Association of serum folate levels during pregnancy and prenatal depression

Abstract

Objective:

To evaluate the association between serum folate levels during pregnancy and prenatal depression and the extent to which obesity may modify this relationship.

Methods:

This secondary data analysis leveraged data from a previous study of pregnant Kaiser Permanente Northern California participants who completed a survey and provided a serum sample between 2011 and 2013. Serum folate was assessed using the Center for Disease Control’s Total Folate Serum/Whole Blood Microbiological Assay Method. A score of 15 or greater on the Center for Epidemiologic Studies Depression Scale (CES-D) was defined as prenatal depression. We used Poisson regression to estimate risk of prenatal depression given prenatal serum folate status (low/medium tertiles vs high tertile) in the full sample and in subsamples of women with pre-pregnancy body mass index in the (a) normal range and (b) overweight/obese range.

Results:

Of the sample, 13% had prenatal depression. Combined low/medium folate tertiles was associated with prenatal depression (adjusted relative risk [aRR]=1.97, 95% confidence interval [CI]: 0.93, 4.18), although results did not reach statistical significance. This relationship was stronger among women with overweight/obesity than women with normal weight (aRR: 2.61, 95% CI: 1.01, 6.71 and aRR=1.50, 95% CI: 0.34, 6.66, respectively).

Conclusion:

Results suggest an association between lower pregnancy folate levels and prenatal depression that may be stronger among women with overweight or obesity. Future studies need to clarify the temporal sequence of these associations.

Introduction

Prenatal depression impacts 12–20% of pregnant women world-wide and is associated with adverse maternal and child health outcomes[1]. A recent meta-analysis reported significantly lower folate levels in individuals with depression compared to without[2]. Folate is linked to the synthesis of neurotransmitters involved in depression[3]. However, the limited research investigating the association between folate levels in pregnancy and prenatal depression has produced mixed findings [4–8].

Obesity alters the pharmacokinetic response to folic acid[9] and has been associated with an increased risk of depression[10]. Therefore, an association between folate levels and depression in pregnancy may vary by obesity status. The aims of this study were to (a) evaluate the association between relative serum folate levels during pregnancy and prenatal depression and (b) identify whether obesity is a potential modifier of the association.

Materials and methods

This secondary data analysis leveraged data from all pregnant Kaiser Permanente Northern California (KPNC) participants of the Study of how Pre-pregnancy and Pregnancy Lifestyle Influences the Outcome of Delivery (PEAPOD) (n=1,810) with a serum sample (n=318; mean gestational weeks at serum collection: 15 [SD:9]) [11]. Participants were pregnant between October 2011 and April 2013, English-speaking, and 18 years of age or older. Self-reported survey data included a prenatal depression screening questionnaire completed at a mean of 21 (SD: 8) gestational weeks. Those who provided a serum sample were more likely to have a college education and were primiparous (p<0.05) (Supplemental Table 1). This study was approved by the KPNC Institutional Review Board.

Measures

Serum folate and homocysteine

Total folate was measured using the procedures outlined in the Center for Disease Control’s Total Folate Serum/Whole Blood Microbiological Assay Method[12]. Serum folate levels were categorized as low/medium (the first and second tertile; 87.3 ± 20.1 nmol/L (mean ± SE)) and high (the third tertile; 154.4 ± 45.1 nmol/L (mean ± SE)) and similar to other studies using the microbiological assay to measure serum folate in pregnant women[13]. Serum homocysteine (μmol/L) was measured using Immulite ^™ chemiluminescent immunoassay system (Siemens, PA).

Prenatal depression

The validated Center for Epidemiologic Studies Depression Scale (CES-D) was used to assess prenatal depression [14,15]. Scores range from 0 to 60 and scores of 15 or greater indicate clinically significant depression symptoms and were categorized as prenatal depression.

Pre-pregnancy body mass index (BMI)

Pre-pregnancy BMI (kg/m²) was calculated using electronic health records and supplemented with self-report data when missing (13%).

Covariates

Covariates included in the model were self-reported maternal age, education, marital status, race and ethnicity, parity, any alcohol, smoking or vitamin use during pregnancy. Serum homocysteine was also included.

Data Analysis

We used chi-squared analyses (for categorical data) and t-tests (for continuous data) to compare characteristics between women with and without prenatal depression. We used Poisson regression[16] to estimate risk of prenatal depression in the full sample and stratified by BMI ((a) normal range (18.5 - < 25 kg/m²) and (b) in the overweight/obese range (>=25.0)). Due to small sample size underweight women were excluded from the stratified models. The analysis in the full sample adjusted for all covariates. The stratified models adjusted for all covariates except marital status and alcohol, smoking, and vitamin use during pregnancy which were excluded due to small sample sizes. Data were cleaned using SAS 9.4 (Cary, NC) and analyzed using R version 3.4.4.

Results

Participants with depression had higher pre-pregnancy BMI (p=0.05, 34.1% vs 18.4% overweight, 31.7% vs. 18.4% obese) and had lower levels of education (p=0.13, 61.0% vs. 79.8% graduated from college) (Supplemental Table 2). In the full sample, the combined low/medium folate tertiles were associated with depression during pregnancy (fully adjusted relative risk [RR] =1.97, 95% confidence interval [CI]: 0.93, 4.18; Table 1), although the results did not reach statistical significance. [Table 1 near here.] Analyses stratified by BMI indicated a stronger assocation among overweight/obese women (fully adjusted RR=2.61, 95% CI: 1.01, 6.71) compared to normal weight women (fully adjusted RR=1.50, 95% CI: 0.34, 6.66). However, we note CIs were wide and overlapping and some estimates did not reach statistical significance.

Table 1.

Associations between serum folate^* tertiles and depression during pregnancy

	Depression during Pregnancy n (%)	Unadjusted RR (95% CI)	Adjusted for homocysteine only RR (95% CI)	Fully adjusted RR (95% CI)
Full sample (n=318)
Low/medium folate tertiles (n=212)	33 (15.6)	2.06 (0.99, 4.31)	2.10 (1.00, 4.43)	1.97 (0.93, 4.18)
High folate tertile (n=106)	8 (7.5)	1	1	1
Normal weight (n=163)
Low/medium folate tertiles (n=106)	10 (9.2)	1.65 (0.47, 5.75)	1.66 (0.47, 5.88)	1.50 (0.34, 6.66)
High folate tertile (n=54)	3 (5.6)	1	1	1
Overweight/obese (n=150)
Low/medium folate tertiles (n=100)	22 (22)	2.20 (0.89, 5.46)	2.22 (0.89, 5.53)	2.61 (1.01, 6.71)
High folate tertiles (n=50)	5 (10)	1	1	1

Note: RR=relative risk. CI=confidence interval. ^aThe fully adjusted model fit in the whole sample adjusted for homocysteine (μmol/L), BMI (underweight [<18.5 kg/m2] or normal weight [18.5 to < 25 kg/m2] vs overweight [25 to < 30 kg/m2] or obese [30+ kg/m2]), maternal age at start of pregnancy (<35 versus 35+ years), marital status (married/partner versus single/divorced/widowed), maternal education (high school or less, some college, college or greater), race (Asian/Pacific Islander, Black, White, Other), ethnicity (Hispanic vs non-Hispanic), parity (0 versus 1 or more), alcohol use during pregnancy (yes versus no), smoking during pregnancy (yes versus no), and prenatal vitamin use (yes versus no). Fully adjusted model stratified by body mass index included the same covariates as the fully adjusted model fit in the full sample except for models stratified by pre-pregnancy adjusted for all covariates except marital status and alcohol, smoking during and vitamin use during pregnancy.

^*Serum folate was measured using the procedures outlined in the Center for Disease Control’s Total Folate Serum/Whole Blood Microbiological Assay Method No 4000.04 with the following modifications: 5-Methyl-THF Standard supplied by Sigma-Aldrich (St. Louis, MO, U.S.A.); Folic Acid Casei Medium (Catalog#: M543–100G) supplied by HiMedia (Mumbai, India); Lyphochek Immunoassay Plus Control (Catalog#: 370) supplied by Bio-Rad(Hercules, California, U.S.A.). Although the range of serum folate in our sample was higher than folate values reported in the literature using LC-MS/MS or immune assay with chemiluminescence analyzer, (RANGE 16.5–382.2) our findings are similar to the published literature using the microbiological assay.[13] Folate concentrations overall and by pre-pregnancy BMI status (median [interquartile ratio]): The median and IQR for serum folate among women with overweight/obesity was 103.7 (IQR: 84.2, 121.72) and among women with normal wegith was: 102.6 (IQR: 83.9, 126.4).

Discussion

Study findings suggest lower relative serum folate levels (lower/medium tertiles) may be associated with prenatal depression, yet the association was significant and stronger among pregnant women with overweight or obesity. These findings are consistent with two previous cross-sectional studies showing an association between lower folate levels and prenatal depression [4,5]. However, our findings are contrast with two smaller cross-sectional studies (n=86 and n=365) that did not find an association between serum folate levels and depression in pregnancy [7,8].

Our findings could be explained by several biological mechanisms. First, folate is involved in the synthesis and metabolism of neurotransmitters implicated in depression (serotonin, dopamine, norepinephrine) [3]. Second, folate is important in regenerating tetrahydrobiopterin (BH4), a cofactor essential to formation of neurotransmitters [3,17,18]. Further, the pharmacokinetic response to folic acid is altered by obesity. Women with obesity given the same dose of folic acid as women with normal weight demonstrate lower serum folate levels and lower overall serum folate response [9].

The four-year-plus time-lag between when samples were frozen and assayed may have affected folate results. Fasting was not required prior to serum collection and may explain the higher-than-normal folate levels in the sample. Thus, conclusions regarding serum folate levels are based on relative rather than absolute comparisons. Causality cannot be established as we are unable to determine whether folate levels preceded depressive symptoms. It is plausible that prenatal depression caused poor dietary intake leading to lower serum levels and we do not have information on depression or medical history. We were unable to assess the relationship between serum folate levels and prenatal depression among women with underweight given the small sample size.

Our findings suggest an association between lower relative serum folate levels and prenatal depression that may be stronger for those with overweight or obesity. However, future studies with larger samples of pregnant individuals with fasting serum samples are needed to clarify both the relationship with overweight and obesity and the direction of the association.