Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 May 30.
Published in final edited form as: Psychiatry Res. 2015 Mar 5;227(1):46–51. doi: 10.1016/j.psychres.2015.02.016

Associations between Vitamin D Levels and Depressive Symptoms in Healthy Young Adult Women

David C R Kerr a,*, David T Zava b, Walter T Piper c, Sarina R Saturn a, Balz Frei d, Adrian F Gombart d
PMCID: PMC4420707  NIHMSID: NIHMS670058  PMID: 25791903

Abstract

There have been few studies of whether vitamin D insufficiency is linked with depression in healthy young women despite women’s high rates of both problems. Female undergraduates (n = 185) living in the Pacific Northwest during fall, winter, and spring academic terms completed the Center for Epidemiologic Studies Depression (CES-D) scale weekly for four weeks (W1–W5). We measured serum levels of vitamin D3 and C (ascorbate; as a control variable) in blood samples collected at W1 and W5. Vitamin D insufficiency (<30ng/mL) was common at W1 (42%) and W5 (46%), and rates of clinically significant depressive symptoms (CES-D ≥ 16) were 35–42% at W1–W5. Lower W1 vitamin D3 predicted clinically significant depressive symptoms across W1–W5 (β = −.20, p < .05), controlling for season, BMI, race/ethnicity, diet, exercise, and time outside. There was some evidence that lower levels of depressive symptoms in Fall participants (vs. Winter and Spring) were explained by their higher levels of vitamin D3. W1 depressive symptoms did not predict change in vitamin D3 levels from W1 to W5. Findings are consistent with a temporal association between low levels of vitamin D and clinically meaningful depressive symptoms. The preventive value of supplementation should be tested further.

Keywords: depressive symptoms, vitamin D, seasonal affective disorder

1. Introduction

Depression is a major public health problem that is linked with premature death by suicide and other causes (Chesney et al., 2014), and disability and economic burden (Greenberg et al., 2003). According to a nationally representative study lifetime prevalence of major depressive disorder was 25% among American women (vs. 16% for men; Kessler et al., 2003). Moreover, a longterm repeated measures study through age 30 years recently indicated the cumulative incidence rate was a staggering 63.4% for women (vs. 34.7% among men; Rohde et al., 20131). Thus, depression places an enormous health burden on young women, in particular.

Some epidemiological studies suggest low levels of vitamin D [25-hydroxyvitamin D or 25(OH)D] may contribute to depression. Proposed biological mechanisms for this association include that vitamin D: 1) has receptors that are distributed in brain areas involved in emotional processing and affective disorders (Eyles et al., 2013, Kesby et al., 2011); 2) regulates serotonin synthesis via transcriptional activation of the tryptophan hydroxylase 2 gene (Patrick & Ames, 2014); and 3) impacts innate immunity and the production of proinflammatory cytokines that in turn influence mood by activating the stress response (Capuron & Miller, 2004; Raison et al., 2006; Silverman et al., 2005; Zhang et al., 2012).

However, evidence for an association between vitamin D sufficiency status and depression is inconclusive. Meta-analyses (Li et al., 2013; Shaffer et al., 2014) indicate vitamin D supplementation trials have not shown significant effects on depression. Likewise, a meta-analysis of 10 cross-sectional studies indicated the odds of depression were not significantly elevated among those with low levels of vitamin D compared to others (Anglin et al., 2013). Importantly, researchers often have not adjusted for important confounders, and have used definitions of vitamin D insufficiency or deficiency that are based on risk for medical conditions. These cut-offs may be arbitrary with respect to risk for depression, and thus their use in research may have contributed to null findings. Indeed, levels (vs. cut-offs) of vitamin D in 18–65 year olds were found to be lower among participants with current or remitted depression relative to controls, and associated with symptom severity and a worsened two-year course (Milaneschi et al., 2014). Thus, alternative measurement and design approaches are needed.

Another problem with the current evidence base on relations between vitamin D and depression is that findings may not generalize to healthy young women. Indeed, most experimental and observational research concerns older or medically compromised populations (e.g., Jorde et al., 2008; Parker and Brotchie, 2011; Sanders et al., 2011). In an important exception, Ganji and colleagues (2010) found that rates of current depression were associated with vitamin D deficiency in a nationally representative U.S. sample of young adults. Of note, however, they did not control for season of the year or some other important confounds.

Studying young healthy women minimizes the possibility that observed associations between vitamin D and depression are explained by significant health conditions, age variation, or gender. However, a number of potential confounders remain, including race/ethnicity, health behaviors, and season (Ganji et al., 2010; Hollis, 2005; Norman, 1998). Regarding seasonal effects, humans can generate sufficient vitamin D subcutaneously when exposed to sunlight (Norman, 1998). During non-summer months, however, meteorological and behavioral factors can limit skin exposure to sunlight and lead to decreased production of vitamin D3 and rapid depletion of body reserves. For this reason, individuals from a range of climates often become vitamin D insufficient in the winter and do not regain sufficiency for months (Holick, 2007).

Given that vitamin D levels vary with season of the year and may have an inverse relationship with depressive symptoms, some have suggested vitamin D may play a role in explaining seasonal affective disorder (SAD) and seasonality (the milder spectrum of sensitivity of mood and behavior to seasonal changes). Leading models of SAD emphasize the disruption of vulnerable individuals’ circadian rhythms following seasonal changes in day length (Rohan et al., 2009). However, day length is confounded with other factors (intensity of solar radiation, cloud cover, clothing) that explain seasonal variation in vitamin D. Thus, it is possible that seasonal changes in vitamin D levels account for some of the seasonal variation in depressive symptoms. Despite the intuitive appeal of this theory, it has not been studied extensively. In a case control study Kjaergaard and colleagues (2012) found that participants from Northern Norway showed a range of seasonality scores, but scores did not differ for those with insufficient versus sufficient vitamin D levels. Furthermore, they found that seasonality scores did not decrease more following high-dose vitamin D supplementation than following placebo. Reviews of the topic yield little conclusive support for a role of vitamin D in the etiology and treatment of SAD or seasonality (see Bertone-Johnson, 2009; Parker & Brotchie, 2011). However, further research is needed using larger samples and stronger study designs.

The present study builds on the extant literature in several ways. First, we test whether vitamin D levels are associated with depressive symptoms in healthy young women; such women rarely have been considered in this literature despite their high risk for depression and the straightforward implications such research may have for prevention. Second, we use a latent variable to estimate clinically significant depressive symptoms over a four week period, a measurement approach that is more reliable than using a cut-off score based on a single assessment. Third, in addition to controlling for participant characteristics and health behaviors, we control for vitamin C levels to evaluate the specificity of the vitamin D path of interest. Finally, we integrate prior findings on seasonal and racial/ethnic group differences in the separate literatures on vitamin D sufficiency and depressive symptoms (Eisenberg et al., 2013; Ganji et al., 2010; Hollis, 2005; Norman, 1998). That is, we hypothesize that any differences in depressive symptoms by season or race/ethnicity will be explained in part by group differences in vitamin D levels (i.e., indirect effects).

2. Methods

2.1. Participants

Participants were 185 undergraduate women ages 18–25 years [mean (SD) = 19.66 (1.56)] who were enrolled at a large university in the northwestern United States. Inclusion criteria for participation were: female gender; not currently pregnant; self-reported weight of 110 pounds or greater2; and willingness to complete web-based surveys and contribute two 10-mL blood samples. From the initial sample of 190 enrolled participants, 5 were excluded from the present analyses because they were over the age of 25 years (n = 3) or did not have data on vitamin D level at baseline (n = 2). Participants identified as White (n = 153, 82.7%), African American (n = 4, 2.2%), Asian (n = 20, 10.8%), Pacific Islander (n = 3, 1.6%), Native American (n = 3, 1.6%), Hispanic (n = 21, 11.4%), and “other” (n = 3, 1.6%) race/ethnicity.

2.2. Procedures

Participants were recruited from a psychology department subject pool in the fall, winter, or spring terms (n = 86, 36, and 63, respectively). Participants were recruited each week for the first half of each 10-week term (baseline recruitment dates of 9/26/2011–10/26/2011; 1/16/2012–2/1/2012; 4/3/2012–5/2/2012) and surveyed weekly for 4 weeks. Participants were not recruited after the fifth week of the term to avoid conflicts with final exams and university breaks. In an in-person baseline appointment individuals answered whether they were eligible based on inclusion criteria described above. Eligible women completed informed consent. Participants who enrolled completed the W1 web-based survey on a lab computer. They scheduled non-fasting blood draw appointments for 2–3 days later (W1) and 4 weeks later (W5). Following the first blood draw, participants began receiving a survey web link by email once per week for 4 weeks (W2–W5). Participants were allowed 48 hours to complete each survey before the link became inactive. After the final survey (W5), participants were reminded of their second blood draw appointment.

Time of day of blood draws was variable (9am–4pm). Phlebotomists recorded participants' heights and weights and then collected 10 mL of venous blood. Samples were settled for 20–30 minutes and then centrifuged for 10 minutes at 1500 g at room temperature. Serum was aliquoted and stored at −80°C. Vitamins D and C were assayed in separate labs.

Participants were compensated for completing the five surveys and W1 blood draw with course extra-credit; participants were paid $25 for the W5 blood draw. All procedures were approved by the university IRB. Participation rates on online surveys at W2–W5 were high (98%, 98%, 95%, 98%), and 95% of participants completed all surveys. Participation rate at the W5 blood draw also was high (98%).

2.3 Measures

2.3.1. Depressive symptoms

On the five W1 through W5 surveys, participants completed the 20-item Center for Epidemiologic Studies Depression (CES-D) scale (Radloff, 1977). Participants used a 4-point scale (rarely or none of the time [0–1 day] coded ‘0’ to most or all of the time [5–7 days] coded ‘3’) to report how often they felt symptoms in the past week. Internal consistency was adequate at all waves (α = .89 to .92). Scores also were coded as clinically significant (yes ‘1’ or no ‘0’) if they reached the cut-off score of 16.

2.3.2. Vitamin D and C blood concentrations

Total 25(OH)D concentrations were determined by measuring 25(OH)D2 and 25(OH)D3 in solvent-extracted serum samples by LC-MS/MS at ZRT Laboratory (Beaverton, OR; Newman et al., 2009). This assay was calibrated using National Institute of Standards Technology (NIST) standard reference materials. Continued method accuracy was ensured by participation in College of American Pathologists (CAP) and Vitamin D External Quality Assessment Scheme (DEQAS) proficiency testing schemes that require laboratories to meet performance targets assigned by the CDC Reference Laboratory and NIST Reference Measurement Procedure respectively. Vitamin D2 levels were below detectable levels (<4 ng/mL) in 98% of samples and thus did not contribute to total blood concentration estimates. Therefore, statistical models used vitamin D3 levels only; descriptive statistics on deficiency and insufficiency rates were based on total vitamin D as is the convention. Vitamin C (ascorbate) blood level was analyzed by ion-paired, reverse phase HPLC with electrochemical detection (Suh et al., 2003) at Linus Pauling Institute (Corvallis, OR).

2.3.3. Season

For regression models two variables (Winter vs. Fall, and Spring vs. Fall) were created to represent the academic term of participation. Fall term was the referent (coded 0, 0); winter and spring term participation were coded 1, 0 and 0, 1 respectively.

2.3.4. Body Mass Index (BMI)

Participant height and weight (no shoes) were measured at the time of the W1 blood sample and converted to BMI.

The remaining covariates were collected on the W1 self-report survey.

2.3.5. White, non-Hispanic race/ethnicity

This variable was coded “1” for participants (73.5%, n=136) who did not endorse any race/ethnicity other than White, non-Hispanic, and “0” for all others (also referred to as women of color).

2.3.6. Exercise time

Duration in minutes was calculated from responses to the question, “How many hours and minutes of exercise did you get in total this week?” Reported exercise from one outlying participant who reported 26 hours of exercise was excluded.

2.3.7. Time outside in daylight hours

Duration in minutes was calculated from the question, “Including exercise, how many hours and minutes of time this week did you spend outside during daylight hours?” and then was natural log transformed to reduce skewness.

2.3.8. Dietary intake of vitamin D

Past month dietary intake of vitamin D was estimated using a screening instrument developed by Blalock and colleagues (2003). For each of 22 vitamin D rich or fortified foods (e.g., 2% milk, shrimp, eggs) participants answered how often they consumed the food in the past month and what sized serving they usually consumed (small, medium or large; defined for each food). Intake was then calculated in micrograms per month using tables provided by the National Cancer Institute (2009). A square root transformation reduced skewness.

2.3.9. Antidepressant and multivitamin use

Participants answered the questions (no ‘0’, yes ‘1’), “Do you currently take an antidepressant medication (such as Prozac, Paxil, Zoloft, Lexapro, Celexa, Effexor, Cymbalta, Strattera, or Wellbutrin) for depression, anxiety, inattention, or another reason?” and “Do you take a vitamin D supplement or a multivitamin that includes vitamin D?”

2.4. Data analysis

We first examined descriptive statistics for continuous and binary measures of depressive symptoms and vitamin D3 levels to characterize the sample. We also evaluated univariate associations between these variables, and relations depressive symptoms and vitamin D3 levels had with study covariates. Next we fit a latent variable3 model with binary indicators for clinically significant symptoms at W1–W5 using Mplus version 6.12. We used the default robust weighted least squares estimator, which uses probit regressions to regress the binary indicators on a latent factor (Muthen & Muthen, 2011). The model fit the data when a covariance between W4 and W5 depression scores was specified [Model χ2 (df = 4) = .706, p = .95].

We tested our primary hypothesis using this latent variable representing clinically significant depressive symptoms. To do so we regressed the latent variable on W1 vitamin D3 blood concentration while simultaneously regressing both variables on multiple potentially confounding covariates (including season and race/ethnicity).

Finally, using this model we explored whether any prediction of depressive symptoms from season of participation would be explained by an indirect effect through vitamin D3 levels. Similarly, we explored whether any effects of race/ethnicity on depressive symptoms would be partially explained by racial/ethnic group differences in vitamin D3 levels. These possibilities were evaluated by examining total, direct, and indirect effects of these predictors on depressive symptoms. An indirect effect would be supported if a significant total effect of a predictor on depressive symptoms could be decomposed into a significant indirect effect via vitamin D3 levels and a significant or non-significant remaining direct effect.

3. Results

3.1. Descriptive statistics

Descriptive statistics are reported in Table 1. At each of five observations, 34% to 41% of participants reported clinically significant depressive symptoms. Rates of vitamin D insufficiency [n = 58 (31%) at 20–29.99 ng/mL] and deficiency [n = 19 (10%) at <20ng/mL] at baseline and at four week follow-up [n = 57 (31%) insufficient and n = 27 (15%) deficient, out of 182 participants] were common. Antidepressant medication and multivitamin use were reported by 10% and 42% of participants, respectively. Statistics on other covariates are listed in Table 1.

Table 1.

Descriptive statistics on study measures.

Study measure Continuous Scale Categorical Scale
Depressive symptoms Mean (SD) Clinically significant
  W1 14.04 (9.11) 34%
  W2 15.23 (9.68) 42%
  W3 14.13 (10.15) 39%
  W4 13.21 (9.55) 35%
  W5 13.43 (10.22) 36%
Vitamin D3 (ng/mL), blood concentration Mean (SD) Sufficiency1
  W1 33.40 (12.42) 31% insufficient
10% deficient
  W5 32.36 (12.17) 31% insufficient
15% deficient
Body mass index (kg/m2) Mean (SD) CDC cut-points
  W1 24.79 (4.79) 1% underweight
62% normal
27% overweight
10% obese
Dietary vitamin D intake (µg/month) Mean (SD)
  W1 unadjusted2 182.45 (178.37)
  W1 ln-transformed 4.79 (0.97)
Exercise (minutes/week) Mean (SD)
  W1 282.31 (193.68)
Time outside (minutes/week) Mean (SD)
  W1 unadjusted 522.97 (520.29)
  W1 square root-transformed 20.32 (10.52)
Ascorbate (µM), blood concentration Mean (SD)
  W1 59.31 (21.79)
Open in a new tab

Note.

1

rates based on total vitamin D (D2 + D3).

2

for reference, current U.S. recommended dietary allowance is 15 µg per day (~450 µg/month; Holick et al., 2011)

3.2. Unadjusted associations between vitamin D3 and depressive symptoms

Consistent with our hypothesis, vitamin D3 levels and depressive symptoms generally were negatively correlated (see Table 2). Tests of these associations using cut-off scores for both variables further illustrate these results. That is, chi-square analyses indicated women who were vitamin D sufficient at W1 less often had clinically significant depressive symptoms across W1–W5 [χ2 (1) = 5.40 to 8.48, all p < .05]. For example, 26% (28 of 108) of participants with sufficient vitamin D at W1 reported depressive symptoms that exceeded the clinical cut-off at W1 compared to 45% (35 of 77) of participants with deficient or insufficient W1 levels. Sufficiency at W5 also was consistently related to W1–W5 depressive symptom cut-off scores in the predicted direction, but was only significantly associated at W3 and W4 [χ2 (1) = 7.33 and 5.21, respectively, p < .05].

Table 2.

Correlations among CES-D depressive symptoms (W1–W5) and vitamin D3 (ng/mL, W1 and W5).

CES-D,
W1		 CES-D,
W2		 CES-D,
W3		 CES-D,
W4		 CES-D,
W5		 Vit D3,
W1		 Vit D3,
W5
CES-D, W1 .72*** .68*** .62*** .56*** −.13† −.07
CES-D, W2 .74*** .65*** .59*** −.23** −.15*
CES-D, W3 .77*** .61*** −.18* −.17*
CES-D, W4 .68*** −.24** −.13†
CES-D, W5 −.21** −.13†
Vitamin D3, W1 .82***
Vitamin D3, W5
Open in a new tab

Note. CES-D: Center for Epidemiologic Studies Depression scale; W1–W5: Wave 1 – Wave 5.

†

p < .10.

*

p < .05.

**

p < .01

***

p < .001.

3.3. Associations of covariates with depressive symptoms and vitamin D3

Correlations that study covariates had with depressive symptoms and vitamin D3 levels (see Table 3) supported the need to control for them in the final model. Lower depressive symptoms generally were associated with White/non-Hispanic race/ethnicity, having lower BMI, and not using antidepressants. Regarding seasonal differences in depressive symptoms, mean symptom scores across W1–W5 were lowest among participants recruited in fall term [ANOVA F (2, 179) = 3.73, p < .05]; CES-D means (SD) for fall, winter, and spring term participants were 12.3 (8.2), 16.4 (7.3), and 14.9 (8.6), respectively.

Table 3.

Correlations of study covariates with outcomes.

Depressive symptoms (CES-D) Vitamin D3
W1 W2 W3 W4 W5 W1 W5
White/non-Hispanic −.10 −.22** −.22** −.13† −.12 .33*** .38***
BMI .12 .17* .19* .17* .18* −.17* −.17*
Dietary D intake (ln) −.08 −.08 −.10 −.13† −.08 .09 .01
Exercise .00 .05 −.02 −.01 −.07 .18* .13†
Time outside (square root) −.06 −.12 −.17* −.15† −.08 .02 .04
Vitamin C .01 −.08 −.10 −.14† −.12 .20** .11
Multivitamin use −.02 −.01 −.02 −.02 −.00 .19* .21**
Antidepressant use .17* .12 .14† .20** .29*** .10 .13†
Open in a new tab

Note. CES-D: Center for Epidemiologic Studies Depression scale; BMI: body mass index.

†

p < .10.

*

p < .05.

**

p < .01

***

p < .001.

Correlations that covariates had with vitamin D3 levels also are reported in Table 3; higher levels were associated with being White/non-Hispanic, having lower BMI, exercising more, taking a multivitamin, and having higher levels of vitamin C. Differences in vitamin D3 by season were examined using repeated measures ANOVA. As expected, vitamin D3 level [F (2, 178) = 4.49, p < .05] and change [F (2, 178) = 27.67, p < .001] varied by whether women participated in fall, winter, or spring term. Vitamin D tended to be lower in winter than in fall and spring terms, and from W1 to W5 levels tended to decrease among fall participants, remain stable among winter participants, and increase among spring participants4. Chi-square analyses also showed that W1 vitamin D sufficiency rates differed by race/ethnicity [χ2 (1) = 10.54, p< .01] and vitamin supplement use [χ2 (1) = 11.13, p< .001]; specifically, sufficiency rates were 39% for women of color, compared to 65% for other women; and 73% for those taking supplements versus 48% for others.

3.4. Multivariate prediction model of clinically significant depressive symptoms

Next we built on the unconditional model described in the data analysis section by regressing the latent variable (representing clinically significant depressive symptoms) on W1 vitamin D3 levels, and simultaneously regressing both of these variables on: race/ethnicity, BMI, dietary D intake, exercise, time outside, vitamin C levels, antidepressant use, and the two variables representing season of recruitment. As shown in Table 4, vitamin D3 was negatively associated with the probability of clinically significant depressive symptoms after covariates were controlled5. Antidepressant use was the only other independent predictor of symptoms.

Table 4.

Results of untrimmed path model predicting depressive symptoms from vitamin D and covariates.

Vitamin D3
(observed)		 Depressive symptoms
(latent factor)
Predictor Est (SE) β Est (SE) β
Vitamin D3 −.02 (.01) −.20*
Winter vs. Fall −9.06 (2.21) −.29*** .18 (.22) .07
Spring vs. Fall −5.80 (1.89) −.22** .16 (.19) .08
White/non-Hispanic 8.41 (2.22) .30*** −.25 (.19) −.11
BMI −.29 (.25) −.11 .03 (.02) .15
Dietary D intake (ln) .72 (.72) .06 .04 (.09) .04
Exercise .01 (.00) .16* .00 (.00) .02
Time outside (square root) −.08 (.08) −.07 −.01 (.02) −.14†
Vitamin C .07 (.04) .13* −.00 (.00) −.03
Antidepressant use 4.44 (2.06) .11* .71 (.29) .22*
Open in a new tab

Note. Covariance of W4 and W5 depressive symptoms Est (SE) = .17 (.09), β = .17*. Vitamin D3 intercept Est (SE) = 23.64 (7.79), p < .01. Model χ2 (df = 44) = 40.19, p = .64. Total and indirect effects on depressive symptoms were as follows: Winter term participation: total [Est (SE) = .33 (.21), β = .13, p = n.s.] and indirect [Est (SE) = .15 (.07), β = .06, p < .05] effects; Spring term participation: total [Est (SE) = .25 (.19), β = .12, p = n.s.] and indirect [Est (SE) = .09 (.05), β = .05, p = .06]; White/non-Hispanic race/ethnicity: total [Est (SE) = −.39 (.18), β = −.18, p < .05] and indirect [Est (SE) = −.15 (.07), β = −.06, p < .05] effects.

†

p < .10.

*

p < .05.

**

p < .01

***

p < .001.

We then explored whether the previously identified associations of season and race/ethnicity with depressive symptoms were explained by effects on vitamin D3 levels (see Table 4 notes). Indirect paths were significant for season6 of participation and race/ethnicity suggesting effects of these variables on depressive symptoms were partially explained by their effects on vitamin D3.

4. Discussion

Vitamin D3 levels were negatively related to clinically significant depressive symptoms across five weekly assessments in healthy young adult women. These findings are important given that previous studies often have concerned older adults and special medical populations, or have not controlled for important confounds (Ganji et al., 2010; Jorde et al., 2008; Sanders et al., 2011). Without an experimental design we cannot conclude that low vitamin D levels caused depressive symptoms (or that high levels were protective). However, we found no evidence for the reverse causal sequence; depressive symptoms did not predict change in vitamin D3 across the four week study, whereas other factors did.

Mean depressive symptoms were lower in women who participated in Fall than in other terms, and this pattern was partially explained by their higher levels of vitamin D3 relative to those of Winter and Spring participants. Thus, seasonal changes in vitamin D3 may be an underappreciated mechanism in models of seasonality and SAD. Although we controlled for time spent outside, it remains possible that low vitamin D3 levels are simply a proxy for low sunlight exposure or reduced day length that influence depression through other (e.g., circadian) mechanisms. Studies that disentangle multiple etiological mechanisms by which depressive symptoms are responsive to time of the year are needed, as few high quality studies exist (Bertone-Johnson, 2009). In the present sample, seasonal differences in depressive symptoms were very modest overall, suggesting season is not a prominent influence on depressive symptoms in young adult women. Thus, larger samples of individuals at high risk for SAD will be needed for etiological studies.

Consistent with prior research, our findings indicated that young women of color were more often vitamin D insufficient and depressed than other women (Eisenberg et al., 2013; Ganji et al., 2010; Hollis, 2005; Norman, 1998). Notably, the association between race/ethnicity and depressive symptoms was partially explained by levels of vitamin D3. To our knowledge this is the first study to report such a pattern. Although the reasons for these effects are unknown (e.g., effects of melanin on endogenous vitamin D3 production; Norman, 1998) it is possible that vitamin D insufficiency contributes to mental health problems in young minority women, and that supplementation may help reduce this burden. To date, no supplementation trials have focused on preventing or reducing depression among people of color, despite their separately documented risks for vitamin D insufficiency and depression.

Finally, a few observed patterns in vitamin D3 levels deserve further comment. Vitamin D3 levels showed the expected seasonal patterns in terms of absolute level and change over the four week follow-up; these patterns likely reflected seasonal changes in intensity and duration of skin exposure to sunlight. Still, insufficiency was observed in all three seasons (assessments spanned eight months of the year, end of September to end of May). Additionally, null findings regarding dietary intake of vitamin D were consistent with the notion that even foods that are fortified or naturally rich in vitamin D make negligible contributions to blood levels relative to season (sunlight exposure) or vitamin supplement use (Holick et al., 2011). Thus, supplementation may be relevant to some populations year-round.

4.1. Limitations

Our sample was recruited from a university and a single geographic location; how representative participants were of healthy young adult women is unknown. Another limitation was the reliance on self-report measures of health behaviors and depressive symptoms rather than, for example, a clinical interview to establish a diagnosis of depression. Additionally, although our repeated measures design was an advance over prior studies, the follow-up period was so short and the constructs of interest were so stable during this period that statistical power may have been too low to detect and predict change over time. Finally, we did not screen for health factors that may affect vitamin D levels such as anticonvulsant or antiretroviral medication use, inflammatory bowel disease, or tanning bed use, which is widespread (36% of U.S. women ages 18–24 years; Choi et al., 2010) and associated with vitamin D production (Tangpricha et al., 2004).

4.2. Conclusions

Vitamin D deficiency and insufficiency occur at high rates in healthy young women, and lower vitamin D3 levels are related to clinically significant depressive symptoms. Supplementation is a simple, low cost, and low risk intervention. Given the lifespan health risks associated with insufficiency (e.g., Giovannucci, 2009), supplementation is warranted whether or not the modest role of vitamin D in depression observed here generalizes more broadly. Future research on multi-dimensional etiological models of seasonal depression should include measures of vitamin D levels. Controlled trials are needed to examine the extent to which vitamin D supplementation can prevent seasonal and non-seasonal forms of depression, and whether women of color are especially likely to benefit.

Highlights.

  • Depressive symptoms and vitamin D were measured in 185 healthy women across four weeks.

  • Significant symptoms and vitamin D insufficiency were common, and differed by season.

  • Initially low vitamin D levels were associated with clinically significant depressive symptoms across follow-up.

  • Between-subjects differences in depression by season were partially explained by seasonal changes in vitamin D.

  • Racial-ethnic differences in depression were partially explained by group differences in vitamin D levels.

Acknowledgements

The authors thank phlebotomy and lab staff Mary Garrard, Deborah Hobbs, Taylor Wolf, Teresa Wolfe, and Jill Redmond for their assistance with data and sample collection, and David Kimball for technical assistance on vitamin D assays. This study was supported by a grant from the John C. Erkkila, M.D. Endowment for Health and Human Performance of Good Samaritan Hospital Foundation, and by grant number P30 ES000210 from the National Institute of Environmental Health Sciences (NIEHS). Neither funding source influenced study design; the collection, analysis or interpretation of data; the writing of the report; or the decision to submit the article for publication.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Rates by gender given by P. Rohde & J. M. Gau, personal communication, May 1, 2014.

2

The IRB used this cut-off for defining a study as minimal risk.

3

We initially explored whether a growth model for level and change in clinically significant depressive symptoms from W1 to W5 could be identified. However, a number of different growth models (e.g., linear, linear spline, quadratic) failed to specify patterns of change over the 4 week study period.

4

Specifically, mean (SD) vitamin D3 levels for fall, winter, and spring participants at W1 were = 37.37 (12.55), 28.05 (10.98), and 31.05 (11.38) ng/mL, respectively; and at W5 were 32.58 (12.07), 27.91 (12.11), and 34.64 (11.84) ng/mL, respectively. Considered differently, rates of sufficiency at W1 in fall, winter, and spring terms were 73%, 33%, and 52%, respectively [χ2 (2) = 18.06, p< .001] and at W5 were 55%, 31%, and 65% [χ2 (2) = 11.14, p < .01].

5

Our modeling approach did not provide information about temporal ordering of associations between W1 vitamin D3 and W1–W5 depressive symptoms. Thus, we tested an alternative model that depressive symptoms led to decreases in vitamin D. We regressed W5 vitamin D3 levels on W1 depressive symptoms and covariates (term of participation, BMI, race/ethnicity, dietary vitamin D, exercise, time outside, antidepressant use, and multivitamin use). The strong predictive path from vitamin D3 levels at W1 to levels at W5 was controlled [Est (SE) = .83 (.05), β = .84, p < .001]; thus the W5 vitamin D3 outcome is interpreted as change in vitamin D3 levels from W1 to W5. Relatively positive changes in vitamin D3 from W1 to W5 (more pronounced increases or less pronounced decreases) were predicted by winter and spring participation [vs. fall; Est (SE) = 2.72 (1.35), β = .09, p < .05 and Est (SE) = 6.92 (1.07), β = .27, p < .001, respectively], and White, non-Hispanic race/ethnicity [Est (SE) = 2.61 (1.11), β = .09, p < .05], but not by W1 depressive symptom scores [Est (SE) = .07 (.05), β = .05, p = n.s.] or other predictors. Results did not differ appreciably when the binary depression variable (clinically significant depressive symptoms) was used, or when the same model was run using W2 depressive symptoms as the predictor. Thus, the alternative hypothesis was not supported.

6

The effects of winter vs. fall (significant) and spring vs. fall (marginally significant) must be interpreted jointly.

Conflict of interest

David T. Zava is CEO of ZRT Laboratory. The authors have no other conflicts to disclose.

Contributors

David Kerr conceptualized the study, served as PI on the grant and study, conducted the data analyses, and had primary responsibility for writing the manuscript. David Zava supervised the vitamin D assays, wrote critical portions of the methods and limitations, and edited and approved the manuscript. Walter Piper played a chief role in recruiting participants, and the collection, processing, and storage of blood samples; contributed to study conceptualization and literature review; and reviewed and approved the manuscript. Sarina Saturn participated in conceptualization of the study and grant proposal; contributed to the literature review and wrote sections of the introduction and discussion; and reviewed and approved the manuscript. Balz Frei contributed to study conceptualization; provided institutional support for the project; supervised the vitamin C assays; contributed to the literature review; and edited and approved the manuscript. Adrian Gombart contributed to conceptualization of the study and grant proposal; supervised procedures related to collection and storage of blood samples; contributed to the literature review; wrote sections of the introduction, method, and discussion; and edited and approved the manuscript.

References

  1. Anglin RE, Samaan Z, Walter SD, McDonald SD. Vitamin D deficiency and depression in adults: systematic review and meta-analysis. The British Journal of Psychiatry. 2013;202:100–107. doi: 10.1192/bjp.bp.111.106666. [DOI] [PubMed] [Google Scholar]
  2. Bertone-Johnson ER. Vitamin D and the occurrence of depression: causal association or circumstantial evidence? Nutrition Reviews. 2009;67:481–492. doi: 10.1111/j.1753-4887.2009.00220.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Blalock SJ, Norton LL, Patel RA, Cabral K, Thomas CL. Development and assessment of a short instrument for assessing dietary intakes of calcium and vitamin D. Journal of the American Pharmacists Association. 2002;43:685–693. doi: 10.1331/154434503322642615. [DOI] [PubMed] [Google Scholar]
  4. Capuron L, Miller AH. Cytokines and psychopathology: lessons from interferon-alpha. Biological Psychiatry. 2004;56:819–824. doi: 10.1016/j.biopsych.2004.02.009. [DOI] [PubMed] [Google Scholar]
  5. Chesney E, Goodwin G, Fazel S. Risks of all-cause and suicide mortality in mental disorders: a meta-review. World Psychiatry. 2014;13:153–160. doi: 10.1002/wps.20128. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Choi K, Lazovich D, Southwell B, Forster J, Rolnick SJ, Jackson J. Prevalence and characteristics of indoor tanning use among men and women in the United States. Archives of Dermatology. 2010;146:1356–1361. doi: 10.1001/archdermatol.2010.355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Eisenberg D, Hunt J, Speer N. Mental health in American colleges and universities: variation across student subgroups and across campuses. The Journal of Nervous and Mental Disease. 2013;201:60–67. doi: 10.1097/NMD.0b013e31827ab077. [DOI] [PubMed] [Google Scholar]
  8. Eyles DW, Burne TH, McGrath JJ. Vitamin D, effects on brain development, adult brain function and the links between low levels of vitamin D and neuropsychiatric disease. Frontiers in Neuroendocrinology. 2013;34:47–64. doi: 10.1016/j.yfrne.2012.07.001. [DOI] [PubMed] [Google Scholar]
  9. Eyles DW, Smith S, Kinobe R, Hewison M, McGrath JJ. Distribution of the vitamin D receptor and 1α-hydroxylase in human brain. Journal of Chemical Neuroanatomy. 2005;29:21–30. doi: 10.1016/j.jchemneu.2004.08.006. [DOI] [PubMed] [Google Scholar]
  10. Ganji V, Milone C, Cody MM, McCarty F, Wang YT. Serum vitamin D concentrations are related to depression in young adult US population: the Third National Health and Nutrition Examination Survey. International Archives of Medicine. 2010;3:29–36. doi: 10.1186/1755-7682-3-29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Giovannucci E. Vitamin D and cancer incidence in the Harvard cohorts. Annals of Epidemiology. 2009;19:84–88. doi: 10.1016/j.annepidem.2007.12.002. [DOI] [PubMed] [Google Scholar]
  12. Greenberg PE, Kessler RC, Birnbaum HG, Leong SA, Lowe SW, Berglund PA, Corey-Lisle PK. The economic burden of depression in the United States: How did it change between 1990 and 2000? The Journal of Clinical Psychiatry. 2003;64:1465–1475. doi: 10.4088/jcp.v64n1211. [DOI] [PubMed] [Google Scholar]
  13. Holick MF. Vitamin D deficiency. The New England Journal of Medicine. 2007;357:266–281. doi: 10.1056/NEJMra070553. [DOI] [PubMed] [Google Scholar]
  14. Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, Heaney RP, Weaver CM. Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline. The Journal of Clinical Endocrinology & Metabolism. 2011;96:1911–1930. doi: 10.1210/jc.2011-0385. [DOI] [PubMed] [Google Scholar]
  15. Hollis BW. Circulating 25-hydroxyvitamin D levels indicative of vitamin D sufficiency: implications for establishing a new effective dietary intake recommendation for vitamin D. The Journal of Nutrition. 2005;135:317–322. doi: 10.1093/jn/135.2.317. [DOI] [PubMed] [Google Scholar]
  16. Jaddou HY, Batieha AM, Khader YS, Kanaan SH, El-Khateeb MS, Ajlouni KM. Depression is associated with low levels of 25-hydroxyvitamin D among Jordanian adults: results from a national population survey. European Archives of Psychiatry and Clinical Neuroscience. 2012;262:321–327. doi: 10.1007/s00406-011-0265-8. [DOI] [PubMed] [Google Scholar]
  17. Jorde R, Sneve M, Figenschau Y, Svartberg J, Waterloo K. Effects of vitamin D supplementation on symptoms of depression in overweight and obese subjects: randomized double blind trial. Journal of Internal Medicine. 2008;264:599–609. doi: 10.1111/j.1365-2796.2008.02008.x. [DOI] [PubMed] [Google Scholar]
  18. Kesby JP, Eyles DW, Burne TH, McGrath JJ. The effects of vitamin D on brain development and adult brain function. Molecular and Cellular Endocrinology. 2011;347:121–127. doi: 10.1016/j.mce.2011.05.014. [DOI] [PubMed] [Google Scholar]
  19. Kessler RC, Berglund P, Demler O, Jin R, Koretz D, Merikangas KR, Wang PS. The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R) JAMA. 2003;289:3095–3105. doi: 10.1001/jama.289.23.3095. [DOI] [PubMed] [Google Scholar]
  20. Khoraminya N, Tehrani-Doost M, Jazayeri S, Hosseini A, Djazayery A. Therapeutic effects of vitamin D as adjunctive therapy to fluoxetine in patients with major depressive disorder. Australian and New Zealand Journal of Psychiatry. 2013;47:271–275. doi: 10.1177/0004867412465022. [DOI] [PubMed] [Google Scholar]
  21. Kjærgaard M, Waterloo K, Wang CE, Almås B, Figenschau Y, Hutchinson MS, Jorde R. Effect of vitamin D supplement on depression scores in people with low levels of serum 25-hydroxyvitamin D: nested case–control study and randomised clinical trial. The British Journal of Psychiatry. 2012;201:360–368. doi: 10.1192/bjp.bp.111.104349. [DOI] [PubMed] [Google Scholar]
  22. Li G, Mbuagbaw L, Samaan Z, Falavigna M, Zhang S, Adachi JD, Thabane L. Efficacy of vitamin D supplementation in depression in adults: a systematic review. The Journal of Clinical Endocrinology & Metabolism. 2013;99:757–767. doi: 10.1210/jc.2013-3450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Milaneschi Y, Hoogendijk W, Lips P, Heijboer AC, Schoevers R, van Hemert AM, Penninx BWJH. The association between low vitamin D and depressive disorders. Molecular Psychiatry. 2013;19:444–451. doi: 10.1038/mp.2013.36. [DOI] [PubMed] [Google Scholar]
  24. Muthén LK, Muthén BO. Mplus (version 6.12) [computer software] Los Angeles: Author; 2011. [Google Scholar]
  25. National Cancer Institute. Diet History Questionnaire: current database file. [Accessed December 29, 2011]; [publication online]. Available at: http://riskfactor.cancer.gov/DHQ/database/current.html.
  26. Newman MS, Brandon TR, Groves MN, Gregory WL, Kapur S, Zava DT. A liquid chromatography/tandem mass spectrometry method for determination of 25-hydroxy vitamin D2 and 25-hydroxy vitamin D3 in dried blood spots: a potential adjunct to diabetes and cardiometabolic risk screening. Journal of Diabetes Science and Technology. 2009;3:156–162. doi: 10.1177/193229680900300118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Norman AW. Sunlight, season, skin pigmentation, vitamin D, 25-hydroxyvitamin D: integral components of the vitamin D endocrine system. The American Journal of Clinical Nutrition. 1998;67:1108–1110. doi: 10.1093/ajcn/67.6.1108. [DOI] [PubMed] [Google Scholar]
  28. Patrick RP, Ames BN. Vitamin D hormone regulates serotonin synthesis. Part 1: relevance for autism. The FASEB Journal. 2014;28:2398–2413. doi: 10.1096/fj.13-246546. [DOI] [PubMed] [Google Scholar]
  29. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Applied Psychological Measurement. 1977;1:385–401. [Google Scholar]
  30. Raison CL, Capuron L, Miller AH. Cytokines sing the blues: inflammation and the pathogenesis of depression. Trends in Immunology. 2006;27:24–31. doi: 10.1016/j.it.2005.11.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Rohan KJ, Roecklein KA, Haaga DAF. Biological and psychological mechanisms of seasonal affective disorder: a review and integration. Current Psychiatry Reviews. 2009;5:37–47. [Google Scholar]
  32. Rohde P, Lewinsohn PM, Klein DN, Seeley JR, Gau JM. Key characteristics of major depressive disorder occurring in childhood, adolescence, emerging adulthood, and adulthood. Clinical Psychological Science. 2013;1:41–53. doi: 10.1177/2167702612457599. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Sanders KM, Stuart AL, Williamson EJ, Jacka FN, Dodd S, Nicholson G, Berk M. Annual high-dose vitamin D3 and mental well-being: randomised controlled trial. The British Journal of Psychiatry. 2011;198:357–364. doi: 10.1192/bjp.bp.110.087544. [DOI] [PubMed] [Google Scholar]
  34. Shaffer JA, Edmondson D, Wasson LT, Falzon L, Homma K, Ezeokoli N, Davidson KW. Vitamin D supplementation for depressive symptoms: a systematic review and meta-analysis of randomized controlled trials. Psychosomatic Medicine. 2014;76:190–196. doi: 10.1097/PSY.0000000000000044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Silverman MN, et al. Immune modulation of the hypothalamic-pituitary-adrenal (HPA) axis during viral infection. Viral Immunology. 2005;18:41–78. doi: 10.1089/vim.2005.18.41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Smith MP, Fletcher-Turner A, Yurek DM, Cass WA. Calcitriol protection against dopamine loss induced by intracerebroventricular administration of 6-hydroxydopamine. Neurochemical Research. 2006;31:533–539. doi: 10.1007/s11064-006-9048-4. [PubMed: 16758362] [DOI] [PubMed] [Google Scholar]
  37. Suh J, Zhu B, Frei B. Ascorbate does not act as a pro-oxidant towards lipids and proteins in human plasma exposed to redox-active transition metal ions and hydrogen peroxide. Free Radical Biology and Medicine. 2003;34:1306–1314. doi: 10.1016/s0891-5849(03)00147-3. [DOI] [PubMed] [Google Scholar]
  38. Tangpricha V, Turner A, Spina C, Decastro S, Chen TC, Holick MF. Tanning is associated with optimal vitamin D status (serum 25-hydroxyvitamin D concentration) and higher bone mineral density. The American Journal of Clinical Nutrition. 2004;80:1645–1649. doi: 10.1093/ajcn/80.6.1645. [DOI] [PubMed] [Google Scholar]
  39. Zhang Y, Leung DYM, Richers BN, Liu Y, Remigio LK, Riches DW, Goleva E. Vitamin D inhibits monocyte/macrophage proinflammatory cytokine production by targeting MAPK phosphatase-1. The Journal of Immunology. 2012;188:2127–2135. doi: 10.4049/jimmunol.1102412. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Zhao G, Ford ES, Li C, Balluz LS. No associations between serum concentrations of 25-hydroxyvitamin D and parathyroid hormone and depression among US adults. British Journal of Nutrition. 2010;104:1696–1702. doi: 10.1017/S0007114510002588. [DOI] [PubMed] [Google Scholar]

RESOURCES