By comparing imaging parameters between the two groups, we found that patients in the L4 IS group had smaller LLL than those in the L5 IS group (Table 1). An age-matched analysis did not reveal any significant influence of age on either group (Table 2). Hyun et al. [12] divided lumbar lordosis into three parts, of which LLL describes lordosis of the L4–S1 and counts for the two-thirds of the total lumbar lordosis [17]. As most cases of IS were located at L4–S1, LLL was included in the present study. The difference in LLL between the two groups could be explained by the special compensation mechanism of IS. For most people with complete pars interarticularis of the vertebral arch, the local sagittal imbalance of the lesion segment is always compensated by hyperextension of adjacent segments, restricting the consequence of lumbar kyphosis on the load of axis gravity [18]. However, for patients with spondylolytic lesions, the imbalance of the lesion segment always progresses to anterior slippage of the upper vertebra. With the anterior slippage of the vertebra, the segments above the spondylolisthesis hyperextend and the lordosis increases in the cranial zone of the lumbar spine, limiting the excessive forward shifts of the center of gravity [19]. This also explains the higher LL among patients with L5 IS in many studies [8–10]. As a result, segmental lordosis of L4/5 in L5 IS increased while that of L5/S1 in L4 IS remained nearly unchanged. With the same degree of degeneration and lordosis of the lesion segment, LLL of the IS at L4 was smaller than that at L5 (Fig. 2). Typical changes in radiological features of patients with single-level IS at L4 and L5 are shown in Fig. 3. We observed a straighter low lumbar curvature in the L4 IS.
Roussouly et al. [16] proposed four types of sagittal lumbo-pelvic alignment using lateral radiography. Certain types of sagittal lumbo-pelvic alignment were more frequently associated with specific degenerative diseases [20]. Patients with symptomatic disc herniation were most commonly classified as Type 1 or 2, and patients with spinal stenosis as Type 4 [16]. Funao et al. [21] found that degenerative spondylolisthesis tended to be classified as Type 3 or 4. In our study, most patients with single-level IS were classified as Type 3 or 4 (Table 3). However, the most frequent type was Type 2 in the L4 IS group and Type 3 in the L5 IS group. As Roussouly et al. [16] pointed out, Type 2 patients had flatter low lumbar curvature than Type 3 patients. Different distributions of Roussouly type among patients with L4 IS and L5 IS also suggested a straighter low lumbar curvature among patients with IS at L4.
In the present study, patients with IS at L4 were found to be older than those with IS at L5. We believe that this could be related to the biomechanics of the lumbar spine. To maintain lumbar lordosis, the lower arc of the lumbar spine, mainly L4 and L5, must be tilted downward [18]. This would explain the anterior instability with sliding-producing spondylolisthesis that always occurs at L4 and L5. As L5 is located at the lumbosacral junction with greater downward tilt, the shear force at L5 and its posterior structure is greater than that at L4. A smaller anterior shear force at L4 and its posterior structure results in later fracture of the bilateral pars interarticularis at L4, which is likely why we observed a higher proportion of older patients in the L4 IS group. However, DDG, SR, and ODI showed no significant differences in our study, even when accounting for age. This result was discordant with the conclusions and speculations of previous anatomical studies [6]. The disagreement was caused mainly by difference in study subjects; the subjects in our study were in urgent need of interventions while those in previous anatomical studies had no clinical symptoms or only mild symptoms. Many studies have shown that L4/5 is more unstable than L5/S1 due to lack or less of connections between iliolumbar ligament and L4 [5, 22]. Despite the similar severity of clinical symptoms and slippage, we believe that patients with IS at L4 deserve more attention.
Comparisons between Table 4 and Table 5, reveal that fewer sagittal lumbo-pelvic parameters were significantly correlated to SR in the L4 IS group than the L5 IS group. There were fewer significant relationships among sagittal lumbo-pelvic parameters in the L4 IS group. This could be due to the discontinuity of the lumbar curve in IS at L4. PI, first proposed by Duval-Beaupere [23], has now been proven to be correlated to the grade and progression of slippage in IS as a basic anatomical parameter. However, these studies only focused on IS at L5 or simply ignored different lesion segments [8–10]. Oh et al. [24] pointed out that PI could be a progressive factor for slippage in IS at L5, but not at L4, and concluded that for patients with IS at L4, segmental instability and disc degeneration in L4/5 could have a greater influence on the pathological mechanism of slippage. In the current study, we also observed that PI was positively related to SR in the L5 IS group but not in the L4 IS group. PI was also significantly correlated with all other sagittal lumbo-pelvic parameters in the L5 IS group, but not in the L4 IS group. Thus, PI is an important parameter to be considered in adults with IS at L5.
In terms of the relationship between sagittal lumbo-pelvic alignment and SR, we found that only L5 I showed a significantly positive correlation with SR in both groups. L5 I, first introduced by Roussouly, was often used as a sagittal lumbo-pelvic parameter to assess spinopelvic morphology in high-grade spondylolisthesis and for surgical follow-up, especially when there was sacral doming [13]. Although there were significant positive correlations between L5 I and SR in both groups, the bases of the correlations were different. For L5 IS, this positive correlation was mainly due to the secondary change caused by slippage of L5. With the anterior slippage of L5, the upper end plate of L5 gradually inclines downward and forward, and the vertical distance between L5 and the hip joint decreases, thus increasing L5 I [25]. But for IS at L4, the structure under L4/5 is relatively stable. L5 I is more similar to a constant like PI in IS at L5, which can be considered a progressive factor for slippage in IS at L4. For patients with IS at L5 and high PI, close follow-ups and early interventions are very important, as PI is an effective biomechanical parameter for predicting progression of spondylolisthesis [26]. As for patients with IS at L4, high L5 I is worth examining as it contributes a certain reference value for clinical doctors when making decisions about treatment.
The sagittal lumbo-pelvic alignment was not only closely related to SR but also the degree of clinical symptoms of patients with spondylolisthesis. Tanguay et al. [27], through analysis of 96 patients with L5 spondylolisthesis, proposed that decreased LSA is significantly correlated with a decline in the physical aspect of quality of life. Wang et al. [28] showed that PT, SS, and LL had significant correlations with ODI in patients with severe L5 IS. In our study, only SS showed a significant correlation with ODI in the L5 IS group, while no sagittal lumbo-pelvic parameters were significantly correlated with ODI in the L4 IS group.
Contrary to our expectations, we found no significant correlations between SR and ODI in either group. This revealed that there were no associations between the severity of clinical symptoms and degree of slippage. Considering that most patients in our study had mild to moderate degree of slippage (SR < 50%), it was reasonable to conclude that the severity of clinical symptoms of patients with low-grade IS was more closely related to their tolerance of lumbar spinal stenosis and nerve root compression rather than to the grading of images.
There were some limitations to this study. First, as a retrospective and cross-sectional study, lack of follow-up information impeded our comparison of prognosis between the two groups. However, this did provide a starting point for prospective and longitudinal studies. Second, the subjects were limited to patients who were obviously symptomatic and needed interventions, which affected the comparison results to some extent. Nevertheless, we believe that our study is of significant interest, as we are the first to discover a difference in sagittal lumbo-pelvic alignment between single-level IS at L4 and L5 and identify the reason for the difference. Our correlation analysis between SR and sagittal lumbo-pelvic parameters in IS at L4 and L5 will also be helpful in outlining appropriate interventions for patients with IS.