The human spine is a highly mobile load-bearing column. Its sagittal alignment comprises self-adjusting reciprocal curves [1] that constantly try to maintain balance, so as to achieve energy conservation [2] and good health-related quality of life scores [3]. Although spinal alignment is well documented in standing, it is far less well studied across other commonly assumed physiological postures [4]: in particular, the slump sitting flexion that gives the maximum kyphosis [5] and the backward bending extension that gives the maximum lordosis [4]. Since spinal alignment changes most when adapting between these different postures, it is expected that the load-bearing properties of the axial skeleton will also vary between them.

The lumbar spine’s load-bearing area lies largely within its anterior column throughout most of its range of movement [6] and can be represented by changes in relative disk heights across the endplates [7,8]. In this study, we aimed to identify the load-bearing pattern of each functional spinal unit (FSU) using the changes in relative disk heights that occur during commonly assumed, clinically meaningful physiological postures: flexion (slump sitting), erect (standing), and extension (backward bending). The results can potentially serve as an additional guide for intervertebral cage placement to achieve spinal realignment while respecting the strut properties of cages. Additionally, it may also lend some insight into the mechanical reasons for catastrophic hardware failure, such as rod breakages, after spinal deformity surgeries.

Table 1 shows the baseline characteristics of the study cohort. Eighty-three patients (77 males and six females) participated in the study, and their mean age was 21.7 years. All patients had American Society of Anesthesiologists grade 1–2. The mean PI was 48.8°, and the mean LL was −35.2°. The mean SVA was −1.3 cm, and the mean PT was 16.1°. No lower limb compensation in the sagittal alignment was observed. A total of 415 disks were analyzed, of which, the majority were Pfirrmann grade 1–2 (74.7%).

The change in mean global ROM for the lumbar spine between flexion and extension was 50.2°. Relative to the erect posture, a larger proportion of mean alignment change was found during flexion 45° (89.6%) than during extension 5.2° (10.4%) (Table 2). Therefore, the erect posture resembles the extended lumbar spine to a greater extent (Table 2). Significant differences in ROM were found between all three postures for the L1–S1 global angle and at the L1/2 segment. Significant differences were found only between flexion and erect for the L2/3, L3/4, L4/5, and L5/S1 segments. Subgroup analysis was also performed, which showed no significant differences in disk angles and disk height measurements across Pfirrmann grades and between low and high PI groups (p>0.001). Table 2 compares the measurements of the global and segmental lumbar spinal alignments, as well as the disk heights, between postures.

Fig. 2 illustrates FSU dynamics and the changes in disk heights when transitioning from erect to flexion, as well as from erect to extension. These changes demonstrate portions of the disk that approximate (black arrows), separate (red arrows), or which show no change in height (green block). The portion that shows no significant change in disk height can be described as the fulcrum of movement, and represents the neutral axis. It can be determined for both flexion and extension.

Fig. 3 shows the region of the spine that experiences compressive forces (represented by blue arrows) for recommending the optimal position of cages when fulfilling strut properties. Using the angular changes across postures occurring in each FSU, the load-bearing area in the erect posture is also plotted. The erect posture lies between these two common postures, albeit resembling more the extended posture, explaining why the load-bearing area plots in this posture (yellow circles) approximate the extended posture as shown in Fig. 3.

In recent years, the importance of restoring sagittal spinal alignment [11] has led surgeons to pursue a corrected lumbar lordosis that matches an individual’s pelvic incidence (PI–LL discrepancy) [12]. In pursuing this strategy, previous studies have advocated that intervertebral cages be placed “as anterior as possible” to maximize anterior column distraction [13-15]. However, this may have implications in terms of the cages not fulfilling their coexisting role as load-bearing struts, and potentially accounting for some of the complications seen following deformity correction surgery prior to fusion, such as pseudoarthrosis [16] and rod breakages [17].

Similar to previous studies [18,19], this study showed that the load-bearing properties of the lumbar spine change across postures, are level-specific, and are more anterior with less lumbar lordosis (Fig. 3). However, minor differences in the position of the load-bearing area were found in the current study. These could be attributed to the differences in methodology, with the current study better understanding the physiological lumbar spine alignments, i.e., using flexed (slump sitting), erect (standing), and extended (backward bending) postures, which gives greater clinical relevance to spinal realignment surgeries. This study uses different postures and their associated different alignments in each individual to conclude the relationship between load-bearing properties and spinal shape. Its methodology removes any interindividual confounding factors, thus achieving proof-of-concept. Its extrapolation to recommendation of cage placement across individuals, though logical, should still be tested in future studies.

Along with other studies [18,19], the current study also supports the notion that cage placement should be influenced by the final shape in which the lumbar spine is fused. While respecting the advantages of “as anterior as possible” cage placement for lordosis realignment, a mechanically more stable anterior column reconstruction from cage placement may be preferred. This includes situations when the lumbar spine is required to be fused in lordoses of varying magnitudes (over- and undercorrection strategies [20-22]) and curve shapes (restoring Roussouly curve type strategies [1]).

Various alignment targets for the lumbar spine have been proposed in the literature [23,24]. To date, there is ambiguity as to what an ideal spinal alignment entails, as different surgeons strive to fuse the lumbar spine in different degrees of lordosis. In accordance with the SRSSchwab classification [25], which considers only the magnitude of lordotic restoration, the physiological shape of the lumbar spine is disregarded. Since a cage placed with the sole aim of maximizing lordotic restoration may not simultaneously provide optimal anterior support, as shown in our results, paying attention to the levelspecific restoration of load-bearing properties of the spine is important. Alternatively, the Roussouly classification emphasizes the need to restore individualized curve morphologies [1], and recommends that patients with type 1 sagittal profiles require shorter hyperlordotic segments in the lower lumbar spine. This implies that a larger extent of lordosis must be generated in the lower lumbar spine, urging the need to respect proper cage placement for anterior support in these segments.

Intervertebral cages, in the authors’ opinion, should be placed at least in the neutral axis, if not in the load-bearing area, to allow optimal spinal realignment and load bearing. In the event that massive lordosis is warranted in a particular FSU, the cage needs to be placed “as anterior as possible” with increased risk of rod fractures, and posterior constructs should be reinforced [26]. Alternatively, cages with reinforced anterior fixation can also be employed to improve the overall strength of the construct [27].

To fulfil optimal load-bearing properties of the lumbar spine, intervertebral cages should at least be situated on the erect weight-bearing axis, if not on the load-bearing area. Fig. 3 illustrates a snapshot of the plotted loadbearing areas at each lumbar intervertebral level in different alignments and recommends ideal, level-specific cage positions, assuming fusion is performed in these FSU alignments. Awareness of the varying load-bearing areas occurring in different extents of lumbar lordosis can help surgeons achieve their desired amount of correction while minimizing complications. As determined in this study, this position is more anterior in a less lordotic FSU, and more posterior in a more lordotic FSU (Fig. 3). For example, L1/2 will require the furthest anterior cage placement if fused in flexion, while L3/4 will require the furthest posterior cage placement if fused in the erect or extension postures. With a larger lordosis, it is also likely that facet joints may participate more in load-bearing. This is demonstrated at L3/4, where the fulcrum of movement from erect to extension is found to be posterior to the vertebral body. Therefore, we recommend that for lordotic spinal realignment in accordance to the erect posture, to optimize load-bearing function, intervertebral cages should generally be placed in the mid-zone of the endplates, or slightly posterior (Fig. 3B).

While the use of young patients with nonspecific lowerback pain, and without advanced disk degeneration, allows reliable measurement of disk heights to determine the neutral axis, and subsequently, the load-bearing area of the spine, it can be argued that the spinal dynamics of young healthy individuals may not represent those of patients with adult spinal deformity and degeneration. Nevertheless, this is offset by the fact that a realigned spine should approximate that of a normal spine, in terms of both disk height and spinal shape.

Although this method of studying spinal dynamics may differ from the conventional technique used to plot instantaneous axes of rotations (IARs) and centers of rotation (COR) in the sagittal plane, which classically compares intersecting perpendicular bisectors of the moving underlying vertebra on lateral radiographs of flexed and extended spines in the standing posture [18,28], its simplified, yet logical, methodology will allow easier understanding of the anterior column in load bearing. Further quantification of IARs to characterize COR can be performed in future studies to allow a more complete understanding of cage placement and its effect on spinal mechanics.

This study details disk height pattern changes in various key sagittal spinal profiles reflective of physiological loadbearing behavior. It is advantageous in demonstrating the clinically relevant, alignment-specific and level-specific nature of the load-bearing areas in the lumbar spine. It suggests that intervertebral cage placement should not be routinely “as anterior as possible,” disregarding the need to optimize load-bearing properties of the lumbar spine. However, to support its use to guide intervertebral cage placement, further clinical studies should be conducted to substantiate how the positioning of cages could affect clinical fusion rates and surgical outcomes.