Abstract
Background
Prolonged inactivity has been found to be a major contributor to poor bone health. However, limited research has examined the specific relationship between pelvic bone mineral density (BMD) and sedentary time. The study aims to clarify the association between sedentary time and pelvic BMD in adult Americans.
Methods
Data from the National Health and Nutrition Examination Survey (NHANES) from 2011 to 2018 were examined in this cross-sectional investigation. Participants who had comprehensive information on their pelvic BMD and amount of sedentary time were included. We used multivariable linear regression to assess the association between sedentary time and pelvic BMD, and logistic regression to estimate the odds of low BMD across sedentary levels. Potential effect modification and contributing factors were explored using interaction tests and prespecified subgroup analyses. All analyses accounted for the complex NHANES sampling design with survey weights, strata, and primary sampling units.
Results
The analysis includes 15,328 individuals. Longer sedentary time was associated with lower pelvic BMD. In the fully adjusted model, the ≥ 8 h/day group had lower pelvic BMD compared with the < 4 h/day reference group (β = -0.018, 95% CI: -0.031, -0.006; P < 0.01). Subgroup and interaction analyses suggested effect modification by body mass index and glucocorticoid use (P for interaction = 0.047 and 0.003, respectively).
Conclusion
Longer sedentary time is associated with lower pelvic BMD. This finding underscores the importance of controlling sedentary behavior to stop the decline of bone health. Nevertheless, large prospective studies are required to confirm the results.
Keywords: Sedentary, Pelvis, Bone mineral, NHANES
Introduction
Sedentary behavior has become increasingly prevalent in modern lifestyles, with growing evidence linking it to adverse health outcomes. Increasing evidence shows an inverse dose-response between sedentary time and all-cause mortality, whereas higher physical activity confers graded reductions in risk [1]. Insights from inactivity and bedrest models further delineate the physiology of sedentary exposure, including loss of muscle mass and strength, insulin resistance, low-grade inflammation, and fluid‑shift-related adaptations, that collectively diminish skeletal loading and can perturb bone remodeling [2–5]. Consistent with these pathways, sedentary behavior is associated with adverse metabolic and musculoskeletal outcomes. While most prior work has examined the effects of sedentary behavior on bone health at major skeletal sites such as the spine and femur [6, 7], relatively little research has focused on pelvic bone mineral density (BMD).
Prior studies of youth reported dose‑dependent effects of physical activity on both BMD (volumetric bone mineral density [vBMD], bone mineral content [BMC]) and bone geometry, with site‑ and compartment‑specific patterns (cortical: density and size; trabecular: predominantly density) [8]. Mechanistic and review evidence further indicates that physical activity can also improve BMD through dynamic loading pathways [9]. In older adults, higher physical activity is associated with greater muscle mass and leg strength, which may secondarily support skeletal integrity [10]. A study from Ricci et al. further suggests that replacing sedentary time with light activity reduces the risk of spinal osteoporosis, with stronger associations among overweight women and those older than 65 years [11]. However, most of these data were derived from the proximal femur or lumbar spine and may not generalize to other skeletal sites.
The pelvis is a pivotal bony hub linking the trunk and lower limbs that underpins athletic performance, postural stability, and female reproductive and pelvic‑floor function. Alterations in pelvic BMD can therefore carry broad functional consequences [12, 13]. Mechanobiologically, the pelvis operates within a distinct lumbopelvic-hip load‑transfer system, in which it sustains multidirectional joint‑reaction forces and shear across the sacroiliac and acetabular regions, with strain fields modulated by gluteal and pelvic‑floor muscle activity [14, 15]. Prolonged sitting redistributes loads through the pelvic ring and suppresses key muscle activations, thereby altering local strain environments that modulate osteocyte‑mediated mechanotransduction [9, 16]. Given region‑specific trabecular architecture and cortical‑to‑trabecular ratios, such redistribution may elicit remodeling responses in the pelvis that differ from those at the spine or femur, providing a biologically plausible basis for site‑specific effects of sedentary time on pelvic BMD [17, 18]. Clinically, lower pelvic BMD is relevant to pelvic fragility fractures and mobility impairments and is particularly important in women - owing to sex‑dimorphic pelvic morphology and hormonal regulation of bone remodeling - and in older adults at heightened osteoporotic risk [19, 20]. These mechanistic and clinical considerations underscore the need to evaluate pelvic BMD directly rather than extrapolate from other skeletal sites. Thus, we hypothesized that greater sedentary time would be associated with lower pelvic BMD. Accordingly, we examined this association in a nationally representative sample.
Methods
Study design and population
NHANES is a comprehensive, nationally representative survey of the non‑institutionalized civilian population in the United States. As the principal health statistics agency in the United States, the National Center for Health Statistics (NCHS) ensures that the NHANES data are collected using standardized, high‑quality procedures. All participants provided written informed consent. For this research, we obtained data from four consecutive NHANES cycles (2011–2012, 2013–2014, 2015–2016, 2017–2018) through the NHANES website (https://www.cdc.gov/nchs/nhanes/index.htm). We analyzed participant data from the 2011–2018 NHANES cycles. Because whole‑body dual‑energy X‑ray absorptiometry (DXA), including pelvic BMD, was performed using consistent protocols and publicly released only for participants younger than 60 years in these cycles, subsequent NHANES cycles do not provide whole‑body DXA data that are fully comparable or publicly available. Each cycle’s dataset includes five categories: demographic, dietary, examination, laboratory, and questionnaire data. Across these cycles, 39,156 individuals were surveyed. NHANES applied program‑level DXA exclusions at the examination centers: pregnancy (urine test and/or self‑report), self‑reported barium contrast within 7 days, and scanner size limits (> 6’5” [195 cm] or weight > 450 lb [204 kg]). To avoid bias from conditions that substantially alter bone metabolism or indicate atypical daily mechanical loading, we excluded individuals with chronic kidney disease (n = 627), a history of malignancy (n = 613), thyroid disease (n = 1391), rheumatoid arthritis (n = 366), or severe mobility limitations indicative of abnormal habitual weight-bearing (inability to walk a quarter mile or to climb 10 steps without stopping; (n = 768). We then removed participants with missing pelvic BMD (n = 6563) or age < 18 (n = 1,3256), followed by those lacking sedentary time data (n = 244). After these sequential exclusions, the final analytic sample comprised 15,328 participants (Fig. 1).
Fig. 1.
Flow chart of study participants
Definitions of sedentary and pelvic bone mass density
The exposure variable, sedentary time, was derived from the Physical Activity (PA) questionnaire, which asks participants: “How much time do you usually spend sitting in a typical day?” This measure included time spent sitting at a desk, traveling, reading, watching TV, or using a computer, but excludes sleep time [21]. The outcome variable, pelvic BMD, was assessed using Whole‑body DXA. According to the NHANES DXA protocol for the examined cycles, whole‑body and pelvic DXA were performed in participants aged 8–59 years; therefore, the analytic sample did not include adults over 60 years.
Covariate
Covariates were prespecified based on prior epidemiologic evidence regarding determinants of BMD [22–24] and included age, gender, race, body mass index (BMI), smoking status, alcohol consumption, PA, glucocorticoid use, and total blood calcium levels [25, 26]. Data on glucocorticoid use were obtained from prescription medication questionnaires and pill bottle reviews. Participants were asked, “In the past 30 days, have you taken any prescription medications?” Those who answered “yes” were asked to list all medications used; individuals reporting prednisone, prednisolone, methylprednisolone, triamcinolone, methylprednisolone acetate, dexamethasone, or betamethasone were classified as glucocorticoid users [27]. Smoking behavior was defined using the questionnaire item “Smoked at least 100 cigarettes in life?” Participants answering “yes” were identified as ever smokers; otherwise, they were classified as never smokers. Alcohol consumption was determined from two items in the Alcohol Use questionnaire: “Had at least 12 alcoholic drinks in 1 year?” and “Had at least 12 alcoholic drinks in lifetime?” Participants answering “yes” to either item were classified as alcohol consumers. To standardize the assessment of PA levels, metabolic equivalents of task (METs) were calculated and used. The questionnaire covered five types of physical activities with different intensities: (i) high-intensity work-related activities (MET = 8); (ii) moderate-intensity work-related activities (MET = 4); (iii) transportation-related activities (MET = 4), including walking or cycling; (iv) high-intensity leisure-related activities (MET = 8); (v) moderate-intensity leisure-related activities (MET = 4). Each activity type was quantified in MET-minutes per week, calculated as: MET value * weekly frequency * duration of each activity (minutes). Overall PA level was defined as the sum of MET values of the above five types of activities. Sedentary behavior was assessed with the question: “How much time do you usually spend sitting or lying down each day?” Responses were used to calculate each participant’s daily sedentary time. In the subsequent regression analysis and visualization, to more clearly show the effect size, the PA units were converted to thousand MET-minutes per week [28]. Missing data were handled as follows: variables for BMI and glucocorticoid use were analyzed using complete-case analysis, in which observations with missing values were excluded. For all other covariates obtained from questionnaires - such as smoking status, alcohol use, and the family income-to-poverty ratio (PIR) - missing values were handled using multiple imputation by chained equations (MICE). We performed 30 imputations with 20 iterations each, using predictive mean matching for continuous variables and (multinomial) logistic regression models for categorical variables. Each imputed dataset was analyzed accounting for the NHANES complex survey design (including weights, strata, and primary sampling units), and results were pooled according to Rubin’s rules. Detailed measurement protocols are available on the NHANES website (www.cdc.gov/nchs/nhanes).
Statistical analysis
Statistical analyses followed CDC/NCHS guidelines for analyzing NHANES complex survey data. Categorical variables are presented as counts (percentages), and continuous variables as means and standard deviations. Participants were grouped by quartiles of pelvic BMD to describe baseline characteristics. All estimates incorporated NHANES sampling weights with stratification and clustering to account for the multistage design and yield nationally representative inferences. Because the outcome variable was derived from the examination component, MEC examination weights were used; the 2‑year weights (WTMEC2YR) were divided by 4 for the 2011–2018 combined analyses. Variances were estimated using Taylor series linearization, specifying Survey Design and Demographic Variables - Stratum (SDMVSTRA) and Survey Design and Demographic Variables - Primary Sampling Unit (SDMVPSU), with the standard lonely‑PSU adjustment. Three sampling-weighted linear regression models were used to assess the association between sedentary time and pelvic BMD: Model 1 was unadjusted; Model 2 adjusted for age, sex, and race/ethnicity; and Model 3 further adjusted for total serum calcium, BMI, PIR, education level, glucocorticoid use, smoking status, and alcohol consumption. Stratified analyses were conducted by sex, age, race, education level, PIR, BMI, total serum calcium, smoking status, alcohol consumption and PA, subgroup heterogeneity was assessed using multiplicative interaction terms. We employed survey‑weighted restricted cubic spline (RCS) regression using the rms package in R. Knots were placed at the 5th, 35th, 65th, and 95th percentiles of sedentary time. Both overall and nonlinear likelihood‑ratio tests were performed; P < 0.05 for the nonlinear component indicated deviation from linearity. All tests were two-sided, with p < 0.05 considered statistically significant. Analyses were performed in R version 4.2.2 using survey package for design specification and weighted estimation and mice package for multiple imputation, and were cross‑checked in EmpowerStats 2.0 (X&Y Solutions, Boston, MA).
Results
Baseline population characteristics of participants
This study included 15,328 participants with a survey-weighted mean age of 35.90 years, comprising 52.2% males and 47.8% females. The survey‑weighted mean pelvic BMD was 1.240 g/cm², which declined with increasing sedentary time. Participants with ≥ 8 h/day of sedentary time had a higher survey‑weighted prevalence of low pelvic BMD compared with those with less sedentary time (p < 0.001). Across pelvic BMD quartiles (Q1 - Q4), the unweighted counts within the ≥ 8 h/day sedentary‑time category were 1,823, 1,682, 1,682, and 1,569, respectively; the corresponding survey‑weighted prevalences were 47.7%, 44.1%, 43.7%, and 40.9%. Baseline characteristics differed significantly across pelvic BMD quartiles for sex, age, race, PIR, education level, glucocorticoid use, alcohol use, BMI, total blood calcium and PA (all p < 0.05), whereas smoking status did not differ (p > 0.05). Compared with Q4 (highest BMD), participants in Q1 (lowest BMD) were more likely to be female, have lower PIR, use glucocorticoids more frequently, have lower BMI, report alcohol use, and have higher total blood calcium levels and more PA (all p < 0.05; Table 1).
Table 1.
Baseline characteristics by pelvic BMD quartiles (survey‑weighted)
| Pelvis BMD (g/cm2) | P-value | ||||
|---|---|---|---|---|---|
| Q1 (0.67–1.15, n = 3,825) |
Q2 (1.15–1.22, n = 3,822) |
Q3 (1.22–1.34, n = 3,848) |
Q4 (1.34–2.77, n = 3,833) |
||
| Daily sedentary time (hour) | < 0.001 | ||||
| < 4 | 665 (17.4%) | 747 (19.6%) | 770 (20.0%) | 796 (20.8%) | |
| 4–6 | 786 (20.6%) | 766 (20.0%) | 837 (21.8%) | 895 (23.3%) | |
| 6–8 | 551 (14.4%) | 624 (16.3%) | 559 (14.5%) | 573 (14.9%) | |
| ≥ 8 | 1823 (47.3%) | 1685 (44.1%) | 1682 (43.7%) | 1569 (40.9%) | |
| Sex | < 0.001 | ||||
| Male | 1679 (43.9%) | 1724 (45.1%) | 2044 (53.2%) | 2494 (65.1%) | |
| Female | 2143 (56.1%) | 2096 (54.9%) | 1802 (46.8%) | 1334 (34.9%) | |
| Age (year, continuous) | 36.8 ± 16.2 | 35.8 ± 13.6 | 34.2 ± 13.6 | 33.3 ± 12.9 | < 0.001 |
| Age (year, categories) | < 0.001 | ||||
| 18–25 | 1260 (32.9%) | 952 (24.9%) | 944 (24.5%) | 878 (22.9%) | |
| 25–39 | 803 (21.0%) | 1194 (31.2%) | 1277 (33.2%) | 1244 (37.7%) | |
| 40–49 | 606 (15.8%) | 830 (21.7%) | 900 (23.4%) | 869 (22.7%) | |
| 50–59 | 1156 (30.2%) | 846 (22.1%) | 727 (18.9%) | 642 (16.8%) | |
| Race | < 0.001 | ||||
| Mexican American | 426 (11.1%) | 427 (11.2%) | 419 (10.9%) | 384 (10.0%) | |
| Other Hispanic | 321 (8.4%) | 304 (8.0%) | 271 (7.0%) | 229 (6.0%) | |
| Non-Hispanic White | 2295 (60.0%) | 2311 (60.5%) | 2408 (62.6%) | 2207 (57.6%) | |
| Non-Hispanic Black | 311 (8.1%) | 381 (10.0%) | 443 (11.5%) | 723 (18.9%) | |
| Other Race - Including Multi-Racial | 472 (12.3%) | 399 (10.4%) | 307 (8.0%) | 290 (7.6%) | |
| PIR | 2.80 ± 1.69 | 2.88 ± 1.69 | 2.84 ± 1.66 | 2.97 ± 1.66 | < 0.001 |
| Education level | 0.023 | ||||
| Less than 9th grade | 193 (5.1%) | 161 (4.2%) | 129 (3.4%) | 124 (3.2%) | |
| 9-11th grade | 385 (10.1%) | 348 (9.1%) | 370 (9.6%) | 344 (9.0%) | |
| High school or GED | 822 (21.5%) | 791 (20.7%) | 869 (22.6%) | 839 (21.9%) | |
| Some college or AA degree | 1187 (31.0%) | 1275 (33.3%) | 1251 (32.5%) | 1297 (33.8%) | |
| College graduate or above | 1237 (32.4%) | 1247 (32.6%) | 1229 (32.0%) | 1228 (32.0%) | |
| Glucocorticoid use | 0.025 | ||||
| Yes | 62 (1.6%) | 46 (1.2%) | 36 (0.9%) | 38 (1.0%) | |
| No | 3723 (98.4%) | 3740 (98.8%) | 3779 (99.1%) | 3761 (99.0%) | |
| Smoking status | 0.935 | ||||
| Yes | 1538 (40.2%) | 1547 (40.5%) | 1534 (39.9%) | 1554 (40.6%) | |
| No | 2287 (59.8%) | 2275 (59.5%) | 2314 (60.1%) | 2279 (59.5%) | |
| Alcohol use | < 0.001 | ||||
| Yes | 3182 (83.2%) | 3010 (78.8%) | 3089 (80.8%) | 2884 (75.5%) | |
| No | 643 (16.8%) | 838 (21.2%) | 733 (19.2%) | 938 (24.5%) | |
| BMI (kg/m 2) | 26.0 ± 8.3 | 28.0 ± 7.2 | 29.0 ± 6.4 | 30.1 ± 5.5 | < 0.001 |
| Total blood calcium (mg/dL) | 9.42 ± 0.35 | 9.40 ± 0.35 | 9.40 ± 0.34 | 9.40 ± 0.34 | 0.048 |
| PA | 3.61 ± 5.82 | 4.23 ± 6.26 | 4.47 ± 6.25 | 5.54 ± 7.33 | < 0.001 |
Mean ± SD for continuous variables; P < 0.05 presents significant difference
Unweighted counts are shown; percentages are survey‑weighted
Abbreviations: BMD bone mineral density, h hour, GED General Educational Development, PIR poverty-to-income ratio, BMI body mass index, PA physical activity
Association between sedentary time and pelvic BMD
Table 2 summarizes the association between daily sedentary time and pelvic BMD. Compared with the < 4 h/day reference group, a consistent inverse association was observed for the ≥ 8 h/day category across all three models. In the fully adjusted model, the ≥ 8 h/day group had lower pelvic BMD relative to the reference group (β = −0.018, 95% CI: −0.031 to −0.006, p < 0.01). This coefficient represents the mean difference for the ≥ 8 h/day category versus < 4 h/day, rather than a per‑hour effect. The negative pattern was further examined using RCS modeling. The spline curve was visually linear (P‑nonlinear = 0.198), consistent with a monotonic decrease in pelvic BMD with increasing sedentary time. Using a survey‑weighted RCS model (Fig. 2A), we explored potential nonlinear associations between sedentary time and pelvic BMD. The overall association was significant (P < 0.001), but the test for nonlinearity was not (P‑nonlinear = 0.198), suggesting an approximately linear trend. The linear regression model (Fig. 2B) indicated that each additional hour of daily sedentary time was associated with a decrease of 0.0024 g/cm² (95% CI: −0.0035 to −0.0013; P < 0.001) in pelvic BMD, after adjustment for age, sex, race, education, BMI, PIR, serum calcium, glucocorticoid use, alcohol consumption, smoking status, and PA.
Table 2.
The connection between survey-weighted daily sedentary time and pelvic BMD
| Daily sedentary time (hour) | Crude model (Model 1) | Minimally adjusted model (Model 2) | Fully adjusted model (Model 3) |
|---|---|---|---|
| < 4 (β, 95% CI) | Ref | Ref | Ref |
| 4–6 (β, 95% CI) | −0.006 (−0.017, 0.004) | −0.006 (−0.016, 0.004) | −0.012 (−0.026, 0.001) |
| 6–8 (β, 95% CI) | −0.020 (−0.030, −0.009) *** | −0.003 (−0.013, −0.006) ** | −0.008 (−0.021, 0.006) |
| ≥ 8 (β, 95% CI) | −0.022 (−0.032, −0.013) *** | −0.010 (−0.020, 0.001) * | −0.018 (−0.031, −0.006) ** |
Survey-weighted models accounted for the NHANES complex sampling design
Model 1, No covariates were adjusted;
Model 2, Adjusted for sex, age and race;
Model 3, Adjusted for sex, age, race, education level, BMI, PIR, serum calcium, use of glucocorticoids, alcohol use, smoking status and PA
95% CI, 95% Confidence Interval; *P < 0.05, **P < 0.01, ***P < 0.001
Fig. 2.
Association between sedentary time and pelvic BMD. A RCS model displaying the nonlinear dose‑response pattern between sedentary time and pelvic BMD, with shaded areas representing the 95% confidence interval. B Survey‑weighted linear regression demonstrating the approximately linear relationship between sedentary time and pelvic BMD
Subgroup analysis
Subgroup analyses (Table 3) demonstrate that the link between sedentary time and pelvic BMD varies across different subgroups. In men, sedentary time ≥ 8 h/day was associated with lower pelvic BMD relative to < 4 h/day (β = −0.021, 95% CI: −0.038, −0.003, P < 0.05). In women, the association for the ≥ 8 h/day category was β = −0.017 (95% CI: −0.035, 0.001), which was not statistically significant. Age categories were prespecified to reflect skeletal development and early bone loss phases within the 18 to 59 years range: < 25, 25 to 39, 40 to 49 and 50 to 59 years. Among individuals aged 50 to 59 years, those with ≥ 8 h of sedentary time per day exhibited significantly lower pelvic BMD compared to those with < 4 h/day (β = −0.042; 95% CI: −0.073, −0.010; P < 0.05). This inverse association remained after multivariable adjustment for sex, age, race, BMI, smoking, alcohol use, PIR, educational level, PA and total blood calcium. Interaction analyses indicated that the association between sedentary time and pelvic BMD differed by BMI category and by glucocorticoid use (P for interaction with BMI = 0.047; with glucocorticoid use = 0.003).
Table 3.
Subgroup analysis for the association between survey-weighted daily sedentary and pelvic BMD
| Characteristics | Daily sedentary time(h) | P for interaction | |||
|---|---|---|---|---|---|
| < 4 (β,95% CI) |
4–6 (β,95% CI) |
6–8 (β,95% CI) |
≥ 8 (β,95% CI) |
||
| Sex | 0.650 | ||||
| Male | Ref | −0.006 (−0.029, 0.016) | =−0.004 (−0.025, 0.016) | −0.021 (−0.038, −0.003) * | |
| Female | Ref | −0.020 (−0.037, −0.004) * | −0.013 (−0.031, 0.004) | −0.017 (−0.035, 0.001) | |
| Age | 0.169 | ||||
| 18–25 | Ref | −0.010 (−0.041, 0.021) | 0.007 (−0.025, 0.040) | −0.005 (−0.036, 0.025) | |
| 25–39 | Ref | 0.004 (−0.013, 0.021) | −0.009 (−0.030, 0.013) | −0.010 (−0.029, 0.009) | |
| 40–49 | Ref | 0.017 (−0.039, 0.004) | −0.003 (−0.023, 0.017) | −0.003 (−0.027, 0.021) | |
| 50–59 | Ref | −0.022 (−0.051, 0.007) | −0.013 (−0.044, 0.017) | −0.042 (−0.073, −0.010) * | |
| Race | 0.540 | ||||
| Mexican American | Ref | −0.014 (−0.022, 0.050) | −0.028 (0.001, 0.055) * | −0.002 (−0.029, 0.034) | |
| Other Hispanic | Ref | −0.021 (−0.039, −0.003) * | −0.006 (−0.024, 0.011) | −0.011 (−0.027, 0.004) | |
| Non-Hispanic White | Ref | −0.001 (−0.022, 0.019) | −0.009 (−0.028, 0.009) * | −0.030 (−0.047, −0.013) ** | |
| Non-Hispanic Black | Ref | −0.028 (−0.052, −0.004) * | −0.011 (−0.043, 0.021) | −0.034 (−0.059, −0.008) * | |
| Other Race - Including Multi-Racial | Ref | 0.009 (−0.022, 0.005) | −0.007 (−0.021, 0.007) | −0.015 (−0.028, −0.002) * | |
| Smoking status | 0.117 | ||||
| Yes | Ref | −0.001 (−0.022, 0.019) | −0.009 (−0.028, 0.009) | −0.030 (−0.047, −0.013) ** | |
| No | Ref | −0.021 (−0.039, −0.003) * | −0.006 (−0.024, 0.011) | −0.011 (−0.027, 0.004) | |
| Alcohol use | 0.408 | ||||
| Yes | Ref | −0.009 (−0.022, 0.005) | −0.007 (−0.021, 0.007) | −0.015 (−0.028, −0.002) * | |
| No | Ref | −0.028 (−0.052, −0.004) * | −0.011 (−0.043, 0.021) | −0.034 (−0.059, −0.008) * | |
| BMI | 0.047 | ||||
| < 25 | Ref | −0.002 (−0.022, 0.019) | 0.015 (−0.008, 0.039) | 0.010 (−0.008, 0.029) | |
| ≥ 25 | Ref | −0.014 (−0.032, 0.004) | −0.017 (−0.033, −0.002) * | −0.027 (−0.042, −0.011) ** | |
| PIR | 0.058 | ||||
| < 1.3 | Ref | −0.008 (−0.027, 0.010) | −0.013 (−0.035, 0.008) | −0.028 (−0.051, −0.005) * | |
| 1.3–3.5 | Ref | −0.020 (−0.039, −0.001) * | −0.015 −0.015 (−0.035, 0.006) | −0.001 (−0.023, 0.021) | |
| ≥ 3.5 | Ref | −0.006 (−0.028, 0.016) | 0.001 (−0.024, 0.027) | −0.022 (−0.046, 0.002) | |
| Education level | 0.952 | ||||
| Less than 9th grade | Ref | 0.002 (−0.035, 0.045) | −0.035 (−0.082, 0.025) | −0.018 (−0.071, 0.016) | |
| 9-11th grade | Ref |
0.005 (−0.054, 0.025) |
−0.031 (−0.042, 0.030) | −0.021 (−0.066, −0.004) * | |
| High school or GED | Ref |
0.001 (−0.023, 0.024) |
0.008 (−0.036, 0.021) |
−0.012 (−0.035, 0.012) | |
| Some college or AA degree | Ref |
−0.016 (−0.038, 0.006) |
0.004 (−0.020, 0.029) |
−0.014 (−0.037, 0.009) | |
| College graduate or above | Ref |
−0.015 (−0.041, 0.015) |
−0.006 (−0.041, 0.021) |
−0.035 (−0.043, 0.009) | |
| Total blood calcium | 0.488 | ||||
| < 8.5 | Ref |
−0.089 (−0.326, 0.075) |
−0.075 (−0.351, 0.053) |
−0.093 (−0.325, 0.002) | |
| 8.5–10 | Ref |
−0.005 (−0.026, 0.001) |
−0.013 (−0.020, 0.007) |
−0.018 (−0.030, −0.005) ** | |
| ≥ 10 | Ref |
−0.056 (−0.096, −0.018) ** |
−0.033 (−0.091, 0.017) |
−0.064 (−0.106, −0.021) ** | |
| Glucocorticoid use | 0.003 | ||||
| No | Ref. |
−0.012 (−0.026, 0.001) |
−0.008 (−0.021, 0.006) |
−0.018 (−0.031, −0.005) ** | |
| Yes | Ref |
0.015 (−0.076, 0.113) |
0.065 (−0.056, 0.252) |
−0.066 (−0.126, −0.021) | |
| PA | 0.321 | ||||
| Low | Ref |
−0.016 (−0.039, 0.006) |
−0.011 (−0.037, 0.015) |
−0.028 (−0.052, −0.003) * | |
| Moderate | Ref |
−0.016 (−0.039, 0.006) |
−0.002 (−0.025, 0.020) |
−0.005 (−0.027, 0.017) |
|
| High | Ref |
−0.008 (−0.028, 0.012) |
−0.010 (−0.033, 0.013) |
−0.017 (−0.042, 0.008) |
|
A fully adjusted model: adjusted for sex, age, race, smoking status, alcohol use, BMI, PIR, education level, total blood calcium, use of glucocorticoids and PA
Within each stratification, models were fully adjusted for all covariates except the stratifying variable itself
*p < 0.05, **p < 0.01, ***p < 0.001
Discussion
Sedentary time and pelvic BMD were negatively associated in this cross‑sectional study of 15,328 participants. Gender, BMI, and glucocorticoid use were significant modifiers of this association. These findings imply that prolonged inactivity may increase the risk of low pelvic BMD, underscoring the potential benefit of reducing daily sedentary time.
The results of earlier research about BMD and sedentary time have been mixed. For instance, Chastin et al. did not find a substantial correlation between pelvic BMD and sitting time, likely due to a small sample size and methodological limitations [6]. In contrast, Kopiczko et al. reported an inverse association between sedentary time and BMD of the radius and ulna in 115 adolescent boys aged 14–17 years [25]. Using data on 10,346 adults, Yang et al. identified associations between PA, sleep duration, sedentary time, and BMD at the lumbar spine and femoral neck [7], and Weiss et al. noted that occupations involving prolonged standing or walking were associated with higher BMD than more sedentary jobs [26]. João et al., in a study of female adolescent athletes in Portugal, found that more intensive PA was linked to higher bone density in the lower extremities [29]. Additionally, Zhao et al. reported findings consistent with ours, that sedentary time and lumbar spine BMD are inversely associated [30]. Taken together, although heterogeneity exists across age groups, skeletal sites, and exposure definitions (sedentary time versus PA), a substantial portion of the literature - along with our findings - supports a generally unfavorable relationship between sedentary behavior and BMD; nonetheless, residual heterogeneity remains.
Our study extends this literature by focusing on the pelvis and demonstrating subgroup differences. The inverse association between sedentary time and pelvic BMD was stronger in participants with BMI ≥ 25 kg/m² and in those not using glucocorticoids, with a significant interaction for BMI (P for interaction = 0.047) and for glucocorticoid use (P for interaction = 0.003). However, these findings, particularly within the 4–6 and 6–8 h/day sedentary categories, should be interpreted with caution due to the limited sample size of glucocorticoid users, which may lead to unstable estimates. Several mechanisms could plausibly contribute to these patterns. Localized mechanical stimulation from muscle contraction and joint motion is crucial for osteogenesis [31, 32]. Prolonged sitting minimally activates the quadriceps and hamstring muscles, which are essential for positively influencing the pelvis and lower limb bones [33–35]. Additionally, sitting reduces the routine weight-bearing and mechanical stress on the pelvic and lower limb bones, stimuli that are necessary for bone formation and maintaining bone density [36–38]. In individuals with higher body weight, although greater BMI increases static loading, obesity‑related adipokine dysregulation and low‑grade inflammation can promote bone resorption and deteriorate microarchitecture, potentially offsetting mechanical benefits. Under sedentary conditions - with reduced lumbopelvic muscle activation and diminished dynamic strains - these factors may jointly attenuate osteogenic stimuli at the pelvis, consistent with concepts such as sarcopenic obesity and marrow adiposity [39, 40]. Moreover, glucocorticoids rapidly reduce BMD by suppressing osteoblastogenesis, enhancing resorption, and inhibiting Wnt signaling, and they may induce steroid‑related myopathy [41–43]. Concurrently, sedentary behavior reduces dynamic mechanical loading and lumbopelvic muscle activation, attenuating osteogenic mechanotransduction. These pathways plausibly interact, endocrine suppression of bone formation combined with mechanical unloading, to amplify net bone loss at the pelvis. The observed interaction with glucocorticoids may also be influenced by confounding by indication and bone‑active co‑therapies among glucocorticoid users. Interestingly, findings at the lumbar spine have been mixed, suggesting potential site‑specific differences; the pelvic association may relate to sitting posture and lumbopelvic muscle activation patterns [6, 44–49].
Although a negative association between sedentary time and pelvic BMD was identified, the clinical relevance of the effect size requires careful interpretation. In our fully adjusted model, participants with ≥ 8 h/day of sedentary time had a mean pelvic BMD 0.018 g/cm² lower than the reference group (< 4 h/day). To contextualize this change, we calculated the standardized coefficient for the ≥ 8 h/day group, which was − 0.012 (calculated using the formula: standardized β* = β × (SDx/SDy), where β is the unstandardized coefficient from Table 3 [−0.018 g/cm²], SDx is the standard deviation of sedentary time [0.111], and SDy is the standard deviation of pelvic BMD [0.167 g/cm²]). This indicates a small effect size by Cohen’s conventions (where the absolute value of β* ≈ 0.1 is considered small) [50].
Furthermore, each additional hour of daily sedentary time was associated with a decrease of 0.0024 g/cm² in pelvic BMD. While this per-hour effect is modest, it may accumulate over time and contribute to long-term bone loss. The clinical impact of such BMD changes can be inferred from prospective studies linking BMD to fracture risk. A meta-analysis by Johnell et al. found that each 1 standard deviation (SD) decrease in hip BMD was associated with a 2.6-fold increase in hip fracture risk [51]. In our study population, the SD for pelvic BMD was approximately 0.167 g/cm². Thus, the observed difference of 0.018 g/cm² between the highest and lowest sedentary groups represents approximately 10.8% of 1 SD. While this isolated difference might not translate to a dramatic short-term increase in individual fracture risk, it could be clinically meaningful at a population level. Moreover, for individuals with other risk factors (e.g., older age, glucocorticoid use, high BMI), this sedentary-related BMD reduction could compound their overall risk profile. Therefore, although the effect size is modest, reducing prolonged sedentary behavior may still be a valuable component of a broader strategy for osteoporosis prevention.
This study offers several advantages. First, it systematically explores the association between sedentary behavior and pelvic BMD, whereas previous research has primarily focused on changes in spinal and femoral BMD. The findings indicate that the impact of sedentary behavior on BMD is site‑specific, similar to the regional effects of PA, but that its influence is more mechanical and localized. Second, our study used nationally representative NHANES data and survey‑weighted analyses with standardized protocols, enhancing generalizability. Additionally, adjusted for multiple confounders to improve the robustness of the estimates.
Nonetheless, it is important to recognize the limitations of the study. The cross-sectional design did not allow for a clear inference of causality, and the BMD data were based only on DXA measurements at a single time point, which may not fully capture long‑term changes in BMD. Long-term data on sedentary behavior would be more informative, necessitating future large-scale cohort studies to confirm these findings. Although the dataset is nationally representative, it is based on data from 2011 to 2018 and may not fully capture current trends. The authors attempted to use more recent data, but pelvic BMD information was only available in the 2011–2018 NHANES survey cycles. Moreover, while pelvic bone health is particularly pertinent to older adults, the NHANES DXA protocol in the analyzed cycles restricted measurements to participants aged under 60 years. Therefore, our findings apply to youth and midlife adults, and extrapolation to adults ≥ 60 years should be validated in datasets with DXA assessments at older ages. Secondly, in certain covariate subgroups, such as those based on glucocorticoid use, the small sample sizes led to unstable point estimates and wide confidence intervals. Therefore, the corresponding results should be interpreted with caution, and future studies should include larger sample sizes for validation. Finally, even after controlling for a wide range of possible factors, the study was unable to completely rule out the impact of other confounders such medication usage and substantial comorbidities, which were not taken into consideration during the NHANES study design and might have an impact on how the results are interpreted.
Conclusion
This study demonstrated that increased sedentary time is significantly associated with decreased pelvic BMD, underscoring the importance of controlling sedentary behavior as part of a comprehensive strategy to maintain bone health. Nevertheless, large prospective studies are required to confirm these results.
Acknowledgements
We wish to extend our gratitude to all the participants, medical personnel, and researchers who have contributed to the NHANES study.
Abbreviations
- BMD
Bone mineral
- NHANES
National Health and Nutrition Examination Survey
- CDC
Centers for Disease Control and Prevention
- NCHS
National Center for Health Statistics
- BMI
Body mass index
- PA
Physical activity
- GED
General Educational Development
- PIR
Family income-to-poverty ratio
Authors’ contributions
All authors made significant contributions. Chaoquan Y performed the analyses and wrote the original manuscript. Zhiling H, Yue Q, Wenjun H, Rongyuan L, and Xiajie H involved in preparing and visualization of the results. Yan Chen conceived the study design and revised the manuscript. All authors were involved in the interpretation of the results and approved the final version of the manuscript.
Funding
This study was supported by grants from National Natural Science Foundation of China (82060406, 82360429, and 82260448), Joint Project on Regional High-incidence Diseases Research of Natural Science Foundation of Guangxi (2022JJA141126), Advanced Innovation Teams and Xinghu Scholars Program of Guangxi Medical University, Postdoctoral Science Foundation (2019M650235), and Key R&D Project of Qingxiu District, Nanning, Guangxi (2021003).
Data availability
The NHANES website hosts the datasets created and examined in this study: [https://wwwn.cdc.gov/nchs/nhanes/default.aspx].
Declarations
Ethics approval and consent to participate
The protocols of NHANES were approved by the institutional review board of the National Center for Health Statistics, CDC (https://www.cdc.gov/nchs/nhanes/irba98.htm). NHANES has obtained written informed consent from all participants.
Consent for publication
Consent to publish was given by each NHANES research participant.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Chaoquan Yang, Zhiling Huang and Yue Qiu contributed equally to this work.
References
- 1.Ekelund U, Tarp J, Steene-Johannessen J et al. Dose-response associations between accelerometry measured physical activity and sedentary time and all cause mortality: systematic review and harmonised meta-analysis[J].BMJ. 2019;366:l4570. [DOI] [PMC free article] [PubMed]
- 2.Le Roux E, De Jong NP, Blanc S et al. Physiology of physical inactivity, sedentary behaviours and non-exercise activity: insights from the space bedrest model[J].J Physiol. 2022;600(5):1037–51. [DOI] [PMC free article] [PubMed]
- 3.Nicholas JA, Lo Siou G, Lynch BM, et al. Leisure-time physical activity does not attenuate the association between occupational sedentary behavior and obesity: results from Alberta’s tomorrow project. J Phys Act Health. 2015;12(12):1589–600. [DOI] [PubMed] [Google Scholar]
- 4.Van Der Berg JD, Stehouwer CD, Bosma H et al. Associations of total amount and patterns of sedentary behaviour with type 2 diabetes and the metabolic syndrome: the Maastricht Study[J]. Diabetologia. 2016;59(4):709–18. [DOI] [PMC free article] [PubMed]
- 5.Kim IY, Park S, Chou TH, et al. Prolonged sitting negatively affects the postprandial plasma triglyceride-lowering effect of acute exercise. Am J Physiol Endocrinol Metab. 2016;311(5):E891–8. [DOI] [PubMed] [Google Scholar]
- 6.Chastin SFM, Mandrichenko O, Helbostadt JL, et al. Associations between objectively-measured sedentary behaviour and physical activity with bone mineral density in adults and older adults, the NHANES study. Bone. 2014;64:254–62. [DOI] [PubMed] [Google Scholar]
- 7.Yang H, Li B, Li H, et al. 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. 2024. 10.1016/j.bone.2024.117264. [DOI] [PubMed] [Google Scholar]
- 8.Michalopoulou M, Kambas A, Leontsini D, et al. Physical activity is associated with bone geometry of premenarcheal girls in a dose-dependent manner. Metabolism. 2013;62(12):1811–8. [DOI] [PubMed] [Google Scholar]
- 9.Turner CH. Robling A G.Mechanisms by which exercise improves bone strength[J].J Bone Miner Metab. 2005;3 Suppl: 16–22. [DOI] [PubMed]
- 10.Foong YC, Chherawala N, Aitken D, et al. Accelerometer-determined physical activity, muscle mass, and leg strength in community-dwelling older adults[J]. J Cachexia Sarcopenia Muscle. 2016;7(3):275–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ricci C, Gervasi F, Havemann Nel L, et al. Substitution of sedentary time with light physical activity is related to increased bone density in U.S. women over 50 years old. An iso-temporal substitution analysis based on the National health and Nutrition Examination Survey. Eur J Sport Sci. 2019;19(10):1404–13. [DOI] [PubMed] [Google Scholar]
- 12.Desilva JM, Rosenberg KR. Anatomy, development, and function of the human pelvis. Anat Rec (Hoboken). 2017;300(4):628–32. [DOI] [PubMed] [Google Scholar]
- 13.Rocca Rossetti S. Functional anatomy of pelvic floor[J]. Arch Ital Urol Androl. 201688(1):28–37. [DOI] [PubMed]
- 14.Grimaldi A, Mellor R, Hodges P, et al. Gluteal tendinopathy: a review of mechanisms, assessment and management. Sports Med. 2015;45(8):1107–19. [DOI] [PubMed] [Google Scholar]
- 15.Bo K, Ellstrom Engh M, Hilde G. Regular exercisers have stronger pelvic floor muscles than nonregular exercisers at midpregnancy[J].Am J Obstet Gynecol. 2018;218(4):427 e1-427 e5. [DOI] [PubMed]
- 16.Qin L, Liu W, Cao H, et al. Molecular mechanosensors in osteocytes. Bone Res. 2020;8:23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Iconaru L, Moreau M, Kinnard V et al. Does the prediction accuracy of osteoporotic fractures by BMD and clinical risk factors vary with fracture Site?[J].JBMR Plus. 2019;3(12):e10238. [DOI] [PMC free article] [PubMed]
- 18.Rittweger J. Vibration as an exercise modality: how it may work, and what its potential might be. Eur J Appl Physiol. 2010;108(5):877–904. [DOI] [PubMed] [Google Scholar]
- 19.Smith CT, Barton DW, Piple AS et al. Pelvic fragility fractures: an opportunity to improve the undertreatment of Osteoporosis[J].J bone joint Surg Am. 2021;103(3):213–8. [DOI] [PubMed]
- 20.Srivastava M, Deal C. Osteoporosis in elderly: prevention and treatment. Clin Geriatr Med. 2002;18(3):529–55. [DOI] [PubMed] [Google Scholar]
- 21.Li Y, Di X, Liu M, et al. Association between daily sitting time and kidney stones based on the National health and nutrition examination survey (NHANES) 2007–2016: a cross-sectional study[J]. Int J Surg. 2024;110(8):4624–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Wang J, Li S, Pu H, et al. The association between the non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio and the risk of osteoporosis among U.S. adults: analysis of NHANES data. Lipids Health Dis. 2024;23(1):161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hou W, Chen S, Zhu C, et al. Associations between smoke exposure and osteoporosis or osteopenia in a US NHANES population of elderly individuals. Front Endocrinol (Lausanne). 2023;14:1074574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Afarideh M, Sartori-Valinotti JC, Tollefson MM. Association of sun-protective behaviors with bone mineral density and osteoporotic bone fractures in US adults. JAMA Dermatol. 2021;157(12):1437–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kopiczko A, Czapla M, Juárez-Vela R, et al. Dairy product consumption, eating habits, sedentary behaviour and physical activity association with bone mineral density among adolescent boys: a cross-sectional observational study. BMC Pediatr. 2024. 10.1186/s12887-024-04539-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Weiss M, Yogev R, Dolev E. Occupational sitting and low hip mineral density. Calcif Tissue Int. 1998;62(1):47–50. [DOI] [PubMed] [Google Scholar]
- 27.Skversky AL, Kumar J, Abramowitz MK, et al. Association of glucocorticoid use and low 25-hydroxyvitamin D levels: results from the National Health and Nutrition Examination Survey (NHANES): 2001–2006. J Clin Endocrinol Metab. 2011;96(12):3838–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Du M, Chen L, Li Y et al. Association between sedentary behavior, physical activity, and the triglyceride glucose index and its derived indices: evidence from NHANES 2013–2018[J]. J Exerc Sci Fit. 2025;23 (4):451–61. [DOI] [PMC free article] [PubMed]
- 29.Valente-Dos-Santos J, Tavares OM, Duarte JP et al. Total and regional bone mineral and tissue composition in female adolescent athletes: comparison between volleyball players and swimmers[J].BMC Pediatr. 2018;18(1):212. [DOI] [PMC free article] [PubMed]
- 30.Lin Z, Shi G, Liao X et al. Correlation between sedentary activity, physical activity and bone mineral density and fat in America: National Health and Nutrition Examination Survey, 2011–2018 .Sci Rep. 2023;13(1);10054. [DOI] [PMC free article] [PubMed]
- 31.Whyte EF, Richter C, O’connor S, et al. The effect of high intensity exercise and anticipation on trunk and lower limb biomechanics during a crossover cutting manoeuvre[J]. J Sports Sci. 2018;36(8):889–900. [DOI] [PubMed] [Google Scholar]
- 32.Watson SL, Weeks BK, Weis LJ, et al. High-Intensity resistance and impact training improves bone mineral density and physical function in postmenopausal women with osteopenia and osteoporosis: the LIFTMOR randomized controlled Trial[J]. J Bone Miner Res. 2018;33(2):211–20. [DOI] [PubMed] [Google Scholar]
- 33.Lessi GC, Dos Santos AF, Batista LF, et al. Effects of fatigue on lower limb, pelvis and trunk kinematics and muscle activation: gender differences. J Electromyogr Kinesiol. 2017;32:9–14. [DOI] [PubMed] [Google Scholar]
- 34.Zhang S, Huang X, Zhao X, et al. Effect of exercise on bone mineral density among patients with osteoporosis and osteopenia: A systematic review and network meta-analysis[J]. J Clin Nurs. 2022;31(15–16):2100–11. [DOI] [PubMed] [Google Scholar]
- 35.Naderi A, Degens H, Rezvani MH, et al. A retrospective comparison of physical health in regular recreational table tennis participants and sedentary elderly men[J]. J Musculoskelet Neuronal Interact. 2018;18(2):200–7. [PMC free article] [PubMed] [Google Scholar]
- 36.Dent E, Daly RM, Hoogendijk EO, et al. Exercise to prevent and manage frailty and fragility Fractures[J]. Curr Osteoporos Rep. 2023;21(2):205–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Kemmler W, Shojaa M, Kohl M et al. Effects of different types of exercise on bone mineral density in postmenopausal women: A systematic review and Meta-analysis[J]. Calcif Tissue Int. 2020;107(5):409–39. [DOI] [PMC free article] [PubMed]
- 38.Dutto DJ, Liang MTC, Braun WA et al. Effect of 6-month high-impact step aerobics and resistance training on BMD and tibial bending strength in sedentary premenopausal women[J]. Osteoporos Int. 2022;33(3):695–701. [DOI] [PubMed]
- 39.Zerwekh JE, Ruml LA, Gottschalk F, et al. The effects of twelve weeks of bed rest on bone histology, biochemical markers of bone turnover, and calcium homeostasis in eleven normal subjects[J]. J Bone Min Res. 1998;13(10):1594–601. [DOI] [PubMed] [Google Scholar]
- 40.Iwaniec UT, Turner RT. .Influence of body weight on bone mass, architecture and turnover[J].J Endocrinol. 2016;230(3):R115–30. [DOI] [PMC free article] [PubMed]
- 41.Compston J. Glucocorticoid-induced osteoporosis: an update[J].Endocrine. 2018;61(1):7–16. [DOI] [PMC free article] [PubMed]
- 42.Buckley L, Guyatt G, Fink HA et al. 2017 American college of rheumatology guideline for the prevention and treatment of Glucocorticoid-Induced Osteoporosis[J]. Arthritis care res (Hoboken). 2017;69(8):1095–110. [DOI] [PubMed]
- 43.Weinstein RS. .Glucocorticoid-induced osteoporosis and osteonecrosis[J].Endocrinol metab clin North Am. 2012;41(3):595–611. [DOI] [PMC free article] [PubMed]
- 44.Buchman-Pearle JM, Pinto BL, Barrett JM, et al. Lower limb postures resembling sitting and standing alter lumbar angles along the passive stiffness curve[J]. J Electromyogr Kinesiol. 2023;69:102752. [DOI] [PubMed] [Google Scholar]
- 45.Roussouly P, Pinheiro-Franco JL. Biomechanical analysis of the spino-pelvic organization and adaptation in pathology[J]. Eur spine J. 2011;20 suppl 5 (Suppl 5):609–18. [DOI] [PMC free article] [PubMed]
- 46.Pelegrini A, Klen JA, Costa AM, et al. Association between sedentary behavior and bone mass in adolescents. Osteoporos Int. 2020;31(9):1733–40. [DOI] [PubMed] [Google Scholar]
- 47.Mcmichan L, Dick M, Skelton DA et al. Sedentary behaviour and bone health in older adults: a systematic review[J]. Osteoporos Int 2021;32(8):1487–97. [DOI] [PubMed]
- 48.Gobbo LA, Júdice PB, Hetherington-Rauth M, et al. Sedentary patterns are associated with bone mineral density and physical function in older adults: Cross-Sectional and prospective Data[J]. Int J Environ Res Public Health. 2020;17:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Bruyere O, Scott D, Papaioannou A, et al. The impact of sedentary behavior and physical activity on bone health: A narrative review from the rehabilitation working group of the international osteoporosis Foundation[J]. Calcif Tissue Int. 2025;116(1):109. [DOI] [PubMed] [Google Scholar]
- 50.J C. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. [Google Scholar]
- 51.Johnell O, Kanis JA, Oden A, et al. Predictive value of BMD for hip and other fractures[J]. J Bone Min Res. 2005;20(7):1185–94. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The NHANES website hosts the datasets created and examined in this study: [https://wwwn.cdc.gov/nchs/nhanes/default.aspx].