Lumbar Intervertebral Joint Loads as Determined by Patient Specific Musculoskeletal Modeling for Two Types of Sagittal Spino-Pelvic Alignment with Different Occurrence of Symptomatic Adjacent Segment Degeneration

Abstract

Introduction Biomechanical consequences of fusion leading to adjacent segment degeneration (ASD) have been investigated in different studies, and adverse consequences of sagittal alignment on intervertebral disk (IVD) loads have been proposed as a potential cause. This study investigates IVD loading by means of patient-specific musculoskeletal modeling of potential anatomical factors that may predispose to ASD as recently identified in a retrospective clinical study.1 The current study thus attempts to quantify the reaction forces in the prospective adjacent IVDs as a consequence of spino-pelvic alignment in nonfused spines. Materials and Methods A publicly available model for the human lumbar spine2 was employed and slightly modified for this study. Most notable alterations were adjustments to muscle attachment points, scaling, and changes in the vertebral and sacral reference systems. Joint centers were maintained and length-dependent muscle properties were scaled to maintain the original muscle characteristics. For each patient from the clinical study,1 important landmark coordinates (positions of endplates T12-S1 and acetabulum) were extracted from a digital lumbar spine sagittal radiograph and models were created. Three subjects were excluded due to insufficiently clear radiographs. The models were grouped according to the anatomical relationship between lumbar lordosis (LL) and pelvic incidence (PI) (δPILL = PI-LL). Models above a threshold for ΔPILL, which clinically corresponded to patients having had revision surgery, were assigned to group ASD + (n = 25); whereas models below the threshold and without symptomatic ASD were assigned to a control group (CTRL-,n = 34). To study only the effect of sagittal alignment, the height of the lumbar spine was normalized and segmental weights and center of mass positions were kept the same for all models. Simulations for static postures of neutral upright standing, 30° and 60° flexion and 15° extension were performed. Joint reaction loads between groups were compared in MATLAB with a Wilcoxon rank sum test. Results Computation of sagittal parameters LL and PI based on extracted landmarks matched well with measures determined in Reference1 and led to the same categorization of subjects. Validation of compression forces was performed with in vivo measurements3 (291N in our model vs. experimental values between 191 and 365N). Reported shear forces4 (L5-S1: 134N, L4-L5: 15N) were compared against those predicted by our model (L5-S1: 103N, L4-L5: 25N). Furthermore, load tendency for different postures were compared to intradiscal pressure values from the literature5,6, both normalized to neutral standing. Good agreement was achieved for flexion (239% vs. 220 and 260%, respectively) and extension (160 vs. 120% for both). Comparison of joint reaction loads for the two groups showed the most prominent differences in shear force and joint moments at the L3-L4 IVD joint level (Fig. 1). Significantly higher shear forces were observed in the ASD + group for flexed postures, while differences in neutral position were not significant. Flexion moments were significantly different among the two groups for all postures, while magnitudes were higher for the ASD + group for 60 degrees flexed postures only. (Fig. 1 shear force and flexion moments at the IVD joint L3-L4 for both groups) (Figure presented) Conclusion The simulations showed that the IVDs at the third lumbar level of patients who needed revision surgery due to development of symptomatic ASD were exposed to significantly higher loading regimes before fusion in flexed postures. This might be an indication that not only the biomechanical effect of a fusion, but already a difference in disk loading history prior to fusion could affect the risk of ASD. Simulations of fusion for the two groups will show if the predictive capability of Reference1 can be supported by mechanical consequences on the loading of the IVD.