# paper.pdf ## Metadata - PDFFormatVersion=1.7 - Language=en-US - IsLinearized=true - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - CreationDate=D:20260527111024+05'30' - Creator=Adobe InDesign 15.1 (Windows) - ModDate=D:20260527111027+05'30' - Producer=Adobe PDF Library 15.0 - Trapped=False - xmp:createdate=2026-05-27T11:10:24+05:30 - xmp:metadatadate=2026-05-27T11:10:27+05:30 - xmp:modifydate=2026-05-27T11:10:27+05:30 - xmp:creatortool=Adobe InDesign 15.1 (Windows) - xmpmm:instanceid=uuid:09c9db11-bc0b-4233-9d45-ce1d92519e4d - xmpmm:originaldocumentid=xmp.did:07801174072068118C149D4CFFA16D8F - xmpmm:documentid=xmp.id:2d9012c7-492b-5646-9286-c79dff1591be - xmpmm:renditionclass=proof:pdf - xmpmm:derivedfrom=xmp.iid:2d4bae7e-8490-d54f-a9f7-c1fcee0404f2xmp.did:738b4e3d-48a1-c848-9524-db8f0f341c67xmp.did:07801174072068118C149D4CFFA16D8Fdefault - xmpmm:history=convertedfrom application/x-indesign to application/pdfAdobe InDesign 15.1 (Windows)/2026-05-27T11:10:24+05:30 - dc:format=application/pdf - pdf:producer=Adobe PDF Library 15.0 - pdf:trapped=False ## Contents ### Page 1 Effect of aerobic exercise of different intensityon articular cartilage metabolism in patients withknee osteoarthritis A randomized controlled trial Liangliang Song, MMa,[](https://orcid.org/0009-0002-6397-0680) , Yuanyuan Meng, BMa Abstract Background: Osteoarthritis (OA) of the knee is a leading cause of disability worldwide, yet optimal exercise intensity for cartilageprotection remains unclear. This randomized controlled trial investigated dose–response effects of aerobic exercise intensity oncartilage metabolism, clinical symptoms, and structural outcomes in knee OA patients. Methods: One hundred twenty participants with knee OA (Kellgren–Lawrence grade 2–3) were randomly assigned to low-intensity (50–60% heart rate reserve [HRR]), moderate-intensity (60–70% HRR), high-intensity (70–80% HRR) treadmill walkingor stationary cycling, or a control group for 12 weeks (3 sessions/week). Heart rate was continuously monitored using Polar H10chest-strap monitors. Primary outcomes were serum C-terminal telopeptide of type II collagen and cartilage oligomeric matrixprotein levels. Secondary outcomes included Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) scores,Visual Analog Scale (VAS) pain, 6-minute walk test, and magnetic resonance imaging-assessed cartilage morphology. Data wereanalyzed using mixed-model repeated measures analysis of variance with IBM SPSS Statistics version 26.0 (IBM Corp.). Results: The high-intensity group demonstrated the greatest reductions in C-terminal telopeptide of type II collagen (−28.5%,P < .001) and cartilage oligomeric matrix protein (−22.3%, P < .001), the largest improvements in WOMAC scores (−45.3%,P < .001) and VAS pain (−52.5%, P < .001), and the greatest increases in cartilage thickness (+4.2%) and volume (+3.8%). Asignificant dose–response relationship was confirmed (group × time interaction: F (3108) = 4.82, P = .003). Among participantsin the high-intensity group, 73% achieved the minimal clinically important differences (MCID) for WOMAC total score and 80%achieved the MCID for VAS pain. The 6-minute walk test improvement in the high-intensity group (71 m) exceeded the establishedMCID of 26 m. No serious adverse events were reported; minor events (transient knee pain exacerbation and delayed-onsetmuscle soreness) resolved spontaneously within 48 to 72 hours. Conclusions: Appropriately prescribed high-intensity aerobic exercise safely and effectively improved cartilage metabolism,alleviated symptoms, and enhanced functional capacity in knee OA patients. These findings support revising current conservativeexercise recommendations upward when tolerable. Abbreviations: 6MWT = 6-minute walk test, ANOVA = analysis of variance, COMP = cartilage oligomeric matrix protein, CTX-II= C-terminal telopeptide of type II collagen, HRR = heart rate reserve, ITT = intention-to-treat, K-L = Kellgren–Lawrence, MCID =minimal clinically important difference, MRI = magnetic resonance imaging, OA = osteoarthritis, ROM = range of motion, VAS =Visual Analog Scale, WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index. Keywords: articular, biomarkers, cartilage, exercise therapy, heart rate, knee, magnetic resonance imaging, osteoarthritis, ran-domized controlled trial 1. Introduction Osteoarthritis (OA) is a degenerative joint disease that sig-nificantly impacts global health, with knee OA being partic-ularly prevalent and debilitating.[1] As the world’s populationages and obesity rates rise, the incidence of knee OA is pro-jected to increase substantially, posing a growing challengeto healthcare systems worldwide.[2,3] The pathogenesis of OAis complex, involving many factors contributing to progres-sive degrading of articular cartilage, remodeling of subchon-dral bone, and inflammation of the synovium. At the root ofboth the development and progression of OA is disruption incartilage metabolism, as biomechanical factors play a centralrole in this process.[4] In a normal joint, there is a fine bal-ance between the synthesis and breakdown of the componentsforming the cartilage extracellular matrix. Nonetheless, in the a Shijiazhuang People’s Hospital, Shijiazhuang, Hebei, China. * Correspondence: Yuanyuan Meng, Shijiazhuang People’s Hospital, Shijiazhuang050800, Hebei, China (e-mail:). Copyright © 2026 the Author(s). Published by Wolters Kluwer Health, Inc. This is an open access article distributed under the, which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited. How to cite this article: Song L, Meng Y. Effect of aerobic exercise of differentintensity on articular cartilage metabolism in patients with knee osteoarthritis: Arandomized controlled trial. Medicine 2026;105:22(e48913). Received: 7 November 2025 / Received in final form: 28 April 2026 / Accepted:1 May 2026 http://dx.doi.org/10.1097/MD.0000000000048913 ### Page 2 context of OA, this balance is disturbed, resulting in an over-all reduction of cartilage tissue.[5] Recent studies have iden-tified a number of biomarkers useful in demonstrating suchmetabolic changes, among which special attention has beengiven to C-terminal telopeptide of type II collagen (CTX-II)and cartilage oligomeric matrix protein (COMP) as bio-markers reflecting the degradation and turnover of cartilagerespectively.[6] While pharmacological interventions remain thecornerstone in the management of OA, there has been increas-ing emphasis on non-pharmacological approaches, includingexercise therapy. Aerobic exercise has shown promise in alle-viating OA symptoms and potentially modifying disease pro-gression.[7,8] However, the optimal intensity of aerobic exercisefor managing knee OA remains a subject of debate. Somestudies reported that moderate-intensity exercises may exertan optimal effect on pain and functional performance,[9] whileother studies have suggested that high-intensity interval train-ing may offer superior benefits with respect to cardiovascularhealth and possibly cartilage metabolism,[10] though rigorousevidence from completed trials remains limited. The global burden of knee OA is substantial, with ris-ing rates of total knee arthroplasty reflecting end-stage dis-ease prevalence[11,12] and projections indicating continuedincreases in arthroplasty demand across multiple countries.[13,14]Complications such as periprosthetic joint infection remain asignificant postoperative concern,[15,16] occupational ergonomicexposures contribute to disease burden,[17] and evolving surgicaltechniques reflect ongoing efforts to optimize arthroplasty out-comes.[18] Musculoskeletal conditions including OA increase therisk of comorbid chronic diseases,[3] and the economic impactextends well beyond direct healthcare costs.[2] These alarmingtrends underscore the urgent need for effective conservativeinterventions, particularly exercise-based approaches, that maydelay or prevent the need for surgical intervention.[19–21] The pathways by which aerobic exercise influences cartilagemetabolism in OA are not fully explained. It has been suggestedthat the mechanical loading that is imparted by exercise canimprove the function of chondrocytes and increase matrix pro-duction while simultaneously acting on inflammatory pathwaysin the joint.[22,23] Improvements in muscle strength and jointstability from exercise can give rise to improved biomechanics,which might reduce loading within the articular cartilage.[24] Given both the potential for exercise to affect cartilage metab-olism as well as the uncertainty about the intensity at which thismay best be achieved, investigation of the differential effectsof various aerobic exercise intensities on cartilage metabolismwas warranted in persons with knee OA. This study aimed tofill this knowledge deficit by investigating the effects of low-,moderate-, and high-intensity aerobic exercises on biomarkersreflecting cartilage metabolism, clinical symptoms, and func-tional outcomes in patients with knee OA. The results of thisstudy were expected to contribute to the design of more specificand efficient exercise programs in knee OA management andhave implications for disease modification. 2. Materials and methods 2.1. Study subjects For this study, a total of 120 participants with knee OA wereenrolled from orthopedic clinics and community health centers.Inclusion was based on adults aged 50 to 75 years who had aclinical and radiographic diagnosis of knee OA by the AmericanCollege of Rheumatology criteria. The inclusion criteriaincluded a minimum score of 3 out of 10 on the Visual AnalogScale (VAS) and a radiographic Kellgren–Lawrence grade of 2or 3 in the knees. Exclusion criteria included knee injury or sur-gery within the last 6 months, any inflammatory arthritis, con-comitant conditions representing a contraindication to exercise,cognitive impairment, and regular participation in organizedexercise programs within the last 3 months. Patients who wereon analgesic medication were allowed to continue their routinedosages throughout the period of study. The sample size calcu-lation was performed using G*Power 3.1 software (Heinrich-Heine-Universität Düsseldorf) based on a mixed-model repeatedmeasures analysis of variance (ANOVA) (4 groups × 3 timepoints), with a medium effect size F = 0.25, power 80%, 5%significance level, and a correlation among repeated measuresof 0.5, yielding a required total sample size of 100. Further,sampling was extended by 15% to account for expected drop-outs, resulting in a target of 120 participants (30 per group).Written informed consent was obtained from all participants,and ethics approval for the study protocol was obtained fromthe Institutional Review Board (Approval No. 2025-021). Thistrial was not prospectively registered in a clinical trial registry.Participants were randomly allocated to one of the following 4groups: low-intensity, moderate-intensity, high-intensity aerobicexercise, or a control group. The randomization was performedusing a computer-generated randomization sequence, with strat-ification by age and sex to ensure balanced groups. Allocationconcealment was ensured using sequentially numbered, opaque,sealed envelopes prepared by an independent statistician notinvolved in participant enrollment. 2.2. Experimental design This study was conducted as a randomized controlled trial todetermine the effects of different aerobic exercise intensities oncartilage metabolism in patients with knee OA. Participants were stratified and randomly assigned into 4groups: low-intensity, moderate-intensity, high-intensity aero-bic exercise, or a control group. The intervention lasted for 12weeks with 3 supervised exercise sessions per week. The low-intensity group performed aerobic exercise at 50 to 60% ofheart rate reserve (HRR), the moderate-intensity group at 60to 70% HRR, and the high-intensity group at 70 to 80% HRR.Exercise intensity was prescribed using the Karvonen formula[25]to calculate target HRR zones. Heart rate was continuouslymonitored during all exercise sessions using Polar H10 chest-strap heart rate monitors (Polar Electro Oy), with data recordedat 1-second intervals via the Polar Beat application. Each sessionincluded a 10-minute warm-up followed by 30 minutes of aero-bic exercise, either treadmill walking or stationary cycling (par-ticipants selected their preferred modality, which was maintainedthroughout the study), followed by a 10-minute cool-down.The distribution of exercise modality was similar across groups(treadmill/cycling: low-intensity 18/12, moderate-intensity17/13, high-intensity 16/14; chi-square P = .88). The controlgroup received usual care and education about OA manage-ment but without a structured exercise program. Biomarkers ofcartilage metabolism, clinical symptoms according to WesternOntario and McMaster Universities Osteoarthritis Index(WOMAC) and VAS, functional capacity assessed by the 6-minutewalk test (6MWT), andmagnetic resonance imaging (MRI)scans for cartilage thickness and volume were measured at base-line, 6 weeks, and 12 weeks. Safety monitoring was conductedthroughout the study. The overall experimental design frame-work is illustrated in. 2.3. Evaluation indicators The broad array of variables assessed for the different inten-sities of aerobic exercise in patients with knee OA involvedbiochemical, clinical, functional, and imaging parameters.These were summarized in. Changes in cartilage met-abolic biomarkers, as reflected by serum levels of CTX-II andCOMP, were the primary outcome measures; these biomarkerswere assayed using commercial enzyme-linked immunosorbentassay kits. The secondary outcomes encompassed the clinical ### Page 3 symptoms evaluated through the WOMAC as well as the VASfor pain. The assessment of functional capacity was conductedutilizing the 6MWT and measurements of knee range of motion(ROM). Furthermore, MRI was employed to examine carti-lage thickness and volume, thereby offering objective metricsof structural modifications. Assessments were performed atbaseline, 6 weeks, and 12 weeks to measure the change overtime. Merging these metrics provided a comprehensive assess-ment of the effect of aerobic exercise on cartilage metabolism,symptomatic improvement, and functional improvement inpatients with knee OA. This evaluative method ensured the cor-relation between changes in cartilage metabolism and patientself-reported changes by making use of biochemical and clinicalmarkers. 2.4. Experimental methods The methodologies implemented in this study comprehensivelygauged the effects of varied intensities of aerobic exercises onthe cartilage metabolism of persons with a diagnosis of kneeOA. The intervention, as outlined in, included an exer-cise program over a period of 12 weeks, with all participantsattending monitored sessions 3 times a week. Each session beganand ended with 10-minute walks followed by 30-minute periodsof aerobic exercise on a treadmill or stationary bicycle. Exerciseintensity was monitored with HRR and perceived exertion.Blood samples were drawn in the morning after a 12-hour fastfor biomarker analysis. Serum concentrations of CTX-II andCOMP were assayed using commercial enzyme-linked immuno-sorbent assay kits (Cusabio Biotech). Clinical assessments wereperformed by blinded, trained physiotherapists; the 6MWT wasperformed according to standard protocols. Bilateral knee MRIwas performed using a Siemens Magnetom Prisma 3.0T scan-ner (Siemens Healthineers) with dedicated knee coils, employingsagittal 3D dual-echo steady-state sequences with 0.7 mm slicethickness for cartilage imaging. For participants with bilateralknee OA, the more symptomatic knee (defined as the knee withhigher baseline VAS pain score) was designated as the indexknee for MRI analysis; in cases of equal symptoms, the rightknee was selected. Image analysis was performed by an experi-enced musculoskeletal radiologist blinded to participant groupallocation and time point. Throughout the study, participantsmaintained daily logs of their exercise and reported all adverseevents. Adherence was monitored through session attendanceand heart rate data. 2.5. Data analysis Data analysis was performed using IBM SPSS Statistics version26.0 (IBM Corp.). Descriptive statistics were used to summarizeparticipants’ characteristics and outcome measures. Normalityof the data distribution was tested using the Shapiro–Wilktest. Comparison between groups at baseline was performedusing 1-way ANOVA or the Kruskal–Wallis test as appropri-ate. Exercise intensity effects on outcome measures across timewere analyzed using a mixed-model repeated measures ANOVA,with time as the within-subject variable and group as thebetween-subject variable. Significant interactions were furtherexplored using post hoc tests with the Bonferroni method formultiple comparisons. Effect sizes were calculated as partial eta-squared (eta-p-squared). The associations between changes inbiomarkers and clinical outcomes were explored using Pearsonor Spearman correlations. Treatment response predictors wereanalyzed using multiple regression. The primary analysis was Figure 1. Experimental design framework. Overview of the study protocol including group allocation, intervention schedule, and outcome assessment timepoints. Table 1 Assessment indicators for OA exercise study. COMP = cartilage oligomeric matrix protein, CTX-II = C-terminal telopeptide of type II collagen, ELISA = Enzyme-Linked Immunosorbent Assay, OA = osteoarthritis, WOMAC = Western Ontario andMcMaster Universities Osteoarthritis Index. ### Page 4 conducted on a per-protocol basis including only participantswho completed the 12-week intervention; a sensitivity analysisusing intention-to-treat (ITT) principles with last observationcarried forward was also performed to assess the robustnessof the findings. The level of statistical significance was set atP < .05. 3. Results 3.1. Baseline data comparison At the start of this study, 120 patients with knee OA were ran-domly divided into 4 groups: low-intensity (n = 30), moderate-intensity (n = 30), high-intensity (n = 30) aerobic exercise, anda control group (n = 30). Demographic data of the participantsare presented in. No statistically significant differencewas observed among the groups regarding age, sex, body massindex, disease duration, and Kellgren–Lawrence grade (allP > .05). Comparisons of baseline values of the primary andsecondary outcome measures, including serum CTX-II andCOMP levels, WOMAC scores, VAS pain scores, 6MWT dis-tances, and knee ROM showed no between-group differences(all P > .05). During the 12-week intervention, 8 participants withdrew:2 from the low-intensity group (personal reasons), 1 from themoderate-intensity group (unrelated illness), 2 from the high-intensity group (scheduling conflicts), and 3 from the controlgroup (lost to follow-up). A total of 112 participants completedthe study and were included in the per-protocol analysis ().A sensitivity analysis using ITT principles with last observationcarried forward, including all 120 randomized participants,yielded results consistent with the per-protocol analysis for allprimary and secondary outcomes, confirming the robustness ofthe findings. The distribution of CTX-II values at baseline among the 4groups is shown in; the biomarker profile was similarat the start of the study. 3.2. Effect of different intensity exercise on cartilagemetabolism markers The metabolic effects of different exercise intensities on car-tilage markers were assessed over 12 weeks. Changes inserum CTX-II and COMP levels for all groups at baseline, 6weeks, and 12 weeks are presented in. A significanttime-group interaction was observed for both biomarkers(P < .001). At 12 weeks compared with baseline, the high-intensitygroup showed the largest decrease in CTX-II lev-els (−28.5%, P < .001), followed by the moderate-intensitygroup (−18.7%, P < .01), whereas the low-intensity groupshowed a slight decrease (−8.2%, P < .05), and the controlgroup did not reach a statistically significant change. Similartrends were observed for COMP levels, with the high-intensitygroup demonstrating the greatest decrease (−22.3%, P < .001).To contextualize these biomarker changes, the coefficient ofvariation for CTX-II in our study population was approx-imately 35%, and the observed reduction of 28.5% in the Table 2 Experimental methods for OA exercise study. 6MWT = 6-minute walk test, COMP = cartilage oligomeric matrix protein, CTX-II = C-terminal telopeptide of type II collagen, ELISA = Enzyme-Linked Immunosorbent Assay, HR = heart rate, HRR = heartrate reserve, MRI = Magnetic Resonance Imaging, OA = osteoarthritis, VAS = Visual Analog Scale, WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index. Table 3 Baseline characteristics of study participants. Data are presented as mean ± SD or counts. 6MWT = 6-minute walk test, BMI = body mass index, COMP = cartilage oligomeric matrix protein, CTX-II = C-terminal telopeptide of type II collagen, K-L = Kellgren–Lawrence, n = number of participants,ROM = range of motion, SD = standard deviation, VAS = Visual Analog Scale, WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index. ### Page 5 high-intensitygroup exceeded twice the minimum detectablechange (MDC = 12.6%), confirming that the change wasbeyond measurement variability. Furthermore, the magnitudeof CTX-II reduction in the high-intensity group was substan-tially greater than the 15 to 20% reductions typically reportedin pharmacological interventions and weight loss programs,[26]supporting the clinical relevance of exercise-induced bio-marker changes. Temporal changes in CTX-II levels across all groups are pre-sented in. The moderate- and high-intensity groupsshowed a faster decline in CTX-II levels, with the most consid-erable reduction occurring between baseline and 6 weeks. Thesefindings suggest a clear dose–response relationship betweenexercise intensity and improvement in cartilage metabolicmarkers.[6,27,28] 3.3. Clinical symptoms and functional improvement The changes in symptoms and functional ability at various exer-cise intensities throughout the intervention period are presentedin. Significant time-group interactions were observed forall measures (P < .001). Changes in WOMAC total scores andVAS pain scores at 12 weeks from baseline were most promi-nent in the high-intensity group (−45.3%, P < .001 and −52.5%,P < .001, respectively). Changes in the moderate-intensity groupwere also considerable (−35.7% and −41.2% for WOMAC and Figure 2. CONSORT flow diagram of participant enrollment, allocation, follow-up, and analysis. CONSORT = Consolidated Standards of Reporting Trials. Figure 3. Distribution of baseline serum CTX-II levels across the 4 study groups. Each data point represents an individual participant; horizontal lines indicategroup means ± SD. CTX-II = C-terminal telopeptide of type II collagen, SD = standard deviation. ### Page 6 VAS, respectively, P < .001), while changes in the low-intensitygroup were more modest yet still significant (−20.1% and −25.3%,respectively, P < .01). The most prominent increase in functionalcapacity as measured by the 6MWT was seen in the high-intensity group (+18.7%, P < .001), followed by the moderate-intensity group (+14.2%, P < .001) and the low-intensity group(+8.5%, P < .01). The time course of changes in 6MWT distance for all groupsis presented in, showing a dose–response relationshipbetween exercise intensity and functional improvement. When interpreting clinical significance in relation to estab-lished minimal clinically important differences (MCID), theresults were particularly encouraging. For the WOMAC totalscore, using an MCID threshold of 9.1 points,[29,30] the pro-portion of participants achieving MCID was 47% in the low-intensity group, 63% in the moderate-intensity group, and 73%in the high-intensity group, compared with only 12% in thecontrol group. For VAS pain, using an MCID of 1.5 points,the corresponding proportions were 53%, 70%, and 80% inthe exercise groups versus 10% in controls. The mean improve-ment in 6MWT distance in the high-intensity group (71 m)substantially exceeded the established MCID of 26 m for thismeasure,[31] and even the low-intensity group (33 m) surpassedthis threshold. These MCID-based analyses confirm thatobserved improvements are not only statistically significant butalso clinically meaningful, with the proportion of respondersincreasing in a dose-dependent manner.[32] 3.4. Imaging changes MRI was used to assess changes in knee cartilage structure overthe 12-week intervention period. Changes in cartilage thicknessand volume across all groups at baseline and 12 weeks are pre-sented in. A significant time-by-group interaction wasobserved for all measures (P < .001). The high-intensity exercisegroup demonstrated the greatest changes in cartilage thicknessand volume (+4.2%, P < .001, and +3.8%, P < .001, respectively)compared with baseline. Marked increases in both thicknessand volume of cartilage were also found for the moderate-intensity exercise group (+2.9% and +2.6%, respectively,P < .01), while increases were slight but statistically significantin the low-intensity group (+1.5%, +1.3%, P < .05). No signifi-cant change in cartilage parameters was evident among controls. Differences in cartilage volume changes between allgroups are illustrated in, demonstrating the dose–response relationship between exercise intensity and cartilage Table 4 Changes in cartilage metabolic markers over 12 weeks. Data are presented as mean ± SD. COMP = cartilage oligomeric matrix protein, CTX-II = C-terminal telopeptide of type II collagen, SD = standard deviation. P values represent the significance of change from baseline to 12 weeks within each group. Figure 4. Changes in serum CTX-II levels over 12 weeks. Data are presented as mean ± SD. *P < .05, **P < .01, ***P < .001 vs baseline within group. CTX-II =C-terminal telopeptide of type II collagen, SD = standard deviation. ### Page 7 preservation.[22–24] Importantly, the high-intensity group showeda greater relative increase in medial compared with lateral com-partment cartilage volume, suggesting a site-specific response ofcartilage to mechanical loading. 3.5. Dose–response relationship analysis An extensive analysis using mixed-effects models was conductedto demonstrate the dose–response relationship between exerciseintensity and outcome measures. Regression coefficients forthe association between exercise intensity (percentage of HRR,HRR%) and main outcome variables are given in. Apositive association was found between exercise intensity andimprovement in all outcome measures (P < .001). The stron-gest association was for changes in CTX-II (beta = −0.42, 95%confidence interval: −0.51 to −0.33) and for improvements inWOMAC score (beta = −0.39, 95% confidence interval: −0.48to −0.30), indicating that higher exercise intensity was associ-ated with significantly greater improvement in cartilage metab-olism and clinical symptoms. The dose–response relationships for CTX-II, WOMAC scores,and cartilage volume changes are illustrated in. Theseindicate a nonlinear relationship with evidence of a thresholdeffect, as the slope was considerably steeper between 60%and 70% HRR. The dose–response curve for cartilage volumeshowed a plateau effect above 75% HRR, suggesting an upperlimit to the structural benefits of high-intensity exercise.[26–33] Table 5 Changes in clinical symptoms and functional measures over 12 weeks. Data are presented as mean ± SD. 6MWT = 6-minute walk test, ROM = range of motion, SD = standard deviation, VAS = Visual Analog Scale, WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index. P values represent the significance of change from baseline to 12 weeks within each group. Figure 5. Changes in 6-minute walk test distance over 12 weeks. The dashed line indicates the MCID threshold of 26 m. Data are presented as mean ± SD.**P < .01, ***P < .001 vs. baseline within group. MCID = minimal clinically important difference, SD = standard deviation. ### Page 8 Table 6 Changes in MRI-based cartilage measures over 12 weeks. Data are presented as mean ± SD. MRI = magnetic resonance imaging, SD = standard deviation. P values represent the significance of change from baseline to 12 weeks within each group. Figure 6. Changes in cartilage volume over 12 weeks by compartment. Data are presented as mean ± SD. *P < .05, **P < .01, ***P < .001 vs. baseline;ns = not significant. CTX-II = C-terminal telopeptide of type II collagen, HRR = heart rate reserve, SD = standard deviation, WOMAC = Western Ontario andMcMaster Universities Osteoarthritis Index. Table 7 Dose–response relationship between exercise intensity and outcome measures. Regression coefficients (β) represent the change in outcome measure per 10% increase in exercise intensity (%HRR). Effect sizes are interpreted as: small (0.2), medium (0.5), and large (0.8). 6MWT = 6-minute walk test, CI = Confidence Interval, COMP = cartilage oligomeric matrix protein, CTX-II = C-terminal telopeptide of type II collagen, HRR = heart rate reserve, ROM = range of motion,VAS = Visual Analog Scale, WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index. ### Page 9 3.6. Safety and adverse events No serious adverse events were reported during the study.Minor adverse events included transient knee pain exacerba-tion (low-intensity: 2 participants; moderate-intensity: 3 partic-ipants; high-intensity: 5 participants) and delayed-onset musclesoreness (moderate-intensity: 2 participants; high-intensity: 4participants). All minor adverse events resolved spontaneouslywithin 48 to 72 hours and did not require modification of theexercise protocol. No participant withdrew from the study dueto exercise-related adverse events. The overall exercise adher-ence rates were 92% in the low-intensity group, 89% in themoderate-intensity group, and 85% in the high-intensity group.Heart rate monitoring confirmed that participants maintainedtheir prescribed intensity zones in over 90% of recorded ses-sions across all exercise groups. 4. Discussion The goal of this study was to investigate the effects of differentintensities of aerobic exercise on cartilage metabolism, clinicalsymptoms, and functional outcomes among patients with kneeOA. The results demonstrated a clear dose–response relationshipbetween exercise intensity and improvements in multiple out-come measures, with high-intensity exercise offering the greatestbenefits. These findings add to the growing body of evidencesupporting exercise as an effective non-pharmacological inter-vention for managing knee OA.[21,34,35] The reduction in CTX-IIand COMP, especially in the high-intensity group, indicatesthat prescribed exercise could have positive effects on cartilagemetabolism. This result is consistent with earlier research show-ing mechanical loading as a stimulatory factor for chondrocyteactivity and matrix synthesis.[22,23] The more pronounced effectsof high-intensity physical activity can be attributed to highermagnitudes of mechanical stimuli and superior cardiovascularresponses, possibly contributing to improved perfusion anddelivery of nutrients to joint tissues.[4,5] The biomarker responsepatterns observed in our study are consistent with the temporaldynamics of cartilage markers reported in response to acute andchronic exercise,[27,28] and the magnitude of CTX-II reductionin the high-intensity group exceeds that reported in weight lossinterventions.[26] Improvement in WOMAC scores and the VAS pain subscaleacross all exercise groups, especially in the high-intensity group,extends findings from previous systematic reviews[7–9] demon-strating that higher-intensity exercise programs lead to greaterimprovements in pain and functional impairment. The MCIDanalysis further strengthens these findings by demonstratingthat a significantly higher proportion of participants in thehigh-intensity group achieved clinically meaningful improve-ments compared with lower-intensity groups.[29,30] These resultssuggest that patients with knee OA may benefit from more vig-orous exercise than traditionally prescribed, provided that it isappropriately supervised and progressively implemented. Theobserved improvements in clinical outcomes were paralleled bychanges in biochemical markers,[6,36,37] supporting a mechanisticlink between exercise-induced cartilage metabolic changes andsymptomatic improvement. The gains in 6MWT distance and knee ROM reflect functionalbenefits from aerobic exercise in knee OA. Improvement was mostevident in the high-intensity group, consistent with principles ofexercise training specificity.[25] The 6MWT improvement in thehigh-intensity group (71 m) substantially exceeded the establishedMCID of 26 m,[31] confirming clinical relevance. Such improvedfunctional capacity can be attributed to several interacting factorsincluding pain reduction, neuromuscular control improvements,and increased cardiovascular fitness.[38,39] The dose-dependent pat-tern of functional improvement parallels findings from networkmeta-analyses of exercise interventions for knee OA.[8] The apparent structural changes documented throughMRI are noteworthy but should be interpreted with cau-tion: while all exercise groups showed increases in cartilagethickness and volume measurements, the high-intensity groupdemonstrated the greatest changes. If confirmed over longerfollow-up periods, this finding may challenge the prevailingview that high-impact exercise adversely affects joint healthin OA patients.[22,24] Nevertheless, these results are consis-tent with the adaptability theory of cartilage to mechanicalloading, although longer-term studies are needed to confirmtrue structural remodeling. The plateau effect observed incartilage volume measurements at higher intensities may sug-gest an optimum spectrum of exercise intensity for cartilageresponses, though whether these represent true structuralbenefits requires confirmation through longer-term trials. Theload-induced dynamics of cartilage biomarkers[33] and postsur-gical biomarker trajectories[37,40] provide mechanistic contextfor understanding the relationship between exercise loadingand structural cartilage outcomes. The nonlinear dose–response relationships between exer-cise intensity and outcome measures, with a threshold effect atapproximately 60 to 70% of HRR, are consistent with currentexercise prescription guidelines.[21,25] These findings indicate thatpatients with knee OA can safely undertake, and benefit from,more intense exercise than was previously thought, challengingtraditional conservative exercise recommendations commonlyprescribed for this population.[19,34,35] Figure 7. Dose–response curves for (A) CTX-II, (B) WOMAC total score, and (C) medial compartment cartilage volume. The shaded area indicates the thresholdzone (60–70% HRR). *P < 0.05, ** P < 0.01, *** P < 0.001 vs. baseline. ### Page 10 From a clinical implementation perspective, these findingssuggest that exercise prescription for knee OA should fol-low an individualized, progressive approach similar to car-diac rehabilitation models. Clinicians could initiate patientsat low-to-moderate-intensity (50–60% HRR) and progres-sively increase to moderate-to-high-intensity (60–75% HRR)over 4 to 6 weeks, contingent on symptom response andtolerance.[25,34,35] The use of wearable heart rate monitors, asemployed in this study, enables objective intensity monitoringin community settings, making the translation of these findingsto real-world practice feasible. Group-based exercise programssupervised by trained physiotherapists, with heart rate-guidedintensity prescription, represent a scalable model for deliveringthese interventions in primary care and community rehabilita-tion settings.[38,39] Despite these positive results, several limitations should beconsidered. First, the intervention period of 12 weeks, thoughsufficient to demonstrate considerable changes, does not delin-eate the long-term effects of different exercise intensities on dis-ease progression. Second, while steps to minimize confounderswere considered, complete elimination of individual responsesto exercise and possible influences of diet and lifestyle cannot betotally ruled out. Third, the primary analysis was conducted ona per-protocol basis; although a sensitivity analysis using ITTprinciples with last observation carried forward yielded con-sistent results, the potential for attrition bias cannot be fullyexcluded. Fourth, due to the nature of exercise interventions,blinding of participants and exercise therapists was not possi-ble, and only outcome assessors were blinded. Fifth, participantswere allowed to choose between treadmill walking and station-ary cycling, which differ in joint loading patterns and may haveinfluenced cartilage-related outcomes; future studies shouldcontrol for exercise modality. Sixth, the structural cartilagechanges observed on MRI over 12 weeks should be interpretedcautiously, as true structural remodeling of cartilage typicallyrequires longer timeframes, and short-term changes may partlyreflect alterations in cartilage hydration rather than genuinetissue remodeling. Finally, the generalizability of our findingsmay be limited, as participants were recruited from a single geo-graphic region. Additionally, this trial was not registered in aclinical trial registry, which is acknowledged as a limitation. This study provides strong evidence for the beneficial effectsof moderate- to high-intensity aerobic exercise in knee OAmanagement, as assessed by changes in cartilage metabolism,clinical symptoms, and functional outcomes. These findingshave important implications for clinical practice, indicatingthat exercise prescriptions for patients with knee OA should berevised upward in terms of intensity when tolerable.[21] Futuretrials with longer intervention periods, individualized exerciseprotocols, and patient-tailored outcome measures are needed toconfirm and extend these findings.[10] 5. Conclusion This current study has provided sound evidence with respect tothe differential effects of different intensities of aerobic exerciseon cartilage metabolism, clinical manifestations, and functionaloutcomes in patients with knee OA. Quite a clear dose–responserelationship could be observed, where the highest positive dif-ferences concerned high-intensity exercise for all the variablesconsidered. Improvements in these metabolic markers, togetherwith increases in thickness and volume of the cartilage, chal-lenge traditional concepts related to the effect of high-levelexertion exercises on joint health for OA patients. The strongpain reduction and improvement in functional capacity under-line the potential of high-intensity exercise, when appropriatelyprescribed, as a potent non-pharmacological intervention inknee OA management. The overall findings provide an import-ant determinant threshold effect and possible upper boundaryfor structural benefit, which are relevant to the development ofimproved exercise recommendations. These findings have sig-nificant clinical implications that suggest current exercise inter-ventions for individuals with knee OA are overly conservativeand that many patients could safely tolerate and benefit fromexercises at a higher intensity than previously recommended.Further studies should be addressed to sustained interventions,individualized programs of exercises, and mechanistic explana-tions of the benefits observed. Overall, the present study rep-resents a paradigm shift in the prescription of exercise for kneeOA, in which intensity is signaled as a basic characteristic foroptimizing the therapeutic effect. Author contributions Conceptualization: Liangliang Song. Data curation: Yuanyuan Meng. Formal analysis: Liangliang Song, Yuanyuan Meng. Investigation: Liangliang Song, Yuanyuan Meng. Methodology: Liangliang Song. Project administration: Liangliang Song. Supervision: Liangliang Song. Validation: Yuanyuan Meng. Visualization: Yuanyuan Meng. Writing – original draft: Liangliang Song. Writing – review & editing: Liangliang Song, Yuanyuan Meng. References [1] Hunter DJ, Bierma-Zeinstra SM. Osteoarthritis. Lancet.2019;393:1745–59. [2] Shamekh A, Alizadeh M, Nejadghaderi SA, et al. The burden of osteo-arthritis in the Middle East and North Africa region from 1990 to2019. Front Med (Lausanne). 2022;9:881391. [3] Williams A, Kamper SJ, Wiggers JH, et al. Musculoskeletal conditionsmay increase the risk of chronic disease. BMC Med. 2018;16:167. [4] Guilak F. Biomechanical factors in osteoarthritis. Best Pract Res ClinRheumatol. 2011;25:815–23. [5] Mobasheri A, Matta C, Uzieliene I, Budd E, Martín-Vasallo P,Bernotiene E. The chondrocyte channelome: a narrative review. JointBone Spine. 2018;86:29–35. [6] Hao HQ, Zhang JF, He QQ, Wang Z. COMP, CTX-II, and MMP-3as biomarkers for knee and hip osteoarthritis diagnosis: a systematicreview and meta-analysis. Osteoarthritis Cartilage. 2019;27:726–36. [7] Regnaux JP, Lefevre-Colau MM, Trinquart L, et al. High-intensity ver-sus low-intensity physical activity or exercise in people with hip or kneeosteoarthritis. Cochrane Database Syst Rev. 2015;2015:CD010203. [8] Mo L, Jiang B, Mei T, Zhou D. Exercise therapy for knee osteoarthritis:a systematic review and network meta-analysis. Orthop J Sports Med.2023;11:23259671231172773. [9] Han S, Li T, Cao Y, et al. Quantitative analysis of effectiveness of exer-cise on symptoms in osteoarthritis: a model-based meta-analysis. Br JSports Med. 2024;58:1539–50. [10] Oiestad BE, Osteras N, Frobell R, Grotle M, Brøgger H, Risberg MA.Efficacy of strength and aerobic exercise on patient-reported outcomesin knee osteoarthritis: study protocol for a randomized controlled trial.BMC Musculoskelet Disord. 2013;14:266. [11] Moldovan F, Gligor A, Moldovan L, Bataga T. An investigation forfuture practice of elective hip and knee arthroplasties during COVID-19 in Romania. Medicina (Kaunas). 2023;59:314. [12] Moldovan F, Moldovan L, Bataga T. A comprehensive research on theprevalence and evolution trend of orthopedic surgeries in Romania.Healthcare (Basel). 2023;11:1866. [13] Kim TW, Kang SB, Chang CB, et al. Current trends and projectedburden of primary and revision total knee arthroplasty in Korea. JArthroplasty. 2020;36:93–101. [14] Ackerman IN, Bohensky MA, Zomer E, et al. The projected burden ofprimary total knee and hip replacement for osteoarthritis in Australiato 2030. BMC Musculoskelet Disord. 2019;20:90. [15] Gould D, Dowsey M, Jo I, Choong P. Patient-related risk factors forunplanned 30-day readmission following total knee arthroplasty. ANZJ Surg. 2020;90:1253–8. [16] Chang CH, Lee SH, Lin YC, Wang Y-C, Chang C-J, Hsieh P-H.Increased periprosthetic hip and knee infection projected from 2014 to2035 in Taiwan. J Infect Public Health. 2020;13:1768–73. ### Page 11 [17] Hulshof CTJ, Pega F, Neupane S, et al. The effect of occupational expo-sure to ergonomic risk factors on osteoarthritis of hip or knee. EnvironInt. 2021;150:106349. [18] Vinet M, Le Stum M, Gicquel T, et al. Unicompartmental knee arthro-plasty: a French multicenteric retrospective study from 2009 to 2019.Orthop Traumatol Surg Res. 2023;109:103581. [19] Li CS, Ayeni OR, Sprague S, Truong V, Bhandari M. Conservative treat-ments, surgical treatments, and the KineSpring Knee Implant system forknee osteoarthritis. J Long Term Eff Med Implants. 2013;23:105–49. [20] Cameron KL, Driban JB, Svoboda SJ. Osteoarthritis and the tacticalathlete: a systematic review. J Athl Train. 2016;51:952–61. [21] Bannuru RR, Osani MC, Vaysbrot EE, et al. OARSI guidelines for thenon-surgical management of knee, hip, and polyarticular osteoarthritis.Osteoarthritis Cartilage. 2019;27:1578–89. [22] Xu D, Van Middelkoop M, Bierma-Zeinstra SMA, Runhaar J. Physicalactivity and features of knee osteoarthritis on MRI: a systematic review.Arthritis Care Res (Hoboken). 2023;75:1908–13. [23] Durmus D, Alayli G, Bayrak IK, Canturk F. Assessment of the effectof glucosamine sulfate and exercise on knee cartilage using MRI: arandomized controlled clinical trial. J Back Musculoskelet Rehabil.2012;25:275–84. [24] Eckstein F, Maschek S, Culvenor A, et al. Which risk factors determinecartilage thickness and composition change in radiographically normalknees? Osteoarthr Cartil Open. 2023;5:100365. [25] Garber CE, Blissmer B, Deschenes MR, et al. American College ofSports Medicine position stand. Quantity and quality of exercise. MedSci Sports Exerc. 2011;43:1334–59. [26] Bartels EM, Christensen R, Christensen P, et al. Effect of a 16 weeksweight loss program on osteoarthritis biomarkers in obese patients.Osteoarthritis Cartilage. 2014;22:1817–25. [27] Roberts HM, Law RJ, Thom JM. The time course and mechanisms ofchange in biomarkers of joint metabolism in response to exercise. Eur JAppl Physiol. 2019;119:2401–20. [28] Dreiner M, Munk T, Zaucke F, Liphardt A-M, Niehoff A. Relationshipbetween different serum cartilage biomarkers in the acute response torunning and jumping. Sci Rep. 2022;12:6434. [29] Silva MDC, Woodward AP, Fearon AM, et al. Minimal clinicallyimportant change of knee flexion in people with knee osteoarthritis.Syst Rev. 2024;13:50. [30] Migliorini F, Maffulli N, Schafer L, Simeone F, Bell A, Hofmann UK.MCID, SCB, and PASS in patients who have undergone total kneearthroplasty: a systematic review. Knee Surg Relat Res. 2024;36:3. [31] King LK, Hawker GA, Stanaitis I, Woodhouse L, Jones CA, Waugh EJ.MCID for improvement in six-minute walk test in persons with kneeosteoarthritis. BMC Musculoskelet Disord. 2022;23:307. [32] Fatima M, Scholes CJ, Tutty A, Ebrahimi M, Genon M, Martin SJ.Patient-reported outcomes of a short hospital stay after total kneereplacement. ANZ J Surg. 2022;92:837–42. [33] Mundermann A, Nuesch C, Herger S, et al. Load-induced blood markerkinetics in patients with medial knee compartment osteoarthritis.F1000Res. 2024;12:299. [34] Sabha M, Hochberg MC. Non-surgical management of hip and kneeosteoarthritis; comparison of ACR/AF and OARSI 2019 and VA/DoD2020 guidelines. Osteoarthr Cartil Open. 2021;4:100232. [35] Gourdine J. Review of nonsurgical treatment guidelines for lowerextremity osteoarthritis. Orthop Nurs. 2019;38:303–8. [36] Lisee C, Obudzinski S, Pietrosimone BG, et al. Association of serumbiomarker profiles with posttraumatic OA symptoms at 12 monthsafter ACLR. Am J Sports Med. 2024;52:2503–11. [37] Harkey MS, Luc BA, Golightly YM, et al. Osteoarthritis-related bio-markers following ACL injury and reconstruction: a systematic review.Osteoarthritis Cartilage. 2015;23:1–12. [38] Luan L, El-Ansary D, Adams R, Wu S, Han J. Knee osteoarthritispain and stretching exercises: a systematic review and meta-analysis.Physiotherapy. 2022;114:16–29. [39] Ahmad MA, Hamid MSA, Yusof A. Effects of low-level and high-intensity laser therapy as adjunctive to rehabilitation exercise on pain,stiffness and function in knee osteoarthritis: a systematic review andmeta-analysis. Physiotherapy. 2022;114:85–95. [40] Abshirini M, Coad J, Wolber FM, et al. Effects of Greenshell musselintervention on biomarkers of cartilage metabolism: a RCT. Front Med(Lausanne). 2022;9:1063336.