Test-retest repeatability of lumbar sagittal alignment and disc height measurements with or without axial loading: a computed tomography study.

STUDY
DESIGN: In vivo study using young healthy volunteers, CT scan, and an axial compression device.
OBJECTIVE: This study was conducted to evaluate the test-retest repeatability of the measurements of sagittal alignment and disc height of the lumbar spine with and without an axial compression device in the supine position. SUMMARY OF BACKGROUND DATA: Dynamic radiologic examinations have been used to investigate the relationships between disc degeneration and lumbar kinematics or disc height changes under different loading conditions in vivo. There have been a number of investigations of axial loading changes in the lumbar spine during CT and MRI using axial compression devices. Axial compression devices are widely used for simulation of standing position during magnetic resonance imaging (MRI) or computed tomography (CT) scans of the lumbar spine. However, the test-retest repeatability of lumbar morphological changes using axial loading devices is not well understood.
METHODS: The study population consisted of 14 asymptomatic healthy young volunteers (7 men, 7 women: age 21 to 32, mean 27 y). Lumbar CT scan with axial loading using a DynaWell compression device (axial loading condition), and CT scan without loading (lying down condition) were carried out. Each participants was evaluated on 2 occasions, 1 month apart at about the same time of the day. Lumbar spinal length, disc height, disc angles, and total lumbar angle were measured by a single observer. Test-retest repeatability was assessed using the intraclass correlation coefficients (ICC). The dependability coefficient ranged between 0 and 1, in which 0 implies null repeatability and 1 implies perfect repeatability. A value of ~0.75 indicates good repeatability, a value between 0.50 and 0.75 indicates moderate repeatability, and values under 0.5 indicate poor repeatability. Test-retest repeatability (intraclass correlation coefficients: ICC) in spinal length, lumbar disc height, disc angle, total lumbar angle, and pelvic angle in both conditions were assessed between the first and second examinations. To evaluate the effects of the device, these parameters under axial loading and lying down conditions were compared statistically using the paired t test.
RESULTS: Although spinal length was significantly decreased with axial loading, test-retest ICC of spinal length under lying down and axial loading conditions was ≥0.995, suggesting good repeatability. Although the average disc height showed a significant decrease at L5/S, test-retest ICC of disc heights under lying down and axial loading conditions was ≥0.739, suggesting moderate to good repeatability. Although disc angles at L2/3 and 3/4 showed a significant increase and disc angle at L5/S and pelvic angle showed significant decreases in axial loading, test-retest ICC of these angle parameters were ≥0.877, suggesting good repeatability. The differences in these parameters between lying down and axial loading conditions showed the same tendencies in the first and second examinations.
CONCLUSIONS: Spinal length was significantly decreased under conditions of axial loading. Segmental lordotic angle at L2/3 and L3/4 was significantly increased under axial loading conditions. However, disc lordotic angle at L5/S and the pelvic angle were significantly decreased under conditions of axial loading. Axial loading CT of the lumbar spine, and CT without axial loading, provided reproducible measurements of lumbar spinal anthropometric parameters within a 1-month test-retest interval. Axial compression devices are potentially reliable to examine lumbar spinal alignment changes.