Introduction

Combined with multi-planar malformation and concomitant lumbar stenosis, DLS could cause intractable pain and neurological deficits, leading to marked loss in HRQOL. Surgical intervention could be recommended when conservative treatment fails to sufficiently alleviate symptoms. However, the postoperative HRQOL was not always to the DLS patients’ satisfaction [1]. Therefore, various kinds of affecting factors have been evaluated for postoperative quality of life [2,3,4]. Among them, it is generally acknowledged that ideal postoperative performance relies heavily on the optimal spinal alignment reconstruction [5, 6]. The SRS-Schwab classification provided a fundamental comprehension of the normative thresholds that PT, PI-LL and SVA should fall within, and DLS patients with overlarge PT and SVA showed poor quality of life and functional disability [7,8,9]. The importance of sagittal realignment was deeply investigated in the past studies; however, few are available to describe the clinical relevance between coronal plane parameters and HRQOLs, especially among patients with fixed coronal malalignment after long-level fusions.

Coronal malalignment is generally defined as a distance more than 2 cm from the coronal C7-plumb-line (C7PL) to the central sacral vertical line (CSVL) [10]. One recent study found that fixed coronal malalignment independently impacts SRS-22 total, social life and self-image in adult spinal deformity (ASD) patients [11]. However, the author only took postoperative CVA into account. Considering the etiological feature of DLS, asymmetric intervertebral disc and facet joint degeneration in local lumbar spine is regarded as the initial factor [12]. Thus, the remnant Cobb angle that could represent regional coronal malalignment should be taken into consideration as well. Moreover, osteotomy is not necessarily performed in DLS patients without severe spinal imbalance, but only decompression and fusion could not always sufficiently correct local scoliosis, therefore remaining Cobb angle among postoperative DLS patients is quite common. Whether it is necessary to fully correct the coronal curve after adequate neurological decompression in DLS should be figured out.

The preoperative association between spinal malalignment and HRQOL in the non-fused spine have been fully analyzed, however, the malalignment in the fused spine, probably caused by decompensation, correction loss or inadequate operation, has been rarely described. This study aims to investigate the prevalence of global and regional malalignment in coronal plane among surgically treated DLS patients, and their impacts on HRQOLs. As for sagittal plane, we also examine the clinical relevance between sagittal modifiers of SRS-Schwab classification and HRQOLs after long-level fusion.

Methods

Patient population

Our study was approved by the ethics committee of the Peking University Third Hospital. A retrospective review of 121 consecutive patients with DLS who underwent long-segment fusion surgery was performed, including 23 males and 98 females, with an average age of 62.8 ± 7.3 (range:40–79). Eligible were patients over 40 years of age who received isolated posterior instrumented arthrodesis, with ≥ 4 fused levels. Patients with following diagnoses were excluded, such as tuberculosis, tumor, and ankylosing spondylitis. Patients with previous lumbar surgery were excluded. Patients were also excluded if they did not reach the threshold of 1-year follow-up.

Data collection

Demographic and clinical data at baseline were collected using the electronic medical system in our hospital. Health-related quality of life data was collected including VAS Back pain, Oswestry Disability Index (ODI), Scoliosis Research Society (SRS) -22 questionnaire, and the 36-item Short Form Health Survey (SF-36) scores. The questionnaires were filled out at the final follow-up. The residual Cobb angle was defined as the major Cobb angle in coronal plane after long-level fusion surgery. Two assessors independently measured spinopelvic parameters including CA, CVA, PT, PI-LL and SVA with the Centricity Enterprise Web V3.0, and these parameters were collected from the last follow-up visit. Full-spine anteroposterior and lateral radiography was used.

In terms of coronal alignment, the patients were allocated into three groups according to different postoperative radiographic parameters. For postoperative residual Cobb angle, patients were divided into subgroups as follows: group (0) (CA < 10°), group (+) (CA 10°∼20 °), and group (++) (CA > 20°). For postoperative CVA, patients were separated into subgroups as follows: group (0) (CVA < 2 cm), group (+) (CVA 2 ∼ 3 cm), and group (++) (CVA > 3 cm).

In terms of sagittal alignment, the patients were grouped by the sagittal modifiers (PT, PI-LL, and SVA) as group (0), group (+), and group (++) according to the SRS-Schwab classification, respectively.

Statistical analysis

Categorical variables were analyzed using the chi-square test, while continuous variables were analyzed using the independent-sample t-test. Utilizing the normality test and distribution curve, the postoperative HRQOL data did not satisfy normal distribution. Spearman analysis was used to analyze the correlation between HRQOL and radiological parameters (CA, CVA, PT, PI-LL and SVA). The rank sum test was performed to compare HRQOL among subgroups by setting the thresholds of coronal and sagittal parameters. All statistical analyses were performed using IBM SPSS (version 27, IBM Corp.). A p value < 0.05 was considered statistically significant.

Results

A total of 121 patients were involved in our research, including 23 males and 98 females, with a mean age of 62.8 ± 7.3 (range:40–79). The mean follow-up was 34.8 ± 22.8 months (range:12–98). Demographical and surgical data of the patients grouped according to different radiographic parameters is listed in Tables 1 and 2. The residual coronal angularity in the postoperative DLS patients was demonstrated in Fig. 1.

Table 1 Demographical and surgical data among subgroups classified by CA and CVA
Full size table
Table 2 Demographical and surgical data among subgroups classified by SVA
Full size table
Fig. 1
figure 1

Preoperative and postoperative radiographs of a 76-year-old female with DLS. She presented with low back pain, lower limb pain, and neurogenic intermittent claudication. (A) Preoperative coronal view showed severe coronal imbalance with a main Cobb angle of 42.2°. (B) Preoperative sagittal view showed sagittal imbalance and thoracolumbar hyperkyphosis. (C) Despite global balance achieved in the coronal plane, the patient showed regional malalignment with a residual Cobb angle of 28.6° (D) Postoperative sagittal view showed an significant improvement in sagittal alignment

Full size image

The postoperative SVA showed significant correlation with a series of HRQOL after surgery. The ODI was positively correlated with postoperative SVA. The SRS-22 (subscales for Pain, Function, Self-image, Mental health, and Satisfaction) and SF-36 scores (subscales for PCS and MCS) were all negatively correlated with the postoperative SVA. CA and CVA did not show remarkable linear correlation with HRQOL. The correlation analysis between postoperative HRQOLs and radiographic parameters is shown in Table 3.

Table 3 Correlation between postoperative HRQOLs and radiographic parameters
Full size table

The results of rank sum test are shown in Table 4. The VAS back pain, ODI, SRS-22 total, and SF-36 PCS were significantly different among CA subgroups as shown in Fig. 2. When comparing SRS-22 subscales among CA subgroups, a statistical significance of the SRS-pain score was detected with differences in the CA (+) - CA (++) groups (p = 0.002) and CA (0) - CA (++) groups (p = 0.028). The SRS-function score was statistically significant with differences in the CA (+) - CA (++) groups (p = 0.006) and CA (0) - CA (++) groups (p = 0.043). The SRS-self-image score was statistically significant with differences in the CA (+) - CA (++) groups (p = 0.016) and CA (0) - CA (++) groups (p = 0.024). The SRS-mental health score and satisfaction score were statistically significant with differences in the CA (0) - CA (++) groups (p = 0.041) and CA (0) - CA (++) groups (p = 0.046), respectively.

Table 4 Comparison of HRQOL at follow-up between subgroups by each radiographic parameter
Full size table
Fig. 2
figure 2

Differences of health-related quality of life between subgroups. CA, Cobb angle; SVA, sagittal vertical axis; VAS, visual analog scale; ODI, Oswestry Disability Index; SRS-22, Scoliosis Research Society-22 questionnaire; SF-36, the 36-item Short Form Health Survey scores; MCS, Mental Component Scores; PCS, Physical Component Scores

Full size image

The statistical significance of ODI, SRS-22 total and SF-36 MCS and PCS scores was also observed among SVA subgroups with differences as shown in Fig. 2. For SRS-22 subscales, The SRS-pain score was statistically significant with differences in the SVA (0) - SVA (++) groups (p = 0.008). The SRS-function score was statistically significant with differences in the SVA (0) - SVA (++) groups (p = 0.001) and SVA (0) - SVA (+) groups (p = 0.016). The SRS-satisfaction score was statistically significant with differences in the SVA (0) - SVA (++) groups (p = 0.018).

Discussion

Due to the aging population, the prevalence of degenerative scoliosis would be on the rise worldwide. Many clinicians have investigated the radiological reference values of spinopelvic alignment in preoperative ASD patients and developed surgical guideline based on the global and regional parameters [9, 13]. Previous studies mainly focused on the sagittal alignment, but the coronal component had been considered rather little. Up to 30% of postoperative coronal malalignment happened after ASD surgery [14], and its impacts on HRQOLs has not been rigorously studied. Additionally, the existing literature related to coronal malalignment mainly concentrated on global coronal imbalance, however, residual coronal curve after corrective surgery were hardly mentioned. Coexisting lumbar spinal stenosis is quite common in Chinese patients with DLS, probably due to anatomically smaller spinal canal in Chinese population [13]. Radicular symptoms in the legs and feet could be remarkably alleviated after adequate decompression without complete correction of regional angulation. Therefore, patients did not routinely have complete correction to perfect straight in return for less surgical trauma, shorter operation time and less blood loss. But the impact of resulting residual angulation on postoperative HRQOL was unknown. In this study, DLS patients were divided into three groups by postoperative CA and CVA, respectively. Considering sample size, preoperative severity of scoliosis and corrective effect of surgery, we selected 10° and 20° as the cut-off values of postoperative CA in DLS patients. The results showed that VAS back pain was significantly lower in CA 10°∼20° group compared with CA < 10° and CA > 20° groups. Furthermore, the statistical difference in other HRQOLs also displayed an underlying trend that DLS patients would present better clinical outcomes when their postoperative residual CA was 10°∼20° (Fig. 2).

In clinical practice, many spine surgeons believed that regional angulation and linear displacement should be restored to the possible minimum, strictly realigning a straight spine in coronal plane. However, this view was not entirely applicable to DLS patients. The findings of this study suggested that a certain degree of remnant Cobb angle after surgery was perfectly acceptable, even better than no residue. This finding could be reasonably explained. On the one hand, structural support by internal fixation stabilized the local lumbar spine and decreased the muscular work, improving patients’ functional ability even if regional angularity remained over 10 degrees. On the other hand, extensive osteotomy for total correction not only would lead to greater iatrogenic injury and extended rehabilitation time, but also possibly aggravate postoperative coronal imbalance particularly for patients imbalanced towards convexity [15], which might be the potential reason for slight outperformance in group (+) relative to group (0). Furthermore, long duration of truncal inclination resulted in correspondingly reciprocal changes in spinopelvic structures as well as lower limbs [16]. However, due to the deterioration of compensatory capacity in the elderly, the long-adapted adjacent structures could not immediately respond to the postoperative curvature changes. Therefore, a certain residual angle might be more acceptable to these patients. But patients with residual CA over 20 degrees manifested significantly worse performance in VAS back pain, ODI, SF-36 PCS and SRS-22 Total scores compared with other two groups. This implied that under-correction with more than 20° residual CA in DLS patients might be intolerable even if the spine was fused. These findings could facilitate surgeons in the preoperative planning by properly correcting the main Cobb angle to improve the postoperative HRQOLs. Furthermore, the study reinforced the necessity for surgeons to realize “how much can patients tolerate?”, rather than harshly pursing ideal alignment [17].

Based on the work of Louis et al. [11] and Bao et al. [18], 2 and 3 cm were defined as the cut-off values of CVA. However, unlike previous studies, no significant difference was observed among the subgroups in all HRQOLs. Similar to SVA, CVA represents the global balance status in coronal plane. Bao et al. [18] reported that 34.8% of DLS patients had coronal imbalance, and those with a C7PL shifted to the convex side of the curve may predispose to postoperative coronal imbalance after posterior osteotomy. In the recent study of Louis et al. [11], the prevalence of postoperative coronal imbalance was reported as 26% in ASD patients after long-level fusion, and the patients with marked coronal imbalance showed worse ODI and SF-36 PCS. Based on their conclusions, the prevalence of coronal imbalance (CVA > 3 cm) in fixed spine was 19.8% in our study. However, no significant difference of HRQOL was found among subgroups of CVA. We speculated that the possible cause of discrepancy was inclusion criteria of all subtypes of ASD in their research. Severe coronal imbalance occurred less frequently in DLS patients compared to other types of ASD. Although the result did not reach statistical significance, patients with CVA > 3 cm still showed a tendency of lower SRS-22 and higher ODI scores than other two groups in our data.

It is commonly accepted that ASD patients benefit from sagittal balance restoration [19, 20]. The SRS-Schwab classification was a practical criterion for ASD, proved helpful for making an osteotomy proposal and predicting clinical outcomes [21]. In this study, we also investigated the sagittal modifiers including PT, PI-LL and SVA in the long-level fixed spine. Restoring an SVA < 50 mm in degenerative scoliosis patients has been suggested for ideal postoperative outcomes [10]. In accordance with past reports, the present study demonstrated that group (0) reported lower ODI scores than other two groups, and both the SF-36 and SRS-22 scores were statistically different between group (0) and group (++). The results also confirmed that postoperative SVA significantly correlated with ODI, SF-36 and SRS-22 subscales. For the reasons above, adequate reconstruction of SVA might be suggested.

Though several past research had underlined the clinical value of keeping PT and PI-LL within proper range [22, 23], this study presented no significant difference between subgroups for all postoperative HRQOLs. Similar to our results, Srikanth et al. [24] found that postoperative PT ≥ 20° or PI-LL > 10° do not indicate worse clinical results after lumbar decompression and fusion surgery by evaluating postoperative SF-12, ODI, and VAS for back and leg. These findings implied that for DLS patients after long-level fusion, postoperative PT and PI-LL might not reliably predict worse HRQOLs. There are two possible reasons contributed to such results. First, due to the anatomical characteristics of Chinese ethnic, setting PT ≥ 20° and PI-LL > 10° as the thresholds for postoperative sagittal imbalance was not quite optimal [25]. Significant variations were observed between the Chinese population and Japanese, American, Caucasian populations in terms of PI and LL in normal posture [26,27,28]. A recent study pointed out that asymptomatic Chinese volunteers had smaller PI (47.4°±10.4°), compared to normative PI of the Americans (52.1°±10.2°) [29, 30]. Thus, such criteria possibly resulted in false negatives. Second, to our knowledge, the lumbar lordosis decreases due to degenerative factors and then the PI-LL increases, therefore the body needs to activate various compensatory mechanisms to avoid leaning the trunk forward, of which retroverted pelvis plays a key role. The SVA, which represents overall balance of the trunk, is often significantly restored after orthopedic operation, and limbs function and quality of life were considerably improved, therefore the local sagittal imbalance such as residual PI-LL mismatch and large PT may not be sufficient to affect patients’ subjective feelings postoperatively.

Admittedly, several weaknesses exist in our study. First, the inherent flaws of retrospective study at a single center are inevitable, and the universality of our conclusions is limited. Second, the minimum follow-up time is relatively short, and the further progress of deformity would be neglected. The next step will conduct prospective studies with larger sample sizes and longer follow-up period. Third, we subjectively chose cut-off values for postoperative CA and CVA based on the literature and clinical experience, so further research is necessary to calculate more precise threshold values of significantly worse postoperative HRQOL in DLS patients. Fourth, this study did not account for potential confounders that could affect residual angularity and post-surgical quality of life, such as the curve types, severity of preoperative deformities or the length of the spinal fusion. Consequently, a more detailed exploration with a larger dataset is necessary to fully understand the impact of postoperative residual Cobb angles on patient quality of life.

Conclusion

Restoring global alignment to normal range is important for DLS patients, but regional malalignment also deserves more attention than it currently does, especially for the common issue of residual Cobb angle. Overlarge postoperative Cobb angle, more than 20 degrees in our study, could result in worse HRQOL and was recommend avoiding during surgery. Notably, on the other hand, 10° to 20° residual Cobb angle was acceptable in DLS patients and even better than Cobb angle < 10° in several HRQOLs, therefore strictly pursing upright coronal alignment seems unnecessary. Postoperative SVA also showed effectiveness in assessing HRQOL, whereas PT and PI-LL might be less reliable and stable in the long-level fusion spine.