Am Fam Physician. 2009;80(10):1145-1147
Author disclosure: Nothing to disclose.
Clinical Question
What is the best way to diagnose lumbar spinal stenosis in patients with leg pain or numbness?
Evidence Summary
Lumbar spinal stenosis is an important cause of pain and disability, and surgery is beneficial for appropriately selected patients.1,2 Therefore, it is important for primary care physicians to distinguish patients with spinal stenosis from those with musculoskeletal low back pain, peripheral vascular disease, or spinal disk disease.
Individual signs and symptoms suggestive of lumbar spinal stenosis include older age at onset, longer duration of symptoms, symptoms that worsen with walking or standing, numbness of the lower legs with activity, symptoms that improve with bending forward, and symptoms that worsen with bending backward. The findings that most strongly suggest lumbar spinal stenosis are symptoms that improve with bending forward, urinary disturbance, and intermittent claudication.
Table 1 includes accuracy data for individual signs and symptoms.3 These data are derived from the best study to date of the clinical diagnosis of lumbar spinal stenosis.3 The study included patients presenting to an orthopedic surgeon with a primary complaint of pain or numbness in the legs. All patients had plain radiography and magnetic resonance imaging (MRI) of the lumbar spine, as well as a standardized history and physical examination. The reference standard was diagnosis of lumbar spinal stenosis by the referring orthopedic surgeon and the study coordinator; a consensus panel established the final diagnosis when the surgeon and coordinator disagreed (this occurred with 243 patients). Of the 468 patients in the study, 222 patients received a final diagnosis of spinal stenosis. This percentage is higher than in a typical primary care population, suggesting that their prediction tools may overestimate the risk of spinal stenosis.
| Patient-reported signs and symptoms | Sensitivity (%) | Specificity (%) | LR+ | LR– |
|---|---|---|---|---|
| History of urinary symptoms | 14 | 98 | 7 | 0.14 |
| Symptoms improve when bending forward | 52 | 92 | 6.5 | 0.52 |
| Intermittent claudication | 82 | 78 | 3.7 | 0.23 |
| Symptoms worsen when standing up | 68 | 70 | 2.3 | 0.46 |
| Bilateral plantar numbness | 27 | 87 | 2.1 | 0.84 |
| Symptoms induced when bending backward | 70 | 55 | 1.6 | 0.55 |
The authors of the study developed several clinical prediction tools based on this data set. First, the authors created an integer-based scoring system using 10 history and physical examination findings.3 This tool has good accuracy, with a likelihood ratio of 3.3 for a positive test result and 0.1 for a negative test result. However, it has not been prospectively validated. A second rule (Table 2), which does not include physical examination findings, was developed using 80 percent of the data set and validated using the remaining 20 percent.4 Finally, the researchers created a self-administered, 10-item patient survey (Figure 1) to identify patients with lumbar spinal stenosis and then distinguish between those with radicular lumbar spinal stenosis and those with cauda equina syndrome.5
| Findings | Points | |
|---|---|---|
| Age | ||
| < 60 years | 0 | |
| 60 to 70 years | 2 | |
| > 70 years | 3 | |
| Onset of symptoms occurred more than six months ago | 1 | |
| Symptoms improve when bending forward | 2 | |
| Symptoms improve when bending backward | –2 | |
| Symptoms worsen when standing up | 2 | |
| Intermittent claudication present | 1 | |
| Urinary incontinence present | 1 | |
| Total: | ——— | |
| Score | Probability of lumbar spinal stenosis* | |
| ≤ 2 | 11/66 (16.7%) | |
| 3 or 4 | 35/120 (29.2%) | |
| 5 or 6 | 78/151 (51.7%) | |
| ≥ 7 | 98/131 (74.8%) | |
A recent systematic review evaluated the accuracy of diagnostic tests for lumbar spinal stenosis.6 Fifteen studies of imaging for the diagnosis of the condition were identified. Although most of the studies were of poor quality and the accuracy of the tests varied considerably between studies, the authors of the review concluded that myelography, computed tomography, and MRI appear to have similar accuracy. Evidence from two studies showed that three-dimensional magnetic resonance myelography may be somewhat more sensitive than other tests, but it is more expensive.6
It is important for physicians to consider cauda equina syndrome in the differential diagnosis of back pain and numbness. Any patient with signs or symptoms of possible cauda equina syndrome (e.g., saddle anesthesia, bowel or bladder symptoms) should receive emergent referral to a neurosurgeon.
Applying the Evidence
A 64-year-old man presents with leg pain that has gradually worsened since its onset eight months ago. The pain worsens when he walks or stands up, improves when he bends forward, and does not change when he bends backward. He denies having urinary incontinence. What is the patient's risk of lumbar spinal stenosis?
Answer: Using the clinical decision rule in Table 2,4 the patient receives a score of 8 points (two points for age, one for time of onset, two for improved pain with bending forward, two for worsening pain with standing, and one for worsening pain with walking [claudication]). This puts him in the highest risk category for lumbar spinal stenosis (75 percent probability). After ruling out peripheral vascular disease by confirming that his ankle brachial indices are normal, you order an MRI to confirm lumbar spinal stenosis.