Changes in Disc Structure with Aging and Degeneration

Changes that almost universally occur in the IVD with aging include reduction in disc volume, shape, and content. Alterations in gene expression and transcription factors may be responsible for cell senescence within the disc. These senescent cells lose biochemical and synthetic capabilities, which ultimately diminishes the ability of the disc to recover from deformation and renders the matrix more vulnerable to progressive fatigue failure. The nucleus pulposus gradually becomes less hydrated, and usually by the 3rd decade of life there is already a significant decline in the number of viable cells and a loss of proteoglycans.

The early degenerative process affects the nucleus pulposus and the endplate more than the anulus fibrosus. Anabolic and catabolic processes are upregulated during the early stages of degeneration; however, anabolic repair processes fail to keep up with the catabolic processes, and matrix degeneration ensues over time. As the process progresses, the inner layers of the anulus and the nucleus pulposus gradually become indistinguishable and change into a stiff desiccated fibrocartilaginous material.

The number of arterioles supplying the peripheral disc diminishes significantly as remaining blood vessels are obliterated by calcification of the cartilaginous endplates. Loss of endplate vascularity and porosity leads to a reduction in the influx of nutrients and efflux of waste products. Lactate levels increase locally within the hypovascular disc secondary to increased production and decreased removal. Cell apoptosis occurs as a result of decreased tissue pH,¹⁸ and the biosynthetic reparative capability of the disc is impaired further.

Thinning or microfracture of the endplate alters its properties and may allow rapid outflow of fluid from the cartilage endplate on loading, rendering the hydrostatic pressure mechanisms involved in load transference less effective and uniform. Focal elevations in shear stresses at the disc level may adversely affect the disc structure further and result in annular damage. Over time, cracks develop between and through the annular lamellae. Eventually, communicating channels may develop between the peripheral layers of the anulus and the nucleus, and disc material can herniate through these fissures. The weakened anulus can develop a full-thickness defect and allow near-complete herniation of the nucleus pulposus, particularly when the disc is loaded in flexion and torsion.

The degenerative process resulting from matrix changes and internal structural disruption sets the stage for abnormal motion at the degenerated segment. Changes in disc structure alter the loading response and alignment of the spinal column. These changes can influence the facet joints, ligaments, and paraspinal muscles, which may also become pain generators. Pain does not always correlate with morphologic changes in the disc and mechanical compression, however.¹⁹ Macnab²⁰ described traction osteophytes around the vertebrae originating 2 mm from the anterior endplate, at the site of attachment of the outermost annular fibers. These osteophytes were thought to be signs of abnormal biomechanics, caused by traction at the insertion of the annular fibers into the vertebral bodies. Subsequent studies found these osteophytes to be inconsistently present.

Multiple factors lead to disc degeneration, including insufficient nutritional supply, reduction in the amount of viable cells, degradative enzymatic activity, and cell senescence and apoptosis. Alteration in loading patterns between the endplate and disc leads to annular damage and the potential for herniation. Perturbation of the disc also leads to degeneration and pain in other segmental structures such as the facet joints, ligaments, and paraspinal muscles. The initiating event leading to the onset of degeneration is unknown.

Associated Factors

Various risk factors have been implicated in the pathogenesis of lumbar disc degeneration. In a review of factors associated with IVD degeneration in elderly adults, Hangai and colleagues²¹ cited increased age, high body mass index, occupational lifting, sporting activities, and factors associated with atherosclerosis as risk factors. Multiple studies show the genetic contribution to degenerative low back pain.²² Battie and colleagues²³ estimated the familial contribution to IVD degeneration to be 34% to 61%. Cigarette smoking has also been implicated and seems to have an adverse vasoconstrictive and atherosclerotic effect on the nutrition of the IVD.^24,25 Type of occupation has also been shown to have an adverse effect on lumbar spinal segment degeneration, increasing the risk of symptomatic DDD. Studies have implicated occupations that require repetitive lifting or pulling, prolonged sitting²⁶ such as motor vehicle driving,²⁷ and whole-body vibration.²⁸

Arun and colleagues²⁹ used serial postcontrast MRI to study the effect of prolonged mechanical load on diffusion into the IVD. The authors reported that 4.5 hours at a load corresponding to 50% body weight significantly retarded the diffusion of small solutes into the center of the IVD, and it required 3 hours in an unloaded recovery phase to return the diffusion rate to that seen in the unloaded disc. Prolonged mechanical load can cause a disruption of diffusion, which may accelerate disc degeneration; however, this hypothesis has not been confirmed clinically.

The genetic predisposition to lumbar DDD and lifetime exposures were studied in a classic monozygotic twin study by Battie and colleagues.²³ These investigators reviewed 115 male identical twin pairs for exposures to common risk factors such as occupation, recreational activities, driving, and smoking. Disc degeneration was determined by MRI and clinical evaluation. In the upper lumbar spine, only 7% of the variability was explained by occupation, 16% was explained by age, and 77% was explained by familial aggregation. In the lower lumbar spine, recreational physical loading explained 2% of variability, age explained 9%, and familial aggregation explained 43%. Battie and colleagues²³ concluded that primarily genetic and other unexplained factors result in DDD, whereas commonly implicated environmental factors have only modest effects. In a 5-year follow-up study of the same twin population, the same investigators reaffirmed that genetics have a dominant role in progression of DDD, whereas occupational lifting and leisure activity had only modest effects.³⁰ The important role of genetic factors has been corroborated in other twin studies,^31,32 but it seems to be less of an explanatory factor for back pain in older people.³³

Several gene loci have been discovered that are associated with increased risk for DDD. Type IX collagen was one of the first gene loci identified with some aberrant alleles imparting a threefold or fourfold increase in relative risk.^34–36 More recent publications also implicate collagen type XI, interleukin (IL)-1, aggrecan, vitamin D receptor, matrix metalloproteinase (MMP)-3, and cartilage intermediate layer protein (CILP) as candidate genes.³⁷ The discovery of these genetic risk factors has yet to result in new useful diagnostic and treatment modalities, however.

Internal Disc Disruption

Crock¹⁹ coined the term internal disc disruption in 1970 and defined it as a painful increase in biologic activity of the IVD after injury with normal radiographic, computed tomography (CT), and myelogram examinations but an abnormal discogram. IDD as a cause of discogenic back pain is controversial. The advent of MRI has dramatically improved the detection of this entity—IDD manifests as a dark disc with relatively preserved height and contour on MRI. Pain in IDD is thought to be caused by mechanical and chemical stimulation of nociceptors within the anulus or on the surface layers of the anulus and the overlying ligamentous tissue. The hallmark of IDD is the absence of disc herniation, prolapsed disc material, segmental instability, or other radiographic abnormality.^19,38 Nerve root irritation, radicular pain, and neurologic deficits are also absent.

Radiographic changes associated with DDD—significant disc space narrowing, endplate osteophyte formation, endplate sclerosis, and gas formation within the disc space—are not seen in IDD.³⁹ MRI (dark disc) and positive discography (concordant pain in the abnormal level and not at normal adjacent levels) are required to make the diagnosis of IDD. Because of the poor sensitivity and specificity of discography, many clinicians question the existence of IDD as a clinical entity.

Degenerative Disc Disease

The relationship between degenerative disc changes and low back pain is poorly understood. Two potential sources that have been implicated as contributors to pain in patients with disc degeneration are (1) sensitization of nerve endings by release of chemical mediators and (2) neurovascular ingrowth into the degenerated disc.

The precise pathophysiologic mechanism for chemically mediated induction of hyperalgesia within the disc has yet to be fully elucidated. Radial annular tears provide a route for nuclear material and noxious chemicals to leak from the disc and contact the dural sac and nerve roots; some studies have shown that autologous nucleus pulposus alone has the capacity to produce an inflammatory response. Additionally, degradative changes can occur within nerve roots exposed to nuclear material even in the absence of mechanical compression.^38,40–42 Weinstein and colleagues⁴³ investigated the reproduction of pain on discography and concluded that various neurochemical changes within the disc are expressed by sensitized annular nociceptors. These nociceptors are terminal nerve endings of sensory neurons that selectively respond to painful stimuli by the release of substance P.⁴⁴ These chemicals are leaked into the epidural space and are transported into the axons of the exiting nerve roots. Within the nerve root, they alter the excitability of type C nerve fibers and initiate the production of inflammatory agents such as prostaglandins, which results in radicular pain.^20,45,46

In addition to material from the nucleus pulposus, many other substances in the degenerated disc have been implicated in pain generation. The role of nitric acid and phospholipase A₂ in irritation of nerve roots has been well documented.^45,47–50 Phospholipase A₂ has been implicated in multiple aspects: (1) direct activation of nociceptors, (2) nerve injury from degradation of cell membrane phospholipids, and (3) nerve injury from inflammatory mediators created from the arachidonic acid cascade (i.e., prostaglandins and leukotrienes).^51–53 Burke and colleagues⁵⁴ reported on the elevation of inflammatory mediators within the disc, such as IL-6, IL-8, and prostaglandin E₂. Other studies have shown the presence of inflammatory cytokines in the facet joints,⁵⁵ suggesting facet involvement as a pain generator via a biochemical mechanism as well. Ohtori and colleagues⁵⁶ reported on ingrowth of nerve tissue immunoreactive for tumor necrosis factor and prostaglandin P in 18 surgically harvested vertebral endplates of patients with Modic stage I and II changes who had undergone surgery. Their findings suggest that axon ingrowth into the vertebral endplate in association with Modic changes was induced by tumor necrosis factor and may be related to pain generation.

Neurovascular proliferation within and around degenerated disc elements has been proposed as another mechanism of pain generation. Normal IVDs have sparse innervation and vascularity that is distributed mainly within the outer lamellae (3 mm) of the anulus fibrosus,^57,58 whereas degenerated discs have significant neurovascular ingrowth within the inner anulus and nucleus pulposus.⁵⁹ Immunoreactive staining and acetylcholinesterase studies have shown penetration of nerve fibers within the inner third of the anulus in association with neovascularized granulation tissue.^47,57 Peng and colleagues⁶⁰ reported a histologic study of 19 IVDs harvested from surgery compared with normal control discs. The distinctive histologic characteristic of painful discs was a zone of richly innervated vascular granulation tissue extending from the outer anulus to the nucleus along the edges of fissures. Proliferation of vascular channels and sensory nerve endings rich in calcitonin gene-related peptide has also been observed in the endplate region and vertebral body adjacent to the degenerated disc. These findings suggest a role for the vertebral endplate and body as additional pain generators in DDD.⁶¹

Other studies suggest that the sensory nerve supply within the IVD is similar to visceral innervation patterns,⁶² with calcitonin gene-related peptide immunoreactive fibers that pass through the sympathetic trunks.⁶³ This visceral pattern of innervation is potentially susceptible to central sensitization, which may complicate chronic low back pain further with psychosomatic overtones.⁶⁴ Psychosocial and chronic non–back pain syndromes have been implicated in more recent publications as having a significant effect in patients with low back pain.^65–69

Internal Disc Disruption

The diagnosis of IDD is not readily apparent on a routine clinical workup. The patient is typically a younger individual 20 to 50 years old, with recurrent or persistent back pain. There may be a history of antecedent trauma or a forceful provocative event such as heavy lifting or unexpected flexion or compression force on the lumbar spine, but more often the pain is gradual in onset with no associated event or date.

The pain is characterized as a deep, dull ache in the lower lumbar region, exacerbated by rotation, flexion, and side-bending movements, and partially relieved by rest. Sitting intolerance may be a primary complaint, and pain is often relieved in a lateral recumbent position. Occasionally, there is a complaint of pain in the buttock or posterior thigh, but there is a conspicuous lack of radiculopathic symptoms. In the rare instances of associated leg pain, it is usually a late finding and pain does not follow any dermatomal pattern. In a study involving intradiscal electrothermal annuloplasty in 25 patients, O’Neill and colleagues⁷⁰ showed that stimulation of the IVD may result in low back and referred leg pain in patients presenting with symptoms of IDD. The distal distribution of pain was found to depend on the intensity of stimulation, and occasionally pain extending below the knee was produced.

Physical examination reveals decreased range of motion of the back and tenderness of the paraspinal musculature but is otherwise normal. The straight-leg raise test may reproduce back pain but not leg pain. Low back pain may also be reproduced at 20 to 30 degrees of flexion when rising from a flexed position. The sensorimotor examination is unremarkable, and deep tendon reflexes are normal and symmetrical.

Degenerative Disc Disease

Patients with DDD typically present with a history of persistent low back pain over the lumbosacral spine and sacroiliac joints and radiating into the buttocks and posterior thighs. Symptoms are often exacerbated with sitting and prolonged walking, but signs of neurologic claudication in the legs are not seen unless associated with concomitant lumbar stenosis. Radicular symptoms are rarely seen in the early stages of the disease. In end-stage disc degeneration, significant disc collapse may result in foraminal stenosis and late-onset radicular symptoms.

The physical examination is typically unremarkable except for point tenderness over the lumbar spine in the midline and over the sacroiliac joints. Range of motion of the lumbar spine may be reduced, most specifically in flexion. Extreme flexion and returning to upright from a flexed position usually cause significant discomfort. Extension is usually the least painful maneuver and may relieve pain. The straight-leg raise test may elicit some posterior thigh pain, which is often described as a stretching or pulling sensation, but there is no true radicular pain distal to the knee, unless there is coexisting foraminal stenosis. The sensorimotor examination is usually unremarkable, and deep tendon reflexes are normal and symmetrical.

Plain Radiography

Plain radiographs are the recommended initial imaging modality for patients with a complaint of low back pain. Classic comparative and cost benefit studies have been done to determine when and what radiographs to obtain.^71,72 In 1982, Liang and Komaroff⁷³ published a comparison study between performing radiographs on all patients versus performing radiographs only on patients whose pain did not improve within 8 weeks of presentation. They found that risks and costs did not justify obtaining radiographs on initial presentation. Scavone and colleagues⁷⁴ reviewed the radiographs of 782 patients and found that spot lateral and oblique films added diagnostic information in only 2% of patients. They recommended that a spine series in patients with low back pain should consist only of anteroposterior and lateral films. Generally, flexion-extension and oblique views are necessary only in patients suspected to have instability or a pars fracture. The presence of “red flags” increases the chances of diagnostic radiographic findings and may prompt the physician to obtain early radiographic studies. These “red flag” indications are summarized in Table 45–1.⁷⁵

TABLE 45–1 Practice Guidelines—“Red Flags”

From Institute for Clinical Systems Improvement (ICSI): Adult Low Back Pain. Bloomington, MN, ICSI, 2008.

Typical radiographic findings for patients with DDD include narrowing of the disc space (loss of height), endplate sclerosis, and the presence of osteophytes. Degenerative spondylolisthesis and scoliosis may occur secondarily. Advanced stages of disc degeneration may show vacuum phenomenon within the discs, a finding that represents nitrogen collection within voids in the disc.

Radiographs in patients with IDD typically show well-preserved height in the IVD and appear normal except for occasional benign spinal alignment changes. Nonstructural scoliosis and loss of lumbar lordosis may be observed in patients with sciatic list and paraspinal spasm.

Computed Tomography

CT is an excellent study to delineate osseous pathology, but it is generally not the imaging modality of choice for IDD and DDD because they are primarily soft tissue disorders. Addition of contrast material into the vertebral canal—CT myelography—significantly improves the accuracy of CT for showing pathology within the canal such as masses or stenosis, which is not a primary feature of DDD but can occur secondarily. CT myelography is the diagnostic imaging study of choice in patients with significant scoliosis and patients who are unable to undergo MRI because of implanted metal, aneurysm clips, pacemaker, obesity, or claustrophobia.

Magnetic Resonance Imaging

MRI is the best imaging modality to visualize and evaluate the neuronal and discal elements and is the most valuable adjunctive diagnostic tool in assessing disc pathology. IVDs are a highly unlikely cause of pain if MRI is completely normal and all discs are well hydrated. General MRI findings indicative of DDD include loss of water, loss of disc height, disc bulges, and signal or morphologic irregularity within the nucleus pulposus. In addition to these, MRI scans are typically examined for three specific types of findings: (1) a high-intensity zone (HIZ) in the posterior anulus, (2) dark disc with or without loss of height, and (3) endplate signal changes.

The MRI finding of a HIZ was originally described by Aprill and Bogduk⁷⁶ in 1992 and is believed to be specific for an annular tear (Fig. 45–1). Postmortem studies have shown three types of tears that can occur in the anulus: concentric, transverse, or radial.^77,78 A concentric tear is a crescentic or oval cavity created by a disruption in the short transverse fibers interconnecting the annular lamellae and is usually not visible on MRI. Concentric tears are occasionally referred to as delamination. A transverse tear represents a rupture of Sharpey fibers near their attachments to the ring apophysis at the disc periphery; these tears are typically thought to be clinically insignificant. A radial tear extending from the nucleus pulposus to the outermost surface of the posterior anulus is manifest on MRI as a HIZ.⁷⁹ HIZ is visualized on spin-echo T2-weighted images as a high-intensity signal located within the anulus fibrosus and is clearly distinguishable from the nucleus pulposus.

FIGURE 45–1 A and B, Sagittal (A) and axial (B) MRI showing high-intensity zone at L5-S1 level.

Decreased signal within the IVD on T2-weighted images with relative preservation of disc height is a relatively common finding in asymptomatic individuals. Such a disc appearance is frequently referred to as dark disc disease; however, whether these discs constitute a potential pain generator is unclear. In the absence of any psychometric abnormalities, an isolated dark disc in a patient with no other identifiable causes of back pain is considered by many clinicians to be a source of back pain.

Endplate changes (Fig. 45–2) that occur with disc degeneration have been well described by Modic and colleagues.⁸⁰ Stage I change represents edema and is characterized by decreased signal on T1 and bright signal on T2 within the endplate. In stage II, fatty degeneration in the bone adjacent to the endplates is represented by bright signal on T1 and intermediate signal on T2 sequences. Stage III changes correspond with advanced degenerative changes and endplate sclerosis and are characterized on MRI by decreased signal intensity on T1-weighted and T2-weighted images.

FIGURE 45–2 A and B, T1-weighted (A) and T2-weighted (B) sagittal MRI showing Modic stage I endplate changes.

When interpreting MRI findings, the clinician must be careful to consider the high prevalence of clinically false-positive findings. Abnormal disc findings on MRI are often found in clinically asymptomatic individuals. Boden and colleagues⁸¹ showed that approximately 30% of asymptomatic individuals have a major finding on lumbar MRI scans. In patients older than 60 years, such abnormal findings are almost universally present regardless of symptoms. Jensen and colleagues⁸² reported on 98 asymptomatic patients 20 to 80 years old and found that 52% overall had disc bulge in at least one level on MRI. Stadnik and colleagues⁸³ showed an unusually high rate of disc bulge (81%) and annular tears (56%) on MRI in 30 asymptomatic volunteers.

Abnormal MRI findings in asymptomatic patients are not indicators of future problems. Borenstein and colleagues⁸⁴ reported on 50 of the 67 patients from the study by Boden and colleagues⁸¹ at a 7-year follow-up interval and found that incidental MRI findings were not predictive of the development or duration of low back pain. Jarvik and colleagues⁸⁵ studied 148 veterans with no symptoms of low back pain for at least 4 months. They found an incidence of moderate to severe desiccation in at least one disc in 83% of patients, disc bulge in 64%, and loss of disc height in 58%. In a 3-year follow-up of the same cohort, the investigators found no association between the development of new back pain and incidental MRI findings such as Modic changes, disc degeneration, annular tears, or facet degeneration. The greatest risk factor for developing low back pain in the 3-year interval was depression.⁶⁵

Jarvik and colleagues⁸⁶ also published a report on the use of early MRI in the primary care setting. They randomly assigned 380 patients with low back pain to receive initial spine imaging via rapid MRI or plain radiography. Jarvik and colleagues⁸⁶ reported that substituting rapid MRI for x-ray studies in the primary care setting offered little additional benefit to patients in terms of secondary outcomes measures at 1 year and had the potential to increase the cost of care by $320 per patient (in 2002 dollar value). Carragee and colleagues⁸⁷ performed a prospective study of 200 asymptomatic patients to determine the rate at which new episodes of low back pain are associated with changes on MRI. On follow-up MRI in 51 patients who developed an episode of low back pain, 84% had no new finding. The most common new findings were disc signal loss (dark disc), progressive facet arthrosis, and increased endplate changes. New findings were not more common in patients developing back pain after minor trauma. The conclusion was that new findings on MRI within 12 weeks of onset of a serious episode of low back pain were unlikely to represent any significant structural change and preexisted the onset.

In consideration of the high prevalence of false-positive MRI findings, the clinician should remember that MRI does not stand alone in the evaluation of spinal pathology. When combined with a patient’s history, physical findings, and plain radiographs, selective use of MRI can provide valuable information on the source of lumbar complaints.

Discography

There is significant controversy in the literature surrounding the usefulness of discography for the evaluation of the integrity of the lumbar disc. Some investigators consider discography to be the most important tool in the diagnosis of IDD,^43,126 but more recent outcome studies¹²⁷ and a practice guideline by the American Pain Society¹²⁸ have recommended against the use of provocative discography in the diagnosis of discogenic back pain.

Discography is the only physiologic modality used to determine if a specific disc is a pain generator. Although several attempts have been made to explain the pathogenesis of pain provocation during discography, the precise pathomechanism is not well understood. There are four components to the evaluation of a discogram: (1) the pressure and volume of fluid injected into the disc, (2) the morphology of the disc being injected, (3) the subjective pain response at the level of interest, and (4) the pain response when adjacent control levels are injected.^129,130 The subjective pain response to low-pressure provocation is the most important determinant of disc derangement; reproduction of the patient’s symptoms on injection of the diseased level is essential to a positive test. A normal disc can accept 1 to 1.5 mL of contrast medium. If 2 mL or more of contrast agent is easily introduced, some degree of disc degeneration is assumed.

The use of postdiscography CT has also been reported to increase the sensitivity for the diagnosis of radial tears of the anulus.¹³¹ Because of low specificity and sensitivity, postdiscography CT is not as helpful, however, in the diagnosis of IDD. Most authors believe that to be diagnostic, not only should the pain be concordant on low-pressure injection, but also a normal control disc should be pain-free (Fig. 45–3).

FIGURE 45–3 A-C, CT discogram showing sagittal (A) and axial (B) images of normal disc morphology at L4-5 level and at adjacent level, L5-S1, disc fissure (C).

Despite being used since 1948, discography remains controversial. Holt¹²⁶ and Massie¹³² published in the 1960s on the high false-positive rate of lumbar discography, which was found to be 26% by Holt. Walsh and colleagues¹³³ later published a rigorous study on the reliability of lumbar discography. Their study compared 10 normal volunteers with 7 symptomatic patients. Although 17% of the normal discs were found to be morphologically abnormal, there were no positive pain responses. Walsh and colleagues¹³³ concluded that with modern techniques the false-positive rate of lumbar discography is not as high as reported by Holt.¹²⁶

Derby and colleagues¹³⁴ found similar results in a more recent study of 90 patients with low back pain and 16 controls. Morphologically, the prevalence of grade III annular tears was 58% among the asymptomatic control population. Presumably, asymptomatic discs in symptomatic individuals on pressure-controlled discography showed pain levels and responses similar to the control group, whereas patients with true-positive discography showed pain characteristics concordant with their usual symptoms. Derby and colleagues¹³⁴ concluded that pressure-controlled discography can differentiate between asymptomatic discs and morphologically abnormal discs.

Carragee and colleagues⁶⁸ studied the false-positive rate of low-pressure discography in a comparison of 69 volunteers with no significant low back pain and 52 patients undergoing discography in consideration for treatment of discogenic pain. Low-pressure discography was positive in at least one level in 27% of the patients with low back pain and in 25% of the controls. The false-positive rate of discography was 25% and correlated with psychosocial factors and history of chronic pain of a non-lumbar origin. In another publication from Carragee’s group,⁶⁹ psychosocial factors and chronic nonlumbar pain, such as cervical pain and somatization disorder, also correlated with positive discography in patients without symptoms of low back pain. These authors concluded that false-positive rates can be low with strict application of the Walsh protocol¹³³ in patients who do not have positive psychometric issues or other chronic pain syndromes.

In contrast to reports of high false-positive rates, two more recent meta-analyses of low-pressure discography report strong evidence to support the role of discography in identifying patients with discogenic pain.^135,136 Combined data from all studies showed an overall false-positive rate of 9.3% per patient and 6% per disc. False-positive rates among asymptomatic patients were 3% per patient and 2.1% per disc. Chronic pain was not found to be a confounder, and strength of evidence was reported as level II-2 in support of the diagnostic accuracy of discography.

Finding a “gold standard” with which discography results can be compared remains a problem. Few studies have compared the use of discography and outcomes after surgical fusion, which is perhaps the best measure for the validity of discography. Colhoun and colleagues,¹³⁷ in a study of 137 patients, reported 89% favorable outcomes in patients with positive concordant pain on discography versus 52% favorable outcomes among patients who had no painful response. Madan and colleagues¹³⁸ had different findings; 81% of 41 patients who underwent fusion based on MRI findings had satisfactory outcomes versus 76% of 32 patients who had surgery based on discography. Perhaps the most rigorous study to date was published by Carragee and colleagues.¹³⁹ In their study, success of surgical fusion was compared in 32 patients with single-level positive discogram and a matched cohort of 34 patients with single-level spondylolisthesis; 72% of the patients with spondylolisthesis met the highly effective success criteria for surgery versus only 27% of the patients with discogenic pain. Minimal acceptable success criteria were 91% and 43%. Carragee and colleagues¹³⁹ calculated a best case positive predictive value for discography of 50% to 60% and concluded that provocative discography was not highly predictive of single-level discogenic back pain.

In an attempt to improve on the poor reliability of discography, interest has turned to functional anesthetic discograms, also called discoblocks. A discoblock is a modification of discography, in which a local anesthetic, usually bupivacaine, is infused with the contrast agent into the disc to enhance the diagnostic capability of the procedure. Relief of pain after discoblock is considered diagnostic for discogenic pain. Ohtori and colleagues¹⁴⁰ published a randomized controlled study comparing standard provocative discogram with discoblock in diagnosing discogenic low back pain. Anterior lumbar interbody fusion (ALIF) procedures were performed in 15 patients whose discogenic pain was diagnosed with the aid of discography and 15 patients whose pain was diagnosed with the aid of discoblock. Outcome measures (ODI, VAS, and Japanese Orthopaedic Association score) at 3-year follow-up showed better results that were statistically significant in the group in which diagnosis was aided by discoblock.

Regardless of the details of how discography is performed, some authors have posed the question of potential ill effects resulting from perforating the lumbar disc. Carragee and colleagues¹²⁷ more recently published a report on the effect of lumbar discography in precipitating accelerated degeneration in a matched cohort study. The 10-year follow-up showed that discs that had been punctured had a greater progression of disc degeneration—35% versus 14% in the control group. There were 55 new disc herniations in the discography group versus 22 in the control group. Carragee and colleagues¹²⁷ concluded that despite using modern discography techniques with small-gauge needles, there is still an increased risk of disc degeneration, disc herniation, changes in disc and endplate signal, and loss of disc height when discography is performed.

Although discography has the potential to assist in diagnosing disc derangement, its reliance on the patient’s subjective pain response can also be problematic where secondary gain may be an issue. Psychosocial factors and chronic nonlumbar pain have also been shown to alter the diagnostic capabilities of the procedure. Finally, consideration of the consistent reports of the high false-positive rates and new findings of accelerated degeneration in discs that undergo discography make it difficult to recommend the procedure for the diagnosis of discogenic back pain. The validity of lumbar discography is very much in doubt, which is underscored by a more recent practice recommendation published by the American Pain Society. The society’s current recommendation is that provocative lumbar discography should not be used for making the diagnosis of a discogenic source of pain in the setting of nonradicular low back pain.¹²⁸ The value of using discography to assess the levels to be operated on in patients with multilevel disc degeneration has not been adequately established scientifically.

Bed Rest and Advice to Stay Active

The use of bed rest and its duration has long been debated in the literature. Treatment schedules ranging from 2 days to 6 weeks have been described.^142–144 The currently accepted recommendation⁷⁵ is limited bed rest for a maximum of 2 days because longer durations of bed rest may be detrimental to the patient’s general health while offering no benefit to the back pain. Allen and colleagues¹⁴⁵ published a review of studies documenting bed rest as treatment for 15 different conditions and found that for patients with acute low back pain there was significant worsening of outcome measures. The updated Cochrane Review of bed rest for treatment of acute low back pain reported that there is high-quality evidence that advice to rest in bed is less effective than advice to stay active.¹⁴⁶ Progressive return to activity and the initiation of a formal physical therapy or home exercise program are recommended after any initial short period of rest.

Verbunt and colleagues¹⁴⁷ explored reasons why patients sometimes use prolonged bed rest in the setting of acute episodes of low back pain. Among the study population of 282 patients, 33% reported using bed rest, and 8% remained in bed for longer than 4 days. Behavioral factors, catastrophizing, and fear of injury were associated with use of prolonged bed rest. History of back pain and pain intensity were not associated with patient use of prolonged bed rest. Additionally, patients who used prolonged bed rest in the early phase of acute low back pain were more disabled after 1 year.

Patient education and advice to stay active is now the favored recommendation. A more recent Cochrane review¹⁴⁸ of patient education and advice to stay active showed strong evidence that individual instructional sessions of 2.5 hours are more effective in returning patients to work than no intervention; however, in the setting of chronic back pain, patient education was less effective than more intensive interventions. Education sessions of shorter duration and written information were no more effective than no intervention. Another meta-analysis¹⁴⁹ of 39 randomized controlled studies evaluated advice to stay active alone or as an adjunct to other interventions such as back school or specific exercise routines. Advice as an adjunct to a specific exercise program was the most common form of treatment implemented and the best supported of the treatments studied for chronic low back pain. Outcomes among patients with acute low back pain were generally poor, but advice to stay active alone was found to be the best recommendation.

Brox and colleagues¹⁵⁰ published a systematic review of brief education in the clinical setting involving examination, information, reassurance, and advice to stay active. The authors found strong evidence that brief education was more effective for return to work but was no more effective than usual care for reduction of pain. There was limited evidence that dissemination of a back book or an Internet session was less effective than exercise. The authors concluded that a back book should not be distributed to patients as an alternative to another form of treatment.

Brace Wear and Other Orthotics

Another common conservative management technique involves the use of limited brace wear either with a recommendation to stay active or in conjunction with another form of nonoperative therapy. Calmels and colleagues,¹⁵¹ based on the results of a randomized clinical trial, recommended the limited use of a lumbar belt to improve functional status, pain, and medication use. Oleske and colleagues¹⁵² performed a randomized clinical trial of back supports and patient education in work-related back pain. These authors found no effect on patient self-report of recovery or lost work time between brace use and controls, but back supports were found to have some value in preventing recurrence of work-related back pain. A more recent Cochrane systematic review¹⁵³ of brace treatment for low back pain failed, however, to find sufficient evidence to support the use of lumbar supports to treat low back pain. Moderate evidence was found that braces are no more effective than no treatment or physical training in preventing episodes of back pain.

Use of shoe insoles has been recommended in the past for treatment and prevention of nonspecific low back pain. A more recent Cochrane systematic review¹⁵⁴ of six randomized controlled trials reported strong evidence that use of insoles does not prevent episodes of low back pain. There was limited evidence that insoles alleviated low back pain, but no conclusions or recommendations were made for use in the treatment of patients with low back pain.

Physical Therapy

Numerous physical therapy modalities and routines are described in the literature, including land-based and aquatic programs, specific protocols and exercise routines, and group treatment programs—so-called back schools. Adjunctive modalities include pain-relieving treatments such as ultrasound, iontophoresis, transcutaneous electrical nerve stimulation (TENS) unit, and heat therapy. Exercise programs commonly employ aerobic exercise, stretching, flexion and extension routines, core conditioning, and back stabilization protocols. The goal of all of these therapy regimens is to improve core strength, flexibility of the trunk and hip muscles, and conditioning. Patients often respond differently to physical therapy, so treatment programs commonly must be tailored on an individual basis. Periods of activity modification may be necessary. Patients should also be educated on proper body biomechanics; lifestyle change; and healthy living habits, such as weight control, proper nutrition, stress relaxation, and cessation of smoking.

There are multiple randomized controlled trials in the literature in support of many therapy routines or programs. Although a comprehensive review of all the various programs is not undertaken here, there have been some important updates in recent years worthy of discussion. More recent prospective randomized trials comparing physical therapy with fusion have emphasized the importance of a multidisciplinary approach with cognitive therapy, fear avoidance counseling, and intensive exercise programs.^155–157 A systematic review¹⁵⁰ found moderate evidence that fear avoidance training emphasizing exposure is more effective than graded activity increase for improvement of pain, disability, and fear avoidance.

Intensive interdisciplinary rehabilitation with emphasis on cognitive and behavioral intervention was one of the treatment recommendations made by the American Pain Society.¹²⁸ Interdisciplinary rehabilitation was defined by the society as an integrated intervention with rehabilitation plus a psychological or social or occupational component. The American Pain Society recommended that interdisciplinary therapy should be offered as a viable alternative before proceeding to surgical treatment. Noninterdisciplinary or “traditional” physical therapy is also efficacious in this patient population, but no one specific program, method, or technique is significantly better than another.

Back schools are another commonly discussed therapy modality, and there is some indication that low-intensity back schools may have some efficacy. Heymans and colleagues,¹⁵⁸ in a randomized controlled trial, found that patients who attended low-intensity back school experienced fewer sick leave days (68 days versus 75 days and 85 days) than usual care patients and patients who attended high-intensity back school. Functional status and kinesiophobia were improved at 3 months, but there was no difference in pain intensity and perception of recovery between the groups. In another randomized controlled trial, Kaapa and colleagues¹⁵⁹ found no significant benefit of back school, however, compared with physical therapy combined with cognitive therapy at 6-month, 1-year, and 2-year follow-up.

A systematic review from the Cochrane Database in 2004¹⁶⁰ concluded that there was moderate evidence suggesting that back schools in an occupational setting reduce pain and improve function and return-to-work status compared with other forms of therapy, such as exercises, manipulation, myofascial therapy, advice, placebo, and waiting list controls. Brox and colleagues¹⁵⁰ published a separate systematic review of back schools and found moderate evidence that back schools were no better than waiting lists, no intervention, placebo, or general exercises for reduction of pain.

A European economic evaluation of a randomized controlled study¹⁶¹ of intensive group therapy found no significant cost difference between intensive group therapy and standard physiotherapy. There was also no difference in clinical effect between the groups at 1-year follow-up.¹⁶² To the authors’ knowledge, there are no economic studies to date of group therapy back schools in the United States. Although low-intensity back school and programs in a work setting may have benefit versus other forms of nonoperative treatment, most of the current literature shows that back schools offer little benefit over standard physical therapy and cognitive therapy.

Adjunctive Modalities

Another treatment option for low back pain includes adjunctive physical therapy modalities such as TENS, electrical muscle stimulation, ultrasound, and iontophoresis. Poitras and Brosseau¹⁶³ reviewed randomized controlled data on the use of TENS and found that it may be useful for immediate short-term pain reduction but has little impact on patient perception of disability or on long-term pain control. A 2008 Cochrane systematic review of TENS versus placebo¹⁶⁴ concluded that there is currently not enough evidence to support the routine use of TENS for management of chronic low back pain. Even less literature is available on the use of iontophoresis and ultrasound in the setting of discogenic back pain. The few randomized controlled trials that exist focus on ultrasound in conjunction with other physical therapy regimens. The efficacy of these modalities in isolation has not been determined.

Chiropractic and Complementary and Alternative Medicine Therapies

Several studies have reported the potential beneficial effects of chiropractic treatment for acute nonspecific low back pain.^165–167 The role of chiropractic manipulations for the treatment of IDD or DDD of the lumbar spine has not been studied. Chiropractic manipulation is generally not considered effective in the treatment of chronic back pain resulting from disorders of the IVDs.¹⁶⁸ A Cochrane systematic review¹⁶⁹ failed to find evidence that spinal manipulative therapy was superior to general practitioner care, analgesics, physical therapy, exercises, or back school in the treatment of acute and chronic low back pain.

Eisenberg and colleagues¹⁷⁰ published a randomized trial of usual care therapy versus the addition of the patient’s choice of alternative therapy—chiropractic, acupuncture, or therapeutic massage—in the treatment of acute low back pain. Outcomes based on the Roland-Morris scale and subjective assessment of symptoms showed no statistically significant improvement in patients who underwent alternative therapies compared with patients treated with the usual care of limited bed rest, nonsteroidal anti-inflammatory drugs (NSAIDs), education, and activity modification. The study did show, however, an increase in patient satisfaction with care, which came at an average $244 net increase in cost per patient.

Hurwitz and colleagues¹⁷¹ had similar findings in a randomized prospective study of 681 patients with chronic low back pain comparing chiropractic care with medical treatment with 18 months of follow-up. Although less than 20% of the patients overall experienced pain relief and differences in outcome measures were not clinically significant, patients in the chiropractic group were more likely to perceive that their symptoms had improved.

Other alternative medical therapies include acupuncture, prolotherapy, and massage. The Cochrane systematic review of acupuncture¹⁷² showed superiority to placebo sham therapy and a short-term benefit that did not extend beyond first follow-up when acupuncture was used in conjunction with other conventional therapies. A more recent systematic review by Ammendolia and colleagues¹⁷³ questioned inconclusive evidence of the success of acupuncture versus sham acupuncture and called for further randomized trials to rule out the possibility of a placebo effect.

Prolotherapy is a technique that attempts to regenerate ligamentous and tendinous structures of the spine via injections of various irritant solutions. The treatment is usually performed in conjunction with spinal manipulation. There is no consensus on method, type of solution injected, or frequency of sessions. Most practitioners use various combinations of saline, dextrose, glycerin, phenol, and lidocaine. Many randomized trials and systematic reviews report conflicting efficacy of prolotherapy.^174–176 No evidence has been reported for the efficacy of prolotherapy without cointerventions such as spinal manipulation or exercise.

The efficacy of complementary and alternative modalities for the treatment of low back pain remains doubtful. The benefit of spinal manipulative therapy is also controversial, but it may improve patient satisfaction with care and perception of symptoms.

Pharmacotherapy

Judicious use of narcotic pain medications, oral steroids, and NSAIDs in patients with severe, acute back pain can provide good pain relief. Most patients with painful degenerative discs can be treated adequately on an outpatient basis. NSAIDs and acetaminophen (Tylenol) are common over-the-counter medications used to treat back pain. A Cochrane review¹⁷⁷ included 65 studies on NSAID use in low back pain. NSAIDs were found to be superior to placebo but had significantly more side effects. There is no documented difference between type of NSAID, including cyclooxygenase-2 inhibitors. Acetaminophen has an effect similar to NSAIDs but reduced risk of associated side effects when taken as directed and in general should be tried before NSAIDs. The Cochrane group concluded that NSAIDs are effective for short-term treatment of acute and chronic low back pain, but the size of the effect is small.

Opioid formulations are commonly used to treat back pain, but considering their widespread use there is a surprising paucity of high-quality randomized controlled data available on their efficacy. A Cochrane database meta-analysis¹⁷⁸ of opioid use found only four studies, three of which focused on the use of tramadol. Pooled data found that tramadol, an atypical opioid, was more effective than placebo for pain relief and showed a slight improvement in functional scores. The only randomized controlled study of classic opioids¹⁷⁹ was a comparison with naproxen. Opioids were found to be more effective for pain relief but were not more effective for improving function than naproxen. The Cochrane review authors concluded that the benefits of opioids for the treatment of chronic low back pain are questionable, and further well-designed randomized controlled studies need to be performed. Two subsequent systematic reviews^180,181 of opioid use in the setting of chronic low back pain concluded that there is evidence to support the efficacy of opioids for short-term relief of pain only. There is little evidence for long-term opioid use, which is fraught with an incidence of aberrant consumptive behavior approaching 25%.

Use of opioid pain medication has many problems ranging from minor side effects such as constipation and nausea to severe complications including respiratory depression, altered mental status, and insidious issues with tolerance and addiction. Another more recent concern with opioid use is related to the combination of opioids and acetaminophen in commonly prescribed formulations.¹⁸² The maximum recommended daily dose of acetaminophen for adults and children older than 12 years is 4 g; thus concern arises when patients inadvertently take larger doses in the setting of prescription drug abuse. The potential to inadvertently take hepatotoxic or lethal doses can be a concern in the setting of prescription drug abuse. An advisory committee from the U.S. Food and Drug Administration (FDA)¹⁸⁴ recommended the addition of a boxed warning on the risk of acetaminophen overdose and suggested elimination of combination opioid-acetaminophen formulations. Care should be exercised when prescribing opioid pain medications. They are best given for only a few days in the setting of severe acute back pain, and their use in patients with chronic back pain is not recommended.

Oral tapering courses of steroids have also been found useful for decreasing symptoms of low back pain, most specifically in patients with disc herniations.^80,185 Steroids can cause gastrointestinal bleeding, and gastrointestinal protective agents should be used simultaneously with oral steroids to reduce the risk of this complication.

Muscle relaxants are another class of medication routinely used in the treatment of muscle spasm associated with low back pain. Their use should be limited to very short courses because of their addictive potential. The Cochrane review¹⁸⁶ of muscle relaxants for the treatment of back pain included 30 trials evaluating the use of benzodiazepines, nonbenzodiazepines, and antispasmodic muscle relaxants. Strong evidence for the efficacy of muscle relaxants over placebo was reported for short-term pain relief in the setting of acute back pain. No difference between the various drugs and classes was discerned. More trials to determine the efficacy of muscle relaxants compared with other analgesics and NSAIDs were recommended.

The last class of medications commonly prescribed in the setting of back pain is antidepressants. Their use may be particularly beneficial in patients presenting with chronic low back pain in association with altered mental status, depression, anhedonia, sleep disturbances, agitation, and anorexia. Clinical studies^187–189 supporting the use of tricyclic antidepressants (TCAs) have shown an improvement in mood and sleep patterns. Low doses of TCAs also affect membrane potentials of peripheral nerves, which may be a mechanism by which they produce pain reduction. A 2003 review¹⁹⁰ of antidepressants in the treatment of chronic low back pain found that TCAs have a moderate effect on pain reduction in patients with no history of depression but reported conflicting evidence for improvement in functional outcomes. Physicians prescribing TCAs should be aware of potentially serious side effects involving orthostatic hypotension and cardiovascular perturbations. In a systematic review,¹⁹¹ selective serotonin reuptake inhibitors, another common class of antidepressants, failed to show efficacy in the treatment of chronic low back pain and should be reserved for emotional or psychiatric disturbances related to back pain and not used as a primary treatment for symptoms of back pain.

Keller and colleagues¹⁹² published a meta-analysis of nonsurgical management options for low back pain. These authors reported that behavioral therapy, exercise therapy, and NSAIDs had the largest effect of the modalities studied. Machado and colleagues¹⁹³ published a separate large meta-analysis of placebo-controlled randomized trials of various forms of nonoperative treatment for nonspecific low back pain. Small improvements in complaints of pain were found in patients treated with traction, physical therapy, antidepressants, and NSAIDs; moderate improvements were found in patients treated with opioid analgesics, muscle relaxants, facet injections, and nerve blocks.

Epidural Spinal Injection

Administration of epidural steroids should be considered by the surgeon and patient before proceeding to a surgical intervention. The advantage of epidural injections over oral steroids is the ability to achieve higher concentrations of steroid at the site of pain while minimizing systemic effects. Epidural steroids typically work well when administered in the setting of radicular pain and do not work well in the setting of axial pain. Patients with foraminal stenosis secondary to loss of disc height may benefit from selective nerve root blocks either as a diagnostic or as a therapeutic tool. Many clinicians recommend epidural steroid injections as second-line therapy in the treatment of lumbar disc disorders. Epidural steroids are commonly administered by three different routes: caudal, interlaminar, and transforaminal. Although discogenic back pain with leg symptoms is considered an indication for all three modes of administration, the transforaminal approach is generally considered best because it achieves a better anterior epidural distribution. Complications from injection exist but are uncommon.^194,195

Reports on the efficacy of epidural injections in the literature are contradictory. Manchikanti and colleagues¹⁹⁶ published preliminary results of a randomized trial of serial caudal epidural injections in patients with discogenic pain without disc herniation or radiculitis. These authors reported greater than 50% pain relief in 72% to 81% of patients and 40% reduction in ODI scores in 81% of patients. Manchikanti and colleagues¹⁹⁶ concluded that caudal epidural injections with or without steroid are effective in treating discogenic back pain in greater than 70% of patients. Two other observational studies by the same authors^197,198 have similar findings for the beneficial effects of caudal epidural injections in the specific setting of discogenic low back pain.

Buttermann¹⁹⁹ studied patients with DDD and back pain of greater than 1 year’s duration who were candidates for fusion. There was initial success of treatment in greater than 50% of patients, but success rate declined to 23% to 29% by the 1- to 2-year follow-up. The study was plagued by a high dropout rate with more than two thirds of the patients seeking another invasive treatment within 2 years. Buttermann¹⁹⁹ concluded that patients with DDD without spinal stenosis may experience a short-term benefit from epidural injections with only one fourth to one third experiencing long-term improvement in pain and function. Other earlier studies of caudal and transforaminal approaches have reported similar good results for short-term efficacy in low back pain, with 59% of patients having greater than 50% improvement in symptoms and function at a 1-year interval.^197,200

A more recent systematic review²⁰¹ criticized the literature on epidural injections for a lack of careful control of route of administration and patient diagnosis. On evaluation of the pooled data, the only evidence found in support of epidural injections was for short-term symptom relief in nonspecific low back pain. No well-designed randomized trials were found specific to discogenic back pain. A 2008 Cochrane systematic review²⁰² of injection therapy for low back pain failed to find sufficient evidence to make a recommendation. A systematic review by Chou and colleagues,²⁰³ as part of the American Pain Society practice recommendations, found fair evidence that epidural steroid injection is moderately effective for short-term pain relief; however, the literature supporting its use in nonradicular low back pain is sparse and has not shown significant benefit. No specific recommendation for the use of epidural steroid injections or the route of administration was made by the American Pain Society.

Intradiscal Injection

Direct intradiscal injection, usually with a steroid solution, is another intervention that has been described in the literature for IDD. The desired effect is suppression of an inflammatory process within the disc, which is thought to be the cause of the discogenic pain. Intradiscal steroid injections were reported in an early case series by Feffer,²⁰⁴ in which 47% of patients reportedly had remission of discogenic symptoms. Similar results were found by Wilkinson and Schuman.²⁰⁵ More recently, Fayad and colleagues²⁰⁶ reported a short-term improvement in VAS score at a 1-month follow-up with intradiscal steroid injection in patients with Modic stage I and I-2 changes on MRI, but there was no long-term benefit. The only two major prospective randomized trials^207,208 of intradiscal steroid injection failed to find a statistically significant benefit versus placebo in the treatment of discogenic back pain. Other authors have attempted intradiscal injection of various other substances, including solutions of chondroitin and dextrose,²⁰⁹ hypertonic dextrose,²¹⁰ methylene blue,²¹¹ and oxygen–ozone gas mixtures.^212,213 Although these studies purport promising results, they have yet to be proven efficacious by rigorous randomized controlled trials.

Thermal Annuloplasty

IDET involves percutaneous insertion of a thermally controlled catheter into the IVD, usually the posterior anulus, and heating the catheter to a specific temperature (usually 90° C) for a proscribed period (4 to 12 minutes depending on the protocol). Multiple variations of the procedure exist differing on the type of energy delivered (e.g., percutaneous radiofrequency thermocoagulation), mode of energy delivery (e.g., bipolar), positioning of thermal probe, timing, and duration of energy delivery.

The proposed mechanism of action for these procedures is twofold: (1) elimination of nociceptive pain fibers and aberrant painful responses to the disrupted disc and (2) collagen rearrangement in the anulus with resultant spinal segment stabilization. The biologic effects are not well understood, and there is a lack of clear consensus regarding the effects on neuronal deafferentation, collagen modulation, and spinal stability. Freeman and colleagues²¹⁴ studied the effect of nociceptor destruction via IDET on experimentally created annular tears in a sheep model. The authors failed to find any difference in the amount of neoinnervation in the anulus between specimens that underwent IDET and specimens that did not, which calls into question the theory of deafferentation of the anulus. Whether collagen rearrangement with resultant shrinkage and stabilization of the discal element is a viable mechanism for IDET also is questioned.^215,216 Cadaveric studies of the effect of IDET on annular collagen have been performed by Kleinstueck and colleagues,²¹⁶ which showed a 10% to 16.7% reduction in tissue volume immediately adjacent to the electrode.

To destroy nociceptors in the anulus fibrosus, temperatures must be increased to a minimum of 42° C to 45° C.^217,218 It is impossible to generate sufficient temperatures in the anulus with a radiofrequency probe placed in the center of the disc as shown by Houpt and colleagues.²¹⁹ Temperature changes at distances farther than 11 mm were insufficient to increase the tissue temperature of the outer anulus to the 42° C needed for neuronal ablation. Ashley and colleagues²²⁰ compared temperature distribution in the disc between a radiofrequency needle and a navigable SPINECATH (Smith & Nephew, Memphis, TN). Using this method, they were able to deliver thermal energy to the anulus more effectively and achieved sufficient temperatures to cause denervation. Karasek and Bogduk²²¹ recommended inserting the IDET electrode so as to remain within 5 mm of the outer surface of the anulus. Placement of the probe in the interlamellar plane rather than inside the innermost layer of the anulus allows for sufficient heat generation to destroy the nociceptors in the outer layers of the anulus.

Complications secondary to any of the thermal annuloplasty procedures are rare. There has been one reported case of postoperative cauda equina syndrome caused by inadvertent placement of the catheter in the spinal canal²²² and a few reports of broken catheters with no resultant adverse effect. There have been no reports of infection, bleeding, or other equipment-related or technique-related complications.

Early uncontrolled clinical trials of IDET were promising, with improvement in 50% to 70% of patients,^42,221–224 but randomized controlled trials have produced conflicting results. Freeman and colleagues²²⁵ found no significant improvement in outcome measures compared with sham surgery at 6 months’ follow-up. The opposite findings were reported by Pauza and colleagues²²⁶ in patients with discographically diagnosed low back pain of greater than 6 months’ duration. Pauza and colleagues²²⁶ found that 40% of their patients who underwent IDET experienced at least 50% relief of pain, whereas a significant portion of the control group experienced symptom progression. These authors concluded that the IDET procedure is an effective intervention for a selective patient population and reported a number needed to treat of 5 to achieve a 75% relief of pain. Barendse and colleagues²²⁷ reported on a trial of intradiscal radiofrequency thermocoagulation in patients with chronic discogenic back pain. An 8-week follow-up assessment showed no difference from sham surgery in VAS score, global perceived effect, and ODI outcome measures.

Andersson and colleagues²²⁸ published a systematic review of IDET versus spinal fusion in patients with disc degeneration and disruption. Similar median percentage improvement was noted between the two interventions for pain severity and quality of life outcomes. Fusion showed better functional improvement but had a higher rate of complications. Andersson and colleagues²²⁸ concluded that IDET offers similar symptom relief with less risk of complications compared with fusion. In a systematic review, Derby and colleagues²²⁹ concluded that IDET is generally safer and cheaper than more invasive surgical techniques despite the fact that the best evidence available shows only modest improvement in pain relief and functional outcomes.

Other systematic reviews of IDET have been more critical. Helm and colleagues²³⁰ reported level II-2 evidence in support of IDET in the setting of discogenic back pain based on two of the above-mentioned randomized trials and numerous observational studies. Two observational studies were found by these authors in support of radiofrequency intradiscal thermocoagulation for a II-3 level of evidence. Evidence in support of biacuplasty was lacking and was assigned level III. Freeman²³¹ published a systematic review of the literature that criticized generally poor outcomes even among studies in support of IDET. Freeman²³¹ concluded that evidence for the efficacy of IDET is weak and has not passed the standard of scientific proof.

Chou and colleagues²⁰³ published a systematic review summarizing all nonoperative interventional therapies as part of the American Pain Society practice recommendations published in 2009. These authors reported fair evidence that epidural steroid injections are effective for short-term pain relief. Good evidence was reported that prolotherapy, facet injection, intradiscal steroid, and intradiscal radiofrequency thermocoagulation are ineffective. For IDET, no conclusions were made because available randomized controlled trials are conflicting. IDET may best be indicated for patients with less functional impairment, with well-maintained disc heights, and with discogenic pain from annular tears.²²⁹ IDET is not universally successful, but roughly 50% of patients can expect significant reduction (>50%) in pain.

Spinal Fusion

Surgical treatment for unremitting discogenic back pain has traditionally been spinal fusion; however, fusion is not universally accepted as the “gold standard” for this condition. Most clinicians find that it is acceptable to perform spinal fusion for DDD in patients who have failed exhaustive conservative care. The role of spinal fusion in the management of IDD is more controversial.²³³

The goal of a fusion procedure is to eliminate motion at the affected spinal segment. Arthrodesis can be accomplished through a posterolateral fusion (PLF), an interbody technique after removal of the IVD, or a combined approach (360-degree). Interbody approaches include ALIF, through either an abdominal or a retroperitoneal approach; transforaminal lumbar interbody fusion (TLIF), through the facet and neuroforamen; posterior lumbar interbody fusion (PLIF), via a canal decompression; fusion from the side (extreme lateral lumbar interbody fusion [XLIF]), via a transpsoas approach; and use of a presacral approach (percutaneous axial lumbar interbody fusion [AxiaLIF]). All fusion techniques can be supplemented with instrumentation. There are various anterior plates to supplement ALIF procedures, and posteriorly pedicle screw and rod constructs are commonly used. Also, various materials and cages are available to place between the vertebral bodies to perform interbody fusion. Each of these fusion techniques is discussed in subsequent sections.

Three high-quality randomized controlled studies in the past decade have evaluated spinal fusion compared with nonoperative treatment in the setting of chronic low back pain and DDD. Fritzell and colleagues²³⁴ published a randomized controlled multicenter study of severe chronic low back pain comparing fusion of the lower lumbar spine with nonsurgical therapy. The study involved 222 operative and 72 nonoperative patients 25 to 65 years old with chronic low back pain of at least 2 years’ duration and radiologic evidence of disc degeneration at L4-5, L5-S1, or both. The nonsurgical group received physical therapy, patient education, and alternative pain control modalities, such as TENS units, acupuncture, and injections. Results at 2 years’ follow-up were found to be significantly better in the fusion group, with back pain reduced by 33% compared with 7% in the nonsurgical group. Pain improvement was most significant during the first 6 months postoperatively and then gradually deteriorated thereafter. Disability according to ODI was reduced by 25% compared with 6% among nonsurgical patients, and 63% of surgical patients rated themselves as “much better” or “better” compared with 29% of nonsurgical patients. The “net back to work rate” was 36% in the surgical group and 13% in the nonsurgical group. The early complication rate in the surgical group was 17%. Fritzell and colleagues²³⁴ concluded that surgical treatment of severe chronic low back pain provides improved results compared with nonoperative treatment in carefully selected patients.

Brox and colleagues¹⁵⁵ published another randomized trial comparing outcomes of lumbar instrumented fusion versus cognitive intervention and exercise in 64 patients with chronic low back pain and DDD. The critical component of this study was the addition of cognitive therapy to an intensive rehabilitation program. The mean change in ODI for the surgical fusion group was from 41 preoperatively to 26 at 1-year follow-up and for the rehabilitation group from 42 to 30. The investigators reported no significant difference in back pain, use of analgesics, emotional distress, and life satisfaction between the groups. Return to work rate at 1 year was 22% in the surgical group and 33% in the rehabilitation group. The rehabilitation group experienced greater improvement in fear avoidance beliefs, and fingertip-to-floor distance, whereas the surgical group had greater improvement in associated symptoms of leg pain. The overall success rate for surgical intervention was 70% and for nonoperative cognitive therapy was 76%. Brox and colleagues¹⁵⁵ concluded that there were near-equivalent outcomes between the groups, which was offset by an 18% complication rate among the surgical group.

Fairbank and colleagues¹⁵⁷ published the last major randomized clinical trial of surgery versus nonoperative therapy. The Medical Research Council (MRC) spine stabilization trial was a randomized controlled trial comparing surgical treatment and intensive rehabilitation in 349 patients with chronic low back pain. Similar to the study by Brox and colleagues,¹⁵⁵ the intensive physical therapy program in the MRC trial also incorporated principles of cognitive behavioral therapy. At 1-year follow-up, the mean ODI scores decreased from 46.5 to 34 in the surgical group and from 44.8 to 36.1 in the rehabilitation group. No significant differences were found between the groups in the shuttle walking test and Short Form-36 General Health Survey (SF-36) outcomes. The authors concluded that although the surgical group enjoyed a small but statistically significant benefit in one of the primary outcome measures (ODI), this was contradicted by the additional cost and potential risk of complication associated with surgery.

In a separate publication on the MRC trial, Rivero-Arias and colleagues²³⁵ performed a cost analysis at 2 years’ follow-up. The cost per patient over the study time frame in the surgical group was estimated to be £7830 ($12,450) versus £4526 ($7200) in the rehabilitation group. There was no significant difference in mean quality-adjusted life-years between the groups. The investigators concluded that surgical treatment was not a cost-effective use of health care funds compared with therapy, although the authors pointed out that ultimate costs could vary depending on the number of patients in either group that require subsequent intervention after the 2-year follow-up period.

Meta-analyses of surgical versus nonoperative treatment have paralleled the findings of Brox and colleagues¹⁵⁵ and the MRC trial.¹⁵⁷ Ibrahim and colleagues²³⁶ pooled the data from these three randomized trials and found that a modest improvement in mean ODI scores among surgical patients should not be used as justification for routine operative treatment in light of a 16% early complication rate. Mirza and Deyo,²³⁷ in a separate systematic review, concluded that surgical outcomes are equivalent to a structured rehabilitation program with cognitive behavioral therapy.

Posterolateral Fusion

PLF is typically performed through a traditional midline approach with exposure of the posterior spinal elements; disruption of the facet joints at the fusion levels; and decortication of the transverse processes, pars, and facets to stimulate fusion. Autograft or allograft bone is typically placed over the decorticated areas, and fusion may be augmented with various osteoconductive and osteoinductive materials. Instrumentation in the form of pedicle screw and rod constructs can also be placed to provide segmental stability to increase the success of fusion.

Fusion rates for the lumbar spine vary in the literature and depending on the type and extent of procedure. For a single-level, uninstrumented PLF in the setting of DDD, fusion rates of 85% to 91% have been reported.²³⁸ McCulloch²³⁹ reported a 91% solid fusion rate with uninstrumented single-level PLF and a satisfactory clinical outcome in 78% of patients.

The improvement in fusion rates and outcomes with instrumentation is debated in the literature. France and colleagues²⁴⁰ reported radiographic fusion rates of 76% among instrumented patients and 64% among noninstrumented patients, but there was no significant difference in outcomes between the groups or any correlation between radiographic union and patient-reported improvement. Other studies have reported a 26% rate of pseudarthrosis in uninstrumented fusion,^241,242 whereas addition of instrumentation has been reported to improve fusion rate, reduce symptoms of pain, and increase return-to-work rate.^241–247 In a prospective study of one-level fusions with and without instrumentation for disabling back pain, Lorenz and colleagues²⁴³ reported superior results with instrumented fusion. There were no reports of pseudarthrosis among the instrumented group, and 75% of patients experienced improvement in pain and were able to return to work. In contrast, 58% of patients in the uninstrumented group had a nonunion, and only one third experienced pain relief and were able to return to work.

In contrast, Thomsen and colleagues,²⁴⁸ in a randomized clinical study, reported no statistical difference in the rate of fusion between instrumented and noninstrumented patients. Instrumentation was related to an increase in operative time, blood loss, and early reoperation rate and a 4.8% risk of pedicle screw misplacement. Bono and Lee²⁴⁹ performed a comprehensive review of studies published on lumbar fusion from 1979-2000 and noted a clear trend toward increasing use of instrumentation—23% of all fusions in the 1980s versus 41% of all fusions in the 1990s. These authors were unable to show any significant improvement in overall fusion rate or clinical outcome. The benefit of supplemental instrumentation in PLF is not clearly documented in the literature, particularly in light of newer biologic materials currently being used to enhance fusion rates.

Lumbar Interbody Fusion

There are many potential benefits of using an interbody technique for lumbar fusion. The lumbar vertebral body represents 90% of the surface area and supports 80% of the load within the spine. The greater amount of compressive force anteriorly and the larger surface area theoretically leads to a greater potential for fusion. Interbody fusion is also a more effective technique for maintenance of sagittal and coronal deformity, which can be particularly important in the setting of loss of lumbar lordosis secondary to disc collapse or postlaminectomy kyphosis.

PLF can lead to the persistence of discogenic pain in some patients despite solid fusion, presumably owing to the presence of micromotion and pain in the involved disc. The disc material itself may be a pain generator, which interbody fusion directly addresses by discectomy.^250,251 Weatherley and colleagues²⁵⁰ reported resolution of pain after an ALIF in five patients with persistent back pain despite solid PLF. Barrick and colleagues²⁵² also reported excellent pain relief after anterior interbody fusion in 20 patients who had persistent low back pain despite previous PLF. For these reasons, interbody fusion is thought by many to provide better and more predictable pain relief in patients with primarily a discogenic source of low back pain, but there are no high-quality studies supporting this view.

Anterior Lumbar Interbody Fusion

ALIF with bone grafting (Fig. 45–4) for the treatment of IDD was the treatment modality originally recommended by Crock³⁸ when he described the disorder. ALIF classically is performed through either an intra-abdominal or a retroperitoneal approach. After the symptomatic disc levels have been exposed, the surgeon performs an annulotomy and complete discectomy. The discal segment is reconstructed with autograft bone, allograft bone, or a cage device.

FIGURE 45–4 A and B, Preoperative sagittal (A) and axial (B) T2-weighted MRI in a patient with L5-S1 discogenic back pain, treated with anterior lumbar interbody fusion using carbon fiber–reinforced cages and rhBMP-2. C and D, Postoperative anteroposterior (C) and lateral (D) radiographs are shown.

Reports of success for ALIF in the literature are high; Loguidice and colleagues²⁵³ reported an 80% rate of fusion and an 80% clinical success rate with ALIF. Newman and Grinstead²⁵⁴ reported similar results with 89% fusion and 86% clinical success in patients with IDD. Other than infection, the early risks in ALIF are mainly associated with the surgical approach, including ileus, injury to the abdominal contents or vasculature,^255,256 incisional hernia, muscular atony, and retrograde ejaculation in men secondary to injury to the autonomic plexus.^257,258

Circumferential Fusion

Combined interbody and posterior fusion, so-called global or 360-degree fusion, is another technique described in the literature. Interbody fusion can be performed via a separate anterior approach (ALIF and PLF) (Fig. 45–5), but a posterior approach (PLIF or TLIF) is often simpler because it involves a single approach for both parts of the fusion procedure (these procedures are discussed subsequently). A combined ALIF and PLF procedure previously was reserved for situations in which the risk of pseudarthrosis was high, such as in patients undergoing revision surgery, patients with preexisting pseudarthrosis, smokers, and diabetic patients; however, today it is commonly used in primary cases as well.

FIGURE 45–5 A and B, Postoperative anteroposterior (A) and lateral (B) radiographs of a patient with L4-5 and L5-S1 discogenic back pain treated with two-level circumferential fusion via an anterior retroperitoneal approach with carbon fiber–reinforced cages and posterior percutaneous transpedicular instrumentation and posterior spinal fusion.

Moore and colleagues²⁵⁹ reported a 95% arthrodesis rate and 86% clinical success rate with combined anterior and posterior fusion for patients with chronic low back pain and DDD who had failed prolonged nonoperative treatment. Gertzbein and colleagues²⁶⁰ reported 97% fusion rate and 77% good clinical outcome with global fusion in a challenging group of patients—62% had previously had surgery, 25% had pseudarthrosis, 55% had two or more levels fused, and 43% were heavy smokers. Kozak and O’Brien²⁶¹ treated 69 patients with circumferential fusion through two incisions for discogram-positive, disabling low back pain. They reported greater than 90% good results with one-level and two-level fusions and 78% good results with three-level procedures. Similarly, Hinkley and Jaremko²⁶² reported greater than 90% positive outcomes in 81 patients who were receiving workers’ compensation and were treated with 360-degree lumbar fusion. Videbaek and colleagues,²⁶³ in a randomized clinical trial involving 148 patients comparing the results of circumferential fusion with PLF at 5 to 9 years’ follow-up, found that the circumferential fusion group had significantly better outcomes as measured by the Dallas Pain Questionnaire (DPQ), ODI, and SF-36. The circumferential group also complained of less back pain than the PLF group.

Although combined ALIF and PLIF may reduce the rate of pseudarthrosis to less than 5%, the morbidity is higher than for PLF alone, which is often warranted in patients undergoing revision spine surgery, diabetics, and heavy smokers. Suratwala and colleagues²⁶⁴ reported retrospectively on 80 complicated patients who underwent circumferential fusion of three or more levels. Encountered complications included 19% pseudarthrosis rate per patient (12% per level), 14% symptomatic pseudarthrosis, and 14% rate of adjacent segment degeneration. Within the 2- to 7-year follow-up, 34% of patients underwent repeat surgery with 20% undergoing implant removal for pain. The rate of deep wound infection was 2.5%, and the rate of superficial infection was 3.8%. Excessive intraoperative bleeding (>3 L) was rare, but 50% of patients required transfusion. Despite the rate of perioperative complications in this complex patient population, the authors reported mean ODI improvement from 50 to 35 and statistically significant improvement in SF-36 and Roland Morris scores.

Posterior and Transforaminal Lumbar Interbody Fusion

Posterior techniques for performing interbody fusion, including PLIF and TLIF (Fig. 45–6), are typically performed with posterior instrumentation and fusion, which makes them by default circumferential fusions. TLIF and PLIF have become increasingly popular techniques for performing circumferential fusion because they can be performed through a single posterior incision, which considerably lessens the morbidity associated with combined anterior and posterior approaches.²⁶⁵

FIGURE 45–6 A-D, Preoperative T2-weighted sagittal (A) and axial (B) MRI and postoperative anteroposterior (C) and lateral (D) radiographs of a patient with L4-5 discogenic back pain after prior discectomy treated with transforaminal lumbar interbody fusion and transpedicular instrumentation.

In PLIF, the IVD is approached through laminectomy, partial facetectomy, and retraction of the dura and its contents. Risks of PLIF include dural tears, conus injury from retraction, nerve root injury, and epidural fibrosis. Success rates of PLIF in the literature are mixed. Madan and Boeree²⁶⁶ reported no difference in the outcome of discogenic back pain treated by ALIF versus instrumented PLIF. Conversely, Vamvanij and colleagues²⁶⁷ compared four fusion procedures and found simultaneous anterior interbody and posterior facet fusion to be superior to PLIF, with an 88% fusion rate. Superior fusion rate did not correlate with a better clinical outcome, however, because only 63% of patients in their study experienced a satisfactory result. Other studies have reported even lower success rates with fusion for discogenic back pain; Knox and Chapman²⁶⁸ reported only 35% good clinical results with one-level fusion for IDD.

TLIF involves placement of a pedicle screw and rod construct by which the disc space is then distracted. A complete facetectomy is performed unilaterally, through which the discectomy is performed, the endplate is meticulously prepared, and a structural graft or cage is placed into the interbody space. The laminectomy and bilateral partial facetectomy required for the PLIF approach becomes unnecessary in TLIF, which shortens the operative time; decreases blood loss, risk of conus injury, and dural tear; and minimizes epidural scarring. A TLIF procedure still places the nerve root at risk, however, and because the procedure requires complete unilateral facetectomy, posterior instrumentation is mandated owing to resultant instability. A TLIF approach has the benefits of an interbody fusion—elimination of the potential disc pain generator, improved fusion rate, restoration of disc height, and improved sagittal alignment—and has fewer attendant risks than PLIF and by virtue of the approach decompresses the neuroforamen. Contralateral foraminotomy and decompressive laminectomy can optionally be performed as the patient’s symptoms require.

Lowe and colleagues²⁶⁹ performed a prospective analysis of 40 consecutive patients who had spinal fusion for degenerative diseases of the lumbar spine using unilateral TLIF with pedicle screw fixation. Universal improvement in segmental lordosis, solid fusion in 90% of patients, and excellent or good clinical outcomes in 85% of patients were reported in this study. Whitecloud and colleagues²⁷⁰ reported that the TLIF approach produces greater than $15,000 of cost savings compared with a combined anterior-posterior procedure. There were no major complications noted in either group in this study, and no patient required repeat surgery for a lumbar spinal complication at the authors’ hospital within the 1-year follow-up period.

Minimally Invasive Surgical Approaches

Minimally invasive surgical approaches to lumbar spine fusion have become increasingly popular. Potential advantages include reduced morbidity, less muscle dissection and blood loss, and shortened inpatient hospital stays; however, minimally invasive surgical approaches typically require specific instrumentation and have a larger learning curve. Several of these minimally invasive techniques are described.

Minimally Invasive Transforaminal Lumbar Interbody Fusion and Posterior Lumbar Interbody Fusion

Various, less invasive modifications of the traditional TLIF technique have been described. One technique involves using pedicle instrumentation only on the side of the facetectomy and placing a transfacet screw on the contralateral side (Fig. 45–7). This technique minimizes dissection on the contralateral side and eliminates the risk of screw abutment on the adjacent facet joint. Operative time and cost of the instrumentation are slightly reduced without adversely affecting the rigidity of the construct.²⁷⁸ Slucky and colleagues²⁷⁹ reported on a biomechanical study of constructs using bilateral pedicle screws, unilateral pedicle screws, and unilateral pedicle screws with a contralateral facet screw. These authors found that unilateral pedicle screw constructs allowed significantly increased segmental motion, less stiffness, and off-axis movement, whereas the addition of a contralateral facet screw produced biomechanics similar to bilateral pedicle screw constructs. Constructs using unilateral pedicle screws with a contralateral facet screw are a viable minimally invasive option and reduce instrumentation costs. Sethi and colleagues²⁸⁰ reported that this technique reduces construct cost by nearly 50% and still has a rate of fusion as high as traditional constructs—100% in the authors’ patient population at 9 to 26 months of follow-up.

FIGURE 45–7 A-D, Preoperative T2-weighted sagittal (A) and axial (B) MRI and postoperative anteroposterior (C) and lateral (D) radiographs of a patient who developed adjacent segment degeneration at L5-S1 segment after previous L3-5 posterior spinal fusion. The patient was treated with transforaminal lumbar interbody fusion, ipsilateral pedicle instrumentation, and contralateral transfacet screw fixation.

Other minimally invasive TLIF approaches employ fluoroscopic assisted percutaneous instrumentation (Fig. 45–8). Interbody fusion is typically performed through a paramedian incision with the assistance of a tubular retractor system. Early experience has been positive, but there is a steep learning curve, and the potential for neurologic injury exists. Schwender and colleagues²⁸¹ reported good success with this minimally invasive TLIF approach in a population with mixed diagnoses; mean improvement of ODI was from 46 to 14, and the fusion rate was reported to be 100%.

FIGURE 45–8 A-C, Preoperative T2-weighted sagittal (A) and postoperative anteroposterior (B) and lateral (C) images of a patient with L4-5 and L5-S1 degenerative disc disease treated with TLIF via minimally invasive approach and percutaneously placed transpedicular instrumentation.

Peng and colleagues²⁸² reported on a comparative prospective analysis of 29 minimally invasive TLIF procedures versus 29 traditional open TLIF procedures. Intraoperatively and perioperatively minimally invasive TLIF was associated with less blood loss (150 mL vs. 681 mL) and shorter postoperative stays (4 days vs. 6.7 days) but greater fluoroscopic times (106 seconds vs. 35 seconds) and longer operative times (216 minutes vs. 171 minutes). Both groups had improvement in ODI and back pain at 6 months and 2 years of follow-up, with no significant difference between the groups. Radiographic evidence of fusion was 80% for minimally invasive TLIF and 87% for open TLIF. Similar results were reported by Scheufler and colleagues²⁸³ and Park and Ha.²⁸⁴

Stevens and colleagues²⁸⁵ performed an MRI analysis of patients who had undergone minimally invasive lumbar fusion versus traditional open PLF to determine the difference in effect on the paraspinal musculature. The measured maximal intramuscular pressure intraoperatively was significantly less with use of the minimally invasive tubular retractor systems versus open retractors. There was also significantly less edema of the paraspinal musculature on MRI in patients who underwent minimally invasive fusion, indicating that minimally invasive surgical techniques produce less muscle and tissue damage than traditional open fusion.

Percutaneously placed facet screws have also been described in the literature. Shim and colleagues²⁸⁶ described the use of fluoroscopic assisted percutaneously placed facet screws as a modification of the Magerl technique. These authors reported 11% violation of the laminar wall with no incidences of neural compression; they also reported 15% rate of imperfect pedicle screw placement. Jang and Lee²⁸⁷ reported on the efficacy of circumferential fusion with percutaneously placed facet screws compared with ALIF and PLF with pedicle screws and found no difference between the two groups in regard to operative outcomes. These authors concluded that percutaneously placed facet screws after ALIF are a viable alternative to PLF and pedicle screws.

Extreme Lateral Lumbar Interbody Fusion and Direct Lateral Lumbar Interbody Fusion

Another novel approach to lumbar interbody fusion is XLIF, sometimes called direct lateral lumbar interbody fusion (Fig. 45–9). This procedure involves an anterolateral interbody fusion performed through a transpsoas approach to the lateral aspect of the IVD. This approach was first described by Pimenta²⁸⁸ in 2001 as a modification of the retroperitoneal approach. In 2004, Bergey and colleagues²⁸⁹ described a similar endoscopically assisted transpsoas procedure.

FIGURE 45–9 A and B, Postoperative anteroposterior (A) and lateral (B) radiographs of a patient with L3-4 discogenic back pain treated with interbody fusion through a direct lateral approach.

In the most common synthesis of the anterolateral approach, the XLIF, the patient is positioned in a right lateral decubitus position with the table flexed to open the left side of the disc space. A dilator is guided through the retroperitoneal space to the psoas muscle, and the fibers of the psoas are spread under electromyographic monitoring to avoid damage to the lumbar nerve roots and plexus. After exposure of the disc, the procedure is completed similar to any interbody fusion procedure: by discectomy, endplate preparation, and graft placement. The approach does not require a laparotomy and avoids most of the attendant complications associated with the anterior approach. Exposure is limited by the inferior border of the 12th rib and the superior edge of the iliac crest. The greatest risk during the procedure involves the dissection through the psoas, which is associated with risk of injury to the lumbar nerve roots and the lumbar plexus, most specifically the genitofemoral nerve.

Benglis and colleagues²⁹⁰ published an anatomic study of the lumbar plexus around the psoas muscle to clarify the approach and risk to the neural structures during a direct lateral approach. The nervous plexus has a progressive dorsal to ventral migration from the L1-2 posterior endplate edge, to a ratio of 0.28 of the width of the disc at the L4-5 level. Ventral migration was most pronounced at the L4-5 level, and risk of injury to the lumbosacral plexus is highest at this level if the dissector cannula is placed too far posterior.

Reports of outcomes for the XLIF procedure are sparse, and short-term data only are available. Knight and colleagues^290a reported complication rates of direct lateral interbody fusion in a cohort of 58 patients; 8 of the 13 reported complications were mild and related to the approach, the most common being meralgia paresthetica. Two patients (3.4%) experienced L4 nerve root injury, one of which lasted longer than 1 year. In a separate, smaller series, Ozgur and colleagues²⁹¹ reported no complications with the XLIF procedure.

Materials and Osteobiologics

To assist in fusion and interbody space reconstruction, multiple structural interbody graft and cage options have been designed. Common materials include machined allograft bone, titanium cages, reinforced carbon fiber, polyetheretherketone (PEEK), and bioresorbable cages. Several cage designs exist, including cylindric, tapered, impacted, lordotic, biconvex, and boomerang-shaped. There is a paucity of randomized controlled trials comparing clinical results between the different designs. A review of the multiplicity of structural grafts and cages currently being marketed is beyond the scope of this chapter.

To improve fusion rates, various graft materials have been studied, including iliac crest autograft, morcellized local laminectomy bone, allograft bone, and several graft extenders. With the advent of bone morphogenetic proteins (BMPs), spinal fusion surgery has entered a new era. Multiple BMPs have been discovered, most of which are members of the transforming growth factor superfamily. BMP-2 and BMP-7 have been studied extensively and currently have FDA approval for limited use in humans.

Use of BMP has many reported benefits. Results for fusion rates so far have been excellent,^300,301 and reduction in pseudarthrosis rates results in decreased requirement for costly revision surgeries. Some centers have attempted placing BMP through the PLIF or TLIF approach; however, the benefit of this use is unclear.³⁰² BMP seems to obviate the need for autograft bone grafting, eliminating the risk of donor site morbidity, decreasing operative time, and decreasing blood loss. Donor site morbidity associated with iliac crest autograph harvest is significant. Sasso and colleagues³⁰³ combined results from four randomized trials comparing iliac crest graft and recombinant human BMP-2 (rhBMP-2) as part of an ALIF procedure. The investigators evaluated 208 patients’ VAS score for intensity and frequency of pain. At 2 years’ follow-up, 31% of patients were still reporting persistent pain at the donor site, and 16% reported fair or poor appearance of the graft site.

Multiple randomized controlled trials have been performed on BMP-2; most have evaluated the effectiveness of the protein compared with iliac crest autograft. Dimar and colleagues³⁰⁴ published the results of a randomized prospective trial with 98 patients who underwent single-level PLF. Fusion rate at 2 years was 73% in the iliac crest autograph group and 88% in the rhBMP-2 group. Average operating time was 21% longer and blood loss was 81% greater in the autograft group compared with the rhBMP-2 group. No significant difference in outcome measures was noted between the groups at any of the follow-up intervals. A later study³⁰⁵ by the same group found decreased operative time and blood loss and higher rates of fusion but no difference in outcome measures. Fusion results in single-level ALIF with allograft dowels were similar as reported by Burkus and colleagues:³⁰⁶ 100% in the rhBMP-2 group and 81.5% in the autologous bone graft group. A 6-year follow-up of patients in the Burkus study³⁰⁷ projected a worst-case scenario fusion rate of 91%, considering reoperation rate for pseudarthrosis within the 6-year postoperative interval.

Vaccaro and colleagues³⁰⁸ compared OP-1/BMP-7 with iliac crest autograft in PLF. This was the first large (335 patients), randomized controlled trial of BMP-7 for use in spine fusion. At 3 years’ average follow-up, there was no significant difference between the two groups with respect to outcome measures, success of fusion, reoperation rate, complications, and fusion. The iliac crest autograft group had statistically longer operative times and blood loss. Vaccaro and colleagues³⁰⁸ concluded that OP-1/BMP-7 was equivalent to iliac crest bone graft in PLF.

The use of BMP is not without complications.³⁰⁹ BMP used in minimally invasive TLIF is reported to be associated with postoperative radiculitis.³¹⁰ Use of BMP is also linked to postoperative bleeding and seroma formation, heterotopic bone formation,³¹¹ and osteolysis. Rates of vertebral osteolysis associated with use of BMP-2 in interbody cages have been reported to be 5.8%.³¹² Two unpublished cases of vertebral osteolysis at the authors’ institution have resulted in cage dislodgment.

The high cost of BMP has led some authors to caution against routine use, whereas others report that the decreased risk of pseudarthrosis and associated revision surgery justifies the expense in certain patient populations. Glassman and colleagues³¹³ reported on the success of rhBMP-2 in single-level posterolateral lumbar fusions in smokers. In the rhBMP-2 group, 20 of 21 patients (95%) achieved fusion, and in the iliac crest autograph group, 16 of 21 patients (76%) achieved fusion; however, the authors noted that other clinical outcomes were still adversely affected by smoking independent of fusion status. Carreon and colleagues³¹⁴ published a cost utility study on BMP-2 versus iliac crest autograft in elderly patients. At 2 years’ follow-up, after accounting for complications, including a higher rate of nonunion (9.6% in the autograft group) and revision surgery, the mean cost in the rhBMP-2 group was $39,967 and in the iliac crest autograph group was $42,286.

A British meta-analysis published by the National Institute of Health Research (NIHR) Health Technology Assessment Programme³¹⁵ evaluated the cost-effectiveness of BMP in spinal fusions. Garrison and colleagues³¹⁵ found evidence that BMP-2 was more effective than autogenous bone graft for radiographic fusion in single-level DDD. BMP was also associated with decreased operative time, improved clinical outcomes, shorter hospital stays, and fewer secondary interventions. The probability that BMP was cost-effective in Great Britain, based on a cost per quality-adjusted life-year of less than £30,000 (roughly $50,000), was only 6.4%, however. Garrison and colleagues³¹⁵ concluded that although BMP improved outcome measures and decreased morbidity associated with autologous graft harvesting, its use was not cost-effective.

The development of osteobiologics such as BMPs has been an exciting advancement in recent years. More substances are likely to be developed in coming years. Several other proteins, such as GDF-5, TIMP, TGF-β, IGF-1, and SOX-9, are already being investigated for their potential application in spinal fusion and disc regeneration.

Dynamic Spinal Stabilization Techniques

Dynamic techniques for spinal stabilization have been described in the literature for many years, but newer techniques and instrumentation designs have made this an area of renewed interest. Several authors have attempted to use soft tissue stabilization techniques that restrict rather than completely eliminate motion, while still relieving mechanical back pain. One of the first dynamic systems was described by Graf³²⁵ and involved an artificial ligament reconstruction using four pedicle screws and two braided polyester bands to stabilize the painful segment in lordosis. Proponents of the technique state that pain is reduced by coaptation of the painful facets; posterior annular compression, which closes annular tears; and stabilization of the motion segment. Theoretically, the fixation relaxes over the first 6 months postoperatively allowing a return of motion after healing has occurred.³²⁶

Outcomes for Graf ligamentoplasty in the literature are sparse and mixed. Kanayama and colleagues³²⁷ reported on 10-year follow-up, with preservation of segmental motion in 80% of patients and improvement in VAS scores. Other studies have reported poor outcomes at 1 year and a high rate of revision at 2 years. Revision for ligamentoplasty has been found to have poor outcomes—similar to the outcomes seen after revision arthrodesis.^328–331 One of the biggest arguments against the procedure has been that the Graf ligament significantly restricts flexion at the spinal segment, which increases load in the problematic posterior anulus of the disc.

Newer techniques and designs, such as the dynamic neutralization system for the spine (Dynesys Dynamic Stabilization System; Zimmer, Warsaw, IN), attempt to reduce movement equally in flexion and in extension. The Dynesys system consists of titanium pedicle screws connected by an elastic band, which controls motion in a more consistent manner than the Graf ligamentoplasty. Although an improvement over the Graf technique, the degree to which the Dynesys system successfully unloads the disc is still unpredictable.³³² Reported outcomes by Grob and colleagues³³³ were poor with only half of patients achieving improved quality of life and less than half experiencing functional capacity improvement. These authors concluded that there was no support for superiority of dynamic stabilization to typical arthrodesis. Few randomized controlled clinical studies have been conducted with this technique, and long-term efficacy is not yet clearly established.³³⁴

Disc Arthroplasty

Another technique for maintenance of segmental motion while treating symptomatic disc disease is TDA. Interest in TDA was inspired by the resounding successes of total hip arthroplasty and total knee arthroplasty in restoration of function and resolution of pain. The goal of TDA is to remove the discal pain generator while maintaining segmental height, stability, and motion. The potential benefits of disc arthroplasty are twofold: (1) Healing does not require fusion, removing the risk of pseudarthrosis, and (2) motion is preserved, which theoretically reduces the risk of adjacent segment degeneration.

The first implanted lumbar total disc replacement prosthesis in the United States was the SB Charité III (DePuy Spine, Raynham, MA). There are now two FDA-approved devices for lumbar disc arthroplasty: SB Charité approved in October 2004 and ProDisc-L (Spine Solutions/Synthes, Paoli, PA) approved in August 2006. At least two other disc replacement systems are undergoing FDA trials: Maverick (Medtronic Sofamor Danek, Memphis, TN), and FlexiCore (Stryker Spine, Allendale, NJ).

Although TDA designs have been used in Europe for quite a while, long-term results are still in dispute. The Charité device has the longest term information based on randomized controlled data. Guyer and colleagues³³⁵ reported on 5-year follow-up for 90 TDA patients in a noninferiority study compared with a control group of 43 ALIF patients. ODI and SF-36 data between the groups was comparable, but 66% of the TDA patients versus 47% of the ALIF patients were back to full-time work at the 5-year interval. Range of motion for the Charité device at 5 years’ follow-up was reported to be 6 degrees.

Blumenthal and colleagues³³⁶ published clinical results of a randomized controlled trial of the Charité disc and reported noninferiority to ALIF controls. Radiographic examination of study patients showed better restoration of disc height and less subsidence in TDA versus ALIF with BAK cages (Spine tech, Minneapolis, MN).³³⁷ Tropiano and colleagues³³⁸ had 75% excellent or good results at an average 8.7 years of follow-up. Lemaire and colleagues³³⁹ reported 90% good or excellent outcomes and 91.5% return-to-work rate at a minimum of 10 years of follow-up in 147 patients.

Randomized controlled device trials of ProDisc-L³⁴⁰ have also shown noninferiority to spinal fusion. Outcomes of TDA in this study show a statistically significant advantage of arthroplasty compared with fusion. Improvement criteria (≥15%) in ODI, VAS, and SF-36 were met in 77% of TDA patients versus 64% of fusion patients. Favorable initial randomized controlled clinical trial results have also been published in support of the FlexiCore system.³⁴¹

Proponents of TDA suggest that preservation of motion at the operated level decreases the incidence of adjacent segment degenerative disease associated with fusion. One of the purported benefits of this technique is for patients who would otherwise require multilevel fusion because of asymptomatic or subacute adjacent level degenerative changes. A significant problem with fusion procedures is the risk of developing adjacent segment degeneration; however, in some reports TDA devices have been associated with increased adjacent segment degenerative changes. Kostuik³⁴² reported that two of the most common complications necessitating revision surgery in TDA were facet degeneration and adjacent level disease.

Park and colleagues³⁴³ reported on radiographic evidence of adjacent level degeneration after ProDisc-L implantation; at 26 months postoperatively, progression of facet degeneration was noted in 29% of 32 patients. Adjacent level progression was positively correlated with female gender, malposition of the prosthesis, and two-level disc replacement. In a comparative study between Charité and ProDisc-L, Shim and colleagues³⁴⁴ found that although the clinical outcomes of both systems were good, the facet joints at the operative level (32% to 36%) and the discs at adjacent levels (19% to 29%) showed advancement of degeneration at 3 years’ follow-up regardless of device used. Harrop and colleagues³⁴⁵ published a systematic literature review comparing published reports of adjacent segment degeneration in lumbar fusion with TDA. In the fusion group, 173 of 1216 patients (14%) developed symptomatic adjacent segment disease compared with 7 of the 595 TDA patients. This study had the benefit of large patient numbers, but none of the published reports used for the analysis were of a high level of evidence.

Another problem with TDA systems is that it is very difficult to create a mechanical device that can mimic all of the properties of the native disc. Several patients who underwent TDA with good pain relief have been found to have experienced spontaneous fusion secondary to heterotopic bone formation. In some reports, heterotopic ossification and spontaneous fusion have been unusually high in cervical disc replacement.³⁴⁶ In lumbar disc replacement, Huang and colleagues³⁴⁷ reported a 13% rate of heterotopic ossification in 65 patients, and 2 patients went on to spontaneous fusion. These authors cited preoperative ossification of the anulus, bony endplate injury, component malposition, and subsidence as potential factors leading to postoperative heterotopic ossification. Tortolani and colleagues³⁴⁸ reported that the rate of heterotopic ossification in 276 Charité patients was only 4.3% and that regardless of the presence of heterotopic ossification, postoperative range of motion was still better than preoperative range of motion in all cases.

Another possible problem with TDA is the potential for a failed prosthesis to cause catastrophic complications and make revisions extremely complex. Reported reasons for failure in the literature include acquired spondylolysis,³⁴⁹ implant subsidence,³⁵⁰ implant loosening, malposition, displacement, early wear, and infection.³⁴² TDA failure by catastrophic wear, similar to that seen in total joint arthroplasty, is a potential concern, but wear rates and behavior of polyethylene debris around the spine are still largely unknown. Managing revision surgeries has the potential to become incredibly complicated dealing with repeat anterior exposures, osteolysis, prosthesis subsidence, and bone loss. In a feasibility study of revision of the Charité device, McAfee and colleagues³⁵¹ reported that TDA did not preclude further procedures at the index level of surgery. Of 589 TDA patients, 52 required revision (9%) compared with 10 of 99 ALIF patients (10%). Through a repeat anterior retroperitoneal approach, 22 of 24 TDA devices were removed successfully. Seven of the 24 (29%) removed discs were revised to another Charité device.

Another concern with TDA is the cost of the device and procedure. Levin and colleagues³⁵² published a charge analysis of one-level and two-level TDA with ProDisc-L versus circumferential fusion. For one-level fusion, average implant cost was nearly identical at $13,990 for fusion and $13,800 for TDA. Operative time for TDA was about one half of the average time for fusion (185 minutes vs. 344 minutes). Total charge for a single level averaged $35,592 for TDA and $46,280 for fusion. For two-level procedures, implant costs were much less for fusion ($18,460) than for TDA ($27,600), and operative times were 242 minutes versus 387 minutes. In a smaller retrospective review, Patel and colleagues³⁵³ compared TDA with fusions and found that implant costs for TDA, ALIF, and TLIF were similar if rhBMP-2 was excluded from the analysis.

Success for TDA has been found in multiple studies with up to 10 years follow-up to be at least equivalent to fusion. Concern for complications specific to TDA exist, such as heterotopic bone formation, spontaneous fusion, spondylolysis, catastrophic failure, displacement, implant subsidence, and potential for complicated revisions. Preliminary studies have shown revisions to be feasible for short-term periods, but long-term effects of polyethylene wear behavior and osteolysis are still unknown. From a cost standpoint, TDA seems to compare favorably with fusion techniques for one-level disease even in the absence of BMP use; however, two-level TDA seems to be more expensive than two-level fusion. TDA offers yet another option for operative treatment of IDD and DDD, but long-term results are yet to be determined.

Nuclear Replacement

Another concept for treating IVD disease involves replacing the nucleus pulposus while retaining the anulus intact. The objective essentially is to reinflate the nucleus. This concept allows for the design of smaller prostheses that can be implanted via a minimally invasive approach. Implant failure would theoretically be less devastating than failure associated with total disc replacements.

The first nuclear replacement procedures were performed by Fernstrom³⁵⁴ in the late 1950s. This technique involved an annulotomy, resection of the nucleus pulposus, placement of a steel ball bearing (dubbed the Fernstrom ball), and preservation of the anulus. Fernstrom³⁵⁴ claimed outcomes similar to fusion, but application of this technique became associated with unacceptable rates of implant subsidence.

Multiple modern nuclear replacement designs currently are under development and investigation, which fall under two main device types. The first type is a mechanical design and includes devices such as the original Fernstrom ball, and newer designs composed of PEEK and pyrolytic carbon. The second type is an elastomeric design usually made with either preformed or injectable materials such as polyurethane, silicone, and various other polymers.³⁵⁵

The Prosthetic Disc Nucleus (PDN) (Raymedica, Bloomfield, MN) is the most studied nuclear replacement device. Klara and Ray³⁵⁶ published a series of 423 patients treated with PDN since 1996. They reported a 90% survival rate and 10% rate of device explantation. Initially, the study was plagued with a high device migration rate, but newer designs have had improved results. Shim and colleagues³⁵⁷ reported 78% good results in 46 patients followed for longer than 6 months; 4 patients required revision surgery because of migration of the implant. Ahrens and colleagues³⁵⁸ published a 2-year prospective outcome study on the DASCOR device (Disc Dynamics, Eden Prairie, MN) involving 85 patients. These authors reported significant improvement in VAS and ODI outcome measures. The rate of explantation was 8%, most commonly for resumption of severe back pain. The outcome data for nucleus pulposus replacement technologies are short-term, and the technique is still in need of careful investigation.

Future Directions

Great interest has been generated in recent years for the development of biologic repair strategies and biopharmaceutical approaches for interception and prevention of the degenerative disc cascade. Several investigators^359,360 have reported successful transplantation of disc chondrocyte cells in animal studies with regeneration of viable matrix and normalized distribution within the disc space. Nomura and colleagues³⁶¹ hypothesized that the extracellular matrix may play a role in slowing the rate of disc degeneration. Meisel and colleagues³⁶² published a canine model for autologous chondrocyte transplant repair of damaged IVDs. A randomized controlled trial is under way comparing chondrotransplant DISC with discectomy.

Another approach under investigation is the use of various growth factors alone or in combination to induce disc regeneration. Metabolically impaired cells in the IVD that exhibit degenerative or age-related changes have been shown to repair their own matrix and disc structure under the influence of OP-1/BMP-7. BMP-7 seems to have an anabolic effect on proteoglycan and collagen synthesis, particularly within the nucleus pulposus. Animal studies have been performed using BMP-7, and GDF-5 with results showing restoration of disc height, increased extracellular matrix, and increased proteoglycan synthesis,^363–365 but human studies have not been performed yet.

Manipulation of genes that regulate the synthesis of specific RNA and protein molecules on a cellular level also has promising applications for prevention of DDD and disc regeneration. Gene delivery, typically by a viral vector, provides for local production of sustainable, high concentrations of the gene product for extended periods. Targeted delivery of a gene product maximizes therapeutic potential, while minimizing side effects. Endogenously produced proteins may also have greater biologic activity than exogenously administered recombinant proteins.³⁶⁶ The IVD is relatively avascular and has poorly characterized, slowly dividing cells, so injection of viral vectors could potentially be maintained for long periods in this encapsulated and immunoprotected environment.

Because of safety concerns, ex vivo methods of gene therapy with the help of a bioreactor or tissue scaffold may be preferable. Ex vivo studies showing successful incorporation, increased matrix production, and restoration of the IVD structure have now been published for multiple gene products, including SOX-9,³⁶⁷ TIMP-1,³⁶⁸ TGF-β1, IGF-1, and BMP-2.³⁶⁹ There are still several concerns, however; for example, the virus may leak through annular fissures in the degenerated disc and evoke an immune response. An ideal treatment program for DDD should also allow for repetitive administration of gene therapy injections at the same or different disc levels.³⁷⁰

Regardless of the method of disc regeneration, significant challenges remain. Regeneration of a severely degenerated disc may be impossible because the environment within the degenerated disc may be too hostile secondary to endplate sclerosis and poor tissue nutrition. Also, an unstable spinal segment would make regeneration difficult because continued abnormal loading during the healing phase would likely lead to failure.