to failure as compared with the expected value. They observed that as the screw began to displace, the hook functions as a pivot and fails to carry any load.

Screw Augmentation with Cement

Cement augmentation of the screw track into the pedicle and vertebral body additionally increases fixation in osteoporotic bone (Fig. 271-1). Zindrick and coworkers demonstrated that even stripped pedicle screws regain within 5% of their baseline pullout strength in cadavers when PMMA is injected into the screw track.⁴⁹ Sarzier and associates observed that PMMA injection resulted in increased pullout resistance in osteoporotic cadavers with a BMD at least 2 SD below normal.⁵⁶ The increase in cancellous bone density with PMMA often led to an increase in pullout strength that exceeded the strength of the cortical bone. As a result, loading to failure caused a fracture of the vertebral body at the pedicle-body junction.

FIGURE 271-1 Lateral intraoperative fluoroscopic image demonstrating augmentation of lumbar pedicle screws with polymethyl methacrylate to increase pullout strength.

Screw augmentation with cement carries the risk of extravertebral or intracanal leakage of cement. The injection needle is generally inserted to a depth that is at least 5 mm less than the intended screw length so that accidental violation of the ventral vertebral cortex is avoided. Higher viscosity cement is used to reduce the likelihood of cement leakage. One to 3 mL of cement is generally recommended, with an increase in cement injection failing to demonstrate any significant benefit in pullout strength.⁵⁷ Large-volume cement injection is avoided to prevent excess extravasation of cement.

Burval and coworkers found that performing a kyphoplasty technique with an inflatable bone tamp before cement augmentation further improves pullout resistance.⁵⁸ Similar to conventional kyphoplasty for osteoporotic vertebral fractures, an inflatable bone tamp is used to create a cavity within the vertebral body. High-viscosity cement is injected into the cavity, followed by the insertion of pedicle screws. Pedicle screws with kyphoplasty cement augmentation demonstrate almost twice the pullout strength of screws augmented with standard cement injection and 255% better fixation than noted with unaugmented screws.

Fenestrated taps and screws facilitate cement injection while reducing the risk for retrograde migration of cement out of the pedicle track. Frankel and colleagues found that injecting PMMA through a fenestrated tap before screw placement increased pullout strength by 119% in primary procedures and by 162% in salvage procedures.⁵⁷ McKoy and An contend that PMMA functions as primarily a bone void filler and not as a true adhesive.⁵⁹ Therefore, injecting cement after preparation of the pilot hole only fills the void of the pedicle track. Subsequent insertion of a pedicle screw into the doughy curing cement simply coats the screw threads, thereby effectively reducing screw purchase. Alternatively, injecting cement through a cannulated fenestrated screw after insertion of the screw allows the cement to infiltrate the bone and ensures that the cement remains within the vertebral body.

Osteobiologic cement is an area of interest and development for screw augmentation. PMMA is brittle, toxic, permanent, and difficult to remove in situations requiring revision surgery. In addition, PMMA exhibits a high exothermic polymerization temperature that can induce thermal injury in surrounding structures. Alternatively, calcium phosphate is an osteobiologic agent that is bioresorbable and therefore becomes integrated in the natural process of bony remodeling. Renner and associates demonstrated that injection of 3 mL of calcium phosphate significantly improves pedicle screw pullout strength in comparison to unaugmented screws.⁶⁰ In a revision model, screw augmentation with calcium phosphate resulted in pullout strength that was similar to that of unaugmented screws placed for the first time. In comparing PMMA and calcium phosphate for screw augmentation, Rohmiller and colleagues observed that calcium phosphate augmentation results in a 167% improvement in screw pullout strength in comparison to unaugmented screws.⁶¹ PMMA demonstrated a 199% increase in pullout strength over unaugmented screws; however, this difference was not statistically significant when compared with calcium phosphate augmentation.

Additional Techniques: Multiple Points of Fixation, Appropriate Release, Anterior Reconstruction

Because of their optimal three-dimensional control, pedicle screws are often used in patients with conditions that place increased stress on the screw-bone interface. Correction of deformity, fracture reduction, and complex spinal reconstruction are indications for pedicle screw fixation associated with a greater need for both screw and bone integrity. In the setting of low BMD, weakened cancellous bone may lower the threshold for screw failure when using pedicle screws for spinal correction or load bearing.

Several strategies exist for decreasing the risk for screw failure when using pedicle screws in osteoporotic bone and conditions of increased stress. Extending the number of segments included in the screw-rod construct distributes the loading forces across multiple fixation points (Fig. 271-2). Increasing the number of points of fixation decreases the stress applied to each individual screw and consequently minimizes the risk for pullout failure at each segment.⁶² This principle is relevant when performing procedures to correct deformity, such as rod derotation maneuvers. With rod derotation, forces are applied directly to the screws by the rotating bent rod to achieve spinal correction. Particularly in situations of anterior column failure such as a vertebral body fracture, pedicle screws are exposed to large cantilever bending loads, which may result in screw breakage or pullout at the distal ends of the construct.⁶²

FIGURE 271-2 A, Preoperative anteroposterior radiograph demonstrating a degenerative coronal lumbar deformity in an elderly patient with osteoporosis. B, Postoperative anteroposterior radiograph demonstrating surgical correction of the deformity with segmental pedicle screw-rod instrumentation. Multiple points of fixation were used to both reduce the deformity and decrease stress at the implant-bone interface for each screw.

When performing spinal corrective procedures, optimizing release of discoligamentous and bony constraints before reduction can decrease the stress applied to the screw-bone interface. Techniques for release such as discectomy, facetectomy, or various osteotomies increase the flexibility of the spinal column. With better spinal mobility, less stress needs to be applied to the instrumentation to achieve and maintain correction.

Anterior column support also decreases the biomechanical loading of pedicle screws in weakened bone. With reconstruction of the anterior column, the load is shared by the graft or cage and less stress is directed toward the pedicle screw–rod construct. However, in osteoporosis, subsidence of the graft or cage into weakened vertebral end plates can lead to collapse of the anterior column, kyphosis, and deformity. Particularly in osteoporotic bone, cage placement should ideally contact the peripheral apophyseal ring, where the stronger cortical bone is more supportive of compressive loads than the weaker central portion. Increasing the diameter of the cage and ensuring at least 30% coverage of the vertebral body maximize the cage-bone contact area and optimize anterior column support.^63,⁶⁴

Novel Screw Designs: Expandable Screws, Hollow Monaxial Screws

Novel screw designs have been developed to improve fixation and pullout resistance in osteoporotic bone. Expandable pedicle screws are designed such that the distal part of the screw enlarges within the vertebral body to resist pullout failure. As a posteriorly directed force is applied to the screw, the flared tip becomes anchored against the inner cortex of the dorsal vertebral body. The integrity of the expandable screw-bone interface is augmented by the relatively uncompromised cortical bone rather than depending solely on weakened osteoporotic cancellous bone.

McKoy and An studied an expandable screw in cadaveric mechanical testing.⁶⁵ The screw design involved a cannulated screw that is placed transpedicularly into the vertebral body. Placement of a smaller inner screw down the cannulated center causes flanges at the distal part of the screw to flare outward and expand within the vertebral body. The authors found that the expandable screw resulted in a 76% increase in holding strength in comparison to conventional pedicle screws. Other studies have demonstrated that expandable pedicle screws markedly improve pullout strength up to 50% in low-BMD bone versus standard pedicle screws.⁶⁶ However, in patients with severely low BMD, expandable screws are unable to overcome the extreme biomechanical disadvantage and result in failure. Augmentation of expandable screws with PMMA injected through the cannulated portion of the screw increases mean axial pullout resistance by 200% in severely osteoporotic bone.⁶⁷ Screw revision, however, remains an issue in the clinical application of both expandable and injectable screws.

Anterior thoracolumbar screw designs have been explored to improve fixation in osteoporotic bone. Continuous cyclic loading of anterior screw constructs in porous, brittle bone can lead to screw cutout. Novel screw designs incorporate increased surface area for the screw-bone interface to improve the load-bearing cross-sectional area. Hollow monaxial screws are designed either with a cylindrical spiral blade or as a hollow-perforated cylinder (Fig. 271-3). This allows increased screw-bone contact, as well as promotes ingrowth of bone within the screw. Various hollow monaxial screws have demonstrated promising results in comparison to conventional screws in biomechanical testing on low-BMD specimens.⁶⁸^–⁷⁰

FIGURE 271-3 A, Preoperative sagittal magnetic resonance image demonstrating a midthoracic compression fracture resulting in a sharp angular kyphosis. B, Postoperative anteroposterior radiograph demonstrating a single-level corpectomy with anterior cage placement and correction of the deformity. Instrumented stabilization was achieved with hollow monaxial screws connected to a rod, which provides greater mechanical stability than possible with traditional cancellous bone screws.

Semirigid Fixation

Implants composed of stainless steel and titanium result in a supraphysiologic degree of stiffness.⁷¹ As the spine is loaded in various axes of motion, the rigid instrumentation restricts mobility at the instrumented levels and creates increased stress at the screw-bone interface. The lack of flexibility with nonmalleable instrumentation can also accelerate degeneration at adjacent motion segments, shield bone grafts from stress, and therefore prevent arthrodesis.

Semirigid fixation allows spinal stabilization while permitting some degree of flexibility. Polyetheretherketone (PEEK) has a modulus of elasticity between that of cortical and cancellous bone, thus theoretically mimicking the modulus of elasticity of the native environment. With PEEK rod instrumentation, some degree of motion is permitted while resisting marked flexion, extension, axial rotation, or lateral bending. As a result, semirigid fixation may provide sufficient stabilization to facilitate bony fusion while permitting enough flexibility to offload stress at the screw-bone interface or adjacent segments (Fig. 271-4).