CHAPTER 285 Lumbar Spine Stenosis
Verbiest1–4 was one of the first to define central lumbar spinal stenosis, noting that laminectomy alone relieved radiculopathic symptoms for patients with narrowed spinal canals. He defined congenital stenosis, characterized by shortened pedicles and a shallow sagittal diameter, in two ways: absolute stenosis (≤10 mm) or relative stenosis (>10 to 12 mm).3–5 In a trefoil-shaped spinal canal, lateral recess or subarticular stenosis also contributes to lateral thecal sac and nerve root entrapment beneath the superior articular facets.6 Acquired stenosis is defined by the development of progressive degenerative changes superimposed on an originally normal or narrowed spinal canal. Single and multilevel stenosis is variously attributed to thickened laminae, arthrotic facets, exaggerated lordotic curvatures, laminar shingling, infolding of the hypertrophied yellow ligament, and ossification of the posterior longitudinal ligament, all of which contribute to central, lateral, or foraminal stenosis.7 Stenosis typically occurs at the L4-5 level, followed in descending order of frequency at L3-4, L2-3, L5-S1, and L1-2.8 Other factors more rarely contribute to spinal canal stenosis. Endocrinopathies, including Paget’s disease, acromegaly, and fluorosis, contribute to increased narrowing of the spinal canal, as does achondroplasia (characterized by trapezoidal vertebrae; short, thick pedicles and vertebral end plates; hypertrophied lamina; and increased periosteal bone formation).9
Anatomy
Hypertrophy or ossification of the ligamentum flavum contributes to dorsal compression of the thecal sac, whereas ventrally and centrally situated disks, spurs, or osteophytes, and, rarely, hypertrophy of the posterior longitudinal ligament may compromise the available canal space. Although congenital benign Tarlov cysts (small or large) may occasionally balloon into neural foramina, they are not considered pathologic and rarely require surgery.10,11
Symptoms of stenosis or neurogenic claudication are attributed to direct mechanical compression or indirect vascular insufficiency involving the lumbar nerve roots or cauda equina.12 Standing and walking increase lordosis and transiently exaggerate infolding of the ligamentum flavum, while sitting and recumbence reverse the lordosis, open the canal, improve blood flow, and often temporarily relieve complaints.
Symptoms and Signs
Patients with congenital lumbar stenosis may become symptomatic early, in the third to fifth decades of life, whereas patients with acquired stenosis typically develop symptoms later, in the sixth to seventh decades. Unilateral or bilateral radiculopathy may be associated with neurogenic claudication, characterized by leg pain, numbness, tingling, and weakness, that is exacerbated by walking specific distances (measured in blocks). When 100 consecutive patients with disk disease, lateral recess stenosis, and central stenosis were evaluated, all three groups of patients demonstrated comparable complaints of pain at rest, pain at night, and pain on coughing.13 Many patients without significant neurological deficits can be managed conservatively with rehabilitation or multidisciplinary pain centers.14
Neurological findings include mechanical, motor, reflex, and sensory signs that reflect the level or levels of involvement. In descending order, these are L5 root syndromes (L4-5 disease), L4 root syndromes (L3-4), L3 root syndromes (L2-3), and S1 root pathology (L5-S1).15 Bladder dysfunction is rare in young patients but is frequently encountered in the geriatric population with lumbar spinal stenosis. Patients averaging 71 years of age and undergoing two- to four-level laminectomies for severe lumbar stenosis exhibited a significant degree of both preoperative and postoperative bladder compromise.16 Six months following laminectomies, although only 45% of patients showed cystoscopic or urodynamic evidence of improved urinary function (i.e., decreased postvoiding residual volumes), 60% of patients reported a subjective recovery of urinary function.
Diagnostic Studies
Radiography
Plain anteroposterior radiographs demonstrate the number of lumbar vertebrae and help determine whether there is a lumbosacral bony anomaly (frequency, 7%).17 Lateral radiographs reveal the curvature of the lumbar spine and the presence or absence of static or dynamic olisthesis or instability, defined by greater than 4 mm of translation and greater than 10 to 12 degrees of angulation at the level of olisthy.18
Magnetic Resonance Imaging
MRI offers the best noninvasive assessment of soft tissue structures but not bone-related pathology (
Figs. 285-E1 to 285-E4). MRI studies best demonstrate the soft tissue compressive changes of the thecal sac or nerve roots attributed to disk disease, ligamentous compression (hypertrophy of the yellow ligament or posterior longitudinal ligament), olisthesis, and other intrinsic neural pathology.19 MRI also offer simultaneous longitudinal, sagittal, and transaxial views of soft tissue pathology that may be located centrally, laterally, foraminally, or far laterally. It does not, however, directly demonstrate calcific or ossific changes often associated with disk or limbus fractures, spondylosis, and ossification of the posterior longitudinal or yellow ligament. Many otherwise asymptomatic lesions are also identified on MRI in the older population. In one series of patients older than 60 years, herniated disks (36%) and spinal stenosis (21%) were symptomatic in only one third of patients.19
Lumbar instability associated with lumbar stenosis can also be documented on MRI by the visualization of increased fluid within the facet joints.20 The volume of facet fluid can be calculated on routine axial MRI by adding the width of the fluid in both facet joints and dividing that by the sum of the width of both facets. Increased lumbar facet fluid (indicating instability) positively correlated with sagittal instability documented on dynamic x-ray studies obtained at the L4-5 level (focus of degenerative changes). Of 50 patients entered in the study, 28 (55%) had facet fluid; 23 (82%) were unstable, and 5 (18%) were not. The investigators concluded that a close linear relationship exists between the facet fluid and instability as seen on MRI and dynamic radiographs, respectively. Enhanced MRI offers “myelographic” views that correlate with operative findings 82.6% of the time.21
Contrast-enhanced MRI also accurately differentiates scar from disk (96%), visualizes the cervicothoracic or thoracolulmbar junction, and helps differentiate among tumor, demyelinating syndromes, adhesive arachnoiditis, and infection.22 Magnetic resonance myelography offers more specific data in the preoperative assessment of foraminal stenosis. In one study it documented foraminal stenosis that was surgically confirmed in 35 of 990 patients.23
Computed Tomography
Routine and two- and three-dimensional reconstructed CT images in multiple planes (axial, coronal, sagittal) confirm the diagnosis of lumbar stenosis (
Figs. 285-E5 to 285-E14). They also contribute to the recognition of accompanying disk disease, limbus fractures, olisthesis or instability, ossification of the posterior longitudinal ligament (OPLL), ossification of the yellow ligament (OYL), and progression of fusion in a posterolateral fusion mass. In a series of 48 patients who were symptomatic following lumbar fusion, CT identified 157 abnormalities; 12 of 27 major lesions included fusion mass fractures, hairline pseudarthroses, and residual spinal stenosis (all confirmed during second operations).24 Myelographic CT studies are now rarely performed because the majority of pathologic findings can be confirmed noninvasively with varying combinations of MRI and CT evaluations.
Pathology Mimicking Lumbar Stenosis
Some patients with lumbar stenosis may have more cephalad tumors contributing to seemingly “lumbar” complaints. These typically include ependymomas, neurofibromas, meningiomas, and metastatic lesions (
Fig. 285-E15). In one study, a patient with an unresolved right footdrop following lumbar surgery was found to have a left-sided parasagittal convexity meninigoma; resection resulted in complete resolution of the deficit.25 Other degenerative, metabolic, endocrine, and vascular disorders may mimic the signs and symptoms of lumbar stenosis. These include thoracic disk herniation, Scheuermann’s disease, Paget’s disease, and arthritis of the hips. In one series, amyloidosis and its characteristic crystals contributed to hypertrophy of the yellow ligament in 12 of 97 patients undergoing lumbar surgery for stenosis.26 Diabetes, contributing to diabetic peripheral neuropathy, femoral amyotrophy, or angiopathy, can also be misdiagnosed as lumbar stenosis. Of note, among diabetics with lumbar stenosis, only 39% exhibited good or excellent outcomes, compared with 95% good or excellent results for those without diabetes.27 Peripheral vascular diseases resulting in vascular rather than neurogenic claudication, characterized by pain associated with ambulation but relieved with rest alone, may be present alone or exist simultaneously with lumbar stenosis.
Associated Conditions
Tandem Stenosis
For patients with tandem lesions, decompression or excision of the cephalad pathology may result in partial and occasionally marked improvement in lumbar complaints.25,28–32 In one series, following an L4-5 instrumented fusion, one patient was found to have additional C4-7 and T11-12 stenosis; he underwent laminectomies to decompress all levels and finally improved.25 In some series of tandem cervical and lumbar stenosis, the order of surgery is dictated by the severity of the stenosis.31 For patients with absolute stenosis (canal ≤ 10 mm) and myelopathy, often with superimposed pathology such as OPLL, cervical surgery should typically precede lumbar intervention (
Fig. 285-E16). For patients with relative stenosis and a canal depth of 11 to 13 mm correlated with radiculopathy rather than myelopathic symptoms, lumbar surgery often takes precedence over cervical intervention. For the former patients, preliminary cervical decompression often results in improvement of seemingly “lumbar” complaints in more than one third. In another series, 8 of 230 patients with lumbar stenosis were also found to have cervical stenosis.32 Three patients underwent cervical surgery first, and five had initial lumbar procedures; all patients exhibited excellent or good results, and none deteriorated. Although tandem cervical-lumbar stenosis occurs in only 10% of patients, it should be anticipated in patients older than 65 years, and these individuals should undergo more stringent screening for tandem cervical disease.
Ossification of the Yellow Ligament
OYL may significantly contribute to lumbar stenosis. OYL presents as an initial ingrowth of fibrocartilage attributed to the proliferation of type II collagen (
Figs. 285-E17 to 285-E20). Hypertrophy or OYL begins laterally at the enthesis and extends medially.33 In 110 predominantly geriatric individuals undergoing multilevel laminectomies (average, five levels) with noninstrumented fusion, the 10 who developed intraoperative dural tears exhibited severe OYL; this extended to or through the dura in 3.34 For the remaining 100 patients, 57 showed moderate OYL, and 22 showed marked OYL.
Ossification of the Posterior Longitudinal Ligament
The frequency of OPLL in the proximal lumbar spinal canal is 10%, with another 10% being found in the proximal thoracic spine (T1-4) spine; the majority (80%) of OPLL is found in the cervical spinal canal.35 OPLL and OYL may both contribute to thoracic or lumbar stenosis.29,30,35,36 Of 1100 patients having surgery for spinal stenosis from 1986 to 1997, 26 (2.3%) had OYL or OPLL (11 OPLL, 12 OYL, 3 OYL and OPLL).35 Analysis of a fluid collection in the lumbar spine using β2-transferrin may help document the presence of a fistula.37
Ankylosing Hyperostosis
Ankylosing hyperostosis may also contribute to lumbar stenosis (
Fig. 285-E21). The characteristic CT findings include anterior or posterolateral marginal, somatic osseous proliferation and proliferative changes involving the posterior facet joints, articular capsules, yellow ligament, or supraspinal ligaments.38
Disk Herniation with and without Olisthesis
Disk herniations occur in up to 45% of patients undergoing surgery for lumbar stenosis with or without olisthesis (in a series of 857 individuals).39,40 Disk herniations in conjunction with stenosis alone were reported in 15% to 45% of cases.39,41–43 In 60 patients undergoing multilevel laminectomies (average, 5.4 levels) for lumbar stenosis using lamina autograft and β-tricalcium phosphate (β-TCP) for one- to two-level noninstrumented posterolateral lumbar fusion, disk herniations were observed in 20 patients (33.3%).44 Of 95 patients undergoing one-level instrumented posterolateral lumbar fusion using lamina autograft supplemented with demineralized bone matrix, 13 patients (13.6%) without preoperative olisthesis underwent unilateral or bilateral facetectomy for far lateral disks or stenosis.45 Among 100 patients undergoing multilevel laminectomy (3.6 levels) with one-level (78 patients) and two-level (22 patients) instrumented fusion using lamina autograft and β-TCP, 57 herniated disks were found in 50 patients: 21 were routine disks, 7 were foraminal disks, 24 were far lateral disks, and 5 were recurrent disk herniations.
Lower frequencies of disk herniation (4.3% to 20%) have been reported for patients undergoing lumbar decompression in the presence of olisthesis.39,42 In one study evaluating 290 patients with olisthesis, 20% had disk herniations; 47 were routine, and 12 were foraminal or far lateral.42
Limbus Vertebral Fracture
Patients with lumbar stenosis may also exhibit one of four types of limbus vertebral fractures, which are better visualized on CT than MRI.46–48 Type I consists of a shelf of cortical bone traversing the canal, type II involves predominantly a large central fragment, type III is characterized by a lateral or far lateral calcified fragment, and type IV is defined by a massive fragment extending across the entire width of the spinal canal from one interspace to the next. These fragments are typically extremely large, include both cortical and cancellous elements, and warrant more extensive resection to afford access for adequate decompression. For foraminal and far lateral lesions, unilateral facetectomy is typically warranted. Furthermore, resection requires piecemeal removal; this is most safely carried out by first creating a defect or depression at the level of the disk space and then morcellating the limbus fracture into fragments using a down-biting curet, tamp, and mallet technique, which allows delivery into the defect and safe removal. In some cases, intraoperative monitoring (somatosensory evoked potentials [SEPs] or electromyography [EMG]) may be useful to minimize undue retraction or manipulation and subsequent neurological injury.
Far Lateral Disk Pathology
In the lumbar spinal canal, lumbar nerve roots may become trapped by disk herniation or stenosis extending into the far lateral compartment, which is bordered superiorly by the pedicle, anteriorly by the disk, medially by the vertebral body and superior articular facet, and laterally by fat (
Figs. 285-E22 to 285-E28).49,50 Far lateral disks, which typically originate at the inferior interspace and migrate superolaterally, constitute 7% to 12% of all disk herniations.51 Other factors contributing to far lateral pathology include spondylostenosis, arthrosis, limbus vertebral fractures, olisthy with or without lysis, and scoliotic deformity.49,52
Three surgical techniques are used to approach far lateral disk herniations that accompany lumbar stenosis.49,51 Only rarely can a far lateral disk be removed through a medial facetectomy at the L5-S1 level; most cases require either the intertransverse approach or a full facetectomy.49,51 Notably, the L5-S1 level is the widest and least stenotic, and very laterally located facet joints may allow adequate access to the foraminal and proximal far lateral portion of a far lateral disk. The intertransverse procedure combines a medial facetectomy-foraminotomy and far lateral exposure (Wiltse approach), thus preserving the pars interarticularis; risks, however, include delayed fracture of the pars interarticularis, retention of disk material, or damage to the nerve root secondary to incomplete exposure. Finally, the full facetectomy, which is the safest approach, fully visualizes the root along its entire course, along with OYL, synovial cysts, spondylostenosis, and olisthesis; however, it does increase the risk of instability.51 Although in the past many patients undergoing full facetectomies for far lateral disks were not fused, these patients typically undergo simultaneous noninstrumented or instrumented fusion today.
Over a 10-year period (1984 to 1994), 170 patients underwent surgery for far lateral lumbar disks accompanied by lateral recess stenosis (134 patients) or central stenosis (36 patients), far lateral stenosis (30 patients), or degenerative spondylolisthesis (23 patients).51 Most far lateral disks were found at L4-5 (68 patients), L3-4 (63 patients), and L5-S1 (33 patients).49,51 Of note, only 4 of 170 patients (2.4%) underwent secondary pedicle-screw arthrodesis.51 All 4 patients had become unstable following L4-5 laminectomy–unilateral facetectomy for decompression of stenosis accompanied by grade I olisthy. Of interest, outcomes of far lateral disk surgery (Odom’s criteria) were comparable for the three approaches to facet excision: good to excellent in 68% with medial facetectomy, 79% with the intertransverse approach, and 70% with full facetectomy.
Of the original 170 patients, 76 were evaluated by both the surgeon and the Short Form 36 (SF-36) questionnaire (preoperatively and 3, 6, and 12 months postoperatively) for an average of 2.8 years.52 For the 56 patients who were examined within 4.5 years of surgery, significant positive correlations were found between the surgeon’s assessment and six of the SF-36 Health Scales; General Health and Social Function were excluded.
Synovial Cysts
Synovial cysts may contribute to the pathology of lumbar spinal stenosis (
Figs. 285-E29 to 285-E36).53,54 In one series, outcomes were assessed in 45 patients with stenosis and synovial cysts versus 35 with stenosis and degenerative spondylolisthesis using the SF-36 questionnaire.53 The procedures in these patients required laminectomy at an average of 3.8 and 3.5 levels, respectively. Five of the 45 with synovial cysts developed instability postoperatively, and 11 of 35 with preoperative olisthesis developed further progression. After a minimum of 2 postoperative years, good or excellent results were observed in only 58% and 63% of patients, respectively (Physical Function Scale +44 and +38). Because synovial cysts indicate intrinsic disruption of the facet joint and, therefore, instability, it is likely that more primary fusions should be considered in these patients to improve outcomes.53,54
FIGURE 285-E33 On this parasagittal T2-weighted image, a right-sided synovial cyst arising from the L4-5 facet joint is visualized.
FIGURE 285-E35 On this parasagittal two-dimensional computed tomography scan (soft tissue window), the right-sided synovial cyst seen on the axial study (see Fig. 285-E34), arising from the L4-5 facet joint and extending into the foramen, is observed.
In a subsequent series of 110 mostly geriatric patients undergoing, on average, five-level laminectomy with primary noninstrumented fusion, a high frequency of synovial cysts was encountered in those who developed dural tears.34 Of note, 5 of 10 patients with dural tears had synovial cysts (with severe OYL in all 10), whereas only eight synovial cysts were encountered among the remaining 100 without dural tears.
Degenerative Spondylolisthesis
In the lumbar spine, degenerative olisthesis or spondylolisthesis (with an intact neural arch) occurs when facet joints are congenitally oriented in a sagittal rather than a coronal position (
Figs. 285-E37 to 285-E43).39,43 Progressive arthrotic changes of the facet joints contribute to a grade I or 25% anterolisthesis or olisthesis, frequently resulting in progressive cauda equina and nerve root compression. Resultant hypertrophied facet joints contribute to dorsolateral intrusion on the thecal sac and superiorly exiting nerve roots as they exit the spinal canal foraminally and far laterally. Simultaneously, the inferiorly exiting nerve root is compressed in the lateral recess by hypertrophied yellow ligament, disk “bulges,” or arthrotic spurs.
FIGURE 285-E37 Lateral three-dimensional computed tomography image of grade I degenerative spondylolisthesis at the L4-5 level.
Patients with degenerative spondylolisthesis are typically women (2 : 1 female-to-male ratio) 50 to 60 years of age whose symptoms have evolved over decades but have been exacerbated over months to years. Neurological deficits appear late in the clinical course and may be correlated with the onset of neurogenic claudication or radiculopathy associated with proximal weakness or a partial footdrop. The L4-5 level is most commonly involved, followed in descending order by L3-4, L2-3, and L5-S1.43 Olisthesis is rare at L5-S1 because this level is typically located below the intercrestal line and is therefore supported by longer transverse processes and the iliotransverse ligaments. Of 290 patients with degenerative spondylolisthesis, 86% had olisthy at one level, and the remaining 14% had two-level olisthesis.42
Nonsurgical Versus Surgical Management
A subset of patients with spinal stenosis of differing severities might be managed conservatively without surgery. For example, both young and old patients with severe comorbidities that may preclude risky surgical intervention and with severe MRI- and CT-documented stenosis or olisthesis may benefit from treatment at multimodality pain centers.14 In one series, the Oswestry Disability Index (ODI) and surgeons’ clinical assessments were used to compare outcomes for 54 matched pairs with lumbar stenosis treated conservatively (nonoperatively) versus with laminectomy.55 No statistically significant differences in outcomes between the two groups were revealed.
Other studies have documented the superiority of surgery for lumbar stenosis, including decompression or decompression with fusion, compared with nonsurgical alternatives. In a randomized controlled trial involving 94 patients with stenosis, the results of no surgery (44 patients) versus surgical decompression (50 patients) were compared; the latter included undercutting laminectomies for stenosis, with 10 undergoing additional fusions.56 Outcomes, based on the ODI and self-reported measures, revealed that patients in both groups improved over the 2-year postoperative interval, but surgical patients achieved greater relief of leg and back pain and demonstrated less overall disability. In a nonrandomized cohort study of patients with lumbar stenosis, better outcomes were observed following decompression (54 patients) and decompression with fusion for degenerative spondylolisthesis (42 patients) compared with nonoperative intervention (29 patients).57 Based on Roland Morris questionnaires to assess outcomes, patients undergoing either of the surgical procedures exhibited higher scores than those treated conservatively (6.9 and 6.1 versus 1.2). Similarly, better outcomes were observed for the decompression and decompression and fusion groups compared with those who chose nonsurgical alternatives (63.3% and 61.5% versus 25%).
In another study, patients with stenosis and degenerative spondylolisthesis from 13 centers, randomly enrolled in two treatment groups, demonstrated substantial gains in pain relief and function following decompressive laminectomy with or without fusion (304 patients) compared with no surgery (303 patients).58 Patents were evaluated with the SF-36 and a modified ODI 1.5, 3, 6, 12, and 24 months postoperatively. Surgical patients demonstrated significant advantages at 3 months, which increased at 1 year and were only minimally diminished at 2 years postoperatively. On the SF-36, Bodily Pain and Physical Function scores showed a net gain of 18.1 and 18.3 points, respectively, whereas the net ODI was –16.7.
Pre- and Postoperative Considerations
Comorbid Factors
In an observational study of 3482 patients undergoing surgery for lumbar stenosis, medical (headache, depression, central nervous system disorders) and psychosocial comorbid factors (active compensation cases, self-reported poor health, smoking) negatively impacted 3-month and 1-year SF-36 and ODI postoperative outcomes.59 In another study involving 122 lumbar decompressions performed for stenosis in geriatric patients averaging 78.8 years of age, a grade I or II rating based on the American Society of Anesthesiologists’ scale indicated limited comorbid factors and a greater likelihood that surgery would prove safe and effective.60 The advisability for such lumbar procedures in patients deemed grade III proved less certain.
Preoperative Psychiatric Clearance
Depression significantly affects the outcome of patients undergoing lumbar spine surgery. In one study of 99 patients undergoing surgery for lumbar stenosis, questionnaires were completed preoperatively and 3 months postoperatively.61 The Beck Depression Inventory, ODI, Stucki Questionnaire, and Visual Analogue Scale were used. Before surgery, 20% were considered depressed. This factor positively correlated with increased postoperative disability based on the multiple questionnaires; those with continuous depression showed less improvement in symptom severity, pain intensity, walking capacity, and overall disability score. When preoperative depression improved or resolved, the postoperative Beck Depression Inventory outcomes were comparable to those in patients without a history of depression. In another series, 66% of 95 patients with lumbar stenosis were satisfied with their postoperative results; they were typically younger and exhibited less severe preoperative symptoms and disabilities. Depression in particular had a uniquely negative impact on their outcomes.62 In another series, the perceived quality of life (expectations, level of optimism) was evaluated in 57 patients before and 3 months after lumbar spine surgery; higher preoperative expectations and optimism correlated with a better quality of life postoperatively.63
Prophylactic Antibiotics
Although the Centers for Disease Control and Prevention recommends prophylactic antibiotics for lumbar surgery to limit spine infections, protocols vary from single-dose to multiple-dose regimens.64–68 In one study involving patients undergoing comparable but varied lumbar procedures, including stenosis with and without olisthesis and with and without instrumentation, the efficacy of multiple- (5 to 7 days) versus single-dose prophylactic antibiotic regimens employing a first-generation cephalosporin were compared.64
