Cross-Sectional Associations of Serum Zinc with Bone Mineral Density in the NHANES 2011-2014 Cohort

Osteoporosis represents a major public health concern worldwide, affecting an estimated 18.3% of the global population^1,2. Epidemiological data indicate that this condition does not impact all demographic groups equally, with women experiencing a markedly higher burden compared to men, particularly after menopause due to hormonal changes that accelerate bone loss^1,2. Bone mineral density (BMD) serves as a fundamental indicator of skeletal health and is widely recognized as a crucial determinant of bone strength³. Clinically, it is considered one of the most reliable predictors of an individual's susceptibility to fractures³. Individuals with higher BMD generally possess stronger, more resilient bones, which confer protection against bone fragility and the occurrence of fractures^4,5. Conversely, reduced BMD is associated with compromised bone integrity, placing individuals at elevated risk for conditions such as osteopenia and osteoporosis, as well as fracture-related morbidity^4,5.

BMD decline is a multifactorial and biologically intricate process influenced by a constellation of factors, including aging, hormonal milieu, lifestyle behaviors, nutritional status, and underlying pathological conditions⁶. With advancing age, the rate of osteoblastic bone formation progressively diminishes, whereas osteoclastic bone resorption becomes relatively enhanced, resulting in a net loss of bone mass over time⁷. Among the biological determinants, hormonal alterations, particularly the marked reduction in circulating estrogen following menopause, constitute one of the principal drivers of rapid bone demineralization⁸. Nutritional inadequacies, such as insufficient intake of calcium, vitamin D, protein, and essential trace elements like magnesium and zinc, further compromise skeletal integrity by impairing bone mineralization and matrix formation^3,8. Lifestyle-related factors, including physical inactivity, cigarette smoking, excessive alcohol consumption, and chronic psychosocial stress, exacerbate bone resorption and accelerate bone loss⁹. Additionally, prolonged administration of glucocorticoids, anticonvulsants, or aromatase inhibitors, as well as chronic systemic disorders such as diabetes mellitus, renal insufficiency, and gastrointestinal diseases, can perturb calcium homeostasis and disrupt normal bone remodeling, collectively contributing to the progressive reduction in BMD and heightened susceptibility to osteoporosis³.

Zinc, as a vital trace element, is integral to the regulation of bone metabolism and the preservation of skeletal integrity¹⁰. Accumulating evidence indicates a positive correlation between zinc status and BMD, although the magnitude and consistency of this relationship are influenced by baseline zinc levels, existing bone health, and demographic characteristics of the study population¹¹. From a mechanistic perspective, zinc facilitates osteogenesis by enhancing osteoblast differentiation, promoting collagen matrix synthesis, and activating critical enzymes such as alkaline phosphatase, which is essential for proper bone mineralization¹⁰. Concurrently, zinc exerts inhibitory effects on osteoclast-mediated bone resorption, thereby contributing to the homeostatic balance between bone formation and degradation^10,12. Observational and interventional studies have shown that higher circulating zinc concentrations are generally associated with increased BMD, whereas zinc insufficiency correlates with elevated osteoporosis risk¹².

In the present study, we leveraged data from the National Health and Nutrition Examination Survey (NHANES) 2011-2014, which provides a large, nationally representative cohort of U.S. adults with comprehensive biochemical, demographic, and lifestyle data. Compared with previous investigations, this approach offers several methodological advantages, including an expanded sample size that enhances statistical power and the ability to detect subtle associations, as well as the implementation of threshold modeling to explore potential nonlinear relationships between serum zinc concentrations and BMD. By systematically examining zinc-BMD associations across different quartiles and adjusting for relevant confounders, our study provides a more nuanced understanding of how zinc status may influence skeletal health. Importantly, the findings derived from this nationally representative dataset can inform clinical decision-making and public health strategies, enabling clinicians and researchers to better assess the potential role of zinc in bone health maintenance and osteoporosis prevention.

The National Center for Health Statistics (NCHS) Research Ethics Review Board (ERB) approved the NHANES data collection protocols. Since the NHANES datasets are publicly accessible, researchers are exempt from obtaining additional approval from their own institutional review boards (IRBs).

Study design and characteristics of the study population

In the present cross-sectional investigation, we sought to examine the potential independent association between serum zinc concentrations and BMD utilizing data derived from the 2011-2014 cycles of the NHANES. NHANES, which is administered by the National Center for Health Statistics (NCHS), employs a carefully designed stratified, multistage probability sampling methodology to generate a nationally representative dataset of the civilian, non-institutionalized population of the United States. Each survey cycle systematically collects extensive information encompassing demographic characteristics, lifestyle factors, dietary intake, clinical parameters, and anthropometric measurements through a combination of standardized interviews, structured questionnaires, and physical examinations. Comprehensive protocols and methodological details for each survey component are publicly accessible on the official NHANES website. For the purposes of this analysis, we focused on adults aged 20 years and older and excluded individuals with missing or incomplete data on key anthropometric measures or other essential covariates required for the study models. The process of participant selection, along with specific inclusion and exclusion criteria, is visually summarized in Figure 1 to provide a clear overview of the analytic sample.

Assessment indicators

Assessment of Anthropometric Measures

Trained and certified health technicians performed the collection of anthropometric measurements, including body weight and standing height, at dedicated NHANES mobile examination centers, following rigorously standardized protocols established by the NCHS. Body weight was recorded to the nearest 0.1 kg using a calibrated digital scale, ensuring consistency and accuracy across all measurements. Standing height was measured to the nearest 0.1 centimeter employing a fixed stadiometer, with participants instructed to maintain an upright posture and heels together for precision. Subsequently, body mass index (BMI), a widely recognized indicator of general adiposity, was computed for each participant by dividing the measured weight in kilograms by the square of the height in meters (kg/m²).

Zinc status in serum

Venous blood specimens were obtained from study participants at the designated mobile examination centers (MECs) by certified phlebotomists following standardized NHANES procedures. Collected samples were promptly processed and transported under controlled conditions by trained personnel to the National Center for Environmental Health (NCEH) laboratory for subsequent biochemical analysis. After initial centrifugation, serum aliquots were kept at °C during transit to preserve sample integrity and were then stored at ultra-low temperatures of either -20 °C or -70 °C until the time of analysis. Quantification of serum zinc concentrations was performed using inductively coupled plasma dynamic reaction cell mass spectrometry (ICP-DRC-MS), a highly sensitive and precise analytical method. Comprehensive details regarding the laboratory protocols, quality control measures, and analytical specifications for serum zinc assessment are documented in the NHANES laboratory data file (LBDZN), which can be accessed on the official NHANES website, with further methodological description provided at: https://wwwn.cdc.gov/nchs/nhanes/2011-2012/PBCD_G.htm.

BMD

All participants who met the inclusion criteria for this study underwent dual-energy X-ray absorptiometry (DXA) scans to obtain precise measurements of BMD. The scans were conducted by certified radiologic technologists using a fan-beam densitometer, with data acquisition and analysis performed through Hologic APEX software (version 4.0) to ensure accuracy and consistency. The standardized DXA examination protocols, including participant positioning, calibration procedures, and quality control measures, are fully detailed on the official NHANES website, with the corresponding link provided in the Table of Materials. Certain individuals were excluded from undergoing DXA assessment based on specific safety and technical criteria, including pregnancy, administration of contrast agents within the preceding seven days, body weight exceeding 450 pounds, multiple prior hip replacements, bilateral hip fractures, or the presence of hip prostheses, in order to maintain the validity and reliability of BMD measurements.

Covariates

To rigorously assess the association between serum zinc concentrations and BMD, it was essential to account for potential confounding variables and the complex interactions among multiple covariates. Demographic factors included participants' age (≥20 years), sex (male or female), race/ethnicity (categorized as Mexican American, non-Hispanic White, non-Hispanic Black, other Hispanic, and other races), educational level (less than high school, high school or equivalent, and above high school), and the ratio of family income to the federal poverty threshold. Anthropometric parameters, including BMI, were collected during standardized physical examinations. Biochemical measures, specifically serum total calcium and serum phosphorus concentrations (both in mmol/L), were also incorporated as key covariates due to their critical roles in bone metabolism and mineralization. In addition, lifestyle and health-related characteristics were assessed through structured questionnaires, capturing self-reported participation in vigorous physical activity (yes/no), history of diabetes (yes/no), hypertension status (yes/no), and alcohol consumption. Alcohol intake was quantified as the average number of alcoholic beverages consumed per day over the preceding 12 months¹³. Participants' smoking status was determined based on their cumulative lifetime cigarette consumption. Individuals who had smoked fewer than 100 cigarettes throughout their lifetime were classified as non-smokers, whereas those who reported having smoked 100 or more cigarettes were categorized as smoker¹⁴.

Statistical analysis

All statistical analyses for this study were performed using EmpowerStats software (version 4.1). Analyses were conducted using the NHANES complex survey design, accounting for sampling weights, stratification, and clustering to produce nationally representative estimates. Specifically, the 4-year fasting subsample weight (WTSAF4YR), stratification variable (SDMVSTRA), and primary sampling unit (SDMVPSU) were applied in all analyses. The demographic, clinical, and biochemical characteristics of the study population were first summarized and compared across quartiles of serum zinc concentrations to provide an overview of population distributions. For continuous variables, group differences were assessed using linear regression models, whereas categorical variables were compared using chi-square tests to identify statistically significant variations between quartiles.

To evaluate the relationship between serum zinc levels and BMD, multivariable linear regression models were employed. Three progressively adjusted models were constructed to account for potential confounding factors. Model 1 represented the unadjusted association. Model 2 controlled for basic demographic covariates, including age, sex, and race/ethnicity. Model 3 further incorporated additional potential confounders such as educational attainment, diabetes status, engagement in vigorous physical activity, smoking status, hypertension, serum total calcium, serum phosphorus, and the ratio of family income to the federal poverty threshold. Furthermore, to investigate potential nonlinear relationships between serum zinc concentrations and BMD, smooth curve fitting analyses were conducted. All statistical tests were two-sided, and a p-value of less than 0.05 was considered indicative of statistical significance.

Descriptive characteristics and baseline profile of the participants included in the study

Since serum zinc concentrations were measured in a fasting subsample, the corresponding 4-year fasting subsample weight (WTSAF4YR) was applied in accordance with NHANES analytic guidelines. Because data from two survey cycles (2011-2012 and 2013-2014) were combined, the 4-year weight was derived by dividing each participant's 2-year fasting subsample weight (WTSAF2YR) by two, following the formula WTSAF4YR = ½ × WTSAF2YR. Initially, a total of 19,931 participants from the 2011-2014 cycles of the NHANES were considered for inclusion in the present analysis. Individuals under the age of 20 years (n = 8,242) were first excluded to focus on the adult population, followed by the removal of participants with missing data on key variables, including serum zinc concentrations and BMD measurements (n = 9,814). This filtering process yielded a subset of 1,875 adults aged 20 years or older who had complete data for the primary variables of interest. To ensure the integrity of anthropometric measurements, participants lacking data on essential parameters such as height, weight, or BMI (n = 201) were subsequently excluded. Following these sequential exclusions, the final analytical cohort comprised 1,674 individuals with comprehensive and complete data across all variables required for the study.

Table 1 summarizes the baseline characteristics of the study population stratified by serum zinc quartiles. The quartile cut-off values were automatically calculated according to the 25th, 50th, and 75th percentiles of serum zinc levels. A total of 1,674 participants from NHANES 2011-2014 were included, with a mean age of 39.5 ± 3.7 years, showing no significant variation across serum zinc quartiles (p = 0.198). Gender distribution differed significantly (p < 0.001), with men more prevalent in the highest quartile (64.9%) than in the lowest (41.1%). Racial/ethnic composition also varied across quartiles (p = 0.003), while education level, vigorous physical activity, alcohol intake, diabetes, hypertension, BMI, and poverty-income ratio did not differ significantly. Serum total calcium showed a modest but significant increase across quartiles (p = 0.044), whereas serum phosphorus remained comparable (p = 0.061). Notably, higher serum zinc concentrations were associated with slightly greater BMD at the lumbar spine, pelvis, and total body (p = 0.030-0.047), indicating a trend toward improved bone density with increased zinc status.

Association between serum zinc levels and total bone mineral density

Building on the descriptive characteristics of the study population, we next examined the relationship between serum zinc concentrations and total BMD (Table 2). In unadjusted analyses (Model 1), higher serum zinc levels were significantly associated with greater total BMD (β = 0.002, 95% CI: 0.001-0.004, p = 0.018). This positive association persisted after adjusting for demographic factors including age, sex, and race/ethnicity in Model 2 (β = 0.003, 95% CI: 0.002-0.005, p = 0.001), and remained statistically significant in Model 3, which additionally accounted for education level, diabetes status, vigorous physical activity, smoking status, hypertension, serum total calcium, and family income-to-poverty ratio (β = 0.002, 95% CI: 0.001-0.004, p = 0.013).

Further analyses stratified participants by serum zinc quartiles, revealing that those in the highest quartile (Q4) consistently exhibited significantly higher total BMD compared to the lowest quartile (Q1) across all models, with β coefficients ranging from 0.018 to 0.023 (all p < 0.001). A clear linear trend was observed across quartiles (p for trend < 0.001), indicating a dose-response relationship. Moderate increases in BMD were also noted in quartiles 2 and 3, although some associations lost statistical significance in the fully adjusted model, suggesting that the effect of zinc may be more pronounced at higher serum concentrations.

Subgroup analyses by sex indicated a significant positive association between serum zinc and total BMD in men (Model 3: β = 0.004, 95% CI: 0.002-0.007, p = 0.055), whereas no significant relationship was observed in women after full adjustment (Model 3: β = 0.001, 95% CI: -0.002-0.004, p = 0.224). The interaction test confirmed that the association differed significantly by sex (p for interaction < 0.001), suggesting that the skeletal benefits of higher zinc levels may be more prominent in men than in women.

Site-specific associations between serum zinc and bone mineral density

In addition to examining total BMD, we investigated the relationship between serum zinc concentrations and site-specific bone mineral density, including the lumbar spine and pelvis (Table 3). In unadjusted analyses (Model 1), higher serum zinc levels were positively associated with lumbar spine BMD (β = 0.001, 95% CI: -0.001-0.003, p = 0.013) and pelvis BMD (β = 0.003, 95% CI: 0.001-0.004, p = 0.046). After adjusting for age, sex, and race/ethnicity in Model 2, the association with lumbar spine BMD remained significant (β = 0.002, 95% CI: 0.001-0.003, p = 0.016), whereas the relationship with pelvis BMD was attenuated and no longer statistically significant (β = 0.005, 95% CI: 0.003-0.007, p = 0.068). In the fully adjusted Model 3, which additionally accounted for education level, diabetes status, vigorous physical activity, smoking status, hypertension, serum total calcium, and family income-to-poverty ratio, serum zinc remained significantly associated with lumbar spine BMD (β = 0.002, 95% CI: 0.001-0.004, p = 0.025), while the association with pelvis BMD was further attenuated and non-significant (β = 0.003, 95% CI: 0.002-0.005, p = 0.118). These results suggest that higher serum zinc concentrations may have a more pronounced influence on trabecular-rich sites such as the lumbar spine compared to cortical-rich sites like the pelvis.

Analysis of nonlinear threshold effects

Smooth curve fitting was performed using a generalized additive model (GAM) with penalized spline functions to explore potential nonlinear associations between serum zinc and bone mineral density. When nonlinearity was detected, a two-piecewise linear regression model was further applied to identify the inflection point (threshold). The fitted smooth curve with 95% confidence intervals (Figure 2, Supplementary Table 1) demonstrated a nonlinear association between serum zinc concentrations and total BMD. As serum zinc levels increased, BMD initially showed a modest upward trend, with a noticeable inflection point around 13.20 umol/L. Beyond this threshold, the positive association between serum zinc and BMD became more pronounced and linear. This finding is consistent with the results obtained from the multivariable linear regression analyses presented in Table 2, which demonstrate that serum zinc concentrations, when categorized into quartiles, exhibit a statistically significant association with bone mineral density even after controlling for a comprehensive set of potential confounding factors.

Data availability

This study made use of data from the publicly available National Health and Nutrition Examination Survey (NHANES) database. These datasets are freely accessible to all individuals and organizations. The data underlying the results of this research can be accessed via the NHANES repository at https://www.cdc.gov/nchs/nhanes/.

Figure 1: Flow diagram depicting the selection and inclusion process of study participants. Please click here to view a larger version of this figure.

Figure 2: Generalized additive model-based smooth curve demonstrating the association between serum zinc concentration and bone mineral density (BMD). The solid red line represents the fitted relationship, while the blue shaded area indicates the 95% confidence interval derived from the model. Please click here to view a larger version of this figure.

Table 1: Distribution of demographic and clinical characteristics according to serum zinc quartiles among NHANES 2011-2014 participants. Please click here to download this Table.

Table 2: Relationship between serum zinc levels and their quartile distribution with total BMD in the study population. Please click here to download this Table.

Table 3: Association of serum zinc levels with BMD in the study population. Please click here to download this Table.

Supplementary Table 1: Model statistics for generalized additive models. Please click here to download this File.

In this cross-sectional study of 1,674 adults from NHANES 2011-2014, higher serum zinc concentrations were positively associated with total and lumbar spine BMD, with the strongest effects observed in the highest zinc quartile. Multivariable analyses confirmed these associations after adjusting for demographic, lifestyle, anthropometric, and biochemical covariates, and subgroup analyses suggested a more pronounced effect in men. Site-specific results indicated that trabecular-rich sites like the lumbar spine were more responsive to zinc than cortical-rich sites such as the pelvis. Threshold effect modeling further revealed a nonlinear relationship, with an inflection point around 13.20 µmol/L, beyond which increases in zinc were linked to greater BMD gains.

Several critical analytical decisions underpinned the robustness of our findings. First, the choice of multivariable linear regression models allowed us to account for a comprehensive set of potential confounders, including demographic, lifestyle, anthropometric, and biochemical factors, thereby isolating the association between serum zinc and BMD. Additionally, stratifying analyses by serum zinc quartiles and performing subgroup analyses by sex enabled the identification of potential dose-response relationships and sex-specific effects. A threshold effect analysis using smooth curve fitting was implemented to detect nonlinear associations and to estimate an inflection point around 13.20 µmol/L, providing insight into potential concentration-dependent effects of zinc on bone density.

From a methodological perspective, several key measures were implemented to ensure the robustness of the analysis. First, participants with missing data on primary variables, including serum zinc, BMD, and anthropometric measurements, were excluded to maintain data completeness. Second, continuous variables such as BMI and serum biochemical parameters were standardized to minimize variability and facilitate comparison across different models. Furthermore, multivariable-adjusted regression models were employed to account for potential confounding factors and to evaluate the association between serum zinc and BMD.

A threshold effect model was also applied to examine the potential nonlinear relationship between serum zinc and BMD. The analysis indicated that when serum zinc concentrations were below approximately 13.20 µmol/L, total BMD increased gradually with rising zinc levels, with relatively modest increments per 1 µmol/L increase in serum zinc. However, once serum zinc reached or exceeded 13.20 µmol/L, the increase in BMD became more pronounced and approximately linear, suggesting a stronger positive effect of zinc on bone density at higher concentrations. Identifying this inflection point not only quantifies a key transition in the zinc-BMD relationship but also provides a concrete explanation for why individuals in the highest zinc quartile (Q4) exhibited higher BMD, thereby reinforcing the observed positive association between elevated serum zinc levels and improved bone density.

Zinc has long been recognized as a pivotal trace element in maintaining skeletal integrity and preventing osteoporosis, primarily through its dual action of promoting osteoblast activity while inhibiting osteoclast-mediated bone resorption^15,16,17,18. Our findings of a positive association between serum zinc levels and BMD, particularly the enhanced effect observed above a threshold of 13.20 µmol/L, provide novel insights into the concentration-dependent nature of zinc's skeletal benefits. Mechanistically, this threshold may correspond to the activation levels of zinc-dependent enzymes crucial for bone metabolism, such as alkaline phosphatase and collagen synthase, which facilitate osteoblast proliferation, collagen matrix synthesis, and subsequent mineralization. Below this threshold, enzyme activity and osteogenic signaling may be suboptimal, limiting the impact of zinc on bone formation. Once serum zinc surpasses approximately 13.20 µmol/L, these enzymatic processes may achieve near-maximal activity, thereby amplifying the positive association with BMD.

Importantly, the identified 13.20 µmol/L threshold lies within the upper-middle segment of the established normal reference range for adult serum zinc in the U.S. (approximately 10.7-17.7 µmol/L). This suggests that optimal osteogenic effects may manifest when serum zinc reaches mid-to-high normal concentrations, aligning with physiological thresholds required for zinc transporter activation and sufficient intracellular zinc availability to support osteoblast proliferation and differentiation. Thus, the nonlinear pattern observed in our threshold analysis not only reflects a statistical inflection point but also corresponds plausibly to the biological requirements for zinc's skeletal functions.

Overall, these observations highlight that maintaining serum zinc at or above the 13.20 µmol/L threshold may be particularly important for maximizing bone formation and preserving BMD, offering a mechanistic context for both clinical assessment and potential nutritional interventions aimed at fracture prevention. Future studies should further investigate whether this threshold aligns with individual variability in zinc transporter expression, enzymatic responsiveness, and age- or sex-specific skeletal demands.

Utilizing the nationally representative NHANES 2011-2014 dataset, this study advances previous research by uncovering a novel threshold effect in the association between serum zinc concentrations and BMD. Our analyses indicate that BMD exhibits a more pronounced increase when serum zinc surpasses 13.20 µmol/L, situated within the upper-middle segment of the normal reference range for U.S. adults (10.7-17.7 µmol/L). This threshold may reflect the activation of zinc-dependent enzymes, such as alkaline phosphatase and collagen synthase, as well as the zinc concentration required to support osteoblast proliferation, providing a mechanistic basis for the observed nonlinear relationship. From a clinical perspective, these findings imply that maintaining serum zinc above this threshold may promote optimal skeletal health, underscoring its potential utility in osteoporosis risk assessment and targeted nutritional interventions. By integrating a large, nationally representative sample with threshold modeling, the present study offers novel insights that extend beyond the previously established positive association between zinc and BMD, with implications for both individualized patient management and broader public health strategies.

First, the cross-sectional design precludes causal inference and cannot establish the temporal direction of the association; reverse causation and residual confounding remain possible. Second, serum zinc concentrations reflect short-term zinc status and may be influenced by acute-phase responses or other transient factors, while dietary zinc intake was not analyzed and could provide complementary context. Third, key confounding factors such as serum vitamin D and other nutrients involved in bone metabolism were not included, which may affect the estimated independent association between zinc and BMD. Fourth, although NHANES provides comprehensive data, certain high-risk subpopulations, including institutionalized elderly individuals, may be underrepresented. Fifth, the study did not explore the mechanistic basis of the identified 13.2 µmol/L threshold or perform subgroup analyses stratified by age (e.g., <50 years vs. ≥50 years) or menopausal status in women. Consequently, it remains unclear whether this threshold is applicable to older adults or postmenopausal women, limiting the generalizability of the threshold finding across different demographic groups. Future prospective and interventional studies are needed to validate these results, clarify underlying mechanisms, and assess the clinical relevance of threshold-specific zinc interventions in diverse populations.

In summary, this large, population-based analysis provides robust evidence that higher serum zinc levels are associated with greater bone mineral density in U.S. adults. The positive dose-response relationship and threshold effect suggest that maintaining adequate zinc status may be important for bone health and osteoporosis prevention.