Severe Disc degeneration is associated with lower caudal subendplate hounsfield unit values in degenerative lumbar spinal stenosis

Abstract

Study design

Retrospective radiological analysis

Objective

To analyze the distribution of computed tomography (CT)-based Hounsfield unit (HU) values of lumbar vertebrae and explore the correlation between HU values and the modified Pfirrmann grades in patients with degenerative lumbar spinal stenosis (DLSS).

Method

A total of 82 DLSS patients were divided into two gender-based groups for this retrospective analysis. Demographic data, including age, gender, body mass index, medical history, smoking status, operative levels, bone turnover markers (BTMs), and bone metabolism indicators, of all the patients were collected. HU values of the L1-midline, L2-midline, L1 lower subendplate, upper and lower subendplates of L2-L5, and S1 upper subendplate were obtained based on lumbar CT in DLSS patients. The modified Pfirrmann grading system was used to classify the disc degeneration observed in the magnetic resonance imaging (MRI) of DLSS patients.

Results

The demographic data, including age, BTMs, bone metabolism indicators, and HU values of the L1- and L2-midlines, showed no significant differences between the male and female DLSS patients. The L1-midline HU value was positively correlated with the Ca levels and negatively correlated with the β-CTX levels in both male and female DLSS patients. The L2-midline HU value was positively correlated with all the lumbar HU values in male DLSS patients. The Pfirrmann grades of all the lumbar discs showed a significant correlation. Additionally, the Pfirrmann grades of L1-L2 to L5-S1 discs were higher in males than in females, but the difference was not statistically significant. Moreover, the Pfirrmann grades showed an increasing trend with age in all lumbar discs, especially in the L1-L2, L2-L3, and L3-L4 segments. The L3-L4 and L5-S1 Pfirrmann grades were positively correlated with the HU values of the corresponding adjacent vertebral subendplates. The differences in the HU values of the adjacent vertebral subendplates were negatively correlated with severe intervertebral disc degeneration (IDD), compared to mild-to-moderate IDD, especially in Pfirrmann grade VIII cases.

Conclusion

The Pfirrmann grades of the lumbar discs were positively correlated in DLSS patients and showed an increasing trend with age, especially in the L1-L2, L2-L3, and L3-L4 segments. Additionally, the HU values of the adjacent vertebral subendplate were positively correlated with the Pfirrmann grades of the L3-L4 and L5-S1 discs. Relatively lower HU values of the caudal subendplate was associated with severe IDD, and bone mineral density(BMD) in the caudal subendplate correlated with preserving disc height.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-026-09833-7.

Introduction

With the increase in population aging, degenerative lumbar spinal stenosis (DLSS) has become one of the most prevalent orthopedic diseases, threatening the quality of life and physical health of middle-aged and elderly people. DLSS is often accompanied by severe intervertebral disc degeneration (IDD) and bone mass loss.This is consistent with reports that severe disc degeneration in the lumbar spine is closely associated with endplate changes and fatty infiltration in the multifidus and erector spinae [1], highlighting the multifaceted nature of spinal degeneration. However, the association between IDD and bone mass, especially in DLSS patients, is poorly understood. Several clinical and cadaveric studies, conducted between 1998 and 2022, found an association between IDD and osteoporosis; however, the nature of this association has not yet been established [2]. IDD is primarily caused by insufficient nutrient supply and biomechanical stress. An intervertebral disc acquires nutrients through the endplate of the adjacent vertebrae. However, bone mass increase or loss around the subendplate area can disrupt the nutrient supply to the disc, hastening IDD. DLSS patients with severe IDD often opt for operative treatment. However, it is essential to determine the bone quality and its association with IDD in the affected region to select the appropriate internal fixation method for surgery. In addition, the study of the association between bone mass and IDD can improve our understanding of the local biomechanical environment and pathophysiological mechanisms of IDD.

Several methods have been used to evaluate the bone mass. Among these, dual-energy x-ray absorptiometry (DXA) is considered to be the gold standard for measuring bone mineral density (BMD); however, DXA results of elderly people are often inaccurate due to the presence of lumbar osteophytes [3]. Quantitative computed tomodensitometry (qCT) demonstrates superior ability in distinguishing between cancellous and cortical bones, providing more accurate BMD results in elderly patients with scoliosis, obesity, aortic calcification, or osteophytes. In qCT, cross-sectional images of the bones are obtained through computed tomography (CT) scans, which are then analyzed using professional phantoms and software to accurately measure bone mineral content and bone density values. Although qCT elevates the diagnostic accuracy of osteoporosis, the need for additional phantoms and analysis software limits its popularity and promotion in practical applications. In comparison, the Hounsfield unit (HU) measurement of lumbar CT has been recognized as a useful technique for assessing bone mass, as it does not require any additional examinations. Ullrich BW et al. [4] found that CT-derived HU measurement exhibited high inter-rater reliability in assessing bone quality and that it was closely related to qCT-derived BMD in the native scans of traumatic thoracolumbar fracture. Similarly, vertebral CT-based HU measurements have been found to be correlated with DXA-derived T-scores [5]. A study reported that HU values can be used to detect undiagnosed spinal osteoporosis in patients with lumbar degenerative diseases, such as DLSS [6]. HU values vary based on the level and position of the measurement [6, 7].

Pfirrmann grading system is a prominent classification strategy for IDD, and it exhibits excellent intra- and inter-observer agreement [8]. In 2007, Griffith JF et al. [9] proposed a modified Pfirrmann grading system with eight grades (I–VIII) to better discriminate the severity of disc degeneration in the elderly population. In the modified Pfirrmann grading system, grades I–V were defined as “mild-to-moderate IDD” with normal disc height, while grades VI–VIII were defined as “severe IDD” with a loss of disc height ranging between < 30% and > 60%. The modified Pfirrmann grading system [9] also showed excellent intraobserver agreement (weighted kappa range: 0.79–0.91) with substantial interobserver agreement (weighted kappa range: 0.65–0.67).

This study aims to analyze the distribution of CT-based HU values in lumbar vertebrae and explore the correlation between HU values and the modified Pfirrmann grades in patients with DLSS.

Methods

Patients

Patients who had opted for lumbar decompression and fusion surgery at the Department of Orthopedics in Shanghai Pudong New Area Peoples’ Hospital (Shanghai, China) from April 2023 to April 2024 were considered for this analysis. Among these, only the patients who were diagnosed with DLSS; those who were subjected to magnetic resonance imaging (MRI) and CT examination; and those who had completed bone metabolism indicator and bone turnover marker (BTM) testing at admission were enrolled in this study. Meanwhile, patients with hyperthyroidism, Cushing’s syndrome, multiple myeloma, chronic renal failure, significant spinal deformity, or Paget’s bone disease; patients subjected to long-term hormonal, chemotherapeutic, or aromatase inhibitor therapy; patients who underwent anti-osteoporosis treatment; patients with a history of fragility fractures; patients with a history of lumbar spinal surgery; or patients with Modic changes or a high-intensity zone in MRI were excluded from this study. Subsequently, a correlation analysis was conducted between the demographic data, BTM, bone metabolism indicators, lumbar vertebrae HU value, and lumbar disc degeneration grades of the enrolled patients. Informed consent was obtained from all patients. This study was approved by the institutional ethics committee of Shanghai Pudong New Area People’s Hospital (approval number: prylz2023-024) and conducted following the Declaration of Helsinki.

Demographic data

Demographic data, including the age, gender, body mass index (BMI), medical history, smoking status, and operative levels, of all the patients were recorded. To explore the correlation between BTMs, bone metabolism indicators, HU value, and disc degeneration grades from L1-2 to L5-S1, all the patients were classified into two groups based on gender. All the patients received blood tests at admission to determine their BTM (total procollagen type I N-terminal propeptide [TPINP], N-MID osteocalcin [N-MID-OC], and beta-crosslaps [β-CTX]) and bone metabolism indicator (calcium [Ca], phosphorus [P], and alkaline phosphatase [ALP]) levels.

Measurement of lumbar CT-based HU values

Following the method described by Schreiber JJ et al. [10], HU values were recorded from the axial CT scans of the midlines of L1 and L2 vertebrae, lower subendplate of L1, upper and lower subendplates of L2-L5 vertebrae, and upper subendplate of S1 vertebrae. CT scans were performed using a 320-slice multi-detector CT scanner (Siemens Sensation 320, Erlangen, Germany) at 1.25-mm slice thickness and 120 kV tube voltage. The HU measurements of L1–S1 vertebrae were recorded independently by a clinician and a radiologist using the Picture Archiving and Communication Systems (PACS) software. The largest possible elliptical region of interest was drawn on the axial slice of the index vertebrae, excluding the cortical margins and basivertebral vein foramen to decrease measurement error (Fig. 1).

Fig. 1.

CT scans of lumbar vertebrae illustrating the method of determining the HU values. (A) HU value measurement position. The red line indicates the LI-midline, the yellow line indicates the L1 lower subendplate, and the green line indicates the L2 upper subendplate. (B–D) Axial planes of L1-midline (B), L1 lower subendplate (C), and L2 upper subendplate (D) for HU value measurement. The largest possible elliptical regions of interest were drawn on the axial slices of the three regions for HU measurement

Determination of the modified Pfirrmann grades for lumbar IDD by MRI scan

Lumbar disc MRI scans were performed using the 3.0 Tesla MRI Scanner (Siemens, Germany) prior to or after admission. The modified Pfirrmann grades were determined independently by a clinician and a radiologist based on the intervertebral disc signal and intervertebral space height on T2-weighted sagittal lumbar MRI of the L1-L2 to L5-S1 discs.

In the modified Pfirrmann grade [9], Grades I, II, and III are based on the signal intensity of the nucleus and inner fibers of the annulus; in Grade I, the disc is uniformly hyperintense, equal to cerebrospinal fluid (CSF); in Grade II, the disc is hyperintense (> presacral fat and < CSF); in Grade III, the disc is hyperintense (although < presacral fat); in Grade IV, the margins between the inner and other fibers of the annulus at the posterior margin of the disc are indistinct; in Grade V, the disc is uniformly hypointense with no loss of disc space height; and in Grades VI, VII, and VIII, there is a progressive loss of disc space height. These grades are broadly classified as mild, moderate, or severe based on the loss of disc space height.

Statistical analysis

Data were analyzed using the Statistical Product and Service Solutions software (v29.0; SPSS, Chicago, IL, USA). Continuous variables, such as age, BMI, BTMs, bone metabolism indicators, and HU, were recorded as the mean ± standard deviation. Categorical variables, including gender, past medical history, smoking status, and operative levels, were expressed as frequency or percentages. The modified Pfirrmann grades were defined as ordinal scale and presented as the Median and Interquartile Range (IQR). The inter-rater agreements between the independent researchers evaluating the HU value and the modified Pfirrmann grades were assessed with the intraclass correlation coefficient (ICC). Independent samples t-test was used to compare the continuous variables between the two gender groups, and analysis of variance was used to compare continuous variables by age from low to high. The chi-squared test was used to analyze the categorical variables between the two groups. A paired samples t-test was used to compare the HU values of two different vertebrae. The correlation between the demographic data, BTMs, bone metabolism indicators, HU values, and the Pfirrmann grades of the disc was assessed using Spearman’s correlation analysis. The Mann-Whitney U test (for comparisons between two groups) and the Kruskal-Wallis test (for comparisons more than two groups) were used to compare the Pfirrmann grades. A p-value < 0.05 was considered statistically significant.

Results

Characteristics of the subjects

A total of 82 DLSS patients, aged 35–92 years (65.40 ± 1.62 y), were included in this study. Among these, 42 (51.2%) were male and 40 (48.8%) were female. The demographic data of the two gender groups are presented in Table 1. As shown in Table 1, there were no significant differences between the BTM (TPINP, N-MID-OC, and β-CTX) and bone metabolism indicator (Ca, P, and ALP) levels in the two groups. Additionally, the HU values (ICC: 0.88; 95% confidence interval [CI]: 0.86 − 0.90) and the modified Pfirrmann grades (ICC: 0.90; 95% CI: 0.87 − 0.92) showed excellent intra-rater reliability between the two independent clinicians.

Table 1.

Demographic data of the subjects

HU values of the lumbar vertebrae

The female patients demonstrated lower HU values of the L1 and L2 midlines compared to the male patients; however, the difference was not statistically significant. Additionally, the two groups showed significant differences in the HU values of the upper subendplates of L2 and L3 (Table 2). Moreover, compared to the female patients, the male patients showed higher HU values from the midline of L1 to the upper subendplate of L5 but lower HU values in the lower subendplate of L5 and the upper subendplate of S1 (Fig. 2). The HU values of the L1 midline and the other vertebrae showed significant differences in individual patients (p < 0.001, Table S1). In male patients, the L1 midline HU value was positively correlated with the L2 midline HU value (r = 0.490, p < 0.001), while the L2 midline HU value was positively correlated with the HU values of all the lumbar vertebrae (Table 3). Meanwhile, in female patients, the L1 midline HU value was positively correlated with the HU values of all the lumbar vertebrae, including the L2 midline (Table 3). Additionally, both the L1 and L2 midline HU values were positively correlated with the Ca levels (r = 0.388, p = 0.015 and r = 0.340, p = 0.034, respectively), while the L1 midline HU values were negatively correlated with the β-CTX levels (r = − 0.227, p = 0.049) in the two gender groups.

Table 2.

HU value of the lumbar vertebrae between male and female

Midline, transverse HU value of one-half height of the vertebrae in the sagittal plane; LSEP, transverse HU value of the lower subendplate of the vertebrae in the sagittal plane; and USEP, transverse HU value of the upper subendplate of the vertebrae in the sagittal plane

*p < 0.05

Fig. 2.

Average HU values from L1 to S1 for male and female DLSS patients. Compared to females, males showed higher HU values from the L1-midline to the L5 upper subendplate and lower HU values in the L5 lower and S1 upper subendplates

Table 3.

Spearman correlation results of HU value in different vertebrae for male and female respectively

Midline, transverse HU value of one-half height of the vertebrae in the sagittal plane; LSEP, transverse HU value of the lower subendplate of the vertebrae in the sagittal plane; and USEP, transverse HU value of the upper subendplate of the vertebrae in the sagittal plane

*p < 0.05, **p < 0.001

Modified Pfirrmann grades for lumbar disc degeneration

A total of 410 lumbar discs were analyzed in 82 patients using the modified Pfirrmann grading system. Among the 410 discs, 30, 94, 96, 92, 66, and 32 were classified as Grades III, IV, V, VI, VII, and VIII, respectively. There was no significant difference between two groups for the modified Pfirrmann grades (Table 4). Additionally, the Pfirrmann grades in all the lumbar segments, especially the L1-L2, L2-L3, and L3-L4 segments, showed an increasing trend with age (Table 5).

Table 4.

Pfirrmann grades between male and female

MRI(L1-2)	Male	Female	statistics	P value
	4(2)	5(2)	-1.115	0.265
MRI(L2-3)	5(2)	5(2)	-0.992	0.321
MRI(L3-4)	5(2)	5(1)	-0.897	0.370
MRI(L4-5)	6(2)	6(2)	-0.904	0.366
MRI(L5-S1)	5(2)	5(3)	-1.227	0.220

Table 5.

Pfirrmann grades trend with aging

MRI(L1-2)	age ≤ 50	50< age ≤ 60	60< age ≤ 70	age>70	statistics	P value
	3(0.75)	4(1.5)	4.5(2)	5(2)	23.408	< 0.001*
MRI(L2-3)	3(0.75)	5(2)	5(2)	5(1.75)	13.884	0.003*
MRI(L3-4)	3.5(1)	5(2)	6(2)	5(1)	13.426	0.004*
MRI(L4-5)	6(0.75)	6(1.5)	6(2)	6(2)	0.402	0.940
MRI(L5-S1)	5.5(1.75)	5(3)	5(3)	5(2.75)	0.640	0.887

*p < 0.05

Correlation between the HU values and modified Pfirrmann grades

Age was positively correlated with the Pfirrmann grades of the L1-L2 (r = 0.479, p < 0.001), L2-L3 (r = 0.291, p = 0.008), and L3-L4 (r = 0.342, p = 0.002) discs. The difference in the HU values of the adjacent vertebral subendplates (calculated by subtracting the HU value of the axial slice inferior to the upper endplate of caudal vertebrae from the HU value of the axial slice superior to the lower endplate of cranial vertebrae) showed no significant correlation with the severity of IDD in the corresponding positions (Table 6). However, there was a significant positive correlation between the Pfirrmann grades of all the lumbar discs, except for L5-S1 (Table 6). Additionally, the L3-L4 Pfirrmann grades were positively correlated with the HU values of the axial slice superior to the lower endplate of L3 (r = 0.289, p = 0.009) and inferior to the upper endplate of L4 (r = 0.260, p = 0.019), and this correlation was more prominent in the female patients (r = 0.411, p = 0.009 and r = 0.446, p = 0.004, respectively). Moreover, the L5-S1 Pfirrmann grades were positively correlated with the HU value of the axial slice superior to the lower endplate of L5 (r = 0.229, p = 0.04) and inferior to the upper endplate of S1 (r = 0.235, p = 0.035). The differences in the HU values of the adjacent vertebral subendplates were primarily negative in cases with severe IDD (− 0.94 ± 31.63), especially in Pfirrmann grade VIII cases, compared to cases with mild-to-moderate IDD (3.66 ± 12.38) (p = 0.030, Fig. 3; Table 7).

Table 6.

Spearman correlation between the difference of HU value and Pfirmann grades

HU value (Ln+1-Ln), the difference between the HU value of the upper subendplate of Ln+1 and theHU value of the lower subendplate of Ln. *p < 0.05 and **p < 0.001

Fig. 3.

Differences in the HU values of adjacent vertebral subendplates in patients with Pfirrmann grades III–VIII. The differences in the HU values of adjacent vertebral subendplates were negatively correlated with severe IDD, especially in patients with Pfirrmann grade VIII

Table 7.

Comparison of differences in HU value differences between mild-to-moderate and severe IDD

Mild-to-moderate IDD, Pfirrmann Grades I V; and severe IDD, Pfirrmann Grades VI VIII. *p < 0.05

Discussion

Distribution and influencing factors of lumbar HU values

Being an opportunistic osteoporosis screening method, CT-based HU value measurement has been widely applied for the evaluation of bone mass without additional examinations [11]. A few studies reported that HU value measurement has high diagnostic and predictive value for osteoporosis [7, 12, 13], promoting its clinical application. Hiyama A, et al. [14] showed that the lumbar and femoral neck T-scores were significantly associated with the HU values in patients with lumbar degenerative disease. Since the L1 vertebra is scanned in both chest and abdominal CT, previous studies have selected the L1 midline HU value to evaluate the lumbar bone mass. In this study, the HU value of the L1 midline was positively correlated with the HU values of the other vertebrae in female patients, suggesting that it can be used for predicting bone mass in female patients. In contrast, the HU value of the L2 midline was positively correlated with the HU values of the other vertebrae in male patients, suggesting that the L2 midline may be a better target for evaluating bone mass in male patients. The HU values of the L2 and L3 vertebrae were significantly different between the male and female patients. Moreover, the HU values of the upper subendplates of the L1–L5 vertebrae were lower in female patients than in male patients. These differences in the lumbar HU values in the male and female patients may be attributed to the differences in the spinopelvic parameters of the two genders. Qi H et al. [15] demonstrated the presence of significant gender-based differences in pelvic tilt, lumbar lordosis (LL), and thoracic kyphosis (TK) in an asymptomatic pediatric population that consisted of 112 males and 105 females, aged between 4 and 15 years (average of 12.19 years). Higher LL and TK values in males are associated with more biomechanical stimuli in their thoracolumbar and lumbosacral region, which may lead to variations in the distribution of lumbar HU values between the genders; however, the underlying causes of these variations need to be determined by comprehensive biomechanical studies. Furthermore, the increased occurrence of osteoporosis among women might result in lower HU values in them. Our study found that the L1 midline HU value was positively correlated with the Ca levels (r = 0.388) in the two groups, suggesting that serum Ca ions may be essential for the maintenance of vertebral bone mass. Previous studies demonstrated that BTMs can be used to monitor the efficacy of anti-osteoporosis drug therapy; however, BTMs have low diagnostic and predictive value for osteoporosis [16, 17]. Being a major degenerative disease in the aging population, DLSS is often associated with the wear-and-tear damage caused by arthritis. In arthritis, both bone formation and absorption are active, which can influence the BTM levels. Our study showed that the L1 midline HU values were negatively correlated with the β-CTX levels (r = − 0.227) in the two groups, indicating that β-CTX can be used to predict vertebral bone mass in DLSS patients.

Distribution of Pfirrmann grades in the lumbar disc degeneration

In this study, none of the DLSS patients were evaluated as Pfirrmann Grade I or II. Our study found that the highest Pfirrmann Grades were recorded for the L4-L5 discs, consistent with the severe IDD observed in this segment. Consistent with the results of previous studies [18, 19], our study found that the Pfirrmann Grades of the L1-L2, L2-L3, and L3-L4 discs were higher in older populations in general. Additionally, our study found a significant correlation between the Pfirrmann Grades of the lumbar intervertebral discs of each segment, suggesting a consistent convergence of IDD across the segments in DLSS patients.

Correlation between lumbar HU values and IDD

Compared to the DXA examination, CT-based HU measurements could reveal the BMD around the disc. In this study, we calculated the differences in the HU values of the adjacent vertebral subendplates and assessed their association with disc degeneration in the corresponding segments. Our results showed that the differences in the adjacent vertebral subendplate HU values were not correlated with the IDD in the corresponding segments, indicating that the asymmetric bone mass distribution between the adjacent vertebral subendplates does not contribute to IDD. Further correlation analysis showed that the HU values of the adjacent vertebrae subendplates were positively correlated with IDD in the L3-L4 (r = 0.289, p = 0.009 and r = 0.260, p = 0.019) and L5-S1 (r = 0.229, p = 0.04 and r = 0.235, p = 0.035) discs.

First, the endplate pathway is the main source of nutrition for the disc [20]. The subendplate is rich in arteries and veins, with numerous branches intertwined at the subendplate and endplate junction [21]. These branches pass through the bone marrow contact channels, creating a vascular network that supplies nutrients to the disc. An increase in BMD within the subendplate can obstruct the channels, blocking the nutrient supply to cells and ultimately contributing to IDD. Second, Tosun O et al. [22] suggested that discs with dense subchondral bone might endure increased mechanical stress under the same loading conditions. Seol D et al. [23] proposed that osteoporotic bone could serve as a cushion, mitigating stress on the cartilage and thereby slowing the progression of disc degeneration. These reports indicate that denser bone in the subendplate region exerts greater biomechanical forces on the disc, ultimately exacerbating IDD. Therefore, we hypothesized that the increase in BMD in the subendplate regions may have contributed to IDD.

In patients with severe IDD, the subendplate HU values of the caudal vertebrae were lower than those of the cranial vertebrae. Therefore, it is postulated that BMD in the subendplate of the caudal vertebrae may play a vital role in maintaining disc height. Our results are consistent with the results of several studies. Upper vertebral endplate fractures are the most common types of spinal fractures. Since endplate fractures disrupt the nutrient supply to the disc, rapid IDD with the loss of disc height is inevitable in cases with endplate fractures [24, 25]. The upper endplate of the caudal vertebrae bears more mechanical stress than the lower endplate of the cranial vertebrae to maintain the optimal disc height. When the BMD of the upper endplate of the caudal vertebrae decreases, the upper endplate may no longer be able to sustain the mechanical stress from the cranial vertebrae. This could further cause upper endplate damage, microfractures, and narrowing of the intervertebral space. Su Q et al. [26] found that the disc height and disc T2 signal decrease significantly with time in a rat caudal vertebral body fracture model. Ding Y et al. [27] conducted a single-center, retrospective cohort study to analyze the association between disc degeneration and bone density around the endplate from 569 lumbar segments in 130 patients with lumbar spinal stenosis or disc herniation. Their results revealed that Pfirrmann Grades are correlated with higher bone density in the subendplates, but not in the whole vertebral body. Additional subgroup analysis showed that the Pfirrmann grades are negatively correlated with caudal subendplate bone quality (r = − 0.163, p < 0.05), consistent with our findings. These results indicate that the optimal BMD of the caudal subendplate is crucial for preventing intervertebral space narrowing in patients with DLSS.

In cases with severe IDD with loss of disc height, surgical interventions should focus on pedicle screw implantation strategies of caudal vertebrae to minimize the risk of screw pullout and the protection of the upper caudal endplate during interbody fusion.

There are some limitations to this study. First, this was a retrospective study. Second, the sample size of this study was small and limited to DLSS patients, without any normal controls. Future case-control or prospective cohort studies, incorporating a healthy control group and comprehensive data collection on potential confounders, are essential to validate these preliminary associations and establish causality.Our study builds upon the pre-established validity of HU measurements, and that the key novelty lies in applying this opportunistic metric to investigate associations with disc health. Future studies that directly correlate DXA or qCT measurements with Pfirrmann grade in a dedicated cohort are warranted to further solidify the scientific foundation of the relationship between bone quality and disc degeneration suggested by this initial, cost-effective approach. Due to concerns over statistical power—age-stratified subgroups would have had insufficient sample sizes for reliable inference, our original analysis did not include age stratification. The sample sizes within such subgroups were insufficient for robust statistical inference. Therefore, future studies with larger, prospectively recruited cohorts are necessary to validate and explore the potential influence of age on this relationship in greater detail.

This study adds several key aspects to the existing literature on lumbar degeneration. Firstly, while previous research has often focused on the relationship between systemic osteoporosis and disc degeneration, our findings highlight the critical role of localized bone quality at the specific vertebral subendplate—a key interface for disc nutrition and biomechanics. Secondly, by demonstrating that a lower HU value in the caudal subendplate is independently associated with severe disc degeneration in DLSS patients, we provide a more nuanced, site-specific understanding of the bone-disc interaction. This insight moves beyond assessing generalized bone mass and underscores the importance of evaluating the local mechanical environment. Lastly, from a clinical perspective, our findings suggest that preoperative CT scans, which are routinely obtained, can offer valuable quantitative data (HU values) on subendplate bone quality. This information could potentially inform surgical planning, such as the selection of fixation points, and contribute to understanding the pathophysiological mechanisms underlying rapid disc degeneration in DLSS.

In conclusion, our study found that Pfirrmann grades of the lumbar discs were positively correlated in patients with DLSS and showed an increasing trend with age, especially in the L1-L2, L2-L3, and L3-L4 discs. Additionally, the HU values of the adjacent vertebral subendplates were positively correlated with the Pfirrmann grades of the L3-L4 and L5-S1 discs. Furthermore, our results showed that relatively lower HU values of the caudal subendplate was associated with severe IDD and that BMD in the caudal subendplate correlated with preserving disc height in DLSS patients.

Authors’ contributions

Conceptualization: Xin Song, Jian Wang; Methodology: Xin Song, Bangwei Qian; Formalanalysis and investigation: Xin Song, Bangwei Qian, Jingfen Zhou; Writing—original draft preparation: Xin Song; Writing—review and editing: Silu Feng, Jian Wang; Funding acquisition: Xin Song, Jian Wang; Resources: Jingfen Zhou, Donglin Ren; Supervision: Silu Feng, Jian Wang. The author(s) read and approved the final manuscript.

Funding

This study was funded by the Academic Leaders Training Program of Shanghai Pudong New Area Health Commission(PWRd2025-11), The Medical Discipline Construction Program of Shanghai Pudong New Area Health Commission (the Key Disciplines Program ) (PWZxk2022-16), Medical and Health Project of Shanghai Pudong New Area Science and Technology Development Fund for Public Institutions’ Special Research on People’s Livelihood(PKJ2025-Y36), and New Quality Clinical Specialty Program of High-end Medical Disciplinary Construction in Shanghai Pudong New Area(2025-PWXZ-10).

Data availability

All data generated or analysed during this study are included in this published article.

Declarations

Ethics approval and consent to participate

This study was approved by the institutional ethics committee of Shanghai Pudong New Area People’s Hospital (approval number: prylz2023-024) and conducted following the Declaration of Helsinki. Informed consent was obtained from all patients.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.