This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Depression is associated with various physical conditions, including bone metabolism. Most previous studies were cross-sectional, and longitudinal findings regarding osteoporosis or fracture risk remain inconsistent. This study examined the association between mild depressive symptoms and annual changes in bone mineral density (BMD) among women undergoing repeated dual-energy X-ray absorptiometry (DEXA).
Methods
We retrospectively analyzed 1,629 women aged ≥20 years who completed the Patient Health Questionnaire-9 (PHQ-9) and underwent at least two DEXA scans between 2017 and 2023. Mild depressive symptoms were defined as PHQ-9 scores of 5–9. Annualized BMD changes at the lumbar spine, femoral neck, and total femur were calculated. Multivariable linear regression analyses adjusted for age, body mass index, smoking, alcohol consumption, physical activity, and diabetes. Women with PHQ-9 ≥10, only one DEXA scan, T-score ≤–2.5, or missing PHQ-9 data were excluded.
Results
Mild depressive symptoms were not significantly associated with annualized BMD change in the overall cohort. However, among women aged ≥60 years (n=1,047), PHQ-9 scores of 5–9 were associated with more negative annualized BMD changes at the femoral neck (β=–3.26; 95% confidence interval [CI], –5.92 to –0.61; P=0.016) and total femur (β=–2.24; 95% CI, –4.39 to –0.09; P=0.041), but not at the lumbar spine.
Conclusion
In women aged ≥60 years, mild depressive symptoms were negatively associated with annual BMD change. Given the close relationship between femoral BMD and hip fracture, even modest negative shifts may be clinically meaningful in older women.
Depression is among the most prevalent psychiatric disorders, and is commonly comorbid with chronic medical conditions including cardiovascular disease, diabetes, and arthritis [1]. The global prevalence of depressive disorders increased by approximately 88.5% between 1990 and 2021 [2] and, in South Korea, the prevalence of both mild and major depressive symptoms increased between 2014 and 2018 [3]. After the novel coronavirus disease 2019 pandemic, the global prevalence of depression and anxiety more than doubled, with South Korea reporting the highest rate among Organisation for Economic Co-operation and Development (OECD) countries (37%), underscoring the urgent need for proactive diagnosis and management [4].
Osteoporosis is a progressive systemic skeletal disease characterized by reduced bone mass and deterioration of the bone microarchitecture [5]. Bone mineral density (BMD), typically assessed using dual-energy X-ray absorptiometry (DEXA), has a well-established inverse association with fracture risk, and a low(er) BMD is now recognized to be a key determinant of fragility fractures [6,7].
Bone remodeling is a continuous process, and bone loss occurs when bone resorption outpaces bone formation, which commonly occur during aging and menopause. Peak bone mass is generally achieved by the late third decade of life, after which bone loss begins and is influenced by genetic factors, sex, nutrition, pubertal health, endocrine status, and physical activity levels [8,9].
Accumulating evidence suggests that depression may influence bone metabolism. Meta-analyses have reported an association between depression and lower BMD at major skeletal sites, including the lumbar spine and femur. However, findings have been inconsistent across age groups and between the sexes; as such, the relationship remains inconclusive [10]. Moreover, Mendelian randomization studies using large-scale genomic data failed to demonstrate a causal association between depression, BMD, and fracture risk, raising concerns that observational findings may be confounded or reflect reverse causation [11].
To date, most studies examining the association between depression and BMD have been cross-sectional in design, while longitudinal studies have predominantly focused on osteoporosis and fracture outcomes. As such, we aimed to investigate longitudinal changes in BMD among adult women who underwent ≥2 DEXA scans during routine health examinations, and to determine whether depressive symptoms are associated with annualized changes in BMD.
Methods
Ethical consideration
This study was approved by the Institutional Review Board (IRB) of Seoul National University Hospital (IRB approval no., E-2603-001-1721), and the requirement for informed consent was waived due to the retrospective study design.
Study data and participants
This retrospective cohort study included data from women ≥20 years of age, who underwent ≥2 DEXA scans and completed the Patient Health Questionnaire-9 (PHQ-9) to assess depressive symptoms during routine health examinations at Seoul National University Hospital Health Promotion Center (Seoul, South Korea) between January 1, 2017, and May 31, 2023.
Among 3,300 individuals who underwent ≥2 DEXA scans and completed baseline questionnaires during this period, the following were excluded: those with baseline T-scores ≤–2.5 (osteoporosis); those without PHQ-9 data; those with missing values for major covariates (e.g., body mass index [BMI], smoking, alcohol consumption, physical activity, history of diabetes); those with PHQ-9 scores ≥10; and males.
Ultimately, data from 1,629 female participants with PHQ-9 scores ranging from 0 to 9 and without missing covariate data were included in the final analysis (Supplement 1).
Definition of key variables
Assessment of depressive symptoms
Depressive symptoms were assessed using the PHQ-9, a self-administered questionnaire used to evaluate the frequency of depressive symptoms over the preceding 2 weeks, with each item scored from 0 (“not at all”) to 3 (“nearly every day”), yielding a total score ranging from 0 to 27. In this study, the PHQ-9 score was used as the independent variable. For the analysis, PHQ-9 scores were treated as continuous variables within the range of 0 to 9, and as categorical variables, grouped into no depression (score <5) and mild depressive symptoms (score 5–9) groups. Participants with PHQ-9 scores ≥10, consistent with clinically significant depression that may require active treatment, such as selective serotonin reuptake inhibitor (SSRI) therapy and, thereby, potentially affect BMD, were excluded from the final analysis [12]
Assessment of changes in BMD
The primary outcome was the annualized change in BMD (mg/cm2/y) of the lumbar (L) spine (L1, L2, L3, L4, and L1–4), femoral neck, and total femur. BMD was measured using DEXA. For each site, the annualized change was calculated by dividing the difference between the baseline and first follow-up measurements by the number of months between the two scans, and multiplying by 12.
Covariates
Covariates included in the multivariable regression models were as follows. Age was defined as the chronological age at the time of examination. BMI was calculated as weight (kg) divided by height (m) squared (kg/m2). Based on self-reported questionnaires, smoking status was classified as ever-smoker or nonsmoker, whereas alcohol consumption was categorized as drinker or non-drinker. Physical activity was defined according to the World Health Organization (WHO) guidelines as ≥600 metabolic equivalent of task (MET)-min/wk (equivalent to at least 150 minutes moderate-intensity activity or 75 minutes of vigorous-intensity activity per week). A history of diabetes was defined as a self-reported physician diagnosis or current use of antidiabetic medication.
Statistical analysis
Baseline characteristics were compared across the PHQ-9 categories using t-tests for continuous variables and chi-square tests for categorical variables. The association between PHQ-9 scores and annualized changes in BMD was examined using multivariable linear regression, with Model 1 adjusted for age and BMI and Model 2 additionally adjusted for smoking status, alcohol consumption, physical activity, and a history of diabetes. Subgroup analyses were performed among women ≥60 years of age to further evaluate the effects of depressive symptoms on changes in BMD. All statistical analyses were performed using Stata ver. 18.0 (StataCorp LLC), and differences with P<0.05 were considered to be statistically significant.
Results
Participant characteristics
Characteristics of the study population are summarized in Tables 1 and 2. Data from 1,629 women were included, 405 (24.9%) of whom had mild depressive symptoms (PHQ-9 score 5–9). The mean age was 62.6±7.3 years, with no significant difference between the two groups (P=0.111).
Baseline BMD of the lumbar spine, femoral neck, and total femur, as well as annualized changes in BMD, did not differ significantly between the groups (Table 1). The mean baseline lumbar spine BMD was 1.09±0.15 g/cm2, with an annualized change of –6.1±22.9 mg/cm2. The mean femoral neck BMD was 0.84±0.11 g/cm2, with an annualized change of –6.4±18.8 mg/cm2, and total femur BMD was 0.91±0.12 g/cm2, with an annualized change of –6.0±15.3 mg/cm2.
Sociodemographic and lifestyle factors are summarized in Table 2. Participants with normal weight (BMI, 18.5 to <23.0 kg/m2) accounted for 49.7%, while underweight (BMI, <18.5 kg/m2) was observed in 4.1%. The proportions of overweight (BMI, 23.0 to <25.0 kg/m2) and obesity (BMI, ≥25.0 kg/m2) were 24.8% and 21.5%, respectively. No significant between-group differences were observed in BMI distribution (P=0.249) or mean BMI (P=0.792).
The proportion of current or former smokers was significantly higher in the mild depressive symptom group (7.16% vs. 3.92%, P=0.008). There was no significant difference in alcohol consumption (P=0.714). However, the proportion of participants who fulfilled the WHO criteria for physical activity (i.e., ≥600 MET-min/wk) was significantly lower in the mild depressive symptom group than that in the non-depressive group (32.8% vs. 39.5%, P=0.016). The prevalence of diabetes was 8.8% in the overall cohort, with no significant difference between the groups (P=0.485).
Association between PHQ-9 score and annualized change in BMD
When the PHQ-9 score (range, 0–9) was analyzed as a continuous variable (Table 3), no statistically significant association with annualized change in BMD was observed in the lumbar spine (L1–L4), femoral neck, or total femur. The results remained non-significant after adjusting for age and BMI (Model 1), and further adjustment for smoking, alcohol consumption, physical activity, and history of diabetes (Model 2). Overall, depressive symptoms (PHQ-9 score) were not significantly associated with annual changes in BMD at any skeletal site.
Comparison of annualized change in BMD according to PHQ-9 category
When PHQ-9 scores were categorized into non-depressive (score, 0–4 [reference]) and mild depressive symptoms (score, 5–9) groups (Table 4), participants with scores of 5–9 exhibited more negative directions in annualized BMD change at the femoral neck and total femur. However, this association was not statistically significant. There were similar patterns in the unadjusted model. In Model 1, adjusted for age and BMI, β values were approximately –1.5 for the femoral neck and –1.1 for the total femur. After additional adjustment for lifestyle factors (Model 2), the β values were –1.3 and –1.0, respectively, indicating consistent negative directions but without statistical significance. No significant associations were observed in the lumbar spine (L1–L4) in any model.
Subgroup analysis of women ≥60 years of age
A subgroup analysis of women ≥60 years of age, who were considered postmenopausal, was performed, with the results presented in Table 5. In this subgroup (n=1,047), participants with PHQ-9 scores of 5–9 exhibited significantly lower annualized chan in BMD in the femoral neck and total femur than those in the non-depressive group, indicating a negative association between mild depressive symptoms and BMD change. These associations were consistent across unadjusted Model 1 and Model 2 analyses. In Model 2, the β values were –3.26 (95% confidence interval [CI], –5.92 to –0.61; P=0.016) for the femoral neck and –2.24 (95% CI, –4.39 to –0.09; P=0.041) for the total femur. There were no statistically significant associations in the lumbar spine (L1–L4) in any model.
Discussion
The present investigation evaluated the association between mild depressive symptoms (PHQ-9 score 5–9) and annualized changes in BMD among 1,629 adult women who underwent ≥2 DEXA scans and completed the PHQ-9 at the Seoul National University Hospital Health Promotion Center. In the overall cohort, no significant associations were observed between mild depressive symptoms and annualized changes in BMD at the lumbar spine or femoral sites.
However, in the subgroup analysis of women ≥60 years of age, those with mild depressive symptoms demonstrated significantly greater reductions in BMD at the femoral neck and total femur than those in the non-depressive group (PHQ-9 score <5). These associations remained consistent after multivariable adjustment. The femoral neck and total femur, where our findings are clinically relevant, are frequent sites of hip fracture(s), which is associated with substantial morbidity and mortality among older women [13]. Therefore, even modest shifts toward a more negative femoral BMD trajectory may be clinically meaningful in this population. The observed association suggests that mild depressive symptoms may contribute to less favorable femoral bone health patterns over time and act as an additional risk factor for bone loss in the context of postmenopausal estrogen deficiency, which is associated with skeletal vulnerability. In contrast, no significant association was observed in the lumbar spine. This may, in part, be explained by the tendency of DEXA measurements of the lumbar spine to be overestimated or distorted due to degenerative changes and joint hypertrophy in older adults [14].
Although previous findings have been inconsistent, a comparison with previous studies provides important insights. Some investigations have reported lower lumbar spine and femoral BMD among women with depression [15], whereas an analysis of data from the National Health and Nutrition Examination Survey demonstrated a significant association between depression and reduced femoral BMD in young adult men, but not in women [16]. South Korean studies have also yielded mixed results, with one reporting a close association between osteoporosis, depression, and self-rated health among postmenopausal women [17], while another observed an inverse relationship in older men but not in women [18]. Our study adds to this heterogeneous body of evidence by providing further support to the link between depressive symptoms and bone loss, specifically in women ≥60 years of age, thereby reinforcing the potential role of depression as a risk factor for skeletal health among postmenopausal women.
The mechanisms by which depressive symptoms influence bone metabolism are likely to be multifactorial. Depression has been associated with reductions in estrogen and growth hormone/insulin-like growth factor-1 levels, as well as dysregulation of the sympathetic nervous system and inflammatory cytokines, such as interleukin-6, a mediator of bone resorption [15]. Elevated cortisol and interleukin-6 levels, and reduced levels of sex hormones and growth factors collectively contribute to decreased bone formation and enhanced resorption. Given that depression is characterized by recurrent episodes of relapse and remission, hormonal and nutritional abnormalities accompanying exacerbations may lead to cumulative effects on peripheral tissues such as the bone. Because it is difficult to recover from loss in bone density, these intermittent yet progressive alterations may accumulate over time, resulting in sustained bone loss [19,20].
In addition, lifestyle factors associated with depression, such as reduced physical activity, poor appetite, and nutritional deficiencies, may further accelerate bone resorption [15]. Epidemiological evidence indicates that antidepressant use, particularly SSRIs, is associated with lower BMD and increased risk for fracture [12]. Collectively, these findings suggest that depressive symptoms contribute to bone loss in postmenopausal women through a combination of biological and behavioral pathways.
This study had several strengths. It used a large health-screening cohort of South Korean women, applied a standardized assessment of depressive symptoms using the PHQ-9, and used repeated DEXA scans to evaluate longitudinal changes in BMD. By focusing on mild depressive symptoms, we explored the potential skeletal implications of subclinical levels of depression commonly encountered in clinical practice.
Nevertheless, the present study also had several limitations. First, fracture history was not included in the dataset, and the fractures themselves may have influenced changes in BMD. Second, participants with PHQ-9 scores ≥10 were excluded to minimize the confounding effects of pharmacological treatment; however, actual antidepressant use or treatment status could not be verified. Given the established association between antidepressant use (particularly SSRIs) and reduced BMD or increased fracture risk, the absence of medication data may have contributed to residual confounding. Third, postmenopausal status was inferred using an age threshold ≥60 years because detailed menopausal data were unavailable. Given that the mean age of natural menopause among South Korean women is approximately 49 to 50 years [21], we applied a conservative cut-off to minimize potential misclassification; however, residual misclassification cannot be ruled out. In addition, information regarding hormone replacement therapy and other medications that may influence bone metabolism (e.g., corticosteroids) was not available, which may have introduced residual confounding factors. Finally, because this was a retrospective, single-center study, the findings may have limited generalizability and causality cannot be established.
In conclusion, among South Korean women ≥20 years of age, mild depressive symptoms were not significantly associated with annualized changes in BMD. However, among women ≥60 years of age, mild depressive symptoms were significantly associated with a greater decline in BMD in the total femur and femoral neck. These findings suggest that depressive symptoms may be an important consideration in risk assessment for osteoporosis in older women and that routine depression screening could support early identification and management of this risk. Future studies with larger cohorts and longer follow-up periods incorporating data regarding antidepressant use, fracture history, and lifestyle factors are warranted.
Notes
Conflict of interest
No potential conflict of interest relevant to this article was reported.
Funding
None.
Data availability
Data of this research are available from the corresponding author upon reasonable request.
Author contribution
Conceptualization: JB, KM, SEH. Methodology: JB, KM, SEH. Formal analysis: JB, HK. Investigation: JB, HK. Data curation: JB, HK. Visualization: JB. Supervision: KM, SEH. Writing–original draft: JB. Writing–review & editing: all authors. Final approval of the manuscript: all authors.
Baseline and annualized changes in BMD among study participants according to depressive symptom status
Characteristic
Total
PHQ-9 (0–4)
PHQ-9 (5–9)
P-value
Lumbar 1
Baseline L1 BMD (g/cm2)
0.98±0.15
0.98±0.15
0.97±0.14
0.483
Annual L1 BMD change (mg/cm2)
–7.37±29.25
–7.4±29.81
–7.27±27.51
0.937
Lumbar 2
Baseline L2 BMD (g/cm2)
1.04±0.16
1.04±0.16
1.03±0.15
0.340
Annual L2 BMD change (mg/cm2)
–6.62±28.88
–7.15±29.03
–5.01±28.37
0.195
Lumbar 3
Baseline L3 BMD (g/cm2)
1.14±0.17
1.14±0.17
1.13±0.16
0.274
Annual L3 BMD change (mg/cm2)
–5.79±29.17
–5.32±29.78
–7.24±27.22
0.251
Lumbar 4
Baseline L4 BMD (g/cm2)
1.16±0.17
1.17±0.18
1.16±0.16
0.376
Annual L4 BMD change (mg/cm2)
–4.71±30.51
–4.83±30.67
–4.34±30.07
0.780
Lumbar vertebrae 1–4
Baseline BMD (g/cm2)
1.09±0.15
1.09±0.16
1.08±0.14
0.317
Annual BMD change (mg/cm2)
–6.06±22.88
–6.1±23
–5.93±22.54
0.895
Femoral neck
Baseline BMD (g/cm2)
0.84±0.11
0.84±0.11
0.85±0.1
0.704
Annual BMD change (mg/cm2)
–6.41±18.76
–6.03±18.81
–7.58±18.6
0.151
Total femur
Baseline BMD (g/cm2)
0.91±0.12
0.91±0.12
0.91±0.11
0.729
Annual BMD change (mg/cm2)
–5.98±15.28
–5.7±15.33
–6.84±15.12
0.194
Values are presented as mean±standard deviation. P-values were calculated using Student t-test for continuous variables.
BMD, bone mineral density; PHQ-9, Patient Health Questionnaire-9.
Table 2.
Sociodemographic and lifestyle characteristics of study participants according to depressive symptom status
Characteristic
Total
PHQ-9 (0–4)
PHQ-9 (5–9)
P-value
Total
1,629 (100.0)
1,224 (75.1)
405 (24.9)
Age (y)
62.61±7.29
62.78±7.24
62.11±7.4
0.111
BMI (kg/m2)
23.01±3.07
23±2.99
23.05±3.31
0.792
Mean difference of exam (DEXA)
2.26±1.11
2.24±1.09
2.31±1.17
0.248
Obesity (BMI), kg/m2
0.249
<18.5
66 (4.1)
47 (3.8)
19 (4.7)
18.5 to <23.0
809 (49.7)
608 (49.7)
201 (49.6)
23.0 to <25.0
404 (24.8)
316 (25.8)
88 (21.7)
≥25.0
350 (21.5)
253 (20.7)
97 (24.0)
Smoking
0.008
Non-smoker
1,552 (95.3)
1,176 (96.1)
376 (92.8)
Ex or current smoker
77 (4.7)
48 (3.9)
29 (7.2)
Alcohol
0.714
No
1,014 (62.2)
765 (62.5)
249 (61.5)
Yes
615 (37.8)
459 (37.5)
156 (38.5)
PA (MET-min/wk)
0.016
<600
1,012 (62.1)
740 (60.5)
272 (67.2)
≥600
617 (37.9)
484 (39.5)
133 (32.8)
Diabetes
0.485
No
1,486 (91.2)
1,120 (91.5)
366 (90.4)
Yes
143 (8.8)
104 (8.5)
39 (9.6)
Values are presented as number (%) or mean±standard deviation. P-values were calculated using Student t-test for continuous variables and chi-square test for categorical variables.
PHQ-9, Patient Health Questionnaire-9; BMI, body mass index; DEXA, dual-energy X-ray absorptiometry; PA, physical activity; MET, metabolic equivalent of task.
Table 3.
Association between Patient Health Questionnaire-9 score (0–9) and annual mean change in bone mineral density
Multivariable linear regression analyses were performed to assess the association between Patient Health Questionnaire-9 score (0–9) and annualized bone mineral density change.
CI, confidence interval; MET, metabolic equivalent of task.
a)Model 1 was adjusted for age, body mass index.
b)Model 2 additionally included smoking status (non-smoker vs. current/ex-smoker), alcohol consumption (non-drinker vs. drinker), physical activity (<600 MET-min/wk vs. ≥600 MET-min/wk), and diabetes, in addition to the variables included in Model 1.
Table 4.
Association between mild depressive symptoms (Patient Health Questionnaire-9 score 5–9) and annual mean change in bone mineral density
Multivariable linear regression analyses were conducted to compare annualized bone mineral density changes according to depressive symptom categories (Patient Health Questionnaire-9 score <5 vs. 5–9).
CI, confidence interval; Ref, reference; MET, metabolic equivalent of task.
a)Model 1 was adjusted for age, body mass index.
b)Model 2 additionally included smoking status (non-smoker vs. current/ex-smoker), alcohol consumption (non-drinker vs. drinker), physical activity (<600 MET-min/wk vs. ≥600 MET-min/wk), and diabetes, in addition to the variables included in Model 1.
Table 5.
Association between mild depressive symptoms (Patient Health Questionnaire-9 score 5–9) and annual mean change in bone mineral density among postmenopausal women (≥60 years) (n=1,047)
Subgroup analyses among women aged ≥60 years were performed using multivariable linear regression.
CI, confidence interval; Ref, reference; MET, metabolic equivalent of task.
*P<0.05 (Statistical significance).
a)Model 1 was adjusted for age and body mass index.
b)Model 2 included smoking status (non-smoker or ex/current-smokers), daily alcohol consumption (none or current drinks), physical activity <600 MET-min/wk vs. ≥600 MET-min/wk, and diabetes as covariates in Model 1.
References
1. Moussavi S, Chatterji S, Verdes E, Tandon A, Patel V, Ustun B. Depression, chronic diseases, and decrements in health: results from the World Health Surveys. Lancet 2007;370:851-8.
2. Chen XD, Li F, Zuo H, Zhu F. Trends in prevalent cases and disability-adjusted life-years of depressive disorders worldwide: findings from the global burden of disease study from 1990 to 2021. Depress Anxiety 2025;2025:5553491.
3. Priyono D, Kim S. Rising prevalence of subthreshold and major depressive symptom in South Korea: a trend analysis from 2014 and 2018. PLoS One 2025;20:e0320980.
4. Organisation for Economic Co-operation and Development (OECD). Health at a glance 2021: OECD indicators. OECD Publishing; 2021.
5. Curtis EM, Harvey NC, Cooper C. The burden of osteoporosis. In: Harvey NC, Cooper C, editors. Osteoporosis: a lifecourse epidemiology approach to skeletal health. CRC Press; 2018. p. 1-20
6. Marshall D, Johnell O, Wedel H. Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 1996;312:1254-9.
7. Wainwright SA, Marshall LM, Ensrud KE, Cauley JA, Black DM, Hillier TA, et al. Hip fracture in women without osteoporosis. J Clin Endocrinol Metab 2005;90:2787-93.
8. Chadha M, Chaddha R, Divakar H, Kalyan H, Seth S, Shah P. Osteoporosis: epidemiology, pathogenesis, evaluation and treatment. Open J Orthop 2022;12:153-82.
9. Weaver CM, Gordon CM, Janz KF, Kalkwarf HJ, Lappe JM, Lewis R, et al. The National Osteoporosis Foundation’s position statement on peak bone mass development and lifestyle factors: a systematic review and implementation recommendations. Osteoporos Int 2016;27:1281-386.
10. Yuan S, Chen J, Zeng L, Zhou C, Yu S, Fang L. Association of bone mineral density and depression in different bone sites and ages: a meta‐analysis. Food Sci Nutr 2021;9:4780-92.
13. Abrahamsen B, van Staa T, Ariely R, Olson M, Cooper C. Excess mortality following hip fracture: a systematic epidemiological review. Osteoporos Int 2009;20:1633-50.
17. Pang DD, Yoon YS, Yang YJ, Lee ES, Lee JH, Seo YB, et al. Association between osteoporosis and mental health status in postmenopausal women: the Korea National Health and Nutrition Examination Survey (2016 and 2018). Korean J Fam Pract 2021;11:449-56.
18. Oh SM, Kim HC, Ahn SV, Rhee Y, Suh I. Association between depression and bone mineral density in community-dwelling older men and women in Korea. Maturitas 2012;71:142-6.
19. Michelson D, Stratakis C, Hill L, Reynolds J, Galliven E, Chrousos G, et al. Bone mineral density in women with depression. N Engl J Med 1996;335:1176-81.
20. Tolea MI, Black SA, Carter-Pokras OD, Kling MA. Depressive symptoms as a risk factor for osteoporosis and fractures in older Mexican American women. Osteoporos Int 2007;18:315-22.
Association between mild depressive symptoms and annual changes in bone mineral density among South Korean women: a retrospective study
Graphical abstract
Graphical abstract
Association between mild depressive symptoms and annual changes in bone mineral density among South Korean women: a retrospective study
Characteristic
Total
PHQ-9 (0–4)
PHQ-9 (5–9)
P-value
Lumbar 1
Baseline L1 BMD (g/cm2)
0.98±0.15
0.98±0.15
0.97±0.14
0.483
Annual L1 BMD change (mg/cm2)
–7.37±29.25
–7.4±29.81
–7.27±27.51
0.937
Lumbar 2
Baseline L2 BMD (g/cm2)
1.04±0.16
1.04±0.16
1.03±0.15
0.340
Annual L2 BMD change (mg/cm2)
–6.62±28.88
–7.15±29.03
–5.01±28.37
0.195
Lumbar 3
Baseline L3 BMD (g/cm2)
1.14±0.17
1.14±0.17
1.13±0.16
0.274
Annual L3 BMD change (mg/cm2)
–5.79±29.17
–5.32±29.78
–7.24±27.22
0.251
Lumbar 4
Baseline L4 BMD (g/cm2)
1.16±0.17
1.17±0.18
1.16±0.16
0.376
Annual L4 BMD change (mg/cm2)
–4.71±30.51
–4.83±30.67
–4.34±30.07
0.780
Lumbar vertebrae 1–4
Baseline BMD (g/cm2)
1.09±0.15
1.09±0.16
1.08±0.14
0.317
Annual BMD change (mg/cm2)
–6.06±22.88
–6.1±23
–5.93±22.54
0.895
Femoral neck
Baseline BMD (g/cm2)
0.84±0.11
0.84±0.11
0.85±0.1
0.704
Annual BMD change (mg/cm2)
–6.41±18.76
–6.03±18.81
–7.58±18.6
0.151
Total femur
Baseline BMD (g/cm2)
0.91±0.12
0.91±0.12
0.91±0.11
0.729
Annual BMD change (mg/cm2)
–5.98±15.28
–5.7±15.33
–6.84±15.12
0.194
Characteristic
Total
PHQ-9 (0–4)
PHQ-9 (5–9)
P-value
Total
1,629 (100.0)
1,224 (75.1)
405 (24.9)
Age (y)
62.61±7.29
62.78±7.24
62.11±7.4
0.111
BMI (kg/m2)
23.01±3.07
23±2.99
23.05±3.31
0.792
Mean difference of exam (DEXA)
2.26±1.11
2.24±1.09
2.31±1.17
0.248
Obesity (BMI), kg/m2
0.249
<18.5
66 (4.1)
47 (3.8)
19 (4.7)
18.5 to <23.0
809 (49.7)
608 (49.7)
201 (49.6)
23.0 to <25.0
404 (24.8)
316 (25.8)
88 (21.7)
≥25.0
350 (21.5)
253 (20.7)
97 (24.0)
Smoking
0.008
Non-smoker
1,552 (95.3)
1,176 (96.1)
376 (92.8)
Ex or current smoker
77 (4.7)
48 (3.9)
29 (7.2)
Alcohol
0.714
No
1,014 (62.2)
765 (62.5)
249 (61.5)
Yes
615 (37.8)
459 (37.5)
156 (38.5)
PA (MET-min/wk)
0.016
<600
1,012 (62.1)
740 (60.5)
272 (67.2)
≥600
617 (37.9)
484 (39.5)
133 (32.8)
Diabetes
0.485
No
1,486 (91.2)
1,120 (91.5)
366 (90.4)
Yes
143 (8.8)
104 (8.5)
39 (9.6)
Characteristic
Unadjusted
Model 1a)
Model 2b)
β (95% CI)
P-value
β (95% CI)
P-value
β (95% CI)
P-value
Lumbar 1
–0.15 (–0.71 to 0.42)
0.608
–0.12 (–0.68 to 0.43)
0.665
–0.08 (–0.64 to 0.48)
0.786
Lumbar 2
0.15 (–0.41 to 0.71)
0.598
0.19 (–0.36 to 0.75)
0.496
0.18 (–0.38 to 0.74)
0.529
Lumbar 3
–0.54 (–1.1 to 0.03)
0.062
–0.49 (–1.05 to 0.07)
0.083
–0.49 (–1.05 to 0.08)
0.09
Lumbar 4
–0.12 (–0.71 to 0.46)
0.678
–0.11 (–0.7 to 0.48)
0.712
–0.1 (–0.69 to 0.49)
0.74
Lumbar 1–4
–0.17 (–0.61 to 0.28)
0.463
–0.13 (–0.57 to 0.3)
0.545
–0.12 (–0.56 to 0.31)
0.579
Femoral neck
–0.27 (–0.63 to 0.09)
0.145
–0.25 (–0.61 to 0.11)
0.171
–0.21 (–0.58 to 0.15)
0.246
Femur total
–0.22 (–0.52 to 0.07)
0.138
–0.22 (–0.51 to 0.08)
0.147
–0.21 (–0.51 to 0.09)
0.164
Characteristic
Unadjusted
Model 1a)
Model 2b)
β (95% CI)
P-value
β (95% CI)
P-value
β (95% CI)
P-value
Lumbar 1
<5
Ref
Ref
Ref
5–9
0.13 (–3.16 to 3.42)
0.937
0.38 (–2.87 to 3.62)
0.820
0.58 (–2.68 to 3.83)
0.728
Lumbar 2
<5
Ref
Ref
Ref
5–9
2.15 (–1.1 to 5.39)
0.195
2.5 (–0.73 to 5.72)
0.130
2.42 (–0.82 to 5.66)
0.144
Lumbar 3
<5
Ref
Ref
Ref
5–9
–1.92 (–5.2 to 1.36)
0.251
–1.6 (–4.86 to 1.67)
0.337
–1.44 (–4.72 to 1.83)
0.387
Lumbar 4
<5
Ref
Ref
Ref
5–9
0.49 (–2.94 to 3.92)
0.780
0.62 (–2.81 to 4.05)
0.722
0.71 (–2.74 to 4.15)
0.686
Lumbar 1–4
<5
Ref
Ref
Ref
5–9
0.17 (–2.4 to 2.75)
0.895
0.44 (–2.11 to 2.98)
0.737
0.52 (–2.04 to 3.08)
0.689
Femoral neck
<5
Ref
Ref
Ref
5–9
–1.54 (–3.65 to 0.56)
0.151
–1.45 (–3.56 to 0.65)
0.176
–1.3 (–3.41 to 0.82)
0.229
Femur total
<5
Ref
Ref
Ref
5–9
–1.14 (–2.86 to 0.58)
0.194
–1.1 (–2.82 to 0.62)
0.211
–1.03 (–2.76 to 0.69)
0.242
Characteristic
Unadjusted
Model 1a)
Model 2b)
β (95% CI)
P-value
β (95% CI)
P-value
β (95% CI)
P-value
Lumbar 1
<5
Ref
Ref
Ref
5–9
–0.79 (–4.92 to 3.34)
0.708
–0.64 (–4.75 to 3.46)
0.759
–0.56 (–4.67 to 3.55)
0.789
Lumbar 2
<5
Ref
Ref
Ref
5–9
2.27 (–1.81 to 6.36)
0.275
2.47 (–1.61 to 6.55)
0.236
2.24 (–1.85 to 6.33)
0.283
Lumbar 3
<5
Ref
Ref
Ref
5–9
–2.38 (–6.38 to 1.61)
0.242
–2.25 (–6.25 to 1.75)
0.27
–2.15 (–6.15 to 1.85)
0.292
Lumbar 4
<5
Ref
Ref
Ref
5–9
0.79 (–3.55 to 5.12)
0.722
0.82 (–3.54 to 5.17)
0.713
0.72 (–3.65 to 5.09)
0.746
Lumbar 1–4
<5
Ref
Ref
Ref
5–9
–0.11 (–3.18 to 2.96)
0.944
0.02 (–3.05 to 3.09)
0.988
–0.02 (–3.1 to 3.06)
0.991
Femoral neck
<5
Ref
Ref
Ref
5–9
–3.24 (–5.88 to –0.6)
0.016*
–3.3 (–5.95 to –0.65)
0.015*
–3.26 (–5.92 to –0.61)
0.016*
Femur total
<5
Ref
Ref
Ref
5–9
–2.3 (–4.44 to –0.16)
0.035*
–2.34 (–4.49 to –0.19)
0.033*
–2.24 (–4.39 to –0.09)
0.041*
Table 1. Baseline and annualized changes in BMD among study participants according to depressive symptom status
Values are presented as mean±standard deviation. P-values were calculated using Student t-test for continuous variables.
BMD, bone mineral density; PHQ-9, Patient Health Questionnaire-9.
Table 2. Sociodemographic and lifestyle characteristics of study participants according to depressive symptom status
Values are presented as number (%) or mean±standard deviation. P-values were calculated using Student t-test for continuous variables and chi-square test for categorical variables.
PHQ-9, Patient Health Questionnaire-9; BMI, body mass index; DEXA, dual-energy X-ray absorptiometry; PA, physical activity; MET, metabolic equivalent of task.
Table 3. Association between Patient Health Questionnaire-9 score (0–9) and annual mean change in bone mineral density
Multivariable linear regression analyses were performed to assess the association between Patient Health Questionnaire-9 score (0–9) and annualized bone mineral density change.
CI, confidence interval; MET, metabolic equivalent of task.
Model 1 was adjusted for age, body mass index.
Model 2 additionally included smoking status (non-smoker vs. current/ex-smoker), alcohol consumption (non-drinker vs. drinker), physical activity (<600 MET-min/wk vs. ≥600 MET-min/wk), and diabetes, in addition to the variables included in Model 1.
Table 4. Association between mild depressive symptoms (Patient Health Questionnaire-9 score 5–9) and annual mean change in bone mineral density
Multivariable linear regression analyses were conducted to compare annualized bone mineral density changes according to depressive symptom categories (Patient Health Questionnaire-9 score <5 vs. 5–9).
CI, confidence interval; Ref, reference; MET, metabolic equivalent of task.
Model 1 was adjusted for age, body mass index.
Model 2 additionally included smoking status (non-smoker vs. current/ex-smoker), alcohol consumption (non-drinker vs. drinker), physical activity (<600 MET-min/wk vs. ≥600 MET-min/wk), and diabetes, in addition to the variables included in Model 1.
Table 5. Association between mild depressive symptoms (Patient Health Questionnaire-9 score 5–9) and annual mean change in bone mineral density among postmenopausal women (≥60 years) (n=1,047)
Subgroup analyses among women aged ≥60 years were performed using multivariable linear regression.
CI, confidence interval; Ref, reference; MET, metabolic equivalent of task.
P<0.05 (Statistical significance).
Model 1 was adjusted for age and body mass index.
Model 2 included smoking status (non-smoker or ex/current-smokers), daily alcohol consumption (none or current drinks), physical activity <600 MET-min/wk vs. ≥600 MET-min/wk, and diabetes as covariates in Model 1.
