INTRODUCTION
Degenerative kyphosis (DK) is an age-related spinal disorder that adversely affects health outcomes, encompassing physical function, pain, disability, and mortality risk [1,2]. Sagittal imbalance (SI) is a pronounced feature of DK. SI not only leads to chronic low back pain, and neurological impairment that substantially compromise quality of life [3,4] but also serves as a critical predictor of surgical outcomes and revision risk [3]. This underscores the clinical imperative to identify SI determinants for optimizing therapeutic strategies in DK management.
Paraspinal musculature is located around the spine that mechanically controls spinal movement, and gesture, thus conferring spinal stability [5]. Accumulating evidence suggests paraspinal muscle (PSM) sarcopenia is associated with the extent of spine degeneration and spine malalignment [6]. However, the specific relationship between paraspinal sarcopenia (PS) and SI in DK remains poorly understood. Multiple studies have explored the correlation between PS and spinopelvic alignment in lumbar spinal stenosis (LSS) [6-8], yet DK’s distinct biomechanical profile featuring thoracolumbar kyphosis and gravitational loading alterations may induce unique patterns of muscular degeneration [9]. Notably, while degenerative scoliosis (DS) demonstrates strong correlations between PSM atrophy and coronal imbalance as well as Cobb angle (R=0.59), the association between cross-sectional area (CSA) of multifidus and sagittal vertical axis (SVA) remains weak (R=-0.17) [10,11]. This suggests that the role of PSM degeneration in DS patients is mainly manifested in the coronal plane rather than the sagittal plane. Previous inclusion of patients with deformities may not fully reveal the implication of PS in DK, as they have both kyphosis and scoliosis [12,13]. To the best of our knowledge, limited research has been investigated the correlation between PS and SI in the scenario of DK.
The PS is defined by the loss of muscle mass and impaired function [14]. Previous studies on the PSM majorly evaluate muscle morphological changes, only reflecting muscle quantity and quality [10,12]. While CSA demonstrates linear correlations with muscle strength in healthy populations [15], this relationship becomes inconsistent in degenerative spinal disorders [16-18], particularly in DK where functional impairment may desynchrony with morphological changes. Furthermore, the clinical utility of these labor-intensive imaging analyses remains constrained by practical implementation challenges.
To address this dilemma, growing attention was given to the evaluation of muscle function of the PSM. Previous studies found the maximal voluntary exertion (MVE) and endurance time (ET) of the PSM were reliable physical indexes to directly evaluate muscle function in several contexts [19]. Preliminary findings suggest differential functional correlates: ET associates with pelvic compensation failure in LSS [8], whereas MVE reduction links to SI progression in spinal deformities [13]. Nevertheless, existing studies exhibit limitations including inadequate control Cobb angle in DK populations and incomplete multimodal assessments simultaneously evaluating muscle morphology and function.
We hypothesized that PSM function loss and morphology change have different effects on the development of SI in DK. Objectives of this study were: (1) to investigate the correlation between PS and SI in DK, (2) and to explore the correlation between PSM function and morphology in DK.
MATERIALS AND METHODS
1. Patient Cohort
We recruited patients with DK and LSS who underwent PSM function testing and radiographic assessments at baseline between December 2023 and February 2025.
Inclusion criteria: (1) aged more than 50 years; (2) DK patients who were defined as regional kyphosis (RK) >30° or SVA >50 mm; (3) LSS patients was diagnosed through a combination of clinical history, physical examination, and radiographic changes showing spinal canal stenosis.
Exclusion criteria: (1) patients with scoliosis that had Cobb angle more than 15°, or with kyphosis caused by trauma, osteoporotic vertebral fractures, tuberculosis, and Scheuermann disease or had previous spinal surgery; (2) acute or chronic back pain that could interfere with the evaluation of endurance and MVE.
After enrollment, the DK patients were divided into the DK sagittal balance (SB) group (SVA≤50 mm) and SI group (SVA >50 mm). LSS patients with SB were included as the control group (SVA≤50 mm).
The demographic data including age, sex, body mass index (BMI). Visual analogue scale (VAS) scores for back pain and leg pain, Oswestry Disability Index (ODI) scores were collected at baseline, the Charlson comorbidity index was used to evaluate the fragility [20].
The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and was approved by the Institutional Review Board (IRB) of Nanjing Drum Tower Hospital, The Affiliated Hospital of Medical School, Nanjing University (IRB No. 2021-389-01). All volunteers were fully informed about the methods, purposes, and risks involved in the study protocol and signed the informed consent.
2. PSM Endurance Evaluation
Participants were positioned prone with a pelvic cushion to reduce lumbar lordosis (Fig. 1A). The test required maintaining sternal elevation while maximizing cervical flexion and gluteal contraction to stabilize pelvic alignment. This standardized posture minimized lumbar curvature while coactivating trunk flexor and extensor musculature. Trials were terminated upon meeting cessation criteria: (1) upper-trunk-to-surface angle <5°, (2) 3 consecutive sternal contacts, or (3) participant-reported discomfort. A 5-minute maximum duration was permitted, with 2 trials conducted (5-minute intertrial rest period; best score recorded). The protocol demonstrated favorable safety and testretest reliability [21-24].
3. MVE Evaluation
MVE of PSM was quantified using a calibrated hand-held dynamometer (Micro FET3, Hoggan Health Industries). Participants assumed a standardized prone position with hips/knees neutrally aligned, feet positioned beyond the bed edge, and arms extended with supinated palms. The dynamometer was secured midline between the scapular superior angles (Fig. 1B). Participants performed maximal back extensions with scapular elevation, sustaining isometric contraction for 3–5 seconds. Peak force values were recorded. A protocol comprising one warm-up trial followed by 3 recorded trials (60-second intertrial intervals) minimized learning effects. If the final measurement exceeded prior results by >5%, an additional trial ensured maximal effort capture. The protocol demonstrated satisfactory safety and reproducibility [25-27]. The measured muscle strength values were normalized to body weight, nMVE=MVE/body weight.
4. PSM Morphology Evaluation
The routine lumbar magnetic resonance imaging (MRI) images of enrolled patients was acquired on the routine 1.5-T sigma imaging system (Magetom Skyra, Siemens healthcare, Germany). After scanning, the transaxial and sagittal sequence of T2-weighted images were saved with DICOM format from PACS (picture archiving and communicating system). The sagittal sequence was used to identify the lower endplate of lumbar vertebrae. A total of 7 transaxial slices, which located at T10 to L5 disc levels, were obtained with 4-mm thickness and 2-mm space. Bilateral CSA of PSM (T10–L5) and psoas (L3–4 disc level) was quantified by outlining the thoracolumbar facial boundary at targeted transaxial slices using ImageJ image analyzing software (Image J ver. 1.3; NIH, USA). Fat infiltration rate (FI%) of PSM (T10– L5) and psoas (L3–4 disc level) was obtained by pseudo-coloring technique as previously reported [28,29], as shown in Fig. 2. Relative CSA (rCSA) was calculated as: (total CSA of PSM/CSA of the intervertebral disc)×100; relative FI% (rFI %) was calculated as: [(CSA of PSM at concave side×FI% of PSM at concave side)+ (CSA of PSM at convex side×FI% of PSM at convex side)]/total CSA of PSM [30].
5. Sagittal Parameters Collection
Standing full-spine x-ray examination was conducted for all participants preoperatively. The following parameters were measured in the sagittal plane: Thoracic kyphosis (TK), regional kyphosis (RK), lumbar lordosis (LL), pelvic incidence (PI), pelvic tilt (PT), sacral slope (SS), SVA. The pelvic compensatory potential was evaluated using the PT to PI ratio (PTr).
6. Statistical Analysis
The data is presented as means±standard deviation. The patients were compared in terms of clinical characteristics using analysis of variance with post hoc least significant difference tests. Spearman correlation coefficient was used to assess the correlation between PS and sagittal parameters. The strength of the correlation was evaluated according to established cutoffs: 0–0.3: negligible, 0.3–0.5: low, 0.5–0.7: moderate, 0.7–0.9: high, and ≥0.9: very high. Binary logistic regression was used to identify the risk factors associated with SI. The relationships between PS and sagittal parameters were assessed by calculating correlation coefficients and by multiple linear regression. The predictive capacity of independent risk factors was evaluated using the area under the receiver operating characteristic (ROC) curve. The Youden index was used to find the cutoff maximizing sensitivity and specificity. A p-value of <0.05 was considered to be statistically significant. All statistical analyses were performed using IBM SPSS Statistics ver. 23.0 (IBM Co., USA).
RESULTS
1. Demographics Data
A total of 342 patients with a mean age of 61.1±7.5 years (273 females) were enrolled. The control group included 204 patients (mean age, 59.0±5.2 years; 156 females). The SB group included 61 patients (mean age, 57.4±6.1 years; 52 females) and the SI group included 77 patients (mean age, 65.3±5.0 years; 65 females). SI group were older compared to both SB and control group (p<0.05). There were significant differences in VAS back pain score between the SI group and control group (p=0.042). The ODI was significantly higher in the SI group (p=0.001). Sex, BMI, Charlson comorbidity index, and VAS leg pain score did not differ significantly between the groups (Table 1).
2. Comparisons of the Sagittal Alignment
SS, PI, and LL in the control group were significantly higher than that in the SB group (SS: 32.2°±8.0° vs. 22.5°±11.3°, p< 0.001; PI: 48.8°±7.5° vs. 38.5°±10.4°, p<0.001; LL: 43.8°± 11.4° vs. 34.3°±24.5°, p=0.039). There were significant differences in sagittal parameters except SS between SB and SI group (SB vs. SI, SVA: 20.9±19.1 mm vs. 119.1±58.5 mm, p<0.001; PT: 15.6°±11.4° vs. 27.3°±12.1°, p<0.001; PI: 38.5°±10.4° vs. 45.1°±13.1°, p=0.006; PTr: 0.4±0.3 vs. 0.6±0.3, p<0.001; LL: 34.3°±24.5° vs. 10.9°±26.9°, p<0.001; TK: 38.2°±24.8° vs. 22.3°±24.1°, p=0.001; RK: 52.7±19.3 vs. 45.5±15.9, p=0.040), shown as Table 2.
3. Comparisons of PSM Paraments
SB group showed significantly lower MVE, smaller rCSA (PSM and psoas), and higher rFI% (PSM) than in the control (MVE: 1.7±0.6 N/kg vs. 2.5±0.5 N/kg, p<0.001; rCSA [L1–5]: 207.0±41.2 vs. 241.4±37.2, p=0.004; rFI% (L1–5): 21.0±10.5 vs. 12.6±3.4, p<0.001). SI group revealed multidimensional functional and morphological deterioration of PSM compared to both SB and control group (p<0.05), shown as Table 3.
4. Correlation Between Sagittal Parameters and PSM
Measurements The correlations between sagittal parameters, PSM function, and morphology shown in Table 4. The analyses showed that ET, MVE, rCSA (L1–5) and rFI% (L1–5) were significantly correlated with SVA, PT, and PTr (p<0.01). ET, MVE, and rCSA (L1–5) had significant correlations with SS and LL (p<0.01). Whereas there was no significant correlation between rFI% (L1–5), SS, and LL. Besides, TK was positively correlated with ET (p< 0.01, rho=0.26) and negatively correlated with rCSA (T10–12).
5. Correlation Between PSM Function and Morphological
Parameters Correlations between PSM function and morphology were generally low, shown as Fig. 3. The strongest correlation was found between MVE and ET (rho=0.66, p<0.01), rCSA (L1– 5) was positively correlated with ET (rho=0.42, p<0.01) and MVE (rho=0.34, p<0.05), the rFI% (L1–5) was negatively correlated with ET (rho=-0.43, p<0.05) and MVE (rho=-0.31, p<0.05).
6. Logistic Regression and ROC Curve Analysis
The analysis found low ET, poor MVE, and small rCSA as significant predictors of SI in DK. ROC curve analysis revealed ET demonstrated the strongest predictive value (area under the curve=0.815, p<0.001) (Fig. 4 and Table 5). Youden index analysis determined optimal cutoff thresholds at ET=15.5 S and MVE=1.3 N/kg, achieving maximal sensitivity-specificity balance.
7. Multiple Linear Regression Analysis
In model 1, LL retained a significant correlation with ET and rCSA of L1–5 (p<0.05, R2=0.23). In model 2, PTr was significantly correlated with rCSA of L1–5 (p=0.003, R2=0.21). In model 3, RK was significantly correlated with rFI% of L1–5 (p=0.020, R2=0.19), shown as Table 6.
DISCUSSION
In this study, we found a significant association between progressive PS and SI development in DK, revealing multidimensional functional and morphological deterioration of PSM that exceeds patterns observed in other degenerative lumbar pathologies (Fig. 5). Diminished ET, reduced MVE, and decreased rCSA of L1–5 were found as potential risk factors for SI in DK. Notably, ET and MVE demonstrated differential influences on spinopelvic compensation. These findings collectively suggest that the potential implications of PS in the progression of SI in DK, given insights into the sagittal compensatory mechanisms thus highlighting the therapeutic potential of targeted PSM augmentation strategies.
Comparative analysis revealed that decreased ET, MVE, smaller rCSA, and higher relative fat infiltration (rFI%) in the SI group compared to SB. Recently, Han et al. [8] revealed that the SI with failure of pelvic compensation (SI-FPC) group had lower PSM endurance than the SB group in patients with LSS. Chen et al. [13] found that MVE in the balance group was significantly higher than in the imbalance group in patients with degenerative spinal deformity. Compare our research with their data, we suggest PSM in DK with SI is more severe than other degenerative lumbar diseases in multiple dimensions of function and morphology. This is also supported by the finding that SVA was negatively correlated with ET, MVE, rCSA (L1–5) and positively correlated with rFI% (L1–5) of PSM. This may reflect DK-specific biomechanical demands, wherein chronic compensatory posturing imposes heightened energy expenditure on paraspinal musculature. This persistent muscular strain likely accelerates fatigue-mediated decompensation, ultimately culminating in progressive deformity through insufficient strengthendurance capacity to sustain alignment.
Current morphological assessments of PSM degeneration, including CSA and FI% [31,32], provide limited functional insights despite widespread clinical use [16-18]. PSM strength and endurance are important functional physical indexes to assess muscle condition, but a few studies investigated PSM functional change in DK. Schlaeger et al. [16] investigated the PSM fat fraction based on chemical shift encoding-based water-fat MRI measurements to improve the prediction of PSM strength beyond CSA. Goodpaster et al. [17] found that the loss of muscle mass is associated with the decline in strength in older adults. In our study, we observed unconventional findings indicating a low correlation between the rFI% and MVE (rho=-0.31). These results suggest that the necessity for direct functional evaluation in DK.
In response to positive SB, patients use musculoskeletal compensatory mechanisms to regain SB and horizontal gaze [33]. One of the first and most effective mechanisms recruited is pelvic retroversion. Emerging evidence suggests PSM degeneration may precede LL loss in driving SI progression [32,34], a hypothesis supported by our findings of reduced SS, and LL flattening in SB and SI patients—indicators of advanced PS with concomitant kyphosis. Previous work has highlighted that the gluteus and hamstring muscles are involved in pelvic retroversion [35,36]. One plausible mechanism is that the DK caused the trend of SI, the extent of degeneration in the hamstring muscle may be com-paratively lower than that in the PSM, then pelvic retroversion as compensatory mechanisms, the spine and pelvis will conform to a shape that is successfully stabilized by the available muscular forces. Besides, in DK patients, PT and PTr were negatively correlated with ET, MVE, rCSA (L1–5), and positively correlated with rFI% (L1–5). The binary logistic regression showed loss of ET, MVE, and rCSA to be significant factors that may be influencing SI. This aligns with documented PT-CSA associations in aging populations [37]. It indicates that PS engaged in loss of SVA, simultaneously correlated with pelvic retroversion.
Another interesting finding was ET and MVE exert distinct implications on the spinopelvic alignment. The multiple linear regression analysis showed ET-LL correlation implies endurancedependent maintenance of lumbar curvature, whereas logistic regression analysis showed MVE’s association with SI suggests its role in acute postural correction. This dichotomy proposes 2 SI progression patterns: (1) Decompensation from critical MVE decline (<1.3 N/kg), manifesting as uncompensated trunk inclination; and (2) Chronic imbalance with preserved MVE but inadequate ET (<15.5 seconds), leading to gradual postural deterioration during sustained standing. Such stratification could guide personalized rehabilitation—endurance training for earlystage patients versus strength-focused interventions for advanced cases (Fig. 6).
We hypothesize the potential PSM compensatory mechanisms in DK SB (Fig. 7). First, different mechanisms cause PSM degeneration including oxidative stress, chronic inflammation, decreased protein intake, lack of anabolic hormones and reduction in the number of neuromuscular junctions [38]. With the decrease of FI% and rCSA, the ET weakened, and loss of LL occurred. RK and decreased LL cause the trend of SI, early-stage pelvic retroversion demonstrates transient stabilization capacity, though sustained compensation correlates with elevated energy demands. As DK patients require higher energy expenditure due to muscle activation to maintain compensatory mechanisms, insufficient muscle strength and endurance can make these mechanisms inadequate to maintain alignment, then increasing deformity. With progression of SVA, compensation of TK decrease involved. Finally, the PT increased to limitation, PSM all failure, MVE has weakened to the point where it is not enough to support the trunk.
This study has several limitations. Firstly, the single-center design and limited sample size may restrict generalizability to broader populations and ethnic groups. Secondly, SVA measurements could be influenced by body positioning, pelvic anatomical variations, potentially limiting their reliability in global balance assessment. Thirdly, the cross-sectional nature prevents causal conclusions regarding PS and SI progression, need longitudinal studies for validation in future. Additionally, the exclusion of gluteal and hamstring muscles from this study limits the reliability of explanations regarding compensatory mechanisms in the sagittal plane, particularly in lower limb compensatory aspects. Future studies should incorporate lower limb muscles to provide a more comprehensive understanding of compensatory mechanisms.
CONCLUSION
In conclusion, this study highlights the strong association between PS and SI in DK. Our findings reveal significant PSM dysfunction and morphological changes, emphasizing the importance of combining functional and morphological assessments for diagnosing PS in DK. These insights provide implications fo early identification and targeted interventions to improve patient outcomes.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
Download Citation
Print
