Abstract
Background
Intervertebral disc degeneration (IDD) is a progressive and debilitating condition associated with aging, inflammation, and metabolic disorders. Although dyslipidemia has been implicated in IDD pathogenesis, large-scale prospective evidence remains limited. This study aimed to investigate the observational association between serum lipid traits and IDD risk using data from the UK Biobank.
Methods
A prospective cohort study was conducted among 298,226 participants (aged 37–73 years; 54.2% female, n = 161,770) without IDD at baseline. Serum lipid levels, including total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), apolipoprotein A (Apo A) and apolipoprotein B (Apo B), were measured at enrollment. IDD cases were identified via ICD-10 codes. Cox proportional hazards models, adjusted for demographic, lifestyle, and comorbid factors, were used to assess associations between lipid levels and IDD risk. Restricted cubic spline analyses explored potential nonlinear relationships, and subgroup analyses examined effect modifications.
Results
Over a median follow-up of 12 years, 8,745 participants developed IDD. Higher levels of TC (> 5.64 mmol/L) and TG (> 1.51 mmol/L) were significantly associated with increased IDD risk. Compared with the lowest quintile, the highest TC quintile was associated with a 10.2% increased risk (HR = 1.102; 95% CI: 1.029–1.181; P = 0.006), and the highest TG quintile with an 11.3% increased risk (HR = 1.113; 95% CI: 1.036–1.195; P = 0.003), after full adjustment. No significant associations were found for LDL-C, HDL-C, Apo A, or Apo B after multivariable adjustment (all P > 0.05). Subgroup analyses revealed significant interactions between age and TG (P for interaction < 0.05), with younger participants (≤ 60 years) showing a stronger association. Additionally, age modified the effects of HDL-C and Apo A. These findings provide observational evidence that lipid traits may be differentially associated with IDD risk across age groups.
Conclusions
Higher TC and TG levels were associated with increased IDD risk. These findings provide observational evidence for the role of lipid levels in stratifying IDD risk. Routine lipid screening may help identify high-risk individuals and guide early prevention strategies.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12944-025-02687-3.
Keywords: Dyslipidemia, Intervertebral disc degeneration, Serum lipids, Triglycerides, Total cholesterol, UK biobank
Background
Intervertebral disc degeneration (IDD) is an irreversible pathological process characterized by progressive deterioration of the intervertebral disc’s structure and function. Key changes include disorganization of the extracellular matrix (ECM) and tears in the annulus fibrosus, potentially leading to herniation of the nucleus pulposus (NP). Additionally, IDD involves remodeling of the cartilaginous and bony endplates [1, 2]. The degenerative process disrupts the balance between catabolic and anabolic activities within the ECM, triggering immune dysregulation, cellular dysfunction, and loss of NP cells [3]. Furthermore, IDD is closely associated with chronic inflammation, oxidative stress, apoptosis, elevated matrix metalloproteinase activity, and other contributing factors [4]. These pathological changes not only compromise the disc itself and destabilize the spine, but also affect adjacent structures such as the facet joints, vertebral endplates, and paraspinal muscles [5, 6].
Dyslipidemia is characterized by abnormalities in one or more circulating lipid parameters, including total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C), when their levels exceed physiological thresholds [7]. It is a common metabolic disorder worldwide, with its prevalence increasing with age. Dyslipidemia has been associated with various diseases, including type 2 diabetes mellitus, atherosclerosis, ischemic stroke, and cardiovascular disease [8]. The UK Biobank, a large-scale population-based prospective cohort study, was established to systematically investigate the impact of genetic, environmental, and lifestyle factors on a wide range of diseases in middle-aged and older adults. This resource has enabled numerous studies exploring potential associations between dyslipidemia and disease risk. For instance, Xiong et al. [9], using data from 414,302 participants, found that higher plasma levels of apolipoprotein A (Apo A) and HDL-C were associated with an increased risk of fractures and osteoporosis, while elevated levels of apolipoprotein B (Apo B), LDL-C, and TG were linked to a reduced risk, possibly mediated through changes in bone mineral density. Similarly, Fang et al. [10] analyzed data from 380,087 participants and reported that elevated TG levels were associated with an increased risk of cecal and transverse colon cancers, whereas higher Apo A levels were inversely associated with hepatocellular carcinoma risk.
Recent studies have increasingly highlighted the potential link between dyslipidemia and in IDD. Dyslipidemia—characterized by elevated levels of TC, LDL-C, TG, or reduced levels of HDL-C—has been associated with systemic inflammation and oxidative stress [11, 12], which may disrupt disc cell metabolism and promote a pro-inflammatory microenvironment within disc tissues [13]. Understanding this relationship may offer new insights into IDD prevention and treatment. Several case-control and animal studies have investigated the role of specific serum lipid components in the risk and progression of IDD [11, 13]. For example, a retrospective case-control study involving 790 patients reported that elevated TC and LDL-C levels were significantly associated with increased risk of lumbar disc degeneration [11]. Similarly, an experimental study using a mouse model suggested that hyperlipidemia may accelerate disc degeneration through mechanisms involving inflammation of the vertebral endplate [13]. Specifically, hyperlipidemia was found to induce neovascularization, macrophage infiltration, and upregulation of matrix metalloproteinases in the cartilaginous endplate, collectively contributing to extracellular matrix degradation and structural deterioration of the disc. These pathological changes may be further exacerbated by lipid-induced oxidative stress and pro-inflammatory signaling, ultimately disrupting matrix homeostasis and accelerating disc degeneration. Nonetheless, further well-designed studies are warranted to validate these findings and identify potential therapeutic targets.
It is important to note that most existing studies are retrospective and have notable limitations, such as small sample sizes, cross-sectional designs [11, 12], or the evaluation of only a single lipid trait. Moreover, few studies have utilized large-scale datasets to comprehensively investigate the associations between serum lipid traits and IDD. To address these gaps, we systematically evaluated the relationships between multiple serum lipid traits and IDD risk using large-scale prospective data from the UK Biobank. Furthermore, we examined whether these lipid traits were independently associated with an increased risk of IDD among middle-aged and older adults.
Methods
Study design and participants
The UK Biobank is a large prospective population-based cohort study that enrolled over 500,000 participants aged 37 to 73 years between 2006 and 2010. Baseline data were collected through touchscreen questionnaires, interviews, physical measurements, biological sample analyses, genotyping, and electronic health records across 22 research centers in England, Scotland, and Wales. The dataset includes comprehensive clinical, biochemical, and genetic information. Self-reported data on race, smoking status, alcohol consumption, medical history, regular medication use, and comorbidities were verified during interviews to enhance data accuracy. For more information on the UK Biobank, please visit https://www.ukbiobank.ac.uk/. The present analysis uses the most recent dataset available as of October 2024. All participants provided written informed consent, and the study was approved by the Northwest Multi-Center Research Ethics Committee. Data access for this study was granted under UK Biobank application No. 117,320. Ethical approvals for the UK Biobank project were issued by the NHS National Research Ethics Service on June 17, 2011 (Ref: 11/NW/0382) and May 10, 2016 (Ref: 16/NW/0274), covering the data used in this analysis.
In this study, we included 502,186 participants from the UK Biobank who attended the initial assessment visit between 2006 and 2010. The exclusion criteria were as follows: (1) participants who reported IDD at baseline; (2) participants with missing or unavailable International Classification of Diseases, 10th Revision (ICD-10) diagnostic codes; (3) participants without available serum lipid measurements. A total of 298,226 eligible participants were included in the final analysis. The participant flowchart for this study is shown in Fig. 1.
Fig. 1.
Flowchart of the study participants
Exposure variable
In the UK Biobank study, baseline blood samples were collected to measure lipid biomarkers, including Apo A, Apo B, TC, HDL-C, LDL-C and TG. In this analysis, baseline lipid levels served as the primary exposure indicators and were considered highly reliable. (Detailed protocols of sample collection, processing, biochemical analysis using the Beckman Coulter AU5800 platform, automated laboratory management systems, and quality control procedures are provided in the Supplementary Methods [14, 15].)
Outcome variable
The primary outcome of the study was the incidence of IDD. IDD was diagnosed based on the International Classification of Diseases, 10th Revision (ICD-10) codes G55.1, M51, M51.0, M51.1, M51.16, M51.19, M51.2, M51.3, M51.37, M51.8, and M51.9 (ICD-10 codes detailed in Supplementary Methods), and was analyzed separately. Sources of reported IDD cases included death registries, primary care records and hospitalization data.
Assessment of covariates
Baseline covariates were collected using touchscreen questionnaires and included age, sex, race, smoking status, alcohol consumption, physical activity, body mass index (BMI), socioeconomic status (Townsend Deprivation Index), dietary intake (processed meat, oily fish, red meat), relevant comorbidities (e.g., hypertension, diabetes, cardiovascular diseases), as well as medical history, including long-term aspirin use and fractures within the past five years. Physical activity was assessed using the International Physical Activity Questionnaire (IPAQ), and diet via the Oxford WebQ, a validated 24-hour dietary recall tool [16–18]. Details of all assessment and classification methods are provided in the Supplementary Methods.
Follow up times
For participants in the IDD group, the onset date was defined as the date of the first recorded IDD diagnosis, which was compared to the baseline date. The follow-up period ended at this onset date. For participants in the Non-IDD group, follow-up ended at death or loss to follow-up. If neither occurred, the follow-up period was censored at September 1, 2021. This censoring date was also compared with the baseline date.
Statistical analysis
Baseline characteristics were summarized, with categorical variables presented as proportions (percentages) and compared using the chi-square test. Continuous variables with non-normal distributions were expressed as medians (interquartile ranges) and compared using the Mann-Whitney U test. Nonlinear associations between serum lipid traits and disc degeneration were assessed using Cox proportional hazards models with restricted cubic splines. Exposure variables were modeled with penalized cubic splines, a variant of basic splines differing from restricted cubic splines in the number and placement of knots [19]. The median value of each lipid index served as the reference point. The likelihood ratio test was used to evaluate nonlinearity and the overall association between exposure and outcome. In the figures, the overall P value indicates the significance of the association, with Overall P < 0.05 suggesting a statistically significant correlation. The nonlinear P value assesses the form of the relationship, with Nonlinear P < 0.05 indicating evidence of nonlinearity. Exposure was treated as a continuous variable, and the hazard ratio (HR) with its 95% confidence interval (CI) was calculated across serum lipid quintiles using multivariate Cox models, with the lowest quintile as the reference. Three analysis models were considered, with the following adjustments:
Model 1: Adjusted for age and sex.
Model 2: Further adjusted for race, BMI, smoking status, drinking status, Townsend deprivation index, and physical activity (MET-mins/week).
Model 3: Based on Model 2, also adjusted for comorbidities such as hypertension and diabetes, as well as fractured in last 5 years, aspirin use, sleep duration, and frequency of meat intake (e.g., processed meat and beef).
Missing data were handled using multiple imputation by chained equations (MICE), a method suitable for various data types [20, 21]. Detailed information on the proportion of missing data for each variable, along with corresponding imputation diagnostics used to assess imputation quality, is provided in Supplementary Table S1. Subgroup analyses were stratified by age (≤ 60 vs. >60 years), sex, and BMI (≤ 25kg/m2 vs. >25kg/m2). Data cleaning and analyses were performed using R software (version 4.4.1), with a two-sided p-value < 0.05 considered statistically significant.
Results
Baseline characteristics
The baseline characteristics of the participants are shown in Table 1. Table 1 presents the baseline demographic characteristics of the 298,226 participants, categorized by sex, including 136,456 males and 161,770 females. Females accounted for 54% of the sample, with an average age of 58 (51–64) years. Females were more frequently associated with the occurrence of IDD. In addition to TG, females had higher concentrations of other serum lipid traits (such as TC, Apo A, Apo B, HDL-C, and LDL-C) compared to males. Furthermore, females had a lower BMI, smoking rates, and alcohol consumption frequency, and aspirin usage. They also reported lower intake of processed meat and pork, and a lower prevalence of most comorbidities, including hypertension and diabetes. In contrast, females engaged in more physical activity, had a higher incidence of fractures in the past five years, and showed a higher prevalence of certain conditions, such as thyroid disease.
Table 1.
The baseline characteristics of the included participants grouped by sex
| Baseline characteristics | Male(N = 136456) | Female(N = 161770) | Total(N = 298226) |
|---|---|---|---|
| Race | |||
| Non-white | 7174(46.5%) | 8253(53.5%) | 15,427(100%) |
| White | 129,282(45.7%) | 153,517(54.3%) | 282,799(100%) |
| Age (years) | 59(51, 64) | 58(50, 63) | 58(51, 64) |
| Townsend deprivation index | (−3.65, 0.6) | −2.13(−3.63, 0.51) | −2.13(−3.64, 0.54) |
| BMI(kg/m2) | 27.35(25.03, 30.10) | 26.22(23.53, 29.84) | 26.82(24.20, 29.98) |
| Total physical activity(MET-mins/week) | |||
| < 500 | 73,655(46%) | 86,741(54%) | 160,396(100%) |
| ≥ 500 | 62,801(45.6%) | 75,029(54.4%) | 137,830(100%) |
| Diagnosis | |||
| Non-IDD | 132,702(45.8%) | 156,779(54.2%) | 289,481(100%) |
| IDD | 3754(42.9%) | 4991(57.1%) | 8745(100%) |
| Follow.up.time (years) | 12(11, 13) | 12(11, 13) | 12(11, 13) |
| Smoking | |||
| Never | 65,364(40.6%) | 95,553(59.4%) | 160,917(100%) |
| Previous | 53,758(51.1%) | 51,438(48.9%) | 105,196(100%) |
| Current | 17,334(54%) | 14,779(46%) | 32,113(100%) |
| Alcohol intake frequency | |||
| Never | 9021(35.9%) | 16,084(64.1%) | 25,105(100%) |
| Less | 57,939(39.7%) | 87,997(60.3%) | 145,936(100%) |
| More | 35,210(52%) | 32,493(48%) | 67,703(100%) |
| Almost daily | 34,286(57.6%) | 25,196(42.4%) | 59,482(100%) |
| Sleep duration, h/day | 7(6, 8) | 7(7, 8) | 7(6, 8) |
| Apolipoprotein A, g/L | 1.403(1.267, 1.557) | 1.604(1.443, 1.781) | 1.508(1.345, 1.693) |
| Apolipoprotein B, g/L | 1.012(0.855, 1.176) | 1.017(0.867, 1.181) | 1.015(0.862, 1.179) |
| Total cholesterol, mmol/L | 5.437(4.69, 6.191) | 5.794(5.074, 6.556) | 5.632(4.894, 6.396) |
| HDL cholesterol, mmol/L | 1.232(1.059, 1.444) | 1.539(1.313, 1.799) | 1.391(1.167, 1.662) |
| LDL cholesterol, mmol/L | 3.45(2.864, 4.037) | 3.564(3.006, 4.168) | 3.513(2.941, 4.107) |
| Triglycerides, mmol/L | 1.693(1.183, 2.432) | 1.348(0.976, 1.914) | 1.492(1.054, 2.152) |
| Aspirin Use | 26,663(62%) | 16,356(38%) | 43,019(100%) |
| Fractured in last 5 years | 12,097(41.2%) | 17,238(58.8%) | 29,335(100%) |
| Hypertension | 56,239(52.9%) | 50,135(47.1%) | 106,374(100%) |
| Diabetes | 17,672(59.2%) | 12,175(40.8%) | 29,847(100%) |
| Kidney failure | 15,601(55.5%) | 12,485(44.5%) | 28,086(100%) |
| Thyroid disease | 5303(21.6%) | 19,285(78.4%) | 24,588(100%) |
| Cardiovascular disease | 30,098(63.2%) | 17,559(36.8%) | 47,657(100%) |
| Aneamia | 7622(43.1%) | 10,079(56.9%) | 17,701(100%) |
| Liver disease | 5791(51.3%) | 5509(48.7%) | 11,300(100%) |
| Cancer | 9162(37.3%) | 15,439(62.7%) | 24,601(100%) |
| Processed meat intake | |||
| Never | 7438(26.8%) | 20,317(73.2%) | 27,755(100%) |
| Less | 120,135(46.5%) | 138,452(53.5%) | 258,587(100%) |
| More | 7017(74.6%) | 2395(25.4%) | 9412(100%) |
| Almost daily | 1866(75.5%) | 606(24.5%) | 2472(100%) |
| Oily fish intake | |||
| Never | 15,584(47.2%) | 17,445(52.8%) | 33,029(100%) |
| Less | 119,244(45.5%) | 142,989(54.5%) | 262,233(100%) |
| More | 1199(53.9%) | 1028(46.1%) | 2227(100%) |
| Almost daily | 429(58.2%) | 308(41.8%) | 737(100%) |
| Beef intake | |||
| Never | 11,287(33.9%) | 22,061(66.1%) | 33,348(100%) |
| Less | 124,682(47.2%) | 139,408(52.8%) | 264,090(100%) |
| More | 353(63.9%) | 199(36.1%) | 552(100%) |
| Almost daily | 134(56.8%) | 102(43.2%) | 236(100%) |
| Lamb intake | |||
| Never | 19,903(37.5%) | 33,160(62.5%) | 53,063(100%) |
| Less | 116,350(47.5%) | 128,465(52.5%) | 244,815(100%) |
| More | 121(62.1%) | 74(37.9%) | 195(100%) |
| Almost daily | 82(53.6%) | 71(46.4%) | 153(100%) |
| Pork intake | |||
| Never | 19,065(36.7%) | 32,903(63.3%) | 51,968(100%) |
| Less | 117,118(47.6%) | 128,729(52.4%) | 245,847(100%) |
| More | 187(70.3%) | 79(29.7%) | 266(100%) |
| Almost daily | 86(59.3%) | 59(40.7%) | 145(100%) |
MET Metabolic equivalent task, BMI Body mass index, HDL High density lipoprotein, LDL Low density lipoprotein
The association of serum lipid traits with IDD
Figure 2 shows the results of the restricted cubic spline plot, illustrating the relationship between serum lipid traits and the risk of IDD. After adjusting for multiple covariates, it was found that Apo A, Apo B, HDL-C, and LDL-C were not significantly correlated with the risk of IDD (Overall P > 0.05). TG and TC, were significantly correlated with the risk of IDD (Overall P < 0.05) and exhibited a linear correlation trend (Nonlinear P > 0.05). Table 2 shows that both total TC and TG have threshold effects on IDD. For TC, the risk of IDD increases when levels exceed 5.64 mmol/L, while levels below this threshold were possibly associated with a reduced risk of IDD. Similarly, TG levels above 1.51 mmol/L are associated with increased IDD risk, whereas levels below this threshold may be related to a lower risk of IDD.
Fig. 2.
Associations between lipid traits and the risk of IDD. Notes: (A) shows the restricted cubic spline(RCS) plot for apolipoprotein A, (B) for apolipoprotein B, (C) for HDL cholesterol, (D) for LDL cholesterol, (E) for total cholesterol, and (F) for triglycerides. RCS models were used to flexibly explore potential nonlinear associations between serum lipid levels and the risk of IDD. Model was adjusted for age, sex, race, townsend deprivation index, smoking, alcohol intake frequency, sleep duration, MET minutes per week, processed meat intake, oily fish intake, beef intake, lamb intake, pork intake, body-mass index, cardiovascular disease, hypertension, diabetes, kidney disease, thyroid disease, liver disease, cancer, anaemia, fractured in last 5 years, and aspirin use. The X-axis represents serum lipid concentrations (mmol/L or g/L), and the Y-axis shows hazard ratios (HRs) with 95% confidence intervals (CIs). An HR > 1 indicates an increased risk of IDD with higher lipid levels, whereas an HR < 1 indicates a decreased risk
Table 2.
Associations between serum lipid concentrations and HR
| TG (mmol/L) |
TC (mmol/L) |
Apo A (g/L) |
Apo B (g/L) |
LDL-C (mmol/L) |
HDL-C (mmol/L) |
|
|---|---|---|---|---|---|---|
| Reference (HR = 1) | 1.51 | 5.64 | 1.18 or 1.52 | 0.52 or 1.02 | 3.52 | 1.4 |
| Higher Risk Associated with | > 1.51 | > 5.64 | < 1.18 or > 1.52 | < 0.52 or > 1.02 | > 3.52 | > 1.4 |
| Lower Risk Associated with | < 1.51 | < 5.64 | 1.18–1.52 | 0.52–1.02 | < 3.52 | < 1.4 |
| Maximum HR | 3.17 | 6.97 | 2.5 | 1.32 | 7.33 | 3.31 |
Hazard ratios (HRs) reflect the relative risk of IDD associated with lipid levels above or below the reference values. Reference values represent the inflection points derived from restricted cubic spline models. Maximum HR values represent the peak risk observed across the lipid concentration range
TG Triglycerides, TC Total cholesterol, Apo A Apolipoprotein A, Apo B Apolipoprotein B, LDL-C Low-density lipoprotein cholesterol, HDL-C High-density lipoprotein cholesterol
Figure 3 provides a comprehensive visualization of the associations between serum lipid traits and IDD risk across all three models, facilitating clearer interpretation and comparison of effect sizes. These associations were estimated using Cox proportional hazards models. Each lipid indicator was divided into quintiles, and Model 3 compared higher quintiles to the lowest after full adjustment for confounders. Compared with the lowest quintile, the highest quintile of TC was associated with a 10.2% increased risk (HR = 1.102, 95% CI: 1.029–1.181, P = 0.006), while the highest TG quintile showed an 11.3% increased risk (HR = 1.113, 95% CI: 1.036–1.195, P = 0.003). In contrast, no significant associations were found for Apo A, Apo B, HDL-C, or LDL-C (all P > 0.05). Although the highest quintile of LDL-C showed a 6.8% increased risk (HR = 1.068, 95% CI: 0.998–1.144), it did not reach statistical significance (P = 0.059). Supplementary Table S2 complements the visual summary in Fig. 3 by presenting a concise and interpretable numerical overview of the Cox model results, aiding in the interpretation of effect sizes and statistical significance.
Fig. 3.
Results from Cox proportional hazards model analyses are visualized as forest plots, showing the associations between serum lipid trait quintiles and the risk of IDD. Hazard ratios (HRs) with 95% confidence intervals (CIs) are presented across Models 1, 2, and 3, with the lowest quintile used as the reference group. Notes: (A) shows the Cox model results for apolipoprotein A; (B) for apolipoprotein B; (C) for HDL cholesterol; (D) for LDL cholesterol; (E) for total cholesterol; and (F) for triglycerides; Model 1: Adjusted for age and sex; Model 2: Further adjusted for race, BMI, smoking status, drinking status, townsend deprivation index, and physical activity (MET min/week); Model 3: Based on Model 2, also adjusted for comorbidities such as hypertension and diabetes, as well as fractured in last 5 years, aspirin use, sleep duration, and frequency of meat intake (e.g., processed meat and beef)
Subgroup analysis
We observed that the association between serum lipid levels and IDD risk varied across subgroups. Figure 4 presents subgroup analyses of TC and TG stratified by age, sex, and BMI, while results for other lipids are shown in Supplementary Figure S1. No significant interactions were found between TC and age, sex, or BMI (all interaction P > 0.05); however, higher TC was an independent risk factor for IDD among individuals with BMI > 25 kg/m² [Q3-Q5 HRs ranging from 1.09 to 1.12; all P < 0.05]. A significant interaction was detected between TG and age (interaction P < 0.05). Among participants aged ≤ 60 years, higher TG levels were associated with increased IDD risk [Q5: HR = 1.18, 95% CI: 1.07–1.29; P < 0.05], while no interactions were observed with sex or BMI. Additionally, age significantly modified the associations of Apo A and HDL-C with IDD risk (interaction P < 0.05), whereas no significant interactions were found for Apo B or LDL-C (all interaction P > 0.05).
Fig. 4.
Subgroup analysis of the association between lipid traits and the risk of IDD. Notes: (A) shows the subgroup analysis results for total cholesterol (mmol/L) and (B) for triglycerides (mmol/L). Model was adjusted for age, sex, race, townsend deprivation index, smoking, alcohol intake frequency, sleep duration, MET minutes per week, processed meat intake, oily fish intake, beef intake, lamb intake, pork intake, body-mass index, cardiovascular disease, hypertension, diabetes, kidney disease, thyroid disease, liver disease, cancer, anaemia, fractured in last 5 years, and aspirin use. Interaction P-values were used to evaluate whether the association between lipid traits and IDD differed significantly across subgroups
Discussion
Based on data from 298,226 participants in the UK Biobank with a median follow-up of 12 years, we found that higher TG and TC levels were linearly associated with an increased risk of IDD. In contrast, ApoA, ApoB, HDL-C, and LDL-C showed no significant associations. The highest risk was observed among individuals in the top quintile of TG and TC levels. Subgroup analyses revealed significant interactions between age and ApoA, HDL-C, and TG (interaction P < 0.05), whereas no significant interactions were found for ApoB, TC, or LDL-C with age, sex, or BMI (all interaction P > 0.05). Previous studies have reported links between elevated serum lipids and disc degeneration risk. For example, Shi et al. [22] identified high TG levels (≥ 1.7 mmol/L) as a significant predictor in a cohort of 678 Chinese participants. Yuan et al. [12] reported high TC (≥ 6.2 mmol/L) and LDL-C (≥ 4.1 mmol/L) as independent risk factors, while Huang et al. [23] emphasized the roles of HDL-C, TG, and the LDL-C/HDL-C ratio in symptomatic lumbar disc disease. In summary, our findings support a potential association between dyslipidemia and IDD, highlighting TG and TC as particularly influential lipid markers. This study contributes to the literature by leveraging large-scale prospective data to clarify the impact of dyslipidemia on IDD and uncover age-related variations in these associations.
Restricted cubic spline analysis demonstrated a significant association between serum TC and TG levels and the risk of IDD. Specifically, TC levels below 5.64 mmol/L and TG levels below 1.51 mmol/L were linked to a lower risk of IDD, whereas levels above these thresholds significantly increased the risk. These cutoffs closely correspond to the 2021 European lipid guidelines [24], which define normal TC as < 5.0 mmol/L and normal TG as < 1.7 mmol/L; values at or above these indicate dyslipidemia and heightened hyperlipidemia risk. Our large-scale cohort enabled a robust evaluation of nonlinear relationships, providing novel insights into lipid metabolism and IDD. Multivariate Cox regression confirmed that higher TC and TG are likely independent risk factors, consistent with prior clinical and animal research [11, 12, 25–27]. Based on these findings, we recommend monitoring and managing lipid levels in individuals with TC > 5.64 mmol/L or TG > 1.51 mmol/L as a potential strategy to prevent IDD. Unlike previous studies limited by small sample sizes and cross-sectional designs, our prospective, time-to-event analysis offers stronger inferential evidence. Nevertheless, residual confounding and reverse causation cannot be fully excluded. Although individuals with baseline IDD were excluded and adjustments made for various lifestyle, dietary, and metabolic factors, unmeasured variables—such as inflammatory biomarkers or genetic predispositions—may still affect both lipid levels and disc degeneration risk. Furthermore, complex interactions involving systemic metabolism, physical activity, and subclinical disc changes could introduce additional variability. Therefore, while our results support a potential causal relationship, they should be interpreted with caution. Further research in independent and diverse populations is warranted to validate these findings and clarify underlying mechanisms.
Restricted cubic spline analysis in this study found no significant association between HDL-C or LDL-C levels and the risk of IDD. Although prior studies have linked dyslipidemia to IDD through mechanisms such as atherosclerosis-induced disc ischemia [11, 12, 23, 25], the evidence remains inconsistent. Animal studies suggest that high-fat diets elevate TG, TC, and LDL-C levels, thereby accelerating disc degeneration [28]. Some research has identified LDL-C ≥ 4.1 mmol/L as an independent risk factor for IDD [11, 12], but these findings mostly derive from small, cross-sectional studies susceptible to bias and reverse causality, limiting their generalizability. Investigations into the roles of apolipoproteins in disc health are limited. ApoA1, the major component of HDL, exhibits anti-atherosclerotic properties, while ApoB, the main LDL component, promotes atherosclerosis [29]. The ApoB/ApoA1 ratio is a recognized marker of lipid metabolism disorders and is linked to various diseases. However, in our study, neither ApoA1 nor ApoB was significantly associated with IDD risk. Further research is warranted to elucidate the role of apolipoproteins in IDD pathogenesis and to develop potential preventive and therapeutic strategies.
Significant interactions between age and Apo A, HDL-C, and TG were observed in subgroup analyses (interaction P < 0.05). Among participants aged 60 years or younger, higher TG levels were more strongly associated with increased risk of IDD, with a similar trend observed for HDL-C. In contrast, among those older than 60 years, Apo A showed a more prominent association with IDD risk. These findings suggest that the influence of lipid metabolism on IDD risk may vary across age groups. However, as this study is observational, the results should be interpreted cautiously, and no causal relationships or intervention strategies can be directly inferred from our data. Although our study did not examine interventions, prior research suggests that healthy lifestyles—balanced nutrition, regular exercise, good posture, and weight management—may support disc health by improving lipid profiles and reducing inflammation [30, 31]. Diets like the Mediterranean diet, rich in antioxidants and omega-3s, have been linked to better metabolic and inflammatory status, potentially slowing disc degeneration [32]. Moderate activities such as swimming, yoga, and walking can strengthen paraspinal muscles and reduce lumbar stress [31, 33]. Additionally, statins have shown protective effects on discs through anti-inflammatory and antioxidative mechanisms [34–36]. These lifestyle and pharmacologic strategies, while not examined in our study, are based on existing literature and may represent valuable directions for future research focused on IDD prevention, particularly among individuals with dyslipidemia.
This study has several strengths, including its prospective design, large sample size, long follow-up (median > 12 years), and comprehensive covariate assessment (e.g., anthropometrics, lifestyle, medical history). Given the progressive nature of IDD [37], long-term follow-up is essential for understanding chronic outcomes. Lipid biomarkers were measured with standardized, validated methods under strict quality control, minimizing measurement errors [14]. This is the largest population-based study to explore the non-linear relationship between serum lipids and intervertebral disc health, with Cox regression and subgroup analyses enhancing result consistency. However, there are limitations. Firstly, residual confounding may still influence the causal inference, despite adjustments for known covariates. For example, potential sources of residual confounding include the use of self-reported dietary data limited to meat and fish intake, which may not fully capture overall dietary patterns (e.g., fiber, saturated fat, micronutrients); physical activity data based solely on self-reported MET scores, which are subject to recall bias; the absence of sensitivity analyses accounting for competing risks (e.g., mortality), which may overestimate IDD risk in older adults; and the lack of comprehensive data on medication use (e.g., lipid-lowering treatments), which could also influence lipid levels and IDD risk. Future studies should consider incorporating more detailed assessments of diet, physical activity (including posture and high-intensity exercise), medication use, and competing risk models to improve causal inference. Secondly, although missing data were handled using the MICE method and most variables showed acceptable imputation performance, certain variables—such as smoking and alcohol intake frequency—had relatively low imputation accuracy, which may introduce residual bias and influence the study results. Lastly, due to the absence of thoracolumbar MRI data in the UK Biobank, IDD was identified based solely on hospital records using ICD codes, which may limit diagnostic accuracy compared to MRI-based evaluations, the current gold standard. Because of the lack of imaging data, we were unable to assess the severity or classification of IDD. However, ICD-coded outcomes are commonly used in large-scale studies and can reasonably capture clinically significant IDD cases. Future studies incorporating imaging data are needed to validate these associations more precisely and to evaluate IDD severity and classification.
Conclusions
This observational study found that higher TC and TG levels were associated with increased risk of IDD, while no significant associations were observed for other lipid markers. The findings also suggest age-related variations in lipid-associated IDD risk. Clinically, routine lipid screening may help identify individuals at higher risk, potentially guiding lifestyle or pharmacologic strategies to reduce risk. However, further interventional studies are needed to confirm these findings and explore preventive approaches.
Supplementary Information
Acknowledgements
The authors would like to thank all UK Biobank participants and the research team for their significant contributions.
Abbreviations
- IDD
Intervertebral disc degeneration
- TC
Total cholesterol
- TG
Triglycerides
- LDL-C
Low-density lipoprotein cholesterol
- HDL-C
High-density lipoprotein cholesterol
- Apo A
Apolipoprotein A
- Apo B
Apolipoprotein B
- ECM
Extracellular matrix
- NP
Nucleus pulposus
- BMI
Body mass index
- MET
Metabolic equivalent task
- MICE
Multiple imputation by chained equations
- HR
Hazard ratio
- CI
Confidence interval
Authors’ contributions
WC, WG, PS and XQ conceived and designed the research; WH, YL, ZZ and NL collected and processed the data; WC, BG, and JC performed the statistical analyses, with additional statistical support from TL; YC, ZD, SL and HS contributed to data interpretation and result visualization; WC, BG, and JC drafted the manuscript; WG, PS and XQ revised the manuscript; DH provided critical suggestions during the study; and all authors read and approved the final manuscript.
Funding
This study was supported by the National Natural Science Foundation of China (No.82072473, No.82202653, No.82302737, No.82372432, No.82103083), Guangdong Basic and Applied Basic Research Foundation (No.2022A1515012459, No.2023A1515012689), Science and Technology Program of Guangzhou, China (No. 2024A04J4684, No. 2025A03J41881).
Data availability
Data from the UK Biobank are available at https://www.ukbiobank.ac.uk/.
Declarations
Ethics approval and consent to participate
Ethical approval for this study was obtained from the Northwest Multi-Centre Research Ethics Committee. All participants provided written informed consent prior to participation. Access to the data was granted through the UK Biobank under application No. 117320. The ethical approval for the UK Biobank was issued by the NHS National Research Ethics Service on June 17, 2011 (Ref 11/NW/0382), and May 10, 2016 (Ref 16/NW/0274), applies to this study.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Wangim Choi, Bo Gao and Jianan Chen contributed equally to this work and the co-first authors.
Contributor Information
Xianjian Qiu, Email: qiuxj6@mail.sysu.edu.cn.
Peijie Shi, Email: shipj5@mail.sysu.edu.cn.
Wenjie Gao, Email: gaowj7@mail.sysu.edu.cn.
References
- 1.Ashinsky BG, Gullbrand SE, Wang C, Bonnevie ED, Han L, Mauck RL, et al. Degeneration alters structure-function relationships at multiple length-scales and across interfaces in human intervertebral discs. J Anat. 2021;238:986–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Adams MA, Roughley PJ. What is intervertebral disc degeneration, and what causes it?? Spine. 2006;31:2151–61. [DOI] [PubMed] [Google Scholar]
- 3.Li W, Zhang S, Wang D, Zhang H, Shi Q, Zhang Y, et al. Exosomes immunity strategy: A novel approach for ameliorating intervertebral disc degeneration. Front Cell Dev Biol. 2021;9:822149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Song C, Cai W, Liu F, Cheng K, Guo D, Liu Z. An in-depth analysis of the Immunomodulatory mechanisms of intervertebral disc degeneration. JOR SPINE. 2022;5:e1233. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gullbrand S, Orozco B, Fainor M, Meadows K, Hilliard R, Boyes M, et al. Intervertebral disc degeneration instigates vertebral endplate remodeling and facet joint pathology in a large animal model. Eur Cell Mater. 2024;47:125–41. [Google Scholar]
- 6.Oichi T, Taniguchi Y, Oshima Y, Tanaka S, Saito T. Pathomechanism of intervertebral disc degeneration. JOR Spine. 2020;3:e1076. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Uzokov J. Influence of abnormal lipid components in statin-naive young adults: is there any gap? Eur J Prev Cardiol. 2020;27:868–9. [DOI] [PubMed] [Google Scholar]
- 8.Qi L, Ding X, Tang W, Li Q, Mao D, Wang Y. Prevalence and risk factors associated with dyslipidemia in chongqing, China. Int J Environ Res Public Health. 2015;12:13455–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Xiong X, Lui DTW, Ju C, Zhou Z, Xu C, Welsh P, et al. Associations of serum lipid traits with fracture and osteoporosis: A prospective cohort study from the UK biobank. J Cachexia Sarcopenia Muscle. 2024;15:2669–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Fang Z, He M, Song M. Serum lipid profiles and risk of colorectal cancer: a prospective cohort study in the UK biobank. Br J Cancer. 2021;124:663–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zhang Y, Zhao Y, Wang M, Si M, Li J, Hou Y, et al. Serum lipid levels are positively correlated with lumbar disc herniation–a retrospective study of 790 Chinese patients. Lipids Health Dis. 2016;15:80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Yuan L, Huang Z, Han W, Chang R, Sun B, Zhu M, et al. The impact of dyslipidemia on lumbar intervertebral disc degeneration and vertebral endplate modic changes: a cross-sectional study of 1035 citizens in China. BMC Public Health. 2023;23:1302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zhang Y, Si M, Li C, Liu Y, Han Y, Nie L, et al. Effect of hyperlipidaemia to accelerate intervertebral disc degeneration in the injured rat caudal disc model. J Orthop Sci Off J Jpn Orthop Assoc. 2019;24:42–9. [DOI] [PubMed] [Google Scholar]
- 14.Elliott P, Peakman TC, on behalf of UK Biobank. The UK biobank sample handling and storage protocol for the collection, processing and archiving of human blood and urine. Int J Epidemiol. 2008;37:234–44. [DOI] [PubMed] [Google Scholar]
- 15.Peakman TC, Elliott P. The UK biobank sample handling and storage validation studies. Int J Epidemiol. 2008;37:i2–6. [DOI] [PubMed] [Google Scholar]
- 16.Guo W, Bradbury KE, Reeves GK, Key TJ. Physical activity in relation to body size and composition in women in UK biobank. Ann Epidemiol. 2015;25:406–e4136. [DOI] [PubMed] [Google Scholar]
- 17.Foster HME, Celis-Morales CA, Nicholl BI, Petermann-Rocha F, Pell JP, Gill JMR, et al. The effect of socioeconomic deprivation on the association between an extended measurement of unhealthy lifestyle factors and health outcomes: a prospective analysis of the UK biobank cohort. Lancet Public Health. 2018;3:e576–85. [DOI] [PubMed] [Google Scholar]
- 18.Liu B, Young H, Crowe FL, Benson VS, Spencer EA, Key TJ, et al. Development and evaluation of the Oxford webq, a low-cost, web-based method for assessment of previous 24 h dietary intakes in large-scale prospective studies. Public Health Nutr. 2011;14:1998–2005. [DOI] [PubMed] [Google Scholar]
- 19.Govindarajulu US, Malloy EJ, Ganguli B, Spiegelman D, Eisen EA. The comparison of alternative smoothing methods for fitting non-linear exposure-response relationships with Cox models in a simulation study. Int J Biostat. 2009;5(1):2. 10.2202/1557-4679.1144. [DOI] [PMC free article] [PubMed]
- 20.Susetyo B, Fitrianto A. Estimating missing panel data with regression and multivariate imputation by chained equations (MICE). CAUCHY J Mat Murni Dan Apl. 2024;9:94–105. [Google Scholar]
- 21.Royston P, White IR. Multiple imputation by chained equations (MICE): implementation in Stata. J Stat Softw. 2011;45:1–20. [Google Scholar]
- 22.Shi S, Zhou Z, Liao JJ, Yang YH, Wu JS, Zheng S, He SS. The impact and distinction of ‘lipid healthy but obese’ and ‘lipid abnormal but not obese’ phenotypes on lumbar disc degeneration in Chinese. J Transl Med. 2020;18(1):211. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Huang Z, Chen J, Su Y, Guo M, Chen Y, Zhu Y, et al. Impact of dyslipidemia on the severity of symptomatic lumbar spine degeneration: A retrospective clinical study. Front Nutr. 2022;9:1033375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Visseren FLJ, Mach F, Smulders YM, Carballo D, Koskinas KC, Bäck M, et al. 2021 ESC guidelines on cardiovascular disease prevention in clinical practice. Eur Heart J. 2021;42:3227–337. [DOI] [PubMed] [Google Scholar]
- 25.Li W, Djuric N, Vleggeert-Lankamp CLA. A systematic review evaluating the association of atherosclerosis and lumbar degenerative disc disease. Brain Spine. 2024;4:103901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Tan B, Xiang S, Zheng Y, Ouyang J, Zhou N. Association of dyslipidemia with intervertebral disc degeneration: a case-control study. Eur J Med Res. 2025;30:194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Qin T, Yan J, Li S, Lin X, Wu J, Huang Z, et al. MicroRNA-155 suppressed cholesterol-induced matrix degradation, pyroptosis and apoptosis by targeting RORα in nucleus pulposus cells. Cell Signal. 2023;107:110678. [DOI] [PubMed] [Google Scholar]
- 28.Bing T, Shanlin X, Jisheng W, Jie H, Ruichao C, Zhiwei Z, et al. Dysregulated lipid metabolism and intervertebral disc degeneration: the important role of ox-LDL/LOX-1 in endplate chondrocyte senescence and calcification. Mol Med. 2024;30:117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Chen F, Wu T, Bai C, Guo S, Huang W, Pan Y, et al. Serum Apolipoprotein b/apolipoprotein A1 ratio in relation to intervertebral disk herniation: a cross-sectional frequency-matched case–control study. Lipids Health Dis. 2021;20:79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, et al. 2018 aha/acc/aacvpr/aapa/abc/acpm/ada/ags/apha/aspc/nla/pcna guideline on the management of blood cholesterol: executive summary: A report of the American college of cardiology/american heart association task force on clinical practice guidelines. J Am Coll Cardiol. 2019;73:3168–209. [DOI] [PubMed] [Google Scholar]
- 31.De Geer CM. Intervertebral disk nutrients and transport mechanisms in relation to disk degeneration: A narrative literature review. J Chiropr Med. 2018;17:97–105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Scicchitano P, Cameli M, Maiello M, Modesti PA, Muiesan ML, Novo S, et al. Nutraceuticals and dyslipidaemia: beyond the common therapeutics. J Funct Foods. 2014;6:11–32. [Google Scholar]
- 33.Shiri R, Coggon D, Falah-Hassani K. Exercise for the prevention of low back pain: systematic review and Meta-Analysis of controlled trials. Am J Epidemiol. 2018;187:1093–101. [DOI] [PubMed] [Google Scholar]
- 34.Chen J, Yan J, Li S, Zhu J, Zhou J, Li J, et al. Atorvastatin inhibited TNF-α induced matrix degradation in rat nucleus pulposus cells by suppressing NLRP3 inflammasome activity and inducing autophagy through NF-κB signaling. Cell Cycle Georget Tex. 2021;20:2160–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Than KD, Rahman SU, Wang L, Khan A, Kyere KA, Than TT, et al. Intradiscal injection of Simvastatin results in radiologic, histologic, and genetic evidence of disc regeneration in a rat model of degenerative disc disease. Spine J Off J North Am Spine Soc. 2014;14:1017–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Chen W, Deng Z, Zhu J, Yuan L, Li S, Zhang Y, et al. Rosuvastatin suppresses TNF-alpha-induced matrix catabolism, pyroptosis and senescence via the HMGB1/NF-kappaB signaling pathway in nucleus pulposus cells. Acta Biochim Biophys Sin (Shanghai). 2023;55(5):795–808. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Roughley PJ. Biology of intervertebral disc aging and degeneration: involvement of the extracellular matrix. Spine. 2004;29:2691–9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data from the UK Biobank are available at https://www.ukbiobank.ac.uk/.