Diagnostic accuracy of the straight leg raise test for lumbosacral radicular pain: a systematic review with meta-analysis

Background Straight leg raise/crossed straight leg raise are recommended tests for diagnosing lumbosacral radicular pain. There are no recent reviews of diagnostic accuracy. Objective To summarize straight leg raise/crossed straight leg raise diagnostic accuracy for lumbosacral radicular pain in patients with low back-related leg pain. Design Systematic review with meta-analysis. Methods Diagnostic accuracy and multivariable diagnostic prediction model studies were eligible. Reference standards included imaging, electrodiagnosis, surgery, or clinical diagnosis. We searched MEDLINE and CINAHL for studies published before July 2025. Risk of bias was assessed with QUADAS-2 or PROBAST. Sensitivity, specificity, and likelihood ratios quantified diagnostic accuracy. Calibration and discrimination further quantified multivariable model performance. We performed Bayesian meta-analysis when appropriate. Certainty of evidence was assessed with GRADE. Results Twenty-four studies reported diagnostic accuracy of straight leg raise (n=7821), crossed straight leg raise (n=3811), or six multivariable models (n=1015). Pooled sensitivity and specificity for straight leg raise were typically <0.80 regardless of test interpretation or reference standard. Crossed straight leg raise was specific (>0.85) but not sensitive (<0.35). Likelihood ratios indicated both tests made small/not important changes in post-test probability. Certainty of evidence was typically low/very low due usually to risk of bias and imprecision. Multivariable models had high risk of bias. Model calibration was inconsistent but discrimination and specificity were relatively high. Conclusion Straight leg raise/crossed straight leg raise had low diagnostic accuracy for lumbosacral radicular pain. Diagnostic accuracy estimates seem unlikely to change until reference standards better capture the nerve mechanosensitivity component of this condition.