www.nature.

com/scientificreports

OPEN The correlation

between high‑density lipoprotein
cholesterol and bone mineral
density in adolescents:
a cross‑sectional study
Gao‑Xiang Wang 1,2,4, Jun‑Tong Li 1,2,4, De‑Liang Liu 2*, Shu‑Fang Chu 2*, Hui‑Lin Li 2*,
Heng‑Xia Zhao 2, Ze‑Bin Fang 2,3 & Wei Xie 2,3

Recent studies have shown a correlation between high-density lipoprotein cholesterol (HDL-C) and
bone mineral density (BMD) in adults, but their relationship is unclear in adolescents. This study
aimed to explore whether a correlation existed between them among adolescents aged 12–19.
Data analyzed in our study was fetched from the National Health and Nutrition Examination Survey
(NHANES) database 2011–2018. The relationship between HDL-C level and total BMD value was
analyzed by multivariate logistic regression models, fitted smoothing curves, and generalized additive
models. 3770 participants participated in this analysis. After adjusting for all relevant covariates
involved in this study, we found a negative correlation between HDL-C levels and total bone density
in male adolescents.Furthermore, the stratified analysis showed that all covariables-adjusted models
retained the negative correlation excepting female, black, or Mexican American subgroups. An
inverted U-shaped curve represented the correlation of HDL-C and total BMD among adolescents aged
16 to 19, and the turning point was 1.06 mmol/L. After adjusting for all relevant covariates involved
in this study, the study found a negative correlation between HDL-C levels and total BMD in male
adolescents aged 12 to 19, particularly among those of races other than Black and Mexican. There was
a saturation effect between HDL-C level and total BMD in 16–19-year-old adolescents. The turning
point was 1.06 mmol/L. Therefore, HDL-C might be a biomarker to detect bone health and further
perform a more detailed examination.

Osteoporosis is a common skeleton disease characterized by decreased bone mass, bone fragility, and increased
fracture ­risk1. It is noteworthy that osteoporosis affects not only elderly subjects or postmenopausal women, but
more attention has been paid to adolescents and c­ hildren2. In adolescence, which is a progressive ontogenesis
period, skeleton microarchitecture and mineralization make a qualitative change, and 40% of the peak bone mass
is accumulated in this ­window3. Making an exact assessment of adolescents’ skeletal health benefits their current
and future quality of life. World Health Organization proposed bone mineral density (BMD) value tested using
dual X-ray absorptiometry (DXA) is the gold standard to diagnose o ­ steoporosis4. Identifying associated and
potential detrimental factors for bone health is significant for preventing and detecting osteoporosis.
Serum lipids, which consist of many substances, play a crucial role in physiopathology and are strongly asso-
ciated with various metabolic diseases. High-density lipoproteins-cholesterol (HDL-C) is considered beneficial
­cholesterol5. Higher HDL-C levels meant better cardiovascular health in the p ­ ast6. However, some researchers
have proposed different perspectives about HDL-C. Hamer et al. indicated that the high HDL-C also increased
mortality in a large population-based s­ ample7.

1
Department of Endocrinology, Shenzhen Traditional Chinese Medicine Hospital Affiliated to Nanjing University
of Chinese Medicine, Shenzhen 518033, Guangdong, China. 2Department of Endocrinology, Shenzhen Traditional
Chinese Medicine Hospital, Shenzhen  518033, Guangdong, China. 3The Fourth Clinical Medical College of
Guangzhou University of Chinese Medicine, Shenzhen  518033, Guangdong, China. 4These authors contributed
equally: Gao-Xiang Wang and Jun-Tong Li. *email: ldl2580@gzucm.edu.cn; chushufanggzhtcm@163.com;
sztcmlhl@163.com

Scientific Reports | (2023) 13:5792 | https://doi.org/10.1038/s41598-023-32885-x 1

Vol.:(0123456789)
 www.nature.com/scientificreports/

Numerous kinds of research revealed that lipids level was closely related to BMD v­ alue8–10. Some studies
showed a negative correlation between HDL-C and adult ­BMD11,12. Despite Ibrahim Duran et al. demonstrated
muscle mass inversely correlated with HDL-C level among a­ dolescents13, no research clarified and defined the
relationship between HDL-C level and BMD value in adolescents. Therefore, exploring the correlation of HDL-C
concentration and BMD value would be beneficial for estimating the risk of osteoporosis, which could effectively
promote the prevention and treatment of osteoporosis.
This study aimed to clarify the relationship between HDL-C and total BMD in adolescents aged 12–19. Pop-
ulation-based subjects were extracted from the National Health and Nutrition Examination Survey (NHANES)
to provide credible evidence for this issue.

Materials and methods
Patient  collection and data extraction.  NHANES database is the largest population-based national
survey in the world and is organized by the National Center for Health Statistics (NCHS) in the US. This sur-
vey is conducted every 2 years to evaluate the nutritional and health status of a representative of the noninsti-
tutionalized US population. A complex stratified and multistage sampling design was applied to extract the
representative population in the NHANES database. More details of the database are available at http://​www.​
cdc.​gov/​nchs/​nhanes/. The current study consisted of four data cycles (2011–2018). In this study, there were
39,156 participants enrolled from NHANES 2011–2018 database. After limiting to adolescents aged 12 to 19, the
participants narrowed to 5215. Subjects with missing total BMD data (n = 1033) or HDL-C data (n = 412) were
excluded. Finally, the data of 3770 cases were retained for analysis (Fig. 1). All NHANES protocols were ratified
by the ethics review board of ­NCHS14.

Covariates.  Our study’s dependent and independent variables were HDL-C and BMD, respectively. The
Roche Modular P chemistry analyzer detected the HDL-C concentration or the Roche modular P and Roche
Cobas 6000 chemistry analyzers in the 2011–2018 cycle. The BMD value of all cases in this study was tested
by using dual-energy X-ray absorptiometry (DXA). Hologic Discovery Model A densitometer (Hologic, Inc.,
Bedford, Massachusetts) and Apex version 3.2 software were used to detect and analyze BMD. The subject was
scanned, and their BMD was calculated using a whole-body scan. According to the relative regulations and
guidelines of WHO, the mean BMD value of a young adult was a reference value.
Furthermore, demographic characteristics in the NHANES database were regarded as potential confounders,
which were analyzed in this study. These covariates included categorical variables (age, gender, race) and continu-
ous variables (ratio of family income to poverty and various physiological and biochemical indexes including
total calcium (Tca), serum phosphorus (P), total protein (TP), total cholesterol (TC), alkaline phosphatase (ALP),
serum uric acid (SUA), triglyceride (TG), glycohemoglobin (HbA1c), urinary albumin creatinine ratio (ACR),
serum creatinine (Scr), and blood urea nitrogen (BUN).

Statistical analysis.  Based on the complex, multistage, stratified probability sampling design, sample
weights were used in all statistical analyses through software R and Empower Stats. The study utilized weighted
multivariate linear regression models to assess the correlation between HDL-C level and BMD value. Covariates
were regulated as possible effect modulators. Therefore, we established three models for statistical inference in
the subgroup analyses. Model 1 was adjusted with no covariates. Three covariates (age, gender, and race) were
adjusted in model 2. In model 3, the ratio of family income to poverty, Tca, P, TP, ALP, TC, TG, HbA1c, Scr,
ACR, SUA, and BUN were further adjusted. If the association of HDL-C with BMD was nonlinear, smooth curve
fittings were applied to address the nonlinearity in these three models. Percentage and means ± standard devia-
tion represented the categorical and continuous variables, respectively. Categorical variables adopted weighted

Figure 1.  Workflow diagram of samples extraction from the NHANES database.

Scientific Reports | (2023) 13:5792 | https://doi.org/10.1038/s41598-023-32885-x 2

Vol:.(1234567890)
www.nature.com/scientificreports/

Male (n = 1979) Female (n = 1791) P value

Age (years) 15.41 ± 2.29 15.46 ± 2.20 0.4569
Race/ethnicity (%) 0.7132
 White 54.11 54.69
 Black 13 12.02
 Mexican American 16.02 15.52
 Other race 16.87 17.77
The ratio of family income to poverty (%) 2.44 ± 1.56 2.38 ± 1.59 0.2485
Total calcium (mmol/L) 2.41 ± 0.07 2.38 ± 0.08  < 0.0001
Serum phosphorus (mmol/L) 1.44 ± 0.23 1.34 ± 0.18  < 0.0001
Alkaline phosphatase (µ/L) 173.40 ± 105.62 98.75 ± 60.54  < 0.0001
Total cholesterol (mmol/L) 3.94 ± 0.72 4.13 ± 0.77  < 0.0001
Triglyceride (mmol/L) 1.17 ± 0.83 1.04 ± 0.63  < 0.0001
Glycohemoglobin (%) 5.23 ± 0.34 5.23 ± 0.40 0.9986
Serum creatinine (µmol/L) 68.94 ± 15.58 58.50 ± 11.53  < 0.0001
Serum uric acid (µmol/L) 333.45 ± 68.56 265.46 ± 58.05  < 0.0001
Urinary albumin creatinine ratio (mg/g) 20.31 ± 86.14 29.70 ± 110.33 0.0035
Blood urea nitrogen (mmol/L) 4.29 ± 1.24 3.74 ± 1.16  < 0.0001
Total protein (g/L) 72.54 ± 4.13 72.11 ± 4.01 0.0013
HDL cholesterol (mmol/L) 1.29 ± 0.30 1.39 ± 0.31  < 0.0001
Total bone mineral density (g/cm2) 1.04 ± 0.13 1.02 ± 0.10  < 0.0001

Table 1.  Weighted characteristics of the sample in this study. Continuous variables were represented via
mean ± SD, calculated using a weighted linear regression model. Categorical variables were presented via %.
Calculated by The chi-square test.

chi-square tests, and continuous variables adopted weighted linear regression models to evaluate differences
between groups.

Ethics approval and consent to participate.  Participants in NHANES must sign an informed consent
form, the data is now publicly available, and the National Center for Health Statistics Ethics Review Board
evaluates and authorizes it. Converting data into a format that can be analyzed is already feasible. The research
methodology will be based on all statistics. All research will conform to applicable laws and standards when fol-
lowing the research’s data usage guidelines.

Results
Characteristics of the study participants.  Table 1 describes the weighted sociodemographic and medi-
cal characteristics of 3770 eligible participants. The variables, including Tca, P, ALP, TC, TG, Scr, SUA,TP, HDL-
C, ACR, BUN,and total BMD, were significantly different among different gender groups.

Associations between HDL‑C and total BMD.  The multivariate regression analysis to reveal the asso-
ciation of total BMD with HDL-C was presented in Table 2. An inversive association existed between HDL-C
level and total BMD value in model 1 [− 0.047 (− 0.059, − 0.035)] with no variables adjusted, model 2 [− 0.038
(− 0.048, − 0.028)] with age, gender and race were adjusted, and model 3 [− 0.025 (− 0.036, − 0.014)] with all
potential covariables were adjusted. Furthermore, there were three stratified analysis (stratified by age, gender,
or race) that indicated the correlation was not significant in female adolescents [− 0.015 (− 0.029, 0.000)], black
subjects [− 0.005 (− 0.028, 0.017)] and Mexican American subjects [− 0.018 (− 0.044, 0.007)]. The current study
adopted smooth curve fittings to present the nonlinear correlation of HDL-C and BMD (Fig. 2). Moreover, the
results of the stratified analysis are visualized with generalized additive models in Fig. 3. In age-stratified analy-
sis, among adolescents aged 16 to 19, an inverted U-shaped curve represented the correlation of HDL-C and
total BMD (Fig. 3a).
When stratified according to gender and age (Table 3), we found that after adjusting all variables, there was a
negative correlation between HDL-C and total BMD in male adolescents of all ages but not in female adolescents.
When stratified according to gender and race, we found that after adjusting all variables, there was a negative
correlation between HDL-C and total BMD between white men and people of other races. Additionally, Table 4
of this study reveals a saturation effect analysis between HDL-C and total BMD. The results indicate a saturation
effect was observed in adolescents aged 16 to 19. The turning point of direct HDL-C was the same in males and
females (1.06 mmol/L). However, we did not find this effect in 12–16 year-old teenagers.

Scientific Reports | (2023) 13:5792 | https://doi.org/10.1038/s41598-023-32885-x 3

Vol.:(0123456789)
 www.nature.com/scientificreports/

Exposure Model 1, β (95% CI) Model 2, β (95% CI) Model 3, β (95% CI)
Direct HDL cholesterol (mmol/L) − 0.047 (− 0.059, − 0.035) − 0.038 (− 0.048, − 0.028) − 0.025 (− 0.036, − 0.014)
Quintiles of direct HDL cholesterol (mmol/L)
 Q1 Reference Reference Reference
 Q2 − 0.019 (− 0.029, − 0.008) − 0.008 (− 0.017, 0.001) − 0.004 (− 0.012, 0.004)
 Q3 − 0.015 (− 0.026, − 0.004) − 0.011 (− 0.020, − 0.003) − 0.005 (− 0.013, 0.004)
 Q4 − 0.038 (− 0.049, − 0.027) − 0.028 (− 0.037, − 0.019) − 0.015 (− 0.025, − 0.006)
 P for trend  < 0.001  < 0.001 0.002
Stratified by age
 12–15 years old − 0.043 (− 0.058, − 0.028) − 0.054 (− 0.069, − 0.039) − 0.022 (− 0.038, − 0.007)
 16–19 years old − 0.036 (− 0.051, − 0.021) − 0.019 (− 0.033, − 0.004) − 0.023 (− 0.039, − 0.007)
Stratified by gender
 Male − 0.071 (− 0.090, − 0.051) − 0.043 (− 0.058, − 0.029) − 0.036 (− 0.053, − 0.020)
 Female − 0.016 (− 0.031, − 0.000) − 0.024 (− 0.037, − 0.010) − 0.015 (− 0.029, 0.000)
Stratified by race
 White − 0.054 (− 0.077, − 0.030) − 0.040 (− 0.059, − 0.021) − 0.030 (− 0.051, − 0.009)
 Black − 0.058 (− 0.083, − 0.032) − 0.030 (− 0.051, − 0.008) − 0.005 (− 0.028, 0.017)
 Mexican American − 0.044 (− 0.070, − 0.018) − 0.030 (− 0.052, − 0.008) − 0.018 (− 0.044, 0.007)
 Other race − 0.062 (− 0.085, − 0.039) − 0.044 (− 0.064, − 0.024) − 0.036 (− 0.057, − 0.014)

Table 2.  Association between high-density lipoprotein cholesterol and total bone mineral density. Model 1:
No adjusted. Model 2: Adjusted age, gender, and race. Model 3: Adjusted all covariates were. The stratification
variable itself was not adjusted in the subgroup analysis.

Figure 2.  The expounded curved line correlation between the level of HDL-C and total BMD. (a) One black
point was equal to a subject. (b) A 95% CI is expressed as an area between two blue dotted lines. Each point
shows HDL-C magnitude, and a continuous line is drawn.

Discussion
We explored the relationship between HDL-C and BMD in adolescents. A higher quality of skeleton in adoles-
cence protects against osteoporosis later in ­life15. Multivariate logistic regression analysis indicated an inversive
correlation between HDL-C concentration with total BMD value. HDL-C was well known as beneficial choles-
terol, promoting reversed cholesterol transportation and decreasing the incidence of various chronic diseases,
including cardiovascular disease and endocrine d ­ isorders16,17. However, more and more researchers suggested
the protective effect of HDL-C in some diseases was ­overestimated7,18. Previous research has confirmed that
the correlation of HDL-C with muscle mass was also negative among a­ dolescents13,19. Therefore, HDL-C level
might be a potential biomarker for completing the evaluation of skeleton health in adolescents and improving
quality of life in the future.
The relationship between HDL-C level and BMD value had been widely studied, but a unanimous agree-
ment was not reached. Yuchen Tang et al. showed that HDL-C levels were inversely correlated with BMD value
­ steoporosis20. Qi Zhang et al. explored the relationship of serum cholesterols
and might predict bone loss and o

Scientific Reports | (2023) 13:5792 | https://doi.org/10.1038/s41598-023-32885-x 4

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 3.  The HDL-C concentration and BMD relationship. (a) The illustrated line was stratified by age. (b)
The illustrated line was stratified by sex. (c) The illustrated line was stratified by race/ethnicity.

Stratified by age Male Female

Stratified by age
 12–15 years old
  Model 1, β (95% CI) − 0.060 (− 0.082, − 0.039) < 0.00001 − 0.027 (− 0.048, − 0.005) 0.01517
  Model 2, β (95% CI) − 0.071 (− 0.092, − 0.050) < 0.00001 − 0.032 (− 0.053, − 0.011) 0.00258
  Model 3, β (95% CI) − 0.034 (− 0.055, − 0.013) 0.00129 − 0.011 (− 0.032, 0.010) 0.31799
 16–19 years old
  Model 1, β (95% CI) − 0.007 (− 0.031, 0.017) 0.57234 − 0.016 (− 0.034, 0.003) 0.09634
  Model 2, β (95% CI) 0.016 (− 0.040, 0.008) 0.19389 − 0.021 (− 0.039, − 0.003) 0.02354
  Model 3, β (95% CI) − 0.027 (− 0.053, − 0.001) 0.04248 − 0.020 (− 0.042, 0.001) 0.05890
Stratified by race
 White
  Model 1, β (95% CI) − 0.094 (− 0.131, − 0.057) < 0.00001 − 0.013 (− 0.042, 0.017) 0.40100
  Model 2, β (95% CI) − 0.060 (− 0.087, − 0.033) 0.00002 − 0.014 (− 0.039, 0.011) 0.28524
  Model 3, β (95% CI) − 0.054 (− 0.085, − 0.024) 0.00050 − 0.008 (− 0.036, 0.021) 0.60236
 Black
  Model 1, β (95% CI) − 0.067 (− 0.107, − 0.026) 0.00126 − 0.045 (− 0.077, − 0.014) 0.00465
  Model 2, β (95% CI) 0.000 (− 0.031, 0.031) 0.99446 − 0.044 (− 0.072, − 0.016) 0.00218
  Model 3, β (95% CI) 0.013 (− 0.021, 0.047) 0.44398 − 0.021 (− 0.050, 0.009) 0.16453
 Mexican American
  Model 1, β (95% CI) − 0.039 (− 0.081, 0.002) 0.06555 − 0.036 (− 0.068, − 0.005) 0.02434
  Model 2, β (95% CI) − 0.027 (− 0.060, 0.006) 0.10693 − 0.032 (− 0.060, − 0.004) 0.02394
  Model 3, β (95% CI) − 0.026 (− 0.066, 0.013) 0.19245 − 0.023 (− 0.057, 0.011) 0.19028
 Other race
  Model 1, β (95% CI) − 0.088 (− 0.125, − 0.051) < 0.00001 − 0.016 (− 0.045, 0.012) 0.25417
  Model 2, β (95% CI) − 0.039 (− 0.069, − 0.009) 0.01086 − 0.030 (− 0.056, − 0.004) 0.02367
  Model 3, β (95% CI) − 0.044 (− 0.076, − 0.012) 0.00695 − 0.023 (− 0.053, 0.006) 0.12544

Table 3.  Association between high-density lipoprotein cholesterol and total bone mineral density stratified by
gender, age, and race. Model 1: No adjusted. Model 2: Age and race were adjusted. Model 3: All covariates were
adjusted. The stratification variable itself was not adjusted in the subgroup analysis,

(HDL-C, LDL-C, and TC) with lumber BMD. These three cholesterols indicated an inversive association with
BMD in Chinese postmenopausal ­participants12. Peng Niu et al. confirmed the inverse correlation of HDL-C level
and BMD through their analysis of the MIDUS II s­ tudy21. However, Irene Zolfaroli et al. revealed that HDL-C
level, but not TC and LDL-C, was positively relevant to the femoral neck and lumbar B ­ MD22. Song-Seng Loke
et al. suggested a positive association existed in Taiwanese older women, which was a reverse correlation in men.
This finding reminded the importantce of ­gender23. Furthermore, Lian-Hua Cui et al. recruited 730 rural pre- or
postmenopausal women and revealed no significant connection between HDL-C and B ­ MD24. The reasons for
controversial issues in previous studies may include small sample sizes, insufficient scientific data collection
process, different populations analyzed, and differences in adjusted covariates.These controversial conclusions
prompted us to develop more exploration, and the current study was more plausible. Firstly, 4 cycles of data were
extracted from the NHANES database, which provided a larger sample. Second, the study considered gender,

Scientific Reports | (2023) 13:5792 | https://doi.org/10.1038/s41598-023-32885-x 5

Vol.:(0123456789)
 www.nature.com/scientificreports/

Male Female
Direct HDL cholesterol (mmol/L) β (95% CI), P-value β (95% CI), P-value
12–15 years old
 Fitting by the standard linear model − 0.032 (− 0.052, − 0.012) 0.0022 − 0.012 (− 0.033, 0.009) 0.2507
 Fitting by the two-piecewise linear model
 Turn point of direct HDL cholesterol (mmol/L) 0.93 0.98
 < K, effect1 − 0.158 (− 0.312, − 0.004) 0.0450 0.104 (− 0.112, 0.320) 0.3463
 > K, effect2 − 0.030 (− 0.050, − 0.009) 0.0045 − 0.015 (− 0.036, 0.006) 0.1712
 The effect difference 0.128 (− 0.027, 0.283) 0.1066 − 0.119 (− 0.339, 0.101) 0.2906
 The predicted value of the equation at the folding point 0.983 (0.972, 0.995) 0.994 (0.983, 1.006)
 Log-likelihood ratio test 0.103 0.285
16–19 years old
 Fitting by the standard linear model − 0.028 (− 0.054, − 0.002) 0.0342 − 0.019 (− 0.040, 0.002) 0.0780
 Fitting by the two-piecewise linear model
 Turn point of direct HDL cholesterol (mmol/L) 1.06 1.06
 < K, effect1 − 0.129 (− 0.216, − 0.041) 0.0042 − 0.232 (− 0.344, − 0.120) < 0.0001
 > K, effect2 − 0.009 (− 0.040, 0.021) 0.5480 − 0.003 (− 0.025, 0.020) 0.8228
 The effect difference 0.119 (0.020, 0.219) 0.0192 0.229 (0.111, 0.348) 0.0002
 The predicted value of the equation at the folding point 1.118 (1.108, 1.129) 1.058 (1.047, 1.068)
 Log-likelihood ratio test 0.018  < 0.001

Table 4.  Saturation effect analysis of high-density lipoprotein cholesterol and total bone mineral density. All
covariates were adjusted.

which might be a possible impact. Adolescents, not only females, were included in the study. Third, more poten-
tial variables that influence total BMD were adjusted. We found a negative association between HDL-C and BMD
in a specific population of adolescents. This result enriches the research field more detailed and comprehensive
to provide a more substantial reference for clinical application.
In this study, after adjusting all variables, the female subgroup did not have an association between HDL-C
and BMD. Previous research had also suggested that gender hormones were possibly influencing factors in
HDL-C’s and BMD’s ­correlation25–27. Ke Xu et al. investigated the relationship between gender hormones with
BMD value. They demonstrated that the relationship between sex hormones such as testosterone, estradiol, or
sex hormone-binding globulin levels and BMD were opposite in the boy and girl s­ ubgroups28. In summary, sex
hormones might be the crucial cause of this association. In addition, a saturation analysis showed an inverted
U-shaped curve representing the correlation between HDL-C and total BMD among adolescents aged 16 to 19,
stratified by gender and age. The turning point was 1.06 mmol/L for both male and female adolescents. This
variation tendency could be explained by a dramatic change in metabolic status, environmental risk factors,
and other variables.

Limitation and strengths
Recognizing the correlation between HDL-C and BMD could help physicians better evaluate bone health status.
The advantages of this study are the relatively large sample size, the use of data from public databases which
ensures scientific data collection, and gender and age stratified analysis. however, there were some limitations
in this study. (i) All of the included participants were American. The inversive association might be inconsistent
in other countries. (ii) Even though research had revealed that the sex hormones were influencing factors of
the correlation, the NHANES databases did not contain detailed information about these factors. (iii) A cross-
sectional design cannot confirm the cause-effect connection. Researchers should conduct more prospective
studies to solve the causality.

Conclusion
This study indicated a negative association between HDL-C and total BMD values among male adolescents.
Therefore, adolescents with higher HDL-C concentrations were needed to monitor the value of BMD and prevent
bone loss in the future.

Data availability
Data from the survey is publicly available online at http://​www.​cdc.​gov/​nchs/​nhanes/ for data users worldwide.

Received: 22 December 2022; Accepted: 4 April 2023

References
1. Clynes, M. A. et al. The epidemiology of osteoporosis. Br. Med. Bull. 133(1), 105–117. https://d
​ oi.o
​ rg/1​ 0.1​ 093/b
​ mb/l​ daa00​ 5 (2020).

Scientific Reports | (2023) 13:5792 | https://doi.org/10.1038/s41598-023-32885-x 6

Vol:.(1234567890)
www.nature.com/scientificreports/

2. Makitie, O. Causes, mechanisms and management of paediatric osteoporosis. Nat. Rev. Rheumatol. 9(8), 465–475. https://​doi.​org/​
10.​1038/​nrrhe​um.​2013.​45 (2013).
3. Weaver, C. M., Peacock, M. & Johnston, C. C. Jr. Adolescent nutrition in the prevention of postmenopausal osteoporosis. J. Clin.
Endocrinol. Metab. 84(6), 1839–1843. https://​doi.​org/​10.​1210/​jc.​84.6.​1839 (1999).
4. Kline, G. A., Lix, L. M. & Leslie, W. D. Patient outcomes in the years after a DXA-BMD treatment monitoring test: Improved
medication adherence in some, but too little too late. J. Bone Miner. Res. 36(8), 1425–1431. https://​doi.​org/​10.​1002/​jbmr.​4333
(2021).
5. Chiesa, S. T. & Charakida, M. High-density lipoprotein function and dysfunction in health and disease. Cardiovasc. Drugs Ther.
33(2), 207–219. https://​doi.​org/​10.​1007/​s10557-​018-​06846-w (2019).
6. Chang, T. I., Streja, E. & Moradi, H. Could high-density lipoprotein cholesterol predict increased cardiovascular risk?. Curr. Opin.
Endocrinol. Diabetes Obes. 24(2), 140–147. https://​doi.​org/​10.​1097/​med.​00000​00000​000318 (2017).
7. Hamer, M., O’Donovan, G. & Stamatakis, E. High-density lipoprotein cholesterol and mortality: Too much of a good thing?.
Arterioscler. Thromb. Vasc. Biol. 38(3), 669–672. https://​doi.​org/​10.​1161/​atvba​ha.​117.​310587 (2018).
8. Tian, L. & Yu, X. Lipid metabolism disorders and bone dysfunction-interrelated and mutually regulated (review). Mol. Med. Rep.
12(1), 783–794. https://​doi.​org/​10.​3892/​mmr.​2015.​3472 (2015).
9. Zheng, J. et al. The effect of plasma lipids and lipid-lowering interventions on bone mineral density: A Mendelian randomization
study. J. Bone Miner. Res. 35(7), 1224–1235. https://​doi.​org/​10.​1002/​jbmr.​3989 (2020).
10. Li, G. H. Y. et al. Positive effects of low LDL-C and statins on bone mineral density: An integrated epidemiological observation
analysis and Mendelian randomization study. Int. J. Epidemiol. 49(4), 1221–1235. https://​doi.​org/​10.​1093/​ije/​dyz145 (2020).
11. Maghbooli, Z. et al. Negative correlation of high-density lipoprotein-cholesterol and bone mineral density in postmenopausal
Iranian women with vitamin D deficiency. Menopause 25(4), 458–464. https://​doi.​org/​10.​1097/​gme.​00000​00000​001082 (2018).
12. Zhang, Q. et al. Association between bone mineral density and lipid profile in Chinese women. Clin. Interv. Aging 15, 1649–1664.
https://​doi.​org/​10.​2147/​cia.​S2667​22 (2020).
13. Duran, I. et al. Inverse association of high-density lipoprotein cholesterol concentration with muscle mass in children. Child. Obes.
15(7), 476–484. https://​doi.​org/​10.​1089/​chi.​2019.​0122 (2019).
14. Zipf, G. et al. National health and nutrition examination survey: Plan and operations, 1999–2010. Vital Health Stat. Ser. 1. Prog.
Collect. Proced. 56, 1–37 (2013).
15. Tamoliene, V., Remmel, L., Gruodyte-Raciene, R. & Jurimae, J. Relationships of bone mineral variables with body composition,
blood hormones and training volume in adolescent female athletes with different loading patterns. Int. J. Environ. Res. Public
Health. https://​doi.​org/​10.​3390/​ijerp​h1812​6571 (2021).
16. Stadler, J. T. & Marsche, G. Obesity-related changes in high-density lipoprotein metabolism and function. Int. J. Mol. Sci. https://​
doi.​org/​10.​3390/​ijms2​12389​85 (2020).
17. Xie, R., Huang, X., Liu, Q. & Liu, M. Positive association between high-density lipoprotein cholesterol and bone mineral density
in US adults: The NHANES 2011–2018. J. Orthop. Surg. Res. https://​doi.​org/​10.​1186/​s13018-​022-​02986-w (2022).
18. Madsen, C. M., Varbo, A. & Nordestgaard, B. G. Extreme high high-density lipoprotein cholesterol is paradoxically associated
with high mortality inmen and women: Two prospective cohort studies. Eur. Heart J. 38(32), 2478–2486. https://​doi.​org/​10.​1093/​
eurhe​artj/​ehx163 (2017).
19. Pietrobelli, A., Lee, R. C., Capristo, E., Deckelbaum, R. J. & Heymsfield, S. B. An independent, inverse association of high-density-
lipoprotein-cholesterol concentration with nonadipose body mass. Am. J. Clin. Nutr. 69(4), 614–620 (1999).
20. Tang, Y., Wang, S., Yi, Q., Xia, Y. & Geng, B. High-density lipoprotein cholesterol is negatively correlated with bone mineral density
and has potential predictive value for bone loss. Lipids Health Dis. https://​doi.​org/​10.​1186/​s12944-​021-​01497-7 (2021).
21. Niu, P. et al. Association between HDL-C and bone mineral density: An cross-sectional analysis. Int. J. Gen. Med. 14, 8863–8872.
https://​doi.​org/​10.​2147/​ijgm.​S3349​72 (2021).
22. Zolfaroli, I. et al. Positive association of high-density lipoprotein cholesterol with lumbar and femoral neck bone mineral density
in postmenopausal women. Maturitas 147, 41–46. https://​doi.​org/​10.​1016/j.​matur​itas.​2021.​03.​001 (2021).
23. Loke, S.-S., Chang, H.-W. & Li, W.-C. Association between metabolic syndrome and bone mineral density in a Taiwanese elderly
population. J. Bone Miner. Metab. 36(2), 200–208. https://​doi.​org/​10.​1007/​s00774-​017-​0826-7 (2018).
24. Cui, L. H. et al. Association between bone mineral densities and serum lipid profiles of pre- and post-menopausal rural women
in South Korea. Osteoporos. Int. 16(12), 1975–1981. https://​doi.​org/​10.​1007/​s00198-​005-​1977-2 (2005).
25. Mohamad, N.-V., Soelaiman, I.-N. & Chin, K.-Y. A concise review of testosterone and bone health. Clin. Interv. Aging 11, 1317–
1324. https://​doi.​org/​10.​2147/​cia.​S1154​72 (2016).
26. Khosla, S., Oursler, M. J. & Monroe, D. G. Estrogen and the skeleton. Trends Endocrinol. Metab. 23(11), 576–581. https://​doi.​org/​
10.​1016/j.​tem.​2012.​03.​008 (2012).
27. Jirapinyo, M., Theppisai, U., Manonai, J., Suchartwatnachai, C. & Jorgensen, L. N. Effect of combined oral estrogen/progestogen
preparation (Kliogest((R))) on bone mineral density, plasma lipids and postmenopausal symptoms in HRT-naive Thai women.
Acta Obstet. Gynecol. Scand. 82(9), 857–866. https://​doi.​org/​10.​1034/j.​1600-​0412.​2003.​00185.x (2003).
28. Xu, K., Fu, Y., Cao, B. & Zhao, M. Association of sex hormones and sex hormone-binding globulin levels with bone mineral density
in adolescents aged 12–19 years. Front. Endocrinol. https://​doi.​org/​10.​3389/​fendo.​2022.​891217 (2022).

Acknowledgements
Support from colleagues and friends from Shenzhen Traditional Chinese Medicine Hospital is gratefully
acknowledged.

Author contributions
G-X.W. and J-T.L. designed this study, analyzed, wrote, and revised the manuscript, H-L.L., D-L.L., S-F.C.,
H-X.Z., Z-B.F., and W.X. supervised and reviewed the manuscript. All authors reviewed and approved the final
manuscript, which had not been previously published. The work was not under consideration for publication
elsewhere.

Funding
This study was funded by the National Natural Science Foundation of China (No. 82104759); The Natural Science
Foundation of Guangdong Provincial (No. 2019A1515110108).

Competing interests 
The authors declare no competing interests.

Scientific Reports | (2023) 13:5792 | https://doi.org/10.1038/s41598-023-32885-x 7

Vol.:(0123456789)
 www.nature.com/scientificreports/

Additional information
Correspondence and requests for materials should be addressed to D.-L.L., S.-F.C. or H.-L.L.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note  Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access  This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, 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 changes were made. 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://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.

© The Author(s) 2023

Scientific Reports | (2023) 13:5792 | https://doi.org/10.1038/s41598-023-32885-x 8

Vol:.(1234567890)