Background:

The common manual measurement technique of spinal sagittal alignment on X-rays is susceptible to rater-dependent variability, which has not been adequately considered in previous publications. This study investigates the effect of those variations in the characterization of patients receiving lumbar spondylodesis.

Methods:

General alignment parameters on pre- and postoperative X-rays were evaluated by four raters in 43 prospectively sampled patients undergoing monolevel spondylodesis. The Intra-class Correlation Coefficient (ICC) for each rater pair and all raters together was calculated for inter-rater reliability. For the operation-induced change of the sagittal alignment in every patient the Wilcoxon test was applied to compare for each rater separately.

Results:

The ICCs were “good” (>0.75) to “excellent” (>0.9) for all raters together and for 45 of the 48 single rater pairs (93.75%). All revealed a significant increase of the addressed segmental lordosis and disc height and no significant change for spinopelvic parameters and sagittal vertical axis from pre- to postoperative. The lumbar lordosis showed a significant increase through the operation of +2.5° (p = 0.014) and +3.7° (p = 0.015) in two raters and no difference for the other ones (+2.1°, p = 0.171; -2.2°, p = 0.522).

Conclusions:

The pre- to postoperative change of lumbar lordosis revealed different significance levels for different raters, although the ICCs were formally good. Accordingly, the evaluation by only one rater would lead to different conclusions. Due to this susceptibility of alignment measurements to rater-dependent variability, the exact evaluation process should be described in every publication and the consistency of significant results be validated through multiple raters.

Trials registration:

The trial was approved by the local ethics committee and listed at the national clinical trials register ( DRKS00004514 , date of registration: 08/11/2012).