Emotional and physical health in older persons: a role for vitamin D?
de Koning, E.J.
2020
document version
Publisher's PDF, also known as Version of record
Link to publication in VU Research Portal
citation for published version (APA)
de Koning, E. J. (2020). Emotional and physical health in older persons: a role for vitamin D?.
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal ?
Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
E-mail address:
Chapter 6
The relationship between serum 25(OH)D levels
and anxiety symptoms in older persons: results
from the Longitudinal Aging Study Amsterdam
Elisa J. de Koning*, Lotte Verweij*, Paul Lips, Aartjan T.F. Beekman, Hannie C. Comijs, Natasja M. van Schoor * Authors contributed equally to the manuscript Published in:
ABSTRACT
Objectives
Low serum 25-hydroxyvitamin D levels (25(OH)D <50 nmol/L) are common in older persons and associated with depressive symptoms. Depression and anxiety are highly interrelated, but only very few studies examined the association between 25(OH)D and anxiety. This study investigated whether 25(OH)D levels are related to anxiety symptoms in older persons, both cross-sectionally and over time.
Methods
Data from two samples of a large population-based cohort study were used (sample 1: N=1259, 64-88 years; sample 2: N=892, 60-98 years). Anxiety symptoms were measured with the Hospital Anxiety and Depression Scale - Anxiety subscale at baseline and after three years; serum 25(OH)D was measured at baseline. Cross-sectional and longitudinal relationships between 25(OH)D and anxiety were examined using logistic regression analysis, taking into account relevant confounding variables.
Results
Of the participants, 48.0% (sample 1) and 26.4% (sample 2) had 25(OH)D levels <50 nmol/L, whereas 8.1% (sample 1) and 6.5% (sample 2) had clinically relevant anxiety symptoms. Cross-sectionally, persons with 25(OH)D <50 nmol/L experienced more anxiety symptoms than persons with 25(OH)D ≥50 nmol/L (sample 1: OR= 1.55; 95% CI: 1.03, 2.32, p=.035; sample 2: OR= 1.74; 95% CI: 1.03, 2.96, p=.040). However, after adjustment for demographic and lifestyle variables and depressive symptoms, significant associations were no longer observed (p=.25-.72). Similarly, 25(OH)D levels were not significantly related to anxiety symptoms after three years in both samples.
Conclusions
After adjustment for confounding, there was no cross-sectional or longitudinal association between 25(OH)D levels and anxiety symptoms, independently from depression, in two large samples of older persons.
6.1 INTRODUCTION
Inadequate serum 25-hydroxyvitamin D (25(OH)D) levels are commonly seen in older persons [1]. In the Netherlands, vitamin D deficiency (defined as serum 25(OH)D <50 nmol/L) [2] occurs in about 50% of independently living older persons [3], with usually lower levels in women than in men [4]. Vitamin D deficiency can be caused by inadequate exposure to sunlight, a decreasing ability of the aging skin to synthesize vitamin D and/or a deficiency in nutrition [1]. Low levels of serum 25(OH)D have been associated with various health consequences in older persons, including reduced physical functioning [5, 6], cardiovascular disease [7] and depression [8, 9].
Anxiety symptoms are also common in older adults. Anxiety is associated with increased disability, reduced wellbeing and lower quality of life [10]. In addition, more severe anxiety symptoms are associated with long-term stress and reduced cognitive functioning in older persons [11]. Anxiety symptoms often coexist with depression [12]. Previous research concluded that 43% of older persons with depression and 15% without depression reported anxiety symptoms [12]. Although much overlap exists between risk factors for anxiety and depression, they are nonetheless distinct conditions [13]. Several studies observed that reduced serum 25(OH)D levels are associated with more depressive symptoms [8, 9]. However, a very limited number of studies examined the relationship between 25(OH)D levels and anxiety. From animal research it is known that a deficiency of the vitamin D receptor (VDR) affects emotional behavior and increases anxiety-like behaviors in mice [14]. In humans, Armstrong et al. assessed the association in patients with fibromyalgia and concluded that serum 25(OH)D deficiency is more common in patients with higher levels of anxiety and depressive symptoms [15]. However, when theyexamined the association between serum 25(OH)D and anxiety separately from depressive symptoms, results were no longer statistically significant. Furthermore, Chao et al. found that persons with lower levels of serum 25(OH)D reported significantly more symptoms of anxiety and depression on a health-related quality of life instrument [16].
As it is not clear whether reduced serum 25(OH)D levels are associated with anxiety, independently from depression,the present study examined both the cross-sectional and longitudinal relationship between serum 25(OH)D and anxiety symptoms in two large population-based samples of older persons, with the second sample as a replication sample.
6.2 METHODS
6.2.1 Design and participants
Data of the Longitudinal Aging Study Amsterdam (LASA) were used [17, 18]. LASA is an ongoing population-based cohort study among Dutch older persons that commenced in 1992. LASA examines the predictors and consequences of aging and focuses on physical, emotional, cognitive and social functioning in later life. Initially, a representative age- and sex-stratified sample of older men and women (55-84 years, N=3107) was selected from municipality registries in three regions of the Netherlands. A second cohort of participants started in 2002 (55-64 years, N=1002), using the same sampling procedure. Measurement cycles are repeated every three years. Interviews are conducted by trained interviewers at the participants’ homes. In addition, clinical measurements are conducted and participants are asked to complete questionnaires. LASA was approved by the Medical Ethics Committee of the VU University Medical Center Amsterdam, the Netherlands. Prior to the start of the study, all participants gave written informed consent. In the present study, the measurement cycle of 1995/96 of the first LASA cohort (64-88 years, defined here as sample 1), and the cycle of 2008/09, comprising participants of the first and second cohort (60-98 years, defined here as sample 2) were used to examine the cross-sectional association between serum 25(OH)D and anxiety symptoms. In addition, the longitudinal association of serum 25(OH)D with anxiety symptoms after three years was examined in both samples. For sample 1, an additional three years of follow-up were available, resulting in a total of six years of follow-up in this sample.
In sample 1, serum 25(OH)D was assessed only in participants of ≥65 years. Of the 1509 eligible persons, data on both serum 25(OH)D and anxiety was available for 1283 persons. In sample 2, serum 25(OH)D and anxiety data was available for 915 out of 1494 eligible participants. Reasons for missing data were failed or refused blood collection or incomplete data on the anxiety questionnaire. One hundred sixty-seven participants (18.7% of sample 2) had data in both samples.
6.2.2 Measurements Anxiety
Anxiety symptoms were measured with the Anxiety subscale of the Hospital Anxiety and Depression Scale (HADS-A) [19, 20] during a structured interview. This subscale consists of seven self-rated items that indicate to what extent the participant felt tense, frightened, worried, relaxed, frightened in the stomach, restless, or had feelings of
panic in the past four weeks. Items were scored on a Likert scale ranging from 0 ‘rarely or never’ to 3 ‘mostly or always’. Scores of item four (feeling relaxed) were reversed. The total anxiety score - the sum of the seven items - range from 0 to 21. Higher scores indicate the presence of more anxiety symptoms. A score of 8 or higher is indicative of an anxiety disorder. The Dutch translation of the HADS-A was found to be a reliable and valid measurement scale across different age groups [21].
Serum 25(OH)D
To measure serum 25(OH)D, morning blood samples of participants were obtained. Prior to the blood collection of sample 1 in 1995/96, participants were allowed to consume a light breakfast of tea and toast, but no dairy products. In 2009/10 (sample 2), blood samples were drawn in a fasting state. Samples were centrifuged and preserved at -20⁰C. Determination of 25(OH)D took place after blood collection of the respective measurement cycle was completed: in 1997/98 for sample 1 and in 2010/11 for sample 2. Analysis of the blood samples was carried out by the Endocrine Laboratory of the VU University Medical Center. For sample 1, serum 25(OH)D was determined by using a competitive protein binding assay (Nichols Diagnostics Capistrano, CA, USA, interassay coefficient of variation (CV): 10%). For sample 2, a radioimmunoassay was used (Diasorin, Stillwater, Minnesota, USA, interassay CV: 10%). The correlation coefficient between these two methods was r=.94 [22].
Covariables
Sex was examined as a potential effect modifier in the association between serum 25(OH)D and anxiety symptoms, as a stronger association in women than in men was expected, similar to depression [23]. Potential confounders were age [24, 25], sex (if not an effect modifier), education level [26, 27], body mass index (BMI) [28, 29], physical activity (PA) [30, 31], alcohol consumption [32], smoking status [33, 34], and depressive symptoms [13, 35]. We did not include skin color as a confounder because over 99% of LASA participants are Caucasian.
Education level, PA, alcohol consumption, smoking status, presence of chronic diseases and depressive symptoms were assessed during a structured interview. Education level was categorized into low (less than elementary, elementary, or lower vocational education), intermediate (general intermediate, intermediate vocational, or general secondary education) and high (higher vocational, college, or university education). Number of chronic diseases was assessed by asking the participants about the presence of seven common somatic diseases: asthma / chronic obstructive pulmonary disease, cardiac disease, peripheral artery disease, diabetes mellitus,
cerebrovascular accident / stroke, osteoarthritis / rheumatoid arthritis and cancer. PA was measured with the LASA Physical Activity Questionnaire (LAPAQ), a validated questionnaire measuring the frequency and duration of participants’ activities in the past two weeks [36]. Measured activities included walking, bicycling, household activities and sports. To calculate the number of kilocalories used per day per kilogram of body weight (kcal/day), a metabolic equivalent value (MET) was linked to each activity. Alcohol consumption was categorized into (very) excessive, moderate, light and no alcohol consumption, based on the number of days per week the participant drank alcohol and the number of alcoholic consumptions per day [37]. Smoking status was classified into never, former and current smoker. Depressive symptoms were measured with the Center for Epidemiologic Studies Depression Scale (CES-D) [38]. This self-report scale consists of 20 items, measuring depressive symptoms the participant experienced in the past week. The answers were rated on a four-point scale, ranging from 0 ‘rarely or never’ to 3 ‘mostly or always’. Body weight was measured with a calibrated scale without heavy clothes and shoes. Body height was measured using a calibrated stadiometer. Subsequently, BMI was calculated by dividing body weight in kilograms by height in meters squared.
6.2.3 Statistical analyses
All statistical analyses were conducted with SPSS (IBM SPSS Statistics, version 22; Armonk, NY: IBM Corp.). Frequencies and percentages were reported of categorical variables. Medians and interquartile ranges (IQRs) were presented for continuous variables with skewed distributions. As the distribution of anxiety symptoms (outcome variable) remained skewed after a natural logarithmic transformation, this variable was dichotomized using the clinical cut-off score of eight (HADS-A <8: normal range; HADS-A ≥8: indicative of anxiety disorder).
Non-response analyses were conducted to examine differences between included and excluded individuals. Differences in age, sex, education level, alcohol consumption, smoking status, BMI and depressive symptoms were tested with chi-square tests in case of categorical variables, or Kruskal-Wallis tests in case of skewed continuous variables. Serum 25(OH)D, age, BMI, physical activity and depressive symptoms were categorized as they were not linearly associated with anxiety symptoms. Serum 25(OH)D was dichotomized using the commonly used cut-off score of 50 nmol/L [2]. Age was dichotomized at the median. BMI was categorized into low (<20 kg/m²), normal (20-25 kg/m²) and high (≥25 kg/m²). Physical activity was divided into tertiles. Depressive symptoms were dichotomized at the cut-off score of 16, with scores of 16 and higher indicating clinically relevant depressive symptoms [39].
Multiple logistic regression analysis was performed to examine the relation between serum 25(OH)D levels (independent variable) and anxiety symptoms (outcome variable). The longitudinal analyses with three years of follow-up were conducted with the follow-up anxiety score as outcome and the baseline anxiety score as covariate. In addition, we used Generalized Estimating Equations (GEE) analyses with an exchangeable correlation structure to assess the association between baseline 25(OH)D and anxiety over six years of follow-up in sample 1.
As an additional analysis to increase the sample size and power, samples 1 and 2 were pooled in one large dataset. As 25(OH)D assays differed between the samples, the 25(OH)D values of sample 1 (Nichols assay) were calibrated towards the 25(OH)D values of sample 2 (Diasorin assay) with the following formula: Diasorin = 3.7778 + 0.8889 * Nichols [22]. We repeated the cross-sectional and longitudinal logistic regression analyses in this pooled dataset and added ‘sample’ as an additional confounder. As mentioned before, 167 participants had data in both samples. Sometimes, these overlapping participants had different scores on measures that were done on the two baseline cycles of this study (1995/96 and 2008/09). If this was the case, the scores of the sample 2 baseline cycle were chosen in order to reach a more evenly distributed representation of the samples in the pooled analysis, because the N of sample 2 was smaller than the N of sample 1.
For all models, a double-sided p-value of <.05 was considered statistically significant. Potential effect modification of sex was examined by including an interaction term (serum 25(OH)D * sex) in the unadjusted models. In case of significant effect modification (p<.10), stratified analyses were performed. Confounders were added in separate groups to the analysis. First, demographic confounders (age, (sex,) education level) were added to the models (Model 1). Second, the model was elaborated by adding lifestyle confounders (BMI, physical activity, alcohol consumption, smoking status) (Model 2). Finally, depressive symptoms were added in Model 3. The influence of depressive symptoms was examined in this separate, third model to investigate the association between serum 25(OH)D and anxiety symptoms independently from depression. As Spearman correlation analyses showed that depression and anxiety were correlated (sample 1: r= .47, p<.01; sample 2: r= .45; p<.01), there was a risk of multicollinearity in the longitudinal logistic regression models, where both depression and baseline anxiety were predictors. However, the variance inflation factors (VIFs) of these predictors were very low (1.25 - 1.38), and the confidence intervals (CIs) and standard errors (SEs) were not inflated, which indicates that the analyses were not distorted by multicollinearity.
6.3 RESULTS
Of the 1283 participants in sample 1 with both a serum 25(OH)D and an anxiety measure, another 24 persons were excluded from the analyses due to missing values on confounding variables. In sample 2, 23 of 915 participants with both a 25(OH)D and anxiety measure were excluded for the same reason. Missing data on confounding variables were mainly due to refusal or incomplete questionnaires. Non-response analyses showed that the excluded persons in sample 1 (N=250) and 2 (N=602) were older (p<.001 in both samples), had more depressive symptoms (p<.001 in both samples), and drank less alcohol (p=.008 in sample 1; p<.001 in sample 2), compared to participants who were included in the present study. In addition, excluded persons in sample 1 were more often current smokers (p=.014) and excluded participants in sample 2 had a lower education level (p=.018) compared to included persons. Baseline characteristics of both LASA samples are presented in Table 1. The median age of the participants was 74.8 years (IQR: 69.6-80.8) in sample 1 and 69.6 years (IQR: 65.1-76.6) in sample 2. About half of the participants in both samples were female (sample 1: 51.5%, N=649; sample 2: 53.6%, N=478). Mean serum 25(OH)D levels were 53.6 (SD: 24.2) nmol/L in sample 1 and 65.5 (SD: 21.8) nmol/L in sample 2. In sample 1, 48.0% (N=604) had serum 25(OH)D levels below 50 nmol/L, as opposed to 26.4% (N=242) in sample 2. In both samples, the median HADS-A score was 2 (IQR: 0-4).
6.3.1 Cross-sectional analyses
Sex was not a significant effect modifier in both samples (sample 1: p= .55; sample 2: p= .62). Hence, sex was added as a confounder in the analyses. Table 2 shows the results of the cross-sectional logistic regression analyses. In sample 1, participants with serum 25(OH)D levels of <50 nmol/L had a 1.55 times higher odds of having clinically relevant anxiety symptoms than participants with serum 25(OH)D levels of ≥50 nmol/L (p= .035) in the unadjusted model. However, after adjustment for demographic and lifestyle confounders, serum 25(OH)D levels were no longer significantly associated with anxiety symptoms (Model 1: OR=1.29, p=.25; Model 2: OR=1.21, p=.42). Further adjustment for depressive symptoms decreased the association even more (OR=1.13, p=.65). In the unadjusted analysis of sample 2, participants with serum 25(OH)D levels below 50 nmol/L had a 1.74 times higher odds of having clinically relevant anxiety symptoms than participants with serum 25(OH)D levels of ≥50 nmol/L (p= .040). This association remained significant after adjustment for demographic confounders (OR: 1.93; p= .019), but disappeared after additional adjustment for lifestyle confounders and depressive symptoms (Model 2: OR=1.12, p=.72; Model 3: OR=0.86, p=.68).
Table 1.
Baseline characteristics of the two LASA
samples. Sample 1 Sample 2 All (N=1259) 25(OH)D <50 nmol/L (N=604) 25(OH)D ≥50 nmol/L (N=655) p All (N=892) 25(OH)D <50 nmol/L (N=235) 25(OH)D ≥50 nmol/L (N=657) p Sex a: Male Female 610 (48.5) 649 (51.5) 236 (39.1) 368 (60.9) 374 (57.1) 281 (42.9) <.001 414 (46.4) 478 (53.6) 99 (42.1) 136 (57.9) 315 (47.9) 342 (52.1) .13 Age (years) b: 74.8 [69.6-80.8] 78.3 [72.1-83.5] 72.0 [68.4-77.5] <.001 69.6 [65.1 - 76.6] 73.6 [67.4-81.1] 68.6 [64.7-74.7] <.001 Level of education a:
Low Medium High 769 (61.1) 341 (27.1) 149 (1 1.8) 390 (64.6) 139 (23.0) 75 (12.4) 379 (57.9) 202 (30.8) 74 (1 1.3) .008 403 (45.2) 306 (34.3) 183 (20.5) 104 (44.3) 79 (33.6) 52 (22.1) 299 (45.5) 227 (34.6) 131 (19.9) .78 Alcohol use a:
None Light Moderate (very) Excessive 304 (24.1) 632 (50.2) 248 (19.7) 75 (6.0) 188 (31.1) 298 (49.3) 91 (15.1) 27 (4.5) 116 (17.7) 334 (51.0) 157 (24.0) 48 (7.3) <.001 108 (12.1) 489 (54.8) 249 (27.9) 46 (5.2) 35 (14.9) 128 (54.5) 56 (23.8) 16 (6.8) 73 (1 1.1) 361 (54.9) 193 (29.4) 30 (4.6) .13 Smoking a:
Never Former Curr
ent 451 (35.8) 583 (46.3) 225 (17.9) 249 (41.2) 237 (39.2) 118 (19.5) 202 (30.8) 346 (52.8) 107 (16.3) <.001 273 (30.6) 500 (56.1) 119 (13.3) 73 (31.1) 121 (51.5) 41 (17.4) 200 (30.4) 379 (57.7) 78 (1 1.9) .073
Body mass index
a: < 20 kg/m² 20 - 25 kg/m² ≥ 25 kg/m² 47 (3.7) 383 (30.4) 829 (65.8) 30 (5.0) 156 (25.8) 418 (69.2) 17 (2.6) 227 (34.7) 411 (62.7) .001 10 (1.1) 244 (27.4) 638 (71.5) 3 (1.3) 55 (23.4) 177 (75.3) 7 (1.1) 189 (28.8) 461 (70.2) .28
Physical activity (kcal/day)
b: 483.8 [276.9-739.9] 424.5 [239.6-669.0] 547.1 [31 1.5-815.2] <.001 613.4 [374.3-879.8] 504.9 [287.8-71 1.2] 651.8 [420.0-933.7] <.001 Number of chr onic diseases b 1 [0-2] 1 [1-2] 1 [0-2] <.001 1 [1-2] 1 [1-2] 1 [0-2] <.001 Anxiety symptoms a: HADS-A < 8 HADS-A ≥ 8 1157 (91.9) 102 (8.1) 545 (90.2) 59 (9.8) 612 (93.4) 43 (6.6) .037 834 (93.5) 58 (6.5) 215 (91.5) 20 (8.5) 619 (94.2) 38 (5.8) .15 Depr essive symptoms a: CES-D < 16 CES-D ≥ 16 1075 (85.5) 183 (14.5) 494 (81.9) 109 (18.1) 581 (88.7) 74 (1 1.3) .001 812 (91.0) 79 (8.9) 201 (85.5) 34 (14.5) 61 1 (93.1) 45 (6.9) <.001 Values pr esented as N (%) or median [inter
quartile range]. Differ
ences between 25(OH)D categories tested with
a Pearson chi squar
e tests or
b non-parametric Kr
uskal-W
allis tests.
6.3.2 Longitudinal analyses
Similar to the cross-sectional analyses, sex was not a significant effect modifier in both samples (sample 1: p=.56; sample 2: p=.90), so sex was added as a confounder in the analyses. Table 3 displays the results of the longitudinal analyses after three years of follow-up in sample 1 and 2. Logistic regression analyses corrected for baseline anxiety scores showed that serum 25(OH)D was not significantly associated with anxiety symptoms after three years, both in the unadjusted and adjusted models of sample 1 and 2.
Table 2. Cross-sectional logistic regression analysis, showing odds ratios (OR) for clinically relevant anxiety symptoms in participants with serum 25-hydroxyvitamin D <50 nmol/L, compared to participants with serum 25-hydroxyvitamin D ≥50 nmol/L.
Crude model Model 1a Model 2b Model 3c
OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Sample 1 1.55* (1.03, 2.32) 1.29 (0.84, 1.98) 1.21 (0.77, 1.90) 1.13 (0.67, 1.91)
Sample 2 1.74* (1.03, 2.96) 1.93* (1.12, 3.35) 1.12 (0.60, 2.11) 0.86 (0.42, 1.78)
a Adjusted for age, sex, and education level.
b Additionally adjusted for body mass index, physical activity, alcohol use, smoking status, and chronic diseases.
c Additionally adjusted for depressive symptoms.
* p<.05.
Table 3. Longitudinal analysis, showing odds ratios (OR) for clinically relevant anxiety symptoms in participants with serum 25-hydroxyvitamin D <50 nmol/L, compared to participants with serum 25-hydroxyvitamin D ≥50 nmol/L.
Crude model Model 1a Model 2b Model 3c
OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Sample 1: - Logistic regression
- GEE 1.19 (0.77, 1.84)1.41* (1.00, 1.99) 1.03 (0.65, 1.63)1.17 (0.82, 1.68) 0.96 (0.59, 1.56)1.10 (0.76, 1.60) 0.97 (0.59, 1.59)1.11 (0.75, 1.63)
Sample 2:
- Logistic regression 1.17 (0.65, 2.10) 1.21 (0.66, 2.20) 0.82 (0.42, 1.63) 0.78 (0.39, 1.57)
Logistic regression analysis: 3 years of follow-up, all models were adjusted for anxiety symptoms at baseline. GEE: Generalized Estimating Equations, 6 years of follow-up, exchangeable correlation structure.
a Adjusted for age, sex, and education level. b Additionally adjusted for body mass index, physical activity, alcohol
use, smoking status, and chronic diseases. c Additionally adjusted for depressive symptoms. * p<.05.
Table 4. Cross-sectional and longitudinal logistic regression analysis of the pooled dataset, showing odds ratios (ORs) for clinically relevant anxiety symptoms in participants with serum 25-hydroxy-vitamin D <50 nmol/L, compared to participants with serum 25-hydroxy25-hydroxy-vitamin D ≥50 mol/L.
Crude model Model 1a Model 2b Model 3c
OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Cross-sectional 1.58** (1.13, 2.21) 1.41 (0.99, 2.01) 1.18 (0.82, 1.72) 1.01 (0.67, 1.54)
Longitudinal 1.00 (0.69, 1.45) 0.91 (0.62, 1.33) 0.81 (0.54, 1.20) 0.78 (0.52, 1.17)
In the pooled dataset, sample 1 and 2 were combined, with cross-calibrated 25-hydroxyvitamin D levels.
a Adjusted for age, sex, education level, and sample. b Additionally adjusted for body mass index, physical activity,
alcohol use, smoking status, and chronic diseases. c Additionally adjusted for depressive symptoms.
The GEE analyses in sample 1 showed that participants with baseline 25(OH)D levels under 50 nmol/L had a 1.41 times higher odds of having clinically relevant anxiety symptoms in the next six years than participants with 25(OH)D levels of ≥50 nmol/L (p=.049). However, this relationship was no longer statistically significant in the adjusted models (p=.39-.61).
6.3.3 Pooled analyses
To increase power, samples 1 and 2 were pooled into one dataset and the cross-sectional and longitudinal logistic regression analyses were repeated in additional analyses. Results (see Table 4) were largely similar to the findings in the separate samples. Cross-sectionally, the unadjusted model showed that a 25(OH)D level below 50 nmol/L is associated with clinically relevant anxiety symptoms (OR=1.58,
p=.008). After adjustment for confounding, this association was no longer statistically significant. The longitudinal analyses showed no statistically significant associations between 25(OH)D at baseline and anxiety symptoms after three years.
6.4 DISCUSSION
The aim of the present study was to investigate whether serum 25(OH)D levels are associated with anxiety symptoms in older persons. To the best of our knowledge, this is the first study that investigated this association in two representative samples of older persons while taking into account relevant confounding variables.
Our results showed that serum 25(OH)D was significantly associated with anxiety symptoms in older adults in the cross-sectional analyses of both samples. However, this association disappeared when adjustments for confounders were made, indicating that the association is explained by demographic and lifestyle factors. Furthermore, after the addition of depressive symptoms in a separate model, the association between 25(OH)D and anxiety became even weaker, indicating that depressive symptoms are an important explaining variable in the relationship between vitamin D levels and anxiety.
Corrected for baseline anxiety levels, longitudinal analyses showed no significant relationship between serum 25(OH)D and anxiety symptoms after three years of follow-up. For sample 1, we also investigated the association between baseline 25(OH)D and anxiety symptoms over six years. Similar to the cross-sectional analyses, the association in these analyses was statistically significant in the unadjusted model only. Finally, as an additional analysis, we repeated the analyses in a pooled dataset including both samples, to increase the sample size and power of the study. The results of the pooled analyses were similar to those of the separate samples.
Our findings are only partly consistent with other research that examined this association [15, 16]. Armstrong et al. found a statistically significant association of lower serum 25(OH)D with anxiety and depressive symptoms among fibromyalgia patients, using the total HADS-score. However, they no longer observed a significant association when anxiety symptoms were measured independently from depression (using the HADS-A subscale) [15]. These findings contradict our significant cross-sectional unadjusted associations in both samples. Furthermore, Chao et al. examined the association between serum 25(OH)D and health-related quality of life in older persons and included anxiety and depression symptoms as a subvariable [16]. Participants with higher serum 25(OH)D levels were significantly less likely to report problems with anxiety and depression. After adjustment for age and sex, Chao et al. still observed a significant association, which is similar to our results in sample 2. However, the HADS-A, that was specifically developed to measure anxiety symptoms, may not be comparable to the more general EuroQol-5D instrument used by Chao et al [16]. Finally, the present study adjusted for many more confounding variables than previous studies on this topic.
Sex was not an effect modifier in the present study. This was contrary to our expectation that women would display a stronger association between serum 25(OH)D and anxiety symptoms than men. However, sex was an important confounder in the analyses.
Although this study did not observe a significant association between vitamin D and anxiety after adjustment for confounding, the existence of such an association cannot be ruled out. Not many studies have investigated the potential underlying mechanisms of a relationship between vitamin D and anxiety. Nevertheless, several studies have found biological evidence for an association between vitamin D and depression. For instance, the VDR has been identified in several areas of the brain that are involved in the pathophysiology of depression, such as the hippocampus [40]. Furthermore, local activation of 1,25-dihydroxyvitamin D (1,25(OH)2D, the
active metabolite of vitamin D) in the brain stimulates the synthesis of the depression-related neurotransmitter serotonin [41]. Regarding anxiety, some animal studies provided evidence of a mechanistic link between vitamin D and anxiety symptoms. For instance, two rodent studies found that the VDR is present in the amygdala [42, 43]. The amygdala is a brain region known for its role in anxiety [44]. In addition, VDR knock-out mice displayed reduced exploration behavior and increased stress-related grooming on anxiety tests [14], suggesting a role for vitamin D in the regulation of emotional behavior.
the association between serum 25(OH)D levels and anxiety symptoms, independently from depression, in older persons. This provides a first step in understanding this potential relationship. Another strength is the use of large representative population-based samples of older persons with the ability to account for many important confounding variables. Finally, we were able to perform a validation check by examining the associations in two different samples of the LASA study. Although the samples had some overlap of participants (18.7%), they were sufficiently distinct from each other, as sample 2 contained persons who were somewhat younger and healthier compared to the participants of sample 1. In additional analyses, we pooled the samples to increase the power of our analyses, but this did not change the conclusions. A potential limitation of the present study is the possibility of selective non-response of the most vulnerable older persons. Excluded participants were older and had more depressive symptoms than persons who were included in the analyses. This may have caused an underestimation of the effect of serum 25(OH)D on anxiety. Secondly, the LASA cohorts contained relatively few persons with clinically relevant anxiety symptoms (8.1% in sample 1; 6.5% in sample 2), which may have reduced the power of our analyses. Therefore, future research should focus on persons with more severe anxiety symptoms and/or diagnosed anxiety disorder. Finally, we were not able to control for some potential confounders, such as sunlight exposure or vitamin D intake from supplements or food, as these measures were not available for our LASA samples. In conclusion, lower serum 25(OH)D levels were significantly associated with more anxiety symptoms at baseline in two large population-based samples of older persons. However, after adjustment for demographic and lifestyle confounders and depressive symptoms, this relationship was no longer significant. Longitudinal analyses revealed no significant association between serum 25(OH)D and anxiety symptoms after three or six years of follow-up. To our knowledge, this is the first study examining the association between serum 25(OH)D and symptoms of anxiety in older persons, independently from depression. More research among persons with more severe anxiety symptoms is needed.
6.5 ACKNOWLEDGEMENTS, FUNDING SOURCES AND
CONFLICTS OF INTEREST
We are grateful to all participants of the Longitudinal Aging Study Amsterdam for their valued contributions.
The Longitudinal Aging Study Amsterdam is largely supported by a grant from the Netherlands Ministry of Health, Welfare and Sports, Directorate of Long-Term Care. The authors have no competing interests to report.
REFERENCES
1. Lips P. Vitamin D physiology. Prog Biophys Mol Biol. 2006;92(1):4-8.
2. Institute of Medicine (IOM). Dietary reference intakes for calcium and vitamin D. Washington DC: The National Academies Press; 2011.
3. Kuchuk NO, Pluijm SMF, van Schoor NM, Looman CWN, Smit JH, Lips P. Relationships of serum 25-hydroxyvitamin D to bone mineral density and serum parathyroid hormone and markers of bone turnover in older persons. J Clin Endocrinol Metab. 2009;94(4):1244-50.
4. Lips P. Worldwide status of vitamin D nutrition. J Steroid Biochem Mol Biol. 2010;121(1-2):297-300.
5. Annweiler C, Schott AM, Berrut G, Fantino B, Beauchet O. Vitamin D-related changes in physical performance: a systematic review. J Nutr Health Aging. 2009;13(10):893-8.
6. Sohl E, de Jongh RT, Heijboer AC, Swart KMA, Brouwer-Brolsma EM, Enneman AW, de Groot CPGM, van der Velde N, Dhonukshe-Rutten RAM, Lips P et al. Vitamin D status is associated with physical performance: the results of three independent cohorts. Osteoporos Int. 2013;24(1):187-96.
7. Wang L, Song Y, Manson JE, Pilz S, Marz W, Michaelsson K, Lundqvist A, Jassal SK, Barrett-Connor E, Zhang C et al. Circulating 25-hydroxy-vitamin D and risk of cardiovascular disease: a meta-analysis of prospective studies. Circ Cardiovasc Qual Outcomes. 2012;5(6):819-29.
8. Ju SY, Lee YJ, Jeong SN. Serum 25-hydroxyvitamin D levels and the risk of depression: a systematic review and meta-analysis. J Nutr Health Aging. 2013;17(5):447-55.
9. Anglin RES, Samaan Z, Walter SD, McDonald SD. Vitamin D deficiency and depression in adults: systematic review and meta-analysis. Br J Psychiatry. 2013;202:100-7.
10. de Beurs E, Beekman AT, van Balkom AJ, Deeg DJ, van Dyck R, van Tilburg W. Consequences of anxiety in older persons: its effect on disability, well-being and use of health services. Psychol Med. 1999;29(3):583-93.
11. Bierman EJM, Comijs HC, Rijmen F, Jonker C, Beekman ATF. Anxiety symptoms and cognitive performance in later life: results from the longitudinal aging study Amsterdam. Aging Ment Health. 2008;12(4):517-23.
12. Mehta KM, Simonsick EM, Penninx BWJH, Schulz R, Rubin SM, Satterfield S, Yaffe K. Prevalence and correlates of anxiety symptoms in well-functioning older adults: findings from the health aging and body composition study. J Am Geriatr Soc. 2003;51(4):499-504.
13. Vink D, Aartsen MJ, Comijs HC, Heymans MW, Penninx BWJH, Stek ML, Deeg DJH, Beekman ATF. Onset of anxiety and depression in the aging population: comparison of risk factors in a 9-year prospective study. Am J Geriatr Psychiatry. 2009;17(8):642-52.
14. Kalueff AV, Lou YR, Laaksi I, Tuohimaa P. Increased anxiety in mice lacking vitamin D receptor gene. Neuroreport. 2004;15(8):1271-4.
15. Armstrong DJ, Meenagh GK, Bickle I, Lee ASH, Curran ES, Finch MB. Vitamin D deficiency is associated with anxiety and depression in fibromyalgia. Clin Rheumatol. 2007;26(4):551-4. 16. Chao YS, Ekwaru JP, Ohinmaa A, Griener G, Veugelers PJ. Vitamin D and health-related quality
of life in a community sample of older Canadians. Qual Life Res. 2014;23(9):2569-75.
17. Deeg DJH, van Tilburg T, Smit JH, de Leeuw ED. Attrition in the Longitudinal Aging Study Amsterdam. The effect of differential inclusion in side studies. J Clin Epidemiol. 2002;55(4):319-28.
18. Huisman M, Poppelaars J, van der Horst M, Beekman ATF, Brug J, van Tilburg TG, Deeg DJH. Cohort profile: the Longitudinal Aging Study Amsterdam. Int J Epidemiol. 2011;40(4):868-76. 19. Snaith RP. The Hospital Anxiety And Depression Scale. Health Qual Life Outcomes. 2003;1:29.
20. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand. 1983;67(6):361-70.
21. Spinhoven P, Ormel J, Sloekers PP, Kempen GI, Speckens AE, van Hemert AM. A validation study of the Hospital Anxiety and Depression Scale (HADS) in different groups of Dutch subjects. Psychol Med. 1997;27(2):363-70.
22. van Schoor NM, Knol DL, Deeg DJ, Peters FP, Heijboer AC, Lips P. Longitudinal changes and seasonal variations in serum 25-hydroxyvitamin D levels in different age groups: results of the Longitudinal Aging Study Amsterdam. Osteoporos Int. 2014;25(5):1483-91.
23. Milaneschi Y, Shardell M, Corsi AM, Vazzana R, Bandinelli S, Guralnik JM, Ferrucci L. Serum 25-hydroxyvitamin D and depressive symptoms in older women and men. J Clin Endocrinol Metab. 2010;95(7):3225-33.
24. Maggio D, Cherubini A, Lauretani F, Russo RC, Bartali B, Pierandrei M, Ruggiero C, Macchiarulo MC, Giorgino R, Minisola S et al. 25(OH)D Serum levels decline with age earlier in women than in men and less efficiently prevent compensatory hyperparathyroidism in older adults. J Gerontol A Biol Sci Med Sci. 2005;60(11):1414-9.
25. Simning A, Conwell Y, Fisher SG, Richardson TM, van Wijngaarden E. The characteristics of anxiety and depression symptom severity in older adults living in public housing. Int Psychogeriatr. 2012;24(4):614-23.
26. Bjelland I, Krokstad S, Mykletun A, Dahl AA, Tell GS, Tambs K. Does a higher educational level protect against anxiety and depression? The HUNT study. Soc Sci Med. 2008;66(6):1334-45. 27. Forrest KYZ, Stuhldreher WL. Prevalence and correlates of vitamin D deficiency in US adults.
Nutr Res. 2011;31(1):48-54.
28. Vimaleswaran KS, Berry DJ, Lu C, Tikkanen E, Pilz S, Hiraki LT, Cooper JD, Dastani Z, Li R, Houston DK et al. Causal relationship between obesity and vitamin D status: bi-directional Mendelian randomization analysis of multiple cohorts. PLoS Med. 2013;10(2):e1001383.
29. Zhao G, Ford ES, Dhingra S, Li C, Strine TW, Mokdad AH. Depression and anxiety among US adults: associations with body mass index. Int J Obes (Lond). 2009;33(2):257-66.
30. Brunes A, Augestad LB, Gudmundsdottir SL. Personality, physical activity, and symptoms of anxiety and depression: the HUNT study. Soc Psychiatry Psychiatr Epidemiol. 2013;48(5):745-56. 31. Gerdhem P, Ringsberg KAM, Obrant KJ, Akesson K. Association between 25-hydroxy vitamin
D levels, physical activity, muscle strength and fractures in the prospective population-based OPRA Study of Elderly Women. Osteoporos Int. 2005;16(11):1425-31.
32. Kushner MG, Abrams K, Borchardt C. The relationship between anxiety disorders and alcohol use disorders: a review of major perspectives and findings. Clin Psychol Rev. 2000;20(2):149-71. 33. Lee DM, Tajar A, O’Neill TW, O’Connor DB, Bartfai G, Boonen S, Bouillon R, Casanueva FF, Finn
JD, Forti G et al. Lower vitamin D levels are associated with depression among community-dwelling European men. J Psychopharmacol. 2011;25(10):1320-8.
34. Mykletun A, Overland S, Aaro LE, Liabo HM, Stewart R. Smoking in relation to anxiety and depression: evidence from a large population survey: the HUNT study. Eur Psychiatry. 2008;23(2):77-84.
35. Hoogendijk WJG, Lips P, Dik MG, Deeg DJH, Beekman ATF, Penninx BWJH. Depression is associated with decreased 25-hydroxyvitamin D and increased parathyroid hormone levels in older adults. Arch Gen Psychiatry. 2008;65(5):508-12.
36. Stel VS, Smit JH, Pluijm SMF, Visser M, Deeg DJH, Lips P. Comparison of the LASA Physical Activity Questionnaire with a 7-day diary and pedometer. J Clin Epidemiol. 2004;57(3):252-8. 37. Garretsen HFL. Probleemdrinken: prevalentiebepaling, beïnvloedende factoren en
preventiemogelijkheden. Theoretische overwegingen en onderzoek in Rotterdam [Problem drinking: determinations of prevalence, influencing factors and prevention possibilities: theoretical considerations and research in Rotterdam]. Lisse, The Netherlands: Swets, Zeitlinger, 1983. [Dissertation in Dutch]
38. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas. 1977;1:385-401.
39. Berkman LF, Berkman CS, Kasl S, Freeman DHJ, Leo L, Ostfeld AM, Cornoni-Huntley J, Brody JA. Depressive symptoms in relation to physical health and functioning in the elderly. Am J Epidemiol. 1986;124(3):372-88.
40. Eyles DW, Smith S, Kinobe R, Hewison M, McGrath JJ. Distribution of the vitamin D receptor and 1 alpha-hydroxylase in human brain. J Chem Neuroanat. 2005;29(1):21-30.
41. Fernandes de Abreu DA, Eyles D, Feron F. Vitamin D, a neuro-immunomodulator: implications for neurodegenerative and autoimmune diseases. Psychoneuroendocrinology. 2009;34 Suppl 1:S265-S77.
42. Musiol IM, Stumpf WE, Bidmon HJ, Heiss C, Mayerhofer A, Bartke A. Vitamin D nuclear binding to neurons of the septal, substriatal and amygdaloid area in the Siberian hamster (Phodopus sungorus) brain. Neuroscience. 1992;48(4):841-8.
43. Walbert T, Jirikowski GF, Prufer K. Distribution of 1,25-dihydroxyvitamin D3 receptor immunoreactivity in the limbic system of the rat. Horm Metab Res. 2001;33(9):525-31.
44. Etkin A, Wager TD. Functional neuroimaging of anxiety: a meta-analysis of emotional processing in PTSD, social anxiety disorder, and specific phobia. Am J Psychiatry. 2007;164(10):1476-88.
