Research Review by Dr. Shawn Thistle©


Nov. 2006

Study Title:

Spinal stenosis, back pain, or no symptoms at all? A masked study comparing radiologic diagnoses to the clinical impression


Haig AJ et al.

Publication Information:

Archives of Physical Medicine & Rehabilitation 2006; 87: 897-903.


Lumbar spinal stenosis can cause neurological deficit, pain, and functional limitation. It is more prevalent in the elderly, and will certainly become more common as the population ages. Some confusion exists in the use of the term stenosis.

Some clinicians use it to refer to measurable deviation from the average size of the spinal canal or neural foramen regardless of symptoms, while others use it to refer to the clinical syndrome that presents classically as neurogenic claudication (that is, pain in the back or legs with ambulation).

Anatomic measures of canal size can be easily made from standard imaging modalities such as plain film radiograph or MRI. However, there is no criterion standard for the diagnosis of spinal stenosis, and no clear relationship exists between clinical symptoms and extent of stenosis on imaging.

Further, no cutoff point for canal size to diagnose this syndrome has been widely agreed upon in the literature.

To make matters more confusing, classic signs and symptoms are not always evident on examination, and similar symptoms can occur in a variety of other conditions such as peripheral vascular disease, polyneuropathy, or even mechanical low back pain.

This study aimed to assess the relationship between imaging findings and clinical diagnosis of spinal stenosis. Utilizing a masked, double-controlled design, patients with or without back pain and with or without MRI-demonstrated stenosis were evaluated by a clinician, radiologist, and electrodiagnostician to determine the extent of agreement in their diagnoses.

126 patients (age 55-80) were included in this study, and were grouped as follows:
  1. stenosis group (n=50) – had a preliminary diagnosis of stenosis (from a previous MRI) regardless of symptoms
  2. back pain group (n=44) – had a preliminary diagnosis of non-radiating back pain with no stenosis (patients were excluded if symptoms were present distal to the knee)
  3. control group – no back pain and no exclusion criteria
Patients were excluded from this study if they had known polyneuropathy, diabetes, heavy alcohol use, previous lumbar surgery (or plans for surgery), or relative contraindications to MRI or electrodiagnostic testing.

Each patient underwent a standard lumbar spine MRI (if they hadn’t already), and standard electrodiagnostic testing (a nerve conduction test). Further, they all completed the Pain Disability Index, Quebec Back Pain Disability Scale, McGill Pain Questionnaire, VAS scale, and pain drawing. Each patient also performed a 15 minute ambulation test, and underwent a comprehensive and codified, but unconstrained physical examination by a physiatrist to arrive at a final “clinical impression”.

These results were then compared with reports from a radiologist and electrodiagnositician for statistical examination of agreement. All examiners were masked to the other aspects of the patients’ diagnosis or imaging findings. It should be noted that the clinician’s diagnosis was not held as the criterion standard in this study, it was merely used for the purpose of determining agreement.

Pertinent Results:

  • radiologist sensitivity for the clinical diagnosis of lumbar stenosis was 59.2% and specificity ranged from 40.9% in the back pain group and 43.8% in the asymptomatic group
  • after combining the back pain and control groups for analysis – there was no significant relation between radiologic and clinical diagnosis of stenosis
  • the sensitivity of the electrodiagnostician was 70.0% and specificity ranged from 46.9% in the back pain group to 47.4% in the control group
  • for the combined control groups (as above) – the electrodiagnostic impression was significantly related with clinical stenosis
  • using a regression analysis, neither MRI no electrodiagnostic testing could predict the clinician’s diagnosis of spinal stenosis

Conclusions & Practical Application:

We have all had patients present to our offices with imaging report in hand stating they have stenosis (or insert disc herniation or bulge, “arthritis”, “degeneration” etc. here). Sometimes we are all guilty of allowing this information to influence our clinical decision making and treatment strategy, regardless of the actual findings or history. This study suggests that we should carefully examine the patient before looking at, or considering the results of, any imaging studies.

The authors provide some interesting insight in the discussion when they say: “Spine specialists, and perhaps more importantly their referral sources, should avoid the temptation to treat the photograph instead of the patient.” [pg. 902] This is a statement that should be at the forefront of our thinking with these patients to say the least.

In conclusion, this study suggests that simple nerve conduction tests may be more useful in diagnosing the clinical syndrome of stenosis, while MRI does not help differentiate patients with clinical spinal stenosis from those with mechanical back pain, or no symptoms at all. Electrodiagnostic testing is generally less expensive and space-consuming compared with MRI, and seems effective in picking up conditions such as peripheral nerve entrapments and polyneuropathies that can mimic spinal stenosis.

That being said, similar to diagnostic ultrasound which is becoming quite popular for diagnosing soft tissue injuries, the consultation is skill-dependent, somewhat subjective, and no meaningful record of the test is generated (as with MRI/x-ray).