Lumbar Spinal Stenosis: How Is It Classified?

Gregory D. Schroeder, MD; Mark F. Kurd, MD; Alexander R. Vaccaro, MD, PhD, MBA


J Am Acad Orthop Surg. 2016;24(12):843-852. 

In This Article

Key Variables in a Spinal Stenosis Classification

Because spinal stenosis is a heterogeneous diagnosis, multiple variables will be required in a new classification. Ishimoto et al[3] reviewed the MRI scans of 938 persons aged ≥40 years in Japan, and 77.9% of participants had more than moderate central stenosis (ie, defined as a decrease between one third and two thirds in the cross-section area of the spinal canal), and 30.4% had severe central stenosis (ie, a decrease of more than two thirds of the cross-section area of the spinal canal); however, even in the cohort with severe central stenosis, only 17.5% of patients were symptomatic. In a smaller study, Boden et al[33] reported asymptomatic spinal stenosis in 21% of Americans aged >60 years who had substantial abnormality on MRI. Because of the high prevalence of asymptomatic patients with radiographic stenosis, a patient's symptomatology must be the cornerstone of a lumbar stenosis classification system.

The history provides the most critical information regarding a patient's symptoms. Patients with central LSS typically have neurogenic claudication,[6] which is pain in the low back, buttocks, and/or posterior thighs that is worse with standing (lumbar extension) and improves with sitting or leaning forward (lumbar flexion). In addition, walking often exacerbates the pain; however, unlike with vascular claudication, the pain is often proximal, and it is not relieved when standing still. Furthermore, activities performed with the lumbar spine in a flexed position, such as bicycling or walking while pushing a shopping cart, often do not exacerbate the symptoms. Importantly, the symptomatology may vary according to the location of the stenosis, with central stenosis predominantly resulting in isolated neurogenic claudication, whereas lateral stenosis may also cause radicular symptoms (eg, dermatomal numbness, weakness). Although classic physical examination findings, such as worsening pain with lumbar extension, new-onset neurologic symptoms with walking, blunted deep tendon reflexes, and a wide-based gait with a flexed lumbar spine, are helpful in clinical evaluation, their absence does not preclude the diagnosis.

After a systematic review of the literature, the North American Spine Society reported that there is insufficient evidence to correlate certain physical examination findings with spinal stenosis;[1] therefore, correlation with the patient history is more important in the diagnosis. The use of patient-reported outcome metrics is increasing, and the incorporation of one of these metrics into a classification system would be helpful. Problematically, many of the existing metrics, such as the SSS Questionnaire and the FLS-25, are complex and thus may not be practical in standard practice. Alternatively, less precise metrics, such as the ODI, or functional outcomes, such as walking distance, may be more useful.

The next important variable to consider is the risk of instability because the treatment algorithms are substantially different for patients with and without this condition (Figure 5). Of the 380 patients with spinal stenosis and degenerative spondylolisthesis treated surgically in the SPORT study, all but 23 (96%) were treated with decompression and fusion;[34] in comparison, none of the surgically treated patients with spinal stenosis without spondylolisthesis were treated with fusion.

Figure 5.

Proposed algorithm for the treatment of a patient with spinal stenosis.

Although no classification is currently available for spinal stenosis, two systems are commonly used for spondylolisthesis: the classification proposed by Wiltse[35] categorizes the condition according to etiology, and the classification proposed by Meyerding[36] is based on the amount of translation of the cephalad vertebra relative to the caudal vertebra. However, these classifications do not provide meaningful distinctions for patients with spinal stenosis and degenerative spondylolisthesis. In 2015, Kepler et al[37] proposed the clinical and radiographic degenerative spondylolisthesis classification, which categorizes patients by the presence of disk collapse, instability, focal kyphosis, and symptoms. This classification allows meaningful separation of patients with degenerative spondylolisthesis; however, because the classification is based largely on radiographs, the location or severity of spinal stenosis is not formally considered.

A third important variable is the site of nerve compression. Because of the lack of an accepted classification system, all patients with any spinal stenosis at any location are grouped together in the literature. Therefore, although Weinstein et al[6] reported that health-related quality of life outcomes improved considerably with surgical treatment of spinal stenosis compared with nonsurgical care, the location of the stenosis may have notably affected the expected results of treatment.

Another important variable for inclusion in a meaningful classification of spinal stenosis is objective and reproducible radiographic parameters that define the presence and severity of neural compression. In a recent systematic review, 63 studies designed to identify the best treatment of LSS were analyzed, and only four of them (6%)[23] used an objective standardized metric to establish the diagnosis. Depending on the amount of nerve compression, some treatments, particularly nonsurgical options, are more efficacious than others; however, with no accepted objective grading system for spinal stenosis, it is difficult to counsel patients on the expected results of these treatments.

Lastly, patient-specific modifiers, such as substantial medical comorbidities, must be included in the classification. Given the advanced age of many patients with spinal stenosis, routine comorbidities (eg, hypertension, hyperlipidemia, diabetes mellitus) increase the risk of postoperative complications, are unlikely to alter the treatment algorithm. However, patients who are taking steroids long term may not be candidates for further oral steroid medications, and patients with significant heart failure or liver failure may not be appropriate candidates for surgical intervention (Figure 5).