Reliability and diagnostic accuracy of clinical special tests for myelopathy in patients seen for cervical dysfunction

Abstract

Study design: Case control study.

Background: Myelopathy is a clinical diagnosis based largely on initial examination findings during a clinical screen, followed by imaging verification of cord injury or compression. At present, few studies have examined the reliability and diagnostic accuracy of clinical examination measures.

Objectives: To determine the reliability and diagnostic accuracy of neurological tests associated with the diagnosis of myelopathy.

Methods and measures: Reliability and diagnostic accuracy of 7 frequently used tests and measures and subjective findings associated with myelopathy were examined on consecutive patients with cervical pain. Interrater reliability and diagnostic accuracy values, including posttest probability, based on a pretest probability of 40%, were calculated for each test and for combinations of tests and measures.

Results: Four of the 7 diagnostic tests were found to have a substantial interrater reliability. None of the single or clusters of tests yielded low negative likelihood ratios. Of the individual tests, the Babinski sign demonstrated the highest positive likelihood ratio (LR+, 4.0; 95% CI: 1.1-16.6) and posttest probability (73%) for diagnosis, but yielded only a moderate negative likelihood ratio (LR-, 0.7; 95% CI: 0.6-0.9). Combinations of tests did not yield improved accuracy values over single test results.

Conclusion: This study demonstrated that 4 of 7 tests used to screen for myelopathy offered substantial levels of interrater agreement when used on individuals with cervical dysfunction. None of the tests when performed individually or in combinations are effective for screening; however, the Babinski sign did alter posttest probability more significantly than combinations of test findings.

Level of evidence: Diagnosis, Level 2b.