Reliability of computed tomography measurements of paraspinal muscle cross-sectional area and density in patients with chronic low back pain

Study Design. A reliability study was conducted. Objective. To estimate measurement errors related to equipment and the observer in computed tomography measurements of cross-sectional area and density of paraspinal muscles. Interobserver reliability was not investigated in the current study. Summary of Background Data. Computer tomography (CT) had been used to measure the cross-sectional area and degeneration of the back muscles in patients with low back pain. Methods. This study included 31 patients, mean age 47 years, with chronic low back pain. The measurements comprised cross-sectional area (cm(2)) and density (Hounsfield units [HU]) of the paraspinal muscles at Th12-L1, L3-L4, and L4-L5. To measure the reliability of the equipment and the observer (total reliability), two independent CT scans were performed for each patient . The radiologist traced the cross-sectional area twice within 2 weeks for measurement of the intraobserver reliability. Results. There were no significant differences in the assessments between the first and second CT scans, or between the radiologist's two measurements of the identical slices. The critical difference for the total reliability ranged from 11.3 to 22.8 for the density and from 10.0 to 16.0 for the cross-sectional area. For the cross-sectional area, the measurement error associated with the observer was higher than for the equipment. For the density, the measurement error related to the equipment was higher. The main measurement error was associated with the radiologist for the cross-sectional area and with the CT scanner for the density. Conclusions. The reliability of the CT scan for measuring the cross-sectional area and density of the back muscles is acceptable. The authors do not know definitely whether their results can be generalized because the interobserver and intermachine reliabilities were not investigated.