Pathophysiology

The normal lumbar intervertebral disc consists of fibrocartilaginous tissue designed to absorb and dissipate load applied to the spinal column. The two components of the disc are the nucleus pulposus and the anulus fibrosus. The nucleus pulposus is composed of proteoglycan aggrecan molecules with 70% to 80% water content. Absorption of water into the nucleus provides disc height and resistance to compression. With loading, water defuses out of the disc, and subsequent reabsorption occurs with unloading. The anulus is an interlacing collagen network that provides tensile strength in axial rotation. With bending or compression of two adjacent vertebrae, the nucleus pulposus changes volumetrically, causing bulging of the disc away from the internal axis of rotation. The anulus also functions to limit and contain expansion or herniation of the nucleus.

With aging, the disc gradually becomes less hydrated, and the concentration of proteoglycans decreases. Normal disc metabolism shifts toward catabolic processes, which further deplete proteoglycans and lead to increasing matrix degeneration. As a result, the disc becomes progressively dysfunctional as the nuclear material is replaced by desiccated fibrocartilaginous material. Loss of fluid results in decreased hydrostatic pressure as a mechanism for effective load transference. Thinning or microfracture of the end plates can occur, and subsequent loss of end-plate vascularity reduces transport of nutrients and waste products out of the disc. Eventually, with cyclic loading of the degenerated disc, radial fissures or cracks propagate through the anulus, with migration of nuclear material peripherally. With complete anular disruption, disc material can herniate into the central canal, lateral recess, or foramen. These degenerative processes are estimated to occur in 90% of normal individuals by 50 years of age.¹⁰

In 1970, Crock first associated back pain with the pathophysiologic process of disc desiccation and subsequent radial fissure formation of the anulus.^11,12 He termed this entity internal disc disruption, which was characterized by the progressive disruption of the internal architecture of the disc while essentially maintaining the external shape such that nerve root compression did not occur. Crock hypothesized that pain is generated when degradation of the disc matrix causes release of inflammatory cytokines, which then migrate through the disrupted inner anular fibers to irritate the high concentration of sensory nerve endings in the outer anulus. His conclusion was supported radiographically by the relative absence of any nerve root compression but the high correlation of concordant pain in discs that exhibited severe radial fissures with intradiscal contrast injection.

Since then, several theories regarding the relationship between degenerative disc disease and pain generation have developed. The mechanical theory suggests that degeneration results in alteration in the biomechanical properties of the disc. As the disc degenerates and the anulus becomes disrupted, increasing instability occurs at the motion segment. Therefore, with normal physiologic loading, the motion segment responds with excessive compression, bending, or rotation, which can trigger pain transmission in surrounding nociceptors. CT and MRI studies have quantified the response of the lumbar spine to rotatory torque and have correlated increased axial rotation in degenerated discs with pain provocation on discography.^13,14 Also as the disc desiccates, it loses hydrostatic pressure, so with normal physiologic loading, more stress is transferred to the anulus and the end plate, where pain-sensitive nerve fibers are in high concentration. Increased stress to the end plate can lead to end-plate fracture and disc herniation into the vertebral body, which may further propagate pain generation.

The chemical theory suggests that catabolism of the disc results in release of proinflammatory chemical mediators. Nitric oxide, phospholipase A₂, prostaglandin E, matrix metalloproteinases, and other cytokines have been implicated as chemical agents that infiltrate through radial fissures to irritate nociceptors that are present in the outer aspect of the anulus and the end plate. Proteoglycan breakdown is known to have a high concentration of the neurotransmitter glutamate, which may in turn stimulate glutamate-specific receptors in the dorsal root ganglion, resulting in back or radicular pain in the absence of nerve root compression. Nociceptors are also known to be present in high concentrations elsewhere within the spinal canal, such as the posterior longitudinal ligament, dura, and blood vessels.

Alternatively, although the disc may demonstrate signs of degeneration, pain may arise from other associated structures. The facet joints, ligaments, fascia, nerve roots, and dura are capable of transmitting pain. While disc degeneration may be the initial inciting pathology, one or more of these structures may in fact be the source of pain. Progressive disc disease results in increased load transference to surrounding structures such as the facet joints, ligaments, and paraspinal muscles, which may eventually exceed their capacity for resistance. Cyclic loading to these structures can lead to increased arthropathy, ligamentous hypertrophy, and muscle fatigue, which may contribute to pain. The medial branch of the dorsal primary rami courses around the facet and innervates the joint capsule and may be a particularly pain-sensitive fiber signaling back pain with increased stress. Diagnostic blockade of various spinal and paraspinal structures with injected anesthetic may be performed to evaluate certain areas as potential pain generators. Studies performed in patients with similar presentations of LBP have demonstrated a wide range of sources of pain, including the disc, facet joints, and sacroiliac joints. Therefore, while the degenerated disc may be implicated in the pathophysiology of LBP, it remains unclear whether the disc itself or other surrounding structures are the actual source of pain.

Diagnosis

Patients with axial LBP generally present with a deep, aching pain localized to the lower back, sacral, or gluteal region. The pain is characteristically worsened with mechanical activities such as bending, twisting, or lifting and is relieved with recumbency. Pain is also increased with lumbar flexion, such as when ascending stairs, and is presumed to be secondary to increased ventral loading on the degenerated discs. Also, patients often describe increased pain with prolonged sitting. Interestingly, patients may also describe buttock or leg pain (generally limited to above the knee) that is not radicular in nature but may be referred sclerotomal pain.

Physical examination rarely provides information in facilitating the diagnosis of axial LBP or the level of disease. Often, the physical and neurologic assessment is normal. The clinical history and the physical and neurologic examination, however, are critical in evaluating other potential etiologies of LBP such as nerve root compression, spinal deformity, fracture, spinal instability, spondylolisthesis, tumor, or infection. Masqueraders that closely mimic axial LBP symptomatology and that can be assessed on physical examination include myofascial pain, sacroiliac joint pain, piriformis syndrome, and hip osteoarthritis.

In light of a normal physical examination and relatively ubiquitous symptomatology, the diagnosis of axial LBP is made clinically using various radiologic studies. Because of the overt dependence on imaging for determining the source of pain, there has been intense scrutiny of radiologic criteria for identifying which discs are painful and which patients may ultimately benefit from fusion surgery.

Plain Radiographs

Plain radiographic findings in patients with axial LBP may demonstrate characteristics consistent with degenerative disc disease. While radiography does not visualize the soft tissue disc, plain films may reveal decreased disc height consistent with a collapsed or dehydrated disc (Fig. 58-1A). Sclerotic end plates or bone-on-bone appearance are commonly seen with severely degenerated discs (Fig. 58-1B). Plain radiographs can be performed with patients in weight bearing, flexion, or extension to demonstrate the alignment of the spine and the nature of the motion segments under normal physiologic loading. The presence of hypermobility or malalignment under such stresses may help to identify which levels are symptomatic or may suggest other pathologic processes.

FIGURE 58-1 A, Lateral radiograph demonstrating a degenerated, collapsed L5-S1 intervertebral disc space. B, Lateral radiograph demonstrating a degenerated L3-4 intervertebral disc space with sclerotic end plates.

In evaluating lumbar plain radiography in symptomatic patients, Scavone et al. observed that radiographs alone were uniquely diagnostic in only 2.4% of patients.¹⁵ Liang and Komaroff found that lumbar radiographs did not provide diagnostic value in differentiating patients with acute versus chronic LBP.¹⁶ Coste et al. reported that there was a high degree of variability in interpretation of plain films performed in LBP patients, underscoring the poor diagnostic value of these studies for this condition.¹⁷ Many clinicians conclude that the degenerative findings that are seen on lumbar plain radiographs may in fact represent normal age-related changes and as such provide little information with regard to differentiating symptomatic degenerated discs from asymptomatic age-appropriate discs. Plain radiographs, however, are useful in the assessment of axial LBP for effectively ruling out other etiologies of back pain. Fractures, osteomyelitis, tumor, spondylolisthesis, and deformity can often be quickly assessed with lumbar weight-bearing or 36-inch standing plain films. Therefore, plain radiographs are indicated in LBP patients who are of pediatric age, are at high risk for osteoporosis, have a history of prior surgery, present with neurologic deficit or gross deformity, or have clinical signs suggestive of trauma, infection, or malignancy.

Discography

The ability of discography to diagnose degenerated disc disease and identify painful symptomatic discs has been the subject of ongoing debate. Discography is an invasive procedure by which a needle is placed percutaneously into the nucleus pulposus under fluoroscopic or CT guidance. Contrast is injected into the nucleus, and images are generated on plain radiographs and/or axial CT images. The flow of contrast within the disc provides information regarding disc morphology and the integrity of the anulus. Contrast filling of a degenerated, desiccated disc may highlight a collapsed narrow disc space. Radial fissures may be revealed by contrast leakage from the nucleus into the periphery and, in the case of anular disruption, spilling of contrast into the canal or foramen. The integrity of the anulus can be inferred from the resistance or volume of injection, normal discs having a high resistance and smaller volume of injection. Low resistance or a high volume of injection suggests disruption or a possible tear in the anulus.

One unique feature of discography, in contrast to other conventional passive imaging modalities, is the behavioral component of the diagnostic procedure. Essential to discography is the evaluation of both the suspected symptomatic disc and control or asymptomatic discs. With injection of contrast into an abnormal disc, the discographer assesses for provocation of concordant pain. Pain with disc injection is likely due to either an increase in intradiscal pressure or the displacement of biochemical agents. With pressurized contrast injection, stretching of pain-sensitive nerve endings in the anulus may be stimulated, or stress may be transmitted to the end plates or vertebral body. Alternatively, contrast injection into the nucleus may cause displacement of inflammatory cytokines, which then diffuse through radial fissures to the periphery, where they may stimulate nociceptors in the outer anulus or the surrounding environment. Ideally, the provoked pain is similar to or an exact reproduction of the patient’s presenting back pain. Otherwise, the injection provokes dissimilar pain or does not produce pain, and the discogram is inconclusive or negative for that level. The gold standard for diagnostic discography is exact reproduction of pain at the morphologically abnormal level with no pain provocation at control or normal discs.

Several studies provide evidence to support discography as an effective method for identifying degenerated, symptomatic discs. Cadaveric studies have shown a strong correlation between discographic contrast distribution and the severity of degeneration on gross examination. There is also a good correlation between disc morphology on discography and intraoperative findings of disc protrusions and herniations. Importantly, studies have identified morphologic findings on discography that correlate with concordant pain provocation. As expected, pain provocation is most commonly observed in discs that demonstrate dorsal anular tears (65.3%) as compared to simply degenerated discs (36.6%), internal anular disruption (20%), or intraosseous disc herniations (0%).¹⁸ The use of axial CT imaging provides even further morphologic assessment of the disc. Using the Dallas Discogram Description (DDD) for CT-based discography, severe and moderately graded degenerated discs show a strong correlation with exact reproduction of pain, anular disruption being the best predictor of concordant pain.¹⁹ McCutcheon and Thomas found that contrast tracking to the periphery of the anulus suggesting radial fissures and anular disruption has an 87% correlation with concordant pain.²⁰

Subsequent studies, however, have failed to consistently reproduce these positive correlations. Several studies have demonstrated pain production with discography in otherwise normal-appearing discs in asymptomatic individuals. Carragee et al. studied discography in a group of patients who did not have back pain but had chronic pain related to iliac crest bone harvest for non–lumbar spine surgery. Of these patients, 37.5% had similar or exact reproduction of donor site pain with lumbar discography. Some have proposed that high-pressure injection is responsible for false-positive pain production in normal discs. However, even with low-pressure discography, pain was produced in the same number of asymptomatic volunteers as LBP patients.²¹ Vanharanta et al. found that morphologically normal discs on DDD grading still provoked some pain response in 24% of individuals.²² Also, they observed that severely degenerated discs resulted in exact reproduction of pain in only 22% of individuals. In their study, even discs with severe anular disruption had exact pain reproduction in only 36% of subjects. The reverse association demonstrated slightly better correlation. Patients who had a positive pain response were also more likely to have evidence of disc degeneration. However, discs with exact reproduction of pain were distributed among a range of DDD grades from slight (20%) to moderate (39%) to severe (37%). The best correlation among patients with exact reproduced pain was for severe anular disruption (77%). Overall, the investigators found that painful discs tend to have higher degrees of degeneration and anular disruption compared to painless discs. However, all discs as they deteriorated were more likely to provoke pain, although often the pain was not similar to the presenting back pain and therefore the degenerating disc could not be conclusively diagnosed as the symptomatic disc. Further adding to diagnostic uncertainty, studies have shown that nonspinal factors such as abnormal psychometric findings, chronic pain states, and disputed compensation claims also strongly correlate with positive discography.²¹

Despite these conflicting studies, many clinicians have adopted discography as an instrument for presurgical screening and patient selection, with the presumption that positive discography can reliably predict which patients and levels will respond favorably to fusion. Specifically, discs that demonstrate abnormal morphology and/or concordant pain provocation are likely the symptomatic levels. Therefore, fusion across the positive discographic motion segment will result in alleviation of pain. Conversely, discs with normal morphology or discs that do not reproduce similar pain can be reasonably excluded from surgery or the fusion construct. As a result, for many surgeons, discography is assigned a critical role in selecting which patients with chronic back pain are surgical candidates and ultimately in determining which levels to fuse.

Certain studies have demonstrated that positive discography reliably predicts good surgical outcomes. Simmons and Segil studied patients who had discography prior to undergoing lumbar discectomy, discectomy and fusion, or fusion alone. They found that preoperative discography demonstrated 82% diagnostic accuracy in identifying the symptomatic level.²³ This study, however, represented a heterogeneous patient population that included not only back pain patients, but also those suffering from herniated discs and nerve root compression. Colhoun et al. found that among patients undergoing lumbar fusion, 89% of those with a positive preoperative discogram had significant improvement postoperatively, including decreased pain, return to work, and cessation of analgesics.¹⁸ Patients with nondiagnostic preoperative discography had a lower rate of success after lumbar fusion, only 52% of patients reporting a similar satisfactory postoperative outcome.

Varying degrees of success with preoperative discography have been observed. Good clinical outcomes have been demonstrated in 64% to 86% of patients with positive discography who undergo anterior lumbar interbody fusion (ALIF).^24–26 Other studies have shown that more than 90% of patients with positive discography improve after posterior lumbar fusion.^27,28 Derby et al. argue that better correlation is observed when chemically sensitive discs are identified on preoperative discography.²⁹ Chemically sensitive discs provoke concordant pain under particularly low-pressure injection, suggesting that pain is generated by the displacement of biochemical agents that then stimulate sensory nerve endings in the outer anulus. Therefore, Derby et al. hypothesize that patients with chemically sensitive discs require complete discectomy with thorough removal of the offending disc for pain relief. Among patients with chemically sensitive discs, successful clinical outcome was observed in 89% of patients who underwent discectomy and interbody fusion, compared to only 20% of patients who underwent dorsolateral fusion alone and 12% of patients who were treated nonoperatively. Similarly, Weatherly et al. used discography to identify painful, symptomatic discs within a fused segment in patients with persistent LBP after posterior lumbar fusion.³⁰ Subsequent ventral discectomy and interbody fusion of the positive discographic levels resulted in complete resolution of pain.

The positive predictive value of discography for success after lumbar fusion, however, has not been borne out through multiple repeated studies. Of particular concern is that the potential for a high false-positive rate may lead to an inappropriate rise in fusion surgery and consequently an unacceptable rate of unsatisfactory outcomes. Carragee et al. evaluated patients with positive single-level discography using low-pressure injection who then underwent lumbar fusion of the abnormal disc.³¹ They observed that only 27% of patients had a highly effective outcome as defined by a visual analogue score (VAS) of 2 or less, Oswestry Disability Index (ODI) of 15 or less, full return to work, and cessation of narcotics and analgesics. A minimal acceptable outcome of VAS of 4 or less, ODI of 30 or less, no narcotic use, and at least some gainful employment was reached in only 43% of patients. The authors concluded that in the best case calculation, the adjusted positive predictive value for a minimal acceptable outcome was only 55%. Other studies have similarly found less promising results, with successful outcomes in only 35% to 46% of patients who had undergone lumbar fusion with discography as the primary diagnostic tool.³² It should be noted, however, that in one of these studies, a particularly low arthrodesis rate was observed (47.9%), which may account for the unexpected poor outcomes.

Overall, discography remains an imperfect instrument for diagnosing and localizing discogenic pain. Particularly, discography has come into doubt as a reliable measure for predicting which patients and what levels will respond well to fusion surgery. Much of this lack of accuracy and consistency may be dependent on discography technique and reporting of pertinent positive and negative findings. Furthermore, patient, discographer, and surgeon expectations may bias toward false-positive results, potentially leading to increasing the number of surgeries or creating unrealistic prospects for successful results. Ultimately, the degree to which discography plays a role in surgical decision making largely depends on the surgeon’s prior experience in drawing upon this diagnostic modality and also establishing clear communication with the patient with regard to reasonable expectations for outcome. Given the available data, discography is best indicated for correlating concordant pain in discs that are morphologically abnormal, as the finding of pain provocation in otherwise normal-appearing discs appears to be clinically irrelevant. Discography may also facilitate assessing prior to fusion whether levels adjacent to the symptomatic level are also abnormal and may be included in the fusion construct. Last, in certain circumstances, discography may play a role in evaluating patients who have persistent back pain after posterior lumbar fusion to assess for a painful pseudarthrosis or the presence of a symptomatic disc within the fused segments.

Magnetic Resonance Imaging

MRI is the radiologic study of choice for visualizing the soft tissue structures of the spine, including the discs, ligaments, joints, and neural elements. As a result, MRI is the preferred test for nerve root compression from disc herniation or lumbar stenosis. MRI is also well equipped for clearly visualizing the intervertebral disc. Besides the ability to image in multiple planes, MRI, with good-quality spin-echo T2-weighted images, provides excellent characterization of the morphology of the disc and superb differentiation between the nucleus pulposus and anulus. The signal intensity within the nucleus is related to the concentration of water in the proteoglycan matrix. Therefore, a reduction in signal intensity correlates with matrix degradation and disc degeneration (Fig. 58-2A). The ability of MRI to detect loss of water content and disc desiccation has prompted consideration of MRI as a sensitive measure of degenerative disc disease.

FIGURE 58-2 A, T2-weighted sagittal MRI demonstrating a severely degenerated L5-S1 disc with abnormal signal changes in the adjacent vertebral body bone marrow. B, T1-weighted sagittal MRI demonstrating decreased signal in the adjacent vertebral bone marrow consistent with vascularized, fibrous tissue.

MRI also well characterizes the effect of disc degeneration on the adjacent end plates and vertebral bodies. With advanced degeneration of the discs, greater load is transferred to the end plates. In the early phase, the normal vertebral body bone marrow is replaced with vascularized fibrous tissue as a reparative response to injury.³³ This appears on T1-weighted MRI as decreased signal (Fig. 58-2B), and conversely increased signal on T2-weighted images. With chronic degeneration, the normally red bone marrow is converted to yellow marrow as the marrow elements are replaced by fat cells, a situation that appears to represent a chronic, stable state. The prevalence of fat is represented by increased signal on T1-weighted MRI.

Because MRI adeptly characterizes the various stages of disc degeneration and its effects on the discovertebral complex, it is often performed as the initial study in evaluating patients who present with axial LBP. However, it is unclear whether findings of disc degeneration on MRI are in fact age-related changes or abnormal processes that are related to pain generation. Several studies have evaluated MRI in asymptomatic volunteers and found that more than 30% of individuals without back pain have evidence of degenerated lumbar discs.^5,34 These include significant findings of herniated discs (24%) and anular defects (14%) in individuals who are asymptomatic. Boden et al. observed that 28% of asymptomatic individuals had an MRI of the lumbar spine that was characterized as abnormal.³⁴ The most common finding was bulging discs, which appeared in 79% of individuals older than 60 years of age and in 54% of those younger than 60 years of age. Degenerated discs were seen in all but one subject older than 60 years of age, many having multiple degenerated discs. Boden et al. concluded that given the high incidence of bulging and degenerated discs in asymptomatic subjects, these findings in part represent the normal process of aging.

Several studies, however, have shown that MRI findings correlate well with abnormal discography that is suggestive of symptomatic degenerative disc disease.³⁵ Horton et al. performed both MRI and lumbar discography in 25 patients presenting with discogenic LBP.³⁶ They observed that normal-appearing discs on MRI with a high fluid content rarely had concordant pain or abnormal morphology on discography. Alternatively, discs that demonstrated a dark nuclear pattern on MRI, suggesting severe disc dehydration and anular disruption, strongly correlated with pain provocation on discography. The more common finding, however, of a dark or speckled nucleus on MRI, which was interpreted as some degree of disc degeneration, had poor correlation with discographic findings. Subsequent studies observed that painful discs on discography tend to show more evidence of degeneration such as anular fissures, disc prolapse, and decreased disc height on MRI. However, there is no significant difference in pattern or frequency of these degenerative findings between asymptomatic individuals and those with LBP to suggest that MRI can effectively screen for painful, symptomatic discs.³⁷

Aprill and Bogduk describe a specific finding on MRI that appears to correlate particularly well with concordant pain on discography.³⁸ The high-intensity zone (HIZ) is an area in the anulus fibrosus that can appear on T2-weighted MRI and has high signal intensity in an otherwise degenerated dark-appearing disc (Fig. 58-3). The HIZ was found in 28% of 500 patients undergoing lumbar MRI for chronic LBP. All discs with an HIZ on MRI were also noted to have a grade 3 or 4 anular disruption on the DDD scale. The HIZ appears to be related pathologically to injury to the anulus. Importantly, the presence of an HIZ correlated strongly with exact reproduction of pain. The sensitivity and specificity of the HIZ for exact pain reproduction on discography were 82% and 89%, respectively. The positive predictive value of the HIZ for concordant pain was determined to be 95%, suggesting that the MRI finding of an HIZ may be a reliable noninvasive measure for anular disruption and pain provocation on discography.

FIGURE 58-3 T2-weighted sagittal MRI of a degenerated disc with a high-intensity zone in the dorsal anulus (arrow), suggestive of anular disruption.

Ultimately, however, there does not appear to be a clear relationship between the mere presence of degenerative changes on MRI and axial LBP. Therefore, one must be careful in interpreting MRI findings and assigning symptomatology to discs that demonstrate radiologic signs of degeneration. Particularly, when evaluating patients for potential surgery, one must consider that the same degenerative changes occur with similar frequency in normal, asymptomatic individuals. Therefore, definitive clinical significance cannot be attributed to abnormal-appearing discs on MRI in the absence of nerve root compression and neurologic symptoms. Modic et al. found that MRI does not appear to have any measurable value in terms of management and outcome of patients with acute LBP.³⁹ With regard to patient selection for surgery, it does appear that MRI is effective for ruling out normal-appearing discs as the source of pain. The decision whether to treat discs that appear abnormal on MRI, however, warrants either further evaluation with provocative discography and/or the experienced assessment of a spine care provider taking into account the patient’s clinical history and symptomatology.

Summary of the Diagnostic Evaluation of Axial Low Back Pain

Diagnosing the source of pain in patients with axial LBP is particularly challenging. Discogenic pain commonly occurs in patients who have a normal physical and neurologic examination. The distribution of pain is somatotopic rather than dermatomal; therefore, identifying the spinal level of pathology is problematic. Imaging findings are also generally nonspecific. Plain radiographs may show decreased disc height and sclerotic end plates. MRI may demonstrate dehydrated, desiccated, or collapsed discs. However, these changes may occur in multiple discs, making it difficult to determine which level is symptomatic. Furthermore, these radiographic and MRI changes commonly occur in asymptomatic individuals, calling into question their clinical relevance.

In 2005, the Joint Section of the American Association of Neurological Surgeons/Congress of Neurological Surgeons published guidelines for the performance of fusion surgery for lumbar degenerative conditions and made recommendations regarding the diagnostic evaluation of axial LBP.³⁵ MRI allows for the assessment of multiple levels simultaneously and noninvasively. Therefore, one can gain an appreciation of the disease pattern of the discs and determine which discs may have evidence of advanced degeneration, disc prolapse, anular disruption, or disc herniation. Once the degenerated levels have been identified on MRI, discography can be performed to further subselect which discs demonstrate abnormal morphology and concordant pain provocation. In their assessment of MRI and discography, they recommended that MRI be performed initially instead of discography in the evaluation of chronic LBP. They also recommended that normal-appearing discs on MRI should not undergo fusion. Conversely, surgery should not be offered on the basis of discography alone, as patients with positive discography but normal MRI had poor outcomes after fusion surgery.⁴⁰ The use of discography to determine which levels to fuse in patients with an abnormal MRI should demonstrate both concordant pain provocation and morphologic abnormalities.

Nonsurgical Treatment

The first line of treatment for axial LBP remains conservative, nonsurgical therapy. While the majority of Americans will suffer an acute episode of LBP at some point in their life, only about 5% to 10% will then continue to develop chronic LBP.³ Basic symptomatic treatment consists of oral analgesics such as nonsteroidal anti-inflammatory agents, narcotics, and muscle relaxants. Injection therapies with caudal epidural steroid injections, selective nerve blocks, direct injections into the peripheral anulus, facet blocks, and trigger point injections may provide significant long-term benefit for some individuals. Alternatively, they may provide enough short-term relief to allow the patient to engage in a physical therapy program and adaptive lifestyle modification, which may afford better long-term benefits. Spinal injections, in addition to being therapeutic, may also help to diagnose which structure is the primary source of pain. Temporary relief of symptoms after selective nerve root injection may help to predict whether decompression will provide more lasting benefit. Pain relief after facet block has not proven to similarly predict long-term success after fusion, although it can be useful in selecting patients who may benefit from medial dorsal rhizotomy. Ohtori et al. observed that preoperative pain relief after intradiscal injection of bupivicaine correlated well with successful improvement in VAS and ODI after lumbar fusion.⁴¹

Bracing has a role in the short-term treatment of acute LBP. In general, however, bracing is not recommended for the treatment of chronic LBP, because there is no evidence to suggest that it provides any long-term benefit. Brace therapy has not been shown to decrease the incidence of LBP or lost work days when used as a preventive strategy.⁴²

Furthermore, long-term bracing may discourage conditioning and strengthening of low back muscles, leading to atrophy and decreased flexibility. Some clinicians contend that symptomatic relief of axial LBP while wearing a brace is a predictor for successful outcome after fusion surgery, although there is no substantial medical literature to support this strategy.

The most significant advances in nonsurgical intervention have been in the development of specialized physical therapy regimens and the emergence of multidisciplinary treatment strategies consisting of lumbar stabilization exercises, endurance training, cognitive behavioral therapy, education, fear avoidance training, adaptive coping skills, and functional restoration. Therapeutic exercises for treating axial LBP have evolved with a recent focus on improving lumbar spine stability. These exercises aim to increase core muscle control, strength, and endurance to maintain dynamic spinal and truncal stability. Specifically, these exercises improve abdominal, spinal, paraspinal, and pelvic muscle function. The potential benefits of lumbar stabilization exercises are based on limited studies showing that patients with unilateral chronic LBP also demonstrate atrophy of the multifidus muscles on the ipsilateral side of their pain. Multicenter prospective clinical studies have shown that patients with chronic LBP who are treated with lumbar stabilization exercises have meaningful improvements in function, pain, and quality of life; however, clinical outcomes are not significantly better when compared to conventional physical therapy, manual therapy, or spinal manipulative therapy.

The McKenzie method is a physical therapy modality that involves an initial assessment of a “directional preference,” in which sustained posture or lumbosacral movement in that direction leads to termination of pain. Proponents of the McKenzie method observe that sequential and eventually lasting abolition of pain can occur in response to repeated movements or sustained postures in this single direction. Once such directional preference has been identified, the patient performs end-range lumbar exercises that match the single direction. In the process, exercises are designed to strengthen muscles, increase soft tissue stability, restore range of movement, improve conditioning, and reduce fear of particular movements.

Multidisciplinary programs have evolved to supplement physical therapy with cognitive behavioral therapy, education, fear avoidance training, adaptive coping skills, and functional restoration. Cognitive behavioral therapy is directed toward replacing maladaptive patient coping skills, emotions, and behaviors with more adaptive ones. Back school is a form of patient education training that teaches patients about lumbar anatomy, physiology, and degenerative disease. Education is directed toward reassurance of safety, fear avoidance, and encouraging activity and self-management. Cognitive behavioral therapy and back school are often provided through group education, training, and exercises. While these interventions do not treat the underlying pathology, they have been demonstrated to reduce pain and anxiety and have become a key component of many multidisciplinary chronic pain management programs.

Functional restoration is a biopsychosocial approach that views chronic pain as an interconnected complex of physiologic, psychological, and social factors, each significantly affecting the patient’s clinical presentation. Functional restoration strategies integrate psychosocial and socioeconomic assessment, cognitive behavioral therapy, pharmacologic therapy, and physical therapy combined with an interdisciplinary medical team to provide a comprehensive treatment plan. Ultimately, functional restoration is directed toward rehabilitation to reduce pain, improve function, and, with early intervention, prevent the progression to chronic disability. Certain studies have shown that functional restoration therapy early in the development of chronic pain syndromes results in fewer indices of chronic pain disability, less narcotic use, earlier return to work, and resolution of medicolegal issues when compared to conventional treatment.

Patient Predictors of Treatment Outcome

Several clinical, psychological, and socioeconomic factors have been identified that correlate with outcome after nonsurgical and surgical treatment of axial LBP. Prior to initiating any intervention, particularly treatments that are irreversible or have associated morbidity, proper consideration of individual patient factors may help to predict the likelihood of success. The benefit of assessing for these predictors is fourfold: first, to appropriately select patients for treatment given the inherent risk/benefit profile; second, to increase the likelihood of success by modifying alterable factors prior to treatment; third, to optimize or tailor the treatment strategy on the basis of the patient’s positive or negative predictors; and fourth, to align patient and physician expectations for realistic treatment outcomes before initiating any intervention.

Patient age has been identified as a risk factor for the occurrence of LBP; however, it is not an individual risk factor for developing chronic disabling LBP. Studies have revealed conflicting results regarding the influence of patient age on outcome after treatment for chronic LBP. Buchner et al. performed a 6-month follow-up on chronic LBP patients who had undergone a multidisciplinary biopsychosocial therapy approach.⁴³ The treatment regimen consisted of cognitive behavioral therapy, lumbar exercises, relaxation training, work-related training, medical training, and therapy directed at developing adaptive coping skills. Significant improvement was observed in all age groups compared to pretherapy except that older patients did not demonstrate significant improvement in functional capacity. Overall, younger patients had significantly better results for both functional capacity and pain level compared to older patients.

Elevated body mass index (BMI) has been associated with LBP symptoms, obesity representing an individual risk factor for the development of chronic LBP.⁴⁴ The relationship between the BMI and the occurrence of LBP is becoming better clarified. Khoueri et al. evaluated morbidly obese patients with chronic axial LBP who were undergoing elective bariatric surgery.⁴⁴ They found a statistically significant improvement in both VAS and ODI in patients who had a decrease in BMI after weight-reduction surgery, despite not undergoing any specific LBP therapy. Obesity is also associated with increased risk for surgical procedures of the spine. A logistic regression analysis demonstrated a positive correlation between risk of significant complication after spine surgery and increasing BMI. While other studies argue that even complex spine surgery can be performed in obese patients with an acceptable morbidity that is comparable to that of nonobese patients, it stands to reason that preoperative weight reduction may improve various aspects of clinical outcome. Overall cardiovascular conditioning, aspects of patient positioning, length of surgery, blood loss, wound healing, and earlier postoperative mobilization and return to activities are potential advantages for patients with an optimal BMI prior to treatment.

Psychological and emotional factors are known to significantly contribute to chronic pain disorders. Anxiety, depression, and maladaptive coping skills that can occur when individuals suffer from longer pain duration result in less improvement in daily activity and work-leisure activity after treatment. Patients with back pain for less than 2 years have significantly better pain relief after treatment than do those with more than 2 years of chronic back pain. The Minnesota Multiphasic Personality Inventory has been well established as a presurgical instrument for predicting outcome after lumbar surgery. Various other measures of psychological and emotional well-being have also been correlated with surgical outcome. Andersen et al. found that the Dallas Pain Questionnaire grading for emotional distress predicted likelihood of significant disability after lumbar fusion surgery.⁴⁵ Wasan et al. observed that psychiatric conditions, particularly depression or anxiety, affected the likelihood of positive outcome from medial branch block injections for LBP.⁴⁶ Carragee et al. reported that absolute psychometric scores, particularly fear avoidance characteristics, were significant predictors of poor outcome.³¹ Derby et al. studied the effect of preoperative psychological and emotional status on patients undergoing single-level lumbar fusion for discogenic LBP.⁴⁷ They found that patients with a normal preoperative mental component summary (MCS) score had significantly better surgical outcomes at 1 and 2 years compared to those with abnormal MCS scores. Regression analyses have also shown that tobacco use, depression, and ongoing litigation are the most consistent presurgical predictors for poor patient outcome.

Work status and pending worker’s compensation claims prior to treatment appear to be significant predictors of outcome. Beals and Hickman demonstrated that work disability for longer than 6 months was a strong predictor of poor outcome.⁴⁸ Only 50% of patients that are disabled from work for more than 6 months eventually return to full work capacity, with only 20% still employed at 1 year and 0% at 2 years after attempting to return to work. Alternatively, patients who are preoperatively employed are more likely to demonstrate good clinical outcomes and return to work function. In a study of patients undergoing dorsolateral fusion for discogenic LBP, all patients who had good or excellent outcomes were employed preoperatively or had been disabled for less than 3 months.⁴⁹ Interestingly, they found that preoperative work status was a better predictor of outcome than was positive discography. Anderson et al. observed similar findings in patients undergoing ALIF for discogenic LBP.⁵⁰ Patients who were employed preoperatively were 10.5 times more likely to be working at follow-up. Only 43% of patients who were unemployed before surgery were employed at follow-up. Alternatively, 90% of patients who were working before surgery were still working at follow-up. Preoperative work status also predicted posttreatment pain relief. Patients who were employed before surgery demonstrated greater improvement in pain scores, while those with worker’s compensation claims had less pain relief. These findings were independent of number of levels treated or other patient demographics. For patients with musculoskeletal injuries in general, it has been shown that continuing even modified work activities prior to surgery had a significant impact on the likelihood of returning to work after treatment.^51,52 Therefore, patients should be encouraged to remain active and continue to work in some capacity, even if work accommodations or restrictions are necessary. Limiting preoperative work disability to less than 3 months may improve outcome and the likelihood of returning to full work capacity postoperatively.

Lumbar lordosis and sagittal balance have recently come to light as independent predictors of LBP and overall health status. An association between LBP and loss of lumbar lordosis has been observed. Jackson et al. characterized lumbar lordosis in asymptomatic individuals and patients with mechanical-type LBP who were matched for age, gender, and size.⁵³ They observed that two thirds of total lumbar lordosis occurs at L4-5 and L5-S1 and that LBP patients have significantly different segmental lordosis for each motion segment, with particularly less distal segmental lordosis. Significant conclusions regarding the effect of decreased lumbar lordosis on back pain must be considered carefully, as it has also been shown that all people, including asymptomatic individuals, demonstrate progressive positive sagittal balance with normal aging. The absence of back pain in individuals with sagittal imbalance may reflect effective compensatory mechanisms, the development of pain representing eventual failure of compensation.

The relationship between sagittal balance and outcome after surgery may potentially be extrapolated from the spinal deformity literature. Glassman et al. investigated predictors of poor health status in adult scoliosis patients.⁵⁴ They found that among both surgical and nonsurgical patients, positive sagittal balance was the most reliable predictor of clinical symptoms. Worse self-assessments in pain, function, and self-image were correlated with positive sagittal balance as compared to other patient factors. They also noted that relative kyphosis of the lumbar spine was correlated with more significant disability than was the case in patients with a normal or lordotic lumbar spine.⁵⁵ Takayama et al. reported the incidence of LBP in patients who had previously undergone spinal deformity surgery.⁵⁶ They observed that LBP was strongly correlated with sagittal balance, more so than latest Cobb angle or evidence of degenerative changes at the adjacent segments. Therefore, when considering operative intervention, one should carefully assess the overall lumbar alignment, particularly the segmental lordosis at L4-5 and L5-S1, as disc degeneration and collapse at these levels factor significantly in the global sagittal balance. Relative straightening of the distal lumbar segments may help to dictate a surgical plan not only to fuse the abnormal motion segment but also to increase lumbar lordosis and improve sagittal balance.

Surgical Outcome

The surgical treatment of axial LBP due to degenerative disc disease remains controversial. Axial back pain is thought to be due to alterations in the biomechanical and biochemical environment created by the degenerated disc. Pathologic motion or stress transference to pain-generating structures under normal physiologic loads can lead to pain. Disruption of the normal architecture of the disc may result in extravasation of biochemical agents that stimulate surrounding nociceptors. Lumbar fusion surgery, by immobilizing the motion segment, is believed to counteract these pathologic processes to reduce pain generation and decrease disability.

The concern with fusion surgery stems from a lack of consensus regarding the appropriate indications for surgical treatment of axial LBP. Because patients commonly present with generalized complaints, a normal neurologic examination, and imaging studies that are often unreliable in identifying the pain generator, determining which patients are candidates for fusion surgery is challenging at best. Unclear surgical indications likely result in a heterogeneous patient population undergoing spine fusion for symptoms of back pain. Many of these patients may in fact have varying underlying etiologies of back pain. Therefore, it is not surprising that surgeons and institutions demonstrate a broad spectrum of management practices for axial LBP, which consequently is reflected in a wide range of patient outcomes in the surgical literature. To compound the issue, the actual correlation between bony fusion, clinical outcome, and pain relief has yet to be definitively established.^57–59 Studies of lumbar spondylolisthesis–related instability, which is commonly treated with arthrodesis, have failed to demonstrate a clear correlation between successful fusion and outcome. In a randomized, prospective study of patients with lumbar spondylolisthesis and stenosis, dramatically increased fusion rates in patients with pedicle screw fixation did not, however, result in a concomitant improvement in clinical outcomes.⁵⁸

Proponents of fusion surgery for axial LBP contend that a certain percentage of patients do respond well to surgery with improved pain, decreased narcotic use, and return to work compared to the natural history of the disease. While the success rate of lumbar fusion for axial LBP may be less than that of decompression for herniated disc or lumbar stenosis, the axial LBP population is a particularly challenging group that is susceptible to poor outcomes. Often, these patients have failed all other treatment options, are disabled from work, and suffer psychosocial disorders related to long-standing pain and anxiety. There is an arguable need, though, to offer a treatment option for those patients who have exhausted all other resources. Therefore, even in the absence of absolute indications, there may be a role for surgery to provide some measure of pain relief and improve the quality of life for a fraction of patients afflicted with chronic axial LBP.

There are a number of clinical studies that have examined outcomes after fusion surgery for axial LBP and degenerative disc disease. Drawing reasonable conclusions from these studies, however, must be done cautiously. The majority of studies are retrospective case series, with only four prospective, randomized controlled studies comparing fusion to nonsurgical treatment for back pain to date. As previously stated, many of these studies employ wide-ranging practices with regard to diagnostic workup, indications for surgery, procedure performed, and method for assessing outcome. Differences with regard to diagnostic evaluation and patient selection vary in the use of discography, MRI, or relying on clinical judgment for determining which patients and what levels to fuse. A broad spectrum of surgical procedures is also described. Currently, there are a number of ways to arthrodese a motion segment, including anterior interbody fusion (ALIF) (Fig. 58-4), posterior interbody fusion (PLIF), transforaminal interbody fusion (TLIF) (Fig. 58-5), lateral interbody fusion (Fig. 58-6), intertransverse fusion (PLF), and facet fusion. Preferences differ with regard to choice of bone graft material, such as autogenous iliac crest, local autograft, allograft, and synthetic osteobiologic agents. The use of segmental instrumentation and interbody devices is also variable. The specific techniques, nuances, advantages, and disadvantages of these surgical procedures are beyond the scope of this chapter. The literature supporting these general fusion techniques in the treatment of axial LBP will be presented separately; however, ultimately, the common singular objective is to achieve solid arthrodesis of the motion segment, and, therefore, the specific surgical measures to perform the fusion are presumed to be secondary.

FIGURE 58-4 A, Lateral radiograph demonstrating an L5-S1 anterior lumbar interbody fusion (ALIF) with an interbody cage. B, Anteroposterior radiograph demonstrating an L5-S1 ALIF with two interbody cages.

FIGURE 58-5 A, Preoperative lateral radiograph demonstrating a degenerated, collapsed L4-5 disc space. B, Preoperative T2-weighted MRI of the same patient demonstrating severe L4-5 disc desiccation with loss of T2 fluid signal within the disc. C, Preoperative T1-weighted MRI of the same patient demonstrating increased signal in the adjacent L4 and L5 vertebral bone marrow consistent with chronic fat infiltration. D, Postoperative lateral radiograph of the same patient after L4-5 transforaminal interbody fusion (TLIF) with dorsal segmental instrumentation. E, Postoperative anteroposterior radiograph of the same patient after L4-5 TLIF with dorsal segmental instrumentation.

FIGURE 58-6 A, Preoperative T2-weighted MRI demonstrating severe disc degeneration of L3-4 disc. B, Postoperative lateral radiograph of the same patient after L3-4 lateral interbody fusion with interbody cage placement. C, Postoperative anteroposterior radiograph of the same patient demonstrating L3-4 lateral interbody fusion with interbody cage placement.

Clinical Series: ALIF, PLIF, PLF, 360-Degree Fusion

Studies evaluating outcome after ALIF for axial LBP demonstrate high rates of clinical improvement. Newman et al. evaluated 36 patients who underwent ALIF with autogenous iliac crest bone graft for degenerated discs diagnosed with MRI and provocative discography.²⁴ Twenty-eight patients underwent single-level fusion, and eight patients had two-level fusions. Overall, 86.1% of patients had a successful clinical outcome, with a fusion rate of 88.9%. Unsuccessful results were observed in 13.9% of patients. Chow et al. similarly reported promising outcomes in 97 patients undergoing ALIF for degenerative disc disease.⁶⁰ Complete relief of symptoms occurred in 60%, with an additional 29% having marked improvement in symptoms. Of note, the primary diagnosis of degenerative disc disease and which levels were symptomatic was made by using only the clinical history and plain radiographs in the majority of patients. Only four patients had preoperative provocative discography to facilitate assessment of the symptomatic discs. Blumenthal et al. found a slightly lower rate of successful outcome after ALIF.²⁶ They observed that 74% of patients with single-level discogenic disease had a good outcome as defined by return to work and cessation of narcotic use.

Lee et al. studied 62 patients undergoing uninstrumented PLIF with autogenous iliac crest bone graft for chronic disabling LBP.²⁸ Symptomatic levels for fusion were determined on the basis of provocative discography and corresponding morphologic abnormalities on MRI. Overall, 59.2% of patients were free of narcotic use at last follow-up. Eighty-one percent of patients had completely returned to work, with over 92% of patients employed in at least some partial capacity. Seventy-two percent of patients reported that they were completely satisfied with their surgical outcomes, with 80.5% stating that they would undergo the operation again.

Parker et al. assessed 23 patients who underwent instrumented PLF with iliac crest autograft for discogenic LBP.⁴⁹ All patients underwent both MRI and discography preoperatively; however, the fusion levels were ultimately determined by which discs demonstrated concordant pain on provocative discography. Overall, only 39% of patients had a good or excellent outcome as defined by a VAS of 4 or less, no analgesic use, and return to at least 75% of premorbid work status. Poor outcomes were observed in 48% of patients, who had a VAS of more than 6, daily narcotic use, and less than 25% of previous work capacity. They noted that pseudarthrosis (22%) and unemployment for more than 3 months preoperatively were associated with poor outcomes. Overall, only 56% of patients reported that they were extremely satisfied with the surgical results.

Combined anterior-posterior or 360-degree fusion for degenerative disc disease has also been reported. Moore et al. studied 58 patients who underwent ALIF with instrumented dorsolateral fusion for discogenic pain.⁶¹ With a minimum of 2 years of follow-up, they observed a 95% arthrodesis rate. At final follow-up, 86% of patients were improved in comparison to their situation preoperatively, with 88% of patients having returned to work. Linson et al. similarly studied patients undergoing anterior-dorsal fusion and reported that 80% of patients had measurable improvement in LBP symptoms.²⁵

Leufven and Nordwall studied 29 patients undergoing combined instrumented PLIF and PLF for axial LBP of greater than 2 years’ duration.⁶² Fusion levels were identified by positive discography. Sixty-nine percent of patients reported either no back pain or improvement of back pain symptoms, with 31% describing complete relief of symptoms, being off all analgesics, and return to full activities. Forty-eight percent of patients had a suboptimal outcome consisting of continued moderate or severe pain, some or total activity restriction, and continued daily use of analgesics. Only 62% of patients were employed at follow-up, and only 76% of patients stated that they would have the surgery again.

Recently, studies have also evaluated surgical outcome for the treatment of disease on three or more levels, challenging the presumption that surgical fusion for degenerative disc disease is best limited to single-level or two-level fusions. Suratwala et al. assessed 360-degree fusion for three or more levels in patients with lumbar degenerative disc disease.⁶³ Surgical procedures consisted of either ALIF or TLIF with dorsal instrumentation. Diagnostic workup consisted of MRI in all patients, with 60 of 80 patients undergoing discography. Overall, the authors reported a 29.5% improvement in ODI and 30.7% improvement in Roland Morris score at a minimum of 2 years of follow-up.

Lettice et al. compared patients undergoing lumbar fusion for single-level or two-level discogenic disease with those with three or more affected levels.⁶⁴ Surgical procedures include PLF plus ALIF or PLIF. Symptomatic discs were identified by using provocative pressure-controlled discography. They found significant improvement in SF-36 physical component scores in both groups at 1-year and 2-year follow-up. The longer-segment group did have a higher pseudarthrosis rate and reoperation rate; however, the number of levels that were treated did not have a significant effect on overall clinical outcome as measured by SF-36.

In 2004, Geisler et al. performed a meta-analysis of the literature to compare various fusion techniques for the treatment of axial LBP.⁶⁵ They reviewed 25 papers that had a minimum of 2 years of follow-up for the standardized outcome measures ODI and VAS. They found that patients undergoing ALIF had a mean 45.5% improvement in VAS and a mean 27.9-point reduction in ODI. Similarly, for 360-degree fusion (PLF plus ALIF, TLIF, or PLIF), patients demonstrated a mean 49.1% improvement in VAS and a mean 20.6-point reduction in ODI. In a prospective, randomized controlled trial, Fritzell et al. compared patients who had undergone uninstrumented PLF, instrumented PLF, and instrumented 360-degree fusion.⁶⁶ At 2 years of follow-up, there was no significant difference in pain, disability, depressive symptoms, or overall rating between the three surgical groups. While they did observe that the use of instrumentation resulted in increased fusion rates, no correlation between fusion rate and clinical outcome was demonstrated. More complications, lengthier hospitalization, increased blood loss, and new incidence of postoperative leg pain were identified in the patients who had undergone pedicle screw instrumentation.

Prospective, Randomized Studies

The Oxford Centre for Evidence-Based Medicine classifies studies investigating medical treatment effect based on the soundness of their scientific method and the value of their clinical relevance.⁶⁷ On the basis of this classification, the majority of existing studies supporting lumbar fusion for the treatment of axial LBP provide only class III or IV evidence (retrospectively collected data; case-control study or case series). To date, there have been only four prospective, randomized controlled trials comparing lumbar fusion to nonsurgical therapy for the treatment of chronic LBP (class I evidence). In 2001, the Swedish Lumbar Spine Study Group published their findings comparing lumbar fusion with nonoperative treatment in patients with chronic LBP.⁶⁸ A total of 294 patients were randomized, 222 patients undergoing surgery and 72 patients treated nonoperatively. Inclusion criteria were patients with chronic disabling LBP with disc degeneration. Exclusion criteria were leg pain, spondylolisthesis, spinal stenosis, fracture, infection, or tumor. Surgical patients were then randomized to either uninstrumented dorsolateral fusion, instrumented dorsolateral fusion, or 360-degree instrumented fusion. Nonsurgical therapy was not standardized, with the patient’s treating physician deciding from a range of interventions including physical therapy, education, transcutaneous electrical nerve stimulation, epidural steroid injections, cognitive and functional training, and coping strategies.

The surgical group as a whole demonstrated a 33% reduction in back pain and 25% reduction in ODI. Sixty-three percent of surgical patients rated themselves as “much better” postoperatively. Return to work was observed in 36% of surgical patients. Interestingly, there was a 20% reduction in depressive symptoms among the surgical group. Conversely, the nonsurgical group demonstrated only a 7% reduction in back pain and a 6% decrease in ODI. Only 29% of nonsurgical patients rated themselves as “much better” after treatment, with only 13% returning to work. On the basis of these findings, the Swedish Lumbar Spine Study Group concluded that fusion surgery not only decreases chronic LBP and disability but is superior to nonsurgical therapies.

Although this study by Fritzell et al.⁶⁸ received the 2001 Volvo Award for clinical studies, it also sparked considerable debate. The primary criticism stems from the nonsurgical group and the lack of a standardized nonsurgical treatment strategy. Because the nonsurgical arm consisted of whichever intervention was offered by the individual treating physician, these patients were subjected to any variety of possible therapies, which are not well detailed in the report. It is certainly possible that the members of the nonsurgical group that had already failed conservative treatment for an average duration of 8 years were potentially being randomized to a treatment arm that consisted of essentially the same continued failed therapy. Therefore, it is not surprising that the nonsurgical group demonstrated modest (if at all) improvement in pain and disability compared to any benefit seen in the surgical group. Additional criticism arises from an unclear algorithm for diagnosing the etiology of back pain and determining the symptomatic levels. Taken a step further, a reasonable concern is that the study could be misconstrued as wholesale advocating of lumbar fusion surgery as a better treatment for chronic LBP than nonsurgical therapy.

In response to these issues, Brox et al. performed a prospective, randomized controlled trial comparing instrumented lumbar fusion and a standardized nonsurgical treatment protocol, which had been shown to be more effective than conventional conservative care.⁶⁹ A total of 64 patients with more than 1 year of LBP and evidence of disc degeneration at L4-5 or L5-S1 were enrolled. Nonsurgical treatment consisted of an intensive 3-week program in which patients stayed at a “back hotel” and participated in physical therapy sessions three times per day, with an average total of 25 hours per week of exercises, cognitive intervention, education, and peer counseling.

Both the surgical and nonsurgical groups demonstrated significant improvements compared to baseline at 1 year of follow-up in nearly all outcome measures. The surgical group had an average 15-point reduction in ODI compared to a 12-point decrease in the nonsurgical group. The surgical group had better reduction in leg pain; however, the nonsurgical group had greater improvement in fingertip-to-floor distance and fear avoidance beliefs. No significant differences in back pain, analgesic requirements, emotional distress, life satisfaction, or return to work were observed between the surgical and nonsurgical groups. Overall, both treatment strategies were considered to be effective, with 70% of surgical patients and 76% of nonsurgical patients deemed successful outcomes.

In a subsequent study, the same research group performed a prospective, randomized clinical trial comparing lumbar fusion versus the same nonsurgical treatment protocol in patients with chronic LBP after prior disc herniation surgery.⁷⁰ A total of 60 patients were enrolled. The investigators found that the surgical group had a mean ODI reduction of 9 points compared to a 13-point decrease in the nonsurgical group. These differences were not statistically significant, and the overall success rate of the surgical group was 50% compared to 48% of the nonsurgical group. Again, the authors concluded that at 1-year follow-up, lumbar fusion failed to demonstrate any significant benefit over cognitive intervention and exercises in the treatment of chronic LBP.

The primary criticism of these two studies is that the follow-up duration was only 1 year and that both studies were significantly underpowered. Fairbank et al. investigated lumbar fusion versus an intensive rehabilitation program.⁷¹ Three hundred forty-nine patients were randomly assigned, with 176 patients undergoing surgical treatment and 173 undergoing rehabilitation. The surgical procedures varied and were dependent on surgeon and institutional practices. The nonsurgical treatment was a standardized protocol that involved 5 days per week for 3 weeks of outpatient therapy. A total of 75 hours of intervention were performed consisting of cognitive behavioral therapy and various exercises directed toward strengthening, flexibility, endurance, and aerobic conditioning. The surgical group demonstrated an average reduction in ODI of 12.5 points compared to an 8.7-point decrease in the nonsurgical group. This difference was marginal although statistically significant. The authors concluded that the modest advantage of fusion surgery did not provide clear evidence that surgery is more beneficial than intensive rehabilitation for the treatment of chronic LBP. Alternatively, they suggest that their study provides evidence that intensive rehabilitation with cognitive behavioral principles is a viable treatment option and alternative to fusion surgery. Critics of this study point out that 28% of nonsurgical patients crossed over to the surgical arm. Therefore, the patients who were demonstrating the worst outcomes with nonsurgical therapy ultimately had surgery. An intent-to-treat analysis was performed that was inherently biased against surgery, as these patients who were failing nonsurgical therapy may have gained benefit after crossing over, although this was not reflected in the final outcomes assessment.

In 2008, Mirza and Deyo performed a Cochrane summary of these three prospective, randomized controlled studies (Fritzell et al.,⁶⁸ Brox et al.,⁶⁹ and Fairbank et al.⁷¹).⁷² Mirza and Deyo concluded that in all three studies, surgical treatment demonstrated similar functional improvement with an 8.9 to 15.6 decrease in ODI, which corresponded to a 19% to 37% change from baseline. The nonsurgical group, which ranged from a nonstandardized variety of treatments to standardized protocols for intensive cognitive therapy and exercise, showed an improvement in ODI of 2.8 to 13.3 points and a 5.8% to 30.1% improvement from baseline. Their conclusion was that surgery provided a fairly modest increase in improvement compared to nonsurgical treatment for discogenic back pain. However, none of the observed differences in average back-specific disability outcome were sufficiently large to meet FDA criteria, as defined by a 15-point or more improvement in ODI, for a clinically meaningful difference between treatments.

Analysis of Lumbar Fusion Outcomes in Studies of Total Lumbar Disc Replacement

It is difficult to generate meaningful practice guidelines from the existing literature regarding spine fusion for axial LBP. The majority of the studies are retrospective case series with varying degrees of follow-up and inconsistent outcome assessment measures. A broad spectrum of practices with regard to patient selection, diagnostic workup, procedures performed, and levels fused is presented. Most important, although many of these studies demonstrate an improvement in back disability indices or pain scores compared to preoperatively, it is unclear whether these changes represent a clinically meaningful improvement.

Over the last 20 years, there has been considerable interest in exploring lumbar disc replacement as a surgical treatment for degenerative disc disease. For a full discussion regarding disc arthroplasty, see Chapter 161. However, heightened enthusiasm for this device technology has led to several FDA-approved investigational device exemption (IDE) studies comparing total disc replacement to lumbar fusion for degenerative disc disease. As part of the FDA IDE protocol, study centers maintained meticulous high standards with regard to subject enrollment, preoperative evaluation, standardization of surgical procedure, prospective outcomes assessment, duration of follow-up, and controlling for confounding variables. Although these studies were intended to evaluate the effectiveness of disc replacement compared to fusion, they also provide valuable information about lumbar fusion outcomes in a highly controlled patient population with axial LBP. Therefore, evaluating only the lumbar fusion treatment arm of these FDA IDE studies provides prospective, controlled data regarding lumbar fusion for the treatment of axial LBP. Based on the Oxford Centre for Evidence-Based Medicine, these studies provide class II evidence regarding the use of lumbar fusion to treat axial LBP.

Blumenthal et al. performed a prospective, randomized controlled trial comparing treatment of degenerative disc disease with the Charité total disc replacement versus ALIF with a BAK cage packed with autogenous iliac crest bone graft.⁷³ Study enrollment included patients suffering from back pain without leg pain, positive discography, single-level disease, failure with conservative therapy, and no prior lumbar fusion surgery. The conclusions of the study were that the Charité total disc replacement demonstrated noninferiority compared to ALIF for the treatment of degenerative disc disease with 2 years of follow-up.

Focusing specifically on only the lumbar fusion group, 99 patients underwent ALIF with outcomes assessment at 24 months. The fusion patients demonstrated a mean 42.4% reduction in ODI and an average 34.1% reduction in VAS at 2 years. However, at 2 years, 80% of patients still required narcotic pain medications. A small increase in employment rate was observed from 57.6% preoperatively to 65% at 2 years. As part of the FDA IDE protocol, a standardized criteria for meaningful clinical success was determined prior to initiating the study. The FDA criteria for overall clinical success were defined as a 25% or greater improvement in ODI, no device failure, no major complications, and no neurologic deterioration. At 2 years, 56.8% of patients undergoing ALIF for degenerative disc disease met FDA criteria for clinical success. A subsequent study that followed a portion of these same patients at 5 years after surgery demonstrated that the employment rate dropped to 46.5% and the overall clinical success rate decreased to 51.2%.⁷⁴ Overall, the study by Blumenthal et al. provides class II evidence that slightly more than half of patients undergoing ALIF for single-level degenerative disc disease meet FDA criteria for clinical success at 2 years.

Zigler et al. studied the Prodisc-L total disc replacement compared to 360-degree fusion for patients with back pain due to single-level degenerative disc disease.⁷⁵ Patients were prospectively, randomized to either disc replacement or ALIF with femoral ring allograft and dorsolateral fusion with iliac crest autograft and pedicle screw fixation. At 24 months of follow-up, patients who were treated with the Prodisc-L had statistically significant better reduction in ODI than did fusion patients, with an improvement in VAS that trended toward significance. Again focusing specifically only on the fusion group, 75 total patients underwent 360-degree fusion for debilitating back pain due to single-level degenerative disc disease. Of these patients, 84.5% reported some improvement in ODI, with 54.9% demonstrating a 25% or greater improvement in ODI and 64.8% having a minimum 15% or greater improvement in ODI. A mean 32-point reduction in VAS was observed. At 24 months, 70% of fusion patients had improvement in their composite SF-36 physical and mental score compared to baseline. A modest improvement in employment rate was noted, with 78.1% working preoperatively and 85.1% employed at 24 months postoperatively. The FDA criteria for success for this study included a 15-point or higher ODI improvement; improvement in SF-36; device success as defined by absence of reoperation to modify, remove implants, or supplemental fixation; radiographic evidence of fusion without graft subsidence, migration, radiolucency, or loss of disc height; and no evidence of neurologic deterioration. Only 40.8% of patients undergoing 360-degree fusion for single-level degenerative disc disease met FDA criteria for overall clinical success.

Together, these two studies provide class II evidence that lumbar fusion provides meaningful clinical success in 56.8% and 40.8% of patients undergoing ALIF or 360-degree fusion for axial back pain, respectively. The strength of these studies is that they are prospective, controlled, multicenter studies with rigorous follow-up using validated outcome assessment measures. Strict inclusion and exclusion criteria were maintained with a well-defined medical condition for study enrollment. Importantly, they also evaluated for clinically meaningful success as determined by FDA criteria. On the basis of these findings, one can conclude that lumbar fusion may significantly benefit approximately half of patients with single-level degenerative disc disease who have failed conservative therapy.

These studies must also be interpreted cautiously. Inherently, these studies were not blind; therefore, involved centers may have been biased toward demonstrating benefit of total disc replacement and downplaying the beneficial effects of lumbar fusion. Furthermore, in an attempt to encourage study recruitment and meet study deadlines, investigators may have aggressively enrolled patients whom they normally might not have considered good candidates for lumbar fusion, knowing that a 2:1 disc replacement to fusion randomization scheme was employed. Furthermore, poor outcomes in the fusion group may have been confounded by disappointment from subjects not being randomized to the newer device technology with disc replacement. Last, the strict inclusion criteria might not in fact represent optimal or realistic candidates for lumbar fusion. Several authors found that only about 0% to 6% of patients treated in their practice with a single-level interbody fusion would have met the inclusion criteria of the Charité study.

Conclusion

Axial LBP remains a significant health-care problem, causing pain and disability for many individuals and placing a heavy financial burden on society. The management of axial LBP is challenging. Difficulties stem from our poor understanding of the pathophysiology of axial LBP and inconsistent and unreliable diagnostic modalities for localizing the source of pain. Therefore, the task of treating axial LBP becomes immeasurably more challenging as our confidence in identifying the pain generator falters and consensus for surgical indications remain unresolved. Furthermore, the overall condition of disabling LBP may in fact be multifactorial, as several nonspinal factors such as psychosocial well-being, employment, and duration of pain appear to significantly contribute to the clinical problem and likelihood of treatment success. The existing literature (class I, II, and III evidence) does, however, provide support for the utilization of lumbar fusion as a treatment option for a properly selected group of patients with chronic axial LBP. In fact, both surgical and nonsurgical interventions have demonstrated improvement in pain relief, back disability index measurements, and return to work, with fusion surgery resulting in better outcomes in a limited number of prospective, randomized controlled trials. However, even in highly controlled studies, fusion surgery may be successful in only about one half to three fourths of patients, with a significant percentage of patients gaining no benefit from surgery. This underscores the heterogeneity in the patient population. Further studies are needed to determine which patients have true mechanical back pain that might be reduced by fusion of the lumbar spine.