Validity of stiffness estimates of individual lumbar motion segments obtained in the intact spine

J.H. van Dieën Degeneration, injury and surgical interventions may alter the mechanical properties of spinal motion segments, but the quantification of these alterations is problematic. Manual or instrumented loading of single segments in the intact spine as applied intra-operatively may overestimate the mechanical properties of this segment, because the applied load is partly sustained by the adjacent segments. The purpose of this study was to investigate the measurement errors that can occur from single segment stiffness tests in intact spines. Therefore, the distribution of stiffness values of individual spinal segments within and across human cadaveric spines was determined and these data were used as input to a model simulation of segment stiffness tests in intact spines. Seventeen human cadaveric lumbar spines were loaded with pure moments in all three directions. An optical system was used to measure the angular rotations of each motion segment and load-displacement curves were used to determine stiffness. With the distribution of measured stiffness data as input, a stochastic mechanical model was constructed, to investigate how the stiffness of adjacent segments influences stiffness estimates obtained by loading a single segment in the intact spine. The results showed that the variance in stiffness values was high for all directions, but covaried between segments within a spine. Model simulations indicated that stiffness estimates obtained by loading a single segment in an intact spine are highly correlated with actual stiffness, but overestimate stiffness by a median of 18% with peak errors of over 300%. In conclusion, current measurement devices and manual assessment substantially overestimate segmental stiffness, due to the effect of adjacent spinal levels. In addition, the variance in stiffness within spines can occasionally cause large errors, which might lead to erroneous surgical decisions.