The difference in the relationship of spinal sagittal alignment between individuals with flat lumbar and normal lordosis posture based on global and regional angles

1.Introduction

Spinal posture assessment is often emphasized by clinicians because of its clinical importance for managing pain, surgical treatment outcomes, prevention of pain or deformity, and functional assessment [1, 2, 3]. In the sagittal plane, the human spine comprises a series of reciprocal curvatures between the cervical and pelvis, with the cervical spine and lumbar spine generally characterized as being lordotic, whereas the thoracic spine is normally kyphotic. These series of segments support the spine and are linked together along a linear chain to maintain a stable posture [4]. An optimally balanced spinal posture contributes to maximal mechanical advantage and minimal energy expenditure during standing and walking by absorbing the shock of the load imposed on the spinal column [4]. However, sagittal profile imbalance is considered one of the variables most highly correlated with adverse health status and outcomes such as significant low back pain, deformity, or gait abnormalities [3, 4, 5, 6].

An excess of lumbar lordosis or thoracic kyphosis is the common driver of sagittal contour imbalance, but a deficiency of normal lumbar lordosis or thoracic kyphosis can also cause spinal sagittal imbalance [2, 3, 7]. Previous studies have reported that such reduced sagittal contours of the spine may be caused by advanced age, as in degenerative lumbar kyphosis, or passive structural problems, such as iatrogenic flat back syndrome and ankylosing spondylitis or scoliosis [2, 7, 8]. However, even healthy people may present with a sagittal profile imbalance such as flat back or sway back postures. These are relatively minor forms of deviated posture and may be a consequence of uncontrolled or faulty movement patterns [8, 9].

Posture is commonly understood to be influenced by a complex interplay of various elements of the human body such as the bone structure, muscular system, and neural system, interacting with external environmental forces such as gravity. In particular, the orientation of the line of gravity relative to the spinal column plays a vital biomechanical role in the stress placed on the region [9, 10]. For instance, if a gravity line passes through the lumbar anterior region, the low back nature lordosis will be reduced due to generating continuous external flexion torques [8]. Thus, the reduction of lumbar lordosis increases under stress, which can lead to mechanical changes in shear forces and coupling motions in the spine. This, in other words, can cause compensatory posture and associated movement in adjacent segments of the vertebral curve during various static and dynamic daily activities [4, 7]. Therefore, it is important that early and precise assessment of spinal curves to alterations of sagittal alignment in individuals with a risk of spinal disorders. In addition, it is necessary to investigate how these spinal sagittal imbalances affect the interconnected spinal chains.

Many previous studies on sagittal assessment of the spine have reported that pelvic morphological parameters, such as pelvic incidence, have a significant influence on the orientation parameters of global lumbar lordosis (GLL) and global thoracic kyphosis (GTK) [3, 4, 7, 11, 12]. These authors reported that greater lordosis angles were correlated with increased sacral slope, pelvic incidence, and pelvic tilt [1, 4, 7, 12, 13]. These relationships between pelvic parameters and lumbar lordosis appear to be clear, although in previous studies, the link between GLL and GTK or that between pelvic parameters and GTK has been inconsistent. According to the results of previous studies, greater GTK is generally associated with greater GLL, although some studies have reported individuals with increased GLL but reduced GTK. Moreover, other studies have concluded that GTK is not influenced by GLL in the maintenance of physiologic balance, as implied by the association between sacral slope and GLL or that between GLL and GTK [4, 7, 14].

It is believed that these conflicting results are due to the wide normal ranges of GLL and GTK, the diversity of participants, and the wide range of age groups studied. Roussouly et al. [3] classified patterns of sagittal curvature associated with lumbar lordosis into four types based on the upper and lower lumbar (LLx) lordotic arcs in 160 participants with healthy backs and ranging in age from 18–48 years. The authors reported that the participant’s sacral slopes varied from 20∘–65∘ and GLL ranged between 41–82∘. Berthonnaud et al. [4] found a GLL range of 16∘–71.9∘ and a GTK range of 22.5∘–70.3∘ in a cohort of 160 asymptomatic volunteers aged 20–70 years. In addition, several studies separated the upper and lower regions of the spine to evaluate posture [3, 4, 15, 16].

Lumbar lordosis primarily includes the LLx vertebrae, with the L5 segment contributing almost 40% to overall lordosis and L1 contributing only 5% [17]. Consistent with greater mechanical stress through the LLx regions, higher risks for degenerative low back pain are more evident in these regions than in the upper lumbar (ULx) regions [15, 16]. Furthermore, the majority of spinal surgeries are in the LLx region [18]. Thus, it is considered that during the assessment of spinal sagittal balance, it is most efficient to evaluate the spine globally and then divide it into sections. Therefore, the purpose of this study was to investigate spinal sagittal alignment and its relationship to global and regional lumbar and thoracic angles, pelvic and sway angles, and C7–S1 distance measurements, followed by a detailed subgroup analysis using an inertial measurement unit system.

2.Methods

3.Results

3.2Linear data

Tables 4 and 4 present the correlations among the global and regional thoracic and lumbar angles and the sacral inclination in the flat lumbar and normal lordosis groups. Figure 1 depicts the data analysis using a stepwise regression model.

Figure 1.

Correlations between global and regional angles of lumbar in flat lordosis group (A) and normal lumbar group (B). Correlations among thoracic global and regional angles and sway angles in flat lordosis group (C) and normal lumbar group (D).

3.2.1Global and regional lumbar data

A strong correlation was observed between the GLL and LLx angles in the flat lumbar group (r= 0.79, r2= 0.62, p< 0.001; Table 4 and Fig. 1A) and the normal lordosis group (r= 0.86, r2= 0.74, p< 0.001; Table 4 and Fig. 1B). Furthermore, a moderate correlation was observed between the GLL and ULx angles in the flat lumbar group (r= 0.48, r2= 0.23, p= 0.011; Table 4 and Fig. 1A) and between the ULx and LLx angles in the normal lordosis group (r=-0.51, r2= 0.26, p= 0.011; Table 4 and Fig. 1B).

A strong correlation was observed between the L3 and ULx angles in the normal lordosis group (r= 0.71, r2= 0.51, p< 0.001; Table 4 and Fig. 1B). Moreover, a moderate correlation was observed between the L3 and ULx angles in the flat lumbar (r=-0.46, r2= 0.21, p= 0.016; Table 4and Fig. 1A) and normal lordosis (r= 0.65, r2= 0.42, p= 0.001; Table 4) groups. A strong correlation was observed between S2 and GLL (r=-0.88, r2= 0.77, p< 0.001; Table 4 and Fig. 1B) and LLx (r=-0.88, r2= 0.77, p< 0.001; Table 4 and Fig. 1B) angles in the normal lordosis group. Additionally, moderate correlations were observed between S2 and GLL (r=-0.65, r2= 0.42, p< 0.001; Table 4 and Fig. 1A) and LLx (r=-0.54, r2= 0.29, p= 0.004; Table 4 and Fig. 1A) angles in the flat lumbar group.

The stepwise regression model revealed that S2 (i.e., sacral inclination) was the best predictor of GLL and LLx angles in both groups, accounting for 42% and 28.8% of the variation in them, respectively, during standing in the flat lumbar group, and for 77.5% and 76.9%, respectively, in the normal lordosis group (Fig. 1A and B). Additionally, L3 was the best predictor of ULx in both groups, accounting for 21% of the variation in the flat lumbar group and for 50.7% in the normal lordosis group (Fig. 1A and B).

3.2.2Global and regional thoracic data

A strong correlation was observed between GTK and UTx angles in the flat lumbar group (r= 0.74, r2= 0.55, p< 0.001; Table 4 and Fig. 1C) and in the normal lordosis group (r= 0.73, r2= 0.53, p< 0.001; Table 4 and Fig. 1D). Additionally, a moderate correlation was found between GTK and LTx angles in the flat lumbar group (r= 0.49, r2= 0.24, p= 0.009; Table 4 and Fig. 1C). The T1 angle showed a strong correlation with the GTK and UTx angles in the flat lumbar (r= 0.90, r2= 0.81, p< 0.001 and r= 0.79, r2= 0.63, p< 0.001, respectively; Table 4 and Fig. 1C) and normal lordosis (r= 0.91, r2= 0.83, p< 0.001 and r= 0.79, r2= 0.63, p< 0.001, respectively; Table 4 and Fig. 1D) groups. The T7 angle demonstrated a strong correlation with the LTx angle in the flat lumbar (r= 0.78, r2= 0.61, p< 0.001; Table 4 and Fig. 1C) and normal lordosis (r= 0.79, r2= 0.63, p< 0.001; Table 4 and Fig. 1D) groups. Furthermore, the T7 angle was moderately correlated with the UTx angle in the normal lordosis group (r=-0.62, r2= 0.38, p= 0.001; Table 4). Additionally, the T12 angle showed a moderate correlation with the LTx and GLL angles in the flat lumbar (r=-0.52, r2= 0.32, p= 0.005 and r= 0.41, r2= 0.17, p= 0.034, respectively; Table 4) and normal lordosis (r=-0.57, r2= 0.32, p= 0.003 and r= 0.46, r2= 0.21, p= 0.025, respectively; Table 4) groups. Furthermore, a moderate correlation was observed between T12 and GTK angles in the normal lordosis group (r= 0.69, r2= 0.48, p< 0.001; Table 4).

The stepwise regression model revealed that T1 was the best predictor of GTK and UTx angles in both groups, accounting for 81% and 62.7% of them, respectively, in the flat lumbar group, and for 82.5% and 62.5%, respectively, in the normal lordosis group (Fig. 1C and D). Additionally, T7 was the best predictor of the LTx angle in both groups, accounting for 60.7% of the variation in the flat lumbar group and 61.1% in the normal lordosis group (Fig. 1C and D).

3.2.3Lumbar, thoracic, and sway angles

In the normal lordosis group, moderate correlations were observed between the GTK and GLL angles (r=-0.48, r2= 0.23, p= 0.048), between sway and GTK angles (r= 0.50, r2= 0.25, p= 0.012), and between sway and UTx angles (r= 0.54, r2= 0.29, p= 0.007) (Table 4). Additionally, in the normal lordosis group, T1 was the best predictor of sway angle, accounting for 29.9% of its variation (Fig. 1D).

4.Discussion

This study assessed the global and regional angles of the lumbar and thoracic spine and pelvis and postural alignment in standing, as well as the relationships among these angles, in individuals with flat lumbar posture and normal lordosis. Our primary findings demonstrate that the global and regional lumbar and thoracic angles were generally reduced in the flat lumbar group compared with the normal lordosis group, except for the upper thoracic (UTx) angle, whereas the S2 angle (i.e., sacral inclination) was increased in the flat lumbar group relative to the normal lordosis group. Moreover, the S2 angle proved to be the single best predictor of GLL and LLx in the standing position for both groups. Additionally, there was a significant and strong correlation between GLL and LLx in both groups. These results indicate that the GLL angle is affected by the orientation of the pelvis and controls sacral inclination in a standing position. Furthermore, the S2 angle was the main contributor to GLL and LLx in these young healthy volunteers. Many previous studies suggested that pelvic alignment, as defined by pelvic incidence, sacral slope, pelvic tilt, and sacral inclination, should be employed to assess the sagittal profile [3, 4, 5, 6] because optimal sagittal contour is influenced by pelvic shape and orientation. The present results are similar to the findings of Hyun et al. [18], who demonstrated a strong correlation between GLL and sacral slope (r= 0.764). Moon at al. [23] also confirmed a strong relationship between GLL and sacral slope in both asymptotic volunteers and symptomatic patients. These previous studies assessed sagittal contour using radiographic parameters, whereas the present study employed an IMU system with skin-surface markers to measure GLL and RLL.

Despite these differences in measurement techniques, the results of the previous studies and this one are similar. Legaye et al. [11] reported that apart from the strong relationship between GLL and sacral slope, a high correlation was also observed between sacral slope and sacral inclination (i.e., S2 angle), and as GLL increased, sacral inclination and sacral horizontal angles also tended to increase. Thus, in light of previous data and the present findings, we consider that the GLL angle was affected by the orientation of the pelvis and its control of sacral inclination during a standing position and that the S2 angle was the main contributor to GLL and LLx in these young healthy volunteers. Furthermore, a previous study by Shin and Yoo [19] also reported that as GLL increased, ULx and LLx tended to increase, whereas the sacral inclination tended to decrease. In that study, the sacral inclination was the single best predictor of GLL in standing participants using IMU measurements. In the present study, individuals with a flat lumbar posture tended toward a posterior pelvic alignment compared with those exhibiting normal lumbar lordosis, possibly due to the reduced range of the reciprocal lower lumbar curve. Additionally, the flat lumbar group also showed a decreased sway angle (i.e., angles between the acromion, greater trochanter, and lateral malleolus) and increased C7-S1 distance compared with the normal lordosis group because the latter group had a well-balanced GLL, whereas the flat lordosis group displayed pelvic posterior tilt. This result indicates that the flat lumbar group demonstrated greater spinal sagittal imbalance than the normal lordosis group. For this reason, we consider that global and regional lumbar and thoracic angles were generally reduced in the flat lumbar group compared with the normal lordosis group.

In this study, although the S2 angle was the key contributor to GLL and LLx in both groups, the groups showed several important differences. GLL and LLx were strongly correlated with S2 in the normal lordosis group, whereas the correlation with S2 was only moderate in the flat lumbar group. Furthermore, the amount of variance in GLL explained by the LLx angle was 38.9% greater than that due to ULx in the flat lumbar group, whereas the amount of variance in GLL explained by LLx was 73.8% in the normal lordosis group. In addition, only the normal lordosis group showed a moderate correlation between GTK and GLL. A possible explanation for this observation is that the GLL angle was more influenced by LLx than by ULx, and the extent of the influence on GLL varied depending on the S2 angle. Król et al. [24] and Berthonnaud et al. [4] reported a significant correlation between the pelvic tilt angle and lumbar lordosis, and it seems that lumbar lordosis is closely linked to pelvic position. When the pelvis is tilted posteriorly, the sacral end is projected posteriorly, displacing the axis of gravity in the same direction. Lumbar lordosis increases linearly with the sacral slope, and it has become apparent that the shape of the pelvis and its contribution to the sacral slope strongly influence lumbar lordosis [8]. Because the lumbar spine is associated with the pelvis through the kinematic lumbopelvic link, anterior and posterior pelvic tilt can lead to lumbar extension and flexion, respectively. In addition, the superior facet surfaces in the ULx region tend to be oriented closest with the sagittal plane, and the inferior articular facets are slightly convex, generally facing laterally to anterolaterally. The inferior facets of L5 articulate with the superior joints, which are typically oriented much closer to the frontal plane than the other lumbar articulations. This anatomical feature of the lumbar vertebra tends to increase the lordotic angle of the LLx compared to the ULx. The S2 angle is the best single indicator of GLL among skin surface measurements; however, a change in the shape or orientation at one level causes a change in the line of gravity, which alters the interrelationship between each vertebral region and directly affects adjacent segments, resulting in a change in posture. In this study, the flat lumbar group demonstrated significantly more posterior tilt than the normal lordosis group, and this result led to reduced GLL in the flat lumbar group compared to the normal lordosis group. Therefore, our results suggest that there was a weaker correlation between ULx and LLx in the flat lumbar group compared to the balanced, normal lordosis group. Thus, we suggest that it is necessary to evaluate sacral inclination (S2), GLL, and RLL for efficient lumbar postural assessment, and IMU-based measurement is necessary to classify subgroups according to GLL.

In this study, GLL showed a stronger correlation with LLx than with ULx, and GTK showed a stronger correlation with LTx than with UTx. In addition, pelvic inclination showed no relationship to GTK and regional thoracic angles. These results indicate that pelvic alignment had little direct influence on GTK, and GLL may not be the ideal parameter to evaluate reciprocal changes in GTK, although we also found a moderate relationship of GLL and GTK with well-balanced LLx and ULx in young healthy participants. In this study, the T1 angle was the single best predictor of GTK and UTx in the standing position for both groups. Our results are supported by Berthonnaud et al. [4] who demonstrated a weak correlation between the lumbar and thoracic spinal segments. The authors found a stronger relationship between shape and orientation parameters that tend to occur in the highly mobile areas of the spine, such as the lumbar and cervical areas. The less mobile thoracic spine does not respond as easily to changes in the shape or orientation of the pelvis and lumbar spine. In addition, based on the results, we consider that GTK is affected more by higher-level influences (e.g., cervical) than by lower-level (lumbar or pelvic) influences. The cervical and lumbar spine maintain lordosis, and the thoracic spine creates reciprocal kyphosis, but the upper thoracic vertebrae are an extension of the lordosis of the lower cervical spine, and the lower thoracic vertebrae are an extension of the lordosis of the upper lumbar spine. Our results are supported by those of several previous studies that examined sagittal balance in the standing position using radiologic methods. These previous studies demonstrated a weak or moderate relationship between GLL and GTK [12, 18, 23], and several studies showed a stronger relationship between the global cervical angle and GTK than between GTK and GLL [25, 26]. Previous findings by Hardacker et al. [25] showed that GTK alignment was strongly correlated with changes in cervical sagittal alignment in asymptotic participants. In addition, Lee et al. [26] and Knott et al. [27] reported that measurement of the T1 slope angle or sagittal angle may be very useful in evaluating cervical spine sagittal balance. An individual with an increased T1 slope requires a large global cervical angle to maintain physiologic sagittal balance. Although we did not measure the cervical angle in this study, our results show that the T1 angle may be a key factor in GTK.

This study has several limitations. First, all of our participants were asymptomatic healthy young adults. Therefore, our findings cannot be generalized to the general population. Second, we did not measure global and regional cervical angles; therefore, we did not assess the entirety of spinal sagittal balance and the relationship between global cervical angles and GTK. Third, although radiographic analysis is considered the gold standard measurement method, it is difficult to measure dynamic movements using this technique. Thus, in this study, we investigated spinal sagittal balance based on global, regional, and pelvic angles as measured by an IMU system using skin-surface sensors; these data were later used to study dynamic movements according to GLL. Finally, we did not examine other variables that might have affected posture, such as muscle length and intra-abdominal pressure patterns; we measured only the spinal angles noted in a standing posture. Future studies should investigate sagittal spinal balance, including cervical spine and muscle activation, in various posture subgroups.

5.Conclusion

The flat lumbar group displayed characteristic imbalances in the sagittal plane and a spinal relationship based on global and regional angles that was distinct from the normal lordosis group. In addition, the flat lumbar group showed no correlation between GLL and GTK, unlike the normal lordosis group. Despite these differences, LLx is useful for predicting GLL angles and UTx is useful for predicting GTK. Therefore, we suggest that when assessing posture, it is necessary to identify the global and regional angles and to classify patients into subgroups according to their spinal sagittal balance and degree of lumbar lordosis. Taken together, our results provide support for the comprehension of sagittal balance in the context of normal and pathological conditions of the spine and pelvis.