Effect of spinal fusion length on hip and knee osteoarthritis: a comparative cohort study of no fusion, short-segment fusion, and long-segment fusion Authors: Yusuf Sülek, İsmail Demirkale, Rodi Ertoğrul, Hacı Mustafa Özdemir Published: 2026 ABSTRACT PurposeLumbar fusion alters spinal–pelvic biomechanics and may influence degenerative changes in the lower extremities. This study investigated whether fusion length is associated with radiographic progression of hip and knee osteoarthritis.MethodsIn this retrospective, single-center cohort study, patients aged 40–65 years who underwent lumbar surgery between 2010 and 2021 were grouped as no fusion (N), short fusion (S; ≤3 levels), or long fusion (L; ≥4 levels). Hip joint space width was assessed using minimum joint space (MJS) and superointermediate joint space (SJS), including femoral head–standardized measures (sMJS and sSJS). Knee medial joint space width (KMJS) was measured when available. Spinopelvic parameters (LL, PI, PT, SS, PI–LL) and functional outcomes (ODI, HHS, WOMAC) were recorded. Multivariable linear and logistic regression analyses were performed to identify independent predictors of joint space narrowing and osteoarthritis risk.ResultsHip joint space narrowing differed significantly among groups, with the highest annual narrowing rates in the long-fusion group, followed by the short-fusion group, and the lowest in the no-fusion group (p < 0.001). In contrast, knee KL grade change and annual KMJS narrowing did not differ significantly between groups (p > 0.05). In logistic regression, each additional fused level increased the odds of postoperative MJS ≤ 2.0 mm by 16–19%, and higher CEA was also associated with increased risk.ConclusionLong-segment lumbar fusion (≥ 4 levels) was associated with greater hip joint space narrowing and greater radiographic progression of hip osteoarthritis, whereas knee degeneration did not differ by fusion status. Patients undergoing extensive fusion may benefit from closer hip surveillance to facilitate earlier recognition of degeneration.Supplementary InformationThe online version contains supplementary material available at 10.1007/s00402-026-06360-x. Introduction As the population ages, an increasing number of patients become candidates for spinal fusion due to degenerative spinal conditions. With appropriate patient selection and indications, spinal fusion can reduce symptoms and improve quality of life[1, 2]. Accordingly, the use of lumbar fusion surgery has increased substantially in recent years [3, 4]. Lumbar fusion can substantially alter the biomechanics of the spine, pelvis, and lower extremities[5, 6]. Maintaining sagittal alignment depends on the coordinated relationship between spinal curvatures and pelvic–lower extremity parameters. This interaction is commonly referred to as spinopelvic alignment, and disruption of this balance was first described as the “hip–spine syndrome”[7]. Previous studies have established a connection between spinopelvic alignment and hip osteoarthritis. A study on spinopelvic alignment in patients with coxarthrosis and low back pain demonstrated that a high pelvic incidence in young individuals may contribute to coxarthrosis at older ages [8]. Although improvement in both low back pain and spinal function have been observed after total hip arthroplasty in patients with coxarthrosis, it has been reported that patients who undergo longer lumbar spinal fusion have a significantly increased risk of undergoing total hip arthroplasty surgery at an older age [9, 10]. It has been reported that restriction of sacral movements due to adjacent segment degeneration after lumbar spinal fusion may accelerate the degenerative processes and cause coxarthrosis [11, 12]. However, most prior studies quantified hip degeneration primarily with MJS and did not routinely evaluate location-specific superior joint space measurements that may detect earlier change. In addition, non-standardized joint space measurements are vulnerable to magnification and body-size effects, potentially reducing comparability across radiographs and patients. This study aims to address these shortcomings by introducing SJS as an additional parameter and using standardized MJS assessments to more precisely evaluate the effect of spinal fusion on hip and knee joint space narrowing. We hypothesized that increasing fusion length would be associated with greater hip and knee joint space narrowing and a higher risk of osteoarthritis. The aim of the study is to determine the effect of the fusion level on hip and knee joint space narrowing and the risk of coxarthrosis in patients who have not undergone lumbar spinal fusion and who have undergone short or long segment fusion. Materials and methods Study design This study was designed as a retrospective, single-center comparative cohort study conducted between 2010 and 2021. The study period was selected to minimize variability in surgical practice and radiographic acquisition, and all included patients had at least 2 years of radiographic follow-up. The study protocol was approved by the institutional ethics committee (approval date: 10/10/2023; ID: 4117) and was conducted in accordance with the Declaration of Helsinki. All participants provided informed consent for the use of their clinical data. The study was designed and reported in accordance with the STROBE guidelines for observational cohort studies. Setting The study retrospectively analyzed medical records of patients treated surgically at a single tertiary referral center between 2010 and 2021. Participants The study included patients aged 40–65 years who underwent lumbar and/or lumbosacral surgery with or without fusion. This age range was selected to reduce confounding related to pre-existing secondary causes of hip osteoarthritis in younger patients and advanced primary degenerative changes in older patients. Eligible patients were required to have standing radiographs obtained preoperatively and at a minimum of 2 years postoperatively. Patients were excluded if radiographs were unsuitable for measurement; if moderate-to-severe hip or knee osteoarthritis was present (KL ≥ 2); if hip dysplasia was suspected (acetabular dysmorphology or center-edge angle < 15°); if prior or interval hip/knee surgery occurred during follow-up; or if patients had spinal tumor, fracture, rheumatic/autoimmune disease, or spinal cord injury (Fig. 1). Fig. 1Flow chart showing patients included and excluded from the study Flow chart showing patients included and excluded from the study Variables The primary variables were radiological and functional outcomes. Radiographs were independently reviewed to confirm eligibility and measurement feasibility. The study size was determined by the number of patients who met the inclusion criteria during the defined study period (2010–2021). The patients were divided into three groups, as illustrated in Fig. 2. Fig. 2Grouping of all patients according to fusion groups, nh: number of hips measured, N: number of patients, nk: number of knees measured, nf: number of patients with functional evaluation Grouping of all patients according to fusion groups, nh: number of hips measured, N: number of patients, nk: number of knees measured, nf: number of patients with functional evaluation Minimum joint width (MJS), superointermediate joint width (SJS), and femoral head diameter (D) were measured at the same location before surgery and at least two years after surgery. Center-edge angle (CEA) and sharp angle (SA) were also measured before surgery. In patients with radiographs of both knees, knee medial joint width (KMJS) and knee lateral joint width (KLJS) were measured preoperatively and after at least two years of follow-up (Fig. 3). KL was used to grade both hip and knee radiographs. Previously described and validated methods were used for these measurements (Fig. 3) [13–17]. Fig. 3Radiographic measurement of hip and knee joint space parameters on standing anteroposterior radiographs. Minimum joint space width (MJS), superointermediate joint space width (SJS), femoral head diameter (D), and medial and lateral knee joint space widths (KMJS and KLJS) are demonstrated Radiographic measurement of hip and knee joint space parameters on standing anteroposterior radiographs. Minimum joint space width (MJS), superointermediate joint space width (SJS), femoral head diameter (D), and medial and lateral knee joint space widths (KMJS and KLJS) are demonstrated Annual narrowing rates were calculated as follows \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{F}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{u}\mathrm{l}\mathrm{a}\:1=\frac{\left[\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{M}\mathrm{J}\mathrm{S}\:\right(\mathrm{m}\mathrm{m}\left)\right]\:-\:\left[\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{M}\mathrm{J}\mathrm{S}\:\right(\mathrm{m}\mathrm{m}\left)\right]}{\mathrm{F}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{o}\mathrm{w}-\mathrm{u}\mathrm{p}\:\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{o}\mathrm{d}\:\left(\mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}\mathrm{s}\right)}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{F}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{u}\mathrm{l}\mathrm{a}\:2=\frac{\frac{\left[\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{M}\mathrm{J}\mathrm{S}\:\left(\mathrm{m}\mathrm{m}\right)\right]}{\left[\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{D}\:\left(\mathrm{m}\mathrm{m}\right)\right]}-\frac{\left[\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\:\mathrm{M}\mathrm{J}\mathrm{S}\:\left(\mathrm{m}\mathrm{m}\right)\right]}{\left[\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\:\mathrm{D}\left(\mathrm{m}\mathrm{m}\right)\right]}}{\mathrm{F}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{o}\mathrm{w}-\mathrm{u}\mathrm{p}\:\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{o}\mathrm{d}\:\left(\mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}\mathrm{s}\right)}\times\:1000$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{F}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{u}\mathrm{l}\mathrm{a}\:3=\frac{\left[\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{S}\mathrm{J}\mathrm{S}\:\right(\mathrm{m}\mathrm{m}\left)\right]\:-\:\left[\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\:\mathrm{S}\mathrm{J}\mathrm{S}\:\right(\mathrm{m}\mathrm{m}\left)\right]}{\mathrm{F}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{o}\mathrm{w}-\mathrm{u}\mathrm{p}\:\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{o}\mathrm{d}\:\left(\mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}\mathrm{s}\right)}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{F}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{u}\mathrm{l}\mathrm{a}\:4=\frac{\frac{\left[\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{S}\mathrm{J}\mathrm{S}\:\left(\mathrm{m}\mathrm{m}\right)\right]}{\left[\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{D}\:\left(\mathrm{m}\mathrm{m}\right)\right]}-\frac{\left[\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\:\mathrm{S}\mathrm{J}\mathrm{S}\:\left(\mathrm{m}\mathrm{m}\right)\right]}{\left[\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\:\mathrm{D}\left(\mathrm{m}\mathrm{m}\right)\right]}}{\mathrm{F}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{o}\mathrm{w}-\mathrm{u}\mathrm{p}\:\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{o}\mathrm{d}\:\left(\mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}\mathrm{s}\right)}\times\:1000$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{F}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{u}\mathrm{l}\mathrm{a}\:5=\frac{\left[\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{K}\mathrm{M}\mathrm{J}\mathrm{S}\:\right(\mathrm{m}\mathrm{m}\left)\right]\:-\:\left[\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\:\mathrm{K}\mathrm{M}\mathrm{J}\mathrm{S}\:\right(\mathrm{m}\mathrm{m}\left)\right]}{\mathrm{F}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{o}\mathrm{w}-\mathrm{u}\mathrm{p}\:\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{o}\mathrm{d}\:\left(\mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}\mathrm{s}\right)}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{F}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{u}\mathrm{l}\mathrm{a}\:6=\frac{\left[\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{L}\mathrm{M}\mathrm{J}\mathrm{S}\:\right(\mathrm{m}\mathrm{m}\left)\right]\:-\:\left[\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\:\mathrm{L}\mathrm{M}\mathrm{J}\mathrm{S}\:\right(\mathrm{m}\mathrm{m}\left)\right]}{\mathrm{F}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{o}\mathrm{w}-\mathrm{u}\mathrm{p}\:\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{o}\mathrm{d}\:\left(\mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}\mathrm{s}\right)}$$\end{document} Spinopelvic parameters measured included lumbar lordosis (LL), pelvic incidence (PI), pelvic tilt (PT), sacral slope (SS), and pelvic incidence–lumbar lordosis mismatch (PI–LL). Lumbosacral fusion was defined as surgical fusion including the L5–S1 segment, whereas patients without L5–S1 inclusion were considered to have a mobile lumbosacral junction. Functional outcomes were assessed using the Oswestry Disability Index (ODI) for spine-related disability, the Harris Hip Score (HHS) for hip-specific evaluation, and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) for knee-related health-related quality of life. Statistical methods Statistical analysis was performed using IBM SPSS Statistics software, version 27. Chi-square test was employed for categorical variables. Shapiro-Wilk test was used to assess the normal distribution of variables within groups. One-way ANOVA was used for normally distributed variables, while non-normally distributed variables were analyzed via Kruskal-Wallis test. Linear and logistic multiple regression analyses were performed to assess independent predictors of joint space narrowing rates and osteoarthritis risk. Intraclass correlation coefficients (ICC) were used to examine inter-rater agreement. In all analyses, p-value < 0.05 was considered statistically significant. Results The cohort comprised 178 patients who were classified into three groups according to fusion status. Figure 2 illustrates the grouping of patients, including the number of hips, knees, and patients who underwent functional evaluation. Table 1 summarizes demographic characteristics, surgical indications, fusion levels, presence of lumbosacral fusion, follow-up duration, and preoperative CEA and SA measurements. Table 1Demographic data of the patients, indications for surgery, number of levels fused, lumbosacral fusion, follow-up periods, preoperative and postoperative spinopelvic parameter measurements (CEA, SA), postoperative and preoperative KL differences, and annual narrowing rates for MJS, SJS, standardized (sMJS, sSJS), KMJS, and KLJS according to patient groupsAll patients (nh=356, N = 178, nk=96)Group N (no fusion) (nh=96, N = 48, nk=35)Group S (≤ 3 levels of fusion) (nh=144, N = 72, nk=37)Group L (≥ 4 levels of fusion) (nh=116, N = 58, nk=35)p valueAge (years)55,9 ± 7,857,3 ± 7,854,7 ± 7,956,1 ± 7,50,3Gender (male: female)40:13815:3314:5811:470,121BMI (kg/m²)26,3 ± 1,226,5 ± 1,326,2 ± 1,126,2 ± 1,30,406Surgical Indications< 0,001 Lumbar spinal stenosis33,10%4,20%45,80%41,30% Disc disease25,80%93,80%1,40%0,00% Spondylolisthesis20,20%2,10%47,20%1,70% Degenerative disc disease10,70%0,00%47,20%25,90% Adult spinal deformity10,10%0,00%0,00%31,00%Number of spinal fusion levels3,5 ± 4,101,8 ± 0,68,6 ± 3,5Lumbosacral fusion22,50%037,50%22,40%Follow-up period (years)4,4 ± 1,75 ± 1,94,1 ± 1,64,2 ± 1,5< 0,001Preoperative CEA (degree)35,8 ± 7,637,7 ± 7,935,1 ± 7,335,1 ± 7,60,014Preoperative SA (degree)36,5 ± 4,635,9 ± 4,137,2 ± 4,636,2 ± 4,90,075Difference between postoperative and preoperative hip KL0,315 ± 0,500,25 ± 0,450,65 ± 0,56< 0,001MJS annual narrowing rates (mm/y)−0,095 ± 0,154−0,008 ± 0098−0,065 ± 0,135−0,0204 ± 0,152< 0,001SJS annual narrowing rates (mm/y)−0,105 ± 0,142−0,032 ± 0,062−0,073 ± 0,204−0,204 ± 0,159< 0,001sMJS annual narrowing rates (/y)−19,299 ± 28,193−2,192 ± 20,293−15,311 ± 24,613−38,407 ± 26,976< 0,001sSJS annual narrowing rates (/y)−21,140 ± 24,081−6,591 ± 8,069−16,987 ± 21,656−38,334 ± 25,624< 0,001Difference between postoperative and preoperative knee KL0,333 ± 0,7630,343 ± 0,6840,459 ± 0,8030,125 ± 0,7970,248KMJS annual narrowing rates (mm/y)−1,912 ± 2,168−1,617 ± 1,708−1,842 ± 2,373−2,451 ± 2,4170,416Preoperative (degree) LL37,4 ± 16,733,8 ± 16,242,7 ± 14,133,9 ± 18,4< 0,001 SS33,7 ± 10,931,1 ± 9,737 ± 9,431,8 ± 12,5< 0,001 PT21 ± 10,319,1 ± 7,319,2 ± 9,324,9 ± 12,3< 0,001 PI56,1 ± 13,250 ± 11,259,8 ± 12,256,6 ± 14,2< 0,001 PI-LL15 ± 14,611,9 ± 13,214,08 ± 12,718,7 ± 17,20,018Postoperative (degree) LL37,3 ± 15,834 ± 1744,6 ± 12,930,9 ± 14,3< 0,001 SS32,5 ± 10,829,8 ± 11,736,5 ± 9,329,7 ± 10,2< 0,001 PT19,4 ± 10,918,9 ± 11,123,2 ± 12,727,4 ± 11,70,232 PI56,63 ± 13,150,9 ± 1159,7 ± 12,456,6 ± 14,1< 0,001 PI-LL17,5 ± 15,515,6 ± 15,813,3 ± 12,821,2 ± 16,2< 0,001(nh: number of hips measured, N: number of patients, nk: number of knees measured) Table 1Demographic data of the patients, indications for surgery, number of levels fused, lumbosacral fusion, follow-up periods, preoperative and postoperative spinopelvic parameter measurements (CEA, SA), postoperative and preoperative KL differences, and annual narrowing rates for MJS, SJS, standardized (sMJS, sSJS), KMJS, and KLJS according to patient groupsAll patients (nh=356, N = 178, nk=96)Group N (no fusion) (nh=96, N = 48, nk=35)Group S (≤ 3 levels of fusion) (nh=144, N = 72, nk=37)Group L (≥ 4 levels of fusion) (nh=116, N = 58, nk=35)p valueAge (years)55,9 ± 7,857,3 ± 7,854,7 ± 7,956,1 ± 7,50,3Gender (male: female)40:13815:3314:5811:470,121BMI (kg/m²)26,3 ± 1,226,5 ± 1,326,2 ± 1,126,2 ± 1,30,406Surgical Indications< 0,001 Lumbar spinal stenosis33,10%4,20%45,80%41,30% Disc disease25,80%93,80%1,40%0,00% Spondylolisthesis20,20%2,10%47,20%1,70% Degenerative disc disease10,70%0,00%47,20%25,90% Adult spinal deformity10,10%0,00%0,00%31,00%Number of spinal fusion levels3,5 ± 4,101,8 ± 0,68,6 ± 3,5Lumbosacral fusion22,50%037,50%22,40%Follow-up period (years)4,4 ± 1,75 ± 1,94,1 ± 1,64,2 ± 1,5< 0,001Preoperative CEA (degree)35,8 ± 7,637,7 ± 7,935,1 ± 7,335,1 ± 7,60,014Preoperative SA (degree)36,5 ± 4,635,9 ± 4,137,2 ± 4,636,2 ± 4,90,075Difference between postoperative and preoperative hip KL0,315 ± 0,500,25 ± 0,450,65 ± 0,56< 0,001MJS annual narrowing rates (mm/y)−0,095 ± 0,154−0,008 ± 0098−0,065 ± 0,135−0,0204 ± 0,152< 0,001SJS annual narrowing rates (mm/y)−0,105 ± 0,142−0,032 ± 0,062−0,073 ± 0,204−0,204 ± 0,159< 0,001sMJS annual narrowing rates (/y)−19,299 ± 28,193−2,192 ± 20,293−15,311 ± 24,613−38,407 ± 26,976< 0,001sSJS annual narrowing rates (/y)−21,140 ± 24,081−6,591 ± 8,069−16,987 ± 21,656−38,334 ± 25,624< 0,001Difference between postoperative and preoperative knee KL0,333 ± 0,7630,343 ± 0,6840,459 ± 0,8030,125 ± 0,7970,248KMJS annual narrowing rates (mm/y)−1,912 ± 2,168−1,617 ± 1,708−1,842 ± 2,373−2,451 ± 2,4170,416Preoperative (degree) LL37,4 ± 16,733,8 ± 16,242,7 ± 14,133,9 ± 18,4< 0,001 SS33,7 ± 10,931,1 ± 9,737 ± 9,431,8 ± 12,5< 0,001 PT21 ± 10,319,1 ± 7,319,2 ± 9,324,9 ± 12,3< 0,001 PI56,1 ± 13,250 ± 11,259,8 ± 12,256,6 ± 14,2< 0,001 PI-LL15 ± 14,611,9 ± 13,214,08 ± 12,718,7 ± 17,20,018Postoperative (degree) LL37,3 ± 15,834 ± 1744,6 ± 12,930,9 ± 14,3< 0,001 SS32,5 ± 10,829,8 ± 11,736,5 ± 9,329,7 ± 10,2< 0,001 PT19,4 ± 10,918,9 ± 11,123,2 ± 12,727,4 ± 11,70,232 PI56,63 ± 13,150,9 ± 1159,7 ± 12,456,6 ± 14,1< 0,001 PI-LL17,5 ± 15,515,6 ± 15,813,3 ± 12,821,2 ± 16,2< 0,001(nh: number of hips measured, N: number of patients, nk: number of knees measured) Demographic data of the patients, indications for surgery, number of levels fused, lumbosacral fusion, follow-up periods, preoperative and postoperative spinopelvic parameter measurements (CEA, SA), postoperative and preoperative KL differences, and annual narrowing rates for MJS, SJS, standardized (sMJS, sSJS), KMJS, and KLJS according to patient groups (nh: number of hips measured, N: number of patients, nk: number of knees measured) Comparative analysis demonstrated statistically significant differences among groups in postoperative–preoperative hip Kellgren–Lawrence (KL) grade differences and in annual hip joint space narrowing rates (MJS, SJS, sMJS, and sSJS) (p < 0.001; Table 1). The highest annual narrowing rates were observed in the long-segment fusion group, followed by the short-segment fusion group, while the no-fusion group demonstrated the lowest rates (Fig. 4). In contrast, no statistically significant differences were identified between groups with respect to knee KL grade differences or annual KMJS narrowing rates (p > 0.05; Table 1). Fig. 4Annual hip joint space narrowing rates according to fusion groups. Box plots demonstrate MJS, SJS, and their femoral head–standardized annual narrowing rates (sMJS and sSJS) in Group N (no fusion), Group S (≤ 3 fused levels), and Group L (≥ 4 fused levels) Annual hip joint space narrowing rates according to fusion groups. Box plots demonstrate MJS, SJS, and their femoral head–standardized annual narrowing rates (sMJS and sSJS) in Group N (no fusion), Group S (≤ 3 fused levels), and Group L (≥ 4 fused levels) Significant differences in spinopelvic parameters were observed between groups in preoperative LL, SS, PT, PI, PI–LL mismatch, and in postoperative LL, SS, PI, and PI–LL values (p < 0.001; Table 1). Interobserver reliability was high, with ICC values of 0.96–0.99 for hip measurements, 0.99 for knee measurements, and 0.98–0.99 for CEA and SA. Multiple linear regression analyses were performed to determine independent predictors of hip joint degeneration. Patient group and number of fusion levels were independently associated with postoperative–preoperative hip KL grade differences. Increasing fusion length and presence of lumbosacral fusion were also significantly associated with greater radiographic progression. For annual narrowing rates, patient group remained an independent predictor of MJS, SJS, sMJS, and sSJS narrowing. Higher BMI was associated with increased SJS narrowing, while postoperative PT demonstrated a negative association with SJS narrowing. Postoperative SS was positively associated with sMJS narrowing (Table 2) (Fig. 5). Increasing number of fusion levels, higher CEA, and higher SA were significantly associated with greater sSJS narrowing (Table 3) (Fig. 5). Table 2Regression model for sMJS where annual narrowing is taken as dependent variable and other variables are taken as independent variablesModelNon-Standard CoefficientsStandard Coefficients t p value BStd. ErrorBetasMJS annual narrowing rate (/y), F = 9,945; p < 0,001 R2 = 0,290(Constant)29,19939,6510,7360,462Age−0,1460,169−0,041−0,8630,389Gender (Female)1,1963,2580,0180,3670,714BMI0,7191,0340,0330,6950,487Side (Right)−3,622,57−0,064−1,4090,16Surgical Indications1,1321,6450,0340,6880,492Patient groups (N, S, L groups)−16,0153,036−0,438−5,276< 0,001Number of spinal fusion levels−0,6010,579−0,089−1,0380,3Lumbosacral fusion−1,823,451−0,027−0,5270,598CEA−0,4170,23−0,113−1,8160,07SA−0,7010,388−0,114−1,8060,072 Postoperative SS 0,8380,3520,3212,3770,018Postoperative PT0,4230,3690,1591,1440,253Postoperative PI−0,6720,344−0,312−1,9510,052Postoperative PI-LL0,1660,1230,0881,3520,177 Table 2Regression model for sMJS where annual narrowing is taken as dependent variable and other variables are taken as independent variablesModelNon-Standard CoefficientsStandard Coefficients t p value BStd. ErrorBetasMJS annual narrowing rate (/y), F = 9,945; p < 0,001 R2 = 0,290(Constant)29,19939,6510,7360,462Age−0,1460,169−0,041−0,8630,389Gender (Female)1,1963,2580,0180,3670,714BMI0,7191,0340,0330,6950,487Side (Right)−3,622,57−0,064−1,4090,16Surgical Indications1,1321,6450,0340,6880,492Patient groups (N, S, L groups)−16,0153,036−0,438−5,276< 0,001Number of spinal fusion levels−0,6010,579−0,089−1,0380,3Lumbosacral fusion−1,823,451−0,027−0,5270,598CEA−0,4170,23−0,113−1,8160,07SA−0,7010,388−0,114−1,8060,072 Postoperative SS 0,8380,3520,3212,3770,018Postoperative PT0,4230,3690,1591,1440,253Postoperative PI−0,6720,344−0,312−1,9510,052Postoperative PI-LL0,1660,1230,0881,3520,177 Regression model for sMJS where annual narrowing is taken as dependent variable and other variables are taken as independent variables Fig. 5Relationship between the number of fused spinal levels and annual hip joint space narrowing rates. Scatter plots illustrate the association between fusion length and MJS, SJS, sMJS, and sSJS annual narrowing rates Relationship between the number of fused spinal levels and annual hip joint space narrowing rates. Scatter plots illustrate the association between fusion length and MJS, SJS, sMJS, and sSJS annual narrowing rates Table 3Regression model for sSJS where annual narrowing is taken as dependent variable and other variables are taken as independent variablesModelNon-Standard CoefficientsStandard Coefficients t p value BStd. ErrorBetasSJS annual narrowing rate (/y), F = 12,588; p < 0,001 R2 = 0,341(Constant)19,56332,6330,5990,549Age0,0830,1390,0270,5950,552Gender (Female)2,8152,6820,0491,050,295BMI1,3540,8510,0731,590,113Side (Right)−3,972,115−0,083−1,8770,061Surgical Indications0,4541,3540,0160,3350,738Patient groups (N, S, L groups)−9,3692,498−0,3−3,75< 0,001 Number of spinal fusion levels −1,5650,477−0,27−3,2820,001Lumbosacral fusion−2,0662,84−0,036−0,7280,467 CEA −0,5740,189−0,182−3,0350,003 SA −1,1830,319−0,225−3,702< 0,001Postoperative SS0,1730,290,0780,5970,551Postoperative PT−0,1580,304−0,069−0,5180,605Postoperative PI−0,0960,283−0,052−0,340,734 Postoperative PI-LL 0,1990,1010,1241,9750,049 Table 3Regression model for sSJS where annual narrowing is taken as dependent variable and other variables are taken as independent variablesModelNon-Standard CoefficientsStandard Coefficients t p value BStd. ErrorBetasSJS annual narrowing rate (/y), F = 12,588; p < 0,001 R2 = 0,341(Constant)19,56332,6330,5990,549Age0,0830,1390,0270,5950,552Gender (Female)2,8152,6820,0491,050,295BMI1,3540,8510,0731,590,113Side (Right)−3,972,115−0,083−1,8770,061Surgical Indications0,4541,3540,0160,3350,738Patient groups (N, S, L groups)−9,3692,498−0,3−3,75< 0,001 Number of spinal fusion levels −1,5650,477−0,27−3,2820,001Lumbosacral fusion−2,0662,84−0,036−0,7280,467 CEA −0,5740,189−0,182−3,0350,003 SA −1,1830,319−0,225−3,702< 0,001Postoperative SS0,1730,290,0780,5970,551Postoperative PT−0,1580,304−0,069−0,5180,605Postoperative PI−0,0960,283−0,052−0,340,734 Postoperative PI-LL 0,1990,1010,1241,9750,049 Regression model for sSJS where annual narrowing is taken as dependent variable and other variables are taken as independent variables Functional outcomes demonstrated significant postoperative improvement in ODI scores across all groups (p < 0.001; Table 4). HHS improved significantly in the no-fusion group, whereas changes in fusion groups did not reach statistical significance. WOMAC scores showed numerical improvement in all groups; however, these changes were not statistically significant within groups. Table 4Comparison of preoperative and postoperative functional outcome scores (ODI, HHS, and WOMAC) according to fusion groupsPreoperativePostoperative p value Group N (no fusion) (nf=32)ODI65,5 ± 22,239,5 ± 41,7< 0,001HHS69,3 ± 18,779,5 ± 25,10,006WOMAC61,3 ± 3,255,8 ± 10,20,082Group S (≤ 3 levels of fusion) (nf=44)ODI70,5 ± 20,344,5 ± 27,3< 0,001HHS69 ± 22,271 ± 21,70,565WOMAC58,5 ± 8,853,2 ± 9,40,065Group L (≥ 4 levels of fusion) (nf=38)ODI67,1 ± 23,945,4 ± 37< 0,001HHS64,1 ± 27,372,1 ± 30,80.101WOMAC64,8 ± 10,159,2 ± 10,50,074All patients (nf=114)ODI68 ± 2243,4 ± 34,9< 0,001HHS67,4 ± 23,173,7 ± 260,008WOMAC61,4 ± 9,755,9 ± 10,30,059Values are presented as mean ± standard deviation. (nf: number of patients who underwent functional evaluation) Table 4Comparison of preoperative and postoperative functional outcome scores (ODI, HHS, and WOMAC) according to fusion groupsPreoperativePostoperative p value Group N (no fusion) (nf=32)ODI65,5 ± 22,239,5 ± 41,7< 0,001HHS69,3 ± 18,779,5 ± 25,10,006WOMAC61,3 ± 3,255,8 ± 10,20,082Group S (≤ 3 levels of fusion) (nf=44)ODI70,5 ± 20,344,5 ± 27,3< 0,001HHS69 ± 22,271 ± 21,70,565WOMAC58,5 ± 8,853,2 ± 9,40,065Group L (≥ 4 levels of fusion) (nf=38)ODI67,1 ± 23,945,4 ± 37< 0,001HHS64,1 ± 27,372,1 ± 30,80.101WOMAC64,8 ± 10,159,2 ± 10,50,074All patients (nf=114)ODI68 ± 2243,4 ± 34,9< 0,001HHS67,4 ± 23,173,7 ± 260,008WOMAC61,4 ± 9,755,9 ± 10,30,059Values are presented as mean ± standard deviation. (nf: number of patients who underwent functional evaluation) Comparison of preoperative and postoperative functional outcome scores (ODI, HHS, and WOMAC) according to fusion groups Values are presented as mean ± standard deviation. (nf: number of patients who underwent functional evaluation) Logistic regression analysis evaluating postoperative MJS (≤ 2.0 mm vs. > 2.0 mm) demonstrated that the number of fusion levels and CEA were the strongest independent predictors (Table 5). Each additional fusion level increased the probability of postoperative MJS ≤ 2.0 mm by 16–19%, while each unit increase in CEA increased this probability by 5.6–7.3%. Table 5Logistic regression model in which the postoperative MJS value was divided into categorical changes and taken as the dependent variable and all other variables were taken as independent variables and testedModelUnivariate analysisMultivariate analysisORCI 95% p value ORCI 95% p value Postoperative MJS; (> 2.0 mm:0, ≤2.0 mm:1) χ² = 43.804; p < 0.001Nagelkerke R2 = 0,186Percentage Correct (> 2.0 mm) = %96,9Percentage Correct (≤ 2.0 mm) = %11,8Percentage Correct (overall) = %80,6Age0,984(0,952-1,018)0,355Gender (Female)1,147(0,599-2,194)0,679BMI1,047(0,854-1,282)0,659Side (Left)1(0,589-1,696)1 Number of spinal fusion levels 1,161(1,095 − 1,230)< 0,0011,191(1,118-1,268)< 0,001Lumbosacral fusion0,762(0,415-1,399)0,381 CEA 1,056(1,024 − 1,089)< 0,0011,073(1,037 − 1,109)< 0,001SA0,973(0,918-1,031)0,348Postoperative LL0,987(0,970-1,003)0,118Postoperative SS0,998(0,974-1,023)0,892Postoperative PT1,014(0,989-1,039)0,27Postoperative PI1,004(0,984-1,024)0,715Postoperative PI-LL1,017(1,000–1,035)0,051 Table 5Logistic regression model in which the postoperative MJS value was divided into categorical changes and taken as the dependent variable and all other variables were taken as independent variables and testedModelUnivariate analysisMultivariate analysisORCI 95% p value ORCI 95% p value Postoperative MJS; (> 2.0 mm:0, ≤2.0 mm:1) χ² = 43.804; p < 0.001Nagelkerke R2 = 0,186Percentage Correct (> 2.0 mm) = %96,9Percentage Correct (≤ 2.0 mm) = %11,8Percentage Correct (overall) = %80,6Age0,984(0,952-1,018)0,355Gender (Female)1,147(0,599-2,194)0,679BMI1,047(0,854-1,282)0,659Side (Left)1(0,589-1,696)1 Number of spinal fusion levels 1,161(1,095 − 1,230)< 0,0011,191(1,118-1,268)< 0,001Lumbosacral fusion0,762(0,415-1,399)0,381 CEA 1,056(1,024 − 1,089)< 0,0011,073(1,037 − 1,109)< 0,001SA0,973(0,918-1,031)0,348Postoperative LL0,987(0,970-1,003)0,118Postoperative SS0,998(0,974-1,023)0,892Postoperative PT1,014(0,989-1,039)0,27Postoperative PI1,004(0,984-1,024)0,715Postoperative PI-LL1,017(1,000–1,035)0,051 Logistic regression model in which the postoperative MJS value was divided into categorical changes and taken as the dependent variable and all other variables were taken as independent variables and tested Postoperative MJS; (> 2.0 mm:0, ≤2.0 mm:1) χ² = 43.804; p < 0.001 Nagelkerke R2 = 0,186 Percentage Correct (> 2.0 mm) = %96,9 Percentage Correct (≤ 2.0 mm) = %11,8 Percentage Correct (overall) = %80,6 Discussion The principal finding of this study was that radiographic progression of hip osteoarthritis differed significantly among the no-fusion, short-segment, and long-segment groups. Significant differences were observed in postoperative–preoperative KL grade change and in the annual narrowing rates of MJS, SJS, sMJS, and sSJS. Importantly, the magnitude of these differences increased in proportion to the number of fused levels. To our knowledge, this is the first study to evaluate SJS and standardized SJS annual narrowing rates (sSJS), in addition to MJS-based measures, across fusion-length strata while also assessing knee degeneration and functional outcomes. A study of a large patient cohort revealed that 2.77% of patients who underwent spinal fusion subsequently required total hip arthroplasty due to coxarthrosis. At the 10-year follow-up, this risk was reported to be as high as 14% [10]. Similarly, our study demonstrated that patients who underwent long-level spinal fusion had significantly increased preoperative and postoperative KL differences and an increased likelihood of a postoperative MJS measurement of 2.0 mm or less. These findings support an association between longer fusion constructs and accelerated hip joint space loss. The incidence of coxarthrosis was found to be higher in patients who underwent long-level spinal fusion[12, 18, 19] In the study by Kawai et al. [12], MJS was evaluated by measuring the narrowest point at the narrowest point of the joint before and after surgery. Similar findings were obtained in our study, and a greater rate of narrowing was detected in the MJS measurements in the L group compared to the S group. A study comparing the annual narrowing rate of MJS between patients who did not undergo spinal fusion and those who underwent short-term and long-term spinal fusion did not find any significant difference between the two groups [18]. However, our study revealed that the annual MJS narrowing rate was higher in the L group than in the N and S groups, and a statistically significant difference was also observed between the N and S groups. This discrepancy may be partly explained by differences in follow-up duration, which was longer in our cohort, particularly in the no-fusion group. In the literature, there are studies indicating that superior joint space measurements are more sensitive than MJS and indicate early coxarthrosis progression [15, 20]. Our study diverges from the existing literature by examining the annual narrowing rate of SJS. In addition, defined change in joint space width as the ratio of joint space width to femoral head size to calibrate radiograph magnification and standardize body size [19]. In our study, as the number of fusion levels increased, the rates of SJS and sSJS narrowing also increased. A hip joint space of 2.0 mm or less is linked to osteoarthritis progression [21–23]. It predicts the need for hip replacement surgery within three years with 12 times the risk[24]. Our study investigated the relationship between this risk and the number of levels fused, based on the knowledge that patients with a hip joint space of 2.0 mm or less are at increased risk of osteoarthritis. The risk of coxarthrosis increased by 16–19% with each additional level of fusion and by 5.6–7.3% with each increase in CEA measurement. The consistency of these findings across both continuous (annual narrowing) and categorical (≤ 2.0 mm) outcomes strengthens the clinical relevance of fusion length as a risk factor for hip degeneration. The relationship between pain in the knee joint and spinopelvic alignment has been examined in the literature and knee-spine syndrome has been defined in this context [25, 26]. In our study, no statistically significant difference was found in the preoperative and postoperative KL difference and the annual KMJS narrowing rates. Although the effect of spinopelvic alignment on knee osteoarthritis has been shown, fusion length may not produce differential knee radiographic progression within the follow-up duration of this cohort. The differences in spinopelvic parameters between groups were not entirely consistent with previously reported associations with hip osteoarthritis. Some associations (e.g., postoperative PT and SJS narrowing; postoperative SS and sMJS narrowing) were directionally consistent with prior work, but overall interpretation is limited because postoperative alignment depends on preoperative deformity magnitude and surgical correction strategy [27, 28]. Accordingly, causal attribution to a single spinopelvic variable is not possible in this design. Long lumbosacral fusion has been reported to affect hip osteoarthritis and a higher rate of joint narrowing in patients who underwent long-level fusion [12, 29]. In our study, lumbosacral fusion was associated with an increased change in hip KL grade; however, no significant relationship was identified between lumbosacral fusion and annual joint space narrowing rates. Fusion including the L5–S1 segment eliminates motion at the lumbosacral junction, which plays a critical role in sagittal spinopelvic alignment and pelvic compensation. When this segment remains mobile, adaptive changes in pelvic tilt and sacral slope may partially compensate for sagittal imbalance. In contrast, fixation of L5–S1 may limit these compensatory mechanisms and potentially alter load transmission to the hip joint. In the present study, however, interpretation of this relationship is limited because surgical indications in group S were predominantly related to L5–S1 pathologies, such as spondylolisthesis and degenerative disc disease. Therefore, separating the independent biomechanical effect of lumbosacral fusion from the underlying segmental pathology remains challenging. In all groups, ODI scores improved postoperatively compared with preoperative values. The improvement in ODI scores was consistent with expected postoperative outcomes. Improvements in HHS were significant only in the no-fusion group. This may reflect difficulty among patients in distinguishing whether lower extremity symptoms originate from spinal or hip pathology. The distributions of several independent variables that may potentially influence the development of hip osteoarthritis exhibited disparate distributions across patient cohorts. However, due to the relatively long follow-up period, we were able to evaluate postoperative and preoperative hip KL differences and joint narrowing with both MJS and SJS, suggesting that many potential confounding factors were excluded. This study has several limitations. First, its retrospective design precludes causal inference and introduces the potential for selection bias. Although multivariable analyses were performed, residual confounding cannot be excluded, particularly as fusion length was not randomly assigned and may reflect underlying deformity severity. Second, this was a single-center study, which may limit generalizability. While all patients had at least two years of follow-up, hip osteoarthritis is a slowly progressive condition, and longer follow-up may be required to fully capture clinically meaningful progression. Third, degeneration was assessed using radiographic joint space measurements and KL grading. Radiographic changes do not necessarily correspond to symptom severity or functional impairment. Finally, despite high interobserver reliability, radiographic measurements may be influenced by positioning and imaging variability. Therefore, the findings should be interpreted as demonstrating associations rather than definitive causal effects. Future research should include prospective, multicenter studies with longer follow-up durations to better clarify the long-term clinical implications of spinal fusion on hip and knee degeneration. Advanced imaging modalities and dynamic spinopelvic assessments may further improve understanding of biomechanical load transmission mechanisms. Additionally, integrating radiographic progression with patient-reported outcomes and the eventual need for hip arthroplasty could provide stronger clinical relevance. Conclusion Spinal fusion was associated with increased hip joint space narrowing, and this risk increased with fusion length. Longer fusion constructs demonstrated greater radiographic progression in the hip joint, whereas knee joint degeneration did not significantly differ according to fusion status in this cohort. These findings suggest that patients undergoing extensive lumbar fusion may warrant closer clinical and radiographic monitoring; however, prospective studies are needed to confirm this association. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1