 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1999 The Journal of Bone and Joint Surgery, Inc.
Current Concepts Review - Lumbar Arthrodesis for the Treatment of Back Pain* , CHARLOTTE, NORTH CAROLINA
Lumbar arthrodesis is a commonly performed operative procedure for the treatment of low-back pain; however, the indications, techniques, and results remain controversial and unclear. The frequency of spinal arthrodesis for the treatment of back pain is increasing in the United States, as are criticism of the procedure and the study of available information on its outcomes. The concept of spinal arthrodesis is based on experience with other regions of the body in which arthrodesis has been used to treat painful joints by eliminating motion. Initially, spinal arthrodesis was used for the treatment of infectious conditions, deformity, and trauma of the spine. On the basis of these successful experiences and because of technical advances in imaging, operative procedures, implants, and bone-grafting, the indications for spinal arthrodesis have been expanded in an attempt to control pain attributed to abnormal or unstable motion between one vertebra and an adjacent vertebra or pain due to mechanical degeneration of the intervertebral disc. Much information on arthrodesis of the lumbar spine has been published, but most investigators have used nonstandardized criteria for inclusion of patients who were operated on for various diagnoses and whose outcomes were assessed with nonvalidated methods. Nevertheless, careful review of the available data may assist in determining which treatments are reasonable, which are unreasonable, and which are (or should be) considered investigational155. Under ideal circumstances, spinal arthrodesis should be performed only after a specific pathoanatomical diagnosis has been identified as being responsible for the patient's symptoms. In addition, the natural history of the diagnosis and the appropriate timing of operative intervention should be understood. By limiting abnormal motion or removing the intervertebral disc, lumbar spinal arthrodesis can potentially minimize or eliminate the source of pain. Procedures that result in minimum disruption of tissue and restore the normal biomechanics and physiological characteristics of the spine and that are safe, cost-effective, and not associated with adverse short or long-term effects should be considered for use in patients who are appropriate candidates for operative intervention. An understanding of the etiology, diagnosis, and natural history of low-back pain and of the nonoperative and operative treatments that are available should guide the surgeon in selecting the best treatment for the patient.
The large body of available information regarding the epidemiological characteristics of low-back pain has been summarized previously3. In industrialized societies, the prevalence of back pain over the course of a lifetime in the entire population is more than 70 percent. The incidence of back pain in the United States is 15 to 20 percent per year. Back pain is the most frequent cause of limitation of activity among individuals who are less than forty-five years old, and it is the second most common symptom necessitating a visit to a physician. Approximately 1 percent of the population in the United States is chronically disabled because of low-back pain77, and an additional 1 percent is temporarily disabled75. Each year, there are approximately 500,000 Workers' Compensation and personal-injury cases dealing with low-back pain. Most of the high cost incurred by society for the treatment of back pain is attributable to chronic, disabling pain78. The rate of operations on the back is higher in the United States than it is in other countries, and there are large variations among regions3,115,174,196,202. The rate of lumbar arthrodesis increased by 200 percent between 1979 and 1987; this increase was greater than those for laminectomy (23 percent) and discectomy (75 percent). The rate of operations on the lower back increased an additional 100 percent, to twenty-six per 100,000, between 1987 and 1990196. Although it is impossible with current databases to give precise costs, it has been estimated that in the United States the direct and indirect cost of all conditions related to the spine ranges from thirty-eight to fifty billion dollars per year79. The specific cost for the treatment of low-back pain is unknown, but it is obvious that the cost of diagnosis and treatment is a major health-care issue.
Facet Joints
Intervertebral Discs The innervation of the intervertebral disc has been well characterized18,19,203,222. The meningeal (sinuvertebral) nerve branches supply the posterior longitudinal ligament and the outer layers of the annulus fibrosus. The outer one-third of the annulus fibrosus is innervated with pain-transmitting free nerve-endings, and production of pain-related neuropeptides has been demonstrated45. In addition, recent evidence suggests that severely degenerated lumbar discs have a more extensive innervation than do normal discs45. Disruption of the annulus fibrosus during degeneration of the disc may cause stimulation of these pain pathways. Cartilage end-plates and underlying cancellous bone have been shown to have an increased density of sensory nerves and neuropeptides in patients who have degenerative disc disease, and these structures may transmit pain29. The pain response to tissue stimulation during procedures on the lumbar spine with use of local anesthesia has been reported65,106,123,161,165,192,214. The findings suggest that direct stimulation of muscle and ligamentous tissue, often thought to be a cause of back pain, does not result in typical complaints of low-back pain. The outer layer of the annulus fibrosus has most often been implicated as the tissue causing axial low-back pain. Kuslich et al. found that the ligamentum flavum, epidural fat, posterior aspect of the dura, nucleus pulposus, laminae, and spinous processes were insensitive to mechanical stimulation in 193 patients123. Stimulation of the central and lateral portions of the annulus fibrosus produced typical symptoms in slightly more than two-thirds of patients. The findings for the posterior longitudinal ligament were similar to those for the annulus fibrosus. The vertebral end plate was sensitive to stimulation in 61 percent of patients, and the facet-joint capsule was sensitive in 30 percent. Only direct pressure or stretch applied to the nerve root produced pain in the lower extremities. Pain in the buttocks could be reproduced only by simultaneous stimulation of the annulus fibrosus and the nerve root. Stimulation of fascia and muscle did not produce typical low-back pain. Indirect evidence of a discogenic source of pain is also found in the literature on the cervical spine. Resolution of symptoms after excision of a cervical disc without interbody arthrodesis has been reported148,182. These findings support the hypothesis that the disc is the source of pain and that continued loading after excision does not preclude a successful outcome. This theory also is supported by reports that continuing pain after successful posterior lumbar spinal arthrodesis can be relieved by excision of the disc and interbody arthrodesis208.
A controversial indication for spinal arthrodesis is instability of one vertebra on an adjacent vertebra as a consequence of degenerative arthritis. This condition is thought to be uncommon225. One theory accounting for the success of this procedure is that selective arthrodesis of a single vertebra to an adjacent vertebra has a high probability of relieving symptoms if the diagnosis of abnormal motion can be made with certainty194. The concept of eliminating motion in order to eradicate pain has been questioned by several authors who found no association between the outcome and the presence of a stable fusion, and a poor outcome has not been found to be associated with pseudarthrosis8,69,72,89. On the basis of observations of pathological specimens, Kirkaldy-Willis and Farfan concluded that degeneration of the lumbar spine occurs in three phases: dysfunction (progressive tearing of the annulus fibrosus, degeneration of the nucleus pulposus, and arthropathy of the facet joints), instability (laxity of the facet joints, ligaments, and discs), and restabilization (formation of osteophytes and hypertrophy of the facet joints)116. The natural history of instability was studied radiographically by Sato and Kikuchi, who reported restabilization in 20 percent of fifty patients after a minimum ten-year follow-up184. The American Academy of Orthopaedic Surgeons defined segmental instability as "an abnormal response to applied loads, characterized by motion in the motor segment beyond normal constraints."1 In a patient who has an abnormal disc, subacute and repetitive mechanical overloading may result in abnormal motion that may cause pain. This concept of segmental instability, which does not indicate an exact source of pain, may be related to, but differs from, the diagnosis of so-called discogenic back pain (pain related to disc degeneration), for which demonstrable abnormal motion is not mandatory. Many classifications have been proposed in an attempt to identify a definitive etiology on the basis of biomechanical, clinical, and radiographic criteria74,76,175,213. Frymoyer and Selby attempted to divide segmental instability into subtypes on the basis of the deformity that would produce symptoms if progression occurred76. They also proposed tailoring a specific technique of arthrodesis to the type of instability as a means of reducing the rate of failure. Stokes and Frymoyer subsequently concluded that this classification could not be used with clinical success even when sophisticated radiographic studies were part of the evaluation process194. Many orthopaedists believe that there are patients with degenerative instability without spondylolisthesis or scoliosis who may respond to operative stabilization. Radiographic signs, such as the traction spur140 and abnormal translational motion171, continue to be used as clinical criteria for this diagnosis. Additional studies have been performed with use of cadavera to more precisely quantify the magnitude of the motion that occurs175 or to define the source of pain with use of facet blocks159 or discometry92. Historically, some surgeons have advocated a trial of immobilization with use of a brace as a means of predicting which patients might benefit from intersegmental arthrodesis. However, biomechanical and radiographic studies have shown that lumbar orthoses have very little effect on motion between vertebrae even when a thigh extension is added. This treatment therefore has been abandoned by most clinicians6,7,67,165. A trial of short-term external fixation has shown some promise in helping to determine which patients will benefit from arthrodesis120, although this technique is impractical in most situations and has not gained widespread acceptance. Knutsson defined instability as "three millimeters or more of anterior translation on flexion/extension bending films."119 Other authors have attempted to define abnormal displacement more precisely5,16,59,73,132,168,175. Spratt et al. analyzed the interobserver and intraobserver variability of these measurements and concluded that, at the third and fourth and the fourth and fifth lumbar vertebrae, a minimum of four millimeters of anterior displacement defined instability, and at the fifth lumbar and first sacral level, a minimum of five millimeters as seen on lateral bending radiographs was required191. The relationship between degeneration of the disc as seen on discography and angular instability (tilting of 15 degrees or more by one vertebra on an adjacent vertebra) as seen on lateral bending radiographs was studied, but no association was found190. Others have noted that angular instability tends to resolve over time unless it is combined with translational instability, in which case it tends to persist184.
In order to direct the best course of treatment for a patient who has persistent symptoms, an understanding of psychological factors and their relationship to low-back pain is important. Issues of secondary gain related to low-back pain include financial gain from disability insurance or Workers' Compensation and sick-role-related psychological gain. A range of disability and distress is commonly seen in patients who have comparable objective physical findings144. The discrepancy between the nociceptive stimuli from abnormal tissues causing low-back pain and the behavioral and affective component has been referred to as desynchrony127. It has been proposed that, in addition to the neurological and psychological responses, the reaction to pain is a taught behavior based on socialization109. The effects of culture on low-back pain in a nonindustrialized, isolated community were examined in a community of Australian aborigines109. Nearly half of the fifty-six adults had long-term back pain; however, because of their cultural beliefs, they chose not to make their pain public. Rose et al. documented a positive effect of cognitive behavioral therapy, regardless of its duration or specific form, on patients' reports of pain, disability, self-esteem, and somatic awareness180. There is an increasing body of literature demonstrating the beneficial role of cognitive behavioral programs in the treatment of chronic pain81. The diagnosis of chronic low-back pain has been accepted in our society, and industry, government, and traditional medical organizations have developed and supported programs and mechanisms that have the potential to be used for secondary gain; these include disability insurance, Workers' Compensation, and litigation after injury67. The practitioner should be able to screen for the presence of psychological stress in a patient's life that may aggravate or cause back pain, and these factors should be addressed before major operative intervention is performed. Several validated psychometric instruments have been developed for use in the clinical setting143,228. The physician should be familiar with trained professionals and community resources, including counselors and self-help groups, that provide treatment choices to patients who have identifiable psychological stress. Several authors have advocated programs incorporating cognitive therapy with physical and vocational training and disability case management as an effective multimodal approach for patients in whom clear operative indications are lacking98,152,173.
The use of plain radiographs to evaluate patients who have low-back symptoms is indicated when the symptoms have persisted for more than four to eight weeks and are associated with pain at night or at rest. These radiographs are used primarily to rule out infection, a malignant lesion, fracture, and inflammatory conditions. Negative results may help to reassure the patient that no major pathological condition is present114. The greatest debate over the use of diagnostic studies to evaluate low-back pain concerns provocative discography. Several operative procedures have been developed and advocated for the treatment of lumbar pain diagnosed on discography. It is unclear how injection of the disc provokes pain, although various theories have been postulated. One such theory is that stimulation of nociceptive fibers in the innervated part of the annulus fibrosus94,216 or the vertebral end plate or body27,100 causes pain as the injection increases intradiscal pressure. Another theory is that injection leads to chemical stimulation that produces pain211. The use of provocative discography apparently was first reported in 1948 by Hirsch, who injected normal saline solution into lumbar discs in an effort to localize lesions106. Lindblom apparently was the first investigator to inject contrast medium into the disc space, and he reported both normal and abnormal morphological characteristics133. Colhoun et al. retrospectively reviewed the outcomes for 137 patients with low-back pain who had had spinal arthrodesis at levels for which discography demonstrated abnormal morphological characteristics and produced symptoms43. At a mean of 3.6 years, 121 patients (88 percent) had a successful result compared with only thirteen (52 percent) of twenty-five who had had an arthrodesis without reproduction of the typical symptom. Simmons and Segil found discography to be accurate for the localization and planning of arthrodesis of symptomatic levels in 89 percent of 272 patients who had discogenic disease of the spine186. Less-than-optimum results also have been reported for patients who had an arthrodesis on the basis of discography. Wetzel et al. performed discography at seventy-seven levels in forty-eight patients who had low-back pain and found a positive response at sixty-three levels212. Posterolateral arthrodesis was performed at one level in eight patients, at two levels in twenty-six patients, at three levels in thirteen patients, and at four levels in one patient. Those authors reported a satisfactory outcome in 46 percent of the patients and a rate of union of 48 percent; however, as the number of levels that were operated on was not correlated with the discographic findings, interpretation of these results is difficult. Opponents of discography often cite a study reported in 1968 by Holt108. He found a 37 percent rate of false-positive results with regard to an association between the production of symptoms and abnormal morphological characteristics of the disc in thirty asymptomatic inmates of a correctional institution who had volunteered to participate in the study. A total of seventy-two segments were evaluated. These included the second and third lumbar vertebrae in one subject, the third and fourth lumbar vertebrae in twenty-seven, the fourth and fifth lumbar vertebrae in twenty-four, and the fifth lumbar and first sacral vertebrae in twenty. The validity of that study has been questioned. In a critical reassessment, Simmons et al. concluded that Holt's data "should no longer be used as scientific or authoritative evidence against the use of discography."188 Major flaws that they identified included use of a study population consisting of so-called volunteer prison inmates whose reliability may be questionable, the high rate of failure of placement of the needle (twenty-three [77 percent] of thirty attempts at the fourth and fifth lumbar vertebrae and nineteen [63 percent] of thirty at the fifth lumbar and first sacral vertebrae were successful), and use of a contrast agent that is very irritating. Using superior radiographic equipment and nonirritating contrast medium, Milette and Melanson reported that discography was associated with only a 5 percent rate of false-positive results in a study of 320 normal discs156. Walsh et al. described the results of discography in ten asymptomatic subjects who had had evaluation of a total of thirty segments (three segments per subject)206. No subject reported pain with injection, but five had abnormal findings at a single level. The North American Spine Society Diagnostic and Therapeutic Committee, in a position statement on lumbar discography, noted that recent literature supports the use of discography in selected applications95. For patients who have nonradicular low-back pain, these applications include determination of which disc level or levels have reproduction of the symptoms on noninvasive studies, in order to select the levels to be included in the arthrodesis, and localization of the discogenic sources of persistent pain after a previous unsuccessful procedure on the lumbar spine. Those authors believed that positive findings on discography alone should not be an indication for an operation, and they advocated the use of computed tomography after discography to further assess the morphological characteristics of the disc. The natural history of discogenic low-back pain that is diagnosed with discography has not been studied extensively. Smith et al. reported on thirty-six patients with severe, so-called discogram-concordant low-back pain (that is, pain that was reproduced by injection) who, for various reasons, did not have an arthrodesis when it was suggested to them189. Although criticized as being flawed (in an editorial at the end of the article), this report indicated that, over time, most of these patients had a decrease in symptoms and resumed a fairly functional lifestyle. The development of magnetic resonance imaging has provided a noninvasive method for evaluation of the morphological characteristics of the disc. Schneiderman et al. retrospectively studied 101 discs in thirty-two patients and found a 99 percent association between the findings on magnetic resonance imaging and those on discography185. Other authors have shown that magnetic resonance imaging cannot be used preoperatively to predict the outcome of arthrodesis. Zucherman et al. reported on eighteen patients who had normal findings on magnetic resonance imaging and abnormal findings on discography; the abnormal results most often suggested a peripheral tear of the annulus fibrosus, with or without a small disc herniation227. Brightbill et al. reported on seven patients with chronic low-back pain who had normal findings on magnetic resonance imaging but abnormal findings on discography and a finding of internal disc disruption at the time of the operation26. Despite these findings, a morphological abnormality on magnetic resonance imaging in conjunction with a provocative pain response on discography might be predictive of which patients would benefit from operative stabilization. Gill and Blumenthal retrospectively reviewed the results in fifty-three patients with internal disc disruption at the fifth lumbar and first sacral vertebrae who had discogram-concordant pain and had had magnetic resonance imaging86. Thirty-six months after an anterior arthrodesis, twenty-nine (74 percent) of thirty-nine patients who had abnormal findings on preoperative magnetic resonance imaging had a decrease in the symptoms compared with only seven of fourteen who had normal findings. Newman and Grinstead reported an 86 percent rate of success (thirty-one of thirty-six patients) two years after an anterior lumbar arthrodesis performed because of abnormal findings on discography and magnetic resonance imaging163. Parker et al. noted less successful outcomes in twenty-three patients; nine patients (39 percent) had a clinically successful result, and eighteen (78 percent) had fusion170.
Extensive material concerning nonoperative treatment has been published and reviewed56,201. In a recent review of randomized, controlled trials, van Tulder et al. identified only twenty-eight trials on intervention for acute back pain and twenty on intervention for chronic back pain that were considered to be of high quality201. Those authors found strong evidence of the effectiveness of nonsteroidal anti-inflammatory drugs and muscle relaxants and of the ineffectiveness of exercise for the treatment of acute low-back pain. With regard to the treatment of chronic low-back pain, strong evidence, especially of short-term improvement, was reported for exercise, back school, and manipulation. Both the Agency for Health Care Policy and Research13 in the United States and the Clinical Standards Advisory Group Committee on Back Pain181 in the United Kingdom recommended use of acetaminophen, nonsteroidal anti-inflammatory drugs, muscle relaxants, manipulation, and active exercise for the treatment of acute low-back pain. The Quebec Task Force on Spinal Disorders176 and the Clinical Standards Advisory Group Committee on Back Pain181 stated that the primary goal of treatment of chronic back pain is return to work or usual activities, with relief of pain being a secondary goal. At least three randomized trials of bed rest for the treatment of acute low-back pain have been reported52,70,215. The findings suggest that no more than a brief period of bed rest (two days) is advisable for most patients who have moderate symptoms. Additional bed rest may lead to deconditioning and the development of illness-related behavior. Many exercise programs, including those involving extension, flexion, a combination of flexion and extension, aerobic conditioning, and stretching, have been advocated for the treatment of low-back pain; however, the best regimen is not known56. At least two large, randomized, controlled trials have failed to support the efficacy of flexion exercises63,84, originally popularized by Williams218. Although widely used and somewhat supported in the literature, randomized trials of extension-exercise regimens, such as those reported by Donelson et al.57 and by McKenzie138, have had problems related to classification and design, making the results difficult to interpret56. The role of aerobic conditioning is not clear; however, one study predicted a lower risk of back problems among individuals who have higher levels of fitness34. The use of traction for the treatment of low-back pain was not supported in at least seven randomized clinical trials46,129,150,151,167,209,210. A comprehensive review of the literature on the use of acupuncture for the treatment of chronic pain concluded that the efficacy of this modality was unclear199, and other studies have suggested that it is a form of placebo therapy56. Several well controlled, randomized studies have failed to provide convincing evidence of the efficacy of transcutaneous electrical nerve stimulation53,102,145. The Quebec Task Force on Spinal Disorders concluded that, despite the common use of braces and corsets, there is no substantial scientific evidence of their efficacy176. A similar conclusion was reached regarding the use of electromyographic biofeedback176. There is evidence that postural devices that maintain lumbar lordosis while the patient is sitting lead to a decrease in back pain and pain in the lower extremities217. Drug therapy is commonly used in clinical practice for the treatment of low-back pain. Although nonsteroidal anti-inflammatory drugs have been shown to be effective for the treatment of acute back pain2,4,12,104, their effectiveness for chronic back pain is unclear. Narcotic agents may be useful for acute back pain, but their use is generally avoided for chronic back pain, although some authors have seen a role for them169. Muscle relaxants appear to have some efficacy for the treatment of acute low-back pain20,105. The literature on spinal manipulation has been reviewed extensively201. The effectiveness of this treatment is controversial, but there is some evidence to support its efficacy for the treatment of acute back pain14. Functional restoration, work-hardening, and back-school programs for the treatment of chronic low-back pain have received much attention. Back school provides a structured educational program leading to behavioral modification and increased function in order to prevent and reduce future episodes of back pain. Cohen et al., in an overview of the literature, found insufficient evidence to recommend back school for patients who have low-back pain42, whereas other authors have found back school to be beneficial for those who have chronic back pain13,117,153,201. Functional restoration and work-hardening programs focus on rehabilitation, with an emphasis on return to work and improved function despite continuing pain. Such programs are intensive (forty to fifty hours per week for three to four weeks), combining education, behavioral training, endurance and physical training, and work simulation, and they have shown some promise. However, properly performed studies of their results are lacking, and the programs are expensive57,98,157,197. Pain centers vary widely with regard to their emphasis: some are oriented to a specific modality, others use a multidisciplinary approach, and still others deal with chronic pain. Therefore, the effectiveness of these programs has been difficult to assess68.
Considerable controversy exists regarding specific operative treatment for low-back pain. Operative intervention usually is considered when nonoperative measures have been exhausted and disability and pain are still present. A moderate amount of experience has been accumulated with use of posterior arthrodesis alone, anterior disc excision and interbody arthrodesis, posterior lumbar disc excision and interbody arthrodesis, and circumferential stabilization procedures14,37,80,86,93,118,134,163,193,224. The use of arthrodesis for the treatment of low-back pain may be associated with morbidity. Deyo et al. reviewed the records of a group of Medicare patients who had had spinal procedures and reported nearly twice as many complications for those who had had an arthrodesis as for those who had not55.
Posterolateral Arthrodesis Techniques for posterolateral spinal arthrodesis have been modified since they were first reported in 1911 by Hibbs103. His technique of posterior arthrodesis involved use of graft from the spinous processes that was applied to the laminae as a local autogenous graft. This technique is now used rarely because it has been associated with a high rate of pseudarthrosis71. Cleveland et al., in 1948, described use of a more laterally placed bone graft for the treatment of pseudarthrosis following a posterior spinal arthrodesis39. Watkins popularized this technique in 1959, using a posterolateral approach to the spine and thus gaining direct access to the posterior articulations, the pars interarticularis, and the transverse processes207. Macnab and Dall reported on the anatomy and blood supply of the facet joints and transverse processes of the lumbar spine141. These observations provided a basis for the enhancement of fusion by facilitation of a maximum blood supply while minimizing blood loss during such procedures. There are many variations of the techniques for posterior and posterolateral arthrodesis. McBride stressed the importance of the facet joint in 1949, when he described his results with use of an adjunctive morseled transfacet bone block for lumbosacral arthrodesis137. This simple technique, which represents an attempt to eliminate intervertebral motion at the facet articulation, recently was questioned by Boden et al., who performed cadaver studies analyzing the axial and torsional stiffness associated with various techniques of arthrodesis17. Those authors concluded that disruption of the facet-joint capsule and placement of transfacet bone blocks as described by McBride resulted in loss of spinal stability as compared with use of posterior intertransverse-process arthrodesis alone.
Pedicle-Screw Instrumentation We believe that adjunctive spinal instrumentation increases the rate of fusion and that Food and Drug Administration class-II devices and pedicular instrumentation are comparable with regard to the rates of fusion, patient satisfaction, and complications. Yuan et al., in 1994, retrospectively analyzed the outcomes for patients with degenerative spondylolisthesis who had been managed with pedicle-screw fixation and compared the results with those for patients who had not had pedicle-screw instrumentation at the time of the operation223. A total of 2684 patients were included in the study, and 2177 (81 percent) had pedicle-screw fixation. There were two control groups, consisting of 456 patients who had an arthrodesis without instrumentation and fifty-one who had instrumentation with non-pedicle-screw devices. Intraoperative events related to the pedicle-screw fixation occurred infrequently, and the rate of breakage of the implant (0.2 percent) was extremely low. The rates of nerve-root, spinal cord, and vascular injuries also were low (less than 0.5 percent). Dural tears related to insertion of the screw were rare (prevalence, 0.1 percent), and the rates of tears that were not associated with placement of the screw were comparable between the groups that were managed with and without instrumentation. The rate of reoperation was higher in the patients who had pedicle-screw instrumentation than in those who did not (18 compared with 15 percent). Most of these procedures were related to removal of the device and reinstrumentation, and the rates of repeat arthrodesis were similar among the groups. At the time of the two-year follow-up evaluation, the patients who had been managed with pedicle-screw instrumentation had a significantly higher rate of fusion than the control group that had not had instrumentation (83 compared with 75 percent; p < 0.01). In addition, the patients who had had pedicle-screw instrumentation tended to have a more rapid rate of consolidation of the fusion and more often had maintenance of spinal alignment. They also had better clinical outcomes with regard to pain, function, and neurological recovery than did the patients in the control group that was managed without instrumentation. The rates of nondevice-related events, reoperation, and death were similar for the treatment groups. Yuan et al. concluded that the benefits of pedicle-screw use were substantially greater than the potential risks. Zdeblick, in a prospective, randomized study of patients who had degenerative conditions of the lumbar spine, compared the results of posterolateral arthrodesis with and without segmental (pedicle-screw) instrumentation224. Superior results were found with use of rigid fixation. The rate of fusion for the patients who were managed with segmental instrumentation was 86 percent (sixty-two of seventy-two) compared with 64 percent (twenty-seven of forty-two) for those in whom no instrumentation was used. Hall et al. described their experience with pedicle-screw instrumentation in 120 patients who had so-called degenerative disc disease97. Various criteria had been used to select the patients, and 63 percent had had a previous operation on the lumbar spine. The arthrodesis included one level in thirty-two patients, two levels in forty-two, three in twenty-nine, four in eleven, and five levels or more in six. Results related to the number of levels included in an arthrodesis are difficult to analyze. The overall rate of clinical success at twenty-four months was 73 percent (seventy-eight of 107 patients), the rate of fusion was 91 percent (ninety-seven of 107), and the rate of intraoperative complications was 11 percent.
Lumbar Interbody Arthrodesis
Posterior Lumbar Interbody Arthrodesis
Anterior Lumbar Interbody Arthrodesis Penta and Fraser172 recently reported that 68 percent of 108 patients had an outcome of fair or better after anterior lumbar interbody arthrodesis, according to the Low Back Outcome Score88, at a minimum of ten years. The overall rate of fusion was 72 percent (91 percent for patients who had had a single-level arthrodesis and 51 percent for those who had had the procedure at multiple levels), which is similar to findings in other studies14,37,86,90,118,163,193. A lower rate of success was reported by Knox and Chapman, who analyzed the results for twenty-two patients who had had an anterior lumbar interbody arthrodesis for the treatment of discogram-concordant pain118. Of the seventeen patients who had had a single-level arthrodesis, six had a good result, three had a fair result, and eight had a poor result. There were no excellent results; all five patients who had had the arthrodesis at two levels had a poor result. In a recent study, Greenough et al.90 prospectively reviewed the results of posterolateral lumbar arthrodesis with instrumentation in 135 patients and compared these outcomes with those from a similar study by Penta and Fraser172, who had performed an anterior interbody arthrodesis in 108 patients; both groups of investigators used the same outcome measure (the Low Back Outcome Score88). Overall, there was a twofold difference in the outcome score (Wilcoxon rank-sum test, Fisher exact probability test, and chi-square test)90, with anterior interbody arthrodesis yielding better results regardless of the patient's Workers' Compensation status.
Interbody Fusion Cages The results of two large trials in which Investigational Device Exemption protocols developed by the Food and Drug Administration were used to evaluate these devices have been reported recently124,177. As part of such a protocol, Ray prospectively followed 236 patients who had disabling low-back pain and had been managed with insertion of threaded titanium interbody-fusion cages (the Ray-TFC device) through a posterior approach177. After a minimum of twenty-eight months, the rate of fusion was 94 percent (221 patients), and functional improvement, according to a validated outcome device, was excellent for 40 percent of the patients, good for 25 percent, fair for 21 percent, and poor for 14 percent. The rate of persistent complications after discharge from the hospital was less than 1 percent. Paralysis was not reported. A deep infection developed in two patients, and two patients needed a reoperation. Kuslich et al., in a similar trial, reported on 947 young and middle-aged patients who had had implantation of the BAK titanium device with use of either an open anterior approach (591 patients) or a posterior approach (356 patients)124. At twelve months 680 patients were available for follow-up, and at twenty-four months 299 were available. There were no instances of paralysis, implant or procedure-related death, or deep infection. Major complications were reported in 2 percent of the patients, and migration of the implant necessitating reoperation occurred in 1 percent. Although the findings of that report were difficult to verify, those authors indicated that fusion occurred in 86 percent of 680 patients at twelve months and in 91 percent of 299 at twenty-four months. At the two-year follow-up evaluation, more than 85 percent of 299 patients reported a decrease in pain, and 91 percent had an improvement in function according to a nonvalidated analog scale. The operative approach and the number of levels included in the arthrodesis were not associated with the functional improvement score. The rate of return to work at twelve, twenty-four, and thirty-six months was 68 percent of 680 patients, 78 percent of 299 patients, and 91 percent of 143 patients, respectively. The availability of these devices has had a major impact. Industry analysts stated that interbody fusion cages would account for 23 percent of the spinal implant market in the United States in 1997 compared with 3 percent in 1996154. Hacker compared the outcomes and costs of posterior lumbar interbody arthrodesis with use of the BAK device in fifty-four patients with those of a 360-degree arthrodesis (anterior lumbar interbody arthrodesis and posterolateral arthrodesis without instrumentation) for the treatment of disabling low-back pain in twenty-one patients96. The patients who had been managed with a posterior lumbar arthrodesis had a more rapid return to work and more rapid closure of their Workers' Compensation claims as well as substantial cost-savings. The outcomes were similar for the two groups. Similar results have been reported with use of the Ray-TFC device178.
Circumferential Arthrodesis
Intervertebral Disc Prosthesis In addition to ensuring adequate material properties and patient outcomes over the long term, the difficulties and complications associated with operative implantation of a prosthetic device and revision or salvage procedures must be acceptable. The many devices that are currently under investigation31,32,49,60,99,125,200,204 may be divided into two basic groups: those used to replace all or part of the annulus fibrosus and nucleus pulposus of the disc, and those used to replace the nucleus pulposus only. Fernstrom retrospectively reported the results for 125 patients in whom 191 spherical alloy implants had been used to replace the nucleus pulposus66. The goal was to restore normal weight-bearing to the annulus fibrosus and the facet joints and to restore disc height. Four to seven years after the operation, 88 percent of the prostheses had subsided into the vertebral end plates. The results associated with the use of a dome and cup-shaped, low-friction gliding prosthesis in different stages of development have been reported by several authors32,38,49,91,226. A satisfactory result was noted in more than 50 percent of patients with use of nonvalidated measures. Complications included migration or dislocation of the implant and fracture of the metal.
The available information on the diagnosis and treatment of low-back pain lacks objectivity, as most studies have been retrospective, uncontrolled, and predominantly focused on operative technique. Most investigators have used nonvalidated outcome measures that have reflected the biases of their proponents. So-called facet disease and degenerative instability are extremely rare, and these diagnoses should seldom be used as indications for operative intervention. When they are present, stabilization with use of posterior instrumentation and arthrodesis is appropriate. The pertinent issues revolve around the treatment of idiopathic and discogenic low-back pain. Major controversies center around which treatment is best when nonoperative measures have failed. Opponents of operative treatment believe that the outcomes of most or all procedures fall below the threshold needed to justify operative intervention, as only about 50 percent of patients have a successful result with regard to decreased pain, increased function, and ability to return to work. Proponents of operative treatment maintain that these patients have no other options, that some improvement in pain and function is important, and that operative intervention is therefore warranted. They believe that full relief of disability is too strict a criterion or goal. We think that most patients who have acute or chronic idiopathic or discogenic low-back pain should be managed nonoperatively. Patients who have refractory pain with severe incapacity, and those who have imaging-confirmed morphological changes and concordant symptoms, may be managed successfully with anterior disc ablation and structural arthrodesis. Some improvement occurs as a result of operative treatment in about 75 percent of patients, but major or complete relief of pain and recovery of function are seen in 50 percent or less. Each physician and each patient must assess these issues in a forthright manner and determine what is appropriate under the circumstances.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
|
Top
Introduction
Epidemiology and Economics