Article type
Year
Abstract
Introduction/Objective: to update a systematic review (published in 1995) for the straight leg raising test (SLR) up to 1997 by using a recent developed criteria list. Secondly, we wanted to summarize quantitatively the diagnostic accuracy of both tests and to explore reasons for its variability in the diagnosis of radiculopathy.
Methods: We extended the literature search used for the former systematic review from 1992 through 1997, using MEDLINE and EMBASE databases and reference tracking. Study selection, assessment of the methodological quality and data extraction were performed independently by two reviewers. The results of the studies were combined by the use of the random effect model. One extreme outlier was excluded for the pooling of the results. Metaregression was used to study the association of a priori defined possible confounders and modifiers with the diagnostic odds ratio (DOR).
Results: The 1995 review retained 19 studies on radiculopathy up to 1992; 12 with data on the SLR were retained for this study. Cauda equina studies were excluded. The additional search identified 4 more studies. All 16 studies were operated case-series in tertiary settings; one study had a non-operated non-diseased group of patients (the outlier). Verification bias was possible present in all studies. Measurement of the SLR was independent of the reference (surgery), but in only 2 studies in both directions. 11 studies presented complete data about test validity. Pooled sensitivity was 0.90 (95% CI 0.82-0.95), pooled specificity 0.23 (95% CI 0.15-0.35). Pooled DOR was 2.66 (95% CI 1.89-3.75; range 1.21-4.53). DOR of the outlier with verification bias was 39.2 . The DOR was not associated with the cut-point, but exclusion of previous surgery (6 studies, p 0.062) and bi-directionally blind interpretation (2 studies, p 0.113) gave a smaller estimate of the DOR. Year of publication was also negatively associated with the DOR (p 0.02). The log (DOR) was homogeneous after exclusion of the outlier (DerSimonian & Laird test of heterogeneity, p 0.2).
Discussion: Internal validity of the studies was limited because of the possibility of verification bias in all studies. The validity of the straight leg raising test was only assessed in specialised settings in a highly referred and operated population. Specificity in these patients is low. Validity of SLR should also be studied in an indicated population in primary care.
Methods: We extended the literature search used for the former systematic review from 1992 through 1997, using MEDLINE and EMBASE databases and reference tracking. Study selection, assessment of the methodological quality and data extraction were performed independently by two reviewers. The results of the studies were combined by the use of the random effect model. One extreme outlier was excluded for the pooling of the results. Metaregression was used to study the association of a priori defined possible confounders and modifiers with the diagnostic odds ratio (DOR).
Results: The 1995 review retained 19 studies on radiculopathy up to 1992; 12 with data on the SLR were retained for this study. Cauda equina studies were excluded. The additional search identified 4 more studies. All 16 studies were operated case-series in tertiary settings; one study had a non-operated non-diseased group of patients (the outlier). Verification bias was possible present in all studies. Measurement of the SLR was independent of the reference (surgery), but in only 2 studies in both directions. 11 studies presented complete data about test validity. Pooled sensitivity was 0.90 (95% CI 0.82-0.95), pooled specificity 0.23 (95% CI 0.15-0.35). Pooled DOR was 2.66 (95% CI 1.89-3.75; range 1.21-4.53). DOR of the outlier with verification bias was 39.2 . The DOR was not associated with the cut-point, but exclusion of previous surgery (6 studies, p 0.062) and bi-directionally blind interpretation (2 studies, p 0.113) gave a smaller estimate of the DOR. Year of publication was also negatively associated with the DOR (p 0.02). The log (DOR) was homogeneous after exclusion of the outlier (DerSimonian & Laird test of heterogeneity, p 0.2).
Discussion: Internal validity of the studies was limited because of the possibility of verification bias in all studies. The validity of the straight leg raising test was only assessed in specialised settings in a highly referred and operated population. Specificity in these patients is low. Validity of SLR should also be studied in an indicated population in primary care.