Extending the straight leg raise test for improved clinical evaluation of sciatica: validity and diagnostic performance with reference to the magnetic resonance imaging

Abstract

Abstract Background The straight leg raise test (SLR) is one of the most utilized and studied physical tests in patients with low back pain (LBP) for the detection of lumbar disc herniation (LDH), showing high sensitivity and heterogeneous or low specificity. The high incidence of asymptomatic ‘pathologic’ findings in the magnetic resonance imaging (MRI) scans may cause verification bias to these results. We studied an extended SLR (ESLR) by adding location-specific structural differentiation movements (hip internal rotation or ankle dorsiflexion) to the traditional SLR for it to better differentiate neural symptoms from musculoskeletal. Previously, the ESLR has shown almost perfect interrater reliability between examiners and ability to detect sciatic patients. In this study, we investigated whether a ’positive’ ESLR finding is associated with pathology seen on MRI. Methods Forty subjects comprised the study population, 20 in sciatic group and 20 in control group. The ESLR was performed ‘blinded’ to the subjects. After the ESLR, each subject’s lumbar MRI was evaluated. The MRIs were analyzed independently by 2 senior radiologists and a spine specialist clinician. The ESLR and MRI results were cross-tabulated. To obtain the odds ratio (OR) with positive ESLR or SLR results for LDH or nerve root compression (NC), a binary logistic regression analysis with subjects’ age, gender, height and weight was performed. ESLR’s validity was assessed by combination of interrater agreement and percentage prevalence of both LDH and NC. Results Of sciatic (ESLR+) patients, 85 % had LDH and 75 % NC in the MRI. Not surprisingly, MRI showed a very high incidence of ‘false-positive’ findings with the ESLR negative group. The ESLR showed 0.85 sensitivity and 0.45 specificity for LDH and 0.75 sensitivity and 0.50 specificity for NC. A positive result in the ESLR was found to be strongly associated with for both LDH and NC: the OR was 8.0 (p = 0.028) and 5.6 (p = 0.041), respectively. Conclusions The ESLR shows high validity in detecting neural symptoms and is strongly associated with pathology seen in the MRI when judged positive. We suggest the use of ESLR in clinical practice as a part of clinical examination, where it may prove to be a valuable tool in detecting patients with sciatic symptoms.