A systematic review of the literature including statistical meta-analysis.
To evaluate published methods of the test of Lasègue or straight leg raising test and the cross straight leg raising test by using a recently developed criteria list and to summarize and explore reasons for variation in diagnostic accuracy.
Summary of Background Data.
Little evidence exists on the diagnostic accuracy of the widely used straight leg raising test and the cross straight leg raising test in diagnosing herniated discs in patients with low back pain.
MEDLINE and EMBASE searches up to 1997 showed 17 diagnostic publications evaluating the straight leg raising test with surgery as reference standard. Quality of methods was assessed with a specific checklist. Eleven studies were selected for statistical pooling. Sources of variation and heterogeneity were studied by meta-regression of the diagnostic odds ratio.
All studies were surgical case-series at nonprimary care level. Verification-bias was obvious in one study. Pooled sensitivity for straight leg raising test was 0.91 (95% CI 0.82–0.94), pooled specificity 0.26 (95% CI 0.16–0.38). Pooled diagnostic odds ratio was 3.74 (95% CI 1.2–11.4). Discriminative power was lower in recent studies, in studies with only inclusion of primary hernias, and with blind assessment of both the index-test (straight leg raising test) and the reference (surgery). For the cross straight leg raising test pooled sensitivity was 0.29 (95% CI 0.24–0.34), pooled specificity was 0.88 (95% CI 0.86–0.90), and the pooled diagnostic odds ratio 4.39 (95% CI 0.74–25.9).
The diagnostic accuracy of the straight leg raising test is limited by its low specificity. Discriminative power decreased with a more valid design, a more homogenous case-mix, and year of publication. Although the studies may reflect everyday clinical practice, they do not enable a valid evaluation of the diagnostic accuracy of both tests. Diagnostic research should evaluate the validity of the complete diagnostic process and study the evidence of the added value of the different tests used.