Genetically Predicted Intervertebral Disc Degeneration and Risk of Osteoarthritis: A Two-Sample Mendelian Randomization Study with Mediation Analysis of Obesity and Sarcopenia Authors: Junyu Zhu, Zunyu Li, Shijie Zhou, Jun Yan, Jiufang He Published: 2026 ABSTRACT BackgroundChronic pain related to spinal degeneration and osteoarthritis (OA) represents a major cause of disability worldwide. However, whether intervertebral disc degeneration (IVDD) contributes to OA and related pain burden remains unclear.ObjectiveTo investigate the potential causal relationship between genetically predicted IVDD-related traits and OA risk, and to explore the mediating roles of obesity and sarcopenia.MethodsWe conducted a two-sample Mendelian randomization study using genome-wide association study (GWAS) data. IVDD-related genetic instruments were obtained from FinnGen, and OA outcomes included knee osteoarthritis, hip osteoarthritis, and joint replacement. Sensitivity analyses were performed using multiple MR methods. A two-step MR framework was applied to assess mediation effects.ResultsGenetically predicted IVDD was associated with an increased risk of HOA (OR = 1.33, 95% CI: 1.20–1.48), THR (OR = 1.37, 95% CI: 1.20–1.56), and KOA (OR = 1.17, 95% CI: 1.09–1.26). Similar associations were observed for lumbar disc prolapse and sciatica, which were consistently linked to higher risks of OA outcomes. Two-step MR analyses indicated that both adiposity-, obesity-, and metabolic-related and sarcopenia-related traits partially mediated these associations. The proportion mediated by adiposity-, obesity-, and metabolic-related traits ranged from approximately 2% to 31%, while sarcopenia-related mediation ranged from 3% to 20%, depending on exposure and outcome pairs.ConclusionThis study provides genetic evidence supporting a potential link between spinal degeneration and chronic pain-related osteoarthritis. Obesity and sarcopenia may serve as partial mediators, highlighting potential targets for prevention and intervention in chronic musculoskeletal pain conditions. Introduction Chronic musculoskeletal pain is one of the leading causes of disability worldwide, significantly impairing physical function and quality of life.1 Among its major contributors, spinal degenerative disorders and osteoarthritis (OA) represent two highly prevalent conditions that account for a substantial proportion of chronic pain burden in aging populations.2,3 Intervertebral disc degeneration (IVDD) is a primary structural cause of spinal pain, frequently associated with both nociceptive and neuropathic pain components, particularly when nerve root compression and inflammatory sensitization are involved.4,5 OA, similarly, is not merely a degenerative cartilage disorder but a complex pain condition driven by synovial inflammation, subchondral bone remodeling, and peripheral and central sensitization mechanisms.6 Clinically, spinal degeneration and peripheral joint OA often coexist, suggesting that these conditions may not be independent disorders but rather interconnected components of a broader chronic pain system.7,8 In particular, patients with IVDD frequently present with lower limb joint pain, while individuals with OA often report concurrent spinal symptoms, indicating a possible spine–lower limb degenerative axis.4,9,10 However, whether spinal degeneration contributes causally to OA burden remains uncertain, as observational studies are limited by confounding factors such as aging, obesity, physical inactivity, and shared biomechanical stressors.11,12 Mendelian randomization (MR) offers a powerful approach to assess potential causal relationships by using genetic variants as instrumental variables, minimizing confounding and reverse causation.13,14 Although previous MR studies have explored metabolic and inflammatory determinants of OA, the potential role of IVDD-related genetic liability in shaping OA risk has not been systematically investigated.15 Moreover, the biological pathways linking spinal degeneration to lower limb degeneration remain poorly defined. Several systemic and functional factors may plausibly participate in the biological pathway linking spinal and peripheral joint degeneration. Obesity may contribute to OA through both increased mechanical loading and adipose-derived metabolic and inflammatory pathways, whereas sarcopenia may reduce muscular support, impair neuromuscular function, and alter joint biomechanics.11,16 Importantly, these traits may represent more than shared risk factors. IVDD-related functional impairment may be accompanied by reduced physical activity, altered gait and loading patterns, and impaired neuromuscular function, which could potentially contribute to adverse changes in muscle mass and adiposity. In turn, sarcopenia and obesity may increase abnormal joint loading, reduce joint stability, and promote metabolic or inflammatory processes that facilitate peripheral joint degeneration. These considerations provide a biological rationale for investigating adiposity-, obesity-, metabolic- and sarcopenia-related traits as potential intermediate pathways between IVDD and OA. Nevertheless, because obesity and sarcopenia may also independently contribute to both spinal and peripheral joint degeneration, their roles should be interpreted as potential mediators rather than established intermediates, and formal causal evaluation is required. Despite growing genetic evidence concerning individual risk factors for OA and musculoskeletal degeneration, an integrated assessment of the relationship between spinal disc degeneration and large-joint OA remains limited. In particular, few studies have simultaneously considered multiple IVDD-related phenotypes, anatomically distinct OA outcomes, and clinically consequential joint replacement endpoints within a unified causal framework. Furthermore, whether adiposity-, obesity-, metabolic- and sarcopenia-related traits may partially account for the relationship between spinal degeneration and peripheral joint degeneration has not been systematically evaluated. Addressing these questions may help determine whether IVDD and peripheral joint OA form part of a broader spine–lower limb degenerative pathway.11,12 In this study, we therefore applied a comprehensive two-sample MR framework to investigate the potential causal relationships between multiple IVDD-related traits and large-joint OA outcomes, including hip OA (HOA), knee OA (KOA), total hip replacement (THR), and total knee replacement (TKR). We additionally examined lumbar disc-related phenotypes to assess the consistency of the associations across related manifestations of spinal degeneration. Finally, using a two-step MR approach, we evaluated whether sarcopenia-related traits and adiposity-, obesity-, and metabolic-related traits may represent potential intermediate pathways linking IVDD-related genetic liability to OA. By integrating spinal degenerative phenotypes, peripheral joint outcomes, and potential metabolic and musculoskeletal mediators, this study aims to provide genetic evidence for a potential spine–lower limb degenerative axis and to clarify possible pathways underlying the coexistence of spinal and peripheral joint degeneration. We also acknowledge our previously published MR study examining the causal relationship between hormonal factors and osteoporosis and the potential mediating role of sarcopenia.17 Although the present study uses a similar MR framework, it addresses a distinct scientific question, with IVDD-related traits as the primary exposures and large-joint OA and joint replacement as the outcomes. The present analysis further differs in its focus on the potential spine–lower limb degenerative pathway and its evaluation of obesity- and sarcopenia-related traits as potential mediators. Thus, the current study extends our previous work to a distinct exposure–outcome system and addresses a separate mechanistic hypothesis. Methods Study Design This study used ethically approved, summary-level data from large genome-wide association studies, and thus did not require further institutional review board approval or informed consent. This study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.18 In this MR study, we identified genetic instruments for the exposures from GWAS summary statistics and other relevant databases, and examined their associations with the outcomes. In the first step, IVDD-related traits were used as exposures and OA subtypes as outcomes. In the second step, the exposure–outcome relationship was reversed. We then performed MR analyses to evaluate the causal associations of IVDD with sarcopenia-related traits and adiposity- and obesity-related traits. Finally, a two-step MR framework was applied to determine whether sarcopenia and obesity mediated the relationship between IVDD and OA, based on the observed causal effects of IVDD on sarcopenia and obesity and their respective causal influences on OA. Data were extracted between December 1, 2025, and July 30, 2025 (Figure 1). Figure 1Schematic Presentation of the Mendelian Randomization Design. The asterisk (*) denotes multiplication. Arrow represents the causal effect of interest. For each genetic instrument, the causal effect is estimated by dividing the genetic variant-outcome association by the genetic variant-exposure association. For this estimate to be valid, three main assumptions need to be met: 1) The genetic variant is associated with the exposure; 2) the genetic variant is not associated with outcome via a confounder; and 3) the genetic variant does not directly affect the outcome.A diagram illustrating Mendelian Randomization with Intervertebral Disc Degeneration (IVDD) as exposure, leading to Osteoarthritis (OA) as the outcome. The exposure path includes subtypes: ELDP, LDP and Sciatica. Two mediators are shown: Sarcopenia and Obesity. Sarcopenia mediators include ALM, WBLM and GRIP STRENGTH, with an indirect effect calculated as beta subscript 1 times beta subscript 2. Obesity mediators include BMI, WC, BFP, BMR and BW, also with an indirect effect of beta subscript 1 times beta subscript 2. The total effect is estimated as beta subscript 0. The mediation paths connect IVDD to OA through Sarcopenia and Obesity. OA phenotypes include HOA, THR, KOA and TKR. Two-step MR estimates are shown for both direct and indirect effects, with beta subscript 1 and beta subscript 2 for indirect effects and beta subscript 0 for total effect.A diagram of Mendelian Randomization showing IVDD exposure, sarcopenia and obesity mediators and OA outcome.Abbreviations: IVDD, intervertebral disc degeneration, ELDP, early lumbar disc prolapse, LDP, lumbar disc prolapse, HOA, hip osteoarthritis, THR, total hip replacement, KOA, knee osteoarthritis, TKR, total knee replacement, WBLM, Whole body lean mass, ALM, Appendicular lean mass, WC,waist circumference, HC, hip circumference, WBFM, whole body fat mass, BFP, body fat percentage, BMR, basal metabolic rate, BW, birth weight, BMI, body mass index. Schematic Presentation of the Mendelian Randomization Design. The asterisk (*) denotes multiplication. Arrow represents the causal effect of interest. For each genetic instrument, the causal effect is estimated by dividing the genetic variant-outcome association by the genetic variant-exposure association. For this estimate to be valid, three main assumptions need to be met: 1) The genetic variant is associated with the exposure; 2) the genetic variant is not associated with outcome via a confounder; and 3) the genetic variant does not directly affect the outcome. Data Source Exposed Data Source Summary data for IVDD-related traits (Intervertebral disc degeneration, Lumbar disc prolapse, Early lumbar disc prolapse(operated), and Sciatica) in this study were obtained from the FinnGen database (https://www.finngen.fi/en) (R12). The details of datasets were shown in Table 1.Table 1Overview of Genetic Data SetsDatabaseDatasets or PMIDCharactersSample Size(Cases/Controls)AncestryFinngenFINNGEN_R12_M13_INTERVERTEBIVDD404,907(51,683/353,224)EuropeanFinngenfinngen_R12_M13_EARLY_LUMBAR_PROLAPSEELDP357,783(4,559/353,224)EuropeanFinngenFinngen_R12_M13_LUMBAR_PROLAPSELDP388,514(35,290/353,224)EuropeanFinngenFinngen_R12_M13_LOWBACKPAINORANDSCIATICASciatica500,348(60,099/440,249)EuropeanOthers34822786HOA36,445MixedOthers34822786THR23,021MixedOthers34822786KOA62,497MixedOthers34822786TKR18,200MixedGWAS Catalogukb-b-9405Waist circumference462,166EuropeanGWAS Catalogukb-b-15590Hip circumference462,117EuropeanGWAS Catalogukb-b-19393Whole body fat mass454,137EuropeanGWAS Catalogukb-b-8909Body fat percentage454,633EuropeanGWAS Catalogukb-b-16446Basal metabolic rate454,878EuropeanGWAS Catalogukb-b-13378Birth weight261,932EuropeanGWAS Catalogukb-a-248Body mass index336,107EuropeanGWAS Catalogukb-b-13354WBLM454,850EuropeanGWAS Catalogebi-a-GCST90000025ALM450,243EuropeanGWAS Catalogukb-b-10215Right-hand grip strength461,089EuropeanGWAS Catalogukb-b-7478Left-hand grip strength461,026EuropeanAbbreviations: GWAS ID, Genome-wide association studies Identity number, PMID, PubMed identification number, IVDD, intervertebral disc degeneration, ELDP, early lumbar disc prolapse, LDP, lumbar disc prolapse, HOA, hip osteoarthritis, THR, total hip replacement, KOA, knee osteoarthritis, TKR, total knee replacement, WBLM, Whole body lean mass, ALM, Appendicular lean mass, NA, Not Applicable. Table 1Overview of Genetic Data SetsDatabaseDatasets or PMIDCharactersSample Size(Cases/Controls)AncestryFinngenFINNGEN_R12_M13_INTERVERTEBIVDD404,907(51,683/353,224)EuropeanFinngenfinngen_R12_M13_EARLY_LUMBAR_PROLAPSEELDP357,783(4,559/353,224)EuropeanFinngenFinngen_R12_M13_LUMBAR_PROLAPSELDP388,514(35,290/353,224)EuropeanFinngenFinngen_R12_M13_LOWBACKPAINORANDSCIATICASciatica500,348(60,099/440,249)EuropeanOthers34822786HOA36,445MixedOthers34822786THR23,021MixedOthers34822786KOA62,497MixedOthers34822786TKR18,200MixedGWAS Catalogukb-b-9405Waist circumference462,166EuropeanGWAS Catalogukb-b-15590Hip circumference462,117EuropeanGWAS Catalogukb-b-19393Whole body fat mass454,137EuropeanGWAS Catalogukb-b-8909Body fat percentage454,633EuropeanGWAS Catalogukb-b-16446Basal metabolic rate454,878EuropeanGWAS Catalogukb-b-13378Birth weight261,932EuropeanGWAS Catalogukb-a-248Body mass index336,107EuropeanGWAS Catalogukb-b-13354WBLM454,850EuropeanGWAS Catalogebi-a-GCST90000025ALM450,243EuropeanGWAS Catalogukb-b-10215Right-hand grip strength461,089EuropeanGWAS Catalogukb-b-7478Left-hand grip strength461,026EuropeanAbbreviations: GWAS ID, Genome-wide association studies Identity number, PMID, PubMed identification number, IVDD, intervertebral disc degeneration, ELDP, early lumbar disc prolapse, LDP, lumbar disc prolapse, HOA, hip osteoarthritis, THR, total hip replacement, KOA, knee osteoarthritis, TKR, total knee replacement, WBLM, Whole body lean mass, ALM, Appendicular lean mass, NA, Not Applicable. Overview of Genetic Data Sets Abbreviations: GWAS ID, Genome-wide association studies Identity number, PMID, PubMed identification number, IVDD, intervertebral disc degeneration, ELDP, early lumbar disc prolapse, LDP, lumbar disc prolapse, HOA, hip osteoarthritis, THR, total hip replacement, KOA, knee osteoarthritis, TKR, total knee replacement, WBLM, Whole body lean mass, ALM, Appendicular lean mass, NA, Not Applicable. Outcome Data Source In the present study, four osteoarthritis (OA) phenotypes affecting the lower extremities were analyzed, including knee osteoarthritis (KOA), hip osteoarthritis (HOA), total knee replacement (TKR), and total hip replacement (THR). Summary-level GWAS data for OA were obtained from the largest meta-analysis conducted to date, comprising 826,690 individuals and covering 11 OA-related phenotypes (Table 1).19 Mediators Data Source For sarcopenia-related phenotypes, we included whole-body lean mass (WBLM), appendicular lean mass (ALM), and hand grip strength. Among these indicators, grip strength reflects muscular function, whereas WBLM is widely recognized as a surrogate of overall muscle mass; both are established predictors of sarcopenia. ALM, largely derived from skeletal muscle mass, provides an additional quantitative measure of muscular content and, when integrated with other indices, improves the assessment accuracy of sarcopenia. The GWAS summary statistics for hand grip strength were derived from the UK Biobank (UKB), comprising 461,089 individuals for right-hand grip strength and 461,026 individuals for left-hand grip strength. Data for WBLM (N = 454,850) were also obtained from the UKB cohort. In addition, ALM-related traits were estimated based on total fat-free mass in 450,243 participants from the UKB, as reported in the corresponding GWAS study (https://www.ebi.ac.uk/gwas/publications/33097823). adiposity-, obesity-, and metabolic-related traits (including waist circumference(WC), hip circumference(HC), whole body fat mass (WBFM), body fat percentage (BFP), basal metabolic rate(BMR), birth weight(BW) and body mass index (BMI)) were acquired from the consortium known as MRC-IEU (MRC Integrative Epidemiology Unit) consorium (Table 1).20 Selection of Instrumental Variables (IVs) For a valid MR analysis, the selection of IVs is critical and must satisfy three core assumptions: (1) The chosen IVs are robustly related to the investigation exposure; (2) No confounding variables exist between the exposure and outcome affecting the chosen IVs; (3) Other than by their association with the exposure, the chosen IVs have no impact on the outcome. Based on the above assumptions, IVs were selected through the following steps: (1) We screened significant IVs with a threshold of P < 5×10−8. (2) We applied linkage disequilibrium (LD) (r2 = 0.001 within a 10000-kb distance) as a threshold for LD testing using PLINK (v1.9) to ensure the independence of the selected IVs.21 (3) The F-statistic for each SNP was calculated, with F < 10 used as a threshold to exclude weak IVs. Statistical Analysis All statistical analyses were performed in R software (version 4.5.1; R Foundation for Statistical Computing), using the TwoSampleMR (v0.6.22) and MendelianRandomization (v0.10.0) packages. The IVW approach was adopted as the primary analytical method, as it yields unbiased causal estimates under the assumption of no heterogeneity and absence of directional pleiotropy between exposure and outcome. Cochran’s Q statistic was applied to assess heterogeneity across instrumental variables. When significant heterogeneity was detected (P < 0.05), a random-effects IVW model was applied instead of the fixed-effects model.22–24 MR-Egger regression was further conducted to estimate causal effects while simultaneously evaluating potential directional and horizontal pleiotropy. In this framework, a non-zero intercept (P < 0.05) indicates the presence of directional pleiotropy.25 Horizontal pleiotropy was further assessed using MR-Pleiotropy Residual Sum and Outlier (MR-PRESSO) test. When potential outliers were detected, analyses were repeated after outlier removal, and the consistency between the original and corrected estimates was evaluated. MR-PRESSO results were interpreted together with MR-Egger intercept tests, heterogeneity analyses, and leave-one-out analyses. To ensure robustness of the findings, a series of sensitivity analyses were performed. In particular, a leave-one-out analysis was conducted to evaluate whether the overall causal inference was disproportionately driven by any single SNP.26 Because this study aimed to investigate biologically related hypotheses rather than independent multiple testing scenarios, strict Bonferroni correction was not applied. Instead, robustness was assessed through concordance among multiple MR methods and sensitivity analyses. Mediation Analysis To investigate whether adiposity-, obesity-, metabolic-, and sarcopenia-related traits may represent potential intermediate pathways linking IVDD-related traits to OA, a two-step MR framework was applied based on the observed causal associations between IVDD-related traits and OA and between IVDD-related traits and the candidate mediators. This approach was used to estimate potential indirect effects rather than to establish a definitive sequential biological pathway. First, univariable MR (UVMR) analysis was performed to estimate the total causal effect of each IVDD-related trait on the corresponding OA outcome, denoted as β_total. Second, the causal effect of each IVDD-related trait on the candidate mediator was estimated as β1. Third, the causal effect of each candidate mediator on the corresponding OA outcome was estimated as β2. The indirect effect through each candidate mediator was estimated using the product-of-coefficients method:β_indirect = β1 × β2The direct effect of the IVDD-related trait on OA was subsequently estimated as:β_direct = β_total − β_indirect. The proportion of the total effect mediated through each candidate mediator was calculated as: Proportion mediated (%) = (β_indirect / β_total) × 100%. The standard error and 95% confidence interval (CI) of the indirect effect were estimated using the Delta method. Mediation estimates were interpreted cautiously because the two-step MR framework relies on the validity of the underlying MR assumptions and may be affected by horizontal pleiotropy, shared genetic architecture, and correlated biological pathways. Therefore, the estimated mediation proportions were considered approximate contributions of potential intermediate pathways rather than definitive proportions of the clinical effect of IVDD-related traits on OA.27 Results Causal Estimates Between IVDD-Related Traits and OA This study indicates that IVDD and its related traits, such as ELDP, LDP, and sciatica, may increase the risk of knee and hip OA. Conversely, knee and hip osteoarthritis may also contribute to the accelerated progression of IVDD-related degeneration. Specifically, IVDD was found to increase or accelerate the occurrence or progression of OA subtypes such as HOA(OR=1.330, 95% CI:1.196–1.479, P=1.45E-07), THR(OR=1.367, 95% CI:1.196–1.561, P=4.17E-06), and KOA(OR=1.169, 95% CI:1.086–1.259, P=3.26E-05). Although IVDD also showed an association with TKR, the result was unstable due to significant pleiotropy and was therefore not included in the main interpretation (Figure 2A). Conversely, only HOA and IVDD(OR=1.152, 95% CI:1.039–1.277, P=0.007) exhibited a positive causal relationship (Figure 2B). Both ELDP and LDP were associated with increased odds of HOA(ELDP-HOA(OR=1.230, 95% CI:1.145–1.320, P=1.19E-08), LDP-HOA(OR=1.438, 95% CI:1.227–1.685, P=7.05E-06)) and THR(ELDP-THR(OR=1.286, 95% CI:1.149–1.440, P=1.24E-05), LDP-THR(OR=1.516, 95% CI:1.246–1.845, P=3.30E-05)) (Figure 2A). In the reverse-direction analyses, HOA and THR demonstrated positive causal associations with LDP(HOA-LDP(OR=1.161, 95% CI:1.034–1.304, P=0.012), THR-LDP(OR=1.074, 95% CI:1.004–1.148, P=0.038)), while ELDP exhibited an inverse association (Figure 2B). In addition, LDP exerted a unidirectional causal effect, increasing the risk of KOA(OR=1.133, 95% CI:1.006–1.276, P=0.040). Similarly, the MR analysis indicated that sciatica is associated with an increased risk of HOA(OR=1.539, 95% CI:1.211–1.956, P=0.000), THR(OR=1.456, 95% CI:1.102–1.925, P=0.008), and KOA(OR=1.390, 95% CI:1.090–1.774, P=0.008) (Figure 2A). After reversing the exposure and outcome, no causal relationship was observed between knee and hip OA and sciatica. (The reverse causal association between KOA and sciatica was unstable, and evidence of pleiotropy was detected; therefore, this result was not included in the main interpretation). The results remained stable across repeated IVW-based tests, after which further analyses were discontinued. Associations supported by multiple MR approaches and sensitivity analyses were considered more reliable. The MR-Egger intercept tests did not indicate significant directional horizontal pleiotropy (P > 0.05); however, given the limited statistical power of this approach, the absence of significant intercepts does not completely exclude potential pleiotropic effects. Cochran’s Q tests revealed significant heterogeneity among the selected instrumental variables in some analyses (P < 0.05). Therefore, random-effects IVW models were applied as the primary analysis approach. MR-PRESSO further identified potential outlier SNPs, and the causal estimates remained directionally consistent and statistically significant after removal of these outliers. For several associations, the MR-PRESSO distortion test showed no significant differences between the original and outlier-corrected estimates (P > 0.05), suggesting that the detected outliers had limited influence on the causal estimates. Conversely, although significant differences between the original and outlier-corrected estimates were observed for some associations (P < 0.05), these associations did not demonstrate statistically significant causal effects either before or after outlier correction (Figure 2A and 2B; Supplementary Figure S1–S4; Supplementary Table S1 and S2). Figure 2Results of Mendelian randomization analyses between IVDD and osteoarthritis, sarcopenia-related traits, and adiposity-, obesity-, and metabolic-related phenotypes. (A) Causal effects of IVDD-related traits on osteoarthritis phenotypes. (B) Reverse causal effects of osteoarthritis phenotypes on IVDD. (C) Causal effects of IVDD-related traits on sarcopenia-related phenotypes. (D) Causal effects of IVDD on adiposity-, obesity-, and metabolic-related phenotypes.The images A, B, C and D each display tables with columns: Exposure, Outcome, Method, P-value, Odds Ratio (95% Confidence Interval) and a forest plot with an odds ratio axis. A vertical reference line at odds ratio = 1 is used. Image A groups rows by exposures like IVDD, ELDP, LDP and sciatica, listing osteoarthritis outcomes. Image B reverses this, listing osteoarthritis phenotypes as exposures and IVDD as outcomes. Image C shows IVDD traits as exposures and sarcopenia-related outcomes. Image D lists IVDD, ELDP, LDP and sciatica as exposures with adiposity, obesity and metabolic outcomes. Across images, methods repeat per exposure-outcome pair. Many odds ratio points cluster near 1, with confidence intervals often crossing the reference line. Some points in Image D are notably to the right of the line. A footnote in images A and D notes P < 0.05 as statistically significant.Mendelian tables: odds ratio plots for IVDD, osteoarthritis, sarcopenia, metabolic traits.Abbreviations: IVW, inverse variance-weighted, OR, odds ratio, IVDD, intervertebral disc degeneration, ELDP, early lumbar disc prolapse, LDP, lumbar disc prolapse, HOA, hip osteoarthritis, THR, total hip replacement, KOA, knee osteoarthritis, TKR, total knee replacement, WBLM, Whole body lean mass, ALM, Appendicular lean mass. Results of Mendelian randomization analyses between IVDD and osteoarthritis, sarcopenia-related traits, and adiposity-, obesity-, and metabolic-related phenotypes. (A) Causal effects of IVDD-related traits on osteoarthritis phenotypes. (B) Reverse causal effects of osteoarthritis phenotypes on IVDD. (C) Causal effects of IVDD-related traits on sarcopenia-related phenotypes. (D) Causal effects of IVDD on adiposity-, obesity-, and metabolic-related phenotypes. The Causal Effect of IVDD-Related Traits on SP and Adiposity-, Obesity-, and Metabolic-Related Traits In the IVW MR analysis, IVDD-related traits may be associated with an elevated risk of sarcopenia.(IVDD-ALM(OR=1.073, 95% CI:1.005–1.147, P=0.036), IVDD-WBLM(OR=1.060, 95% CI:1.025–1.097, P=0.001), IVDD-Left Hand grip strength(OR=1.028, 95% CI:1.013–1.044, P=0.000), IVDD-Right Hand grip strength(OR=1.028, 95% CI:1.011–1.044, P=0.001), ELDP-WBLM(OR=1.029, 95% CI:1.005–1.054, P=0.020), ELDP-ALM(OR=1.058, 95% CI:1.013–1.106, P=0.012), LDP-Left Hand grip strength(OR=1.023, 95% CI:1.006–1.042, P=0.010), LDP-Right Hand grip strength(OR=1.026, 95% CI:1.009–1.044, P=0.003)). Although sciatica was causally associated with WBLM, instability and horizontal pleiotropy led to its exclusion from further interpretation (Figure 2C). Meanwhile, IVDD-related traits also exhibited stable positive causal associations with adiposity-, obesity-, and metabolic-related traits. The details are as follows: IVDD-BMI(OR=1.032, 95% CI:1.004–1.061, P=0.025), IVDD-WC(OR=1.050, 95% CI:1.022–1.079, P=0.000), IVDD-HC(OR=1.081, 95% CI:1.047–1.116, P=1.89E-06), IVDD-WBFM(OR=1.062, 95% CI:1.027–1.098, P=0.000), IVDD-BFP(OR=1.027, 95% CI:1.000–1.054, P=0.047), IVDD-BMR(OR=1.028, 95% CI:1.003–1.054, P=0.029), IVDD-BW(OR=1.024, 95% CI:1.002–1.047, P=0.024), ELDP-BMR(OR=1.028, 95% CI:1.003–1.054, P=0.029), ELDP-WBFM(OR=1.017, 95% CI:1.006–1.029, P=0.002), ELDP-HC(OR=1.014, 95% CI:1.001–1.027, P=0.030), ELDP-WC(OR=1.015, 95% CI:1.005–1.026, P=0.003), LDP-BW(OR=1.047, 95% CI:1.019–1.076, P=0.000), LDP-BMR(OR=1.035, 95% CI:1.000–1.070, P=0.049), LDP-WBFM(OR=1.052, 95% CI:1.006–1.101, P=0.027), LDP-BFP(OR=1.041, 95% CI:1.012–1.071, P=0.005), LDP-HC(OR=1.072, 95% CI:1.025–1.120, P=0.002), LDP-WC(OR=1.041, 95% CI:1.007–1.077, P=0.040), Sciatica-BMI(OR=1.190, 95% CI:1.123–1.260, P=3.08E-09), Sciatica-WC(OR=1.121, 95% CI:1.076–1.167, P=4.10E-08), Sciatica-HC(OR=1.153, 95% CI:1.110–1.197, P=1.23E-13), Sciatica-WBFM(OR=1.149, 95% CI:1.109–1.190, P=1.29E-14), Sciatica-BFP(OR=1.124, 95% CI:1.079–1.171, P=1.86E-08), Sciatica-BMR(OR=1.062, 95% CI:1.021–1.105, P=0.002). Associations supported by multiple MR approaches and sensitivity analyses were considered more reliable. The MR-Egger intercept tests did not indicate significant directional horizontal pleiotropy (P > 0.05); however, given the limited statistical power of this approach, the absence of significant intercepts does not completely exclude potential pleiotropic effects. Cochran’s Q tests revealed significant heterogeneity among the selected instrumental variables in some analyses (P < 0.05). Therefore, random-effects IVW models were applied as the primary analysis approach. MR-PRESSO further identified potential outlier SNPs, and the causal estimates remained directionally consistent and statistically significant after removal of these outliers. For several associations, the MR-PRESSO distortion test showed no significant differences between the original and outlier-corrected estimates (P > 0.05), suggesting that the detected outliers had limited influence on the causal estimates. Conversely, although significant differences between the original and outlier-corrected estimates were observed for some associations (P < 0.05), these associations did not demonstrate statistically significant causal effects either before or after outlier correction (Figure 2D, Supplementary Figure S5–S10 and Supplementary Table S3–S4). The Causal Effect of SP and Obesity on OA The MR analysis revealed that sarcopenia-related phenotypes were significantly associated with an increased risk of developing or progressing hip and knee osteoarthritis and their subtypes.(ALM-HOA(OR=1.138, 95% CI:1.068–1.212, P=5.71E-05), ALM-THR(OR=1.175, 95% CI:1.088–1.269, P=3.94E-05), ALM-KOA(OR=1.117, 95% CI:1.058–1.180, P=6.97E-05), ALM-TKR(OR=1.124, 95% CI:1.032–1.223, P=0.007), Left Hand grip strength-HOA(OR=1.433, 95% CI:1.130–1.817, P=0.003), Left Hand grip strength-THR(OR=1.564, 95% CI:1.155–2.120, P=0.004), WBLM-HOA(OR=1.537, 95% CI:1.370–1.723, P=2.17E-13), WBLM-THR(OR=1.569, 95% CI:1.355–1.817, P=1.80E-09), WBLM-KOA(OR=1.709, 95% CI:1.556–1.878, P=7.78E-29), WBLM-TKR(OR=1.836, 95% CI:1.588–2.122, P=2.01E-16) (Figure 3A). The causal analysis results for adiposity-, obesity-, and metabolic-related traits on OA are presented below: BMI-KOA(OR=1.711, 95% CI:1.591–1.840, P=1.87E-47), BMI-TKR(OR=2.048, 95% CI:1.838–2.281, P=1.36E-38), BW-KOA(OR=1.157, 95% CI:1.049–1.276, P=0.003), BW-TKR(OR=1.227, 95% CI:1.081–1.392, P=0.002), BMR-HOA(OR=1.543, 95% CI:1.377–1.730, P=1.02E-13), BMR-THR(OR=1.543, 95% CI:1.377–1.730, P=1.02E-13), BMR-KOA(OR=1.802, 95% CI:1.648–1.970, P=3.39E-38), BMR-TKR(OR=2.111, 95% CI:1.835–2.428, P=1.39E-25), WBFM-HOA(OR=1.493, 95% CI:1.382–1.611, P=1.47E-24), WBFM-THR(OR=1.505, 95% CI:1.371–1.653, P=1.04E-17), WBFM-KOA(OR=1.797, 95% CI:1.675–1.929, P=1.42E-59), WBFM-TKR(OR=2.146, 95% CI:1.928–2.390, P=3.97E-44), BFP-HOA(OR=1.558, 95% CI:1.389–1.749, P=4.78E-14), BFP-THR(OR=1.478, 95% CI:1.281–1.705, P=8.67E-08), BFP-KOA(OR=2.142, 95% CI:1.941–2.365, P=1.25E-51), BFP-TKR(OR=2.684, 95% CI:2.295–3.139, P=4.18E-35), HC-HOA(OR=1.462, 95% CI:1.333–1.604, P=9.02E-16), HC-THR(OR=1.463, 95% CI:1.305–1.640, P=7.07E-11), HC-KOA(OR=1.757, 95% CI:1.629–1.897, P=1.21E-47), HC-TKR(OR=2.093, 95% CI:1.871–2.342, P=6.01E-38), WC-TKR(OR=2.287, 95% CI:2.003–2.611, P=2.60E-34). Beyond the aforementioned results, several positive associations were not reported in this study because they exhibited horizontal pleiotropy. Associations supported by multiple MR approaches and sensitivity analyses were considered more reliable. The MR-Egger intercept tests did not indicate significant directional horizontal pleiotropy (P > 0.05); however, given the limited statistical power of this approach, the absence of significant intercepts does not completely exclude potential pleiotropic effects. Cochran’s Q tests revealed significant heterogeneity among the selected instrumental variables in some analyses (P < 0.05). Therefore, random-effects IVW models were applied as the primary analysis approach. MR-PRESSO further identified potential outlier SNPs, and the causal estimates remained directionally consistent and statistically significant after removal of these outliers. For several associations, the MR-PRESSO distortion test showed no significant differences between the original and outlier-corrected estimates (P > 0.05), suggesting that the detected outliers had limited influence on the causal estimates. Conversely, although significant differences between the original and outlier-corrected estimates were observed for some associations (P < 0.05), these associations did not demonstrate statistically significant causal effects either before or after outlier correction (Figure 3B, Supplementary Table S5 and S6). Figure 3Results of Mendelian randomization analyses between IVDD and osteoarthritis, sarcopenia-related traits, and adiposity-, obesity-, and metabolic-related phenotypes. (A) Causal effects of sarcopenia-related traits on osteoarthritis phenotypes. (B) Causal effects of adiposity-, obesity-, and metabolic-related phenotypes on osteoarthritis phenotypes.The forest plot tables show Mendelian randomization results for sarcopenia and adiposity-related exposures versus osteoarthritis outcomes. Exposures include ALM, WBLM, Left and Right Hand grip strength, Body mass index, Waist and Hip circumference, Whole body fat mass, Body fat percentage, Basal metabolic rate and Birth weight. Outcomes are HOA, THR, KOA and TKR, analyzed using Inverse variance weighted and MR Egger methods. OR values for sarcopenia exposures range from 0.748 to 1.836, with ALM showing significant associations with all outcomes. Adiposity exposures show OR values from 1.042 to 2.684, with Body mass index and Waist circumference having strong links to all outcomes. P-values less than 0.05 indicate statistical significance. The x-axis represents OR with a 95% confidence interval and the y-axis lists exposures and outcomes. Vertical reference lines at 1 help interpret the data.Two forest plots of odds ratio estimates for osteoarthritis outcomes across sarcopenia and adiposity traits.Abbreviations: IVW, inverse variance-weighted, OR, odds ratio, IVDD, intervertebral disc degeneration, ELDP, early lumbar disc prolapse, LDP, lumbar disc prolapse, HOA, hip osteoarthritis, THR, total hip replacement, KOA, knee osteoarthritis, TKR, total knee replacement, WBLM, Whole body lean mass, ALM, Appendicular lean mass. Results of Mendelian randomization analyses between IVDD and osteoarthritis, sarcopenia-related traits, and adiposity-, obesity-, and metabolic-related phenotypes. (A) Causal effects of sarcopenia-related traits on osteoarthritis phenotypes. (B) Causal effects of adiposity-, obesity-, and metabolic-related phenotypes on osteoarthritis phenotypes. Mediation Analysis The following section summarizes the mediating effects of sarcopenia-related traits and adiposity-, obesity-, and metabolic-related traits in the causal relationship between IVDD and lower-limb OA:IVDD-ALM-KOA(Mediation effect:0.007, 95% CI:0.000–0.015, proportion:5.0%), IVDD-WBLM-KOA(Mediation effect:0.031, 95% CI:-0.019–0.082, proportion:20.0%), IVDD-ALM-THR(Mediation effect:0.011, 95% CI:-0.002–0.025, proportion:3.7%), IVDD-WBLM-THR(Mediation effect:0.026, 95% CI:-0.040–0.092, proportion:8.4%), IVDD-Left Hand grip strength-THR(Mediation effect:0.013, 95% CI:-0.123–0.148, proportion:4.0%), ELDP-ALM-HOA(Mediation effect:0.007, 95% CI:-0.001–0.016, proportion:3.5%), ELDP-WBLM-HOA(Mediation effect:0.012, 95% CI:-0.037–0.062, proportion:6.0%), ELDP-ALM-THR(Mediation effect:0.009, 95% CI:-0.003–0.022, proportion:3.6%), ELDP-WBLM-THR(Mediation effect:0.013, 95% CI:-0.053–0.079, proportion:5.2%), IVDD-BW-KOA(Mediation effect:0.003, 95% CI:-0.011–0.018, proportion:2.2%), IVDD-BMR-KOA(Mediation effect:0.035, 95% CI:-0.018–0.088, proportion:22.4%), IVDD-WBFM-KOA(Mediation effect:0.035, 95% CI:-0.006–0.077, proportion:22.5%), IVDD-BFP-KOA(Mediation effect:0.020, 95% CI:-0.055–0.096, proportion:13.0%), IVDD-HC-KOA(Mediation effect:0.044, 95% CI:0.001–0.087, proportion:28.1%), IVDD-BMI-KOA(Mediation effect:0.017, 95% CI:-0.022–0.056, proportion:10.9%), IVDD-BMR-THR(Mediation effect:0.028, 95% CI:-0.038–0.095, proportion:9.0%), IVDD-WBFM-THR(Mediation effect:0.025, 95% CI:-0.014–0.063, proportion:7.9%), IVDD-HC-THR(Mediation effect:0.030, 95% CI:-0.014–0.073, proportion:9.5%), ELDP-BMR-HOA(Mediation effect:0.012, 95% CI:-0.038–0.062, proportion:5.8%), ELDP-WBFM-HOA(Mediation effect:0.007, 95% CI:-0.024–0.038, proportion:3.4%), ELDP-HC-HOA(Mediation effect:0.005, 95% CI:-0.030–0.040, proportion:2.6%), ELDP-BMR-THR(Mediation effect:0.013, 95% CI:-0.053–0.079, proportion:5.2%), ELDP-WBFM-THR(Mediation effect:0.007, 95% CI:-0.031–0.045, proportion:2.9%), ELDP-HC-THR(Mediation effect:0.005, 95% CI:-0.038–0.049, proportion:2.1%), LDP-BW-KOA(Mediation effect:0.007, 95% CI:-0.008–0.021, proportion:5.4%), LDP-BMR-KOA(Mediation effect:0.020, 95% CI:-0.033–0.073, proportion:16.1%), LDP-WBFM-KOA(Mediation effect:0.030, 95% CI:-0.012–0.071, proportion:24.0%), LDP-BFP-KOA(Mediation effect:0.031, 95% CI:-0.044–0.106, proportion:24.7%), LDP-HC-KOA(Mediation effect:0.039, 95% CI:-0.004–0.082, proportion:31.3%), Sciatica-BMR-HOA(Mediation effect:0.026, 95% CI:-0.023–0.076, proportion:6.1%), Sciatica-WBFM-HOA(Mediation effect:0.056, 95% CI:0.025–0.087, proportion:12.9%), Sciatica-BFP-HOA(Mediation effect:0.052, 95% CI:0.000–0.103, proportion:12.1%), Sciatica-HC-HOA(Mediation effect:0.054, 95% CI:0.019–0.090, proportion:12.5%), Sciatica-BMR-THR(Mediation effect:0.029, 95% CI:-0.038–0.095, proportion:7.6%), Sciatica-WBFM-THR(Mediation effect:0.057, 95% CI:-0.018–0.095, proportion:15.1%), Sciatica-BFP-THR(Mediation effect:0.046, 95% CI:-0.010–0.102, proportion:12.2%), Sciatica-HC-THR(Mediation effect:0.054, 95% CI:0.010–0.098, proportion:14.4%) (Table 2).Table 2The Results of the Mediation Analysis and the Proportion Mediated are ShownExposureOutcomeMediatorTotal EffectDirect EffectMediation Effect (95% CI)Mediated Proportion (%)IVDDKOAALM0.1560.1480.007(0.000,0.015)5.0WBLM0.1560.1250.031(−0.019,0.082)20.0THRALM0.3120.3010.011(−0.002,0.025)3.7WBLM0.3120.2860.026(−0.040,0.092)8.4Left Hand grip strength0.3120.3000.013(−0.123,0.148)4.0ELDPHOAALM0.2070.1990.007(−0.001,0.016)3.5WBLM0.2070.1940.012(−0.037,0.062)6.0THRALM0.2520.2420.009(−0.003,0.022)3.6WBLM0.2520.2390.013(−0.053,0.079)5.2IVDDKOABirth weight0.1560.1530.003(−0.011,0.018)2.2Basal metabolic rate0.1560.1210.035(−0.018,0.088)22.4Whole body fat mass0.1560.1210.035(−0.006,0.077)22.5Body fat percentage0.1560.1360.020(−0.055,0.096)13.0Hip circumference0.1560.1120.044(0.001,0.087)28.1BMI0.1560.1390.017(−0.022,0.056)10.9THRBasal metabolic rate0.3120.2840.028(−0.038,0.095)9.0Whole body fat mass0.3120.2880.025(−0.014,0.063)7.9Hip circumference0.3120.2830.030(−0.014,0.073)9.5ELDPHOABasal metabolic rate0.2070.1950.012(−0.038,0.062)5.8Whole body fat mass0.2070.2000.007(−0.024,0.038)3.4Hip circumference0.2070.2010.005(−0.030,0.040)2.6THRBasal metabolic rate0.2520.2390.013(−0.053,0.079)5.2Whole body fat mass0.2520.2440.007(−0.031,0.045)2.9Hip circumference0.2520.2460.005(−0.038,0.049)2.1LDPKOABirth weight0.1250.1180.007(−0.008,0.021)5.4Basal metabolic rate0.1250.1050.020(−0.033,0.073)16.1Whole body fat mass0.1250.0950.030(−0.012,0.071)24.0Body fat percentage0.1250.0940.031(−0.044,0.106)24.7Hip circumference0.1250.0860.039(−0.004,0.082)31.3SciaticaHOABasal metabolic rate0.4310.4050.026(−0.023,0.076)6.1Whole body fat mass0.4310.3750.056(0.025,0.087)12.9Body fat percentage0.4310.3790.052(0.000,0.103)12.1Hip circumference0.4310.3770.054(0.019,0.090)12.5THRBasal metabolic rate0.3760.3470.029(−0.038,0.095)7.6Whole body fat mass0.3760.3190.057(0.018,0.095)15.1Body fat percentage0.3760.3300.046(−0.010,0.102)12.2Hip circumference0.3760.3220.054(0.010,0.098)14.4 Table 2The Results of the Mediation Analysis and the Proportion Mediated are ShownExposureOutcomeMediatorTotal EffectDirect EffectMediation Effect (95% CI)Mediated Proportion (%)IVDDKOAALM0.1560.1480.007(0.000,0.015)5.0WBLM0.1560.1250.031(−0.019,0.082)20.0THRALM0.3120.3010.011(−0.002,0.025)3.7WBLM0.3120.2860.026(−0.040,0.092)8.4Left Hand grip strength0.3120.3000.013(−0.123,0.148)4.0ELDPHOAALM0.2070.1990.007(−0.001,0.016)3.5WBLM0.2070.1940.012(−0.037,0.062)6.0THRALM0.2520.2420.009(−0.003,0.022)3.6WBLM0.2520.2390.013(−0.053,0.079)5.2IVDDKOABirth weight0.1560.1530.003(−0.011,0.018)2.2Basal metabolic rate0.1560.1210.035(−0.018,0.088)22.4Whole body fat mass0.1560.1210.035(−0.006,0.077)22.5Body fat percentage0.1560.1360.020(−0.055,0.096)13.0Hip circumference0.1560.1120.044(0.001,0.087)28.1BMI0.1560.1390.017(−0.022,0.056)10.9THRBasal metabolic rate0.3120.2840.028(−0.038,0.095)9.0Whole body fat mass0.3120.2880.025(−0.014,0.063)7.9Hip circumference0.3120.2830.030(−0.014,0.073)9.5ELDPHOABasal metabolic rate0.2070.1950.012(−0.038,0.062)5.8Whole body fat mass0.2070.2000.007(−0.024,0.038)3.4Hip circumference0.2070.2010.005(−0.030,0.040)2.6THRBasal metabolic rate0.2520.2390.013(−0.053,0.079)5.2Whole body fat mass0.2520.2440.007(−0.031,0.045)2.9Hip circumference0.2520.2460.005(−0.038,0.049)2.1LDPKOABirth weight0.1250.1180.007(−0.008,0.021)5.4Basal metabolic rate0.1250.1050.020(−0.033,0.073)16.1Whole body fat mass0.1250.0950.030(−0.012,0.071)24.0Body fat percentage0.1250.0940.031(−0.044,0.106)24.7Hip circumference0.1250.0860.039(−0.004,0.082)31.3SciaticaHOABasal metabolic rate0.4310.4050.026(−0.023,0.076)6.1Whole body fat mass0.4310.3750.056(0.025,0.087)12.9Body fat percentage0.4310.3790.052(0.000,0.103)12.1Hip circumference0.4310.3770.054(0.019,0.090)12.5THRBasal metabolic rate0.3760.3470.029(−0.038,0.095)7.6Whole body fat mass0.3760.3190.057(0.018,0.095)15.1Body fat percentage0.3760.3300.046(−0.010,0.102)12.2Hip circumference0.3760.3220.054(0.010,0.098)14.4 The Results of the Mediation Analysis and the Proportion Mediated are Shown Discussion In this MR study, we provide genetic evidence supporting a potential causal relationship between IVDD-related traits and OA, two major contributors to chronic musculoskeletal pain and disability. Our findings suggest that genetically predicted liability to IVDD-related traits is associated with increased susceptibility to OA outcomes, including hip and knee OA and joint replacement phenotypes. Importantly, these findings reflect the effects of lifelong genetic liability to IVDD-related traits rather than the direct consequences of clinically established IVDD. Therefore, they do not necessarily indicate that the presence, severity, or progression of clinical IVDD directly causes OA development or progression, but rather suggest that genetic susceptibility to spinal degeneration may contribute to interindividual differences in OA susceptibility. The observed associations may extend beyond isolated structural degeneration and suggest shared biological and functional processes underlying musculoskeletal degeneration and chronic pain susceptibility. Spinal degeneration can contribute to both nociceptive and neuropathic pain through nerve root compression, local inflammation, and neural sensitization.28,29 Similarly, OA-related pain is driven not only by cartilage degeneration but also by synovial inflammation, subchondral bone changes, and peripheral sensitization.30,31 These overlapping mechanisms may partly explain the relationship between spinal degeneration and large-joint OA. However, the present MR findings do not establish a specific shared molecular mechanism and should therefore be considered hypothesis-generating. Our two-step MR analysis further suggests that adiposity-, obesity-, metabolic- and sarcopenia-related traits may represent potential intermediate pathways contributing to the association between genetic liability to IVDD-related traits and OA outcomes. Obesity may influence musculoskeletal health through increased mechanical loading and pro-inflammatory adipokines, whereas sarcopenia may impair joint stabilization and alter biomechanical stress distribution, potentially contributing to functional impairment and pain perception.32–34 Nevertheless, these mediation estimates should not be interpreted as definitive evidence of a complete biological pathway, as they may also reflect shared genetic determinants, correlated biological processes, or pleiotropic pathways influencing both IVDD-related traits and OA susceptibility. Thus, the findings provide evidence for potential intermediate pathways rather than confirmation of a sequential biological mechanism. Importantly, obesity and sarcopenia are unlikely to account for all pathways linking IVDD-related traits with OA. Other potential mediators, including systemic inflammation, metabolic dysfunction, physical activity, muscle strength and function, adipose tissue-derived signaling factors, and bone or connective-tissue metabolism, may also contribute to this association.11,12 The present study focused on adiposity-, obesity-, metabolic- and sarcopenia-related traits based on their biological plausibility and the availability of suitable GWAS summary statistics and therefore did not comprehensively evaluate all potential mediators. Future studies incorporating a broader range of candidate mediators and using multivariable or sequential mediation MR approaches may help clarify the complex pathways underlying the IVDD–OA association. Although the estimated effects of IVDD-related traits on adiposity-, obesity-, metabolic- and sarcopenia-related traits were relatively modest, they may still have biological relevance. IVDD and OA are complex polygenic disorders influenced by numerous genetic and environmental factors, and individual variants generally exert small effects. Therefore, the observed estimates should be interpreted as modest contributions to disease susceptibility rather than deterministic effects. From a clinical perspective, these findings support the concept that spinal degeneration and OA should not be considered isolated structural disorders, but rather components of an integrated chronic pain system. Targeting modifiable risk factors such as obesity and muscle loss may therefore have potential implications for reducing pain burden and disability in patients with musculoskeletal degeneration. However, several limitations should be acknowledged. First, significant heterogeneity was observed in several MR analyses, suggesting that the genetic instruments may not capture a single homogeneous causal effect. Although random-effects IVW models and sensitivity analyses, including MR-Egger regression, MR-PRESSO, and leave-one-out analyses, were used to assess robustness, residual heterogeneity reduces confidence in interpreting the estimates as reflecting a uniform biological mechanism. The causal estimates should therefore be regarded as average effects of genetic liability across multiple instruments. Second, although MR can reduce conventional confounding and reverse causation, violations of its core assumptions cannot be completely excluded. MR-Egger intercept tests did not provide evidence of statistically significant directional horizontal pleiotropy; however, a non-significant intercept does not definitively exclude horizontal pleiotropy, particularly when statistical power is limited. MR-PRESSO and other sensitivity analyses can identify certain forms of pleiotropy or influential variants but cannot completely eliminate balanced pleiotropy or complex pleiotropic pathways. Potential effects of population structure, assortative mating, and residual bias also cannot be entirely excluded. Third, IVDD, LDP, ELDP, and sciatica represent overlapping but distinct manifestations or consequences of spinal degeneration. Their observed associations with OA may therefore partly reflect shared genetic architecture rather than completely independent biological effects. Although these related phenotypes provide complementary evidence, their individual causal contributions remain difficult to distinguish. Future studies using multivariable MR and individual-level longitudinal data are needed to clarify their independent effects. Fourth, the mediation findings should be interpreted cautiously. The two-step MR framework estimates potential indirect pathways based on genetic associations between IVDD-related traits, candidate mediators, and OA, but does not fully capture the temporal and biological complexity of clinical mediation. The mediation estimates may also be influenced by pleiotropy, correlated genetic architecture, and violations of MR assumptions. Therefore, the estimated mediation proportions should be regarded as approximate contributions of potential intermediate pathways rather than definitive measures of the proportion of clinical OA attributable to obesity or sarcopenia. Finally, MR estimates reflect lifelong genetic liability rather than the consequences of clinically established disease onset, severity, duration, or progression. Consequently, this study cannot determine whether the development or progression of clinical IVDD directly increases subsequent OA risk. In addition, most included GWAS datasets were derived from individuals of European ancestry, which may limit generalizability to other populations. Replication in diverse populations and validation using prospective clinical and mechanistic studies are warranted. Overall, this study provides genetic evidence supporting a potential relationship between IVDD-related traits and OA and suggests that adiposity-, obesity-, metabolic- and sarcopenia-related traits may represent potential intermediate pathways. However, these findings should be interpreted within the limitations of MR and mediation analyses. Further integrative studies combining genetic, longitudinal clinical, imaging, and mechanistic data are needed to clarify whether and how spinal degeneration contributes to the development and progression of large-joint OA. Conclusion This study provides genetic evidence supporting a potential link between spinal degeneration and chronic pain-related osteoarthritis. Obesity and sarcopenia may partially mediate this relationship, suggesting shared biological mechanisms underlying chronic musculoskeletal pain and disability.