Standing lateral radiographic positioning does not represent customary standing balance

Study Design. Normal cohort evaluation of the accuracy of existing methods for radiographic measurement of sagittal spinal balance. Objectives. To examine the validity and reliability of sagittal vertical axis measurements during a variety of standing positions commonly used while obtaining lateral thoracolumbar spine radiographs. Summary of the Background Data. The sagittal vertical axis is a widely accepted radiographic measurement of global sagittal alignment of the spine. However, the sagittal vertical axis has not been measured in normal subjects while in functional positions because the arms must be elevated during acquisition of a lateral spinal radiograph. The purpose of this study was to quantify differences in sagittal vertical axis measurements between repeated functional positions and radiographic positions to identify the magnitude of the potential error in measuring the sagittal vertical axis radiographically. Methods. Reflective markers were attached to 15 healthy female adolescents, overlying the spinous processes of C7 and S1. Marker positions were recorded by a motion capture system during simultaneous acquisition of a lateral radiograph. Sagittal vertical axis calculation, using marker data, was matched to sagittal vertical axis measurement obtained by radiograph using anthropometric corrections to limit radiation to a single exposure. Four standing positions with varying shoulder and knee flexion as well as overground walking were examined. The mean sagittal vertical axis for each standing position and during gait was compared using a repeated measures analysis of variance. Intratrial and intertrial repeatability of sagittal vertical axis measurements was also determined. Results. The sagittal vertical axis was positive (C7 anterior to S1) for the functional positions (relaxed standing: 0.9 +/- 2.0 cm, and throughout gait: 4.5 +/- 2.0 cm), whereas shoulder flexion resulted in a negative sagittal vertical axis (-4.6 +/- 3.2) and posterior rotation of the pelvis. Adding knee flexion resulted in a slight relative shift in the sagittal vertical axis anteriorly. No differences were observed in intertrial and intratrial reliability for relaxed standing and standing with shoulder flexion alone. Increased variability was observed between repeated trials involving knee flexion. Conclusions. Measurement of the sagittal vertical axis on radiographs from commonly utilized standing positions (shoulders flexed) results in an sagittal vertical axis that is at least 3 to 4 cm more posterior than a sagittal vertical axis observed during a functional position. Subject repositioning resulted in an intertrial variability of at least 0.8 cm in sagittal vertical axis, while variation as the subject held each standing posture had little contribution to overall error of measurement. Of the analyzed positions, shoulder flexion (45degrees) alone was the best position for a lateral radiograph due to minimal compromise to repeatability of sagittal vertical axis measurement. However, none of the radiographic positions reproduced the spinal balance of the subject's functional standing posture.