Introduction

Degenerative lumbar scoliosis (DLS) can lead to abnormal curvature in the coronal and sagittal planes, and the concomitant chronic pain and functional disability, subsequently taking a powerful toll on health-related quality of life (HRQOL) [1]. Lumbar fusion surgery was acknowledged as the representative surgical method treating symptomatic DLS, and for severe cases, multi-segmental fusion more than 4 levels were not uncommon, which had been demonstrated to greatly improve health status evaluated by the subdomains of HRQOL measurements [2,3,4].

The ultimate goals of spinal deformity operation in DLS patients were to alleviate pain, reduce disability, and return to daily activities, namely, the overall improvement in HRQOL. Previous studies have already described close correlations between spinal radiographic factors and HRQOLs [5, 6]. Among these studies, sagittal plane alignment was the most notable radiographic factors, which showed strong impacts on various HRQOL measurements, such as Visual Analogue Scale (VAS), the Oswestry Disability Index (ODI), and Scoliosis Research Society-22 (SRS-22) [5, 7, 8]. However, these HRQOLs aimed to evaluate pain-related physical disability and corresponding living quality, but failed to address stiffness-related disability following long segmental posterior instrumentation and fusion surgery. In fact, despite the remarkable clinical benefits from complete decompression and restoration of spinal balance, long-level instrumented fusion may inevitably impose varying degrees of functional limitations due to the loss of thoracic and lumbar motion segments, which were independent of the patients’ pain status. With the increasing maturation of the HRQOL evaluation system, different versions of Lumbar Stiffness Disability Index (LSDI) were designed across countries to provide quantitative evidence for stiffness-related disability after spinal arthrodesis [9,10,11]. Likewise, considering for unique characteristics of elderly Chinese lifestyles, Chinese version of LSDI (C-LSDI) was developed and validated for internal consistency and test-retest reliability [12].

Up to date, studies mainly focused on the association between LSDI and fusion length, involving little about the influence of trunk posture and global alignment after arthrodesis [13, 14]. Furthermore, the interconnections between this new scale and other frequently used HRQOL instruments have not yet been well elucidated. Therefore, the objective of present study were to assess the correlations between C-LSDI and other clinical outcomes among patients with DLS, and to compare the postoperative data between low-stiffness group and high-stiffness group. Furthermore, multiple regression analysis was performed to determine the radiological and clinical factors independently correlated with C-LSDI at final follow-up.

Materials and methods

Patient inclusion

This retrospective, single-center study reviewed the clinical data of a consecutive case series between June 2009 to September 2017 and obtained the approval of ethics committee in Peking University Third Hospital before initiating the study. The inclusion criteria in this study were (1) age ≥ 40 years; (2) clinically diagnosed as degenerative lumbar scoliosis; (3) fusion for four or more levels; (4) available radiological and clinical data; (5) more than one-year follow-up. Exclusion criteria were (1) previous spinal fusion surgery; (2) other types of spinal deformity; (3) undergoing revision surgery for postoperative complications. One-year follow-up was selected based on the recent findings by Glassman et al. [15] that HRQOLs stabilize at one year postoperatively after spinal deformity surgery.

Data collection

Demographic and clinical data included age, gender, body mass index (BMI), follow-up time, fusion levels, lower instrumented vertebrae (LIV), surgical time, total blood loss and hospital stay. Postoperative HRQOLs data included C-LSDI, VAS Back and Leg pain, ODI, SRS-22, Japanese Orthopedic Association-29 (JOA-29) and the Short Form-36 Health Survey Physical Component Score (SF-36 PCS) and Mental Component Score (SF-36 MCS). The following radiological parameters were measured preoperatively and at the final follow-up: Cobb angle (CA), apical vertebral translation (AVT), coronal vertical axis (CVA), thoracic kyphosis (TK), thoracolumbar kyphosis (TLK), lumbar lordosis (LL), pelvic incidence minus lumbar lordosis (PI-LL), pelvic tilt (PT), pelvic incidence (PI), sacral slope (SS), sagittal vertical axis (SVA), T1 pelvic angle (TPA), Global tilt (GT). Patients were divided into two groups by the medium number of postoperative C-LSDI: low-stiffness group (C-LSDI < 48 points) and high-stiffness group (C-LSDI ≥ 48 points).

Statistical analysis

All data were analyzed using SPSS software 27.0 (IBM Corp, USA). Pearson correlation coefficient and corresponding p value were calculated to analyze the clinical relevance between the C-LSDI and clinical and radiological parameters at final follow-up. The strength of correlation was considered as follows: correlation coefficient from 0.8 to 1 (very strong), 0.6 to 0.8 (strong), 0.4 to 0.6 (moderate), 0.2 to 0.4 (weak), and under 0.2 (very weak) [5]. Difference between low-stiffness group and high-stiffness group were assessed using the Student t-test for continuous variables and the chi-square test for categorical variables. Variables with p values < 0.05 were included in the multiple linear regression analysis, in order to determine independent factors associated with high stiffness disability. Significance was established at p value < 0.05.

Result

Overall patient information

In total, 118 patients (96 women and 22 men) were enrolled in the study, with an average follow-up of 34.9 months. The mean age was 62.8 years, and the mean BMI was 25.8 kg/m2. The mean fused levels was 6.2. The LIV was located at the sacral vertebra in 74 patients (62.7%) and the L5 or above in 44 patients (37.3%).

Correlations between C-LSDI and HRQOL scores after surgery (table 1)

Table 1 Pearson’s correlations between C-LSDI and HRQOL scores
Full size table

C-LSDI demonstrated significant correlations with all the HRQOLs. C-LSDI correlated positively with ODI and VAS scores, and negatively with JOA-29 and all subdomains of the SRS-22 and SF-36. Among HRQOLs that indicated moderate associations with C-LSDI, the correlation coefficients of ODI, JOA-29, SRS-22 Function, and SF-36 PCS were over 0.5, followed by VAS Back, SRS-22 Pain, Mental health, and Satisfaction subdomains (0.4 < r < 0.5). In addition, C-LSDI also showed weak correlations with VAS Leg, SRS-22 self-image, and SF-36 MCS (r < 0.4).

Correlations between C-LSDI and radiological parameters (table 2)

Table 2 Pearson’s correlations between C-LSDI and radiological parameters
Full size table

All of preoperative spinopelvic radiological parameters did not show any obvious correlations with C-LSDI. Among postoperative radiological parameters, the coronal parameters including CA, AVT, and CVA determined significant positive correlations with C-LSDI after surgery. As for postoperative sagittal alignment parameters, PI-LL, SVA, TPA, and GT positively correlated with C-LSDI, whereas TK, TLK, LL, PI, PT, and SS did not present significant correlations with C-LSDI. In addition, the correlations between C-LSDI and postoperative CA, SVA, TPA, and GT were weak (0.2 < r < 0.4), and AVT, CVA, and PI-LL showed very weak correlations (r < 0.2). No postoperative change parameter was revealed to have significant correlation, except for the change of PT (r = 0.191) and TPA (r = 0.187).

Difference in demographic and clinical data between the low-stiffness and high-stiffness groups (table 3)

Table 3 Comparison of demographic and clinical data between the low-stiffness and high-stiffness groups
Full size table

Patients in the high-stiffness group had significantly higher BMI (26.52 ± 3.51 vs. 25.05 ± 4.02, p = 0.037), longer fusion length (6.59 ± 1.93 vs. 5.78 ± 1.46, p = 0.011), and more total blood loss during surgery (987.63 ± 498.73 vs. 1441.86 ± 1295.54, p = 0.014). There was no significant difference in age, gender, LIV, surgical time, hospital stay and follow-up duration between the two groups.

Difference in HRQOLs and radiographic parameters between the low-stiffness and high-stiffness groups (table 4)

Table 4 Comparison of the HRQOLs and radiographic parameters between the low-stiffness and high-stiffness groups
Full size table

The high-stiffness group showed significantly worse scores in all HRQOLs compared with the low-stiffness group. In terms of postoperative radiographic parameters, CA, AVT, CVA, PI-LL, SVA, TPA, and SVA were significantly higher in the high-stiffness group than the low-stiffness group. There was no significant difference in TK, TLK, LL, PI, PT, and SS between the two groups.

Multivariate analysis for factors independently affecting C-LSDI (table 5)

Table 5 Multiple linear regression for factors influencing C-LSDI
Full size table

Univariate analyses revealed some potential risk factors including BMI, fusion levels, and total blood loss, as well as several overlarge postoperative radiographic parameters including CA, AVT, CVA, PI-LL, SVA, TPA, and SVA. All the risk factors were incorporated into the following multivariable model. The results of multiple linear regression demonstrated the following independently influencing factors for C-LSDI: postoperative SVA value (β = 0.084, p = 0.025), fusion levels (β = 2.13, p = 0.012), and BMI (β = 0.867, p = 0.022).

Discussion

This study emphasized on stiffness-related disability in DLS patients received long-segment posterior fusion surgery, utilizing the Chinese version of LSDI. This is the first study to offer insight into the association between C-LSDI and a range of HRQOLs frequently used in spinal deformity field. Among HRQOL instruments collected, we noted that C-LSDI demonstrated significant correlations with all HRQOLs, and ODI, JOA-29, SRS-22 Function, and SF-36 PCS were most relevant, with moderate strength of associations. This is logical given the fact that these measurements and C-LSDI were all used to evaluate functional limitations and disability but from different aspects. Such correlations suggested that limitations caused by stiffness could occur in tandem with disability due to low back or leg pain. These clinical outcomes confirmed to actual daily condition, as patients with postoperative pain may intentionally reduce the range of lumbar motion that aggravate their symptoms. This finding provided important clues for the evaluation process of HRQOLs in ASD patients after surgery. It is possible that patients could not precisely distinguish the “stiffness-related disability” and “pain-related disability” when they answer the questions in HRQOLs, so the ODI, JOA-29 and the functional domains of SF-36 and SRS-22 may be affected by issues other than low back pain, such as stiff limitations in trunk flexion. The high-stiffness group showed significantly longer fusion length and worse scores in all HRQOLs. Similarly, studies of Kimura et al. [16] and Lee et al. [17] reported poorer physical function assessed by SF-36 PCS and ODI as the number of fused levels increased. These findings suggested that stiffness-related disability may serve as one of mediating factors between fused levels and pain-related disability, which also support our assumption. In fact, the current HRQOL instruments are still under development and cannot untangle the complicated relationships between pain and stiffness, as well as other influence variables such as psychological factors. Therefore, the emphasis on future research is supposed to develop a more comprehensive postoperative HRQOL measurement tool.

In prior studies, it has been fully stressed that the LSDI after arthrodesis was strongly influenced by the length of fusion [2, 14, 18]. In this study, we not only proved the foregoing conclusion, but also first clarified the interconnections between postoperative radiological parameters and stiffness-related disability, and further confirmed the independent impact of sagittal imbalance. A series of past studies had reported residual positive SVA after correction surgery of ASD may lead to persistent pain and suboptimal HRQOLs [19,20,21]. Mac-Thiong et al. [21] revealed that poor ODI (> 34) was closely related with an SVA greater than 6 cm in patients with degenerative scoliosis. Additionally, according to SRS-Schwab classification, achieving a postoperative SVA < 50 mm in the adult scoliosis corrective surgery has been recommended for optimal postoperative HRQOL [22]. As we speculated, the results of multiple regression analysis proved that SVA also had an independent influence on postoperative C-LSDI. The mean SVA in high-stiffness group was 65.7 mm, higher than the ideal postoperative SVA (< 50 mm) and the value in low-stiffness group (mean SVA = 40.2 mm). Our study revealed the significance of adequate restoration of sagittal balance in postoperative mobility. In addition, other spinal alignment parameters reflecting truncal inclination showed significant associations with C-LSDI as well, such as AVT, CVA, TPA and GT. These findings may be explained in this way: due to C-LSDI comprised of a 12-item questionnaire about activities of daily living (ADL) based on Chinese population lifestyle, such as more household chores, these kinds of ADLs could be bothered by truncal inclination and rigid spinal imbalance after loss of multiple motion segments, and hence patients’ feelings about lumbar stiffness become more intense. Furthermore, the body flexibility was not only decided by the number of vertebra but also paraspinal musculature condition. Since the long-term imbalance position could further stiffen paraspinal muscle and accelerate their degeneration, functional limitation due to stiffness was then aggravated. To our knowledge, the study of Park et al. [23] was the only study to analyze the impact of several radiological parameters on stiffness-related disfunction of spinal deformity disease, however, contrary to their expectations, the sagittal parameters including LL, SS, PT, TK, and SVA did not show a significant correlation with stiffness-related functional disability (SRFD). One reason for such discrepancy was the different enrollment criteria which were all types of ASDs in their research, whereas only DLS patients in our study. Due to different subtypes of spinal deformities, etiology and clinical characteristics could confuse the correlation strengths between radiological parameters and stiffness disability among these subtypes. Moreover, the SRFD was also a modified Korean version of LSDI, which definitely existed difference in the question setting compared with C-LSDI, therefore leading to the discrepancy of results.

It is worth noting that BMI was also identified as a significant influence factor besides SVA and fusion length. Recent research could provide theoretical basis for this conclusion. Linden et al. [24] found that elevated BMI was associated with decreased spine flexibility index for adolescent idiopathic scoliosis (AIS) patients. Additionally, an analysis based on multi-center adult spinal deformity database also showed that increased BMI appeared to be correlated with poorer SRS-22 and ODI scores [25]. Divi et al. [26] observed that patients with higher BMI tended to suffer from worse SF-12 PCS and ODI both before and after lumbar fusion surgery. Hence, it seems reasonable to deduce that higher BMI had an undesirable effect on physical function and flexibility of spinal deformity patients, contributing to postoperative stiffness-related disability.

This study marks the first assessment of the correlations between stiffness-related disability and radiological and clinical outcomes after long-level fusion in Chinese population with the use of C-LSDI. Limitations caused by stiffness should be regarded as one of critical consideration factors in HRQOL. We hope that our findings can provide a better understanding to spinal surgeon preoperatively about the expected effect from loss in available range of motion after spinal arthrodesis. Admittedly, several limitations exist in this study. First, this study is retrospective in nature and conducted in a single-center which is an inherent weakness in the design. More prospective studies will be performed in the future. Second, the male-to-female ratio among DLS patients is relatively high, but we do not perform subgroup analysis by gender which might conceal the impact of gender [23]. Finally, lower extremities parameters and paraspinal musculature condition probably play an important role in the mechanism of postoperative compensation and body flexibility, which are not taken into consideration in the present study.

Conclusion

We revealed that C-LSDI demonstrated significant correlations with all HRQOLs, and ODI, JOA-29, SRS-22 Function, and SF-36 PCS were most relevant with moderate strength of associations. Moreover, the results of study showed that longer fusion levels, higher BMI, and greater postoperative SVA independently affect C-LSDI after long segmental posterior instrumentation and fusion for DLS. These findings will facilitate both surgeons and patients to make a proper preoperative protocol and postoperative evaluation.