Previous studies have investigated the correlation between IDH and other radiographic parameters on the diseased segment [1], while the features of the non-diseased segment were not deeply researched. So, both diseased and non-diseased segments were measured in our study aiming to clarify the overall radiographic differences between the DLSS and healthy individuals.

This paper found that IDH in the DLSS group was significantly higher than that in the control group at all lumbar segments except for L4-5, suggesting IDH in DLSS patients may be higher. Also, IDHL4-5/L3-4 was smaller in the DLSS group than in the control group. Although a significant difference was not observed in this parameter, it still suggested a decrease in IDH at the responsible segment in the DLSS group. As a result, iIDHL4-5 in DLSS group was significantly higher than the IDH of L4-5 in the control group. Bai et al. [15] reported that IDHL4-5/L3-4 in Chinese population was 1.14, and the ratio we calculated in the healthy individuals was 1.12 ± 0.23, which was extremely close to the reported value. So, the iIDHL4-5 we estimated was reliable and it could represent the initial IDH of DLSS patients in this study when they didn’t suffer from DLSS. Anna et al. [16] found that higher intervertebral disks are more prone to degeneration after undergoing greater deformation and stress during extension and rotation movements, and IDH would decrease by 0.98–1.6 mm if DD increased by one level while the non-DLSS individuals with lower disks had smaller intervertebral motion, and disks could keep their height. Even though, the IDH in the Control group was still lower, for the initial disk height of DLSS patients was exaggeratively high.

Vergroesen et al. [17] proposed a vicious circle that the disk begins to degenerate due to long-term excessive stress, enhanced catabolism of nucleus pulposus cells, and disruption of the extracellular matrix structure. Based on the theory, higher disks bear more stress and are easier to enter the vicious circle and IDH begins to lose as a consequence, which is consistent with our conclusion. Furthermore, the injury and degeneration at L3-4 and L4-5 are more severe, because the segments are located at a transition area from the rigid sacrum to the active lumbar spine with the largest motion of the intervertebral space, whereas the intervertebral spaces of L1-2 and L2-3 have relatively less mobility and L5-S1 is below the posterior superior iliac spine with the protection of transverse process and strong ligament [16, 18]. As a result, a significant difference in DD was observed at L4-5 between the DLSS group and the control group. However, DD showed no significant relation with ROM, probably due to the lateral and shear stresses causing a greater impact on the intervertebral disk [16].

Maxim et al. [19] found that mechanical stress caused early degeneration of the intervertebral disk as well as the facet joint. Conversely, the degenerative facet joints lead to abnormal stress and accelerate the degeneration of the intervertebral disk. Interestingly, Liu et al. [20] compared the FJA of lumbar spondylolisthesis and healthy population and found the facet joints of spondylolisthesis patients are more sagittalized. The sagittalized facet joints have less ability to limit the mobility of spine motion [21] and lumbar spondylolisthesis might occur. And it may be also a risk factor for DLSS. This study found that the risk for developing DLSS increased when FJA ≥ 52.03°. Hasegawa et al. [10] reported the volume of the facet joint is associated with lumbar instability. The FJO was significantly larger in the DLSS group than in the control group, suggesting that the lumbar segments in DLSS have greater mobility in our study. And the risk of suffering DLSS significantly increased if FJO > 3.75 mm. However, FJA and FJO did not show a relation with other radiographic parameters, and their function in DLSS needs to be further investigated in the next clinical trial.

The atrophy of the multifidus and psoas major muscles was more severe in the DLSS group. Xia et al. [22] found that the atrophy of paravertebral muscles is associated with the severity of stenosis, which is attributed to the denervation of paravertebral muscles after nerve injury or the influence of inflammation and immune response of DD. The point was also confirmed by a basic experiment. Hodges et al. [23] destroyed the disk and nerve roots of mice, then the atrophy and adipocyte clustering appear in multifidus muscle 3 days after the operation, while the contralateral side just shows adipocyte aggregation. Moreover, there was a relation between IDH and the area of multifidus and psoas muscles. A literature pointed out that decreased stability reflex of multifidus muscles will attribute to the desensitization of mechanical receptors caused by the relaxation of viscoelastic tissue within the disk (narrowing of the disk and formation of asymmetric geometry) [24]. The IDH loss caused the spine canal, lateral recess, and intervertebral foramen to narrow. As a result, nerve roots are compressed and the afferent and efferent pathways of the stability reflex are damaged; eventually, the atrophy of muscles happened [23].

CEF was more severe in the DLSS group than in the control group in all lumbar segments in this paper. Rajasekaran et al. [5] thought CEF might be the initial factor for DD. Beth et al. [25] reported that the transport of small molecules would reduce after CEP was damaged and DD begins subsequently. Uruj et al. [26] found DD was associated with the area of endplate damage by autopsy and μCT examination. And higher disk height allows the greater activity of the intervertebral space, and more stress is applied to CEP [16]. As a result, the progress of DLSS was accelerated by CEF by promoting DD. However, the negative linear relation between CEF and DD was only observed at L1-2, L2-3, and L5-S1, which is attributed to the fact that the intervertebral spaces of L3-4 and L4-5 have greater mobility and are subjected to greater stress [18]. Endplate degeneration or defects results in a range of clinical symptoms and diseases by weakening the transport of nutrients and changing the local or overall stress state of the disk [27]. However, the Modic changes showed no difference between the two groups, indicating that Modic changes might not be associated with DLSS.

Our study found that LF was significantly thicker in the DLSS group than in the control group. Sakamaki et al. [28] pointed out that LF would be thicker at all spine segments if the thickness of LF is larger than 3 mm at L2-3. And the thickness of LF at L2-3 was 4.2 ± 0.5 mm in the DLSS group, which was consistent with the previous study. It might be the reason why the patients are vulnerable to symptoms of nerve compression. Peng et al. [11] confirmed that the thickening of LF is associated with stress while LF had no association with ROM in our study, indicating that the thickness of LF is more relevant with lateral and rotational movement. Yabe et al. [29] found the thickness of LF showed a significant relation with age and segments instead of IDH, which is consistent with the results of this study.