Research Review By Dr. Michael Haneline©


Download MP3

Date Posted:

November 2011

Study Title:

Diagnostic accuracy of the clinical examination in identifying the level of herniation in patients with sciatica


Hancock M, Koes B, Ostelo R et al.

Author's Affiliations:

Faculty of Health Sciences, University of Sydney, Australia; Department of General Practice, Erasmus Medical Center, Rotterdam, Netherlands.

Publication Information:

Spine 2011; 36(11): E712-719.

Background Information:

Patients with sciatica complaints are examined by means of a variety of test procedures to determine if the symptoms are caused by a disc herniation as well as which level is involved. The general tests (e.g., Straight Leg Raise [SLR] and Valsalva’s Maneuver) provide information about whether a disc herniation is present, but they do not provide information about which disc level is involved. Tests that do provide disc-level specific information include dermatome, muscle strength and deep tendon reflex testing.

When the clinical picture points to a disc herniation, patients are commonly evaluated by magnetic resonance imaging (MRI) to confirm the presence and level of involvement. However, if the level of herniation that is seen on MRI does not correlate with the clinical examination findings, the patient may not be expected to improve following a disc-specific intervention (e.g., discectomy).

Even though it is extremely important to have a good correlation between clinical examination findings and MRI findings before proceeding to surgery, little is known about the accuracy of commonly used tests in predicting the level of disc herniation.

The objective of this this study was (in patients with sciatica and confirmed lumbar disc herniation) to investigate the diagnostic accuracy of:
  • individual neurologic tests,
  • combinations of tests, and
  • a clinician’s overall impression in identifying the specific level of a suspected disc herniation.
The study’s secondary objective was to determine if the diagnostic accuracy was influenced by whether the herniation was central or lateral.

Pertinent Results:

  • Patients (n = 283) were on average 42 years-of-age and 66% were males.
  • The mean sciatica pain level was 65 on a 100 point visual analogue scale (VAS).
  • Symptom duration was an average of 9 weeks.
  • The most common level of disc herniation was L5/S1, which occurred in 59.6% of the patients.
  • The L4/L5 level was next most common, affecting 44.6% of the patients.
  • Only 4.6% had L3/L4 involvement.
  • Overall, there was a great deal of variation in the frequency of positive findings between the tests.
  • There was a very high correlation between side of herniation and the side of sciatica with only five patients having pain on the opposite side.
Dermatomal pain location was the most accurate of the neurological tests in identifying the level of disc herniation. However, none of the tests were considered to be very accurate, with all of the respective AUC values being below 0.75. (Area under the curve – an indication of a test’s ability to discriminate, that is, to correctly classify those who have a particular condition from those who do not have the condition. A perfect test would have an AUC of 1, whereas a worthless test that could not discriminate at all would have an AUC of 0.5. Generally, a test that has excellent accuracy would be 0.90 or above, good 0.80-0.90, fair 0.70-0.80 and poor 0.60-0.70.) Patient reports of pain in the L5 and S1 dermatomes were the best tests to identify L4/L5 and L5/S1 herniations, respectively.

Examination by a neurologist was more accurate at identifying the level of disc herniation than any of the individual tests, with the associated AUC being 0.79 at L4/L5 and 0.80 L5/S1 herniations.

When several index tests were combined, diagnostic accuracy was slightly superior to dermatomal pain location, the most accurate index test, but was inferior to the rating of a neurologist.

The specificity for identifying disc herniation was high (L3/L4 = 90, L4/L5 = 83, and L5/S1 = 94) when at least 3 tests were positive, although the sensitivity was poor.

The accuracy of the neurologists in identifying disc herniations was influenced by the location of the disc herniation at the L5/S1 level but not at L4/L5. At L5/S1 level the neurologists’ findings were accurate in differentiating those with and without central herniations (AUC = 0.81), but not very accurate for lateral herniations (AUC = 0.45).

Clinical Application & Conclusions:

The authors concluded that their study did not provide evidence to support the accuracy of any of the individual neurological examination tests in identifying the level of disc herniation in the lower three discs that are demonstrated on MRI. On the other hand, a neurologist utilizing a group of tests was moderately accurate in identifying the level of disc herniation.

When the findings of several positive index tests were combined, accuracy increased slightly.

Because of the imprecision of tests from a neurological examination in determining the correct level of herniation, they cannot be used to directly inform the decision to perform surgery. The most valid information in guiding the level of surgery and to support the decision for surgery would at present be derived from the level of herniation as identified on MRI.

Dermatomal pain location was the most accurate single clinical test in identifying the level of disc herniation, which is similar to what was found by Kortelainen et al. (1). Thus, clinicians should pay particular attention to this finding when present in patients with suspected lumbar disc herniation.

The authors provided an explanation as to why a clinical examination by a neurologist was the most accurate way to identify the level of disc herniation. They suggested that the findings of several tests could be interpreted by a neurologist and then the importance of each individual test could be weighed against each other. For instance, tests with highly positive results could be relied on more heavily than other tests. As evidence-informed manual medicine providers, we should keep the relative importance of each clinical finding in mind in combination with all other historical and physical factors.

Study Methods:

This was an analysis of previously collected data from a randomized controlled trial that compared surgery with prolonged conservative care in sciatica patients. A total of 283 subjects were included, out of 599 who were assessed for eligibility.

Patients were included if they:
  • had a diagnosis of lumbosacral radicular syndrome from a neurologist,
  • had their pain from 6 to 12 weeks,
  • were 18 to 65 years of age, and
  • had a radiologically confirmed disc herniation.
Patients were excluded if they had:
  • cauda equina syndrome,
  • severe paresis,
  • a similar episode of symptoms during the last 12 months,
  • previous spine surgery,
  • bony stenosis,
  • spondylolisthesis,
  • pregnancy, or
  • severe coexisting disease.
Patients underwent two neurologic examinations, one before the MRI (that was conducted by neurologists) and the other one soon after the MRI (that was performed by trained research nurses). The neurologists and nurses were all blinded to the MRI results, although they did know that the patients had a disc herniation, they did not know which level was involved.

Each MRI was interpreted by both a neurosurgeon and a radiologist who were blinded to the results of the neurological examinations. The MRIs were rated for the presence or absence of disc herniation at L3/L4, L4/L5, and L5/S1, as well as whether the herniations were central, subarticular, intraforaminal or extraforaminal.

There were 68 participating neurologists who performed whatever tests they thought were appropriate in order to determine if the patients’ sciatica was caused by a disc herniation. They were also asked to specify which level and which side the herniations occurred on.

There were also six nurses who were trained by a neurologist and neurosurgeon to perform a neurological examination and afterward were required to pass tests similar to what neurology residents would take. The examinations that they administered included tests of sensation, muscle strength and deep tendon reflexes for lumbar dermatomes and myotomes. Sensation was rated as either normal or abnormal. Muscle strength was graded on a 0 to 5 scale with 5 considered normal and 4 or less abnormal. Reflexes were rated as normal, absent, or less than the other side, with special consideration when patients had bilaterally absent reflexes.

The accuracy of each index test in predicting herniations was analyzed by calculating the area under the curve (AUC) as well as sensitivity and specificity.

AUC analysis refers to the ability of a test to discriminate, that is, to correctly classify those who have a particular condition from those who do not have the condition. A perfect test would have an AUC of 1, whereas a worthless test that could not discriminate at all would have an AUC of 0.5. Generally, a test that has excellent accuracy would be 0.90 or above, good 0.80-0.90, fair 0.70-0.80 and poor 0.60-0.70 (2).

Study Strengths / Weaknesses:

All of the patients in this study had a MRI confirmed disc herniation; therefore, the results do not provide information as to the accuracy of neurological tests in identifying the presence of a herniation. This is an important consideration that readers should keep in mind.

The examiners were aware that all of the patients had disc herniation, but they were not aware of which level was involved. Their goal was to identify that level. The study would have been stronger, however, if some symptomatic patients who did not have disc herniation were included because it would have provided information about how well the tests discriminate between patients with and without disc herniation.

Since patients with muscle strength that was less than Grade 3 were excluded, there were fewer patients with more severe symptoms than what would be seen in the general population. Patients with more severe symptoms would most likely have had more obvious findings on the index tests and therefore the diagnostic accuracy in these more severe cases may have been better than what was found in the study.

It is possible that some patients had clinically silent disc herniations which may have affected the accuracy of the neurologic examination. The examiners might have been led astray if these disc levels were not responsible for the patients’ symptoms.

The poor accuracy of the clinical tests that were performed compared with the neurologist’s findings in this study may be related to the fact that the clinical tests were done by research nurses who may not have been as adept at performing the tests as a neurologist.

Additional References:

  1. Kortelainen P, Puranen J, Koivisto E, et al. Symptoms and signs of sciatica and their relation to the localization of the lumbar disc herniation. Spine 1985;10:88–92.
  2. Metz CE: Basic principles of ROC analysis. Semin Nucl Med 1978, 8:283-298.