Introduction The early clinical results of cervical disc replacement surgery are encouraging but the in vivo kinematics of prostheses remains poorly understood. Two recent published reports suggest that use of a prosthesis with an unconstrained (over normal range of motion) biconvex nucleus (Bryan Cervical Disc® – Medtronic Sofamor Danek, Memphis, TN) can be associated with post-operative segmental kyphosis.

This study examines post-operative kyphosis and segmental imbalance following cervical disc replacement using the Bryan Cervical Disc prosthesis and factors which may influence this. In particular, the influence of change in disc space height as a result of surgery was studied.

Methods 67 patients underwent prosthetic disc replacement by one of three surgeons (19, 25 and 23 patients, respectively) using the Bryan prosthesis. 46 single, 20 double and 1 triple level were operated. Neutral pre- and post-op erect and intra-operative x-rays were examined manually and using digital image analysis software (Medical Metrics, Inc. Houston, TX). Possible contributing factors to segmental alignment were studied including: pre-op alignment, angle of prosthesis insertion, disc space degeneration and sacrifice of the posterior longitudinal ligament (PLL). Particular attention was given to changes in disc space height and factors which may influence this.

Inter- and intra-observer agreement was assessed. Non-parametric tests were used for assessment of categorical and skewed continuous variables. Multivariate linear regression was used to adjust significant correlation coefficients. Significance was set at p< 0.05.

Results The median pre-op focal lordosis of +0.5° (range: 21 to −14°, −ve = kyphotic) changed by −1° (+14 to −17°), to post-op: 0° (+11.5 to −16°).

There was a significant difference in the median change in focal lordosis for surgeon 1 (−3°) vs. surgeons 2 & 3 (−1°) (p< 0.005) and in the loss of disc space height. Median loss of disc space height for surgeon 1 was 22% vs. 8% for surgeons 2 & 3 (p< 0.002). Correlation co-efficient (Spearman) for change in disc space height vs. change in disc space angulation was 0.67 (p< 0.0001). No single pre- or intra-operative factor was found to clearly correlate with subsequent loss of disc space height apart from a trend towards a weak correlation with the angle of prosthesis insertion (r=0.24, p=0.06).

Discussion The median change (loss) in focal lordosis was −1.5° but there was considerable range: from +14° to −17°. Attempts to identify contributing factors suggest that a number may be involved but there did appear to be a highly significant correlation between loss of disc space height following surgery and subsequent focal kyphosis.

While the difference in outcomes between Surgeon 1 and Surgeons 2 & 3 is probably not clinically significant, it does suggest that intra-operative factors such as the angle of prosthesis insertion may be important. We are continuing to study these factors.