Abstract
Background
Physical activity may mitigate osteoporosis progression by modulating telomere shortening processes.
Aims
To explore the mediating role of telomere length (TL) in the relationship between physical activity and bone mineral density (BMD).
Methods
This study enrolled 2,394 participants aged 50 years and older from the U.S. National Health and Nutrition examination Surveys. TL was measured by quantitative polymerase chain reaction (qPCR) method (TeloMean) and DNA methylation data (HorvathTelo), and accelerated telomere attrition was assessed through residual-based indices of TeloMeanAccel and HorvathTeloAccel. Physical activity was assessed via questionnaire and BMD was measured at multiple body sites. Multiple linear regression models were utilized to evaluate associations between TL metrics, physical activity, and BMD. Mediation analysis, restrict cubic spline (RCS) modeling, subgroup analyses and sensitivity analyses were further conducted.
Results
After adjusting for covariates, TL metrics of TeloMean and HorvathTelo were found significantly positive correlations with BMD. HorvathTeloAccel, reflecting accelerated telomere shortening, also exhibited significant association with BMD. Physical activity demonstrated a significant positive association with total BMD (β = 0.046, 95%CI: 0.004–0.088). Mediation analysis revealed that TeloMean and HorvathTelo accounted for 4.78% and 20.86% of the total effect of physical activity on BMD, respectively, while HorvathTeloAccel explained 5.24% of the observed association.
Conclusion
Reduced physical activity and accelerated telomere attrition were related with BMD decline, and TL partially mediated the association. These findings suggest that enhancing physical activity could mitigate telomere shortening and promote bone health.
Supplementary Information
The online version contains supplementary material available at 10.1007/s40520-025-03176-4.
Keywords: Bone mineral density, Physical activity, Telomere length, Osteoporosis, Biological aging
Introduction
Osteoporosis, a common type of chronic metabolic bone disease manifesting as decreased bone mineral density (BMD) and an elevated risk of fractures, has raised increasing public attention due to the great disease burdens it might cause for the old population. Over the past 30 years, the osteoporosis incidence tripled and the global deaths caused by osteoporosis and low bone mass has increased by 111.16% reported by 2023 GBD studies [1, 2]. Osteoporosis was considered an age-related progressive bone disease and caused by bone aging which characterized by loss of bone mass and accumulation of bone marrow adipose tissue. As aging and declining of the physiological function, the activity of osteoblasts gradually decreased, while the role of osteoclasts was relatively enhanced, resulting in bone mass gradually lost and finally osteoporosis [3]. Thus, among many potential factors related with the decrease of BMD as lifestyles, genetics, gender, calcium and caffeine intake, and smoking, age was considered one of the most contributors [4]. Besides, it is crucial to identify the role of age in the association between these factors and BMD levels.
Telomere length (TL), commonly used as a reliable indicator of biologic age, cellular aging and the mitotic clock [5, 6], was reported to be associated with BMD levels [7, 8]. Natural aging was commonly discussed as the effect factor of BMD decreasing, while biological age, multiple indexes constructed to show the different health outcomes for someone with the same chronological age, might help better determine the effect of acceleration aging on osteoporosis risk [9]. The age acceleration was termed as the positive age gaps between biological age and chronological age [10] and it was considered related with epigenetic changes including genomic damage [11] and telomere reduction [12]. Epigenetic clocks of TL predictors, modeled by CpG methylation levels at specific genomic sites such as HorvathTelo [5] was widely used to estimate age acceleration. It served as a reasonably performing estimator of telomere and the DNA methylation telomere length (DNAmTL) offers a unique perspective into the complex interplay of TL dynamics [13]. Several population studies have provided positive evidences between biological aging of TL shortening and BMD decreasing. For instance, the KBASE cohort built in South Korea have found osteoporosis was associated with rapid leukocyte TL shortening [8]. While the epidemiological evidence was still limited and the potential impact paths or mechanisms was unclear.
Physical activity was related with heathy and longer lifespan [14] via enhancing aerobic fitness and skeletal muscle strength, improving glucose homeostasis, hepatic fat metabolism, and optimizing metabolic parameters [15]. Thus, regular and moderate physical activity considered a prominent non-pharmacological treatment available for preventing cancer, cardiovascular disease, vascular aging and cognitive aging, et al. [16–18]. Recent years, some studies support the positive role of physical activity in biological aging [19]. For instance, a cohort study based on 284,479 UK Biobank participants found physical activity was associated with reduced biological aging of leukocyte TL and biological aging acceleration [15]. While, a systematic review and meta-analysis showed a weak association between physical activity and TL with 2210 individuals from 11 studies, due to the publication bias, different populations and measurement of exposure and outcomes [20]. Considering the role of physical activity in aging, whether it is possible to slow aging via intervening physical activity is important for the prevention of age-related diseases.
The positive effect of physical activity on osteoporosis was commonly recognized and recent years several consensus statements on physical activity and exercise for the treatment of osteoporosis was issued [21]. A directional Mendelian randomization study supported the causal association that TL shortening influence the bone metabolism [22]. A systematic review enrolled 8 RCTs and 27 observational studies indicated physical activity appeared to help preserve TL [23]. Considering osteoporosis was a result of bone aging and affected by TL shortening, and many studies provided evidences that physical activity could preserve TL, aging acceleration measured by TL might mediate the association between physical activity and osteoporosis.
Previous studies based on NHANES providing positive associations between different types of physical activity and TL [24–26]. This study was conducted to assess the association between biological age indicators of TL and TL shortening and BMD, and identify the role of TL in the effect of physical activity on BMD levels. And the findings may offer valuable insights for prevention of age-related osteoporosis through physical activity, and promote healthy aging in older population.
Methods
Study population
The study population was derived from the U.S. National Health and Nutrition Examination Survey (NHANES). NHANES is a program aimed at evaluating the health and nutritional status of both adults and children. This study enrolled participants who took part in either the 1999–2000 or 2001–2002 surveys, totaling 21,004 individuals, and initially excluded 16,057 aged under 50 years for the DNAm biomarkers of telomere (HorvathTelo) only detected for the participants aged 50 and above. Additionally, 2,421 individuals were excluded due to missing HorvathTelo data and q-PCR method-based data of TL (TeloMean), and another 75 were excluded due to missing BMD data. Further, 57 participants were excluded because of missing physical activity data and covariates of BMI and education level. Ultimately, 2,394 participants were included in the study. The process of the inclusion and exclusion process were showed in Supplementary Figure S1.
The measurement of BMD
For the measurement of BMD, dual-energy X-ray absorptiometry (DXA) scans were conducted on eligible participants aged 8 years and older. Participants with pregnancy, self-reported radiographic contrast material use in past 7 days, self-reported nuclear medicine studies in past 3 days, and self-reported weight over 300 pounds or height over 1.96 m were prohibited from testing for BMD. The whole-body DXA scans were conducted. The DXA scans offered extensive data on various body regions, encompassing the total body, subtotal body (which refers to the total body minus the head), arms, legs, trunk, head, pelvis, and ribs. The BMD values were reported in grams per square centimeter (gm/cm²)..
Measurement of telomere length
In the NHANES 1999–2002 study, adults aged 50 years and over were eligible for examination. Purified blood samples were collected and analyzed for DNA methylation. The tests were conducted at Duke University using the Illumina EPIC BeadChip array. Methylation data matrices were produced, pre-processed, and normalized. The Horvath DNA methylation predicted telomere length (HorvathTelo)[5] was modeled. NHANES 1999–2002 also provides TL data by q-PCR method for adults 20 years and older and the indicator of TeloMean is calculated as the length of telomere relative to standard reference DNA. Data files of the predicted values of HorvathTelo and Telomean were collected and analyzed in this study.
The age-adjusted values which were the residuals from regressing of HorvathTelo and TeloMean on chronological age showed as HorvathTeloAccel and TeloMeanAccel respectively were also calculated and analyzed as the acceleration of TL shortening which was normally considered as a biomarker of biological age acceleration [27].
The assessment of physical activity and related covariates
Information regarding physical activities over the past 30 days was collected through questionnaires. Participants were queried about their “average level of physical activity each day” and given six options to choose from: “not walk about very much”, “walk a lot during the day but do not have to carry or lift things very often”, “lift light loads or climb stairs or hills often”, “heavy work or heavy loads”, “refused”, and “Don’t know”. Responses of “refused” or “Don’t know” were excluded from the analysis. The daily physical activities were then rated on a scale from level 1 to level 4, based on their exercise intensity [28].
In the study, covariates such as age, gender, race, body mass index (BMI), and education level which normally considered as potential confounders of physical activity and BMD were collected. Age, gender, race, educational levels were gathered through household interviews. BMI was calculated using data from the mobile examination center.
Statistics
For the description of the characteristics distribution, categorical variables were described using proportions, while continuous variables were described using means with standard deviations (SD). Correlational analyses were conducted between BMD in different body parts and the whole-body BMD. The BMD differences between groups by characteristics were analyzed using the t-test or an analysis of variance (ANOVA). Weighted multiple linear regression models were employed to investigate the associations between TL, physical activity, and BMD and sampling weights (WTDN4YR) which was provide by the NHANES database were applied. The effect estimate was reported as the coefficient β with 95% confidence intervals (95% CI). Covariates were adjusted in the multiple linear regression analyses. Model 1 served as the crude model, Model 2 was adjusted for age, race, and gender, and Model 3 was further adjusted for BMI and education level.
Dose-response analyses using restrict cubic spline (RCS) were utilized to assess whether the relationships between TL or accelerated TL shortening and BMD was linear. Considering the nonlinear relationships were found for some TL indicators, the TL indicators of HorvathTelo, HorvathTeloAccel, TeloMean, and TeloMeanAccel were segmented into quantiles (Q1, Q2, Q3, Q4) for all linear regression models.
A mediation analysis adjusting for all relevant covariates was conducted to evaluate whether TL indicators which found significantly associated with BMD were mediators in the association between physical activity and BMD. For HorvathTelo and TeloMean, age, gender, BMI, race, and education levels were adjusted in the mediation analyses. While considering accelerated TL shortening of HorvathTeloAccel was chronological age adjusted values, only covariates of gender, BMI, race, and education levels were controlled in the mediation analyses. Besides, mediation analyses with further controlled by calcium intake were conducted as sensitivity analyses.
Subsequently, individuals were stratified by age, BMI, and gender, and subgroup analyses were performed using multiple linear regression models. P value < 0.05 was considered statistically significant. R Studio (1.4.1106) with related packages were used for the analysis.
Results
Population characteristics
A total of 4,947 participants were recruited between 1999 and 2002 to form the initial sample. After the exclusion process, 2,394 participants were ultimately included in the final analyses. The majority of the participants were male (50.58%), non-Hispanic White (40.43%), obese (50%), and had less than a high school education (45.86%). The average age and BMI of the participants were 65.71 (SD = 9.87) years and 28.69 (SD = 5.84) kg/m², respectively. Detailed information regarding the population characteristics is presented in Table 1.
Table 1.
Baseline characteristics of study participants
| Characteristics | n | BMD (gm/cm²) | SD | p-value |
|---|---|---|---|---|
| Gender | < 0.01 | |||
| Male | 1211 | 1.14 | 0.12 | |
| Female | 1183 | 1.01 | 0.12 | |
| Age | < 0.01 | |||
| < 60 years | 677 | 1.12 | 0.12 | |
| ≥ 60 years | 1717 | 1.06 | 0.14 | |
| Race(%) | < 0.01 | |||
| Mexican American | 689 | 1.05 | 0.12 | |
| Other Hispanic | 153 | 1.06 | 0.12 | |
| Non-Hispanic White | 968 | 1.08 | 0.14 | |
| Non-Hispanic Black | 505 | 1.13 | 0.13 | |
| Other Race - Including Multi-Racial | 79 | 1.02 | 0.11 | |
| BMI | < 0.01 | |||
| Low body weight (< 18.5 kg/m2) | 22 | 0.99 | 0.14 | |
| normal body weight (18.5–24 kg/m2) | 460 | 1.03 | 0.13 | |
| Over weight (24–28 kg/m2) | 715 | 1.07 | 0.14 | |
| Obese ( > = 28 kg/m2) | 1197 | 1.10 | 0.13 | |
| Education | < 0.01 | |||
| Less than high school graduate | 1098 | 1.06 | 0.14 | |
| High school graduate or GED | 481 | 1.08 | 0.13 | |
| Some college or above | 815 | 1.10 | 0.13 |
For BMD distribution, high correlations from 0.72 to 0.94 were found between whole body BMD and BMD in different body regions including subtotal, head, left leg, right leg, left arm, right arm, left ribs, right ribs, thoracic spine, limbar spine, pelvic, and trunk bone BMD (Supplementary Table S1). Thus, whole body BMD were used for further analyses. The total BMD were 1.06 (SD = 0.13) gm/cm² in the general population, and higher BMD were observed among males, individuals aged <60 years, non-Hispanic Blacks, those with obese, and college or above education levels (P<0.05).
The distribution of telomere length and physical activity
The average DNAmTL of HorvathTelo was 6.58 (SD = 0.31). When divided by quantiles, the BMD showed an increased trend with the increase of HorvathTelo values, and the highest BMD value in Q4 HorvathTelo was 1.11 gm/cm² (SD = 0.12). Additionally, the age-adjusted value derived from the residuals of the regression of DNAmTL (HorvathTeloAccel) averaged 0.014 (SD = 0.26) and the lowest BMD appeared in the Q4 HorvathTeloAccel (Table 2). For TeloMean, the lowest BMD of 1.06 (SD = 0.14) was found in Q1 of TeloMean and increased with TeloMean (P for trend < 0.05).
Table 2.
BMD levels in physical activity groups and TL quantiles
| Characteristics | n | BMD (gm/cm²) | SD | P for trend |
|---|---|---|---|---|
| Daily physical activity | < 0.01 | |||
| not walk very much (Level 1) | 650 | 1,07 | 0.13 | |
| a lot walk, but not lift things often (Level 2) | 1371 | 1.07 | 0.14 | |
| lift light load or climb often (Level 3) | 288 | 1.09 | 0.13 | |
| heavy work or heavy loads (Level 4) | 85 | 1.14 | 0.11 | |
| HorvathTelo | < 0.01 | |||
| Q1 (≤ 6.38) | 599 | 1.05 | 0.14 | |
| Q2 (6.38∼6.59) | 598 | 1.07 | 0.14 | |
| Q3 (6.59∼6.78) | 598 | 1.08 | 0.13 | |
| Q4 (> 6.78) | 599 | 1.11 | 0.12 | |
| HorvathTeloAccel | 0.32 | |||
| Q1 (≤-0.139) | 599 | 1.08 | 0.13 | |
| Q2 (-0.139∼0.016) | 598 | 1.08 | 0.13 | |
| Q3 (0.016∼0173) | 598 | 1.07 | 0.14 | |
| Q4 (> 0.173) | 599 | 1.08 | 0.14 | |
| TeloMean | 0.01 | |||
| Q1 (≤ 0.77) | 599 | 1.06 | 0.14 | |
| Q2 (0.77∼0.90) | 598 | 1.08 | 0.14 | |
| Q3 (0.90∼1.05) | 598 | 1.08 | 0.14 | |
| Q4 (> 1.05) | 599 | 1.08 | 0.13 | |
| TeloMeanAccel | 0.41 | |||
| Q1 (≤-0.15) | 599 | 1.08 | 0.13 | |
| Q2 (-0.15∼-0.03) | 598 | 1.08 | 0.14 | |
| Q3 (-0.03∼0.12) | 598 | 1.07 | 0.14 | |
| Q4 (> 0.12) | 599 | 1.07 | 0.14 |
Among the participants, mostly (57.27%) were walking a lot, but not lifting things often (level 2 of physical activity). And as the increasing of the physical activity intensity, the BMD increased with the highest BMD of 1.14 gm/cm² (SD = 0.11) in physical activity of heavy work or heavy loads.
Association between TL and BMD
In most analyses, the RCS analyses exhibited a linear relationship, while the relationship between HorvathTeloAccel and subtotal BMD, TeloMean and total BMD, and TeloMean and subtotal BMD were nonlinear (p-nonlinear < 0.05) (Fig. 1). Multiple linear regression analyses revealed significant associations between HorvathTelo and both total bone BMD and subtotal bone BMD. Across all three models, HorvathTelo was statistically linked to total BMD and subtotal BMD. Similarly, HorvathTeloAccel also showed a statistical association with both total bone BMD and subtotal bone BMD. Also, in model 2 and model 3, TeloMean was statistically linked to total BMD and subtotal BMD (Table 3 and Supplementary Table S2).
Fig. 1.
RCS curve of HorvathTelo, HorvathTeloAccel, TeloMean, TeloMeanAccel and BMD
Table 3.
The relationships between TL and total BMD
| TL indicators | BMD | ||
|---|---|---|---|
| Model 1 | Model2 | Model3 | |
| β (95%CI) | β (95%CI) | β (95%CI) | |
| HorvathTelo | |||
| Q1 | Ref | Ref | Ref |
| Q2 | 0.028 (0.007,0.050) | 0.036 (0.020,0.052) | 0.027 (0.012,0.043) |
| Q3 | 0.052 (0.034,0.070) | 0.070 (0.057,0.083) | 0.058 (0.047,0.070) |
| Q4 | 0.070 (0.050,0.091) | 0.095 (0.076,0.113) | 0.081 (0.062,0.100) |
| P for trend | < 0.01 | < 0.01 | < 0.01 |
| HorvathTeloAccel | |||
| Q1 | Ref. | Ref. | Ref. |
| Q2 | -0.001 (-0.026,0.024) | 0.021 (0.000,0.041) | 0.009 (-0.010,0.028) |
| Q3 | -0.007 (-0.032,0.017) | 0.026 (0.007,0.045) | 0.019 (0.001,0.036) |
| Q4 | -0.015 (-0.039,0.009) | 0.029 (0.011,0.046) | 0.021 (0.005,0.038) |
| P for trend | 0.16 | < 0.01 | < 0.01 |
| TeloMean | |||
| Q1 | Ref. | Ref. | Ref. |
| Q2 | 0.027 (0.002,0.052) | 0.027 (0.002,0.051) | 0.022 (0.000,0.044) |
| Q3 | 0.026 (0.005,0.051) | 0.037 (0.016,0.059) | 0.033 (0.015,0.052) |
| Q4 | 0.023 (0.001,0.044) | 0.034 (0.015,0.052) | 0.028 (0.012,0.045) |
| P for trend | 0.08 | < 0.01 | < 0.01 |
| TeloMeanAccel | |||
| Q1 | Ref. | Ref. | Ref. |
| Q2 | -0.005 (-0.023,0.012) | 0.003 (-0.013,0.019) | 0.011 (-0.004,0.027) |
| Q3 | -0.010 (-0.032,0.012) | 0.001 (-0.020,0.021) | 0.006 (-0.010,0.023) |
| Q4 | -0.011 (-0.028,0.005) | 0.004 (-0.014,0.021) | 0.008 (-0.008,0.023) |
| P for trend | 0.21 | 0.77 | 0.53 |
Model 1 served as the crude model, Model 2 was adjusted for age, race, and gender, and Model 3 was further adjusted for BMI and education level
The role of telomere length in the association
The results of multiple linear regression analyses revealed that, across all three models, daily physical activity was positively associated with total BMD and most sites, particularly for the upper and lower limbs (Supplementary Figure S2). However, no statistical association was observed between physical activity and BMD in the head region in all three models.
The TL indicators (HorvathTelo, HorvathTeloAccel, and TeloMean), which showed significantly associations with BMD, were further analyzed for the mediated role in the association between physical activity and BMD. The mediation analyses revealed that HorvathTelo, HorvathTeloAccel, and TeloMean mediated the association between physical activity and BMD. Specifically, the mediation effect of HorvathTelo accounts for 20.86% of the total effect on total BMD and 19.08% on subtotal BMD (Fig. 2).
Fig. 2.
Mediation analyses of HorvathTelo, HorvathTeloAccel, and TeloMean in the association between physical activity and BMD
Subgroup and sensitivity analyses
The subgroup analyses revealed that HorvathTelo was positively correlated with total BMD in all subgroups of gender, age, and BMI (Supplementary Table S3). Furthermore, the analyses indicate that the HorvathTeloAccel is positively associated with total BMD in male participants (p = 0.03), those aged 65 and older (p = 0.04), and individuals with a BMI lower than 28 kg/m² (p = 0.03). In the analysis between TeloMean and BMD, significant association was found in individuals aged 65 and older and with a BMD smaller than 28 kg/m² (p < 0.05). For TeloMeanAccel, significant association was only in the group of aged 65 and older. For sensitivity analyses, mediation analyses further adjusted by calcium intake with 2308 participants showed the mediation role of TL indicators were consistent with the main analyses (Supplementary Table S4).
Discussion
This study found the mediation of TL in the association between physical activity and BMD. Our study supported the evidence that the physical activity plays positive role on BMD levels, and TL and accelerated TL shortening which represents aging acceleration was negatively associated with BMD levels. And further, the mediation role of TL indicated increase moderate exercise might have positive effect on bone health via slow down aging.
Our results also strengthen the evidence that the positive associations between TL-based biological age and BMD. Bone aging is theory for the cause of osteoporosis, and cellular senescence regulated by epigenetic regulation play a vital role in bone health [29]. Thus, clarify the association between cellular aging measured by TL and BMD can better help prevent the bone mass loss. A review about the association between TL and rate of TL shortening in osteoporosis summarized the related studies and found a number of studies had demonstrated the TL shortening might contribute to osteoporosis as an epigenetic factor [30]. Some epidemiological studies had inconsistent results. A two-sample Mendelian randomization study in European ancestry with 5 independent SNPs did not support the causal effect of TL on BMD [31]. The different population characteristics among studies, measurements of TL, study designs were potential reasons of the inconsistence. TL was still considered a predictive biomarker of in osteoporosis based on evidences from both clinical studies and mechanism studies, and a review summarized the potential of TL for predicting onset of osteoporosis [32]. Our study also supported the positive association between TL shortening and BMD decreasing.
Healthy life styles as healthy diet and physical activity were considered potential age-reversing interventions for the negative associations between these healthy behavior and epigenetic age acceleration found in some observational studies [33]. A NHANES study based on 42,625 participants from 1999 to 2018 surveys found regular physical activity were associated with lower clinically defined biological aging, regardless of age, sex, and BMI category [34]. And a further randomized controlled trial with 43 males also found that after 8 weeks intervention, health living style including physical activity, diet, supplements, sleep, and stress management could affect the deceleration of biological age [35].
The effect of physical activity on BMD or osteoporosis was reported inconsistent results based on NHANES [36–38], while a multivariable Mendelian randomization study based on UK Biobank with total 377,234 subjects found habitual vigorous physical activity could increase BMD [39]. The potential influence pathways included the key stimulus role of physical activity for bone remodeling, changes in bone geometry and architecture, indirect load to mediate the endocrine stimulation, regular vitamin D and calcium absorption [40–42]. And an existing Clinician’s Guide to Prevention and Treatment of Osteoporosis strongly suggest physical activity as an optimal way to key bone health [43]. In this study, significant association and positive trend were found between physical activity and BMD. Based on the causal assumption of physical activity with TL, and TL with BMD, we observed that TL partially mediated the association between physical activity and BMD.
TL, as a biomarker of cell aging, was considered a mediator in the association between physical activity and PhenoAge, and telomere-related process representing an important spectrum of biological aging [44]. The different measurements of TL including DNAm prediction models with varied SNPs, direct detection of TL, age-adjusted TL which represents aging acceleration varied among studies. TL was shortened as the aging process and had consistent trend with natural age, while the accelerated TL shortening which always based on the residues of the regression of TL and natural age was a better indicator of biological age acceleration and closely associated with health status as bone mass loss. Besides, many evidence were from the studies based on DNAmTL. Although the epigenetic clocks were all aimed to accurately predict biological age, the variation of the participants, specimen, DNA methylation test methods, and the statistical models made the predictive accuracy and applicability of the clocks had some differences. This study used q-PCR method-based TL (TeloMean), DNAmTL (HorvathTelo), and accelerated aging biomarker of TeloMeanAccel and HorvathTeloAccel, and positive associations were found between the TL indicators and BMD.
This study provided evidence that TL partially mediated the association between physical activity and BMD based on adequate sample size from NHANES database. This study used different TL indicators of TeloMean, HorvathTelo, and accelerated TL shortening measurements and found similar results. Despite these advantages, there were still some limitations to be addressed. First, this study was based on a cross-sectional design that the causal relationship and the fundamental assumptions of mediation analyses could not be confirmed, although consistently linear dose-response relationships were found which might strengthen the causal relationship and further prospective studies are needed to prove the findings. Second, physical activity was assessed by four categories due to the limited samples with MET data, while trend analyses by four level physical activity supported the associations with BMD. Third, potential confounders might still exist and not adjusted in our study, although three models with different covariates adjusted in the multivariable regression models were showed in the results.
Conclusions
In sum, this study found the mediation of TL in the association between physical activity and BMD. The consistent results provide new insights into the relationship between physical activity, biological age, and bone health. With the global trend of aging, actively intervening the acceleration of aging via moderate exercise might help keep healthy bone status and improve healthy aging.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
Special appreciation should be given to the NHANES team and the participants in it.
Author contributions
XZ and MT: conception and design of the study. XZ and YS: collection and analysis data. XZ: drafting the manuscript. MT: editing and approval of the final version of the manuscript.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Data availability
The datasets generated and analyzed in the current study are available in the NHANES website.
Declarations
Statements and declarations
There is no financial or non-financial interests that are directly or indirectly related to the work submitted for publication.
Conflict of interest
None.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Zhu Z, Yu P, Wu Y, Wu Y, Tan Z, Ling J et al (2023) Sex specific global burden of osteoporosis in 204 countries and territories, from 1990 to 2030: an Age-Period-Cohort modeling study. J Nutr Health Aging 27(9):767–774. 10.1007/s12603-023-1971-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Shen Y, Huang X, Wu J, Lin X, Zhou X, Zhu Z et al (2022) The global burden of osteoporosis, low bone mass, and its related fracture in 204 countries and territories, 1990–2019. Front Endocrinol (Lausanne) 13:882241. 10.3389/fendo.2022.882241 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Fang H, Deng Z, Liu J, Chen S, Deng Z, Li W (2022) The mechanism of bone remodeling after bone aging. Clin Interv Aging 17:405–415. 10.2147/CIA.S349604 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Compston JE, McClung MR, Leslie WD, Osteoporosis (2019) Lancet 393(10169):364–376. 10.1016/S0140-6736(18)32112-3 [DOI] [PubMed] [Google Scholar]
- 5.Lu AT, Seeboth A, Tsai PC, Sun D, Quach A, Reiner AP et al (2019) DNA methylation-based estimator of telomere length. Aging 11(16):5895–5923. 10.18632/aging.102173 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Vaiserman A, Krasnienkov D (2020) Telomere length as a marker of biological age: State-of-the-Art, open issues, and future perspectives. Front Genet 11:630186. 10.3389/fgene.2020.630186 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Liang X, Shi W, Zhang X, Pang R, Zhang K, Xu Q et al (2023) Causal association of epigenetic aging and osteoporosis: a bidirectional Mendelian randomization study. BMC Med Genomics 16(1):275. 10.1186/s12920-023-01708-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Han MH, Kwon HS, Hwang M, Park HH, Jeong JH, Park KW et al (2024) Association between osteoporosis and the rate of telomere shortening. Aging 16(14):11151–11161. 10.18632/aging.206034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zhu G, Guo B, Liang J (2024) Evaluating the role of biological age in osteoporosis risk among middle-aged and older adults: A nationwide perspective. Bone 189:117255. 10.1016/j.bone.2024.117255 [DOI] [PubMed] [Google Scholar]
- 10.Rutledge J, Oh H, Wyss-Coray T (2022) Measuring biological age using omics data. Nat Rev Genet 23(12):715–727. 10.1038/s41576-022-00511-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zhang H, Hao D, Gong Y, Xu Y, Ding C, Wang J et al (2025) Associations of twelve DNA methylation algorithms of aging with mortality. hLife. 10.1016/j.hlife.2025.05.010
- 12.Lopez-Otin C, Blasco MA, Partridge L, Serrano M, Kroemer G (2013) The hallmarks of aging. Cell 153(6):1194–1217. 10.1016/j.cell.2013.05.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Liang X, Aouizerat BE, So-Armah K, Cohen MH, Marconi VC, Xu K et al (2024) DNA methylation-based telomere length is associated with HIV infection, physical frailty, cancer, and all-cause mortality. Aging Cell 23(7):e14174. 10.1111/acel.14174 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dempsey PC, Rowlands AV, Strain T, Zaccardi F, Dawkins N, Razieh C et al (2022) Physical activity volume, intensity, and incident cardiovascular disease. Eur Heart J 43(46):4789–4800. 10.1093/eurheartj/ehac613 [DOI] [PubMed] [Google Scholar]
- 15.Zhu J, Yang Y, Zeng Y, Han X, Chen W, Hu Y et al (2023) The association of physical activity behaviors and patterns with aging acceleration: evidence from the UK biobank. J Gerontol Biol Sci Med Sci 78(5):753–761. 10.1093/gerona/glad064 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Gronek J, Boraczynski M, Gronek P, Wielinski D, Tarnas J, Marszalek S et al (2021) Exercise in aging: be balanced. Aging Dis 12(5):1140–1149. 10.14336/AD.2021.0107 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nelson ME, Rejeski WJ, Blair SN, Duncan PW, Judge JO, King AC et al (2007) Physical activity and public health in older adults: recommendation from the American college of sports medicine and the American heart association. Circulation 116(9):1094–1105. 10.1161/CIRCULATIONAHA.107.185650 [DOI] [PubMed] [Google Scholar]
- 18.Erickson KI, Donofry SD, Sewell KR, Brown BM, Stillman CM (2022) Cognitive aging and the promise of physical activity. Annu Rev Clin Psychol 18:417–442. 10.1146/annurev-clinpsy-072720-014213 [DOI] [PubMed] [Google Scholar]
- 19.Shin YA (2019) How does obesity and physical activity affect aging?? Focused on telomere as a biomarker of aging?. J Obes Metab Syndr 28(2):92–104. 10.7570/jomes.2019.28.2.92 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Valente C, Andrade R, Alvarez L, Rebelo-Marques A, Stamatakis E, Espregueira-Mendes J (2021) Effect of physical activity and exercise on telomere length: systematic review with meta-analysis. J Am Geriatr Soc 69(11):3285–3300. 10.1111/jgs.17334 [DOI] [PubMed] [Google Scholar]
- 21.Brooke-Wavell K, Skelton DA, Barker KL, Clark EM, De Biase S, Arnold S et al (2022) Strong, steady and straight: UK consensus statement on physical activity and exercise for osteoporosis. Br J Sports Med 56(15):837–846. 10.1136/bjsports-2021-104634 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Du X, Guo C, Zhang C, Xu B (2024) Causal association of telomere length and loss of bone: a directional Mendelian randomization study of Multi-Outcomes. Appl Biochem Biotechnol 196(10):7045–7063. 10.1007/s12010-024-04899-2 [DOI] [PubMed] [Google Scholar]
- 23.Schellnegger M, Lin AC, Hammer N, Kamolz LP (2022) Physical activity on telomere length as a biomarker for aging: A systematic review. Sports Med Open 8(1):111. 10.1186/s40798-022-00503-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Tucker LA (2020) Walking and biologic ageing: evidence based on NHANES telomere data. J Sports Sci 38(9):1026–1035. 10.1080/02640414.2020.1739896 [DOI] [PubMed] [Google Scholar]
- 25.Tucker LA (2017) Physical activity and telomere length in U.S. Men and women: an NHANES investigation. Prev Med 100:145–151. 10.1016/j.ypmed.2017.04.027 [DOI] [PubMed] [Google Scholar]
- 26.Blackmon CM, Tucker LA, Bailey BW, Davidson LE (2023) Time spent jogging/running and biological aging in 4458 U.S. Adults: an NHANES investigation. Int J Environ Res Public Health 20(19). 10.3390/ijerph20196872 [DOI] [PMC free article] [PubMed]
- 27.Okazaki S, Numata S, Otsuka I, Horai T, Kinoshita M, Sora I et al (2020) Decelerated epigenetic aging associated with mood stabilizers in the blood of patients with bipolar disorder. Transl Psychiatry 10(1):129. 10.1038/s41398-020-0813-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Perez-Lasierra JL, Casajus JA, Gonzalez-Aguero A, Moreno-Franco B (2022) Association of physical activity levels and prevalence of major degenerative diseases: evidence from the National health and nutrition examination survey (NHANES) 1999–2018. Exp Gerontol 158:111656. 10.1016/j.exger.2021.111656 [DOI] [PubMed] [Google Scholar]
- 29.Huo S, Tang X, Chen W, Gan D, Guo H, Yao Q et al (2024) Epigenetic regulations of cellular senescence in osteoporosis. Ageing Res Rev 99:102235. 10.1016/j.arr.2024.102235 [DOI] [PubMed] [Google Scholar]
- 30.Fragkiadaki P, Nikitovic D, Kalliantasi K, Sarandi E, Thanasoula M, Stivaktakis PD et al (2020) Telomere length and telomerase activity in osteoporosis and osteoarthritis. Exp Ther Med 19(3):1626–1632. 10.3892/etm.2019.8370 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Wu F, Huang Y, Hu J, Shao Z (2020) Mendelian randomization study of telomere length and bone mineral density. Aging 13(2):2015–2030. 10.18632/aging.202197 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kakridonis F, Pneumatikos SG, Vakonaki E, Berdiaki A, Tzatzarakis MN, Fragkiadaki P et al (2023) Telomere length as a predictive biomarker in osteoporosis (Review). Biomed Rep 19(5):87. 10.3892/br.2023.1669 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Quach A, Levine ME, Tanaka T, Lu AT, Chen BH, Ferrucci L et al (2017) Epigenetic clock analysis of diet, exercise, education, and lifestyle factors. Aging 9(2):419–446. 10.18632/aging.101168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Thomas A, Belsky DW, Gu Y (2023) Healthy lifestyle behaviors and biological aging in the U.S. National health and nutrition examination surveys 1999–2018. J Gerontol Biol Sci Med Sci 78(9):1535–1542. 10.1093/gerona/glad082 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Fitzgerald KN, Hodges R, Hanes D, Stack E, Cheishvili D, Szyf M et al (2021) Potential reversal of epigenetic age using a diet and lifestyle intervention: a pilot randomized clinical trial. Aging 13(7):9419–9432. 10.18632/aging.202913 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Marin-Puyalto J, Maestu J, Gomez-Cabello A, Latt E, Remmel L, Purge P et al (2019) Frequency and duration of vigorous physical activity bouts are associated with adolescent boys’ bone mineral status: A cross-sectional study. Bone 120:141–147. 10.1016/j.bone.2018.10.019 [DOI] [PubMed] [Google Scholar]
- 37.Vasquez E, Alam MT, Murillo R (2023) Race and ethnic differences in physical activity, osteopenia, and osteoporosis: results from NHANES 2009–2010, 2013–2014, 2017–2018. Arch Osteoporos 19(1):7. 10.1007/s11657-023-01356-1 [DOI] [PubMed] [Google Scholar]
- 38.Yang H, Li B, Li H, Zhou M, Li B, Guo J et al (2024) The independent and joint association between physical activity, sleep duration and daily sitting time with bone mineral density: A real world study from NHANES 2007–2018. Bone 189:117264. 10.1016/j.bone.2024.117264 [DOI] [PubMed] [Google Scholar]
- 39.Xu F, Zhang Q, Wang LK, Tang QX, Sun CQ, Deng HW (2021) Estimates of the effects of physical activity on osteoporosis using multivariable Mendelian randomization analysis. Osteoporos Int 32(7):1359–1367. 10.1007/s00198-020-05786-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Forwood MR (2001) Mechanical effects on the skeleton: are there clinical implications? Osteoporos Int 12(1):77–83. 10.1007/s001980170161 [DOI] [PubMed] [Google Scholar]
- 41.Duncan RL, Turner CH (1995) Mechanotransduction and the functional response of bone to mechanical strain. Calcif Tissue Int 57(5):344–358. 10.1007/BF00302070 [DOI] [PubMed] [Google Scholar]
- 42.Wang Y, Xu D (2017) Effects of aerobic exercise on lipids and lipoproteins. Lipids Health Dis 16(1):132. 10.1186/s12944-017-0515-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.LeBoff MS, Greenspan SL, Insogna KL, Lewiecki EM, Saag KG, Singer AJ et al (2022) The clinician’s guide to prevention and treatment of osteoporosis. Osteoporos Int 33(10):2049–2102. 10.1007/s00198-021-05900-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.You Y, Wang D, Ding H, Wang W, Liu Q, Zhang D et al (2025) Mediation role of telomere length in the relationship between physical activity and phenoage: A population-based study. J Exerc Sci Fit 23(3):149–156. 10.1016/j.jesf.2025.03.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and analyzed in the current study are available in the NHANES website.