Skip to main content
Download PDF

Association between the Triglyceride-glucose index and fragility fractures among US adults: insights from NHANES

Diabetology & Metabolic Syndrome volume 17, Article number: 96 (2025) Cite this article

Abstract

Background

The triglyceride-glucose (TyG) index, a recognized marker for insulin resistance, holds potential implications for skeletal health. However, its relationship with fragility fractures remains uncertain. We aimed to elucidate the association between the TyG index and fragility fractures in the general US population.

Methods

Cross-sectional data of 25,082 participants were obtained from the National Health and Nutrition Examination Survey. The association between the TyG index and fragility fractures was investigated using univariate and weighted multivariate logistic regression as well as restricted cubic spline (RCS) regression models. The least absolute shrinkage and selection operator regression with ten-fold cross-validation was employed to identify key variables, leading to the development of a nomogram model. Calibration and receiver operating characteristic curves were utilized to evaluate the model's validity.

Results

The overall prevalence of fragility fractures among participants was 1.10%. After adjusting for confounders, the TyG index exhibited a robust association with the risk of fragility fractures (odds ratio, 1.94; 95% confidence interval, 1.31–2.88; P < 0.001). RCS regression demonstrated a positive linear relationship between the TyG index and fragility fractures. The predictive nomogram, incorporating the TyG index and other clinical factors, demonstrated favorable predictive performance (consistency index = 0.901).

Conclusions

Elevated TyG index levels were significantly correlated with the risk of fragility fractures in the general US population. These findings suggest that the TyG index may serve as a predictive marker for fragility fractures, underscoring the importance of early intervention and improved fracture risk assessment tools in clinical practice.

Background

Fragility fractures, which result from forces typically insufficient to fracture healthy bone, were defined by the World Health Organization in 2013 as fractures caused by forces equivalent to a fall from standing height or lower [1, 2]. These fractures represent one of the most severe consequences of osteoporosis, contributing significantly to mortality and long-term disability globally [3]. The International Osteoporosis Foundation estimated that 50% of women and 20% of men aged ≥ 50 years are at risk of experiencing fragility fractures [4]. Additionally, approximately 50% of fragility fracture patients are prone to experiencing subsequent fractures, with the risk increasing exponentially [5]. This condition severely impacts the quality of life and imposes substantial burdens on families and societies.

Although previous research has identified low bone mineral density (BMD) as a risk factor for fractures, emerging evidence suggests that a significant proportion of fractures occur in individuals with normal or mildly reduced bone density, rather than in those with osteoporosis [6,7,8]. Therefore, a deeper understanding of associated risk factors is necessary for effective prevention and management. Moreover, although most fragility fractures occur in the elderly population, recent trends indicate a shift toward younger groups. This suggests the presence of potential modifiable risk factors, such as diet, physical activity, and other lifestyle factors, including smoking, alcohol consumption, medication use, and hormonal influences, which may contribute to future fracture risk [9].

The mechanisms underlying increased skeletal fragility are multifaceted, encompassing aging, inflammation, oxidative stress, obesity, and insulin resistance (IR), all contributing to impaired bone homeostasis and pathological changes in bone structure [10,11,12]. In recent years, the triglyceride-glucose (TyG) index has emerged as a simple, reliable, highly sensitive, and specific biochemical marker for identifying IR in the general population [13, 14]. Numerous studies have confirmed that the TyG index is closely associated with diabetes, hypertension, atherosclerotic cardiovascular disease, ischemic stroke, and other diseases, whereas limited research has been performed on the correlation with bone health [15,16,17]. Furthermore, existing studies have focused solely on the relationship between the TyG index and bone density, yielding heterogeneous conclusions and lacking detailed investigations into whether monitoring the TyG index could be applied to assess the risk of fragility fractures [18,19,20].

This study aims to investigate the association between the TyG index and the risk of fragility fractures in the general population using cross-sectional data from the National Health and Nutrition Examination Survey (NHANES) and to investigate the construction of a risk prediction model, addressing a critical gap in the existing literature. By providing a more nuanced understanding of the risk factors influencing the occurrence of fragility fractures, this study seeks to contribute to the development of more effective and targeted prevention measures.

Methods

Study population

The NHANES is an ongoing program conducted by the National Center for Health Statistics of the Centers for Disease Control and Prevention of the United States[21]. It aims to assess the nutritional and physical status of Americans through personal structured interviews at home, health examinations at a mobile examination center, and laboratory specimen analyses. NHANES data is publicly accessible and freely downloadable from: https://www.cdc.gov/nchs/nhanes/index.htm.

We analyzed data from eight NHANES cycles (1999–2010, 2013–2014, and 2017–2018). Initially, 81,589 participants were included in this study, of whom 34,241 individuals aged under 18 years were excluded. Additionally, participants were excluded who: (1) lacked baseline TyG index data; (2) lacked osteoporosis questionnaire data; (3) had any cancer at baseline; (4) were pregnant; (5) had undergone hysterectomy or bilateral oophorectomy; or (6) had a history of hormone or glucocorticoid use. Ultimately, 25,082 participants were included in this study (Fig. 1).

Fig. 1
figure 1

Flow chart of participant recruitment

Full size image

Assessment of the TyG index

The TyG index was calculated as follows: TyG index = Ln [fasting triglyceride (mg/dL) × fasting glucose (mg/dL)/2] [13, 21]. Concurrently, two further IR indicators were calculated [22, 23]: (1) Homeostasis model assessment of insulin resistance (HOMA-IR) index = fasting glucose (mmol/L) × fasting insulin (µU/mL)/22.5; (2) Quantitative insulin sensitivity check (QUICKI) index = 1/[log (fasting insulin, µU/mL) + log (fasting plasma glucose, mg/dL)].

Assessment of fragility fracture

The primary outcome was the occurrence of fragility fractures, defined by self-reported previous diagnosis by a physician using the Computer-Assisted Personal Interview system. Participants who answered “yes” to the question: “Have you ever been told by a doctor that you had a fracture?” were considered to have had fractures. Participants reporting fractures were further queried about sites, times, and reasons for fractures. As reasons, participants could select: (1) A fall from standing height or less, for example, tripped, slipped, or fell out of bed; (2) A hard fall, such as falling off a ladder or step stool or down stairs; (3) A car accident or other severe trauma. Participants who selected reason (1) were considered to have low-trauma bone fractures and were included in the fragility fracture group. Participants who selected other reasons or reported no fractures were included in the non-fragility fracture group [1, 24].

Assessment of covariates

We collected as many covariates with known confounding factors for fragility fractures as possible. Standardized questionnaires obtained information on age, sex, race, educational level, alcohol consumption, smoking status, physical activity, presence of diabetes, hypertension, hyperlipidemia, coronary artery disease, chronic kidney disease, and drug use (including insulin, oral glucose-lowering drugs, antihypertensive drugs, and lipid-lowering drugs). Race was categorized into five groups: Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, and other races. Educational level was classified as below high school, high school, and above high school. Alcohol consumption and smoking status were dichotomously classified as current use or not. Participants were classified into three groups based on their physical activity levels: inactive (no moderate or vigorous physical activity), moderate (engaging in 1–149 min of moderate to vigorous physical activity per week), and vigorous (engaging in at least 150 min of moderate to vigorous physical activity per week), following the Centers for Disease Control and Prevention Physical Activity Guidelines for Americans. Diabetes mellitus was defined as a self-reported status of diabetes mellitus diagnosis, current use of hypoglycemic therapy, glycosylated hemoglobin (HbA1c) level ≥ 6.5%, or fasting glucose level ≥ 126 mg/dl (7.0 mmol/L). Hypertension was defined as having a self-reported history of hypertension, use of antihypertensive drugs, a systolic blood pressure (SBP) ≥ 140 mmHg, or a diastolic blood pressure (DBP) ≥ 90 mmHg. Hyperlipidemia was defined as the current use of lipid-lowering medication or self-reported history of dyslipidemia. Coronary artery disease was defined as self-reported coronary heart disease or myocardial infarction. Chronic kidney disease was defined as an estimated glomerular filtration rate < 60 ml/(1.73 m2/min) or a self-reported history of weak/failing kidneys. All participants were measured for height, waist circumference, weight, and SBP/DBP by trained examiners at the mobile examination center, following standardized NHANES protocols. Blood pressure was measured three times using a calibrated sphygmomanometer, and the average of the three readings was recorded for analysis. The body mass index (BMI) was calculated as weight in kilograms (kg) divided by the square of height in meters (m2). Strict laboratory analyses were performed, including the assessment of fasting glucose, insulin, HbA1c, total cholesterol, triglycerides, low density lipoprotein cholesterol, high density lipoprotein cholesterol, serum albumin, alkaline phosphatase, phosphorus, total calcium, and 25-hydroxyvitamin D [25(OH)D] at baseline. Further details of these measurements were documented in the NHANES Laboratory Medical Technologists Procedures Manual. BMD was assessed using dual-energy X-ray absorption images obtained using Apex 3.2 software and the Hologic Discovery model A densitometer (Hologic, Inc., Bedford, Massachusetts, USA). All measurements were conducted by NHANES radiological technologists with extensive training and certification.

Statistical analysis

All patients were divided into the non-fragility and fragility fracture groups. Baseline characteristics were reported as means and standard deviations (SDs) or median (25th–75th percentile) for continuous variables and n (%) for categorical variables, and differences among groups were compared using the chi-square test, Wilcoxon-rank test, or Kruskal–Wallis test where appropriate. To control the confounding factors, an inverse propensity of treatment weighting (IPTW) was applied, allowing a pseudo-population to be created by assigning individuals with weights that corresponded to the inverse of their probability of receiving treatment dependent on observed covariates. The differences in the prevalence of covariates between statin users and nonusers were evaluated by the standardized mean difference (SMD). SMDs of at least 0.300 were considered to be indicative of a relevant between-group imbalance [25, 26]. Correlations of the TyG index with other insulin resistance indexes were assessed using the Spearman correlation test.

To evaluate the association between the TyG index and fragility fracture events, the TyG index was analyzed as both a continuous and a categorical variable. Univariate and a variety of multivariate logistic regression models were built to estimate the odds ratio (OR) and 95% confidence interval (CI) for the associations. IPTW-weighted multivariate logistic regression models were established to further control confounders. Model 1 did not adjust for any covariate. Model 2 was adjusted for age, sex, and educational level; Model 3 was further adjusted for BMI, alcohol consumption, SBP, DBP, HbA1c, albumin, alkaline phosphatase, 25(OH)D, diabetes, hypertension, hyperlipidemia, coronary artery disease, chronic kidney disease, and medication treatments. In addition, the restricted cubic spline (RCS) regression model with five knots was applied to investigate the potential nonlinear association between the TyG index and fragility fracture events. To screen key fragility fracture-related factors and eliminate the collinearity among different variables, the least absolute shrinkage and selection operator (LASSO) regression algorithm was applied. In this, prediction error was minimized for a quantitative response variable by imposing a constraint on the model parameters that caused regression coefficients to shrink toward zero. When some coefficients became zero, they were finally eliminated from the model. The compression of the coefficients of variables helps to avoid the occurrence of model overfitting. Ten-fold cross-validation of included variables was applied to assess the model’s performance and determine suitable parameter values. Significant and proper variables with nonzero coefficients identified by the LASSO regression model were chosen for binary multivariable logistic regression analysis. On the basis of this, statistically significant predictors were applied to develop a nomogram prediction model. A calibration curve was plotted to evaluate the calibration of the nomogram, and a consistency index (C-index) was reported to assess the concordance between the predicted and actual outcomes. An internal validation for the nomogram was performed using 1,000 times bootstrapping. The discriminative capabilities were evaluated by using the receiver operating characteristic (ROC) curve and calculating the area under the ROC curve (AUC).

Subgroup and sensitivity analyses in terms of sex, age, race, education level, alcohol consumption, smoking status, physical activity, BMI, comorbidities, including diabetes, hypertension, hyperlipidemia, coronary artery disease, and chronic kidney disease, medication treatments, including insulin drugs, oral hypoglycemic agents, antihypertensive drugs, and lipid-lowering drugs, were conducted to examine the presence of significant interactions of these covariates with the association between the TyG index and fragility fracture events. All analyses were performed using R software version 4.3.3 (http://www.R-project.org, The R Foundation, Vienna, Austria). P < 0.05 was considered to be statistically significant.

Results

Baseline characteristics

A total of 25,082 participants were included in the analysis, with a mean age of 48.5 ± 17.6 years. The baseline characteristics of the study participants were summarized in Table 1. The overall prevalence of fragility fractures was 1.10% in the study population. There were significant differences in sex, race composition, education level, drinking status, and physical activity between the two groups. Compared to the non-fragility fracture individuals, patients with fragility fractures were older, more likely to be female, and had a higher prevalence of comorbidities, including diabetes, hypertension, hyperlipidemia, coronary artery disease, and chronic kidney disease. Additionally, they were more likely to have a larger waist, higher SBP, fasting blood glucose, HbA1c, the TyG index, alkaline phosphatase, phosphorus, and 25(OH)D levels, and a higher proportion of medication treatments, as well as lower fasting blood insulin and albumin levels and BMD indexes (all P < 0.05).

Table 1 Baseline characteristics of participants grouped by fragility fracture status
Full size table

To further control for the confounding effects, an IPTW was employed. A comparison of participants' characteristics before and after weighting is provided in Supplemental Table 1. Covariates were successfully balanced between groups (Supplemental Fig. 1). Furthermore, Fig. 2 illustrates significant positive correlations between the TyG index and fasting glucose (r = 0.44), fasting triglyceride (r = 0.94), and HOMA-IR (r = 0.42), whereas a negative correlation was observed between the TyG index and QUICKI (r =  − 0.42).

Fig. 2
figure 2

Relationship between the TyG index and other insulin resistance indexes. TyG, triglyceride-glucose; HOMA-IR, homeostasis model assessment of insulin resistance; QUICKI, quantitative insulin sensitivity check index

Full size image

Association of the TyG index with fragility fractures

Results of the association between baseline factors and fragility fractures are shown in Supplemental Table 2. The TyG index as a continuous variable was strongly associated with fragility fractures (P < 0.001) in the univariate logistic regression analysis. In the IPTW-weighted multivariate logistic regression analysis (Table 2), the highest TyG index quartile exhibited a positive association with fragility fractures (OR: 1.75, 95% CI: 1.23–2.48, P = 0.002) compared to the lowest quartile in the non-adjusted model. Similar associations were observed after adjusting for potential confounders, whereby, the TyG index exhibited a robust association with the risk of fragility fractures (OR, 1.94; 95% CI, 1.31–2.88; P < 0.001). RCS regression revealed a positive linear association between the TyG index and fragility fractures (P for nonlinearity = 0.247; Fig. 3). We also investigated the potential sex differences in RCS regression, and the linear relationship between the TyG index and fragility fractures seemed to appear earlier and more significant in females (Supplemental Fig. 2).

Table 2 IPTW-weighted logistic regression analysis on the association between the TyG index and fragility fractures
Full size table
Fig. 3
figure 3

Restricted cubic spline model for the association between the TyG index and fragility fractures among all the study participants. RCS regression was adjusted for age, sex, educational level, BMI, alcohol consumption, SBP, DBP, HbA1c, albumin, alkaline phosphatase, 25(OH)D, diabetes, hypertension, hyperlipidemia, coronary artery disease, chronic kidney disease, and medication treatments, including insulin drugs, oral hypoglycemic agents, antihypertensive drugs, and lipid-lowering drugs. TyG, triglyceride-glucose; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HbA1c, glycated hemoglobin; 25(OH)D, 25-hydroxyvitamin D

Full size image

Key factors of fragility fractures

To identify key factors associated with fragility fractures, we employed a LASSO binary logistic regression model. This approach minimized prediction error by systematically shrinking regression coefficients toward zero, thereby reducing overfitting and enhancing model interpretability (Fig. 4A). Variables with nonzero regression coefficients were retained, indicating a stronger association with fragility fractures. Through this analysis, 20 optimal predictors were identified, including age, sex, race, educational level, physical activity, SBP, waist circumference, HbA1c, the TyG index, albumin, 25(OH)D, diabetes, hypertension, hyperlipidemia, coronary artery disease, chronic kidney disease, and the use of antihypertensive and lipid-lowering medications (Fig. 4B). These predictors collectively reflect the multifactorial nature of fragility fractures, integrating demographic, clinical, biochemical, and lifestyle factors to improve risk stratification.

Fig. 4
figure 4

The least absolute shrinkage and selection operator (LASSO) binary logistic regression model for identifying key fragility fracture-related factors. A LASSO coefficient profiles of 24 clinical variables. A coefficient profile plot was produced against log (lambda). B The optimal lambda in the LASSO model was validated using tenfold cross-validation

Full size image

Development and validation of the prediction nomogram

Multivariable logistic regression analysis incorporating features selected by the LASSO regression model revealed that age, sex, the TyG index, albumin, coronary artery disease, antihypertensive treatment, and lipid-lowering treatment were statistically significant (Table 3). A risk prediction nomogram was then developed based on these variables (Fig. 5A). As an example to better explain the nomogram model, if the subject is a woman, aged 80, with coronary artery disease, with medication treatments of antihypertensive drugs and lipid-lowering drugs, albumin of 43 g/L, and the TyG index of 4.87, the probability of fragility fractures is estimated to be 12.7% (Fig. 5B). The prediction nomogram exhibited a C-index of 0.901 after 1,000 bootstrapping internal validations, indicating good agreement between predicted and actual outcomes. Additionally, the pooled AUC confirmed the model’s robustness (Fig. 6).

Table 3 Binary multivariable logistic regression analysis for fragility fracture-related feature selection
Full size table
Fig. 5
figure 5

Development of the risk nomogram (A) and the dynamic nomogram for an example (B). The fragility fracture risk nomogram was developed with the predictors including age, sex, albumin, the TyG index, coronary artery disease, antihypertensive drugs, and lipid-lowering drugs. TyG, triglyceride-glucose; ** P < 0.01, *** P < 0.001

Full size image
Fig. 6
figure 6

Calibration curves and ROC curve of the fragility fractures risk nomogram model. ROC, receiver operating characteristic; AUC, area under the ROC curve

Full size image

Subgroup and sensitivity analyses

Subgroup and sensitivity analyses confirmed that the association between the TyG index and fragility fractures remained robust. As illustrated in Supplemental Fig. 3, none of the stratifying variables, including age, race, education level, alcohol consumption, smoking status, physical activity, BMI, comorbidities, including diabetes, hypertension, hyperlipidemia, coronary artery disease, chronic kidney disease, or medication treatments, including insulin drugs, oral hypoglycemic agents, antihypertensive drugs, and lipid-lowering drugs, showed significant interactions (all P for interaction > 0.05). However, when the P value for the interaction of sex was < 0.05, the TyG index appeared to have more prominence of its predictive value in females, which was consistent with the conclusion we observed in RCS regression.

Discussion

In this study, we included 25,082 individuals participating in the NHANES survey and conducted a comprehensive assessment of the relationship between the TyG index and fragility fractures. Our study identified a significant positive linear association between the TyG index and the occurrence of fragility fractures in the general population of the United States. This association remained robust after adjusting for potential confounders. Furthermore, we identified seven clinical risk factors strongly associated with fragility fractures, including the TyG index, and constructed a nomogram predictive model. This model demonstrated high predictive value. Our findings provide novel insights into the factors associated with fragility fractures, offering valuable implications for clinical practice and public health strategies.

The TyG index, a widely recognized marker of insulin resistance, has been shown to be a useful predictor for various metabolic disorders and cardiovascular diseases [14, 16, 22]. Previous studies also suggested a significant correlation between the TyG index and bone density as well as markers of bone metabolism, indicating its potential utility in bone-related diseases [20, 28, 29]. A cross-sectional study in Korea found a negative correlation between the TyG index and femoral neck BMD in non-diabetic individuals aged ≥ 50 years and postmenopausal women, being particularly significant in women with a normal BMI [19]. Similarly, a study involving a cross-sectional and longitudinal cohort of Chinese health examination participants also found a significant negative correlation between the TyG index and BMD, with a significant positive correlation with a high risk of low bone mass and osteoporosis [18]. By contrast, Zhan et al. reported a positive correlation between the TyG index and BMD in elderly non-diabetic Chinese individuals, particularly when combined with the BMI [30]. The heterogeneity in study findings may stem from variations in study populations, BMD measurement sites, and confounding factors such as smoking, alcohol consumption, blood levels of calcium, phosphorus, and vitamin D, as well as comorbidities and medication use. Recent research leveraging NHANES data has highlighted the association between the TyG index and osteoporosis. Zhou et al. reported that a higher TyG index was significantly associated with increased odds of osteoporosis, particularly in males under 65 years, with BMI < 25 kg/m2, or without diabetes, and in females with BMI > 30 kg/m2 [31]. These findings complement our study, indicating that the TyG index, as a surrogate for insulin resistance, could be associated with bone health deterioration and serve as a useful marker for identifying individuals at risk of both fragility fractures and osteoporosis. However, research on the relationship between the TyG index and fragility fractures remains limited. Only one cohort study investigated the relationship between the TyG index and the risk of fragility fractures in 220 postmenopausal women with type 2 diabetes mellitus and osteoporosis, finding a 20.9% incidence of fragility fractures over a 6-year follow-up, with an increased risk associated with higher TyG index levels [32]. Our study, based on large-scale cross-sectional data, identified a significant positive linear association between the TyG index levels and fragility fractures, particularly prominent in females. Notably, our study benefits from a larger sample size and encompasses positive events across the entire population, rather than being limited to patients with diabetes or elderly women, thereby mitigating potential selection bias. In terms of the interplay between insulin resistance and skeletal fragility, multifaceted mechanisms may be implicated [10, 33]. Existing evidence indicates that insulin resistance adversely impacts the proliferation, differentiation, and viability of osteoblasts, thereby attenuating their activity [34, 35]. Furthermore, under conditions of insulin resistance, there is a significant increase in the population of activated CD4 + T cells within the bone marrow, fostering a pro-inflammatory milieu characterized by elevated levels of pro-inflammatory cytokines (e.g., tumor necrosis factor and interleukin-6), which, in turn, instigate osteoclastogenesis and augment bone resorption, ultimately culminating in bone loss [36]. Recent studies suggest that insulin resistance impacts bone health through a dynamic interplay between anabolic and catabolic pathways. Insulin receptor (InsR) signaling promotes osteoblastic activity and bone formation, while increased pro-inflammatory cytokines under insulin resistance stimulate osteoclastic activity through the RANK/RANKL/osteoprotegerin pathway, leading to bone resorption. The degree of insulin resistance and cytokine activity determines the net effect on bone mass, ranging from increased bone mass in mild insulin resistance to significant bone loss in severe cases [37]. Additionally, alterations in lipid levels may impact skeletal health, although the exact mechanisms remain unclear [38]. Further research is needed to elucidate the specific molecular mechanisms behind these associations.

It is acknowledged that focusing solely on bone density results may not effectively predict fracture risk. Multicenter cohort studies suggested that only 10%–44% of fragility fractures can be attributed to baseline osteoporosis, necessitating the investigation of other related risk factors [7]. The widely used fracture risk assessment tool, FRAX®, estimates the 10-year probability of major osteoporotic fractures (clinical spine, hip, forearm, and proximal humerus) and hip fractures in individuals aged 40–90 years based on clinical risk factors and BMD [39]. However, not all relevant clinical risk factors are included in the FRAX tool, and it is recognized that clinical risk factors significantly increase fracture risk. A deeper understanding of these risk factors is beneficial for effective prevention and management. Our study comprehensively analyzed the population, considering demographic indicators, physical measurements, laboratory tests, comorbidities, and medication use, identifying seven key factors, including the TyG index. These factors are readily accessible in clinical practice, facilitating clinical decision-making. We found that besides the TyG index, serum albumin levels were significantly negatively associated with the risk of fragility fractures. Previous studies have suggested an independent correlation between low serum albumin concentration and the risk of osteoporosis in postmenopausal rheumatoid arthritis patients [40, 41]. Kanazawa et al. developed a fracture risk assessment tool for individuals with type 2 diabetes, demonstrating a significant negative correlation between low serum albumin levels and vertebral fractures [42]. The serum albumin concentration appears to be associated with skeletal health through mechanisms similar to bone collagen, promoting bone mineralization through the adsorption of charged amino acids and enhancing biological activity such as osteoblast adhesion and proliferation by specific cell binding to hydroxyapatite [43,44,45]. These mechanisms may partly explain the association between low serum albumin and fragility fractures.

To our knowledge, this population-based study is the first to reveal the association between the TyG index levels and fragility fractures in a large, representative population, leveraging advanced LASSO regression to construct a robust predictive model. This provides the possibility of early intervention to reduce the risk of fragility fractures by addressing these risk factors and offers new insights for optimizing or developing more comprehensive fracture risk assessment tools in clinical practice.

However, our study had several limitations. First, due to its retrospective design, causal relationships could not be determined. Although multivariable adjustment and subgroup analyses were performed, residual confounding factors may still have influenced clinical outcomes. The constructed model was only internally validated and not externally validated in external cohorts, limiting its medical value. Further validation is warranted in larger prospective studies. Second, the relatively small sample size of positive events in this study, coupled with numerous influencing factors, may have introduced bias despite the use of LASSO regression to avoid overfitting. Additionally, we did not assess the impact of dietary factors, medication for osteoporosis, menstrual status in females, or serum hormone levels on outcome events due to incomplete records in the NHANES database, leading to further sample size reduction. Third, only the occurrence of fragility fractures was assessed in this study without analyzing the fracture sites or occurrence frequency, among others, warranting further detailed investigations. Fourth, this study only evaluated baseline TyG index values, not considering their changes over time. Fifthly, differences in age and sex between fragility fracture patients and the control group may have introduced limitations to our study results, although efforts were made to apply matching to balance the two groups. The inherent imbalance in baseline data may have affected the comparability of the two groups, which should be considered when interpreting the significance of our study. Lastly, self-reported information on fragility fractures and other variables, such as physical activity and medical history, may have inevitably led to recall bias and measurement inaccuracies. Furthermore, the lack of questionnaire data for 2011–2012 and 2015–2016 in the NHANES database compromised data consistency. Therefore, caution should be exercised in analyzing and interpreting the data.

Conclusions

In summary, our study demonstrates a significant association between elevated TyG index levels and fragility fracture risk in the general US population. Through comprehensive analysis, the TyG index was identified as likely to be a key predictor of fragility fractures. Our findings offer valuable insights into the prevention and management of fragility fractures, underscoring the potential for early intervention and the development of more accurate fracture risk assessment tools in clinical practice. Further studies are warranted to validate these findings.

Data availability

The datasets generated and analysed during the current study are available in the NHANES repository (https://www.cdc.gov/nchs/nhanes/index.htm).

Abbreviations

BMD:

Bone mineral density

BMI:

Body mass index

DBP:

Diastolic blood pressure

HbA1c:

Glycosylated hemoglobin

HDL-C:

High density lipoprotein cholesterol

HOMA-IR:

Homeostasis model assessment of insulin resistance

IR:

Insulin resistance

LDL-C:

Low density lipoprotein cholesterol

NHANES:

National Health and Nutrition Examination Survey

QUICKI:

Quantitative insulin sensitivity check

SBP:

Systolic blood pressure

TC:

Total cholesterol

TG:

Triglycerides

TyG:

Triglyceride-glucose

25(OH)D:

25-Hydroxyvitamin D

References

  1. Rabar S, Lau R, O’Flynn N, Li L, Barry P, on behalf of the Guideline Development Group. Risk assessment of fragility fractures: summary of NICE guidance. BMJ. 2012;345: e3698–e3698. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/bmj.e3698.

  2. Siris ES, Adler R, Bilezikian J, Bolognese M, Dawson-Hughes B, Favus MJ, et al. The clinical diagnosis of osteoporosis: a position statement from the National Bone Health Alliance Working Group. Osteoporos Int. 2014;25:1439–43. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s00198-014-2655-z.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. LeBoff MS, Greenspan SL, Insogna KL, Lewiecki EM, Saag KG, Singer AJ, et al. The clinician’s guide to prevention and treatment of osteoporosis. Osteoporos Int. 2022;33(10):2049–102. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s00198-021-05900-y.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Geusens P, Bours SPG, Wyers CE, Van Den Bergh JP. Fracture liaison programs. Best Pract Res Clin Rheumatol. 2019;33:278–89. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.berh.2019.03.016.

    Article  PubMed  Google Scholar 

  5. IOF Fracture Working Group, Åkesson K, Marsh D, Mitchell PJ, McLellan AR, Stenmark J, et al. Capture the Fracture: a Best Practice Framework and global campaign to break the fragility fracture cycle. Osteoporos Int 2013;24: 2135–52. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s00198-013-2348-z.

  6. Liu J, Curtis EM, Cooper C, Harvey NC. State of the art in osteoporosis risk assessment and treatment. J Endocrinol Invest. 2019;42:1149–64. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s40618-019-01041-6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Stone KL, Seeley DG, Lui L-Y, Cauley JA, Ensrud K, Browner WS, et al. BMD at multiple sites and risk of fracture of multiple types: Long-term results from the study of osteoporotic fractures. J Bone Miner Res. 2003;18:1947–54. https://doiorg.publicaciones.saludcastillayleon.es/10.1359/jbmr.2003.18.11.1947.

    Article  PubMed  Google Scholar 

  8. Chiodini I, Morelli V, Masserini B, Salcuni AS, Eller-Vainicher C, Viti R, et al. Bone Mineral Density, Prevalence of Vertebral Fractures, and Bone Quality in Patients with Adrenal Incidentalomas with and without Subclinical Hypercortisolism: An Italian Multicenter Study. J Clin Endocrinol Metab. 2009;94:3207–14. https://doiorg.publicaciones.saludcastillayleon.es/10.1210/jc.2009-0468.

    Article  CAS  PubMed  Google Scholar 

  9. Wilson-Barnes SL, Lanham-New SA, Lambert H. Modifiable risk factors for bone health & fragility fractures. Best Pract Res Clin Rheumatol. 2022;36: 101758. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.berh.2022.101758.

    Article  PubMed  Google Scholar 

  10. Khosla S, Samakkarnthai P, Monroe DG, Farr JN. Update on the pathogenesis and treatment of skeletal fragility in type 2 diabetes mellitus. Nat Rev Endocrinol. 2021;17:685–97. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41574-021-00555-5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Kanis JA, Johansson H, Oden A, McCloskey EV. Assessment of fracture risk. Eur J Radiol. 2009;71:392–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.ejrad.2008.04.061.

    Article  PubMed  Google Scholar 

  12. Napoli N, Conte C, Pedone C, Strotmeyer ES, Barbour KE, Black DM, et al. Effect of insulin resistance on BMD and fracture risk in older adults. J Clin Endocrinol Metab. 2019;104:3303–10. https://doiorg.publicaciones.saludcastillayleon.es/10.1210/jc.2018-02539.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala MG, Hernández-González SO, et al. The product of triglycerides and glucose, a simple measure of insulin sensitivity. Comparison with the euglycemic-hyperinsulinemic clamp. J Clin Endocrinol Metab 2010;95: 3347–51. https://doiorg.publicaciones.saludcastillayleon.es/10.1210/jc.2010-0288.

  14. Khan SH, Sobia F, Niazi NK, Manzoor SM, Fazal N, Ahmad F. Metabolic clustering of risk factors: Evaluation of triglyceride-glucose index (TyG index) for evaluation of insulin resistance. Diabetol Metab Syndr. 2018;10:74. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13098-018-0376-8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Huo R-R, Liao Q, Zhai L, You X-M, Zuo Y-L. Interacting and joint effects of triglyceride-glucose index (TyG) and body mass index on stroke risk and the mediating role of TyG in middle-aged and older Chinese adults: A nationwide prospective cohort study. Cardiovasc Diabetol. 2024;23:30. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12933-024-02122-4.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Hong S, Han K, Park C-Y. The triglyceride glucose index is a simple and low-cost marker associated with atherosclerotic cardiovascular disease: A population-based study. BMC Med. 2020;18:361. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12916-020-01824-2.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Lopez-Jaramillo P, Gomez-Arbelaez D, Martinez-Bello D, Abat MEM, Alhabib KF, Avezum Á, et al. Association of the triglyceride glucose index as a measure of insulin resistance with mortality and cardiovascular disease in populations from five continents (PURE study): A prospective cohort study. The Lancet Healthy Longevity. 2023;4:e23-33. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/S2666-7568(22)00247-1.

    Article  PubMed  Google Scholar 

  18. Zhuo M, Chen Z, Zhong M-L, Lei F, Qin J-J, Liu S, et al. Association of insulin resistance with bone mineral density in a nationwide health check-up population in China. Bone. 2023;170: 116703. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.bone.2023.116703.

    Article  CAS  PubMed  Google Scholar 

  19. Yoon JH, Hong AR, Choi W, Park JY, Kim HK, Kang H-C. Association of triglyceride-glucose index with bone mineral density in non-diabetic koreans: KNHANES 2008–2011. Calcif Tissue Int. 2021;108:176–87. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s00223-020-00761-9.

    Article  CAS  PubMed  Google Scholar 

  20. Zhan H, Liu X, Piao S, Rong X, Guo J. Association between triglyceride-glucose index and bone mineral density in US adults: a cross sectional study. J Orthop Surg Res. 2023;18:810. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13018-023-04275-6.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Centers for Disease Control and Prevention. About the National Health and Nutrition Examination Survey. Accessed May 1,2023. https://www.cdc.gov/nchs/nhanes/about_nhanes.htm.

  22. Simental-Mendía LE, Rodríguez-Morán M, Guerrero-Romero F. The product of fasting glucose and triglycerides as surrogate for identifying insulin resistance in apparently healthy subjects. Metab Syndr Relat D. 2008;6:299–304. https://doiorg.publicaciones.saludcastillayleon.es/10.1089/met.2008.0034.

    Article  CAS  Google Scholar 

  23. Matthews DR, Hosker JR, Rudenski AS, Naylor BA, Treacher DF, Turner RC, et al. Homeostasis model assessment: Insulin resistance and fl-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412–9. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/BF00280883.

    Article  CAS  PubMed  Google Scholar 

  24. Katz A, Nambi SS, Mather K, Baron AD, Follmann DA, Sullivan G, et al. Quantitative insulin sensitivity check index: A simple, accurate method for assessing insulin sensitivity in humans. J Clin Endocrinol Metab. 2000;85(7):2402–10. https://doiorg.publicaciones.saludcastillayleon.es/10.1210/jcem.85.7.6661.

    Article  CAS  PubMed  Google Scholar 

  25. Warriner AH, Patkar NM, Curtis JR, Delzell E, Gary L, Kilgore M, et al. Which fractures are most attributable to osteoporosis? J Clin Epidemiol. 2011;64:46–53. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jclinepi.2010.07.007.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Naimi AI, Moodie EEM, Auger N, Kaufman JS. Constructing inverse probability weights for continuous exposures: a comparison of methods. Epidemiology. 2014;25:292–9. https://doiorg.publicaciones.saludcastillayleon.es/10.1097/EDE.0000000000000053.

    Article  PubMed  Google Scholar 

  27. Kusamura S, Barretta F, Yonemura Y, Sugarbaker PH, Moran BJ, Levine EA, et al. The role of hyperthermic intraperitoneal chemotherapy in pseudomyxoma peritonei after cytoreductive surgery. JAMA Surg. 2021;156: e206363. https://doiorg.publicaciones.saludcastillayleon.es/10.1001/jamasurg.2020.6363.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Gu P, Pu B, Xin Q, Yue D, Luo L, Tao J, et al. The metabolic score of insulin resistance is positively correlated with bone mineral density in postmenopausal patients with type 2 diabetes mellitus. Sci Rep. 2023;13:8796. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41598-023-32931-8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Sun W, Xing Y, Zhou F, Ma Y, Wan X, Ma H. Association analysis of triglyceride glucose-body mass index and bone turnover markers in patients with type 2 diabetes mellitus. Diabetes, Metabolic Syndrome and Obesity. 2023;16:1435–47. https://doiorg.publicaciones.saludcastillayleon.es/10.2147/DMSO.S406849.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Wen Z, Li Y, Xu L, Yue C, Wang Q, Chen R, et al. Triglyceride glucose–body mass index is a reliable indicator of bone mineral density and risk of osteoporotic fracture in middle-aged and elderly nondiabetic Chinese individuals. J Clin Med. 2022;11:5694. https://doiorg.publicaciones.saludcastillayleon.es/10.3390/jcm11195694.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Association between triglyceride-glucose index and osteoporosis: A cross-sectional study based on the national health and nutrition examination survey. Frontiers in Medical Science Research 2024. https://doiorg.publicaciones.saludcastillayleon.es/10.25236/fmsr.2024.060208

  32. Pan J, Huang X, Wang Q, Sun J, Zhai Z, Mo J, et al. Triglyceride glucose index is strongly associated with a fragility fracture in postmenopausal elderly females with type 2 diabetes mellitus combined with osteoporosis: a 6-year follow-up study. Clin Interv Aging. 2023;18:1841–9. https://doiorg.publicaciones.saludcastillayleon.es/10.2147/CIA.S434194.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. De Paula FJA, Horowitz MC, Rosen CJ. Novel insights into the relationship between diabetes and osteoporosis. Diabetes Metab Res Rev. 2010;26:622–30. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/dmrr.1135.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Cao JJ. Effects of obesity on bone metabolism. J Orthop Surg Res. 2011;6:30. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/1749-799X-6-30.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Fulzele K, Riddle RC, DiGirolamo DJ, Cao X, Wan C, Chen D, et al. Insulin receptor signaling in osteoblasts regulates postnatal bone acquisition and body composition. Cell. 2010;142:309–19. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.cell.2010.06.002.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Cifuentes-Mendiola SE, Solis-Suarez DL, Martínez-Dávalos A, Godínez-Victoria M, García-Hernández AL. CD4+ T-cell activation of bone marrow causes bone fragility and insulin resistance in type 2 diabetes. Bone. 2022;155: 116292. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.bone.2021.116292.

    Article  CAS  PubMed  Google Scholar 

  37. Khanmohammadi S, Kuchay MS. Effects of metabolic dysfunction-associated steatotic liver disease on bone density and fragility fractures: Associations and mechanisms. Journal of Obesity & Metabolic Syndrome. 2024;33:108–20. https://doiorg.publicaciones.saludcastillayleon.es/10.7570/jomes24004.

    Article  Google Scholar 

  38. Orozco P. Atherogenic lipid profile and elevated lipoprotein (a) are associated with lower bone mineral density in early postmenopausal overweight women. Eur J Epidemiol. 2004;19:1105–12. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s10654-004-1706-8.

    Article  CAS  PubMed  Google Scholar 

  39. Kanis JA, Johnell O, Oden A, Johansson H, McCloskey E. FRAX™ and the assessment of fracture probability in men and women from the UK. Osteoporos Int. 2008;19:385–97. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s00198-007-0543-5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Nagayama Y, Ebina K, Tsuboi H, Hirao M, Hashimoto J, Yoshikawa H, et al. Low serum albumin concentration is associated with increased risk of osteoporosis in postmenopausal patients with rheumatoid arthritis. J Orthop Sci. 2022;27:1283–90. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jos.2021.08.018.

    Article  PubMed  Google Scholar 

  41. Afshinnia F, Pennathur S. Association of hypoalbuminemia with osteoporosis: analysis of the national health and nutrition examination survey. J Clin Endocr Metab. 2016;101:2468–74. https://doiorg.publicaciones.saludcastillayleon.es/10.1210/jc.2016-1099.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Kanazawa I, Tanaka K, Takeo A, Notsu M, Miyake H, Sugimoto T. A scoring assessment tool for the risk of vertebral fractures in patients with type 2 diabetes mellitus. Bone. 2019;122:38–44. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.bone.2019.02.003.

    Article  PubMed  Google Scholar 

  43. D’Elia NL, Gravina N, Ruso JM, Marco-Brown JL, Sieben JM, Messina PV. Albumin-mediated deposition of bone-like apatite onto nano-sized surfaces: effect of surface reactivity and interfacial hydration. J Colloid Interface Sci. 2017;494:345–54. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jcis.2017.01.047.

    Article  CAS  PubMed  Google Scholar 

  44. Jin M, Zhu S, Hou Y. Insight on serum albumin: from structure and biological properties to functional biomaterials for bone repair. ACS Biomater Sci Eng. 2023;9:2235–50. https://doiorg.publicaciones.saludcastillayleon.es/10.1021/acsbiomaterials.2c01553.

    Article  CAS  PubMed  Google Scholar 

  45. Tavafoghi M, Cerruti M. The role of amino acids in hydroxyapatite mineralization. J R Soc Interface. 2016;13:20160462. https://doiorg.publicaciones.saludcastillayleon.es/10.1098/rsif.2016.0462.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank the participants and staff of the National Health and Nutrition Examination Surveys for their valuable contributions. They were not compensated.

Funding

This study was funded by the National Key Research and Development Program of China [2020YFC2009004].

Author information

Authors and Affiliations

  1. Department of Endocrinology, Institute of Geriatric Medicine, Beijing Hospital, National Center of Gerontology, Chinese Academy of Medical Sciences, No. 1 Dahua Road, Dongcheng District, Beijing, 100730, China

    Yuan Lou, Huan Chen, Fuli Man, Lina Zhang & Qi Pan

  2. Chinese Academy of Medical Sciences, Graduate School of Peking Union Medical College, Beijing, 100000, China

    Yuan Lou, Huan Chen & Qi Pan

Authors
  1. Yuan Lou
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Huan Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Fuli Man
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Lina Zhang
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Qi Pan
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Y. L. conceived the study, analyzed the data, drafted the manuscript; H.C., FL. M. and LN. Z. reviewed the results; Q. P. contributed to the funding support; Y. L and Q. P. contributed to critical revisions of the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Qi Pan.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lou, Y., Chen, H., Man, F. et al. Association between the Triglyceride-glucose index and fragility fractures among US adults: insights from NHANES. Diabetol Metab Syndr 17, 96 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13098-025-01669-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13098-025-01669-w

Keywords

Diabetology & Metabolic Syndrome

ISSN: 1758-5996

Contact us