Impact of Mood Disorder History and Bone Health on Cognitive Function Among Men Without Dementia

Abstract

Background:

Poor cognitive function, a major disabling condition of older age, is often considered a prodromal feature of dementia. High mortality and the lack of a cure for dementia have necessitated a focus on the identification of potentially modifiable risk factors. Mental and physical health conditions such as mood disorders and bone loss have been previously linked with poor cognition individually although their combined effect remains largely unknown.

Objective:

Considering the multifactorial nature of dementia pathology, we investigated whether mood disorders, bone health and their interaction are associated with cognitive function in a population-based sample of men.

Methods:

Four hundred and forty-two male participants were drawn from the Geelong Osteoporosis Study. Cognitive function was assessed using the CogState Brief Battery, which measured cognitive performance across four domains and was used to compute overall cognitive function. Mood disorders and hip bone mineral density (BMD) were determined using a semi-structured clinical interview and dual-energy X-ray absorptiometry, respectively.

Results:

Hip BMD (B_coeff = 0.56, 95% CI: [0.07, 1.05], p = 0.025) but not mood disorder (B_coeff = –0.50, 95% CI: [–0.20, 0.10], p = 0.529) was associated with overall cognitive function after accounting for potential confounders. Interaction effects were observed between the two exposures (B_coeff = –1.37, 95% CI: [–2.49, –0.26], p = 0.016) suggesting that individuals without a mood disorder displayed better cognitive performance with increasing BMD, while those with a lifetime history of mood disorder displayed poorer cognitive function with increasing BMD.

Conclusions:

These findings highlight the importance of exploring interactions among potentially modifiable health conditions associated with cognitive function.

Keywords

Alzheimer’s disease bone mineral density cognitive function interaction mood disorder

INTRODUCTION

The number of older adults is growing at an unprecedented rate. By 2050, the older population is expected to total 1.5 billion globally [1]. One of the most debilitating conditions of older age is cognitive decline, which has a profound bearing on individuals and society at large [2]. The decline in cognitive function impacts the instrumental activities of daily living ranging from work to more personal activities and often reflects an underlying dementia pathology [3, 4], which is a devastating neurodegenerative disorder and the seventh leading cause of mortality worldwide [5]. High dementia prevalence and associated disability, compounded by the lack of a cure, have shifted the focus to identifying potential risk factors. Older age, family history of dementia, and APOE ɛ4 genetic risk allele are some of the commonly reported non-modifiable risk factors [6]. On the other hand, potentially modifiable risk factors include poor cognitive reserve, lifestyle factors, and certain mental and physical health conditions [6]. Examples of the latter include mood disorders and poor bone health as they have been previously linked with poor cognitive function. The common mental disorders, depression and anxiety, have been widely reported to be associated with cognitive decline [7 –10]. Contrary to mood disorders, less is known about the relationship between bone health and cognitive function. Although the literature suggests that low bone mineral density (BMD) is associated with a higher risk of cognitive decline [11 –13], most of these studies comprised female populations only, highlighting a gap in our understanding of any underlying relationship among men. Establishing physical and mental health conditions associated with cognitive function may shed light on biological mechanisms, provide therapeutic targets, and help design preventative strategies and screening programs.

Studies so far have investigated individual associations between mood disorders/bone loss and cognitive decline, despite the fact that these conditions display frequent co-occurrence in older adults and share common physiological and environmental processes such as low vitamin levels and hormonal disturbances [14 –16]. Furthermore, several reports have found mood disorders to be associated with low bone mass and quality [14 , 17–20]. Although not much is known about the mechanism behind the association between mood disorders and bone health, it is believed to be driven by pharmacological treatments (e.g., antidepressants and hormonal therapy), physiological factors (e.g., hormonal imbalance and inflammation), and behavioral factors (e.g., physical activity, sun exposure and loneliness) [20 –24]. Given their association, we hypothesized that mood disorders and bone health may interact in predicting an individual’s cognitive function. Hence, this study investigated them as individual factors as well as their combined effects on cognitive function in a healthy aging male population.

METHODS

Study cohort

The Geelong Osteoporosis Study (GOS) is an ongoing, prospective population-based study that was established in the early 1990 s to study the epidemiology of osteoporosis [25]. However, the scope has subsequently expanded to include mental health disorders. Age-stratified cohorts of men and women were selected at random from electoral rolls for the Barwon Statistical Division in south-eastern Australia. A listing on the electoral roll for the region was the inclusion criterion. Exclusion criteria were residence in the area for less than six months and inability to provide informed consent. A total of 1,540 men were recruited from 2001 to 2006 (67% participation) and returned for follow-up 5- and 15-years post-recruitment. The present study comprises a cross-sectional analysis of data acquired from 442 men during the 15-year follow-up phase from 2016 to 2020. Participants provided self-reported data on demographic, health, and lifestyle characteristics in addition to undergoing mental and physical health assessments. All participants provided written informed consent to participate in the study, which was approved by the Human Research Ethics Committee at Barwon Health (00/56-E7) and Deakin University (2013-320). All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committees and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Outcome assessment procedures

The CogState Brief Battery (CBB), a computer-based neuropsychology battery, was used to evaluate cognitive function as described previously [26 –28]. In brief, participants were required to respond to stimuli cards as a part of detection (DET), identification (IDN), one-card learning (OCL), and one-back (OBK) tasks that measured cognitive performance across the following four domains: psychomotor function, visual identification/attention, recognition memory/learning, and working memory, respectively. For each task, both a practice trial and a real test were included that were completed in a quiet room accompanied by a researcher. Scores for DET, IDN, and OBK were calculated by measuring the time taken to answer correctly, which was subsequently normalized using a log₁₀ transformation. OCL task scores, on the other hand, were calculated based on the accuracy of participant response and normalized using an arcsine square-root transformation. Following this, primary measures in the four domains were combined to generate an overall cognitive function (OCF, unitless) score. Overall, lower scores for DET, IDN, and OBK tasks indicate better cognitive performance, whereas lower scores for OCL and OCF indicate worse performance. The current analysis includes individual scores on the four tasks as well as the composite score. In addition to CBB, the Mini-Mental State Examination (MMSE) was also administered to assess the overall cognitive function of the participants [29]. However, due to inherent limitations associated with MMSE, scores from the CBB were used in the analysis [30].

Clinical and demographic measures

The presence of lifetime mood disorders was determined using the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Non-Patient Edition (SCID-I/NP), which has been described previously [31]. Areal BMD (g/cm²) was measured at the femoral neck using dual-energy X-ray absorptiometry (GE Lunar, Prodigy Pro, Madison, WI, USA) [32, 33]. Information on sociodemographic characteristics such as education, domestic partnership status, smoking, and mobility was acquired through self-reports. Based on secondary education completion, education was defined as a nominal factor. Domestic partnership status was defined as living with a partner (coded “1”) or not (coded “0”) and current smokers as those who reported smoking at least one cigarette per day. Participants were classified as physically active if they walked regularly, performed household or other work and engaged in light or vigorous exercise, and sedentary otherwise. Alcohol intake was estimated using a validated food frequency questionnaire [34]. Furthermore, area-based socioeconomic status was determined by linking participants’ addresses to the IRSAD (Index of Relative Socio-economic Advantage and Disadvantage) scores from the Australian Bureau of Statistics census data. IRSAD scores were categorized into quintiles whereby quintile 1 was indicative of the greatest disadvantage.

Blood collection, DNA extraction, and APOE genotyping

Blood plasma was collected in EDTA tubes at the Australian Clinical Labs after overnight fasting and stored at –80 °C until use. Buffy coats were used to isolate genomic DNA using the QIAamp^® DNA Mini Kit (Qiagen, Hilden, Germany) as per the manufacturer’s instructions. Genotyping was performed for the APOE variants rs429358 (ɛ4 allele) and rs7412 (ɛ2 allele) at the Australian Genome Research Facility, Brisbane, using the Agena Bioscience MassARRAY^® platform. The carrier status was defined by the presence of at least one copy of the risk allele. The allelic distribution did not depart from the Hardy-Weinberg equilibrium.

Statistical analyses

Participants were stratified based on median OCF scores and characteristics were compared using Student’s t-tests or Mann–Whitney U tests for continuous variables and Chi-squared tests for categorical variables. Linear regression analyses were conducted to investigate whether a lifetime history of a mood disorder, hip BMD, or two-way interaction effects between them were associated with cognitive performance in the four domains and overall cognitive function. The simultaneous main effects of mood disorder history and hip BMD were investigated first, followed by a model that included the interaction term. The two-way interaction was tested as an exploratory sub-group analysis investigating the association between bone health and cognition based on mood disorder history. This was preferred to generating separate regression models (one for mood disorder history and otherwise) to avoid type I error inflation due to multiple comparisons and enhance statistical power.

Further, multivariable linear regression models adjusted for sociodemographic and lifestyle factors such as age, education, socioeconomic status, mobility, smoking status, alcohol consumption, and APOE ɛ4 carrier status were developed [35, 36]. As APOE ɛ4 risk allele is considered the biggest genetic risk factor for late-onset Alzheimer’s disease [37], it was included as a potential confounding variable. Following this, secondary analyses were conducted wherein mood disorders were restricted to major depressive disorder. Predictive margins plots were also generated to visualize the interaction effects for the covariate-adjusted models, and further stratified analyses were performed to aid the interpretation of interaction effects. Age was dichotomized as less than 55 years, and 55 years and above to avoid collinearity. Corrections for multiple testing were not applied. All statistical analyses were performed using Stata/SE 17.0 and a two-sided p-value ≤0.05 was considered statistically significant.

RESULTS

Participant characteristics

Participant characteristics are presented in Table 1. The study participants had a mean age of 63.5 years (SD 13.0) and more than three quarters had completed secondary education (76.2%). Characteristics were compared between two groups stratified using median OCF scores. Participants who scored below the median value were older (p < 0.001), less likely to complete secondary education (p < 0.001), less physically active (p < 0.001), and less likely to be living with a partner (p = 0.008) as compared to the other group. In addition, they also had a lower MMSE score (p < 0.001) and femoral neck BMD (p = 0.004). Participant characteristics by other cognitive domains are shown in Supplementary Table 1.

Table 1

Characteristics compared between participants stratified based on median OCF scores. Data are presented as mean (SD), n (%) or median (IQR)

Variable		Overall	Low OCF	High OCF	p
		n = 442	n = 219	n = 218
Age [y], mean (SD)		63.5 (13.0)	69.3 (11.7)	57.4 (11.3)	<0.001
Education, n (%)	Secondary education completed	336 (76.2)	145 (66.5)	189 (86.7)	<0.001
	Secondary education not completed	105 (23.8)	73 (33.5)	29 (13.3)
Current smoker, n (%)	Yes	30 (6.8)	14 (6.4)	15 (6.9)	0.838
	No	412 (93.2)	205 (93.6)	203 (93.1)
Alcohol intake (g/day), median (IQR)		11.1 (2.8–25.8)	10.4 (1.3–23.8)	12.6 (4.1–29.4)	0.006
APOEɛ4 carriage, n (%)	Yes	113 (26.9)	54 (25.7)	59 (28.8)	0.483
	No	307 (73.1)	156 (74.3)	146 (71.2)
Mobility, n (%)	Active	338 (76.5)	153 (69.9)	184 (84.4)	<0.001
	Sedentary	104 (23.5)	66 (30.1)	34 (15.6)
Socioeconomic status, n (%)	Quintile 1 (most disadvantaged)	69 (15.8)	45 (20.7)	23 (10.8)	0.055
	Quintile 2	81 (18.6)	37 (17.1)	43 (20.1)
	Quintile 3	102 (23.4)	49 (22.6)	50 (23.4)
	Quintile 4	124 (28.4)	61 (28.1)	63 (29.4)
	Quintile 5	60 (13.8)	25 (11.5)	35 (16.4)
Domestic partnership status, n (%)	Living with a partner	358 (81.0)	167 (76.3)	188 (86.2)	0.008
	Not living with a partner	84 (19.0)	52 (23.7)	30 (13.8)
Lifetime history of mood disorder, n (%)	Yes	105 (23.8)	52 (23.7)	52 (23.9)	0.979
	No	337 (76.2)	167 (76.3)	166 (76.1)
Femoral neck BMD [g/cm²], mean (SD)		1.0 (0.1)	0.9 (0.1)	1.0 (0.1)	0.004
MMSE, mean (SD)		28.7 (1.6)	28.2 (1.8)	29.2 (1.0)	<0.001

Associations between lifetime history of mood disorder, hip BMD, and overall cognitive function

Table 2 summarizes the results obtained from the four regression models; unadjusted main effects only model, unadjusted model with interaction effects, adjusted main effects only model and adjusted model with interaction effects. In the unadjusted main effects only model, increasing BMD at the hip was associated with better overall cognitive function and with every 1 g/cm² rise in the BMD, the average cognitive scores increased by 0.90 units (B_coeff = 0.90, 95% CI: [0.39, 1.42], p = 0.001). Although no association was observed between overall cognitive function and lifetime mood disorder (B_coeff = –0.02, 95% CI: [–0.19, 0.14], p = 0.764), a significant negative interaction was observed between hip BMD and mood disorder (B_coeff = –1.44, 95% CI: [–2.63, –0.24], p = 0.019). Among individuals with no lifetime mood disorder, increasing hip BMD was associated with better cognitive function. However, this association was reversed for people with a lifetime mood disorder as higher BMD was associated with poorer cognitive function. Similar results were obtained in the multivariable model adjusted for age, education, socioeconomic status, mobility, smoking status, alcohol intake, and APOE ɛ4 carrier status. In the adjusted main effects only model, hip BMD alone was positively associated with cognitive function (B_coeff = 0.56, 95% CI: [0.07, 1.05], p = 0.025), while no association was found between lifetime mood disorder and cognitive function (B_coeff = –0.50, 95% CI: [–0.20, 0.10], p = 0.529). However, a significant negative interaction effect (B_coeff = –1.37, 95% CI: [–2.49, –0.26], p = 0.016) was detected suggesting that hip BMD had a differential association with cognitive function based on lifetime mood disorder as shown in Fig. 1A. Further stratified analyses comparing predicted marginal mean differences in cognitive scores while adjusting for confounding factors between people with and without mood disorders at different hip BMD values are shown in Supplementary Table 2. The adjusted mean differences between those with and without mood disorders were mainly significant at the extreme values of hip BMD. At low BMD values, individuals without a mood disorder had a higher overall cognitive function score than those with a mood disorder. However, at higher BMD values, those without a mood disorder displayed a lower score.

Fig. 1

Predictive margins plots displaying interaction between lifetime mood disorder and hip BMD in the covariate-adjusted model. A) OCF (overall cognitive function) scores as the outcome. B) IDN (visual identification) scores as the outcome. C) OBK (working memory) scores as the outcome. D) OCL (recognition memory) scores as the outcome. E) DET (psychomotor function) scores as the outcome.

Table 2

Linear regression analyses assessing the association between hip BMD, lifetime mood disorder and cognitive function

	Unadjusted results				Adjusted results*
Exposure	B_coeff (95% CI) from main effects only model	B_coeff (95% CI) from interaction model	p	Partial η^2†	B_coeff (95% CI) from main effects only model	B_coeff (95% CI) from interaction model	p	Partial η^2†
Overall cognitive function (OCF)
Hip BMD	0.90 (0.39, 1.42)		0.001	0.03	0.56 (0.07, 1.05)		0.025	0.01
Lifetime mood disorder	–0.02 (–0.19, 0.14)		0.764	<0.01	–0.50 (–0.20, 0.10)		0.529	<0.01
Hip BMD#Lifetime mood disorder		–1.44 (–2.63, –0.24)	0.019	0.01		–1.37 (–2.49, –0.26)	0.016	0.01
Visual attention (IDN)
Hip BMD	–0.08 (–0.13, –0.02)		0.005	0.02	–0.049 (–0.101, 0.003)		0.067	0.01
Lifetime mood disorder	–0.0002 (–0.0168, 0.0165)		0.986	<0.01	0.0004 (–0.0158, 0.0167)		0.958	<0.01
Hip BMD#Lifetime mood disorder		0.13 (0.01, 0.26)	0.033	0.01		0.13 (0.02, 0.25)	0.027	0.01
Working memory (OBK)
Hip BMD	–0.12 (–0.19, –0.04)		0.003	0.02	–0.09 (–0.16, –0.01)		0.027	0.01
Lifetime mood disorder	0.01 (–0.02, 0.03)		0.685	<0.01	0.01 (–0.02, 0.03)		0.486	<0.01
Hip BMD#Lifetime mood disorder		0.181 (0.001, 0.361)	0.048	0.01		0.16 (–0.02, 0.33)	0.081	0.01
Recognition memory (OCL)
Hip BMD	0.05 (–0.02, 0.12)		0.139	<0.01	0.01 (–0.06, 0.08)		0.788	<0.01
Lifetime mood disorder	0.01 (–0.01, 0.03)		0.494	<0.01	0.004 (–0.018, 0.026)		0.732	<0.01
Hip BMD#Lifetime mood disorder		–0.15 (–0.31, 0.01)	0.073	0.01		–0.15 (–0.31, 0.01)	0.068	0.01
Psychomotor function (DET)
Hip BMD	–0.12 (–0.19, –0.05)		0.002	0.02	–0.08 (–0.16, –0.01)		0.029	0.01
Lifetime mood disorder	0.01 (–0.01, 0.04)		0.205	<0.01	0.02 (–0.01, 0.04)		0.130	0.01
Hip BMD#Lifetime mood disorder		0.09 (–0.09, 0.26)	0.323	<0.01		0.08 (–0.09, 0.25)	0.367	<0.01

The individual main effects are from a model without the interaction term. ^*Adjusted models included age, education, socioeconomic status, mobility, smoking status, alcohol consumption, and APOE ɛ4 carrier status. ^†An η² value of <0.02 is considered a small effect size.

Associations between lifetime history of mood disorder, hip BMD, and cognitive performance in individual domains

Next, interaction effects between lifetime mood disorder and hip BMD were investigated for the individual cognitive domains using unadjusted main effects only model, unadjusted model with interaction effects, adjusted main effects only model and adjusted model with interaction effects. For the IDN task assessing visual attention, increasing hip BMD was associated with lower scores (and hence better cognitive performance) in the unadjusted main effects only model (B_coeff = –0.08, 95% CI: [–0.13, –0.02], p = 0.005), while this association was approaching significance after adjustment for covariates (B_coeff = –0.049, 95% CI: [–0.101, 0.003], p = 0.067).

Lifetime mood disorder was not found to be associated with visual attention in both unadjusted (B_coeff = –0.0002, 95% CI: [–0.0168, 0.0165], p = 0.986) and adjusted (B_coeff = 0.0004, 95% CI: [–0.0158, 0.0167], p = 0.958) main effects models. However, the interaction between lifetime mood disorder and hip BMD was statistically significant in the unadjusted (B_coeff = 0.13, 95% CI: [0.01, 0.26], p = 0.033) and adjusted models (B_coeff = 0.13, 95% CI: [0.02, 0.25], p = 0.027). The predictive margin plots revealed a negative association between hip BMD and identification task scores among participants without a lifetime mood disorder (Fig. 1B). This association was reversed for participants with a lifetime mood disorder, for whom increasing hip BMD was associated with higher IDN scores, which corresponded with poorer performance on the task. The stratified analysis as shown in Supplementary Table 2 also suggested interaction effects between BMD and mood disorders. At low BMD values, individuals without a mood disorder displayed better cognitive performance than those with a mood disorder.

A similar pattern was observed for working memory wherein increasing hip BMD alone was associated with lower scores on the task (B_coeff = –0.12, 95% CI: [–0.19, –0.04], p = 0.003) and thus, better cognitive performance in the unadjusted main effects only model. This remained the case for participants without a lifetime mood disorder, while a lifetime mood disorder reversed this association as seen in Fig. 1C (B_coeff = 0.181, 95% CI: [0.001, 0.361], p = 0.048). This interaction term, however, was not statistically significant in the covariate-adjusted model (B_coeff = 0.16, 95% CI: [–0.02, 0.33], p = 0.081).

Next, interaction effects were investigated for recognition memory and a differential effect of BMD on recognition memory was observed depending on mood disorder status (Fig. 1D). Individuals without a mood disorder had better cognitive performance with increasing BMD, whereas those with a lifetime mood disorder displayed an inverse association between BMD and recognition memory. However, the interaction term was approaching statistical significance in both the unadjusted (B_coeff = –0.15, 95% CI: [–0.31, 0.01], p = 0.073) and adjusted models (B_coeff = –0.15, 95% CI: [–0.31, 0.01], p = 0.068).

For the final outcome psychomotor function, hip BMD individually was associated with lower task scores and hence, better cognitive performance (B_coeff = –0.08, 95% CI: [–0.16, –0.01], p = 0.029) in the adjusted model but no statistically significant interaction effect was found (B_coeff = 0.08, 95% CI: [–0.09, 0.25], p = 0.367). A closer inspection of the predictive margins plot revealed that although the directionality of association remained unchanged in men with a lifetime mood disorder, the association between hip BMD and psychomotor function became weaker.

Associations between lifetime history of major depressive disorder, hip BMD, and cognitive performance

Analyses were repeated with mood disorders restricted to major depressive disorder; however, the results remained virtually unchanged as shown in Table 3 and Fig. 2. Although statistically significant interactions were observed only in case of overall cognitive function and visual attention, the predictive margin plots suggested trends for interaction effects for all five outcomes.

Fig. 2

Predictive margins plots displaying interaction between lifetime major depressive disorder and hip BMD in the covariate-adjusted model. A) OCF (overall cognitive function) scores as the outcome. B) IDN (visual identification) scores as the outcome. C) OBK (Working memory) scores as the outcome. D) OCL (recognition memory) scores as the outcome. E) DET (psychomotor function) scores as the outcome.

Table 3

Linear regression analyses assessing the association between hip BMD, lifetime major depressive disorder and cognitive function

	Unadjusted results				Adjusted results
Exposure	B_coeff (95% CI) from main effects only model	B_coeff (95% CI) from interaction model	p	Partial η^2†	B_coeff (95% CI) from main effects only model	B_coeff (95% CI) from interaction model	p	Partial η^2†
Overall cognitive function (OCF)
Hip BMD	0.88 (0.36, 1.40)		0.001	0.03	0.52 (0.03, 1.01)		0.038	0.01
Lifetime major depressive disorder	–0.03 (–0.20, 0.14)		0.741	<0.01	–0.03 (–0.19, 0.13)		0.724	<0.01
Hip BMD#Lifetime major depressive disorder		–1.40 (–2.69, –0.12)	0.033	0.01		–1.38 (–2.58, –0.17)	0.025	0.01
Visual attention (IDN)
Hip BMD	–0.08 (–0.13, –0.02)		0.006	0.02	–0.05 (–0.10, 0.01)		0.085	0.01
Lifetime major depressive disorder	0.01 (–0.01, 0.02)		0.514	<0.01	0.005 (–0.013, 0.022)		0.602	<0.01
Hip BMD#Lifetime major depressive disorder		0.136 (0.004, 0.269)	0.044	0.01		0.15 (0.02, 0.28)	0.022	0.01
Working memory (OBK)
Hip BMD	–0.11 (–0.19, –0.04)		0.004	0.02	–0.082 (–0.159, –0.005)		0.037	0.01
Lifetime major depressive disorder	0.01 (–0.02, 0.03)		0.566	<0.01	0.01 (–0.02, 0.03)		0.563	<0.01
Hip BMD#Lifetime major depressive disorder		0.15 (–0.04, 0.35)	0.123	0.01		0.14 (–0.05, 0.33)	0.136	0.01
Recognition memory (OCL)
Hip BMD	0.05 (–0.02, 0.12)		0.148	<0.01	0.01 (–0.06, 0.08)		0.858	<0.01
Lifetime major depressive disorder	0.019 (–0.004, 0.042)		0.102	0.01	0.02 (–0.01, 0.04)		0.128	0.01
Hip BMD#Lifetime major depressive disorder		–0.14 (–0.32, 0.03)	0.108	0.01		–0.13 (–0.31, 0.04)	0.141	0.01
Psychomotor function (DET)
Hip BMD	–0.11 (–0.19, –0.04)		0.003	0.02	–0.076 (–0.150, –0.002)		0.043	0.01
Lifetime major depressive disorder	0.02 (–0.01, 0.04)		0.141	<0.01	0.020 (–0.005, 0.044)		0.111	0.01
Hip BMD#Lifetime major depressive disorder		0.10 (–0.08, 0.28)	0.290	<0.01		0.09 (–0.09, 0.27)	0.341	<0.01

DISCUSSION

In this study, we examined cross-sectional associations between lifetime mood disorder, hip BMD and cognitive function among 442 men without dementia (mean age 63.5 years). Hip BMD alone was positively associated with overall cognitive function, visual attention (only in the unadjusted model), working memory, and psychomotor function. In contrast, the main effects of lifetime mood disorder were not associated with any cognitive outcome; however, mood disorder was an effect modifier in the association between hip BMD and cognition. Although significant interaction effects were only observed for overall cognitive function, visual attention and working memory (only in the unadjusted model), the predictive margins plots suggested interaction trends for all five outcomes. The results remained virtually unchanged when analyses were repeated with mood disorders limited to only major depressive disorder.

These findings are consistent with previous reports that have linked low BMD with a higher risk of cognitive decline [11 –13]. Although not much is known about the association between bone health and cognition, it is postulated to be mediated by physiological factors such as vitamin deficiencies, calcium levels, APOE ɛ4 allele, and hormonal disturbances [38, 39]. However, the studies mentioned above included only female participants and there remains a paucity of literature investigating the relationship of bone health with cognitive function among men. Our results indicated that a lifetime mood disorder alone was not associated with cognitive function. The post-hoc power calculation suggested sufficient power in the study to detect small effect sizes and thus, non-significant associations between mood disorders and cognitive function may just be due to a low effect size. This underscores the importance of investigating other health conditions in relation to cognitive function as there have been previous reports that could not find any longitudinal association between cognitive decline and mood disorders such as depression [40 –42]. Another large cohort study of UK Biobank participants (n = 143,828) reported that participants with a lifetime history of a single episode or moderate recurrent depression features outperformed controls on cognitive measures [43]. Furthermore, other studies have found that depression among women is not associated with subsequent risk of cognitive impairment [35, 44].

Despite this, a lifetime mood disorder displayed an interaction with hip BMD as evidenced in the predictive margins plots. Hip BMD by itself was positively associated with cognitive function but this association was reversed in people with a lifetime mood disorder for whom increasing hip BMD was associated with poorer cognitive performance. Although recent studies have suggested that mood disorders are associated with poor bone health [14 , 17–20], the present study extends these findings by demonstrating how the presence of mood disorders can also negatively impact the relationship between bone health and cognitive function. This negative association observed between BMD and cognitive function among individuals with a lifetime mood disorder could be due to the use of antidepressants or antipsychotics. Previous reports from our group have found certain psychotropic drugs to be associated with poor bone health, providing growing evidence of the central nervous system’s role in regulating bone metabolism [32 , 45–48]. However, further studies are required to illustrate whether psychotropic medication usage per se can drive a negative association between BMD and cognitive function.

To the best of our knowledge, this is the first study that reported an interaction between mood disorders and BMD, and its impact on cognitive function, although further studies are warranted to understand the underlying biochemical nexus. Given the complex etiology of dementia, these findings stress the importance of investigating interaction effects among different risk factors such as mood disorders and poor bone health that display frequent occurrence in older age and often co-exist. These findings also highlight the need for clinicians to be vigilant in identifying mood disorder history as it may modify the association between bone health and cognitive function. However, this was an exploratory study that needs to be replicated across bigger cohorts with a prospective study design to investigate whether these interactions are also useful in predicting longitudinal cognitive trajectories.

Strengths and limitations

Overall, our study was strengthened by the use of an ethnically homogenous population-based cohort where participants were drawn at random from the general population and did not comprise individuals with severe cognitive impairment or dementia. In addition, the exposures were determined based on semi-structured clinical interviews and objective measures rather than self-reports. This was a one-of-a-kind study that revealed that interaction between modifiable health conditions can affect cognitive function among men, independent of lifestyle and sociodemographic factors.

A major limitation of our study was the lack of female participants in our study; however, we are collecting similar data for women. Our findings may not be generalizable to other populations and need to be investigated in ethnically diverse cohorts. Due to the cross-sectional study design, we could not examine any temporal associations or address causality. As the GOS is an ongoing prospective study, we plan to re-assess cognitive function among these participants in the next follow-up phase. Furthermore, as this was an exploratory study, corrections for testing multiple outcomes were not applied. However, future confirmatory studies with adjustments for multiple comparisons are expected. Another limitation of our study was the lack of brain imagingdata.

In conclusion, the findings lend support to the hypothesis that an interaction between mood disorders and bone health exists that could be accounted for in establishing risk factors for cognitive decline. However, further research is required to understand the underlying biological mechanisms driving these interactions. The present study also underlines the need to be vigilant for the co-occurrence of mood disorders and poor bone health in the older population. Given the non-significant association observed between mood disorders and cognitive function, exploring other health conditions as risk factors for cognitive decline becomes increasingly pertinent. These potentially modifiable risk factors may enable the identification of high-risk groups for whom early interventions can be made before irreversible neurological damage occurs.

Footnotes

ACKNOWLEDGMENTS

We acknowledge the men who participated in the study and the staff who contributed to data collection. We would also like to thank the Australian Genome Research Facility for conducting the genotyping analysis. Finally, we thank Professor Graham Giles of the Cancer Epidemiology Centre of The Cancer Council Victoria, for his permission to use the Dietary Questionnaire for Epidemiological Studies (Version 2), Melbourne: The Cancer Council Victoria 1996.

FUNDING

The Geelong Osteoporosis Study was funded by the National Health and Medical Research Council (NHMRC) Australia [Grant Nos. 299831, 628582]. K.M. was supported by Deakin University Postgraduate Research Scholarship (DUPRS). L.J.W. is supported by an NHMRC Emerging Leadership Fellowship (1174060). V.B.G. is supported by NHMRC (RM34909). The funding bodies had no involvement in the study design, data collection, data analysis, and preparation and submission of the manuscript.

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

The genetic data analyzed in this study are provided in . Other participant data are available upon reasonable request to the senior authors for non-commercial academic studies but may be subject to some restrictions according to consent and confidentiality.

The supplementary material is available in the electronic version of this article: .

References

Novotny

, Klimova

, Valis

(2018) Nitrendipine and dementia: Forgotten positive facts? Front Aging Neurosci 10, 418.

Singh-Manoux

, Kivimaki

, Glymour

, Elbaz

, Berr

, Ebmeier

, Ferrie

, Dugravot

(2012) Timing of onset of cognitive decline: Results from Whitehall II prospective cohort study. BMJ 344, d7622.

Johansson

, Marcusson

, Wressle

(2015) Cognitive impairment and its consequences in everyday life: Experiences of people with mild cognitive impairment or mild dementia and their relatives. Int Psychogeriatr 27, 949–958.

Deary

, Corley

, Gow

, Harris

, Houlihan

, Marioni

, Penke

, Rafnsson

, Starr

(2009) Age-associated cognitive decline. Br Med Bull 92, 135–152.

GBD 2019 Collaborators (2021) Global mortality from dementia: Application of a new method and results from the Global Burden of Disease Study 2019. Alzheimers Dement (N Y) 7, e12200.

Sosa-Ortiz

, Acosta-Castillo

, Prince

(2012) Epidemiology of dementias and Alzheimer’s disease. Arch Med Res 43, 600–608.

Mehta

, Thandavan

, Mohebbi

, Pasco

, Williams

, Walder

, Ng

, Gupta

(2022) Depression and bone loss as risk factors for cognitive decline: A systematic review and meta-analysis. Ageing Res Rev 76, 101575.

Cherbuin

, Kim

, Anstey

(2015) Dementia risk estimates associated with measures of depression: A systematic review and meta-analysis. BMJ Open 5, e008853.

Gallagher

, Kiss

, Lanctot

, Herrmann

(2016) Depressive symptoms and cognitive decline: A longitudinal analysis of potentially modifiable risk factors in community dwelling older adults. J Affect Disord 190, 235–240.

10.

John

, Patel

, Rusted

, Richards

, Gaysina

(2019) Affective problems and decline in cognitive state in older adults: A systematic review and meta-analysis. Psychol Med 49, 353–365.

11.

Laudisio

, Fontana

, Rivera

, Ruggiero

, Bandinelli

, Gemma

, Ferrucci

, Incalzi

(2016) Bone mineral density and cognitive decline in elderly women: Results from the InCHIANTI study. Calcif Tissue Int 98, 479–488.

12.

Lui

, Stone

, Cauley

, Hillier

, Yaffe

(2003) Bone loss predicts subsequent cognitive decline in older women: The study of osteoporotic fractures. J Am Geriatr Soc 51, 38–43.

13.

Peng

, Li

, Guan

, Wang

, Huang

(2016) A study of cognitive functions in female elderly patients with osteoporosis: A multi-center cross-sectional study. Aging Ment Health 20, 647–654.

14.

Schweiger

, Schweiger

, Huppe

, Kahl

, Greggersen

, Fassbinder

(2016) Bone density and depressive disorder: A meta-analysis. Brain Behav 6, e00489.

15.

Belanoff

, Gross

, Yager

, Schatzberg

(2001) Corticosteroids and cognition. J Psychiatr Res 35, 127–145.

16.

Piscitelli

, Metozzi

, Benvenuti

, Bonamassa

, Brandi

, Cavalli

, Colli

, Fossi

, Parri

, Giolli

, Tanini

, Fasano

, Di Tanna

, Brandi

(2012) Connections between the outcomes of osteoporotic hip fractures and depression, delirium or dementia in elderly patients: Rationale and preliminary data from the CODE study. Clin Cases Miner Bone Metab 9, 40–44.

17.

Williams

, Pasco

, Hodge

, Berk

(2016) Is there a nexus between mental and bone health? Aust N Z J Psychiatry 50, 829–830.

18.

Williams

, Pasco

, Jacka

, Henry

, Dodd

, Berk

(2009) Depression and bone metabolism. A review. Psychother Psychosom 78, 16–25.

19.

Williams

, Stuart

, Berk

, Brennan-Olsen

, Hodge

, Quirk

, Koivumaa-Honkanen

, Honkanen

, Heikkinen

, Chandrasekaran

, Cleminson

, Pasco

(2022) Bipolar disorder and bone health: A case-control study. J Affect Disord 308, 39–43.

20.

Mollard

, Bilek

, Waltman

(2018) Emerging evidence on the link between depressive symptoms and bone loss in postmenopausal women. Int J Womens Health 10, 1–9.

21.

Kashfi

, Abdollahi

, Hassanzadeh

, Mokarami

, Khani Jeihooni

(2022) The relationship between osteoporosis and depression. Sci Rep 12, 11177.

22.

Laudisio

, Marzetti

, Cocchi

, Bernabei

, Zuccala

(2008) Association of depressive symptoms with bone mineral density in older men: A population-based study. Int J Geriatr Psychiatry 23, 1119–1126.

23.

Vestergaard

, Prieto-Alhambra

, Javaid

, Cooper

(2013) Fractures in users of antidepressants and anxiolytics and sedatives: Effects of age and dose. Osteoporos Int 24, 671–680.

24.

Yuan

, Chen

, Zeng

, Zhou

, Yu

, Fang

(2021) Association of bone mineral density and depression in different bone sites and ages: A meta-analysis. Food Sci Nutr 9, 4780–4792.

25.

Pasco

, Nicholson

, Kotowicz

(2012) Cohort profile: Geelong Osteoporosis Study. Int J Epidemiol 41, 1565–1575.

26.

Sui

, Holloway-Kew

, Hyde

, Williams

, Leach

, Pasco

(2020) Muscle strength and gait speed rather than lean mass are better indicators for poor cognitive function in older men. Sci Rep 10, 10367.

27.

Sui

, Williams

, Holloway-Kew

, Hyde

, Anderson

, Tembo

, Addinsall

, Leach

, Pasco

(2021) Skeletal muscle density and cognitive function: A cross-sectional study in men. Calcif Tissue Int 108, 165–175.

28.

Sui

, Williams

, Holloway-Kew

, Hyde

, Leach

, Pasco

(2022) Associations between muscle quality and cognitive function in older men: Cross-sectional data from the Geelong Osteoporosis Study. J Clin Densitom 25, 133–140.

29.

Pasco

, Stuart

, Sui

, Holloway-Kew

, Hyde

, Tembo

, Rufus-Membere

, Kotowicz

, Williams

(2021) Dynapenia and low cognition: A cross-sectional association in postmenopausal women. J Clin Med 10, 173.

30.

Devenney

, Hodges

(2017) The Mini-Mental State Examination: Pitfalls and limitations. Pract Neurol 17, 79–80.

31.

Williams

, Jacka

, Pasco

, Coulson

, Quirk

, Stuart

, Berk

(2016) The prevalence and age of onset of psychiatric disorders in Australian men. Aust N Z J Psychiatry 50, 678–684.

32.

Chandrasekaran

, Pasco

, Stuart

, Brennan-Olsen

, Berk

, Hodge

, Samarasinghe

, Williams

(2021) Anticonvulsant use and bone health in a population-based study of men and women: Cross-sectional data from the Geelong Osteoporosis Study. BMC Musculoskelet Disord 22, 172.

33.

Rufus-Membere

, Holloway-Kew

, Diez-Perez

, Kotowicz

, Pasco

(2021) Associations between bone material strength index, calcaneal quantitative ultrasound, and bone mineral density in men. J Endocr Soc 5, bvaa179.

34.

Giles

, Ireland

(1996) Dietary questionnaire for epidemiological studies (version 2). The Cancer Council Victoria, Melbourne.

35.

, Niti

, Zaw

, Kua

(2009) Depressive symptoms and incident cognitive impairment in cognitively well-functioning older men and women. J Am Geriatr Soc 57, 1058–1063.

36.

Kim

(2019) Effect of depressive disorder on cognitive decline. Arch Gerontol Geriatr 83, 211–216.

37.

Liu

, Liu

, Kanekiyo

, Xu

, Bu

(2013) Apolipoprotein E and Alzheimer disease: Risk, mechanisms and therapy. Nat Rev Neurol 9, 106–118.

38.

Downey

, Young

, Burton

, Graham

, Macfarlane

, Tsapakis

, Tsiridis

(2017) Dementia and osteoporosis in a geriatric population: Is there a common link? World J Orthop 8, 412–423.

39.

Bednarski

, Gasińska

, Rucińska

, Turżańska

(2014) Alzheimer’s disease and osteoporosis – correlation increasing the risk of life-threatening fractures? MEDtube Sci 2, 32–38.

40.

Bassuk

, Berkman

, Wypij

(1998) Depressive symptomatology and incident cognitive decline in an elderly community sample. Arch Gen Psychiatry 55, 1073–1081.

41.

Bunce

, Batterham

, Mackinnon

, Christensen

(2012) Depression, anxiety and cognition in community-dwelling adults aged 70 years and over. J Psychiatr Res 46, 1662–1666.

42.

Ganguli

, Du

, Dodge

, Ratcliff

, Chang

(2006) Depressive symptoms and cognitive decline in late life: A prospective epidemiological study. Arch Gen Psychiatry 63, 153–160.

43.

Cullen

, Nicholl

, Mackay

, Martin

, Ul-Haq

, McIntosh

, Gallacher

, Deary

, Pell

, Evans

, Smith

(2015) Cognitive function and lifetime features of depression and bipolar disorder in a large population sample: Cross-sectional study of 143,828 UK Biobank participants. Eur Psychiatry 30, 950–958.

44.

Potvin

, Forget

, Grenier

, Preville

, Hudon

(2011) Anxiety, depression, and 1-year incident cognitive impairment in community-dwelling older adults. J Am Geriatr Soc 59, 1421–1428.

45.

Rauma

, Pasco

, Berk

, Stuart

, Koivumaa-Honkanen

, Honkanen

, Hodge

, Williams

(2015) The association between use of antidepressants and bone quality using quantitative heel ultrasound. Aust N Z J Psychiatry 49, 437–443.

46.

Williams

, Berk

, Hodge

, Kotowicz

, Stuart

, Chandrasekaran

, Cleminson

, Pasco

(2018) Selective serotonin reuptake inhibitors (SSRIs) and markers of bone turnover in men. Calcif Tissue Int 103, 125–130.

47.

Williams

, Henry

, Berk

, Dodd

, Jacka

, Kotowicz

, Nicholson

, Pasco

(2008) Selective serotonin reuptake inhibitor use and bone mineral density in women with a history of depression. Int Clin Psychopharmacol 23, 84–87.

48.

Chandrasekaran

, Stuart

, Pasco

, Brennan-Olsen

, Berk

, Hodge

, Samarasinghe

, Williams

(2021) Anticonvulsant use and fracture: A case-control study. J Musculoskelet Neuronal Interact 21, 422–428.