Objectives: The aim of this study was to compare flexicurve surface contour measurements of the cervical spine with radiographic measurements of cervical lordosis.
Methods: One examiner evaluated 96 patients with chronic neck pain in neutral posture using a flexible ruler, flexicurve, to measure sagittal contour of the skin over the cervical spine from the external occipital protuberance to the vertebra prominens. The flexicurve skin contour and neutral lateral radiographs were digitized and compared. The flexicurve and radiographs were categorized into height-length ratio, curve angle, curve depth, sum of depths, modified Ishihara's index, and inverse of radius. Mean values, SDs, mean differences, and limits of agreement were calculated. The differences between flexicurve measurement mean values and x-ray mean values were deemed significant if the lower limit of agreement exceeded 15% of the mean values for the x-ray measurements.
Results: For all variables, except the height-length ratio, the mean values of the flexicurve variables differed significantly from the corresponding mean values of the radiographic measurements. All Pearson correlation coefficients were in the very poor range (r < 0.15).
Conclusion: The flexicurve sagittal skin contour measurement has poor concurrent validity compared with established radiographic measurements of the cervical lordosis. The flexicurve tracings always predicted lordosis, overestimated the lordosis compared with x-ray values, and cannot discriminate between radiographic lordosis, straightened, S curves, and kyphotic alignments of the cervical curve.