Vertebral centroid measurement of lumbar lordosis compared with the Cobb technique.

Abstract

An observational study in which vertebral centroid measurement of lumbar lordosis, developed in this study, was used to examine lumbar curvature. The intra- and interobserver reliability of the vertebral centroid measurement of lumbar lordosis and the Cobb technique were compared. To evaluate the reliability of a new method of measuring lumbar lordosis and to examine the changes in the lordotic curve from 0 degree to 90 degrees flexion of the trunk. Several different methods are used to measure lumbar lordosis. The Cobb technique, based on measurement of vertebral endplates, is the method most frequently adopted for clinical diagnosis. However, because of the variations in the vertebral endplate architecture, the vertebral surface angle is difficult to identify. This reduces the reliability of the Cobb technique. Lateral radiographs of 16 study participants were taken from the upright position to a trunk flexion of 90 degrees in 30 degrees increments. The lumbar lordotic curve was measured by three observers individually using two applications of the traditional Cobb technique and the vertebral centroid measurement of lumbar lordosis. Correlation coefficients of lumbar lordosis between the two methods ranged from 0.589 to 0.772 with participants standing upright (all P < 0.05). Interobserver reliability coefficients were 0.903 for vertebral centroid measurement of lumbar lordosis, 0.826 for Cobb (L1-L5), and 0.784 for Cobb (L1-S1), although the three measurements all revealed an excellent intraobserver reproducibility (r greater than 0.9). The vertebral centroid measurement of lumbar lordosis showed the smallest mean absolute differences between any two observers' measurements (< 1.7 degrees). The findings from this study indicate that the vertebral centroid measurement of lumbar lordosis is more reliable than the Cobb method for assessing lumbar lordosis. The vertebral centroid measurement of lumbar lordosis also can be used to evaluate the actual lumbar curvature in outline at various angles of trunk flexion.