Clinical manifestations

Presentation and course

Lumbar spinal stenosis is more common in older individuals and males, and clinical manifestations range from asymptomatic to disabling symptoms. Pain can be located in the lower back, buttocks, thighs, legs, and, in some cases, feet. Pain may be characterized as aching, burning, or cramping. Symptoms can present unilaterally with foraminal stenosis or bilaterally with central canal stenosis (57). In more severe cases, loss of balance and decreased vibration sensation can be seen (43). The distribution of motor and sensory symptoms is not dermatomal unless lumbar radiculopathy is also present. Hall and colleagues described symptoms involving the whole limb in 78% of cases, only above the knee in 15%, and only below the knee in 6% (32). Moderately sensitive symptoms for the diagnosis of lumbar stenosis are pain below the buttocks (88%), worsening of pain when walking (71%), and poor balance (70%), but the specificity of these symptoms for lumbar spinal stenosis is low (42).

The hallmark presentation of lumbar spinal stenosis is the waxing and waning of symptoms, known as neurogenic claudication. Pain and discomfort typically present when the patient stands erect or walks and subside after sitting or leaning forward. Continued walking is often accomplished with a progressively stooped posture. Often, a point is reached where it is impossible to continue because of the pain, and the patient must stop walking, sit down, and rest. Sitting will usually relieve symptoms sufficiently to allow further walking. Relief of symptoms with flexion of the spine explains why it is often easier to walk up an incline than on a level surface and forms the basis of the bicycle test. A patient with neurogenic claudication will be able to cycle (spine flexed) but will not be able to walk erect (spine extended) for an equivalent time (71).

Neurogenic claudication should be differentiated from vascular claudication. Patients with vascular claudication often improve while standing rather than needing to sit or lean forward, as seen in neurogenic claudication. When symptoms are present above the knee, triggered by erect posture, and there is a positive shopping cart sign (leaning forward), there is a likelihood ratio of 13 for neurogenic claudication. When symptoms occur below the knee, and the patient feels relief while standing, vascular claudication is strongly predicted, with a likelihood ratio of 20 (64). Neurogenic claudication is usually of gradual onset and, once established, results in symptoms that may become disabling (progressively lower symptom threshold or diminished exercise tolerance). A multicenter Delphi process identified a series of clinical criteria when taken together as the N-CLASS, having a sensitivity of 80% and specificity of 92%. This weighted clinical scoring model is based on the following criteria: age greater than 60 years, positive 30-second extension test (development of symptoms with fully extending the back with the knees straight), pain in both legs, leg pain relieved by sitting, leg pain relieved by leaning forward, and SLR positive at greater than 60 degrees (26).

Neurologic examination is normal in most patients, but decreased reflexes or mild motor or sensory deficits can be seen after provoking maneuvers. In a study of 93 patients older than 40 years of age who had back pain with or without radicular symptoms, wide-based gait, thigh pain following 30 seconds of lumbar extension, abnormal Romberg test, and the presence of neuromuscular deficits were significantly associated with a final clinical diagnosis of lumbar spinal stenosis (42).

Prognosis and complications

The few natural history studies of lumbar spinal stenosis are characterized by small sample sizes and are difficult to interpret because of methodological bias in the stratification of surgically and conservatively managed patients. In general, these studies suggest that established neurogenic claudication is often nonprogressive. Saal and colleagues documented the outcome of 52 patients treated conservatively: the majority were unchanged clinically at follow-up and could function within their symptomatic limits without surgery. Only four patients eventually had surgery (73). Johnsson and colleagues followed the clinical course of 19 conservatively treated (mean follow-up time of 31 months) and 44 surgically treated (mean follow-up time of 53 months) patients with lumbar spinal stenosis. The two groups had no substantial outcome difference (39).

Although the symptoms of neurogenic claudication may eventually become intolerable, lumbar spinal stenosis typically does not progress to fixed neurologic deficit. Many patients with lumbar spinal stenosis find that they can remain mobile within the specific constraints of their symptoms. Patients with neurogenic claudication often become remarkably well-adapted to their symptoms and simply stay within the limits of pain-free mobility (ie, they may walk in a stooped fashion with the aid of a cane or walker and sit when necessary).

Clinical vignette

A 65-year-old man with coronary artery disease and chronic, mild lower back pain presented with about 3 months of disabling leg pain when walking. The pain was located in the buttocks and upper legs in a symmetric distribution and began after some 10 minutes of walking on the level. The pain would ease after sitting and resting for several minutes, and he could usually continue walking. He could ascend the two flights of stairs to his apartment and continued to ride a bicycle. The initial referral was for suspected vascular claudication, but no clinical evidence of vascular insufficiency was found, including a normal ultrasound examination of blood flow in the lower limbs.

Neurologic examination showed limited lumbar spine mobility but no local tenderness or deformity. The right straight leg raising sign elicited pain in the back and right buttock at 50° elevation of the right leg above the horizontal. Muscle bulk, tone, and power in the lower limbs were normal. Tendon reflexes were symmetric at the knees and depressed at the right ankle. The vibration sensation was diminished at the toes, consistent with the patient’s age.

MRI of the lumbosacral spine showed widespread degenerative spine disease, central disc bulges at L3-L4 and L4-L5, and a focal, right-sided, posterolateral disc protrusion at L5-S1. Deformation of the thecal sac at the lower lumbar levels, and multilevel neuroforaminal narrowing were present. Needle EMG revealed mild chronic partial denervation with reinnervation in the medial head of the right gastrocnemius muscle. A diagnosis of neurogenic claudication secondary to lumbar spinal stenosis associated with a mild right-sided S1 radiculopathy was made. The patient underwent a trial of conservative treatment (physical therapy and anti-inflammatory medication) but felt that no symptomatic improvements had been made after 6 months. His reduced mobility significantly diminished his quality of life. A L3-S1 decompressive laminectomy with posterior fusion was performed. Nine months after surgery, he could walk continuously for 30 minutes on a level platform without pain or needing to rest.

Biological basis

Etiology and pathogenesis

Lumbar spinal stenosis results from narrowing of the intervertebral foramina, lateral recesses, the central canal, or a combination of all due to congenital or acquired causes, leading to compression of neurovascular structures. Typically, multiple levels of the lumbosacral spine are involved in degenerative discogenic, ligamentous, or spondylotic change. It has been proposed that the stenosis process begins with bulging or loss of intervertebral discs with subsequent involvement of the facet joints (03). Spondylolisthesis (the anteroposterior slippage of a vertebra on the one adjacent to it), most often at L4-L5, is frequently coexistent. Central canal stenosis, lateral recess stenosis (15), and foraminal stenosis may coexist and contribute to neurogenic claudication. A common scenario is a progressive, chronic, degenerative lumbar spine disease on a background of a congenitally narrow lumbar spinal canal. Cross-sectional narrowing of the central canal occurs at adjacent segmental levels or central narrowing at a single segment together with contiguous multilevel neuroforaminal narrowing. Subjects of normal stature with short vertebral pedicles that narrow the anteroposterior diameter of the spinal canal may become symptomatic in their third to fourth decades of life. In the Framingham population study, the prevalence of lumbar spinal stenosis was approximately 70% in patients older than 60 years who underwent a CT scan and were found to have a congenital narrowing of the spinal canal (41). Congenital stenosis is frequently exacerbated by secondary degenerative disease and has been associated with high rates of multilevel involvements and re-operation as they have shorter pedicles and narrower anteroposterior spinal canal diameter (50).

The pathogenesis of neurogenic claudication resulting from lumbar spinal stenosis is not completely understood. Symptoms are thought to be caused by direct mechanical compression or indirect vascular compression of the nerve roots or the cauda equine (21). Microvasculature compression leading to root ischemia has been proposed to explain paresthesias, pain, and weakness (99). Another theory is the accumulation of metabolites and inadequate oxygenation of the cauda equina at multiple levels (71).

Reduction in the cross-sectional dimensions of the lumbar spinal canal and an increase in epidural pressure are the result of an erect posture. Extension of the lumbar spine, axial loading with standing, and the exertion of walking are all factors that contribute to symptoms. Takahashi and colleagues have demonstrated increased epidural pressure in human subjects with walking and venous congestion of the cauda equina by myeloscopy (92).

Porter, among others, has suggested that with low-pressure compression of two adjacent spinal segments, exercise may induce congestion in the venous plexus draining the cauda equina and secondary arterial insufficiency to these neural structures (71). In Olmarker's pig model, 2-level compression diminished blood flow to the cauda equina and reduced nerve conduction velocity when the nerve roots were stimulated directly (68). Similar mechanisms may precipitate neurogenic claudication: walking in the presence of lumbar spinal stenosis results in ischemic neurapraxia of the cauda equina and produces pain in the buttocks or legs. Pain subsides when the stenosis is relieved by sitting and flexing the spine, and the muscular activity of walking ceases. Ikawa and colleagues performed an electrophysiologic analysis on a chronic lumbar spinal stenosis model of rats (35). They demonstrated that ectopic firing was elicited by venous stasis only in the lumbar spinal canal stenosis animals, thereby suggesting that venous stasis may play a factor in the pathogenesis of neurogenic intermittent claudication.

Epidemiology

Lumbar spinal stenosis is a clinical syndrome with associated typical radiographic findings. Several series have established the prevalence of asymptomatic lumbar spinal stenosis in the older population. Boden and colleagues found that 21% of asymptomatic patients older than 60 years had lumbar spinal stenosis on MRI (09). Others have reported incidence of radiological stenosis by MRI as high as 65% in patients without symptoms of lumbar spinal stenosis (30; 28). The worldwide incidence of lumbar spinal stenosis was estimated at 102 million cases per year, with 3.5 times more cases occurring in low- and middle-income countries than in high-income countries (72). Lumbar spinal stenosis is particularly prevalent in older people and is the most frequent cause of spinal surgery in this population (91). Large prospective studies of the natural history of lumbar spinal stenosis or neurogenic claudication are lacking. Only a few studies have examined the prevalence of lumbar spinal stenosis in the general population. Yabuki and colleagues estimate the prevalence of lumbar spinal stenosis to be 5.7% in Japanese subjects aged 40 to 79 years, affecting about 3,650,000 Japanese (106). In that study, lumbar spinal stenosis was associated with an advanced age (60 years or older), diabetes mellitus, depressive symptoms, urological disorders, and osteoarthritis or fracture. Studies show a dose relationship between smoking and lumbar spinal stenosis requiring surgical treatment in a cohort of construction workers (48).

Prevention

Anecdotal experience suggests that practical public health advice (posture, correct lifting technique, avoidance of trauma) may reduce the likelihood of developing spinal degenerative disease and lumbar spinal stenosis (62).

Differential diagnosis

The diagnosis of neurogenic claudication as a manifestation of lumbar spinal stenosis requires the recognition of the highly characteristic history and radiologic evidence of lumbar spinal stenosis. An International Delphi study noted that the symptoms in ascertaining the diagnosis of symptomatic lumbar spinal stenosis are leg pain with walking, leaning forward to relieve pain, use of a shopping cart or bicycle for relief, and the presence of motor or sensory symptoms. Additional features include the presence of foot pulses, weakness, or low back pain, walking with a limp, and lacking a history of diabetes (93).

Degenerative lumbar spine disease in the absence of lumbar spinal stenosis can cause intrinsic low back and leg pain that may be exacerbated by activity; however, it does not demonstrate a stereotyped relationship to walking and does not improve with sitting. Usually, the pain from lumbar spinal stenosis is less intense and more diffuse. Radiculopathies frequently result in focal neurologic findings, whereas the neurologic examination in lumbar spinal stenosis is most often normal. More often than not, straight leg raising signs are absent in lumbar spinal stenosis and often present with L5 or S1 nerve root injury. Facet joint-associated ganglions, spondylodiscitis, paravertebral abscesses, neoplasms, and intraspinal hematoma (spontaneous or posttraumatic) are less common causes of radiculopathies. Hematomas in the lumbar ligamentum flavum are rare and might mimic neurogenic claudication (87).

Conditions such as hip osteoarthritis or greater trochanteric pain syndrome may mimic or accompany lumbar spinal stenosis. Greater trochanteric pain syndrome, also known as trochanteric bursitis, is characterized by tenderness over the trochanteric bursa, whereas hip osteoarthritis pain may be elicited by internal rotation of the flexed hip. Symptom relief after injection of anesthetic or corticosteroids in the trochanteric bursa or hip joint may also help differentiate from lumbar stenosis symptoms (43).

Confusing conditions

Vascular claudication may be confused with neurogenic claudication. The symptoms of vascular claudication are evident after a relatively constant walking distance. Pain tends to begin in the calves, spread proximally, and resolve rapidly with rest. Typically, vascular claudication causes pain in the calves when walking up an incline. Evidence of peripheral vascular disease (diminished distal pulses, poor capillary filling, reduced skin temperature, shiny skin, or hair loss) is apparent on examination of the feet. Low back pain is usually absent. Proximal arterial stenosis (bifurcation of the aortic, common iliac, hypogastric, or hip arteries) can be easily confused with neurogenic claudication, as hip or buttock pain is often the leading symptom and involves the area of the trochanter, groin, or thigh. Arteriography, the examination of choice, is performed during treatment to guide angioplasty (53).

On the other hand, neurogenic claudication is less symptomatic when walking uphill due to the flexed position of the spine. The bicycle test facilitates differentiation of these two distinct entities: the leg pain of vascular claudication occurs with any muscular activity of the legs that increases demand for increased regional blood flow to muscle (vascular claudication is triggered by the exertion of pedaling a bicycle). In contrast, neurogenic claudication is not caused by cycling as the increased demand for regional blood flow to nerve roots is mitigated by the increased cross-sectional area of the spinal canal with the spine in flexion.

Dabasia and colleagues reported a case of neurogenic claudication from lumbar epidural varices secondary to obstruction of the inferior vena cava due to follicular lymphoma (16). Successful chemotherapy resulted in resolution of the varices and the symptoms of neurogenic claudication, suggesting that epidural venous engorgement can result in neurogenic claudication in the absence of spinal stenosis.

Distal polyneuropathy often presents with distal symmetric pain in the feet or legs that is worse when walking but also present at rest. In contrast to neurogenic and vascular claudication, the pain of polyneuropathy has a distinct distribution (typically the feet) and quality (burning or painful paresthesias) and is often associated with distal symmetric sensory or motor deficit.

Associated or underlying disorders

Some of the other causes of acquired lumbar spinal stenosis include achondroplastic dwarfism, hereditary exostosis, Paget disease (103), diffuse idiopathic spinal hyperostosis, fluorosis (seen in parts of India with excessive fluoride supplementation of drinking water), postsurgical changes, lipomatosis, trauma, acromegaly (20), ankylosing spondylitis, and transthyretin related amyloidosis.

Previous studies have suggested that 5% to 15% of patients with wild-type transthyretin amyloidosis have a history of lumbar spinal stenosis, which may precede cardiomyopathy (the cardinal manifestation of that disease), by 5 to 10 years (05). As lumbar spinal stenosis may precede cardiomyopathy or polyneuropathy (the cardinal manifestations of transthyretin-related amyloidosis), pathological testing to assess for amyloidosis is suggested in patients with lumbar spinal stenosis who have red-flag symptomatology for amyloidosis, such as carpal tunnel syndrome, polyneuropathy, cardiomyopathy, or cardiac arrhythmia or a positive family history for these conditions. This is especially important because effective disease-modifying treatments are now available for transthyretin-related amyloidosis (19). A pathological study on ligamentum flavum specimens harvested from patients with lumbar spinal stenosis aged 62 to 82 years showed transthyretin-related amyloid deposits in 43 of 95 specimens (108). In that study, the burden of amyloid deposits correlated with the degree of spinal instability, as assessed by lateral functional flexion-extension radiography, and with thickness of ligamentum flavum on the MRI. Another study on ligamentum flavum samples from patients with lumbar spinal stenosis showed amyloid deposition in 74 of 94 samples, the subtyping of which showed transthyretin-related amyloidosis in 65% of the cases (59). Patients with amyloid deposits were older (73.1 ± 9.2 vs. 64.6 ± 10.1 years) and had thicker ligamentum flavum.

Diagnostic workup

Although there is no gold standard test, MRI imaging is the study of choice given the high quality of soft tissue delineation. CT myelography is still utilized, especially in patients with adjacent spine fusion surgery and in patients for whom MRI is contraindicated (those with some pacemakers, spinal cord stimulators, and other implanted neuromodulation and stimulation devices). Flexion extension radiographs are used to evaluate segmental instability. In a systematic review assessing the accuracy of imaging testing, MRI and 3D-MRI showed the highest sensitivity, but overall, the study showed no superiority between CT, MRI, and myelography (18).

Measures of spinal canal and thecal sac anteroposterior diameter or cross-sectional area are used to make the radiologic diagnosis. CT is superior for visualizing the bony elements of the spine, lateral recess, and intervertebral foramen but exposes patients to ionizing radiation.

MRI provides soft tissue detail (including neural structures) without exposure to ionizing radiation.

A variety of measurements have been used to establish the diagnosis of lumbar spinal stenosis. Schonstrom and colleagues considered the following dimensions of the cross-sectional area of the thecal sac at its narrowest point significant: greater than 100 mm² (normal), 76 mm² to 100 mm² (moderate stenosis), and less than 76 mm² (severe stenosis) (79). Anteroposterior diameters of the lumbar spinal canal ranging from 10 mm to 15 mm have been used as a cut-off between stenosis and a normal canal (44). There is also a qualitative classification of lumbar spinal stenosis based on the morphology of the dural sac observed on T2 axial MRI. The classification grades from A to D where A and B show CSF presence while C and D do not (most severe) (75).

Imaging of lumbar spinal stenosis is surveyed in reviews by Saifuddin (74) and Schonstrom (78). However, when evaluating for symptomatic lumbar stenosis pain, neuroradiologists diagnosed lumbar stenosis in 65% of asymptomatic volunteers, and the central canal diameter was no different in symptomatic and asymptomatic patients (30; 28). In adults over 65 with no red flags, another study showed that the severity of disc and facet degeneration was not correlated with neurogenic claudication and pain (34). Other studies indicated a poor relationship between pain and a variety of anatomic measures of stenosis accounting for facet hypertrophy, ligamentum flavum hypertrophy, and stenosis centrally at the foramen and at the lateral recess (11; 100).

MRI sensitivity and specificity in clinical mild to moderate lumbar stenosis were 59% and 42%, respectively (30; 28). No substantial relationship exists between the degree of stenosis and the severity of symptoms. Li and Yen also observed a poor correlation between clinical and radiologic diagnoses of spinal stenosis (55). In their study, the final number of surgical candidates remained unchanged despite an increase in the number of MRI and CT scans. Clinical judgment is required in each case to decide on the probable contribution of the stenosis to patient symptoms. A more recent study on CT evaluation of L5-S1 foraminal stenosis indicated that completing this evaluation in the sagittal plane (compared to coronal and axial) had the best inter-observer agreement and was the best predictor of symptoms and operability defined as no clinical improvement with 3 months of conservative treatment (82).

Electromyography is the most useful test to diagnose nerve root compromise in lumbar spinal stenosis. Typically, multiple root involvement is seen at the paraspinal level as opposed to a single root in radiculopathy. In particular, this is useful when lumbar stenosis and radiculopathy coexist. Abnormal electrophysiologic findings at the paraspinal levels seen in spinal stenosis can be mimicked by the combination of diabetic polyneuropathy and polyradiculopathy (104) and may require thoracic paraspinal examination to distinguish these mimics. Incongruence between imaging and EMG findings can be seen in patients with upper lumbar stenosis who often present with bilateral electrophysiological abnormalities, which are more commonly seen in the lower lumbar rather than the upper lumbar myotomes (69). Paraspinal EMG mapping is suggested to be a more sensitive indicator of lumbar stenosis than limb electromyography, especially in mild or asymptomatic cases (31; 107).

In one study, electromyography distinguished mild and moderate lumbar stenosis in asymptomatic subjects and back pain controls with 79% sensitivity and 50% specificity, whereas paraspinal mapping score showed 30% sensitivity and 100% specificity (31). Another study demonstrated a lack of correlation between the presence of denervation detected in paraspinal mapping and clinical evidence for lumbar radiculopathy in patients with lumbar spinal stenosis, suggesting that such abnormalities are indicative of stretch or damage to the posterior rami due to radiculopathy (29).

The standard nerve conduction studies and needle EMG are, however, frequently normal in symptomatic patients with lumbar spinal stenosis. This observation complements animal models that suggest the neural basis of neurogenic claudication is ischemic neuropraxia of sensory more than motor roots of the cauda equina. This proposed mechanism is congruent with the observed time course of recovery of neurogenic claudication symptoms after rest with the spine flexed. This time course is too rapid for demyelination and remyelination or axonal loss with regeneration. As a result, electrophysiologic investigation has a role limited to investigating focal symptoms and signs of nerve or nerve root injury and evaluating mimics of lumbar stenosis. Nerve conduction studies performed after exercise may reveal slowing in proximal motor and sensory segments (08). London found a reduced H-wave recruitment curve after exercise (56). Similarly, dynamic F-waves showed prolonged latencies and chronodispersion. Stolov found that somatosensory evoked responses were abnormal in several patients with normal nerve conduction studies. F wave studies after walking stress tests provide more information for diagnosing neurogenic claudication (07). However, post-exercise nerve conduction study is not routinely used in the clinical setting to diagnose lumbar spinal stenosis.

Additional techniques that have been considered include measuring distal motor latency with root-level stimulation in foraminal versus central lumbar stenosis (37). Cauda equina motor conduction time has also shown promise in distinguishing patients with lumbar spinal stenosis (80). Studies rely on using cauda equina conduction time calculated from the latencies of compound muscle action potentials, F-waves, and motor-evoked potentials via magnetic stimulation, showing that cauda equina conduction time is prolonged in patients with symptomatic lumbar spinal stenosis compared with age-matched controls (36). A study identified differences in patients with more diffuse cauda equina impairment versus radicular symptoms using cauda equine conduction time to differentiate them (65). Neither technique has been used enough to know its added utility in patient evaluation.

Treadmill exercise testing with recording of the mean time to the development of claudication pain was helpful in quantifying the functional impact of spinal stenosis, as well as predicting response to surgery (85).

Selective nerve root block had a high sensitivity, specificity, and positive predictive value in identifying patients with foraminal stenosis who go on to have successful clinical outcomes with endoscopic decompression surgery (54).

Management

Operative treatment for lumbar spinal stenosis is the most common spine surgery performed in adults over 45 years of age in the United States. The primary goal of surgical treatment is to decompress the neural elements affected by the stenosis. This can often be performed with a decompression alone, such as a laminectomy, laminotomy, or discectomy. However, in cases of instability, whether pre-existing or iatrogenic from surgical decompression, a concurrent fusion operation may be warranted to ensure a stable spine.

A wide range of literature discusses the pros and cons of concurrent fusion in the treatment of lumbar stenosis and spondylolisthesis. A study of 65 patients with symptomatic lumbar stenosis and stable grade 1 spondylolisthesis randomized to complete laminectomy and medial facetectomy or decompression with pedicle screw fixation showed an improved SF-36 out to four years in the fusion group and a reoperation rate in the decompression only group of 34% (27). Reoperation in the fusion group began at year 3. However, a study of 247 randomly assigned patients to decompression versus fusion with 1- to 2-level disease and stratified for spondylolisthesis indicated no difference in pain or function (measured by the Oswestry Disability Index) out to six and a half years with similar reoperation rates (22). A PRISMA-compliant meta-analysis did not show superiority of decompression with fusion over decompression alone with regard to pain, Oswestry Disability Index, and patient satisfaction, although follow-up rates ranged from 12 to 120 months in the studies included (105).

Studies that look at long-term surgical fusion outcomes also need to account for the observation that surgery for adjacent segment disease escalates after the tenth year postfusion to 25.6% and to 37.5% at year 15 (61). A study shows a reoperation rate of 6.2% at 2 years, 10.8% at 5 years, and 18.4% at 10 years, with a 10-year accumulated reoperation rate of 20.6% after anterior fusion, 12.6% after posterior fusion, and 18.6% after decompression (40). European registry data of 4768 patients indicate that outcome is not impacted by age at the time of surgery and that quality of life outcomes were most related to the ASA status of the patient (86). Similarly, a Canadian Spine Outcomes and Research Network database epidemiological study indicated that patients undergoing surgery for lumbar spinal stenosis had a better outcome (defined as better leg pain scores and better overall outcome) if participating in regular preoperative exercise, and patients had a poorer outcome if receiving compensation, experiencing depression, and had poorer overall health (ASA greater than 2). Worse outcome was noted as well if there was a delay to surgery, a greater preoperative level of disability, preoperative anticonvulsant use, and prior history of spine surgery (33). However, pain and functional outcome are not negatively impacted by increased time for conservative therapy compared to proceeding immediately to surgery (110). Amundsen and colleagues concluded that for lumbar spinal stenosis, the outcome was most favorable for surgical treatment (02). But they still recommended conservative management for many symptomatic patients from lumbar spinal stenosis, with the hope that those with unrelieved symptoms can be treated surgically later with a good outcome as demonstrated by Zweig and colleagues (110). Weinstein and colleagues, though their study was somewhat limited, note that in patients with symptomatic spondylolisthesis, nonrandomized comparisons of surgical intervention versus nonsurgical interventions revealed substantially better outcomes in pain and function, as well as satisfaction, 3 months after surgery (101). In a study published a year later, it was found that patients with degenerative spondylolisthesis and spinal stenosis treated surgically showed substantially greater improvement in pain and function during a period of 2 years than patients treated nonsurgically. Prior to the subjective opinion to proceed to surgery, 68% received physical therapy, 54% had spine injection therapy, 55% were treated with NSIADs, and 27% received opioids. In the conservative treatment group, only 39% received spine injection therapy (102), indicating that conservative therapy was not a standard or robust intervention prior to surgical treatment in this study. A 3-year study concluded the overall benefit of surgery over conservative management in lumbar spinal stenosis over this time period, using the EuroQol five-dimensional questionnaire (EQ-5D) and the Spinal Stenosis Measure (SSM) symptoms, physical function, and satisfaction subscales as outcome measures (12). Study design was limited by patient self-selection, which resulted in the nonsurgical patients being older and less socially supported, more joint-related disorders, and less back and leg pain, with a more stable or improving recent course than the surgical group. The type of surgery done (decompression vs. decompression with fusion) was by physician preference rather than randomization or set protocol, with a reoperation rate of 12.4%. Surgical complications included dural tear (6.1%), epidural bleeding (2 of 412), and infection (1 of 412), with reoperation in 46 out of 412 due to re-stenosis. Clinical cauda equina syndrome with appropriate symptoms and signs is an absolute indication for surgery. A long-term outcome study of results after laminectomy for lumbar spinal stenosis in an elderly group of patients found that the Oswestry Disability Index is more sensitive than the Pain Visual Analog Scale score in assessing prognostic value and that patient satisfaction is difficult to prognosticate, underscoring the particularities that this population presents regarding functionality assessment (10). Predictors for lower Oswestry Disability Index and visual analogue pain scores after lumbar decompression with or without fusion over a 10-year follow up included nonsmoking status, absence of prior lumbar surgery, self-rated health, lower BMI, and use of painkillers for less than 12 months preoperatively (96). Prognostic value of preoperative factors, when considered, report a negative influence of low-back pain and female sex. One study concludes that lumbar laminectomy appears to be the most cost-effective treatment strategy for patients with symptomatic lumbar spinal stenosis (13). In efforts to analyze the role and outcome of spinal fusion in conjunction with decompression for lumbar spinal stenosis without concomitant degenerative spondylolisthesis, Sigmundsson and colleagues concluded that predominance of back pain over leg pain is associated with inferior outcome (84). Although additional spinal fusion improves unadjusted outcome, the benefit is small and not clinically significant and generally disappears in the adjusted analysis.

There remains controversy about the best surgical approach for patients with degenerative scoliosis and concomitant lumbar stenosis. Depending on the deformity and the location of the stenosis within the deformity, surgical treatment can vary from decompression alone, decompression with short fusion, and decompression with long fusion. Surgical decision-making is also dependent on the co-morbidities of the patient and the invasiveness of the proposed surgical treatment. Studies have demonstrated that restoration of proper spinal alignment with long fusion has favorable outcomes regarding back pain and back pain with walking, although the comparison group had significantly lower preoperative pain scores such that a floor effect may have been confounding. The doubling of the surgical time and quadrupling of blood loss with long fusion is an important consideration in patient selection of persons with significant comorbidities (06).

For less invasiveness, the technique of unilateral laminotomy for bilateral decompression was developed. This also offers better preservation of spinal stability after decompression. Oertel and associates, in their 10-year experience with unilateral laminotomy for bilateral decompression in symptomatic lumbar spinal stenosis, confirm that unilateral laminotomy for bilateral decompression allows good and long-lasting operative results in patients; furthermore, postoperative deterioration, recurrences, and spinal instability are infrequent in their experience (67). By preserving the posterior ligament complex integrity, spinoplasty is a good alternative to long-segment fusion to enable desired decompression and, at the same time, avoid iatrogenic instability (14; 95). However, two systematic reviews of minimally invasive lumbar decompression show no difference in outcome compared to open decompression with fusion, with complication rates of 3.3% versus 12.8%, and reoperation rates of 5.8% versus 16.3%, respectively (66; 77).

Interspinous process devices have been used to treat lumbar stenosis as well. The insertion of a device, with or without bone graft between the lumbar spinous processes can provide distraction and, thus, indirectly decompress the spine. These devices can be surgically placed with relatively little morbidity but rely on the bony integrity of the spinous processes as well as indirect decompression. Postoperative complications may include infection, iatrogenic instability, pseudarthrosis, hardware failure, and the need for future surgery because of the development of new diseases at the same or adjacent levels (49).

The surgical goals of direct decompression with or without stabilization can also be achieved with minimally invasive techniques. The clinical outcomes are comparable with an equivalent open procedure, but the advantages include less soft tissue injury, minimized blood loss, reduced postoperative pain, and shortened length of stay. However, they are reliant on more imaging modalities and an additional learning curve. In their study of MIS-transforaminal lumbar interbody fusion, Jhala and colleagues report good clinic-radiological outcomes in addition to finding it superior in terms of postoperative back pain, blood loss, hospital stay, recovery time, and medication use (38). On the contrary, minimally invasive transforaminal lumbar interbody fusion requires more radiation time during surgery and has been reported to require longer operative time than a conventional open lumbar fusion, and it is more technically challenging to treat bilateral symptoms using a unilateral approach (51; 98).

With continued technological advances and using the least invasive methods, a clinical series showed noninferiority of biportal endoscopic decompressive laminectomy to microscopic lumbar decompressive laminectomy over a 1-year follow-up (70). Comparable outcomes in low back-associated disability and quality of life, back pain, leg pain, and neuropathic pain were accomplished in this limited series. The microendoscopic techniques continue to evolve. This represents another level of less invasive surgery, but further comparative studies of effectiveness and efficiency are still needed.

Nonoperative treatment for patients with various lumbar spinal stenosis conditions is a reasonable alternative to spinal surgery. Several studies involving patients with nonsurgical therapy suggest that 15% to 43% will continue to enjoy improvement over a 1- to 5-year follow-up period (04; 90). Epidural steroids were not better than lidocaine alone in a study of 400 patients treated with epidural injection with central stenosis with leg pain scores greater than 4. Transforaminal approach was directed based on patient symptoms and intralaminar injection at the level below greatest stenosis. No sham treatment was used as comparison (24). This includes observational studies of epidural injection, which show a 32% operation rate in the first 24 months and 44% satisfaction with a single injection protocol with no further treatment needed over 2 years (17). Improvement with caudal and interlaminar epidural approaches with local anesthetic only or with steroids was seen in a long-term follow-up of up to 2 years in patients suffering with chronic lumbar spinal stenosis (58). They also reported significantly better results with the interlaminar approach. Symptomatic medical treatment of lumbar spinal stenosis may include limited bedrest, antiinflammatory and analgesic medication, and physical therapy. Fluoroscopically guided caudal epidural steroid injection was effective for the management of degenerative lumbar spinal stenosis, especially central canal stenosis, with excellent short-term and good long-term pain relief, without significant outcome predictors (52). Epidural steroids can lead to sustained cortisol suppression at least through the 3-week measurement window of a study in patients with lumbar spinal stenosis. This effect is more pronounced when using particulate steroids (23).

Siebert and colleagues emphasized the need for multimodal conservative management including patient education, pain medication, de-lordosing physiotherapy, and epidural injections for symptomatic patients with mild to moderate lumbar spinal stenosis (83). They feel surgery is indicated only in patients where conservative management proves ineffective after 3 to 6 months, or in patients with severe symptoms. Epidural injections, group exercise, and manual therapy resulted in improvement in walking distance to a similar degree. Manual therapy resulted in statistically greater improvement in Spinal Stenosis Score than the other treatments but the change was not to a degree that would be considered clinically meaningful (76). A comprehensive conservative treatment program of manual therapy, education, and exercise was better than self-directed exercise with improved walking distance and statistically but not clinically significant pain score improvement over the one-year follow-up (01). In one study, patients who successfully completed conservative treatment for lumbar stenosis were more likely than not to have sustained these benefits for five years or more posttreatment (94). In military healthcare beneficiaries with low back pain applying lumbar manipulation techniques, both lumbopelvic and lumbar neutral gap groups experienced statistically significant reductions in pain and disability at 48 hours posttreatment (88). The authors claim the outcomes are predictable if the patients satisfy the clinical prediction rule.

Lipoprostaglandin E1 and EP4 agonist at high concentrations might be potential therapeutic agents because they are expected to increase blood flow in nerve roots in patients with spinal canal stenosis, although their value has not been demonstrated as better than or additive to pregabalin in a controlled trial (45). The basis for this effect is not clear. Murakami found that lipo prostaglandin E1 increased cauda equina blood flow and relieved symptoms of neurogenic claudication in a small series of patients with lumbar spinal stenosis (63). EP4 agonist at high concentrations might be a potential therapeutic agent as it is expected to increase blood flow in nerve roots in patients with spinal canal stenosis (81). Limaprost, a prostaglandin E1 analog, was approved for the treatment of lumbar spinal stenosis in Japan in 2001 (89). However, limaprost is not a marketed drug in the United States, and studies typically needed for FDA approval for this class of drugs have not yet been conducted (60).

Outcomes

Knutsson and colleagues reported that although obese patients achieved significant pain reduction and improved ambulation and QoL after surgical treatment for lumbar spinal stenosis, obesity was associated with a higher degree of dissatisfaction and poorer outcomes after surgery (47).

After analyzing randomized controlled trials comparing surgical versus nonoperative treatments in participants with lumbar spinal stenosis confirmed by clinical and imaging findings, Zaina and colleagues concluded that they have very little confidence in concluding whether surgical treatment or a conservative approach is better for lumbar spinal stenosis, and provided no new recommendations to guide clinical practice (109). However, they drew attention to the fact that the rate of side effects ranged from 10% to 24% in surgical cases, and no side effects were reported for any conservative treatment. They emphasized that clinicians should be very careful in informing patients about possible treatment options, especially given that conservative treatment options have resulted in no reported side effects.

Gauthé and associates reported a case of symptomatic postoperative pneumocephalus, an uncommon complication after lumbar decompression (25). After one day, computed tomography was performed to explore intense lumbar pain and revealed a voluminous pneumorachis. This was followed by a generalized tonic-clonic seizure. Imaging revealed a voluminous pneumocephalus responsible for a significant space-occupying effect on the frontal lobe. A conservative treatment was initiated, including bed rest, oxygen therapy, neurologic monitoring, and antiepileptic therapy, followed by progressive improvement and a total radiological regression occurring in 21 days. This case highlights the fact that prevention of postoperative pneumocephalus should include a systematic repair of iatrogenic dural tear, and even in the presence of severe symptomatic manifestations, conservative treatment is associated with total recovery.

Special considerations

Anesthesia

The existence of spinal stenosis may affect a decision to use epidural anesthesia.