 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1998 The Journal of Bone and Joint Surgery, Inc.
Current Concepts Review - Degenerative Lumbar Spinal Stenosis* , NEW YORK, N.Y.
Over the last ten years, the median age of the United States population has increased by 1.4 years, with more than two million additional people who are sixty-five years old or more91. As the population continues to age, more older people are maintaining an active lifestyle. Consequently, functional limitation and pain due to symptomatic degenerative disease of the spine is becoming more common. Lumbar spinal stenosis remains one of the most frequently encountered, clinically important degenerative spinal disorders in the aging population.
Lumbar spinal stenosis is defined as a narrowing of the spinal canal that produces compression of the neural elements before their exit from the neural foramen3,84. The narrowing may be limited to a single motion segment (two adjacent vertebrae and the intervening intervertebral disc, facet joints, and supporting ligaments) or it may be more diffuse, spanning two motion segments or more. Lumbar spinal stenosis can be classified on the basis of either etiology or anatomy. The original etiological classification, described by Arnoldi et al., distinguishes congenital or developmental stenosis from acquired or degenerative spinal stenosis3. Congenital or developmental stenosis is due to either idiopathic narrowing of the spinal canal or developmental narrowing secondary to a bone dysplasia such as achondroplasia. Acquired or degenerative stenosis also may occur as a result of an underlying metabolic disorder such as Paget disease, a tumor, an infection, post-traumatic osteoarthrotic changes, or instability with spondylolisthesis following a previous operation. Anatomical classifications of lumbar spinal stenosis are used to identify specific areas of narrowing of the spinal canal and are particularly useful as guides for operative decompression. The anatomy of the spinal canal at each vertebral segment can be understood better by dividing the canal into a series of transverse regions (three levels from cephalad to caudad) and sagittal regions (three zones from the midline laterally) (Fig. 1). From cephalad to caudad, the three transverse levels are the pedicle level, the intermediate (vertebral body) level, and the disc level. The pedicle level extends from the superior to the inferior border of the pedicle. The intermediate level begins at the inferior border of the pedicle and extends caudally to the inferior end plate of the vertebra. The disc level begins at the inferior end plate and extends caudally to the superior border of the next pedicle. From the midline laterally, the three sagittal zones are the central zone, the lateral-recess zone, and the pedicle zone. The central zone is the area between the normal lateral borders of the non-compressed dural sac. The lateral-recess zone is the area between the lateral border of the dural sac medially and a longitudinal line connecting the medial edges of the pedicles laterally. The pedicle zone is the area between the medial and lateral borders of the pedicle. This grid-like subdivision of the spinal canal is similar to a system that was recently described for the reporting of the location of spinal lesions94.
Two important anatomical areas, the spinal nerve-root canal and the intervertebral nerve-root foramen, lie bilaterally within this multisegmental grid. The spinal nerve-root canal lies within the lateral-recess zone, beginning with the origin of the nerve-root sleeve at the disc level and ending as the nerve root passes along the inferomedial border of the pedicle at the cephalad aspect of the intermediate level. The intervertebral foramen lies within the pedicle zone; the superior portion of the foramen is located at the intermediate level and the inferior portion is located at the disc level. The intraspinal pathway of the nerve root and the spinal nerve is formed by the lateral-recess zone at the disc level, the pedicle level as it extends to the cephalad aspect of the intermediate level, and the intervertebral foramen at the intermediate level. These three areas correspond to the entrance zone, the mid-zone, and the exit zone described by Lee et al.54. Central spinal stenosis commonly occurs at the disc level, as a result of overgrowth in the region of the facet joint (mainly involving the inferior articular process of the cephalad vertebra) and thickening and redundancy of the ligamentum flavum. Lateral spinal stenosis affects the lateral-recess zone and the intervertebral foramen. Lateral-recess stenosis, which occurs as a result of degenerative changes similar to those associated with central spinal stenosis, affects the spinal nerve-root canal at the disc level and the superior aspect of the pedicle level. Lateral-recess stenosis at the inferior aspect of the pedicle level is uncommon, but it can occur secondary to hypertrophic granulation tissue from the posteriorly located pars interarticularis in patients who have spondylolytic defects. Stenosis of the intervertebral foramen is most common at the disc level; thus, it usually begins in the inferior portion of the intervertebral foramen. Such stenosis becomes clinically important, however, only when it involves the superior aspect of the foramen at the intermediate level. It is at this level that the exiting nerve root passes laterally, inferior to the pedicle, where it can be compressed by disc material or hypertrophic bone in the form of an osteophyte from the inferior aspect of the cephalad vertebra or from the superior articular process of the caudad vertebra. Stenosis is rarely encountered at the pedicle level, except in patients who have extruded disc material in this region, congenitally short pedicles, or spondylolytic defects (as already mentioned).
Adult degenerative lumbar spinal stenosis is almost always associated with osteophytic enlargement in the region of the degenerating synovial facet joints. The general anatomical alignment of each articular process is similar for all lumbar facet joints and is responsible for the pathoanatomical changes associated with this condition. The inferior articular process of the cephalad vertebra of the motion segment is located posteriorly and medially and contributes to the lateral osseous wall of the central spinal canal. The formation of osteophytes on that process results in focal narrowing of the central spinal canal. The superior articular process of the caudad vertebra of the motion segment is located laterally and anteriorly. Osteophytic enlargement of that process in response to degeneration of the surfaces of the facet joints results in narrowing of the lateral recess and the neural foramen. For example, the formation of osteophytes in the region of the facet joints between the fourth and fifth lumbar vertebrae can affect the proximal origin of the sleeve of the fifth lumbar nerve root in the lateral recess and the fourth lumbar nerve root as it exits through its neural foramen. Degeneration of the facet joints may be the result of segmental instability, usually in the form of abnormal segmental motion rather than excessive translational or angular motion49. The initiating factor is believed to be desiccation and degeneration of the disc followed by collapse of the disc space. This results in abnormal kinematics of the motion segment and subsequent osteoarthrosis of the facet joints, with the formation of osteophytes along the margins of the vertebral bodies and facet joints. The orientation of the facet joints becomes more sagittal with the formation of new bone in the anteromedial aspect of both the superior articular process and the inferior articular process. The new bone forms in response to sagittal, anteriorly directed shear stresses imposed by flexion71. Degeneration of a motion segment with narrowing of the disc space causes a relative shortening of the lumbar spinal canal, decreasing the overall volume of space available for the cauda equina. Additional loss of volume occurs as a result of the segmental narrowing caused by bulging of the incompetent discs and infolding of the ligamentum flavum. Other anatomical factors can also play a role in the development of degenerative lumbar spinal stenosis. In most symptomatic patients, the cross-sectional area of the spinal canal is at the low end of the normal range of sizes and there is a limited capacity to accommodate the additional narrowing produced by degenerative disease32,71. An initially diminished size of the spinal canal, especially at the level of the third lumbar vertebra, recently was linked to a variety of antenatal factors; the strongest link was to a shortened gestational age66. Focal protrusions of intervertebral discs often contribute to the overall narrowing of the spinal canal or the intervertebral foramen, or both. Degenerative cysts in the region of a synovial facet joint also can produce a mass lesion, which can contribute to additional narrowing of the canal (Figs. 2-A and 2-B). These cysts often compress nerve roots and produce radiculopathy similar to that caused by focal herniation of a disc. They also can produce intermittent or activity-related pain that is more typical of lumbar spinal stenosis.
Lumbar spinal deformity can develop as a result of degenerative disease in adulthood and can be a major contributing factor to the narrowing of the spinal canal. Unlike idiopathic adolescent lumbar scoliosis, which occurs more commonly on the left side, adult degenerative scoliosis occurs on the right and left sides with equal frequency31. Adult degenerative scoliosis develops as a result of asymmetrical narrowing of the disc space and vertebral rotation secondary to the instability caused by degeneration of the disc. These degenerative changes of worsening scoliosis also can occur in the presence of an underlying mild idiopathic lumbar curve or at the base (the lumbar component) of a King type-II idiopathic thoracic curve, both of which occur more commonly on the left side48. Neural compression associated with adult lumbar scoliosis is commonly manifested as radicular pain that may be related to physical activity. The rotation at the apex of the curve is associated with hypertrophy and subluxation of the facet joints in the concavity of the curve. Also, collapse in the concavity results in narrowing of the neural foramen between adjacent pedicles. As a result, symptoms in the anterior portion of the thigh and leg (resulting from compression of the cephalad and middle lumbar nerve roots) are more common on the side of the concavity of the major lumbar curve. Radiating pain in the posterior portion of the lower extremity is more common on the side of the convexity of the lumbar curve; such pain is due to compression of the caudad lumbar nerve roots and the sacral nerve roots well caudad to the apex, as the spine curves back to meet the pelvis. Degenerative spondylolisthesis, seen most commonly at the level between the fourth and fifth lumbar vertebrae, is a common cause of focal lumbar spinal stenosis. Sanderson and Fraser, in a study of patients who were more than fifty years old, reported that the prevalence of degenerative spondylolisthesis was nearly twice as high in parous women as it was in nulliparous women76. It remains unclear whether this trend was due to the hormonal or mechanical effects of pregnancy. The prevalence in men was less than half of that in nulliparous women76. Degenerative spondylolisthesis is associated with anatomical variations that result in limited mobility of the motion segment between the fifth lumbar and first sacral vertebrae as well as with a relatively sagittal orientation of the facet joints between the fourth and fifth lumbar vertebrae30. It is thought that these anatomical variations result in increased shear force on the motion segment between the fourth and fifth lumbar vertebrae during flexion and that the sagittally oriented facet joints at this level are less able to resist this translating force. The anterior vertebral subluxation results in severe narrowing of the spinal canal between the inferior aspect of the lamina and inferior articular process of the fourth lumbar vertebra and the superior aspect of the posterior part of the fifth lumbar vertebral body. The spondylolisthesis usually is limited to a maximum of 25 to 30 per cent of the anterior-posterior diameter of the superior end plate of the fifth lumbar vertebral body when the intact posterior elements of the fourth lumbar vertebra abut the fifth lumbar vertebral body. The resultant narrowing, combined with hypertrophy of the facets, affects both the lateral recesses and the central spinal canal. The fourth lumbar nerve roots usually are spared but may be involved when there is severe disc-space collapse and foraminal narrowing.
The classic presentation of lumbar spinal stenosis is that of bilateral neurogenic claudication as well as intermittent pain radiating to the thigh or the leg, or both, that is worse with prolonged standing, activity, or lumbar extension and that is relieved by sitting, lying down, or lumbar flexion31. In addition, patients often describe intermittent burning, numbness, heaviness, or weakness radiating to the lower extremity. Neurogenic claudication can be differentiated from claudication due to arterial insufficiency by evaluating the effect of standing; standing worsens the symptoms of neurogenic claudication but not those of vascular claudication. Also, the symptoms of vascular claudication are not affected by changes in the posture of the lumbar spine. Riding a stationary exercise bicycle (while maintaining lumbar flexion) elicits pain in patients who have vascular claudication but not in those who have neurogenic claudication21. Exercise treadmills have been described for use both as a diagnostic tool for detecting the presence of neurogenic claudication and as an indicator of functional status and of the response to operative treatment18,26. On examination, patients with vascular claudication have weak or absent pulses and are likely to have the typical skin changes (mottled discoloration, thinning, and shininess) that are indicative of vascular insufficiency in the lower extremities. Many patients with lumbar spinal stenosis have unilateral symptoms, which may or may not be related to physical activity. A more constant radicular pain resulting from compression of a nerve is common and often occurs in addition to the more long-standing symptoms of numbness, weakness, and pain in the lower extremity that are related to activity. The specific symptoms that are referable to the lower extremity differ somewhat between central spinal stenosis and lateral spinal stenosis41. In the study by Jönsson and Strömqvist, patients with lateral stenosis were less limited in terms of walking ability but had a greater prevalence of pain at rest and at night41. Back pain also is common in patients who have degenerative lumbar spinal stenosis and is more likely to be related to degenerative disc disease than to compression of a nerve. Back pain associated with changes in position and with lifting or bending may be indicative of underlying spinal instability associated with degenerative scoliosis or spondylolisthesis. Pain in the groin or the knee may be a sign of osteoarthrosis of the hip or the knee, and the symptoms of underlying spinal stenosis may become apparent only with increased activity after total joint arthroplasty57. Patients who have lumbar spinal stenosis tend to walk with a stooped forward gait and even maintain this posture while standing. Characteristically, the pain is exacerbated with lumbar extension and is relieved with lumbar flexion. Studies of cadavera have demonstrated that the capacity of the dural sac is significantly larger during flexion than during extension (p < 0.001)15. Baba et al. reported on a small series of patients with lumbar spinal stenosis who had priapism in addition to neurogenic claudication6. This unusual symptom of autonomic dysfunction, like sphincteric dysfunction, indicates severe compression and usually is a consequence of long-standing stenosis. The priapism was relieved postoperatively in six of the seven patients in that series6. The clinical signs of spinal stenosis overlap substantially with those of symptomatic herniation of a lumbar disc41. Limited spinal mobility and a positive nerve-root tension sign (such as pain in the lower extremity during straight-leg raising) are seen more commonly in patients who have a herniated disc than in those who have spinal stenosis. Weakness of the extensor hallucis longus is found in about half of the patients in each of these groups, but reduced reflexes in the knee and the ankle are more common in patients who have central spinal stenosis than in those who have lateral spinal stenosis or a herniated disc. Changes in reflexes in the lower extremity also may occur as part of the normal process of aging. Disturbances in sensory function (such as light touch, two-point discrimination, and proprioception) are more common in patients who have degenerative lumbar spinal stenosis. The pain in the back and the lower extremity that is associated with lumbar spinal stenosis must be differentiated from similar pain that is associated with other conditions not previously mentioned. Disorders of the spinal column that cause a similar type of pain include tumors (primary, metastatic, and epidural), Paget disease, and infection (vertebral, disc-space, and epidural) (Figs. 3-A and 3-B). Extraspinal causes of a similar type of pain include lumbar plexopathy, soft-tissue tumor with neural compression, and peripheral neuropathy. Non-neural causes of a similar type of pain include trochanteric bursitis and osteoarthrosis of the hip or the knee. Peripheral vascular disease may coexist with lumbar spinal stenosis, making it difficult to determine the cause of the pain85. All patients who have neurogenic claudication, even those in whom the symptoms may be related to posture, should be evaluated for vascular insufficiency if peripheral pulses cannot be palpated in the lower extremity.
Radiographic examination of a patient in whom lumbar spinal stenosis is suspected often demonstrates multilevel spondylosis, which may not be associated with stenosis of the spinal canal. Radiographic findings that are more suggestive of spinal stenosis include degenerative spondylolisthesis and degenerative lumbar scoliosis. For patients who have degenerative spondylolisthesis, lateral radiographs that are made with the patient recumbent and the spine in flexion and extension are useful for determining whether there is abnormal motion at the involved level. For those who have associated degenerative scoliosis, anteroposterior and lateral radiographs that are made with use of a long plate and with the patient standing erect are helpful for assessing the magnitude of the curve as well as the relative spinal balance in both the coronal plane and the sagittal plane. Correction of any imbalance is an important goal in the operative management of these patients. The diagnosis of lumbar spinal stenosis can be confirmed with use of computerized tomography or magnetic resonance imaging. Computerized tomography is the most cost-effective single test for establishing the diagnosis of lumbar spinal stenosis. It provides excellent osseous detail, especially in the region of the lateral recess. It also can differentiate the disc and ligamentum flavum from the thecal sac and can provide excellent visualization of far lateral disc abnormalities that may coexist with stenosis of the spinal canal. Magnetic resonance imaging provides superior visualization of the soft-tissue elements of the spinal canal and is especially useful for the evaluation of abnormalities of the intervertebral disc. Its diagnostic accuracy is superior to those of myelography and plain computerized tomography68, and it is as accurate and sensitive as myelography followed by computerized tomography7. The combination of axial and sagittal images allows for complete evaluation of the central spinal canal and the neural foramen (Figs. 4-A and 4-B). Magnetic resonance images of the lateral recess can be more difficult to interpret than computerized tomographic scans. Regions of very low signal intensity on T2-weighted images caused by sclerotic osteophytes can lead to overestimation of the amount of true osseous stenosis. Magnetic resonance imaging of the scoliotic lumbar spine often is suboptimum as axial images cannot be made in the proper plane parallel to the involved disc spaces68. The combination of high-quality magnetic resonance imaging and plain computerized tomography can provide a complete evaluation of the lumbar spine and often obviates the need for preoperative myelography.
Lumbar myelography followed by computerized tomography is currently recommended as the most effective way to obtain a complete anatomical evaluation of compression of neural elements in the lumbar spine. Although myelography is the most specific of the advanced imaging techniques for the diagnosis of lumbar spinal stenosis7, it is invasive and is used mainly as a preoperative study, when needed; the diagnosis of lumbar spinal stenosis is initially established by either computerized tomography or magnetic resonance imaging performed without the use of contrast medium. Lateral myelograms made with the spine in flexion and extension often demonstrate a dynamic component of the stenosis that may be due to segmental instability (more common during flexion) or encroachment on the spinal canal by bulging discs and ligamentum flavum (more common during extension). A computerized tomographic scan made after a myelogram can provide a better evaluation of the degree of central spinal and lateral-recess stenosis, and, as previously stated, provides visualization of foraminal and far lateral disc abnormalities. The importance of correlating radiographic abnormalities with clinical symptoms and signs must be emphasized. In one study, magnetic resonance imaging demonstrated evidence of anatomical stenosis of the lumbar spine in three of fourteen asymptomatic individuals who were sixty years old or more8. A patient in whom the predominant symptom is back pain, with or without symptoms or signs referable to the lower extremity, is much less likely to be satisfied after operative decompression than a patient in whom the predominant symptoms are in the lower extremity, even if spinal stenosis is found on an advanced imaging study46. Electrophysiological studies, including electromyography, testing of nerve-conduction velocity, and evaluation of somatosensory evoked potentials, are not considered part of the routine assessment for establishing the diagnosis of lumbar spinal stenosis. In patients who have diabetes and lumbar spinal stenosis, however, electromyography and evaluation of somatosensory evoked potentials can be useful for differentiating between radiculopathy due to compression of a nerve by the spinal canal and diabetic neuropathy affecting the peripheral motor and sensory nerves. Electromyography also can be useful for differentiating active denervation from chronic inactive changes in peripheral nerves. Normal findings on electrophysiological studies do not rule out the presence of symptomatic lumbar spinal stenosis. However, the more typical pattern in symptomatic patients is that of a polyradiculopathy, often with bilateral involvement of multiple levels43. Also, the evaluation of somatosensory evoked potentials both before and after exercise may help to determine which nerve roots are most involved in a patient who has stenosis of the central spinal canal at the lumbar level.
In the past, early operative intervention was recommended for the treatment of symptomatic lumbar spinal stenosis as it was thought that this disorder was always progressive; however, in a recent study of thirty-two patients who had been advised to have operative treatment but who were managed with observation only, the condition was found to be relatively stable39. Specifically, fifteen (47 per cent) of the thirty-two patients had clinical improvement, twelve (38 per cent) had no change in symptoms, and five (16 per cent) had worse symptoms after a mean duration of follow-up of four years. Twenty-seven of the thirty-two patients described their symptoms with use of a visual-analog scale; nineteen patients (70 per cent) indicated that there had been no change in their clinical status, four (15 per cent) indicated that their clinical status had improved, and four indicated that their clinical status had worsened. No new neurological deficits were found, although eight (38 per cent) of twenty-one patients had evidence of neural deterioration on follow-up electrophysiological examination. The mainstays of non-operative treatment include the use of medication, the modification of activities, the use of a brace, and the epidural injection of steroids; however, to my knowledge, none of these methods has been demonstrated to be efficacious on its own in a prospective clinical trial. Non-steroidal anti-inflammatory medications may have an analgesic effect in addition to their potential to mitigate the inflammatory response associated with neural compression. This group of drugs is commonly prescribed or recommended, although I am not aware of any study that has shown them to be of definite benefit for the treatment of lumbar spinal stenosis. The injection of calcitonin was found, in a double-blind, placebo-controlled crossover study, to be effective for diminishing pain and increasing walking distance22. It was less effective and thus was not recommended for patients who had pain at rest and severe limitation with regard to the distance that they could walk (that is, patients who could walk less than 300 meters) before treatment. The most useful forms of physical therapy in the treatment of lumbar spinal stenosis are considered to be therapeutic stretching and strengthening exercises for the lumbar spine as well as aerobic fitness training. Riding a stationary bicycle can be very useful, as patients often can better tolerate exercises that are done with the lumbar spine in flexion. Walking on a treadmill with use of a harness designed to unload the lumbar spine may also be useful for patients who have lumbar spinal stenosis25. Physical therapy modalities designed for the localized treatment of soft tissues, including the application of heat or ice, ultrasound, massage, electrical stimulation, and traction, are commonly used but have not yet been proved to be effective for any disorders of the lumbar spine. However, they may be beneficial as adjuncts to help to mobilize the lumbar spine in preparation for more intensive exercise therapy. The use of a rigid plastic brace to decrease lumbar lordosis was evaluated by Willner in a study of forty-eight patients who had low-back pain93. After wearing the brace for an average of one year, two of the seven patients who had lumbar spinal stenosis were totally free of pain, four had less pain, and one had no change in pain. The brace was found to be more effective for the treatment of painful spondylolisthesis (thirteen of fifteen patients had complete relief of pain and two had less pain) and less effective for the treatment of chronic low-back pain of unknown etiology (four of twenty-six patients had complete relief of pain, five had less pain, and seventeen had no change in pain). The use of epidural injections of a steroid to treat disorders of the lumbar spine, including stenosis, remains controversial23,51,62,75. In general, the results have been poor when such injections have been used to treat radicular pain. Cuckler et al., in a prospective, randomized, double-blind series of seventy-three patients, found no significant difference between patients who had been managed with epidural injections of a steroid (the experimental group) and those who had been managed with epidural injections of saline solution (the control group) with regard to the relief of radicular pain14. Rosen et al., in a retrospective series of forty patients who had had epidural injections of a steroid, reported that twenty-four patients (60 per cent) had short-term relief of symptoms and only ten (25 per cent) had long-term relief73. Hoogmartens and Morelle reported that an excellent or good result was obtained for only twelve (32 per cent) of thirty-eight patients who had received an average of five epidural injections in the sacral region for the treatment of lumbar spinal stenosis38. Ciocon et al., in a study of thirty patients with lumbar spinal stenosis who had received caudal epidural injections of a steroid each week for three successive weeks, reported that pain was relieved for as long as ten months after treatment13. Only three of the thirty patients in that study had severe anatomical stenosis; in the remaining twenty-seven patients, the stenosis was considered to be mild or moderate. Warfield and Crews reported that a favorable response (even a transient one) to the epidural injection of a steroid is a good prognostic indicator of the pain relief that can be expected after the injection of chymopapain for the treatment of radiculopathy associated with a herniated lumbar disc92. In the study by Derby et al., a favorable response to the injection of a steroid was associated with a favorable outcome after operative treatment in eleven of thirteen patients who had had radicular pain for more than one year (a positive predictive value of 85 per cent)19. An unfavorable response to the injection of a steroid was associated with an unfavorable outcome after operative treatment in thirty-six of thirty-eight such patients (a negative predictive value of 95 per cent). The response to treatment with the injection of a steroid was not predictive of the operative outcome in patients who had had radicular pain for less than one year. Epidural injections of a steroid into the lumbar region have been associated with serious complications, including transient paresis and paralysis, epidural hematoma and infection, and chemical meningitis38,56,62,75. Successful results have been reported after the non-operative treatment of lumbar spinal stenosis. Herno et al., in a retrospective study of fifty-four case-matched pairs of patients, reported that non-operative treatment was as effective as operative treatment overall, although the outcome in men was significantly better after operative treatment (p = 0.027)35. In that study, however, the non-operative treatment did not follow a specific protocol and it was not described. In addition, the operatively and non-operatively treated groups were not randomized and therefore it is unlikely that the two groups had the same average severity of symptoms before treatment. Onel et al., in a prospective study of 145 patients, reported that clinical parameters, such as pain and walking distance, improved after a regimented program of physical therapy and injections of salmon calcitonin during a one-month stay in the hospital65. As lumbar spinal stenosis is not life-threatening and rapid catastrophic neurological deterioration is very rare, the decision to perform an operation should be made after non-operative treatment has failed to relieve pain and to improve function. The presence of non-progressive neurological deficits has been shown to relate poorly with physical function and therefore is not a reason to operate86. Progression of a neurological deficit or the development of a cauda equina syndrome, although rare in association with lumbar spinal stenosis, are two indications for urgent operative decompression. Non-operative treatment has been shown to be less successful in patients who have more severe pain and functional limitation, but the anatomical stenosis in these patients may not be more severe than that in patients who have less severe symptoms22,35.
A variety of techniques for decompression of the degenerative stenotic lumbar spine have been described. Standard wide decompressive laminectomy involves removal of the lamina and ligamentum flavum from the lateral border of one lateral recess to that of the other at all involved transverse levels of the spine. Affected nerve roots then are directly visualized and decompressed from their origin at the thecal sac and throughout their course as they exit the neural foramen. All lateral recesses in which nerve roots may be entrapped should be decompressed. Alternative techniques of decompression for the treatment of degenerative central and lateral-recess stenosis of the lumbar spine were designed in an attempt to preserve more of the posterior osseous and ligamentous structures, theoretically diminishing the problem of postoperative instability. These techniques include a beveled laminectomy with angular resection of only the anterior portion of the lateral aspect of the lamina42, selective single or multiple unilateral or bilateral laminotomy4,28,63,70,77, multilevel partial laminectomy83,95, and lumbar laminaplasty60,87. The early clinical results of each of these techniques have been satisfactory, but reports of long-term results are lacking. The problem of regrowth of bone with clinically important recurrent stenosis may be more frequent in association with methods of decompression involving limited resection of bone67. Only the technique of multiple laminotomy has been compared with standard wide laminectomy in a prospective, randomized study70. The clinical results of these two methods were shown to be similar after a mean duration of follow-up of 3.7 years. Multiple laminotomy took longer to perform, and it was associated with nerve injury in three (12 per cent) of twenty-six patients. In nine (26 per cent) of the thirty-five patients in whom multiple laminotomy was attempted, adequate decompression was not achieved and intraoperative conversion to a standard wide laminectomy was necessary. Recently, selective laminectomy was advocated for patients with two-level stenosis in whom the one level that is responsible for the symptoms can be identified by means of a neurological examination (both before and after a walking test) and with use of a selective nerve-block79. Of the twenty-eight patients in that study who had anatomical stenosis at two levels, twenty-three (82 per cent) were believed to have symptoms caused by one level and five (18 per cent) were believed to be symptomatic at both levels. Decompression was performed only at the level or levels that were believed to be responsible for the symptoms. The postoperative improvement in the functional scores of the patients who had stenosis at two levels but decompression at only one level was similar to that of other patients who had stenosis at one level. Postoperative spondylolisthesis has been reported as a complication after lumbar decompression performed without simultaneous arthrodesis of the decompressed segments40,55,72. Postoperative spondylolisthesis is twice as common among patients who have an unsatisfactory result after lumbar decompression and is more common after decompression that includes total facetectomy40. Postoperative progression of spondylolisthesis at sites at which there was preoperative degenerative spondylolisthesis is more common than postoperative development of spondylolisthesis at sites at which there was no preexisting spondylolisthesis55,88. Progression can occur even when concomitant arthrodesis without instrumentation has been performed; such progression has been associated with a poor clinical outcome10. Preoperative radiographic and anatomical risk factors associated with the postoperative development or progression of spondylolisthesis at the level of the fourth and fifth lumbar vertebrae include a well maintained disc height, the absence of degenerative osteophytes, and a smaller, sagittally oriented facet joint72. The results of operative decompression for lumbar spinal stenosis generally have been good2,37,52,61,80,89,90. In one study, the outcome was described as good or excellent by 191 (59 per cent) of 324 patients who were interviewed by telephone at a mean of 4.6 years postoperatively89. In another series, a good or excellent outcome was reported for 272 (62 per cent) of 438 patients2. Mauersberger and Nietgen, in a study of sixty-six patients who had been managed with operative decompression, reported that the symptoms improved clinically in fifty-three patients (80 per cent) and resolved completely in twenty-four (36 per cent)61. Silvers et al. reported that ninety-six (75 per cent) of 128 patients remained satisfied at a mean of 4.7 years after operative decompression80. Similarly good results have been reported for patients who are more than sixty-five years of age61,78. A critical analysis of seventy-four studies that were published before 1990 revealed a good-to-excellent outcome, according to standardized criteria, in an average of 64 per cent of patients who had had operative treatment of lumbar spinal stenosis; however, the percentage of patients who had a good result varied widely among studies (range, 26 to 100 per cent)90. A comparison of sixty-seven patients who had had non-operative treatment of lumbar spinal stenosis and eighty-one patients who had had operative treatment (either decompression alone [seventy-eight patients] or decompression with arthrodesis [three patients]) showed that the operatively managed patients reported significantly better outcomes at one year, despite the fact that they tended to have been more symptomatic before treatment (p = 0.05)5. The risk factors associated with a worse outcome after laminectomy for lumbar spinal stenosis include diabetes, osteoarthrosis of the hip, preoperative fracture of a lumbar vertebra, and preoperative degenerative scoliosis2,24. More recent studies of the long-term outcome after decompression for lumbar spinal stenosis have suggested that the initial clinical improvement deteriorates over time11,44,45,69. In one study, eleven (20 per cent) of fifty-four patients who had had a satisfactory short-term clinical outcome had an unsatisfactory outcome at a mean of 8.2 years postoperatively69. In another series, twenty-six (27 per cent) of ninety-six patients who had had a good result initially had a clinical failure by five years postoperatively11. Katz et al. found that, at seven to ten years after decompression with or without arthrodesis, forty-one (75 per cent) of fifty-five patients were satisfied with the result and twenty (23 per cent) of the original cohort of eighty-eight patients had had a reoperation44. The risk factors associated with a poor long-term outcome include greater medical comorbidity (as measured with use of an illness-rating scale) and an initial laminectomy involving a single interspace46. Recurrent symptoms may be caused by recurrence of stenosis at a level that was previously decompressed, progression of stenosis at an adjacent level, or mechanical back pain with instability44. In contrast, one group of authors reported that the clinical result of lumbar decompression was better at an average of thirteen years than it had been at an average of seven years34. Reports have varied with respect to the effect of diabetes on the clinical result of decompression for lumbar spinal stenosis. In one series, diabetes was associated with a greater percentage of poor outcomes and postoperative wound complications82. An excellent or good clinical result was reported for only ten (42 per cent) of twenty-four patients who had diabetes compared with twenty (91 per cent) of twenty-two matched controls who did not. In another series, an excellent or good clinical result was reported for eighteen (72 per cent) of twenty-five patients who had diabetes and twenty (80 per cent) of twenty-five matched controls who did not12. In my experience, the relief of activity-related symptoms after decompression has been as reliable in patients who have diabetes as in those who do not. The relief of constant pain and abnormal sensations in the lower extremity has been less reliable in patients who have diabetes, presumably because of residual symptoms of underlying diabetic neuropathy. It is important for a diabetic patient to have realistic expectations when operative decompression is being considered. A relationship between the outcome of lumbar decompression and the presence of preoperative posture-related pain was seen in one investigation27. A good postoperative result was reported for twenty-two of twenty-three patients in whom the preoperative pain had been related to posture and in only five of ten patients in whom it had not. Patients in whom back pain is the predominant preoperative symptom have been less satisfied with the operative result46. The outcome after operative treatment of lumbar spinal stenosis also has been shown to be adversely affected by a history of previous operations on the lumbar spine36. Early failures have been reported to occur after decompressive laminectomy in patients who did not have evidence of severe stenosis on preoperative imaging studies and in those who did not have preoperative neurogenic claudication17. This finding emphasizes the importance of appropriate selection of patients. An operation is appropriate when the patient has neurological symptoms (not only back pain) that are associated with evidence of spinal stenosis on imaging studies. Fracture of a lumbar facet has been described as a potential source of late pain after lumbar laminectomy74. Rosen et al. reported that, in patients who had such a fracture, the amount of bone that had been removed immediately cephalad to the flare of the inferior articular process at the level of the laminectomy was greater than that in patients who did not have such a fracture74. Although primarily done to relieve pain, decompression for lumbar spinal stenosis also can affect the function of the bladder. Deen et al., in a prospective study, reported that standard decompressive laminectomy resulted in subjective improvement of symptomatic dysfunction of the bladder in twelve (60 per cent) of twenty patients16. The results of cystometrography and electromyography did not change after decompression, but the postvoiding residual volume improved in all nine patients in whom it had been elevated preoperatively. The role of arthrodesis in the operative treatment of lumbar spinal stenosis has been the subject of much discussion in the recent literature47. After routine decompression in the absence of spinal deformity or instability, concomitant spinal arthrodesis generally does not provide additional clinical benefit29,64. Instability caused by the resection of facets at the time of the operation may be an indication for simultaneous arthrodesis to prevent persistent instability and pain. It is generally accepted that spinal stability will be maintained if a total of at least one facet joint is preserved9. However, biomechanical evidence has suggested that instability (as indicated by markedly increased segmental motion) occurs after unilateral total facetectomy, even if the remaining facet has been left intact1. Unilateral or bilateral medial partial facetectomy has been shown to have little effect on segmental motion1. The role of arthrodesis in the operative treatment of lumbar spinal stenosis with associated degenerative spondylolisthesis also has been the subject of much discussion. Laus et al. reported clinical success after decompression alone in a group of patients who had degenerative spondylolisthesis53. Those authors suggested that the segments were naturally stabilized by narrowing of the disc space and by the presence of spondylotic osteophytes. However, the findings of other studies have suggested that concomitant arthrodesis of the spondylolisthetic segment may improve the clinical outcome in such patients10,11,33,50,59,67,70. A recent meta-analysis of the literature concerning the operative treatment of degenerative lumbar spondylolisthesis demonstrated that the outcome was significantly better when concomitant spinal arthrodesis had been done (p < 0.0001)58. According to a weighted pooled proportion analysis (which gives more weight to larger studies), the clinical result was satisfactory for 140 (69 weighted pooled per cent) of 216 patients who had had decompression without concomitant spinal arthrodesis, with progression of the spondylolisthesis in sixty-seven (31 weighted pooled per cent) of the 216 patients. A similar weighted pooled proportion analysis of patients who had had operative decompression with in situ posterolateral arthrodesis showed that the clinical outcome was satisfactory for fifty-nine (90 weighted pooled per cent) of eighty-four patients, with progressive anterolisthesis in only fourteen (17 weighted pooled per cent) of the eighty-four patients and a solid fusion in sixty-two (86 weighted pooled per cent)58. Postacchini and Cinotti observed that the regrowth of bone after decompressive laminectomy at the lumbar level was more common in patients with degenerative spondylolisthesis who had not had an arthrodesis67. The use of instrumentation in arthrodesis for lumbar spinal stenosis with degenerative spondylolisthesis remains controversial, but increasing evidence supports the benefits of immediate fixation. In a series of patients who had arthrodesis with fixation with use of a Luque rectangular rod and sublaminar wires after decompression, nineteen (90 per cent) of twenty-one patients had relief of back pain and twenty (91 per cent) of twenty-two had relief of pain in the lower extremity; the overall rate of fusion was 96 per cent (twenty-three of twenty-four)50. This technique necessitates the inclusion of one level cephalad to the most cephalad decompressed segment; the least extensive arthrodesis is one that spans the third, fourth, and fifth lumbar vertebrae after laminectomy of the fourth lumbar vertebra. Pedicle-screw fixation has been shown to improve the rate of fusion and the clinical outcome after decompression and arthrodesis for degenerative lumbar spondylolisthesis10,59,96, but the technique has not yet been approved for marketing by the Food and Drug Administration. Early postoperative relief of back pain often is attributed to the immediate stabilization provided by the instrumentation. Internal fixation also reduces the prevalence of postoperative progression of the spondylolisthesis, which can occur even after arthrodesis without instrumentation10. Rigid constructs have been associated with a better clinical result than semi-rigid constructs, which allow motion between the fixation screws and the rod or the plate59,96. In patients who have lumbar spinal stenosis and associated degenerative lumbar scoliosis, wide decompression can destabilize the spine; this makes further progression of the deformity likely and concomitant arthrodesis necessary20,81. Vertebral osteophytes that bridge spinal segments and narrow disc spaces may signify that there has been some spontaneous stability provided in the region of the degenerative spinal stenosis and asymmetrical collapse. In some instances, a single symptomatic nerve root can be isolated by means of selective diagnostic injections, allowing for a more limited decompression79. This also may obviate the need for arthrodesis, especially in a patient who has had no back pain or history of marked progression of the deformity. In a review that included forty patients, adjunctive fixation with pedicle screws and rods was shown to result in marked relief of pain in the lower extremity (thirty-four patients [85 per cent] had severe pain preoperatively and thirty-eight [95 per cent] had little or no pain at an average of forty-four months postoperatively) as well as relief of back pain (thirty-five patients [88 per cent] had severe pain preoperatively and thirty-two [80 per cent] had little or no back pain at the time of follow-up)81. After the removal of compressive bone and ligamentous tissue, additional decompression can be accomplished by realignment of the spine with use of segmental distraction along the concavity of the curve. This is done after setting of the convex rod first, with mild compression between screws in order to preserve or improve the amount of lumbar lordosis. Lordosis also can be improved by in situ bending of the rod after the rod has been inserted and connected to the screws. In patients who have preoperative imbalance in the sagittal plane because of loss of lordosis as a result of disc-space collapse, consideration should be given to anterior or posterior lumbar discectomies with structural interbody bone-grafting to restore disc height and lordosis before posterior decompression and stabilization. Preoperative lateral radiographs, made with the patient non-weight-bearing and the spine in maximum flexion and extension, can help in the assessment of the overall flexibility in the sagittal plane.
Despite a heightened awareness of the anatomical abnormalities, clinical symptoms, and natural history of degenerative lumbar spinal stenosis, many treatment methods remain scientifically unvalidated. While the efficacy of physical therapy modalities, medications, and epidural injections of steroids remains unknown, non-operative treatment continues to be effective for the relief of symptoms in most patients. Prospective, randomized studies are still needed to show improvement in what has been recognized as the relatively benign natural history of lumbar spinal stenosis in most patients. Current trends toward more limited operative decompression, with retention of the stabilizing elements and a decrease in short-term morbidity, may lead to a higher rate of long-term failure due to recurrent stenosis or the development of stenosis at an adjacent level. There is convincing evidence that the outcome is improved when spinal arthrodesis is done after operative decompression in patients who have degenerative lumbar spinal stenosis with spondylolisthesis or degenerative scoliosis. Pedicle instrumentation can provide segmental fixation even after laminectomy, improving the rate of fusion and avoiding the need to extend the instrumentation to adjacent levels with intact laminae. The best way to determine which patients will benefit from this advanced technology, in order to minimize operative morbidity and limit the overall costs of treatment, remains the subject of much current clinical investigation.
*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
Definition and Classification
