Chapter 41 - Other Disordes of Spine

Douglas A. Linville

Spinal Stenosis

Before the development of roentgenograms, there were scattered reports of paralysis caused by narrowing of the spinal canal. The first verifiable report of lumbar spinal stenosis relieved by two-level laminectomy was that of Sachs and Fraenkel in 1900. Bailey and Casamajor in 1911 and Elsberg in 1913 wrote similar descriptions of the symptoms, pathological findings, and relief after surgery. The syndrome was not widely diagnosed until Verbiest in 1954 described the classic findings of middle-aged and older adults with back and lower extremity pain precipitated by standing and walking and aggravated by hyperextension. He delineated congenital narrowing of the spinal canal as a contributing factor in many and the secondary development of degenerative changes that further narrowed the lumbar canal and precipitated symptoms. Myelographic block in the midlumbar region with the characteristic degenerative hypertrophic changes about the discs, facets, and ligamentous structures was described in detail. A subsequent article by Verbiest on lumbar spondylosis documented detailed measurements of the spinal canal obtained during surgery. Since that time the syndrome has been well recognized, and numerous well-documented series have been reported. An excellent four-part review of the entire topic was published by Ehni et al. in 1969.

Dunlop, Adams, and Hutton in a cadaver study found a significant increase in pressure on the facet joints with disc space narrowing and increasing angles of extension. Degenerative spinal stenosis has been attributed to simple hypertrophic overgrowth of the superior articular facets that is the result of a progressive degeneration of the disc with resultant instability or hypermobility in the facet joints. As joint destruction progresses, the hypertrophic process finally results in local ankylosis. Calcification and hypertrophy of the ligamentum flavum and venous hypertension resulting in generalized bone overgrowth also may be contributing factors. Mild trauma and occupational activity do not appear to significantly affect the development of this disease, but they may exacerbate a preexisting condition.

ANATOMY

Spinal stenosis can be categorized according to the anatomical area of the spine affected, the region of each vertebral segment affected, and the specific pathological entity involved ( Table 41-1 ). Stenosis can be generalized or localized to specific anatomical areas of the cervical, thoracic, or lumbar spine. It is most common in the lumbar region, but cervical stenosis also occurs frequently. It has been reported rarely in the thoracic spine. Spinal stenosis can be localized or diffuse, affecting multiple levels as in congenital stenosis. Degeneration of the disc occurs with disc narrowing and subsequent ligamentous redundancy, which compromises the spinal canal area. Instability may ensue, aggravating symptoms by the formation of facet overgrowth and ligamentous hypertrophy. The ligamentum flavum may be markedly thickened into the lateral recess where it attaches to the facet capsule, causing nerve root compression. These phenomena occur alone or in combination to create the symptomatic complex characteristic of spinal stenosis.

A description of spinal stenosis requires an understanding of the anatomy affected and the use of consistent terminology ( Fig. 41-1 ). Central spinal stenosis denotes involvement of the area between the facet joints, which is occupied by the dura and its contents. Stenosis in this region usually is caused by protrusion of a disc, annulus, or osteophytes or by buckled or thickened ligamentum flavum. Symptomatic central spinal stenosis results in neurogenic claudication. Lateral to the dura is the lateral canal, which contains the nerve roots; compression in this region results in radiculopathy. The lateral recess, also known as Lee's entrance zone, begins at the lateral border of the dura and extends to the medial border of the pedicle. This is where the nerve root exits the dura and courses distally and laterally under the superior articular facet ( Fig. 41-2, A ). The borders of the lateral recess are the pedicle laterally, the superior articular facet dorsally, the disc and posterior ligamentous complex (PLC) ventrally, and the central canal medially. Ciric et al. reported that facet arthritis most frequently causes stenosis in this zone, along with vertebral body spurring and disc or annulus pathology. Lee's mid-zone describes the foraminal region, which lies ventral to the pars. Its borders are the lateral recess medially, the posterior vertebral body and disc ventrally, the pars distally, and the lateral border of the pedicle laterally. The dorsal root ganglion and ventral motor root occupy up to 30% of this space. This also is the point where the dura becomes confluent with the nerve root as epineurium. Causes of stenosis in this area are pars fracture with proliferative fibrocartilage or a lateral disc herniation ( Fig. 41-2, B ). Thickening of the ligamentum flavum sometimes persists into the foramen and can be associated with a spur from the undersurface of the pars, especially if foraminal height is less than 15 mm and posterior vertebral body height is less than 4 mm. The exit zone is identified as the area lateral to the facet

TABLE 41-1 -- Classification of Spinal Stenosis

Anatomical AreaAnatomical Region (Local Segment)

ANATOMICAL

CervicalCentral

Foraminal

ThoracicCentral

LumbarCentral

Lateral recess

Foraminal

Extraforaminal (far-out)

PATHOLOGICAL

Congenital

Achondroplastic (dwarfism)

Congenital forms of spondylolisthesis

Scoliosis

Kyphosis

Idiopathic

Degenerative and Inflammatory

Osteoarthritis

Inflammatory arthritis

Diffuse idiopathic skeletal hyperostosis (DISH)

Scoliosis

Kyphosis

Degenerative forms of spondylolisthesis

Metabolic

Paget disease

Fluorosis

joint. The nerve root is present in this location and can be compressed by a far-lateral disc, spondylolisthesis and associated subluxation, or facet arthritis.

The most common type of spinal stenosis is caused by degenerative arthritis of the spine, including Forestier syndrome, and is characterized by hyperostosis and spinal rigidity in elderly patients. Congenital forms caused by disorders such as achondroplasia and dysplastic spondylolisthesis are much less frequent. Finally, other processes such as Paget disease, fluorosis, kyphosis, scoliosis, and fracture with canal narrowing

have been reported to result in spinal stenosis. Hypertrophy and ossification of the posterior longitudinal ligament in the cervical spine (diffuse idiopathic skeletal hyperostosis [DISH] syndrome) also may result in an acquired form of spinal stenosis. This disease usually is confined to the cervical spine.

Differentiation between the pathological processes causing spinal stenosis is relatively simple. Congenital spinal stenosis usually is central and evident on imaging studies. In achondroplasia the canal is narrowed in both the anteroposterior plane due to shortened pedicles and in lateral diameter because of diminished interpedicular distance. These findings occur in addition to the other characteristic features of achondroplasia. Idiopathic congenital narrowing usually involves one dimension of canal measurement, and the patient otherwise is normal.

Acquired forms of spinal stenosis usually are degenerative ( Box 41-1 ) and are localized to the facet joints and ligamentum flavum, with additional arthritic changes in the joints visible on roentgenographic studies. Frequently these abnormalities are symmetrical. The L4-5 level is the most commonly involved, followed by L5-S1 and L3-4. Disc herniation and spondylolisthesis may further exacerbate the narrowing. Spondylolisthesis and spondylolysis rarely cause spinal stenosis in young patients. The combination of degenerative change, aging, and spondylolisthesis or spondylolysis in patients 50 years or older frequently results in lateral recess or foraminal stenosis. Paget disease and fluorosis have been reported to result in central or lateral spinal stenosis. Paget disease is one form of spinal stenosis that responds well to medical treatment with calcitonin.

NATURAL HISTORY

The natural course of all forms of spinal stenosis is the insidious development of symptoms occasionally exacerbated by trauma or heavy activity. Many patients have significant roentgenographic findings with minimal complaints or physical findings. Johnsson, Rosn, and Udn reported that 19 (70%) of 27 patients with moderate, untreated spinal stenosis (11 mm or more anteroposterior canal diameter) remained unchanged after 4 years of observation; 4 (15%) improved, and 4 deteriorated without serious sequelae. In a separate study, Johnsson et al. found that 11 of 19 (58%) untreated patients were unchanged at 31-month follow-up, 6 were improved, and only 2 were worse.

Box 41-1. Classification of Spinal StenosisCONGENITAL

Idiopathic

Achondroplastic

ACQUIRED

Degenerative

Central canal

Lateral recess, foramen

Degenerative spondylolisthesis

Degenerative scoliosis

Combination of congenital and degenerative stenosis

Iatrogenic

Postlaminectomy

Postfusion

Postchemonucleolysis

Spondylolytic

Posttraumatic

Miscellaneous

Paget disease

Fluorosis

Diffuse idiopathic skeletal hyperostosis syndrome

Hyperostotic lumbar spinal stenosis

Oxalosis

Pseudogout

In a prospective, randomized study by Amundsen et al. of 100 patients with symptomatic spinal stenosis, 19 patients with severe symptoms were treated operatively, 50 patients with moderate symptoms were treated conservatively, and 31 patients were randomized to receive operative (13) or conservative (18) treatment. Pain relief was noted after 3 months in most patients regardless of treatment but took as long as 12 months in a few patients. However, results in conservatively treated patients deteriorated over time, and at 4 years were excellent or fair in 50% of patients treated nonoperatively; 80% of those treated operatively still had good results. Results were not worse if surgery was done 3 years after failed conservative treatment, and significant deterioration did not occur during the 6 years of follow-up in any of the three groups of patients. Predictors of poor outcomes could not be identified. These authors concluded that conservative treatment is appropriate for patients with moderate pain, 50% of whom will have pain relief in less than 3 months, but operative treatment probably is indicated for patients with severe pain and those in whom conservative treatment fails.

Reported studies suggest that for most patients with spinal stenosis a stable course can predicted, with 15% to 50% showing some improvement with nonoperative treatment. Worsening of symptoms despite adequate conservative treatment is an indication for operative treatment.

CLINICAL EVALUATION

In a group of 100 patients, Amundsen et al. found back pain and sciatica present in 95% and claudication present in 91%. Sensory disturbance in the legs was present in 70%, with motor weakness in 33%, and voiding disturbance present in only 12% of patients. Despite the coexistent symptoms, back pain had been present for a median duration of 14 years and sciatica for a median duration of 2 years before presentation. Bilateral leg complaints were present in 42%, and unilateral leg symptoms were present in the other 58%. Distribution of symptoms was L5 in 91%, S1 in 63%, L1-4 in 28%, and S2-5 in 5%. Forty-seven patients (47%) had symptoms specific for two nerve roots, and 35% had monoradiculopathy. Three- and four-level radicular complaints were recorded in 17% and 1%, respectively. In patients with central spinal stenosis, symptoms usually are bilateral and involve the buttocks and posterior thighs in a nondermatomal distribution. With lateral recess stenosis, symptoms usually are dermatomal because they are related to a specific nerve being compressed. Patients with lateral recess stenosis may have more pain during rest and at night but more walking tolerance than patients with central stenosis.

Differentiation of symptoms of vascular claudication from those of neurogenic claudication is important ( Table 41-2 ). Vascular symptoms typically are felt in the upper calf, are relieved after a short rest (5 minutes) while still standing, do not require sitting or bending, and worsen despite walking uphill or riding a stationary bicycle. Neurogenic claudication improves with trunk flexion, stooping, or lying but may require up to 20 minutes to improve. Patients often report better endurance walking uphill or up steps and tolerate riding a bicycle better than walking on a treadmill because of the flexed posture that occurs. Pushing a grocery cart also allows spinal flexion, which enhances endurance in most patients with neurogenic claudication.

Generally, physical findings with all forms of spinal stenosis are inconsistent. Distal pulses should be felt and confirmed to be strong, and internal and external rotation of the hips in extension should be full, symmetrical, and painless. Straight leg raising and sciatic tension tests usually are normal. The neurological examination usually is normal, but some abnormality may be detected if the patient is allowed to walk to the limit of pain and is then reexamined. The gait and posture after walking may reveal a positive stoop test. This test is done by asking the patient to walk briskly. As the pain intensifies, the patient may complain of sensory symptoms followed by motor symptoms. If the patient is asked to continue to walk, he may assume a stooped posture, and the symptoms may be eased; or, if he sits in a chair bent forward, the same resolution of symptoms will occur.

TABLE 41-2 -- Differentiation of Symptoms of Vascular Claudication from Those of Neurogenic Claudication

EvaluationVascularNeurogenic

Walking distanceFixedVariable

Palliative factorsStandingSitting/bending

Provocative factorsWalkingWalking/standing

Walking uphillPainfulPainless

Bicycle testPositive (painful)Negative

PulsesAbsentPresent

SkinLoss of hair; shinyNormal

WeaknessRarelyOccasionally

Back painOccasionallyCommonly

Back motionNormalLimited

Pain characterCrampingdistal to proximalNumbness, achingproximal to distal

AtrophyUncommonOccasional

DIAGNOSTIC IMAGING

Roentgenography

Although plain roentgenography cannot confirm spinal stenosis, findings such as short pedicles on the lateral view, narrowing between the pedicles on the anteroposterior view, ligament ossification, narrowing of the foramen, and hypertrophy of the posterior articular facets can be helpful hints. Leroux et al. outlined hypertrophic roentgenographic changes associated with hyperostosis on plain and computed tomography (CT) ( Box 41-2 ).

The roentgenographic identification and confirmation of lumbar spinal stenosis have improved with the development of new imaging techniques. Initially only central spinal stenosis was recognized, with canal narrowing to 10 mm considered absolute stenosis. This could be measured using roentgenograms or preferably myelography. In 1985 Schnstrm, Bolender, and Spengler compared the identification of central spinal stenosis with anteroposterior canal measurement by CT to identification by measurement of the dural sac with myelography in patients undergoing surgery for spinal stenosis. They found no correlation between the transverse area of the bony canal in normal patients and patients with spinal stenosis. A dural sac transverse area of 100 mm2 or less did correlate with symptomatic spinal stenosis. This method allows the inclusion of soft tissue in the determination of spinal stenosis. The analysis of this area can be calculated relatively easily using standard CT scanning software.

Currently, axial imaging has supplanted standard roentgenograms in the diagnosis of spinal stenosis, although roentgenograms are important in the initial evaluation of patients with persistent pain of more than 6 weeks' duration or of those with red flags of other disease, including recent trauma, history of cancer, immunosuppression, age more than 50 or less than 20 years, neurological deficit, or previous surgery. Flexion and extension views are useful to identify preexisting instability before laminectomy and may be useful in determining the need for subsequent fusion. Translation of more than 4 to 5 mm or rotation of more than 10 degrees to 15 degrees is indicative of instability. A reversal of the normal trapezoidal disc geometry with widening posteriorly and narrowing anteriorly also may indicate instability.

Box 41-2. Hypertrophic Roentgenographic Changes Associated with Hyperostosis * PLAIN ROENTGENOGRAPHS

Dorsal level

Intervertebral osseous bridge

Lobster claw

Cervical level

Exuberant osteophytosis

Narrow cervical canal

Lumbar level

Marginal somatic osseous proliferation

Candle flame

Lobster claw

Intervertebral osseous bridge

Discarthrosis

Acquired vertebral block

Hypertrophy of the posterior articular processes

Bulb appearance of the posterior articular hypertrophy

Anterior subluxation

Posterior subluxation

LUMBAR CT SCAN

Herniated disc

Disc protrusion

Vacuum disc sign

Hypertrophy of the posterior articular processes

Osteoarthritis of apophyseal joints

Osseous proliferations of the nonarticular aspects of the superior apophyseal joint

Osseous proliferations of the nonarticular aspects of the inferior apophyseal joint

Calcification and/or ossification (C/O) of the posterior logitudinal ligament (PLL)

C/O of the yellow ligament (YL)

C/O of the supra-spinal ligament (SL)

Anterior C/O of the posterior articular capsule

Posterior C/O of the posterior articular capsule

Anteroposterior diameter of the spinal canal

Transverse diameters of the spinal canal

*Modified from Leroux JL, Legeron P, Moulinier L, et al: Spine 17:1214, 1992.

MRI

Boden et al. noted abnormal findings in 67% of asymptomatic patients evaluated by MRI. In patients older than 60 years, 57% of MRI scans were abnormal, including 36% of patients with herniated nucleus pulposus and 21% with spinal stenosis. MRI is helpful in identifying other disease processes, such as tumors and infections, and is a good noninvasive study for patients with persistent lower extremity complaints after roentgenographic screening evaluation. MRI should be confirmatory in patients with a consistent history of neurogenic claudication or radiculopathy, but it should not be used as a screening examination because of the high rate of asymptomatic disease. Sagittal T2-weighted images are a good starting point because they give a myelogram-like image. Sagittal T1-weighted images are evaluated with particular attention focused on the foramen. An absence of normal fat around the root is indicative of foraminal stenosis. Axial images provide a good view of the central spinal canal and its contents on T1- and T2-weighted images. Far-lateral disc protrusions are identified on axial T1-weighted images by obliteration of the normal interval of fat between disc and root ( Fig. 41-3 ). The foraminal zone is better evaluated with sagittal T1-weighted sequences, which confirm the presence of fat around the nerve root. Absolute anatomical measures also can be used, as previously discussed ( p. 2065 ). Schnebel et al. noted a 96.6% agreement in pathological anatomy defined by postmyelogram CT and MRI. They noted the greatest disagreement in the MRI L5-S1 axial cuts because of the inability of the cut to be aligned parallel to the disc space. This can be a problem with spinal deformity, including scoliosis and significant spondylolisthesis, because true axial images are difficult to obtain. A disadvantage of MRI is the cost; nonetheless, MRI has become a useful, noninvasive diagnostic tool for the evaluation of patients with extremity complaints.

CT and Myelography

Despite the prevalence of MRI, myelography followed by CT scanning is still accepted and widely used for operative planning in patients with spinal stenosis; it has a diagnostic accuracy of 91%. The addition of CT scanning after a myelogram allows detection of as many as 30% more abnormalities than with myelography alone. Because of the dynamic nature of the study, stenosis not visible on MRI with the patient recumbent may be identified on standing flexion and extension lateral views. CT scanning after myelography characterizes the bony anatomy better than MRI, which helps the surgeon plan decompression surgery. However, imaging of the nerve roots in the foraminal region lateral to the pedicle is not possible because of the confluence of the dura with the epineurium at this point. Thus myelography followed by CT scanning is best suited for dynamic stenosis, postoperative leg pain, severe scoliosis or spondylolisthesis, around metallic implants, or in patients with lower extremity symptoms in the absence of findings on MRI.

Wiesel et al. evaluated 52 patients without low back symptoms or spinal disease and discovered that 50% over 40 years old had abnormal CT scans. They identified herniated nucleus pulposus in 29.2%, facet degeneration in 81.5%, and spinal stenosis in 48.1%. It must be reiterated that abnormal findings occur in as many as 24% to 34% of asymptomatic individuals evaluated with CT and myelography, just as with MRI, so clinical correlation is a must.

Ciric et al. and others used CT to further define lateral recess stenosis and foraminal stenosis. These types of stenosis rarely are identified with myelography. The lateral recess is anatomically the area bordered laterally by the pedicle, posteriorly by the superior articular facet, and anteriorly by the posterolateral surface of the vertebral body and the adjacent intervertebral disc. The superior border of the corresponding pedicle is the narrowest portion of the lateral recess. Measurement of the recess in this area using the tomographic cross section usually is 5 mm or greater in normal patients, but in symptomatic patients the diagnosis is confirmed if the height is 2 mm or less ( Fig. 41-4 ). The foramen is the area of the spine bordered by the inferior edge of the pedicle cephalad, the pars interarticularis with the associated inferior articular facet and the superior articular facet from the lower segment posteriorly, the superior edge of the pedicle of the next lower vertebra caudally, and the vertebral body and disc anteriorly. This area rarely can be seen with myelography. A standard CT scan in the cross-sectional mode suggests narrowing if the foraminal space immediately after the pedicle cut is present for only one or two more cuts (provided the cuts are close together). The best way to appreciate foraminal narrowing is to reformat the lumbar scan, which can create sagittal views through the pedicles and structures situated laterally. Riew et al. reported that myelography influenced operative decision making more than MRI. When requested by the treating physician, CT myelography altered the operative plan in 74% of patients. Even when CT myelography was considered unnecessary, it altered the treatment plan in 49% of patients.

Wiltse et al. described a far-out compression of the root that occurs predominantly in spondylolisthesis when the root is compressed by a large L5 transverse process subluxed below the root and pressing the root against the ala of the sacrum. This diagnosis is best confirmed with a reformatted CT scan with coronal cuts ( Fig. 41-5 ).

Some studies have attempted to correlate clinical outcomes with pathological findings on myelography and CT. In a retrospective review of 251 patients who underwent surgery for lumbar spinal stenosis, operative outcomes, as determined by Oswestry questionnaires, were evaluated and correlated with myelographic results. Patients who had a block on myelogram had a better chance of obtaining a good outcome. A prospective study by Herno et al. confirmed postoperative stenosis in 64% of 191 patients at 4-year follow-up. Slight differences between those with and without stenosis were noted on the Oswestry questionnaire, but not in walking distances; instability was present in 21% without demonstrable clinical effect. In fact, the degree of decompression on CT myelography did not correlate at all with outcomes. Herno et al. reported similar results whether all stenotic levels were decompressed, only one level was decompressed leaving adjacent stenotic levels alone, or incomplete decompression of stenotic levels was done. Other studies by Paine and Surin et al. contradicted this, reporting worse results in patients with severe postoperative stenosis. Nonetheless, decompression of all symptomatic levels with evidence of compression is recommended to enhance neural circulation and function and to avoid reoperation for recurrent spinal stenosis.

Other Diagnostic Studies

Electrodiagnostic studies should be used if the diagnosis of neuropathy is uncertain, especially in patients with diabetes mellitus. The diagnostic use of such studies, including somatosensory evoked potentials, is limited by the lack of prospective studies to determine sensitivity or specificity. Vascular Doppler examinations are useful in identification of inflow problems into the lower extremities and should be accompanied by a vascular surgery consultation when indicated.

Differential diagnosis also can be aided by the use of exercise testing. Tenhula et al. described a bicycle-treadmill test that stresses the patient in an upright position on an exercise treadmill and subsequently in a seated position on an exercise bicycle that allows spinal flexion. This study showed significant postoperative improvement in treadmill walking duration, lower visual analog pain scores, and a later onset of pain. Of the 32 patients evaluated preoperatively, 88% had symptoms during treadmill testing, and 41% had symptoms during bicycle testing. After decompression with or without fusion, only 9% had positive exercise treadmill findings, and 17% developed symptoms on the bicycle. Fritz et al. reported similar results with a two-stage exercise treadmill test. Earlier onset of leg symptoms with level walking and delayed onset of symptoms with inclined treadmill walking were significantly associated with stenosis. Exercise treadmill testing also is useful to help determine baseline function for quantitative evaluation of functional status after surgery.

NONOPERATIVE TREATMENT

Symptoms of spinal stenosis usually respond favorably to nonoperative management. Despite symptoms of back pain, radiculopathy, or neurogenic claudication, conservative management is successful in most patients. Conservative measures should include rest not exceeding 2 days, pain management with antiinflammatory medications or acetaminophen, and participation in a trunk-stabilization exercise program, along with good aerobic fitness. Other methods should be reserved for patients who are limited by pain and should be used to maximize participation in the exercise program. Traction has no proven benefit in the adult lumbar spine. For a patient with unremitting symptoms of radiculopathy or neurogenic claudication, epidural steroid injections may be useful in alleviating symptoms to allow better participation in physical therapy. Epidural steroids can give significant symptomatic relief, although no scientific study has documented long-term efficacy. If spinal stenosis is present with coexistent degenerative arthritis in the hips or knees, some permanent limitation in activity may be necessary regardless of treatment.

Simotas et al. treated 49 patients with an aggressive nonoperative protocol of therapeutic exercise, analgesics, and epidural steroid injection. All had spinal stenosis documented by CT or MRI and symptoms of disabling back, buttock, or leg pain. Except for a few patients with acute neurological changes who initially were prescribed 1 to 2 weeks of bed rest, patients were given one course of oral corticosteroids on a 7-day tapered schedule. An epidural steroid injection was given if symptoms persisted, with a repeat injection given if necessary by the transforaminal route at the point of most severe constriction. A third injection was administered only at the treating physician's discretion, usually for flare-ups during follow-up. For less severe symptoms, nonsteroidal antiinflammatory medications were used for 4 to 6 weeks, and this occasionally was repeated. All patients participated in physical therapy that included postural exercises, gentle lumbopelvic mobilization exercises, and a daily flexion lumbar stabilization program. At 3-year follow-up, only 9 patients (18%) had required surgery after this treatment regimen. At final follow-up, 42% rated overall pain as none or mild, and 56% rated leg pain as none or mild. Sustained improvement was reported in 24% and mild improvement in 28%, with 13% definitely worse. Regarding walking, 40% reported improvement, 35% reported no change, and 25% reported worsening at final follow-up. This study documented the effectiveness of a structured nonoperative treatment regimen in patients with spinal stenosis, reporting satisfactory results in 69% at 3 years.

Epidural Steroid Injection

Although epidural steroid injections have been used in the treatment of spinal stenosis for a number of years, reports in the literature have not substantiated a positive effect, and most prospective reports show no statistically significant benefit. The technique of placementcaudal, translaminar, or transforaminalalso is debated, as is whether fluoroscopy should be used. Using anatomical landmarks for caudal injections, Stitz et al. reported accurate placement in 65% to 74% of patients, with intravascular placement in 4%. Accurate placement of translaminar injections appears to be equally difficult, with successful placement reported in 70%. Complications are infrequent but can occur and include hypercorticism, epidural hematoma, temporary paralysis, retinal hemorrhage, epidural abscess, chemical meningitis, and intracranial air. A 5% incidence of dural puncture has been reported, and, if it occurs, subarachnoid injection of steroids or local anesthetic should be avoided to prevent mechanical or chemical nerve root irritation. Headaches occur in 1% to 5% of patients and are related to dural puncture or the use of the caudal injection route. In patients with headaches associated with caudal injections, the cause has not been determined because dural puncture should not occur at this level, since the dural sleeve has terminated at midsacrum.

No scientifically validated long-term outcomes have been reported to substantiate the use of epidural steroid injections. A meta-analysis showed that epidural steroids have little short-term advantage over placebo for the treatment of leg pain. Unfortunately, studies are divided on the long-term results and the avoidance of surgery. The ideal candidate for epidural steroid injection appears to be a patient who has acute radicular symptoms or neurogenic claudication unresponsive to traditional analgesics and rest, with significant impairment in activities of daily living. We have used this technique successfully in our treatment algorithm for neurogenic claudication and radiculopathy.

OPERATIVE TREATMENT

The primary indication for surgery in patients with spinal stenosis is increasing pain that is resistant to conservative measures. Since the primary complaint is back pain and some leg pain, pain relief after surgery may not be complete. Most series report a 64% to 91% rate of improvement, 42% in patients with diabetes, but most patients still have some minor complaints, usually referable to the preexisting degenerative arthritis of the spine. Neurological findings, if present, improve inconsistently after surgery. In a series reported by Guigui et al., only 30% had complete improvement in motor symptoms after laminectomy, with 58% regaining grade 4 strength or better at a mean follow-up of 3 years. Reoperation rates vary from 6% to 23%. Prognostic factors include better results with a disc herniation, stenosis at a single level, weakness of less than 6 weeks' duration, age less than 65 years, and monoradiculopathy. Reversal of neurological consequences of spinal stenosis seems to be a relative indication for surgery unless the symptoms are acute.

Roentgenographic findings alone are never an indication for surgery, as clearly indicated in the studies by Wiesel et al. and Boden et al. Localized lesions without general involvement respond best. Ganz reported 86% good results in his series of 33 patients treated by decompressive surgery. In patients whose preoperative symptoms were relieved by postural change, the success rate was 96%, compared with only 50% in those not relieved by postural change. Factors predicting outcome are variable, and correlation of imaging with symptoms seems to be the best guarantee of improvement after surgery.

A patient's inability to tolerate the restricted lifestyle necessitated by the disease and the failure of a good conservative treatment regimen should be the primary determining factors for surgery in a well-informed patient. The patient should understand the potential for the operation to fail to relieve pain or to actually worsen it, especially back pain. In addition to the general risks of spinal surgery, the severity of symptoms and lifestyle modifications should be considered. Lumbar spinal stenosis does not result in paralysis, only decreased ambulatory capacity, and conservative management is warranted indefinitely in a patient with good function and manageable symptoms. Cervical and thoracic spinal stenoses, on the other hand, are associated with painless paralysis in the form of cervical and thoracic myelopathy and require closer attention and follow-up.

Principles of Spinal Stenosis Surgery

Decompression by laminectomy or a fenestration procedure is the treatment of choice for lumbar spinal stenosis ( Fig. 41-6 ). Fusion is required if excessive bony resection compromises stability or if isthmic or degenerative spondylolisthesis, scoliosis, or kyphosis is present. Other important indications for fusion include adjacent segment degeneration after prior fusion and recurrent stenosis or herniated disc after decompression. Laminectomy may be preferable in older patients with severe, multilevel stenosis, whereas fenestration procedures, consisting of bilateral laminotomies and partial facetectomies that preserve the midline structures, are an alternative in younger patients with intact discs. Whenever possible, the source of pain should be localized with root blocks preoperatively to allow a more focal decompression. At surgery, specific attention should be directed to the symptomatic area, which may result in more extensive decompression than would normally be done with the pain source unconfirmed. If radical decompression of only one root is necessary, additional stabilization by fusion with or without instrumentation is unnecessary. The removal of more than one complete facet joint usually requires instrumented fusion unless the patient is elderly or has a narrow disc at that level. It is advisable to prepare the patient for fusion in case the findings at surgery require a more radical approach than anticipated. Positioning the patient with the abdomen hanging free minimizes bleeding. If fusion is likely, the hips should remain extended to prevent positional kyphosis. As in disc surgery, a microscope or magnifying loupes and a head lamp are helpful. When proceeding with the decompression, care should be taken to watch for adhesions that can result in dural tears, even if no previous surgery has been done. Frequently the narrowing in the lateral recess and foramen is so great that a Kerrison rongeur cannot be used without damaging the root. Dissection in the lateral recess and foramen usually requires a small, sharp osteotome or a high-speed burr, which allows the surgeon to thin the bone sufficiently to allow removal with angled curets. Unlike disc surgery, for decompression the lateral recess is best seen from the opposite side of the table. The operating surgeon may find it necessary to switch sides during the operation to better view the pathology and nerve roots. Blunt probes with increasing diameters also are useful for determining adequate foraminal enlargement. McLaren and Bailey reported that five of six patients with postoperative cauda equina syndrome had a disc removed from a stenotic spine without adequate decompression of the stenosis. Spinal stenosis should be treated at the same time as the disc herniation. A good approach is to start the decompression at a point of lesser stenosis and work toward the area of most severe stenosis. This often frees the neural structures enough to make the final decompression simpler and decreases the risk of damage to dura or nerve roots.

Adjacent Segment Degeneration

Adjacent disc degeneration and stenosis, or the transition syndrome, deserves special mention. It is known that disc degeneration occurs adjacent to a fusion in 35% to 45% of patients because of the ensuing hypermobility of the unfused joint. Lehmann et al. reported a 45% prevalence (15 of 33) of adjacent segment instability, defined as more than 3 mm of translation, and a 42% occurrence of spinal stenosis (14 of 33), usually above the fusion mass. Adjacent segment stenosis below the fusion mass, although less frequent, always occurred along with stenosis above the fusion.

Adjacent segment breakdown may cause symptoms that require surgery in 30% of patients. Pathology, including spinal stenosis, herniated nucleus pulposus, and instability, may require treatment years after successful surgery. Breakdown is possible one or two levels above lumbosacral fusions and above or below thoracolumbar and floating lumbar fusions. Schlegel et al. reported 58 patients who developed spinal stenosis, disc herniation, or instability at a segment adjacent to a previously asymptomatic fusion that was done an average of 13.1 years earlier. They found that floating lumbar fusions had

a much shorter symptom-free interval (6.3 years) than did other fusions (range, 13 to 16.6 years). Of 37 patients followed for more than 2 years after identification of adjacent segment breakdown and subsequent surgery, 26 (70%) had good or excellent results, and 7 (19%) required additional operative procedures. Some contribution of postoperative loss of lordosis after index surgery was suggested but was not statistically significant in their data. These clinical findings have been substantiated by subsequent biomechanical studies that confirmed kinematic changes in segments adjacent to spinal fusions. Simple malalignment that occurs during patient positioning when the hips are not extended may result in hypolordosis and increase the load across implants, as well as increase posterior shear and laminar strain at adjacent levels. These changes may help to explain the cause of adjacent segment breakdown. Posterior lumbar interbody fusion (PLIF) also resulted in adjacent segment changes in all patients, but this did not affect results at 5 years in the series of Miyakoshi et al.

Rigidity of instrumentation has been hypothesized to correlate with motion at adjacent segments. A study by Even-Sapir et al. using single photon emission computed tomography (SPECT) revealed increased uptake in adjacent motion segments 4 years after surgery in 46% of patients with posterolateral fusions and in 67% of patients with circumferential fusions. However, Doeer et al. found little difference in adjacent segment motion between circumferential fusion and posterior fusion in a calf spine model. Motion at the adjacent segment increased by 29% after posterior fusion and by 21% after circumferential fusion. Thus it is undetermined whether more rigid fusion increases the likelihood of adjacent segment changes.

Fusion is more difficult as the number of levels fused increases, with L4-5 being the most frequent site of pseudarthrosis. Addition of a second level of fusion should be avoided if possible, and fusing a degenerative disc as a prophylactic measure does not seem to be supported by the data available. The actual source of transition syndromes cannot be determined from published studies; however, postoperative hypolordosis and rigidity of the fused segment probably both contribute to the problem. Surgery should attempt to maintain normal segmental lordosis and global sagittal balance, in addition to fusing the fewest segments possible.

Complications are relatively infrequent after decompression for spinal stenosis and occur more often in patients who are older than 80 years or age or who have diabetes. Complications are three times more likely after the age of 75 than before age 40 and are twice as likely in those older than 80 years as in patients younger than 70 years. The mortality rate is 2.3% after spinal stenosis in patients older than 80 years compared with 0.8% in those younger than 75 years. Comorbidities also contribute to poorer patient satisfaction and increased operative complications. Deep venous thrombosis also must be considered in patients after decompression. The incidence of this complication varies but is likely higher than reported. Pulmonary emboli, however, are exceedingly rare. Prophylaxis is best limited to pneumatic compression devices of the foot or calf and early ambulation, since the risk of epidural hematoma from pharmacological agents is greater than the risk of a significant pulmonary event or deep venous thrombosis. Reoperation is necessary in 9% to 23% of patients with spinal stenosis.

Decompression

There are no universal indicators of outcome after decompression. Outcome studies of lumbar decompression for spinal stenosis by Herron and Mangelsdorf and McCullen et al. identified subgroups associated with decreased rates of improvement. The factors they associated with poorer outcomes included questionable roentgenographic confirmation of stenosis, female sex, litigation, previous failed surgery, and the presence of spondylolisthesis. Katz et al. stated that for patients treated with laminectomy with or without fusion, the patient's self-assessment of health was the best predictor of satisfaction, with cardiac comorbidity also being predictive.

A 5-year follow-up study by Jnsson et al. described the ideal patient as one who has a pronounced constriction of the spinal canal, insignificant lower back pain, no concomitant disease affecting walking ability, and a symptom duration of less than 4 years. In their study, successful results were present in 63% to 67% of patients, deteriorating to 52% at 5 years, with reoperations necessary in 18%. Specifically, patients with 6 mm or less anteroposterior canal diameter preoperatively had better results. Similar results were reported by Airaksinen et al., with 62% good or excellent results at 4.3 years after surgery in 438 patients who had undergone decompression only. Results were worse in patients with hip arthritis, diabetes mellitus, previous surgery, vertebral fracture, or a postoperative complication. At an average 8.1-year follow-up, Katz et al. reported that 75% of patients were satisfied with the results of decompression despite severe back pain in 33% and inability to walk two blocks in 53%.

Prospective analysis of the Maine Lumbar Spine Group (Atlas et al.) identified a cohort of 67 patients with spinal stenosis studied 4 years after surgery and compared with a less severely affected group of 52 patients treated conservatively. Better outcomes were noted in the operatively treated group despite their being more severely affected and worse functionally before surgery: 70% of the operatively treated group and 52% of the conservatively managed group reported improvement in back or leg pain. Some loss in functional improvement was noted in the operatively treated group over time, with no worsening in the conservative group. This also was confirmed in a later study by the same group. Of 148 patients with lumbar spinal stenosis followed for 4 years after treatment, 81 had been treated with surgery and 67 were treated without surgery. Those treated with surgery had more severe symptoms and worse function preoperatively but had better outcomes at 4-year follow-up. Back and leg pain were much improved or completely gone in 70% of the operatively treated group compared with 52% of those treated nonoperatively. Of those having surgery, 63% were satisfied with their current status, compared with 47% of those treated nonoperatively. Maximal benefit was noted 3 months after surgery. A modest decline in outcomes over the 4-year follow-up was noted in the operatively treated group, with no significant changes, either positive or negative, in those treated nonoperatively.

Progressive instability after decompression does not predict poor results. Poor ambulation has been correlated with roentgenographic signs of instability; however, Frazier et al. found that normal walking, sensory deficits, and ability to perform activities of daily living improved despite instability. Some further anterolisthesis is tolerated well after decompression, and it is appropriate to observe these patients for further symptoms before recommending fusion because as many as 30% of patients develop anterolisthesis after decompression.

Midline Decompression (Neural Arch Resection)

TECHNIQUE 41-1Perform the procedure with the patient under general endotracheal anesthesia. Position the patient prone using the frame of choice. Make the incision in the midline centered over the level of stenosis. Localizing roentgenograms should be taken if the level of dissection is uncertain. Carry the incision in the midline to the fascia. Strip the fascia and muscle subperiosteally from the spinous processes and laminae to the facet joints to expose the pars interarticularis. Take care to avoid damaging facet joints that are not involved in the bony dissection. Identify and remove the spinous processes of the levels to be decompressed. Then clear the soft tissue with a sharp curet. Dissect the lower edge of ligamentum flavum from the lamina with a curet and remove the lamina with a Kerrison rongeur. If the lamina is extremely thick, a high-speed drill with a diamond or side-cutting burr can be used to thin the outer cortex to allow easier removal of the inner portion with a Kerrison rongeur. Take special care in removal of the lamina and ligamentum flavum. The neural structures will be found compressed, and the usual space for instrument insertion may not be

available. Remove the lamina until the pedicles can be felt. Using the pedicle as a guide, identify the nerve root and trace it out to the foramen. With a chisel or rongeur, carefully remove the medial portion of the superior facet that forms the upper portion of the lateral recess ( Fig. 41-7 ). Check the foramen for patency with an angled dural elevator or graduated probes. If there is further restriction, carry the dissection laterally and open the foramen, taking care not to remove more than half of the pars. Undercutting into the foramen is especially helpful in this regard. Inspect the disc and remove gross herniations unilaterally but try to avoid bilateral annulotomy because this compromises stability. Usually the disc is bulging, and the annulus is firm. Remove the annulus and bony ridge ventrally if it is kinking the nerve. This procedure involves some risk of nerve injury and requires a bloodless field. If safety is a concern, a complete facetectomy may be better. Complete the dissection at all symptomatic levels. Decompression should be from the caudal aspect of the most proximal pedicle to the cephalad aspect of the most distal pedicle, allowing observation of the lateral margins of the dura in the lateral recesses. This can be done with preservation of the proximal portion of the lamina and the intervening ligamentum flavum at the level above and below. Many failed decompressions are the result of inadequate decompression of the foraminal region, so probing the foramen is mandatory to determine if the decompression is adequate. If no obstructions are noted and all areas have been decompressed adequately, close the incision. If desired, take a large fat graft from the incision or buttock and place it over the laminectomy defect and then close the incision.

Less-Invasive Decompression.

The consequences of bone and ligament removal must be considered when performing decompression for spinal stenosis. Removal of the spinous processes, laminae, variable portions of the facets and pars, supraspinous and interspinous ligaments, ligamentum flavum, and portions of facet capsules is routine during these operative procedures. Denervation of the paraspinal musculature occurs with wide exposures, which results in altered muscle function. A minimally invasive technique allows decompression of the significant compressing anatomy while preserving paraspinal muscles, the spinous processes, and intervening supraspinous and interspinous ligaments.

Spinous Process Osteotomy (Weiner et al.)

TECHNIQUE 41-2 ( Fig. 41-8 )Patient positioning and localization of spinal levels are as described in the previous technique. Make a midline incision to expose the dorsolumbar fascia. Make a paramedian incision in the fascia, preserving the supraspinous and interspinous ligaments with subperiosteal dissection of the paraspinal muscles from the spinous process and laminae. Take care to avoid lifting the multifidus muscles beyond the medial aspect of the facet joint to preserve their innervation. With a curved osteotome, free each spinous process from the lamina at its base. Release only the levels shown to be affected on preoperative imaging. Once the spinous process is freed, retract it to one side with the paraspinal muscles beneath the retractor and the other blade of the retractor beneath the multifidus muscles to expose the midline ( Fig. 41-8, A ). Resect approximately half of the cephalad lamina and one fourth of the caudal lamina along with the underlying ligamentum flavum. Using a loupe or microscope for magnification, undercut the lateral recess and open the foraminal zone ( Fig. 41-8, B ). Complete laminectomy is recommended for severe stenosis or congenital stenosis involving all anatomical zones (central, lateral recess, and foraminal zones). Close the incision in routine fashion, allowing the spinous process to return to its normal position with suture of the fascia ( Fig. 41-8, C ).

Weiner et al. reported a 47% improvement in the Low Back Outcome Score and a 66% improvement in average pain level in 46 of 50 patients evaluated 9 months after surgery. Spinous process osteotomy was done at one to four levels; the only complications were dural tears in four patients. Although 3 patients died of unrelated causes, 38 of the 46 remaining patients were satisfied or very satisfied with their operative results. On reexploration or postoperative CT scans, spinous processes usually united with the remaining lamina in patients with short decompressions, although nonunion did not correlate with poor results. Complete laminectomy may be necessary if adequate decompression is not possible through the limited laminotomy in patients with severe involvement.

Microdecompression

Microdecompression can be done in patients without disc herniations or instability including degenerative spondylolisthesis. This is a technically demanding procedure and is not recommended for patients with severe stenosis or congenital stenosis, which require complete laminectomy. McCulloch reported that decompressions were done at one to five levels without intraoperative complications in 30 patients treated for neurogenic claudication unresponsive to nonoperative measures. One superficial wound infection occurred. Of the 30 patients, 26 were very satisfied or fairly satisfied with their results; all but one stated that they would recommend the procedure to a friend with a similar problem.

TECHNIQUE 41-3 (McCulloch) ( Fig. 41-9 )Place the patient in a kneeling position to increase interlaminar distance and identify the operative level on standard roentgenograms. Make a midline incision centered over the affected levels documented on preoperative imaging studies. Make a paramedian fascial incision on the most symptomatic side 1 cm from the midline. Elevate the multifidus muscles subperiosteally from the spinous process and laminae but do not retract them beyond the medial aspect of the facet joint. Obtain unilateral interlaminar exposure and maintain it with a discectomy retractor. Under microscopic magnification, perform laminotomy cephalad until the origin of the ligamentum flavum is encountered. Use undercutting to preserve as much dorsal bone as possible; angle the microscope to accomplish this. In a similar fashion, resect the proximal one fourth of the caudal lamina, thus completing removal of the ligamentum flavum from origin to insertion. Again, angling of the microscope into the lateral recess allows further decompression of the cephalad and caudad nerve roots and lateral dura. Once decompression is completed on one side, angle the microscope toward the midline for contralateral decompression ( Fig. 41-9, A ). Rotation of the operative table allows better viewing of the contralateral structures ( Fig. 41-9, B ). Use a no. 4 Penfield elevator or similar instrument to release adhesions between the dura and opposite ligamentum flavum, which is then resected in a similar fashion. Remove the bone at the base of the spinous processes of the cephalad and caudal levels to provide adequate vision of the opposite side ( Fig. 41-9, C ). Some removal of the deepest portions of the interspinous ligament also will be necessary to adequately view the structures across the midline.

AFTERTREATMENT.

Special considerations are not necessary after a simple decompression. The patient should be examined carefully for the first few days for new neurological changes that may indicate the formation of an epidural hematoma. The patient is encouraged to walk on the first day.

Sutures are removed at 14 days if nonabsorbable sutures have been used. The same limitations as after disc surgery ( Chapter 39 ) apply to decompressions without fusion. For patients engaged in heavy manual labor, a permanent job change may be required. Return to work also is similar to return after disc surgery.

Decompression with Spinal Fusion.

The indications for spinal fusion with decompression for spinal stenosis are becoming more clearly defined. The prevalence of postoperative problems related to instability is highly variable, possibly because of the great variations in the extent of the operative decompression. Shenkin and Hash noted a 6% occurrence of postoperative spondylolisthesis in patients with bilateral facetectomy and a 15% occurrence when three or more facets were removed. White and Wiltse noted subluxation after decompression in 66% of patients with degenerative spondylolisthesis. They suggested that a fusion be done in conjunction with decompression in (1) patients younger than 60 years of age with instability caused by the loss of an articular process on one side, (2) patients younger than 55 years of age with a midline decompression for degenerative spondylolisthesis that preserves the facets, and (3) patients younger than 50 years of age with isthmic spondylolisthesis. Others have suggested that patients with spinal stenosis and concomitant scoliosis should be treated with fusion because of the potential for instability and progression of deformity. The complete removal of one facet, or more than 50% resection of both facets, may result in instability. In addition, generalized spinal stenosis that requires extensive decompression with the loss of multiple articular processes may require fusion. When complete bilateral facetectomies are necessary, the addition of a lateral fusion may be difficult, and the bone graft may impinge on the exposed nerve roots. In this instance, an anterior interbody fusion is warranted to prevent postoperative instability. New posterior segmental fixation instrumentation for posterior spinal fusion has decreased the high incidence of pseudarthrosis after these long lumbar fusions.

The complications of this procedure are similar to those of disc surgery; however, the risk of nerve root damage and dural laceration is greater. The rates of infection, thrombophlebitis, and pulmonary embolism also are slightly higher. When a facet has been partially resected, later facet or pars fracture may account for a recurrence of symptoms. However, Getty et al. found that in their series the most important cause of failure to relieve symptoms was inadequate decompression. Postacchini and Cinotti noted bone regrowth at an average follow-up of 8.6 years in 88% of 40 patients who had total laminectomy for spinal stenosis. Bone regrowth was noted in all patients with associated spondylolisthesis.

Degenerative Spondylolisthesis and Scoliosis

Junghanns first described degenerative spondylolisthesis in 1930, separating the pathology from isthmic spondylolisthesis. He described this entity as pseudospondylolisthesis, which was later renamed by Newman as degenerative spondylolisthesis because of the associated arthritic changes noted on roentgenograms. Degenerative spondylolisthesis often is accompanied by spinal stenosis, which usually is the cause of aggravation of symptoms and must be considered in patients with degenerative spondylolisthesis.

Unstable stenosis, as described by Hanley, exists in two forms: unisegmental, which is manifested by degenerative spondylolisthesis, and multisegmental, which is represented by degenerative scoliosis. Thus treatment of either of these entities is really a continuum, and therefore they are considered together.

INCIDENCE

Degenerative spondylolisthesis occurs in patients older than 40 years and rarely is identified before that time. It is characterized by a posterior slip of one vertebra on another of less than 33%, with deformity occurring at L4-5 six times more often than at other lumbar levels and four times more often above a sacralized L5. Other levels may be similarly affected, with L3-4 affected more often than L5-S1. This disorder was found in approximately 4% of autopsy specimens and was identified in 10% of women older than 60 years. Women are affected by this condition four to six times more often than men, probably because of ligamentous laxity and abnormal facet morphology. Diabetes has been present in a disproportionate number of study patients, and Imada et al. found that oophorectomized patients had a three times greater rate of degenerative spondylolisthesis than did non-oophorectomized patients.

Degenerative scoliosis develops in patients with previously straight spines after the age of 40 years, typically affecting the lumbar spine with an associated lumbar hypolordosis, lateral olisthesis, and spinal stenosis. Men and women are affected more equally than patients with idiopathic curves, with about 60% to 70% of those affected being women. Degenerative scoliosis also affects fewer segments (2 to 5 segments) than does adult idiopathic scoliosis (7 to 11 segments), with an equal distribution of right and left lumbar curves. Symptoms of spinal stenosis occur most often in degenerative curves that have defects in both the convexity and concavity, possibly because significant degenerative changes preceded the development of the scoliosis. As a result, treatment for degenerative scoliosis often is necessary to relieve spinal stenosis by decompression, with instrumented fusion to prevent instability and further progression of deformity.

ANATOMY AND BIOMECHANICS

Degenerative spondylolisthesis is differentiated from isthmic spondylolisthesis by the presence of an intact pars. Because the arch is intact and moves forward with the L4 vertebral body, progressive spinal stenosis occurs in addition to facet degenerative changes. The true deformity of degenerative spondylolisthesis does not appear to be pure translation but rather a rotary deformity that may distort the dura and its contents and exaggerate the appearance of spinal stenosis. Existing theories to explain the development of degenerative spondylolisthesis include the primary occurrence of sagittal facets and disc degeneration, with secondary facet changes accounting for anterolisthesis. The sagittal facet theory suggests a predilection for slippage because of facet orientation that does not resist anterior translation forces and, over time, results in degenerative spondylolisthesis. The disc degenerative theory proposes that the disc narrows first, and subsequent overloading of the facets results in accelerated arthritic changes, secondary remodeling, and anterolisthesis.

It appears that facet arthritic changes are more severe than disc space narrowing, with the most advanced anterolisthesis present when disc narrowing is more pronounced. Thus a continuum seems to exist as degeneration progresses. In addition, facets that are aligned in a more sagittal orientation provide less stability at the involved level, but whether these changes result from chronic instability or from a primary anatomical variant is debatable. Boden et al. showed that sagittal facet angles of more than 45 degrees at L4-5 predicted a 25 times greater likelihood of degenerative spondylolisthesis. Despite the increased frequency of degenerative spondylolisthesis in women, there appears to be no gender-specific difference in facet orientation, which calls into question the theory that sagittal facet joints are a primary cause of degenerative spondylolisthesis. Sagittal facet orientation has been correlated with disc space narrowing, suggesting that disc narrowing increases loading of the facet, resulting in secondary facet changes. Regardless of the exact nature of the first inciting event, this instability causes facet arthritis, disc degeneration, and ligamentous hypertrophy, which all contribute to produce symptoms. Facet orientation therefore may be part of the consideration of potential instability when evaluating a patient for surgery, especially decompression alone.

In degenerative scoliosis, severe degenerative changes result in significant back pain. Mechanical insufficiency of the lumbar spine with a decrease in lordosis is caused by the degenerative changes. Curves usually are less than 60 degrees, with 81% developing a lateral olisthesis. Spinal stenosis occurs most commonly at the apex of the primary curvature (either concave or convex), with variable amounts of structural rotary deformity, although as many as 33% of patients develop symptoms because of stenosis within the distal compensatory curve. Characteristically, patients with neurogenic claudication and degenerative scoliosis do not obtain relief simply by bending forward or sitting. These patients require support of the torso by the upper extremities to obtain relief of their neurogenic claudication. Finally, osteoporosis is a significant consideration in the treatment of these patients. Although there is no cause-effect relationship between bone density and degenerative scoliosis, these patients usually are older and more predisposed to senile osteoporosis, and compression fractures may complicate the treatment.

NATURAL HISTORY

Most of the literature describing natural history concerns spinal stenosis rather than degenerative spondylolisthesis. Matsunaga, Ijiri, and Hayashi reported that in 145 patients examined annually for a minimum of 10 years progressive spondylolisthesis occurred in 34%, and further disc space narrowing continued in those without further slip. There was no correlation between roentgenographic findings and a patient's clinical picture. In fact, low back pain actually improved in those with continued disc space narrowing, which may imply autostabilization. Of the 145 patients, 76% remained without neurological deficits; however, 83% of patients with neurological symptoms, including claudication and vesicorectal disorder, deteriorated and had a poor prognosis. This is in agreement with an earlier study by Matsunaga et al, which showed that, over 60 to 176 months, progressive slipping occurred in 30% of patients without significant effect on clinical outcome.

Degenerative scoliosis occurs in 6% to 30% of the aged population, with most curves being minor. Rotary subluxation is variable and appears to be worse after decompression surgery without fusion. Men and women appear nearly equally affected, although a propensity for women has been noted. Curves can be progressive, but the natural history has not been elucidated conclusively. Grubb et al. reported progression of 1 to 6 degrees a year in 8 patients, but the small number of patients limits generalizations.

It appears from the literature that observation and nonoperative treatment are warranted initially in these patients. For those who develop neurological symptoms or who fail to improve despite nonoperative treatment, operative intervention should be considered.

CLINICAL EVALUATION

Symptoms of degenerative spondylolisthesis include back pain, neurogenic claudication, radiculopathy, and rarely bowel and bladder dysfunction. Although symptoms of spinal stenosis are more common, with leg pain and claudication in 68%, 32% have axial back pain only. Radiculopathy occurs in 32%, and cauda equina is rarely noted (3%). Overlap of symptoms of neurogenic claudication and vascular claudication requires a careful evaluation. Peripheral neuropathy also must be considered in the differential diagnosis.

Symptoms of neurogenic claudication are present in 71% to 90% of patients with degenerative scoliosis and usually cause patients to seek medical attention, with deformity incidentally noted. These symptoms usually do not improve with forward bending, and to obtain relief patients support the trunk with the arms or assume a supine position. This is in contrast to the usual patient with spinal stenosis and neurogenic claudication. Radiculopathy from facet overgrowth, foraminal stenosis within the concavity of curvature, or nerve root tension along the curve convexity can occur, although neurological deficits are rare and back pain is ubiquitous. Thus primary treatment is directed at decompression of spinal stenosis, with fusion or instrumentation indicated based on the potential for instability.

Physical findings are nonspecific in most patients. Motion usually is preserved, but patients guard against hyperextension. Symptoms may be reproduced by this maneuver in the presence of spinal stenosis. Neurological examination rarely identifies significant motor, sensory, or reflex deficits; however, any abnormal findings should be documented. Evaluation of distal pulses is necessary to help rule out peripheral vascular disease and vascular claudication. Bilateral absence of tendo calcaneus reflexes may be an indicator of peripheral neuropathy. Flattening of the lumbar spine is representative of degenerative change, and a lumbar prominence on forward bending accentuates the convexity of a coronal deformity, which is absent in those with degenerative spondylolisthesis. Sagittal and coronal balance should be estimated clinically.

DIAGNOSTIC IMAGING

Roentgenography

Roentgenographic imaging is straightforward in the diagnosis of degenerative spondylolisthesis and scoliosis. Anterolisthesis typically is seen at L4-5 without an associated pars defect. Standing roentgenograms are imperative because 15% of deformities spontaneously reduce on supine imaging. Variable disc space narrowing is indicative of degenerative changes. Flexion-extension lateral views may reveal instability, which is considered to be present when 4 to 5 mm of translation or more than 10 to 15 degrees of sagittal rotation is identified. Degenerative scoliosis is evident, with a solitary lumbar curve and associated degenerative disc changes. Flattening of the normal lumbar lordosis is present, and rotary subluxation may be evident in some patients, with lateral olisthesis of varying degrees.

Bending films add information for operative decision making. Flexion-extension views should be obtained with any spondylolisthesis or kyphosis, and lateral bending films should be obtained for coronal deformities. These studies are useful in determining whether supplemental anterior surgery is necessary or if all decompression and stabilization can be done from a posterior approach alone.

The Ferguson anteroposterior view shows any significant degenerative changes in the lumbosacral joint and allows better visualization of the transverse processes of L5. Hypoplastic transverse processes also should prompt the consideration of interbody fusion because of the paucity of bony substrate for fusion, especially for lumbosacral fusions.

CT Myelography and MRI

CT myelography and MRI are used as indicated for the evaluation of spinal stenosis ( p. 2065 ) and may show facet overgrowth, hypertrophy of the ligamentum flavum, and rarely disc herniation. Because of the arthritic component affecting the facet joints, synovial cysts also have been documented. When identified, synovial cysts may alter the treatment plan, necessitating more extensive decompression into the foraminal zone.

Deformity, specifically scoliosis, is still a common indication for the use of CT scanning after myelography. Because of the difficulty of obtaining MR images parallel to the disc spaces in patients with scoliosis or spondylolisthesis and leg pain or claudication, myelography is better for evaluating the spinal canal and nerve roots. Flexion and extension myelography studies should be obtained with the patient standing and should be followed by CT scanning. Dynamic images may reveal nerve compression not apparent on supine images. The axial images show the bony deformities clearly, including the facet subluxations and bony spurring that often cause lower extremity complaints, lateral recess stenosis from ligamentum flavum hypertrophy, and pedicular kinking of the concave nerve roots. Pathology lateral to the dorsal root ganglion cannot be seen, however, because of the absence of subarachnoid space. MRI may supplement the information obtained by CT myelography, especially regarding the soft tissues, and often is helpful in decision making.

Although axial images are limited in larger deformities, with attention to the neural foramen and far-lateral regions, MRI can be very useful in preoperative evaluation. Disc hydration is a useful piece of information provided by T2-weighted sequences and may influence whether surgery stops at or includes a lower degenerative segment. Desiccation of the L5-S1 disc should prompt extra consideration to end at the adjacent disc above or include the degenerative segment. Adjacent segment pathology is common, and in older patients it may be prudent to include such diseased discs. Discography also may be warranted in such cases.

Discography

Kostuik advocated the use of discography to decide if instrumentation must include the lumbosacral disc or other discs not anticipated to be included in the fusion for adult idiopathic scoliosis. Although no prospective randomized study has proved the usefulness of discography for this purpose, Kostuik reported satisfying results with this testing protocol. Grubb and Lipscomb also used this protocol for determining fusion levels and reported good results, although they did not find discography necessary in patients with degenerative scoliosis. In fact, in their study, discography was abandoned for the evaluation and treatment of patients with degenerative scoliosis because fusion with instrumentation to the L5 or S1 levels was done in most patients, and all degenerative levels evident on plain roentgenograms were included in the fusion. Results were only slightly less satisfactory in patients with degenerative scoliosis than those in patients with idiopathic scoliosis.

Other Studies

EMG and nerve conduction velocity studies are useful in differentiating peripheral neuropathy from other abnormalities, especially in patients with diabetes mellitus. Nerve conduction studies should be used if the diagnosis is uncertain because the results of surgery are less predictable in patients with neuropathy. In addition, arterial Doppler studies, angiography, or vascular surgery consultation may be necessary in some patients. A noninvasive functional study using the bicycle-treadmill test ( p. 2068 ) also may help to differentiate neurogenic from vascular claudication.

NONOPERATIVE TREATMENT

Conservative treatment is warranted for patients with neurogenic claudication from spinal stenosis caused by degenerative scoliosis and spondylolisthesis. Standard interventions should include short periods of rest, antiinflammatory medications, and, rarely, bracing. The use of specific exercise programs has not been evaluated; however, trunk-stabilization exercises and low-impact aerobic exercises seem to benefit this patient group.

Epidural Steroid Injection

Epidural steroid injections have been advocated, but randomized or placebo-controlled trials are lacking to determine their effectiveness in the treatment of spinal stenosis. The antiinflammatory effect of corticosteroids is the basis for their use. Long-term effects also may be a result of the local anesthetic agent administered with the steroid. Epidural steroid injections appear to be most beneficial in patients with radiculopathy; fluoroscopic guidance is helpful because a dorsal medium septum may be present, which can restrict injectate diffusion. No literature exists to support the use of a series of three epidural steroid injections unless symptoms improve partially after the first injection. If the first injection is done without fluoroscopy and is ineffective, a second injection can be done with fluoroscopy to ensure proper placement and diffusion. Further injections are not warranted if there is not a favorable response after a single well-placed injection. If epidural steroid injection is successful, physical therapy should be instituted.

OPERATIVE TREATMENT

For patients with unremitting back and leg pain after adequate nonoperative treatment, operative intervention is indicated. Only 10% to 15% of patients with degenerative spondylolisthesis require surgery. Caution should be exercised in proceeding with surgery for back pain alone because adjacent level pathology may account for symptoms.

Decompression

Degenerative Spondylolisthesis.

Similar criteria for decompression alone are used for degenerative and isthmic spondylolisthesis. For patients with significant disc collapse and no pathological motion on dynamic roentgenograms and for young patients with multiple level congenital stenosis, decompression alone is indicated. Epstein reported the results of decompression in 290 patients. Laminectomy was done in 249 patients over an average of 3.4 levels, and limited decompression by fenestration, hemilaminectomy, or coronal hemilaminectomy was done over an average of 1.7 levels. Progressive slips required fusion in only 5 patients (1.7%), and secondary decompressions were necessary in 7 (2.8%). Roentgenographic findings did not correlate with clinical results, which were good or excellent in 82%. Several studies reported successful 5-year results in 90% of patients after fenestration procedures and in 71% to 81% after laminectomy.

Subsequent fusions were required in only 0.8% to 4%. Epstein reported a higher rate of reoperation for instability in patients with decompression and simultaneous disc excision. Of the eight patients undergoing reoperation for instability, five had discectomy at the time of the index decompression. The violation of the disc may have contributed to instability, which has been documented biomechanically in ligament-sectioning studies.

In the meta-analysis of Mardjetko, Connolly, and Shott, a review of 11 papers representing 216 patients showed that decompression without fusion was satisfactory in 69%, with progressive slips in 31% postoperatively. This occurrence is similar to the natural history of degenerative spondylolisthesis and did not compromise the operative results in most reports.

Degenerative Scoliosis.

Decompression is a viable option for a patient with symptomatic spinal stenosis and minimal kyphosis, a neutral sagittal vertical axis, and no instability on dynamic imaging. Attempts should be made to avoid destabilizing the spine during decompression, and fusion should be done if more than 50% of both facets or an entire facet is resected. Degenerative scoliosis should be considered the coronal variant of spondylolisthesis; if the ligamentous structures are violated, facet and disc resection is necessary and fusion should be considered.

Decompression and Fusion

Degenerative Spondylolisthesis.

Patients with preserved disc height are prone to instability after decompression, and fusion should be considered. Other indications for the addition of fusion include osteoporosis (which predisposes to pars fracture), absence of osteophytes on roentgenograms, and minor nonpathological motion on roentgenograms. After the report of Herkowitz and Kurz, the addition of fusion to decompression has become the standard in the treatment of degenerative spondylolisthesis and concomitant spinal stenosis. Outcomes are improved with the addition of in situ fusion regardless of whether there is a solid roentgenographic fusion. Most studies report that fusion rates are better with instrumentation, but clinical results are not different from those after uninstrumented fusion. Bridwell et al., however, reported better functional results, correction of sagittal alignment, and a fusion rate of 87% in patients with instrumented fusions compared with a 30% fusion rate in those with uninstrumented fusion. Similar findings were documented by Mardjetko et al. in their meta-analysis, with 90% satisfactory outcomes in instrumented fusions compared with 69% satisfactory results with decompression alone at 12- to 84-month follow-up. Booth et al. reported 6.5-year follow-up of 36 patients who had decompression, autogenous iliac crest bone grafting, intertransverse fusion, and pedicle screw instrumentation for degenerative spondylolisthesis. Stenosis at the level of previous decompression did not recur, but five patients developed symptomatic adjacent level transition syndromes, with another seven having asymptomatic adjacent level degenerative changes. Sagittal alignment was maintained at the fused segments in all patients. Eighty-three percent were satisfied with their results, 86% reported that back and leg pain was better than before surgery, and 77% stated that they would have the surgery again. Dissatisfaction was associated with more than four medical comorbidities. Major complications were rare (2%) and involved screw breakage without symptoms. In Nork et al.'s report of 30 patients evaluated after decompression and instrumented posterior fusion, 93% were satisfied with their outcomes, based on SF-36 questionnaires, and showed improvement in abilities to perform heavy and light activities, participate in social activities, sit, and sleep. Pain, depression, and medication use also were decreased in this group. The fusion rate was 93% at an average of 128 days. Complications were rare, and poorer outcome was associated with more severe preoperative stenosis or the occurrence of a complication.

Pedicle screw implants appear to be better for maintaining anatomical alignment than distraction constructs and Luque rods or rectangles, which are reported to worsen anterolisthesis. Fusion rates with pedicle screws also are higher (86%) than with rod constructs (69%). Although fusion is not a necessity for successful results, it does appear to improve results when added to adequate decompression. Long-term results, however, are necessary to determine if adjacent segment transition syndromes are more likely in patients with more rigid instrumented fusions than in those with posterolateral in situ fusions. Therefore decompression and fusion, with or without instrumentation, are recommended for patients with degenerative spondylolisthesis and spinal stenosis in the absence of multiple medical comorbidities.

Degenerative Scoliosis.

In a patient with a flexible spine and mild to moderate scoliosis, symptomatic spinal stenosis is treated with traditional decompression. Occasionally, facetectomy or partial pedicle resection is necessary to decompress symptomatic nerve roots. In this situation, posterior intertransverse fusion is recommended. If laxity is present, instrumentation should extend from neutral and stable vertebrae at each end of the construct. Ending the instrumented fusion at a level of kyphosis potentiates adjacent segment changes and may result in an adjacent compression fracture. This occurs at the thoracolumbar junction in as many as 15% of patients, which may be an indication for prophylactic polymethylmethacrylate augmentation. Rotary subluxation also can occur after decompression alone in the presence of unstable degenerative scoliosis. The fusion should end at a neutrally rotated, level, and stable end vertebra and should restore sagittal and coronal balance. Simmons and Simmons advocated avoiding fusion into the thoracic spine. Care should be taken to avoid stopping instrumentation and fusion at any apical segments or at the apex of a kyphosis or scoliosis because this creates deformity later.

Posterior Lumbar Interbody Fusion (PLIF) and Transforaminal Lumbar Interbody Fusion (TLIF)

Recent advances in instrumentation and techniques have resulted in an increased use of the PLIF technique with threaded interbody fusion cages. These cages may be allograft bone dowels or metal cylinders filled with bone graft. Biomechanically, bone dowels or metal cages appear equivalent.

The use of these as stand-alone implants inserted through a posterior approach for grade I degenerative spondylolisthesis has provided only moderate stability in in vitro studies. Both implants appear to be susceptible to loosening with cyclic fatigue when used alone. This likely is a result of the posterior decompression and the resection of the disc for implant placement. It appears that significant disruption of the disc results in increased motion when loaded. If these implants are used posteriorly, further stabilization is necessary and is best provided by pedicle screw implants ( Fig. 41-10 ). No reports have yet compared the outcomes of PLIF with traditional intertransverse fusion, with or without instrumentation, for degenerative spondylolisthesis. Ideal indications for interbody fusion in a patient with degenerative spondylolisthesis include discographically concordant single-level axial back pain with radiculopathy, minimal disc degenerative changes, and preserved disc height; it also is indicated for revision surgery with an inadequate posterior fusion mass. PLIF or TLIF also is appropriate for small or absent transverse processes at the levels to be fused, which would make intertransverse fusion impractical. The technique for PLIF is described on p. 2094 .

Anterior Spinal Fusion

Anterior spinal fusion can be used for the treatment of degenerative spondylolisthesis if some indirect spinal decompression is provided by eradication of the disc, restoration of disc height, and ligamentotaxis by placement of structural bone graft or cage after distraction of the disc space and tensioning of the posterior ligamentous structures. Mardjetko et al. reported that the fusion rate with anterior procedures in pooled studies was 94%, with an 86% satisfaction rate, suggesting that there may be a place for anterior lumbar interbody fusion in this patient population because the results are similar to those obtained with decompression and instrumented fusion.

Decompression and Combined Fusion (360-Degree Fusion)

Any adult spinal deformity that requires fusion of L5-S1 also requires anterior fusion of the L5-S1 disc. This can be accomplished by an anterior procedure or a posterior interbody procedure. In patients with degenerative scoliosis, kyphosis, and a positive sagittal vertical axis, anterior interbody fusion usually is necessary to restore the height of L5-S1 and restore

the normal segmental lordosis at this level. Because of the disproportionate amount of lumbar lordosis at this and the adjacent L4-5 level, the kyphosis that ensues often can be corrected with restoration of anterior disc height and without osteotomy. Posterior interbody procedures often do not allow safe excision of a contracted annulus and anterior longitudinal ligament for height restoration. Flexion and extension roentgenograms help with this aspect of the decision-making process by showing a narrow, immobile disc space.

For sagittally neutral or lordotic spines and an intact disc, posterior interbody fusion techniques can be considered. If disc height is maintained, it is easier to maintain normal lordotic disc anatomy during interbody fusion.

Regardless of the technique, restoration of normal segmental anatomy is of paramount importance at the L4-5 and L5-S1 levels. The objective of interbody grafts at these levels is to recreate the segmental lordosis of 20 to 28 degrees. This can be done with mesh cages, femoral ring allografts, iliac crest tricortical autograft wedges, fibular rings, or carbon fiber trapezoidal cages. Threaded interbody cylinder devices should be used cautiously in these applications, especially at the L4-5 and L5-S1 levels. Because of the geometry of the cylinder, the lordotic contour of the intervertebral disc cannot be replicated with this device ( Fig. 41-11 ).

Adult Idiopathic Scoliosis

Treatment of adult scoliosis, both idiopathic and degenerative, requires a different approach from that used for typical adolescent idiopathic scoliosis. By definition, adult idiopathic curves are curves that are diagnosed in a patient older than 20 years of age. These curves have been present since adolescence but, for any number of reasons, the diagnosis was delayed. The prevalence of adult scoliosis is from 4% to 6%, with a similar female-to-male ratio as idiopathic adolescent scoliosis ( Chapter 38 ).

In contrast, approximately 60% of patients with late-onset degenerative scoliosis are female. Degenerative curves are short segment, usually lumbar, and less severe than those in idiopathic scoliosis. Symptoms of spinal stenosis are more common than in patients with degenerative scoliosis, and myelographic defects, if present, are within the primary curve. The goal of treatment of degenerative scoliosis is to relieve back pain and the symptoms of spinal stenosis, whereas the treatment goals for adult scoliosis usually are pain control and deformity correction.

SAGITTAL BALANCE

In the treatment of adult spinal deformities, it is important to understand the normal sagittal relationships. In a normal spine, the primary curvature is kyphosis of the thoracic spine, which develops first in infants. Subsequent to upright posture, the secondary lordotic curvatures in the cervical and lumbar spine develop between the ages of 5 and 15 years. Curves in males and females are similar at the cessation of growth, although the curves develop more quickly in females.

Sagittal balance is the alignment that is necessary to center the head over the pelvis or hips in both the sagittal and coronal planes. A plumb line dropped from the center of the C7 vertebral body is referred to as the sagittal vertical axis (SVA). On the lateral long-cassette view, the plumb line normally falls through or behind the sacrum. Normal values for the SVA are 3.2 3.2 cm for adult populations, with negative values indicating a position behind the sacral promontory. The coronal vertical axis (CVA) is represented in a similar fashion with a plumb line on the posteroanterior roentgenogram, with a normal balance through the center of the sacrum. Any translation of the coronal vertical axis to either side of the midline is considered decompensation.

Sagittal and coronal vertical axes are used to evaluate and estimate global balance. Global balance is the result of segmental alignment of the functional spinal unit, and regional alignment of the cervical, thoracic, and lumbar segments. Bernhardt and Bridwell measured 102 roen