Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Horm Metab Res. Author manuscript; available in PMC 2011 June 22.
Published in final edited form as:
PMCID: PMC3120042

Depression and Osteoporosis: A Research Synthesis with Meta-Analysis


Major depressive disorder has been associated with low bone mineral density. The strength of this association, however, varies greatly among studies; the direction of the causative link is still controversial, and the etiology remains unclear. We aimed to confirm this association, assess its magnitude and estimate its clinical relevancy. A total of 535 articles were initially identified and the research synthesis was based on 33 qualified articles. Of these, 25 articles (or 76 %) showed an inverse relationship between major depression or minor depression or depressive symptoms and bone mineral density or bone turnover. Meta-analysis could be performed on 20 of the initially selected 33 articles. Standardized weighted differences in mean AP spine, total femur and femoral neck bone mineral density, each from at least 10 studies, were computed in g/cm 2 and transformed into percent differences. At each site, bone mass was lower in subjects with depression as compared to controls: AP spine bone mineral density was 4.73 % lower (95 % CI −7.28 % to −2.19 %, p < 0.0001; n = 16 studies), total femur bone mineral density was 3.53 % lower (95 % CI −5.66 % to −1.41 %, p < 0.001; n = 13 studies), and femoral neck bone mineral density was 7.32 % lower (95 % CI −10.67 % to −3.96 %; p < 0.0005; n = 8 studies). In conclusion, major depressive disorder was associated with lower bone mineral density at the AP spine, femoral neck, and total femur. The deficits in bone mineral density in subjects with depression are of clinical significance and likely to increase fracture risk over the lifetime of these subjects.

Keywords: bone, fractures, stress, antidepressants, women, evidence-based medicine, leptin


It is increasingly clear that a bidirectional link exists between major depressive disorders (MDDs), several other mood disorders, and various medical conditions, such as osteoporosis and cardiovascular disease [1]. In 2001, we published the first meta-analysis on the relationship between MDDs and osteoporosis [2]. In spite of growing evidence, however, neither depressive symptoms nor MDDs currently appear among the risk factors for osteoporosis in official statements [3]. In the last NIH consensus conference on osteoporosis, depression was not listed as a risk factor for osteoporosis. A 2006 issue of Endocrine News, an official publication of the U.S. Endocrine Society, contained an exhaustive table listing 19 causes of secondary osteoporosis, including bone loss secondary to space flight in astronauts, but failed to mention depression for secondary osteoporosis! [4]. The sustained need to change this incorrect, long-standing perception has prompted us to perform another meta-analysis of published studies regarding bone mass in subjects with depression. The main hypothesis tested in this meta-analysis was whether subjects suffering from depression have lower bone mineral density (BMD) than controls. Furthermore, we examined whether the use of different diagnostic criteria for depression and/or clinical severity, duration, and number of depressive episodes may have any effect on the relationship between depression and bone mass. In the studies considered eligible for our meta-analysis, depression was often defined according to the DSM IV criteria, whereas some of the studies explored the potential association between osteoporosis and depressive symptomatology, defined as an aggregate of symptoms that may or may not have met all the DSM criteria for MDDs.

Contextually to the meta-analysis, we performed a research synthesis on the same topic. We felt that this integrated approach would capture the essence of the problem better than either approach alone. In the last 2–3 years there has been a rising trend in the literature of combining meta-analysis with research synthesis. The potential advantages of this novel approach have not been formally tested.

Materials and Methods

Study identification and retrieval

The PubMed, Embase, Biological Abstracts, PsycInfo, Web of Science, and PASCAL electronic databases were searched by experienced librarians from their inception to December 2007 for relevant human studies. Search terms included depression; depressive disorders; osteoporosis; osteopenia; bone density; bone mineral density; dual energy X-ray absorptiometry; DEXA; and DXA. No language restriction was imposed on the search. Two authors (GC and MC) examined each title generated from the search and identified the most relevant articles for which abstracts were obtained.

Study eligibility criteria

To be eligible for the meta-analysis, the study: a) had to be conducted in subjects with major or minor depression, or depressive symptoms along with a group of appropriate controls; b) should contain BMD data in an extractable form; and c) should be published before December 31, 2007 [524]. Studies that did not meet the above criteria but were still deemed to be relevant to the present article were part of the research synthesis [2537].

Data collection

Two authors (GC and MC) extracted, using ad hoc developed forms, the following information from each article eligible for meta-analysis: author, year and journal of publication, sample size, gender, age, menopausal state, characterization of depression, presence of bone mass measurements, and main findings.

Author contact

In cases of incomplete data, authors were contacted whenever possible. Using up to two electronic mail contacts to the corresponding and/or first author of each eligible article, we requested missing data on BMD. All authors who were contacted replied.


Meta-analysis was performed with the Comprehensive Meta-analysis software (Biostat™, Englewood, NJ, USA) using the random effects model, a more conservative approach than the fixed effects model [38]. Standardized differences were plotted in mean BMD (expressed in g/cm2) with 95 % confidence intervals (CIs) between subjects with MDDs and controls. Weighted mean differences and the associated standard deviations (SDs) of BMD values were transformed into percent differences by using the control population as a reference.

We analyzed the difference in BMD between subjects with depression and controls at the AP spine, total femur, and femoral neck. We also compared studies that used a diagnosis of depression based on DSM III or IV criteria (according to the Diagnostic and Statistical Manual of Mental Disorders) versus studies that used a diagnosis of depression based on other criteria. Additionally, a meta-analysis of studies with exclusively male populations was performed to evaluate whether depression is associated with low BMD in men as well. The potential for publication bias was examined by constructing a funnel plot, in which effect size of each study was plotted against the standard error. For each study we determined the relative weight. Evidence of heterogeneity across studies was examined by using the I2 test [39]. To explain reasons for large differences in the effect size between studies (heterogeneity), a priori hypotheses were developed relating to the methodological quality of the study and the study population: (a) different methodologies for diagnosis of depression (DSM vs. other criteria) and (b) gender interaction with BMD. The criterion required for significance, alpha, was set at 0.05. In addition, classical fail-safe n values, that is, the estimated number of studies that in theory would make the p-value greater than alpha, were calculated for each of these three skeletal sites.


The electronic search initially revealed a total of 535 entries in the databases, of which 33 original articles were found to be eligible for assessing the relationship between depression and osteoporosis, based on analysis of the abstract. Fig. 1 describes the flow of candidates and eligible articles for meta-analysis and research synthesis. Twenty articles were suitable for meta-analysis and are summarized in Table 1 [524]. Of these, 16 articles contained AP spine data [5, 712, 14, 15, 17, 18, 2024], 13 articles contained BMD total femur data [68, 1011, 1315, 1721], and 10 articles assessed BMD at the femoral neck [5, 7, 8, 12, 1518, 20, 23]. Sites with too few articles available for BMD assessment (5 articles for Ward triangle [8, 17, 18, 21, 23], 5 articles for trochanter [8, 17, 18, 21, 23], and 2 articles for radius [7, 23]) were not used for meta-analysis. Thirteen original articles that could not be included in the meta-analyses as they were published after the cutoff date of December 2007 [25, 26] or did not have BMD data in an extractable form [2737] are summarized in Table 2.

Fig. 1
Disposition diagram for bone density/depression studies.
Table 1
Synopsis of 20 eligible articles for meta-analysis of bone mineral density in subjects with depression
Table 2
Synopsis of 13 articles on bone mineral density in subjects with depression not included in meta-analysis

In principle, the best way to address the question of causality between depression and osteoporosis would be to disprove that treating depression has effects on bone mass (i.e., accepting the null hypothesis). Thus, subjects never treated with depression would be randomized to antidepressants or placebo and followed over time with serial measurements of BMD. Obviously this approach is neither ethical nor feasible. It is therefore not surprising that the studies identified for detailed assessment were all observational in nature. We did evaluate every article for quality and made general comments across the manuscript. We, however, did not use a formal quality scoring system, as these systems are usually applicable only to randomized controlled trials [40] and the articles reviewed here were mostly cross-sectional or prospective natural history studies.

Research synthesis: characteristics of studies identified by literature search for assessing a relationship between depression and osteoporosis

Sample size

With regard to sample size, the available studies could be broadly categorized into 2 different kinds. The larger studies (n = 13; nmeta = 6) had more than 500 subjects in total, typically with several hundred depressed subjects and up to several thousand controls and were usually population-based studies of BMD or fractures or clinical trials of osteoporosis [6, 11, 13, 15, 19, 21, 2730, 32, 34, 36]. These studies measured relatively few parameters. The remaining studies (n = 20; nmeta = 14) were smaller in size and usually clinical series of 100 subjects or less, conducted in patient populations with mood disorders [5, 710, 12, 14, 1618, 20, 2226, 31, 33, 35, 37]. Generally these studies measured larger numbers of biological parameters.

Gender/menopausal state

Consistent with the epidemiology of both depression and osteoporosis, the majority of studies (n = 21; nmeta =12) included only women [59, 12, 14, 17, 18, 20, 21, 23, 2527, 2933, 35] and 10 studies (nmeta = 6) included both genders [10, 13, 16, 19, 22, 24, 28, 34, 36, 37]. Only 2 studies focused on men: one on Chinese men (nmeta = 1) [11] and another on white men (nmeta = 1) [15]. Women were all premenopausal in 13 studies (nmeta = 9) [5, 7, 9, 10, 1214, 17, 18, 25, 26, 31, 35], all perimenopausal in 1 study (nmeta= 1) [8], all postmenopausal in 14 studies (nmeta= 6) [6, 16, 1921, 24, 2730, 3234, 37], and 3 studies (nmeta = 2) enrolled both pre- and postmenopausal women [22, 23, 36].


Two studies (nmeta = 1) were conducted in pediatric populations [9, 25]. In 15 studies (nmeta = 12), the mean age of the participants was between 18 and 55 years [5, 7, 8, 10, 1214, 17, 18, 2224, 26, 31, 35, 37] and in 13 studies (nmeta = 6) the mean age of the population was greater than 55 years [6, 11, 15, 16, 19, 21, 2730, 32, 34]. Three studies (nmeta = 1) overlapped the age categories, defined above [20, 33, 36].


Twelve studies had population-based cohorts (nmeta= 7) [68, 11, 15, 19, 21, 25, 27, 28, 30, 32], 2 studies had probability samples (nmeta = 1) [13, 36], 15 studies were conducted in subjects seeking psychiatric care (nmeta = 11) [5, 9, 10, 12, 14, 1618, 2224, 26, 31, 35, 37] and only 3 studies (nmeta = 1) involved subjects seeking medical attention for osteoporosis [20, 29, 33]. For a remaining study (nmeta = 0), the recruitment could not be determined [34].

Geographic location

Twelve studies represented populations from the United States (nmeta = 8) [6, 7, 13, 15, 19, 2123, 25, 27, 30, 36], 11 studies were conducted in Europe (nmeta = 5) [9, 14, 16, 20, 24, 26, 3133, 35, 37], 6 studies (nmeta=5) were conducted in the Middle East [5, 10, 12, 17, 18, 34], 2 studies (nmeta = 1) in the Far East [11, 28], 1 in Australia (nmeta = 1) [8], and 1 clinical trial (nmeta = 0) was international [29].

Diagnostic criteria for depression

Close to half of the studies (n = 16, nmeta = 12) used DSM III or IV criteria for the diagnosis of depressive disorder [5, 7, 8, 10, 1214, 17, 18, 2224, 26, 31, 35, 37]; 9 studies (nmeta = 4) used the Geriatric Depression Scale (GDS) – a 15-item self-report validated scale consisting of 15 yes or no questions regarding symptoms of depression experienced in the previous week [6, 11, 15, 21, 2729, 32, 34]; 2 studies (nmeta = 1) used the Center for Epidemiology Studies Depression Scale (CES-D) – a 20-item scale of symptoms over the preceding week [19, 30]; 1 study (nmeta = 1) used the General Health Questionnaire – a 28-item self-rating questionnaire covering depression, anxiety, somatic factors and physical dysfunctions [20]; 1 study (nmeta = 1) on anorexic girls used the Montgomery-Asberg Depression Scale and the Hamilton depression scale [9]; 1 study (nmeta = 1) used Cornell’s Scale [16]; 1 study (nmeta = 0) used the Children’s Depression Inventory – a 27-item self-reported measure [25]; and another one (nmeta = 0) used the General Well Being Schedule (GWB-S) [36]; the remaining study (nmeta = 0) used the Zung depression scale – a self-reported scale [33].

Research synthesis: assessment of an association between depression and osteoporosis

Effect of depression on BMD

The original study in 1994 by Schweiger [24] reporting a large (15 %) deficit in spine bone mass by CT in patients with major depression has been followed by a growing number of reports assessing an association between osteoporosis and depression [523, 2537]. Out of a total of 33 relevant articles, 25 (or 76 %) showed an inverse relationship between depression and BMD or bone turnover [59, 11, 14, 16, 17, 19, 2337]. As many of these studies have already been described in detail in a previous review article [2], we provide in this review only highlights and focus on potential mechanisms, clinical consequences, and future research questions.

Osteoporotic fractures in subjects with depression

From a public health point of view, it would be critical to determine whether lower bone mass in subjects with depression translates into increased fracture risk. Bone mineral density is an accepted predictor of osteoporotic fractures; however, it does not account for additional factors involved in fracture risk, such as biomechanical bone strength and anatomical properties [41].

Because of the large sample size and the need for long follow-up, very few studies have addressed the question of fracture prevalence in subjects with depression. While depressive symptoms were more common in women with prevalent vertebral fractures in a prospective study, the direction of the causal link (i. e., whether depression in this study was related to fractures or was caused by osteoporosis) was unclear [29]. In a large study on older Mexican-American women, higher levels of depression were predictive of new fractures, suggesting that depression is a true causative factor for fractures [30].

A prospective, population-based study in Norway found that, after adjustment for other known risk factors for osteoporosis, women with the highest level of mental distress were at greater risk for femur fracture (relative risk, 1.95) [42]. Likewise, data from the NHANES I study showed that, after adjusting for age, gender, race, BMI, smoking, alcohol, and physical activity, depressive symptomatology remained predictive of femur fracture with a hazard ratio (HR) of 1.7 (95 % CI: 0.99–2.91) [36]. In addition, in multiple adjusted models, an association between elevated urinary free cortisol (UFC) and subsequent bone fractures was seen in a cohort of generally healthy older subjects [43].

In the Tromso study, a longitudinal cohort study conducted in Norway, women with depression had an adjusted odds ratio (OR) of 2.5 (95 % CI: 1.3–4.9) for sustaining a nonvertebral fracture, and an OR of 3.1 (95 % CI: 1.3–7.2) for an osteoporotic fracture (femur, pelvis, proximal humerus, and wrist/forearm) [44, 45]. However, similar associations were not found in men. Whooley et al. [21] reported that women with depression had a 40 % increased risk for sustaining a nonvertebral fracture compared to women without depression. This association remained strong even after adjusting for various confounding factors and for the increased propensity to falls among women with depression. In this study, the OR for a vertebral fracture was 2.3 (95 % CI 1.6 – 3.2) among women with depression. Interestingly, no significant differences were found in the spine or femur BMD between depressed women and controls in this study.

Evidence for an association between depression and osteoporosis: results of the meta-analyses

Twenty articles with control groups for meta-analysis are summarized in Table 1 [524]. The primary outcome for meta-analysis was the difference in BMD between subjects with MDD and nondepressed controls at the 3 skeletal sites for which there were sufficient number of articles (arbitrarily defined as 10 or more). A number of articles reported BMD data for various skeletal sites: AP spine 16 [5, 712, 14, 15, 17, 18, 2024], total femur/hip 13 [68, 10, 11, 1315, 1721], femoral neck 9 [5, 7, 8, 12, 1618, 20, 23], trochanter 5 [8, 17, 18, 21, 23], Ward triangle 5 [8, 17, 18, 21, 23], and radius 2 [7, 23]. Of note is that, because many articles reported more than one extractable study by bone site, gender, or menopausal state, the number of studies used for meta-analysis is often higher than the number of articles.

Measurement of BMD and adjustment for confounding factors in subjects with depression

Apart from the very first study reporting an association between depression and osteoporosis [24] that used quantitative CT, all other studies in this meta-analysis involved the use of DXA scanning for BMD measurements. Bone mineral density values adjusted for factors known to affect BMD were extractable only from 4 articles [8, 13, 15, 21]. In 2 articles adjustments had been made, but only the unadjusted BMD values were reported [9, 24].

Bone mineral density at AP spine

According to meta-analysis of 18 studies on AP spine BMD data from 16 articles [5, 712, 14, 15, 17, 18, 2024], subjects with depression had 4.73 %. lower bone mass (95 % CI –7.28 to −2.19 %, p < 0.0005) than controls (Fig. 2). The classical fail-safe, defined as the number of nonsignificant studies that would be necessary to reduce the effect size to an nonsignificant value, defined in this study as an effect size of 0.05 was quite large totaling 228 (Table 3), making highly unlikely that the conclusion were the result of publication bias. The deficit in bone mass compared to controls was more accentuated in the 11 reports [5, 7, 8, 10, 12, 14, 17, 18, 2224] that used DSM criteria to diagnose depression (−4.68 %; 95 % CI: −7.81, −1.44; p < 0.004 vs. controls) than in the remaining 5 reports [9, 11, 15, 20, 21] that used non-DSM criteria to diagnose depression (−2.27 %; 95 % CI: −4.41 to −0.14; p < 0.04 vs. controls).

Fig. 2
Meta-analysis forest plots for AP spine BMD, total femur BMD, and femoral neck BMD. a : AP spine BMD: Most studies show lower AP spine BMD in depressed subjects, with 3 studies indicating that BMD was higher in controls. b : Total femur: Most studies ...
Table 3
Qualitative summary of reports on depression and osteoporosis

Bone mineral density at total femur

Meta-analysis of 15 studies [68, 10, 11, 1315, 1721] on BMD at the total femur from 13 articles showed that subjects with depression have 3.53 %. lower bone mass than controls (95 % CI: –5.66 % to –1.41 %, p < 0.001) (Fig. 2) with the classical “failsafe n” equal to 133 studies (Table 3). At this site, the deficit in bone mass was similar in the 7 reports [7, 8, 10, 13, 14, 17, 18] using DSM criteria (−2.69 %; 95 % CI: −5.38 to 0.004; p < 0.005) and in the 6 reports [6, 11, 15, 1921] using other criteria to diagnose depression (−2.37; 95 % CI: −4.69 to −0.044; p < 0.046).

Bone mineral density at femoral neck

Based on 8 studies [5, 7, 12, 1618, 20, 23] on bone mass at the femoral neck from 8 articles, subjects with depression had a 7.32 %. lower bone mass than controls (Fig. 2), with the classical “fail-safe n” equal to 130 studies (Table 3). Compared to the AP spine and the total femur (see above) this site revealed the greatest deficit in subject with depression versus controls (−7.32 %; 95 % CI: −10.67 to −3.96; p < 0.0005). Also of note is that 6 out of these 8 studies used DSM criteria for the diagnosis of depression [5, 7, 12, 17, 18, 23].

Gender differences in BMD

At the AP spine, the deficit in bone mass was significant in women with depression (−5.57 %; 95 % CI: −8.51 to −2.64; p < 0.0005) [5, 710, 12, 14, 17, 18, 20, 21, 23, 24] but not in men with depression (−1.50 %; 95 % CI: −4.28 to 1.28; p < 0.29) [10, 11, 15, 24]. At the total femur, the deficit in bone mass was evident in both genders and was of similar magnitude (women: −3.89 %; 95 % CI: −6.85 to −0.93; p < 0.01 [68, 10, 13, 14, 1721]; men: −3.27 %; 95 % CI 5.34 to −1.19; p < 0.002; [10, 11, 13, 15]). There were not enough studies to perform a formal analysis at the femoral neck of bone mass in men.

Differences in osteocalcin levels

Osteocalcin is a biomarker of bone formation and 8 articles reported on osteocalcin levels [5, 7, 12, 17, 18, 23, 31, 35]. According to meta-analysis of the 6 articles with extractable data [5, 7, 12, 17, 18, 23], there were no differences in osteocalcin levels between subjects with depression and controls (−7.65 %; 95 % CI: −21.23 to 5.93; p < 0.27).

Qualitative global assessment of the studies reviewed

A large proportion of the studies reviewed (25/33) reported an association between depression and low BMD (Table 3). An examination of the 8 remaining studies that confirmed the “null hypothesis” (i. e., no association between depression and osteoporosis) failed to reveal any specific pattern in reference to the age of the study sample, size, menopausal status, or geographic location.


The results of this research synthesis with meta-analysis prove that MDDs are associated with low bone mass and should therefore be considered a risk factor for osteoporosis (Table 3). According to our meta-analysis, BMD in subjects with MDDs was 4.7 % lower at the AP spine, 3.5 % lower at the total femur, and 7.3 % lower at the femoral neck as compared to healthy controls. These differences translate to an approximate 1 SD deficit in spine and 0.6 SD in total femur and are similar or greater than those produced by recognized risk factors for osteoporosis, including life style factors. For example, smoking is associated with an approximately 0.1 SD deficit in BMD for combined skeletal sites [46], lack of exercise (aerobics and weight bearing) with a 0.3 SD deficit in the spine and 0.1 SD deficit in the femur [47], and absence of calcium supplementation with a 0.4 SD deficit in whole body BMD [48].

BMD is a strong predictor of fracture risk. Large, prospective studies conducted mostly in postmenopausal women showed that the absolute risk for fractures increases with increasing age for the same level of BMD. As an example, the NORA study convincingly demonstrated that for a T score of −2.0 to −1.0 there is a four-fold increase in the risk for femur and vertebral fractures in women aged 70 to 79 compared to women aged 50 to 59 [49]. Therefore, the percent loss in BMD observed in subjects with depression likely translates to increasing fracture risk with age, even in the absence of further bone loss. Of note is that the largest deficit in bone mass was observed at the femoral neck in subjects with depression. This important finding makes depression-related osteoporosis a particularly serious medical problem, since femur fractures are far more serious than fractures at other skeletal sites.

The 6 studies available for assessment detected no difference in osteocalcin levels between subjects with depression and controls. More information on changes in biochemical markers of bone turnover is needed to clarify whether the bone deficit in depression is caused by increased bone resorption, decreased bone formation, or both. Such information could guide pharmacological treatments for osteoporosis in this population. In a recently published meta-analysis of 14 studies on the same topic, Wu et al. [50] found a similar relationship of BMD and depression after pooling together skeletal sites into two regions: femur and spine. In the present meta-analysis of 20 studies, we determined the effects of depression on BMD at the AP spine, the total femur, and the femoral neck. In another recent meta-analysis, Yirmia et al. [51] found a similar relationship between depression and decreased BMD; in their study they did not, however, perform subanalyses of males, females and mixed populations as we did here. In addition, unlike Wu et al. [50] and Yirmia et al. [51], we report the effects of depression on BMD in percent differences, a value that is clinically more meaningful and intuitive than the absolute difference in BMD or the effect size. Despite an earlier cutoff date, our meta-analysis includes more studies, with data from a greater number of depressed subjects than the two recently published articles [50, 51].

Potential mechanisms of bone loss in depression

Several endocrine alterations observed during the depressive state may induce bone loss [52]. Depression is associated with alterations of the hypothalamic-pituitary-adrenal (HPA) axis at multiple levels, including altered secretion of hypothalamic corticotropin-releasing hormone (CRH), as indicated by CRH levels in the cerebrospinal fluid, and change in the set point threshold for negative feedback; these changes generally result into hyper-cortisolism, as recently reviewed [52]. In an animal model, increased sympathetic tone induced via chronic and severe stress caused bone loss [53]. Evidence in humans is scanty but a recent survey conducted in Israel reported an association between generalized anxiety disorder and osteoporosis [54]. A secular trend to increased psychosocial stress has been reported which may be contributing to the obesity epidemics as well [55]. Although alterations in leptin levels, often of opposite sign, have been reported in subjects with depression, the effects of hyper-leptinemia on bone mass in humans are not clear [52]. In the ob/ob mouse, a genetic model of leptin deficiency, intracerebrov-entricular leptin administration causes bone loss by inhibiting bone formation [52]. Proinflammatory cytokines are increased in depression and IL-6 is a potent activator of the osteoclast [52]. Both estrogen deficiency in women and androgen deficiency in men may affect bone mass and there is at least theoretical evidence for decreased sexual hormones in both genders during the acute phases of depression. Serotonin transporter receptors are present on the osteoblast and use of antidepressants has been associated with more fractures [52]. Commonly accepted life style risk factors for osteoporosis include smoking, inadequate calcium intake, excessive alcohol intake, and physical inactivity. Although in theory these factors may all contribute to bone loss in subjects with depression [52] at the present time, there is little evidence from studies of depressed subjects that poor lifestyle was more prevalent in subjects with depression than controls.

Recently, an unexpected link between bone and energy metabolism has been reported: physiological amounts of osteocalcin increase the expression of insulin in the pancreatic beta cell, increase glucose utilization in muscle and other peripheral tissues, and stimulate the production of adiponectin by the adipocye [56]. Major depressive disorders are known to increase the risk for type II diabetes with an hazard ratio of 1.62, an effect that does not appear to be mediated by poor health behaviors [57] and osteocalcin levels are inversely related to metabolic syndrome [58] and diabetes [59]. Further studies are warranted in subjects with depression to investigate the clinical implications of these novel pathways linking bone and energy metabolism.

Study Limitations

The studies included in our meta-analysis had several limitations. There was a significant heterogeneity among the various studies in terms of differences in BMD that was most likely real, rather than attributable to measurement error. In the case of DXA, the error is remarkably small (about 1 %). Even after removing the most extreme studies from the meta-analysis for AP spine and total femur, a significant amount of heterogeneity remained (with I2 above 60 %). Possible reasons for the heterogeneity include: a) Differences in methodology. All articles had a cross-sectional design, but some were case series collected prospectively mostly for reasons different from observing the natural history of bone turn-over in depression, whereas others represented retrospective analyses. b) Differences in the diagnosis and definition of depression. As an example, the DSM criteria for depression have evolved over time and were used inconsistently in various studies. c) Differences in age, gender, sample size, geographic location, and use of antidepressant medication among study populations. d) Differences in adjustment for known osteoporosis risk factors (Fig. 3).

Fig. 3
Funnel plots for AP spine BMD, total femur BMD and femoral neck. a: AP Spine BMD: Funnel plot shows slight asymmetry with several outliers, which may be due to different effect sizes of depression in men and women, or depression treatment effects. To ...


Implications for practice

The results of this research synthesis with meta-analysis confirm that major depressive disorder is an important but still unrecognized risk factor for osteoporosis. Premenopausal women with major depression should undergo DXA screening. A similar recommendation may be made for postmenopausal women with depression, especially in the presence of one or more known risk factors for osteoporosis. More studies on bone turn-over are needed in men with major depression, before being able to make specific recommendations about screening by DXA. In addition, given the cost and the fact that DXA involves use of radiation, albeit limited, carefully designed studies of health economics should establish the cost/benefit of large scale screenings. Once a diagnosis of osteoporosis is made in subjects with major depression, DXA measurements should be performed with a frequency based on the current WHO algorithm; this model takes into account the presence of other risk factors and the age of the subjects [60]. As far as the diagnosis of depression is concerned, it is well known that this condition may easily go undiagnosed especially in a primary care setting, and that physicians’ recognition of depression is relatively poor, approximately 30–50 % [62]. Several self-reported screening questionnaires are available that perform quite well and take little time; therefore, we recommend screening tools for depression in any setting (medical orthopedic, ob/gyn) in which a subject with decreased BMD or a nontraumatic fracture may present [62].

Implications for research

Since most of the available evidence comes from cross-sectional studies, we advocate longitudinal controlled studies of BMD and bone turnover to establish causality. Data on fractures in subjects with depression are limited; it would therefore be interesting to see whether fracture risk in subjects with depression is totally accounted for by changes in BMD or whether use of medications, alterations in balance, and other unknown factors may play an additional role. It is important to conclusively establish the potential effect of antidepressant medications on the bone mass. Irrespective of the study on specific population of subjects with depression, studies on bone turnover are needed in men and women with depression. As there is biological evidence to suggest that subjects with post-traumatic stress disorder (PTSD) may be at increased risk for osteoporosis, this question should be further explored with clinical studies. In light of the recent developments in the endocrinology of the adipose tissue [63], the role of leptin on bone mass and changes in osteocalcin levels on energy metabolism in subjects with depression should be examined. Although it is commonly accepted that both depression and osteoporosis have a clear genetic component, the search for a genetic background common to these two conditions remains elusive. Identical twin studies examining the lifetime association between depression and osteoporosis may help shed light on the role played by genetic versus environmental factors in the comorbidity of these two conditions.


This study is supported by the Intramural Program of the National Institute of Diabetes, Digestive, and Kidney Diseases, and the Warren Magnuson, National Institutes of Health Clinical Center.

Glossary of Terms

A forest plot
is a graphical display of the results of a meta-analysis.
The relative weight
of each individual study is a critical information that allows more emphasis (weight) to be given to larger studies, as smaller studies may be more subjected to chance and sampling bias. This method allows avoidance of bias inherent in mere arithmetical average of each individual study result.
The funnel plot
is used primarily to provide a visual aid to detect bias or systematic heterogeneity. In the absence of bias, a symmetrical inverted scatterplot derived by plotting treatment effect (in this particular case the standardized differences in bone mineral density) against a measure of study size (in this case the standard error of bone mineral density). Of note, larger studies tend to have a smaller standard error, as a result of greater precision and to aggregate towards the vertex of the triangle.


1. Evans DL, Charney DS, Lewis L, Golden RN, Gorman JM, Krishnan KR, Nemeroff CB, Bremner JD, Carney RM, Coyne JC, Delong MR, Frasure-Smith N, Glassman AH, Gold PW, Grant I, Gwyther L, Ironson G, Johnson RL, Kanner AM, Katon WJ, Kaufmann PG, Keefe FJ, Ketter T, Laughren TP, Leserman J, Lyketsos CG, McDonald WM, McEwen BS, Miller AH, Musselman D, O’Connor C, Petitto JM, Pollock BG, Robinson RG, Roose SP, Rowland J, Sheline Y, Sheps DS, Simon G, Spiegel D, Stunkard A, Sunderland T, Tibbits P, Jr, Valvo WJ. Mood disorders in the medically ill: scientific review and recommendations. Biol Psychiatry. 2005;58:175–189. [PubMed]
2. Cizza G, Ravn P, Chrousos GP, Gold PW. Depression: a major, unrecognized risk factor for osteoporosis? Trends Endocrinol Metab. 2001;12:198–203. [PubMed]
3. Anonymous Management of osteoporosis in postmenopausal women: 2006 position statement of The North American Menopause Society. Menopause. 2006;13:340–367. quiz 368-349. [PubMed]
4. Kristiansen C. Catching secondary causes of osteoporosis. Endocrine News. 2006;1
5. Altindag O, Altindag A, Asoglu M, Gunes M, Soran N, Deveci Z. Relation of cortisol levels and bone mineral density among premenopausal women with major depression. Int J Clin Pract. 2007;61:416–420. [PubMed]
6. Diem SJ, Blackwell TL, Stone KL, Yaffe K, Haney EM, Bliziotes MM, Ensrud KE. Use of antidepressants and rates of hip bone loss in older women: the study of osteoporotic fractures. Arch Intern Med. 2007;167:1240–1245. [PubMed]
7. Eskandari F, Martinez PE, Torvik S, Phillips TM, Sternberg EM, Mistry S, Ronsaville D, Wesley R, Toomey C, Sebring NG, Reynolds JC, Blackman MR, Calis KA, Gold PW, Cizza G. Premenopausal Osteoporosis Women Alendronate Depression (POWER) Study Group. Low bone mass in premenopausal women with depression. Arch Intern Med. 2007;167:2329–2336. [PubMed]
8. Jacka FN, Pasco JA, Henry MJ, et al. Depression and bone mineral density in a community sample of perimenopausal women: Geelong Osteoporosis Study. Menopause. 2005;12:88–91. [PubMed]
9. Konstantynowicz J, Kadziela-Olech H, Kaczmarski M, Kotowicz MA, Dodd S, Nicholson GC, Berk M. Depression in anorexia nervosa: a risk factor for osteoporosis. J Clin Endocrinol Metab. 2005;90:5382–5385. [PubMed]
10. Ozsoy S, Esel E, Turan MT, Kula M, Demir H, Kartalci S, Kökbudak Z. [Is there any alteration in bone mineral density in patients with depression?] Turk Psikiyatri Derg. 2005;16:77–82. [PubMed]
11. Wong SY, Lau EM, Lynn H, Leung PC, Woo J, Cummings SR, Orwoll E. Depression bone mineral density: is there a relationship in elderly Asian men? Results from Mr. Os (Hong Kong) Osteoporos Int. 2005;16:610–615. [PubMed]
12. Yazici AE, Bagis S, Tot S, Sahin G, Yazici K, Erdogan C. Bone mineral density in premenopausal women with major depression. Joint Bone Spine. 2005;72:540–543. [PubMed]
13. Mussolino ME, Jonas BS, Looker AC. Depression and bone mineral density in young adults: results from NHANES III. Psychosom Med. 2004;66:533–537. [PubMed]
14. Petronijevic MR, Ivkovic M, Paunovic VR, Glisic BC, Petronijevic ND, Stefanovic DZ. Chronic depression as risk factor for osteoporosis. Osteoporosis International. 2004;15:S80.
15. Whooley MA, Cauley JA, Zmuda JM, Haney EM, Glynn NW. Depressive symptoms and bone mineral density in older men. J Geriatr Psychiatry Neurol. 2004;17:88–92. [PubMed]
16. Thomas PDA, Thomas CH. Depression and osteoporosis. Measurements by femoral dual energy absorptiometry in elderly patients with or without depression. Age and Nutrition. 2003;14:17–23.
17. Yazici KM, Akinci A, Sutcu A, Ozcakar L. Bone mineral density in pre-menopausal women with major depressive disorder. Psychiatry Res. 2003;117:271–275. [PubMed]
18. Kavuncu V, Kuloglu M, Kaya A, Sahin S, Atmaca M, Firidin B. metabolism and bone mineral density in premenopausal women with mild depression. Yonsei Med J. 2002;43:101–108. [PubMed]
19. Robbins J, Hirsch C, Whitmer R, Cauley J, Harris T. The association of bone mineral density and depression in an older population. J Am Geriatr Soc. 2001;49:732–736. [PubMed]
20. Reginster JY, Deroisy R, Paul I, Hansenne M, Ansseau M. Depressive vulnerability is not an independent risk factor for osteoporosis in postmenopausal women. Maturitas. 1999;33:133–137. [PubMed]
21. Whooley MA, Kip KE, Cauley JA, Ensrud KE, Nevitt MC, Browner WS. Depression, falls, and risk of fracture in older women. Study of Oste-oporotic Fractures Research Group. Arch Intern Med. 1999;159:484–490. [PubMed]
22. Amsterdam JD, Hooper MB. Bone density measurement in major depression. Prog Neuropsychopharmacol Biol Psychiatry. 1998;22:267–277. [PubMed]
23. Michelson D, Stratakis C, Hill L, Reynolds J, Galliven E, Chrousos G, Gold P. mineral density in women with depression. N Engl J Med. 1996;335:1176–1181. [PubMed]
24. Schweiger U, Deuschle M, Korner A, Lammers CH, Schmider J, Gotthardt U, Holsboer F, Heuser I. Low lumbar bone mineral density in patients with major depression. Am J Psychiatry. 1994;151:1691–1693. [PubMed]
25. Dorn LD, Susman EJ, Pabst S, Huang B, Kalkwarf H, Grimes S. Association of depressive symptoms and anxiety with bone mass and density in ever-smoking and never-smoking adolescent girls. Arch Pediatr Ado-lesc Med. 2008;162:1181–1188. [PMC free article] [PubMed]
26. Petronijevic M, Petronijevic N, Ivkovic M, Stefanoviæ D, Radonjiæ N, Glisiæ B, Ristiæ G, Damjanoviæ A, Paunoviæ V. bone mineral density and high bone metabolism turnover in premenopausal women with unipolar depression. Bone. 2008;42:582–590. [PubMed]
27. Diem SJ, Blackwell TL, Stone KL, Yaffe K, Cauley JA, Whooley MA, Ensrud KE. Study of Osteoporotic Fractures. Depressive symptoms and rates of bone loss at the hip in older women. J Am Geriatr Soc. 2007;55:824–831. [PubMed]
28. Niti M, Ng TP, Kua EH, Ho RC, Tan CH. Depression and chronic medical illnesses in Asian older adults: the role of subjective health and functional status. Int J Geriatr Psychiatry. 2007;22:1087–1094. [PubMed]
29. Silverman SL, Shen W, Minshall ME, Xie S, Moses KH. Prevalence of depressive symptoms in postmenopausal women with low bone mineral density and/or prevalent vertebral fracture: results from the Multiple Outcomes of Raloxifene Evaluation (MORE) study. J Rheumatol. 2007;34:140–144. [PubMed]
30. 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;18:315–322. [PubMed]
31. Kahl KG, Greggersen W, Rudolf S, Stoeckelhuber BM, Bergmann-Koester CU, Dibbelt L, Schweiger U. mineral density, bone turnover, and osteo-protegerin in depressed women with and without borderline personality disorder. Psychosom Med. 2006;68:669–674. [PubMed]
32. Korpelainen R, Korpelainen J, Heikkinen J, Vaananen K, Keinanen-Kiu-kaanniemi S. Lifelong risk factors for osteoporosis and fractures in elderly women with low body mass index--a population-based study. Bone. 2006;39:385–391. [PubMed]
33. Bianchi ML, Orsini MR, Saraifoger S, Ortolani S, Radaelli G, Betti S. Quality of life in post-menopausal osteoporosis. Health Qual Life Outcomes. 2005;3:78. [PMC free article] [PubMed]
34. Cankurtaran M, Halil M, Yavuz BB, Dagli N, Cankurtaran ES, Ariogul S. and concomitant diseases in a Turkish geriatric outpatient setting. Arch Gerontol Geriatr. 2005;40:307–315. [PubMed]
35. Kahl KG, Rudolf S, Dibbelt L, Stoeckelhuber BM, Gehl HB, Hohagen F, Schweiger U. Decreased osteoprotegerin and increased bone turnover in young female patients with major depressive disorder and a lifetime history of anorexia nervosa. Osteoporos Int. 2005;16:424–429. [PubMed]
36. Mussolino ME. Depression and hip fracture risk: the NHANES I epidemiologic follow-up study. Public Health Rep. 2005;120:71–75. [PMC free article] [PubMed]
37. Schweiger U, Weber B, Deuschle M, Heuser I. Lumbar bone mineral density in patients with major depression: evidence of increased bone loss at follow-up. Am J Psychiatry. 2000;157:118–120. [PubMed]
38. Egger M, Smith GD, Phillips AN. Meta-analysis: principles and procedures. BMJ. 1997;315:1533–1537. [PMC free article] [PubMed]
39. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557–560. [PMC free article] [PubMed]
40. Chalmers TC, Smith H, Jr, Blackburn B, Silverman B, Schroeder B, Reitman D, Ambroz A. A method for assessing the quality of a randomized control trial. Control Clin Trials. 1981;2:31–49. [PubMed]
41. Syed Z, Khan A. Bone densitometry: applications and limitations. J Obstet Gynaecol Can. 2002;24:476–484. [PubMed]
42. Forsen L, Meyer HE, Sogaard AJ, Naess S, Schei B, Edna TH. Mental distress and risk of hip fracture. Do broken hearts lead to broken bones? J Epidemiol Community Health. 1999;53:343–347. [PMC free article] [PubMed]
43. Greendale GA, Unger JB, Rowe JW, Seeman TE. The relation between cortisol excretion and fractures in healthy older people: results from the MacArthur studies-Mac. J Am Geriatr Soc. 1999;47:799–803. [PubMed]
44. Sogaard AJ, Joakimsen RM, Tverdal A, Fønnebø V, Magnus JH, Berntsen GK. Long-term mental distress, bone mineral density non-vertebral fractures. The Tromso Study. Osteoporos Int. 2005;16:887–897. [PubMed]
45. Milliken LA, Wilhelmy J, Martin CJ, et al. Depressive symptoms and changes in body weight exert independent and site-specific effects on bone in postmenopausal women exercising for 1 year. J Gerontol A Biol Sci Med Sci. 2006;61:488–494. [PubMed]
46. Ward KD, Klesges RC. A meta-analysis of the effects of cigarette smoking on bone mineral density. Calcif Tissue Int. 2001;68:259–270. [PubMed]
47. Bonaiuti D, Shea B, Iovine R, Negrini S, Robinson V, Kemper HC, Wells G, Tugwell P, Cranney A. Exercise for preventing and treating osteoporosis in postmenopausal women. Cochrane Database Syst Rev. 2002 CD000333. [PubMed]
48. Shea B, Bonaiuti D, Iovine R, Negrini S. Cochrane Review on exercise for preventing and treating osteoporosis in postmenopausal women. Eura Medicophys. 2004;40:199–209. [PubMed]
49. Siris ES, Brenneman SK, Barrett-Connor E, Miller PD, Sajjan S, Berger ML, Chen YT. effect of age and bone mineral density on the absolute, excess, and relative risk of fracture in postmenopausal women aged 50–99: results from the National Osteoporosis Risk Assessment (NORA) Osteoporos Int. 2006;17:565–574. [PubMed]
50. Wu Q, Magnus JH, Liu J, Bencaz AF, Hentz JG. Depression and low bone mineral density: a meta-analysis of epidemiologic studies. Osteoporos Int. 2009;20:1309–1320. [PubMed]
51. Yirmiya R, Bab I. Major depression is a risk factor for low bone mineral density: a meta-analysis. Biol Psychiatry. 2009;66:423–432. [PubMed]
52. Cizza G, Primma S, Csako G. Depression as a risk factor for osteoporosis. Trends Endocrinol Metab. 2009;20:367–373. [PMC free article] [PubMed]
53. Yirmiya R, Goshen I, Bajayo A, Kreisel T, Feldman S, Tam J, Trembovler V, Csernus V, Shohami E, Bab I. Depression induces bone loss through stimulation of the sympathetic nervous system. Proc Natl Acad Sci U S A. 2006;103:16876–16881. [PubMed]
54. Muhsen K, Lipsitz J, Garty-Sandalon N, Gross R, Green MS. Correlates of generalized anxiety disorder: independent of co-morbidity with depression: findings from the first Israeli National Health Interview Survey (2003–2004) Soc Psychiatry Psychiatr Epidemiol. 2008;43:898–904. [PubMed]
55. Siervo M, Wells JC, Cizza G. The contribution of psychosocial stress to the obesity epidemic: an evolutionary approach. Horm Metab Res. 2009;41:261–270. [PMC free article] [PubMed]
56. Fukumoto S, Martin TJ. Bone as an endocrine organ. Trends Endocrinol Metab. 2009;20:230–236. [PubMed]
57. Mezuk B, Eaton WW, Golden SH, Ding Y. The influence of educational attainment on depression and risk of type 2 diabetes. Am J Public Health. 2008;98:1480–1485. [PMC free article] [PubMed]
58. Pittas AG, Harris SS, Eliades M, Stark P, Dawson-Hughes B. Association between serum osteocalcin and markers of metabolic phenotype. J Clin Endocrinol Metab. 2009;94:827–832. [PubMed]
59. Kanazawa I, Yamaguchi T, Yamamoto M, et al. Serum osteocalcin level is associated with glucose metabolism and atherosclerosis parameters in type 2 diabetes mellitus. J Clin Endocrinol Metab. 2009;94:45–49. [PubMed]
60. Blake GM, Fogelman I. The clinical role of dual energy X-ray absorptiometry. Eur J Radiol. 2009 [PubMed]
61. Lowe B, Spitzer RL, Grafe K, et al. Comparative validity of three screening questionnaires for DSM-IV depressive disorders and physicians’ diagnoses. J Affect Disord. 2004;78:131–140. [PubMed]
62. Alexander JL, Richardson G, Grypma L, Hunkeler EM. Collaborative depression care, screening, diagnosis and specificity of depression treatments in the primary care setting. Expert Rev Neurother. 2007;7:S59–S80. [PubMed]
63. Fischer-Posovszky P, Wabitsch M, Hochberg Z. Endocrinology of adipose tissue - an update. Horm Metab Res. 2007;39:314–321. [PubMed]