Folate, Vitamin B12, Homocysteine, and the MTHFR 677C→T Polymorphism in Anxiety and Depression: Title and subTitle BreakThe Hordaland Homocysteine Study

HOMOCYSTEINE and the vitamins involved in 1-carbon metabolism, folate and vitamin B₁₂, have been associated with a diversity of diseases, including cardiovascular disease,^{1 - 2} malignancies,³ Alzheimer disease, impaired cognitive functioning,^{4 - 5} and birth defects and pregnancy complications.^{6 - 7} Several studies have investigated the 677C→T polymorphism of the methylenetetrahydrofolate reductase (MTHFR) gene as a risk factor for these conditions, as the C-to-T transition causes reduced enzyme activity, and elevated plasma total homocysteine (tHcy) levels under conditions of impaired folate status.⁸

Both anxiety and depression are common symptoms or disorders with a major impact on public health.^{9 - 10} Although a possible role of nutritional factors in the pathogenesis of neuropsychiatric disorders has long been debated,^{11 - 12} clinical studies have shown an inverse relationship between folate status and depression.¹³ Such a relationship has been inferred from studies showing increased frequency of folate deficiency among depressed patients^{14 - 18} ;more severe^{18 - 21} and prolonged²² depressive episodes and weaker treatment response to antidepressants in patients with low folate status^{14 ,18 ,21 - 25} ;and enhanced antidepressant response with folic acid supplementation.^{26 - 29}

Investigations on a possible role of vitamin B₁₂ status in neuropsychiatric disorders have been motivated by the central nervous system damage caused by overt or subtle vitamin B₁₂ deficiency.^{30 - 31} Data regarding the association between vitamin B₁₂ status and depression are scarce.^{14 ,18 ,32 - 34}

While some studies have demonstrated a positive relationship between depression and plasma tHcy levels³⁵ or the MTHFR 677C→T polymorphism,³⁶ such relationships have not been confirmed in other studies.^{14 ,32 ,34 ,37}

Folate, vitamin B₁₂, and homocysteine are involved in processes important for central nervous system function (Figure 1). Folate metabolism is linked to biopterin-dependent neurotransmitter synthesis³⁸ and S-adenosylmethionine–dependent methylation of biogenic amines and phospholipids in the central nervous system.³⁹ Homocysteine, or its metabolites such as homocysteic acid, may have a direct excitotoxic effect on the N-methyl-D-aspartate glutamate receptors, or may inhibit the methylation processes in the central nervous system.³⁹ This biochemical knowledge adds to the clinical data cited already, suggesting a possible role of 1-carbon metabolism in mental disorders, particularly in depression and dementia.⁴⁰ However, despite the extensive comorbidity between depression and anxiety,^{41 - 43} we have found no more than 3 studies^{33 ,44 - 45} addressing the possibility of impaired 1-carbon metabolism in anxiety disorders. Only one study suggests such an association, namely, between low vitamin B₁₂ levels and anxiety.³³

The present study is part of the Hordaland Health Study, which included self-administered questionnaires of anxiety and depression according to the Hospital Anxiety and Depression Scale (HADS),⁴⁶ determination of tHcy and related vitamins, and C677T MTHFR genotyping in approximately 6000 middle-aged and elderly subjects. The aim of the current study was to examine whether key components of the 1-carbon metabolism were associated with anxiety disorders and/or depression in this large population.

The Hordaland Health Study 1997-1999 was conducted from 1997 to 1999 as a collaboration between the National Health Screening Service, the University of Bergen (Bergen, Norway), and local health services. A subsample of the study included 2291 men and 2558 women aged 46 to 49 years (mean, 47.4 years) and 1868 men and 2470 women aged 70 to 74 years (mean, 71.9 years) who had participated in the Hordaland Homocysteine Studies in 1992 to 1993.⁴⁷ The younger age group was originally a part of the ordinary national cardiovascular risk survey program, while the older group was included to examine age effects. Participation rate was 77% of those invited. At attendance, the subjects underwent a brief physical examination (height, weight, blood pressure), and a blood sample was drawn and stored at −20°C until the biochemical analyses were performed 1 to 3 years later. A self-administered questionnaire providing information on demographic, socioeconomic, and psychosocial parameters; health behaviors; subjective health; present or former diseases; and use of medication was delivered. Among those who participated, 5948 individuals(84%) returned questionnaires with valid ratings of anxiety and depression.

The study protocol was approved by the Regional Committee for Medical Research Ethics and the Norwegian Data Inspectorate.

The HADS is a self-administered questionnaire consisting of 14 items, 7 for anxiety (HADS-A subscale) and 7 for depression (HADS-D subscale).⁴⁶ To avoid making individuals feel that they are being tested for mental disorders, symptoms of severe psychopathology are not included. The HADS-A contains items mainly related to restlessness and worry, and 1 item reflecting panic attacks. The HADS-D focuses mainly on the reduced pleasure response aspect (anhedonia) of depression, but also psychomotor retardation and impaired mood.

The HADS-A and HADS-D are intercorrelated, most often in the range of 0.50 to 0.60.⁴⁸ Hence, to identify more homogeneous groups with anxiety disorders or depressions, restrictions were put on the other subscale when cases were defined. Thus, anxiety disorder was defined as a HADS-A score of 8 or more restricted to a HADS-D score less than 8 to avoid comorbid disorders. Accordingly, depression was defined as a HADS-D score of 8 or more restricted to a HADS-A score less than 8. Analyses were also carried out on comorbid cases, which were defined as subjects with both subscale scores of 8 or more. To enhance the specificity of anxiety disorder and depression, analyses were carried out on cases defined to HADS-A and HADS-D scores of 11 or more, using the same restrictions as mentioned in this paragraph.

Plasma tHcy was analyzed by high-performance liquid chromatography and fluorescence detection.⁴⁹ Serum folate was determined by a Lactobacillus casei microbiologic assay⁵⁰ and serum vitamin B₁₂ by a Lactobacillus leichmannii microbiologic assay.⁵¹ Both the folate and vitamin B₁₂ assays were adapted to a microtiter plate format and carried out by a robotic workstation (Micro-lab AT plus 2; Hamilton Bonaduz AG, Bonaduz, Switzerland). The MTHFR C677T genotyping was performed by a real-time polymerase chain reaction as described elsewhere.⁵²

Plasma tHcy level was divided into 4 categories (<9.0 µmol/L[<1.22 mg/dL] [reference], 9.0-11.9 µmol/L [1.22-1.61 mg/dL], 12.0-14.9 µmol/L [1.62-2.01 mg/dL], and ≥15.0 µmol/L [≥2.02 mg/dL]).⁵³ Plasma folate and plasma vitamin B₁₂ levels were similarly divided into 4 categories (plasma folate: <3.80 nmol/L [<1.7 ng/mL], 3.80-4.99 nmol/L [1.7-2.2 ng/mL], 5.00-8.49 nmol/L [2.3-3.7 ng/mL], and ≥8.50 nmol/L [≥3.8 ng/mL] [reference]; plasma vitamin B₁₂: <230.0 pmol/L [<312 pg/mL], 230.0-279.9 pmol/L [312-379 pg/mL], 280.0-414.9 pmol/L [380-562 pg/mL], and ≥415.0 pmol/L [≥563 pg/mL] [reference]). Logistic regression analyses were used to estimate odds ratios (ORs) for being a case, comparing each category with the reference category of the metabolites and the MTHFR polymorphism. The representation of covariates as indicator variables was used to allow for assessment of nonlinear dose-response relationships, while a linear (1 df) representation was used to test for linear trends. Two logistic regression models were used, one with adjustment for age and sex (model 1) and one with additional adjustments for smoking status and educational level (model 2). The effect of other possible confounders, such as coffee consumption, physical exercise, body mass index, and self-reported cardiovascular disease, was examined by adding these one by one to model 2. To evaluate possible effect modification of age and sex, product terms were added separately to the models. Possible effect modification of B-vitamin supplementation and tranquilizer and antidepressant use was evaluated by stratification.

To examine whether use of B-vitamin supplements was associated with anxiety or depression, logistic regression analyses were used to estimate OR for being a case, comparing nonusers with users after adjusting for age, sex, smoking status, and educational level.

The precision of the OR estimates was expressed with 95% confidence intervals (CIs). Generalized additive logistic regression was used to provide a graphic representation of the dose-response relationship between folate, vitamin B₁₂, and tHcy and anxiety or depression. This technique is based on the generalized additive model⁵⁴ and allowed adjustment for age, sex, smoking status, and educational level.

A 2-sided P<.05 was chosen to indicate statistical significance. The statistical analyses were conducted with the software package SPSS 11.0 (SPSS Inc, Chicago, Ill) and S-Plus 6.0 (Insightful Corp, Seattle, Wash).

Table 1 summarizes some demographic characteristics and blood indexes according to age groups and sex. The prevalences of anxiety disorder and depression; use of psychoactive drugs; distribution of tHcy, vitamin B₁₂, and folate levels; and some lifestyle factors are also presented. The total numbers of subjects with anxiety disorder and depression were 622 (11.5%) and 243 (4.8%), respectively. Anxiety disorder was most prevalent among middle-aged women (15.3%) and depression among older men (7.5%).

Plasma tHcy level correlated well with plasma folate value and modestly with plasma vitamin B₁₂ level (Spearman coefficient r = −0.40, P<.001, and r = −0.25, P<.001, respectively), and plasma folate and plasma vitamin B₁₂ levels correlated poorly(r = 0.06, P<.001), as has been reported elsewhere.⁵⁵ In the MTHFR T/T genotype, compared with the C/C and C/T genotypes, the mean value of plasma tHcy was significantly increased (13.66 µmol/L [1.85 mg/dL] vs 10.62 µmol/L [1.44 mg/dL] and 10.98 µmol/L [1.48 mg/dL]; P<.001 and P<.001, respectively) and mean plasma folate significantly decreased (7.49 nmol/L [3.3 ng/mL] vs 8.46 nmol/L [3.7 ng/mL] and 8.15 nmol/L [3.6 ng/mL]; P<.001 and P<.05, respectively), while mean plasma vitamin B₁₂ did not show significant differences between the genotypes (data not shown).

The proportions of the sample with anxiety disorder or depression according to the increasing concentrations of tHcy, folate, and vitamin B₁₂, and according to MTHFR genotype and some lifestyle factors, are listed in Table 2, stratified by age and sex. The distributions indicated associations with both anxiety disorder and depression for a variety of the variables listed. Plasma tHcy and folate levels and MTHFR genotype all seemed associated with depression, as did educational level, smoking status, coffee consumption, use of antidepressants and tranquilizers, physical exercise, and body mass index. Associations with anxiety disorder were seen for smoking status, use of antidepressants and tranquilizers, and body mass index, but only in some strata within plasma tHcy, folate, and vitamin B₁₂ levels.

Plasma levels of tHcy were not significantly associated with anxiety disorder (Table 3) but were significantly related to depression (Table 4), as demonstrated by logistic regression models adjusting for age and sex (model 1), or further adjusting for smoking status and educational level (model 2). The association with depression was, however, strongest at high plasma tHcy level (≥15.0 µmol/L [≥2.02 mg/dL]) (OR, 1.90; 95% CI, 1.11-3.25), and test for linear trend was significant only in model 1 (P = .03). Including folate level, vitamin B₁₂ level, or MTHFR genotypes in the analyses did not essentially influence the estimates.

Figure 2 shows a graphic representation of the relationship between tHcy and the vitamins, and anxiety disorder and depression. A dose-response relationship was seen only between tHcy level and depression.

Neither folate nor vitamin B₁₂ was significantly related to anxiety disorder or depression. These results were obtained by logistic regression also after adjustments (models 1 and 2, Table 3 and Table 4) and were supported by generalized additive logistic regression (Figure 2). However, the dose-response curves obtained by the latter approach suggested a weak negative relationship between folate and depression but not anxiety disorder, and the curve for the vitamin B₁₂ and depression relationship was actually U-shaped (Figure 2).

The estimates did not change markedly when analyses were performed separately for the 4 age and sex strata, except for middle-aged women. In this subgroup, ORs (95% CIs) were 3.41 (1.02-11.42) and 3.08 (0.97-9.72) (model 1) and 3.15(0.94-10.60) and 3.05 (0.96-9.65) (model 2) for the 2 lower folate categories, respectively. The P for trend was .02 in both models.

Depression but not anxiety disorder was related to the C677T MTHFR genotype. The OR (95% CI) for depression by logistic regression was 1.69 (1.09-2.62) for the T/T genotype (using C/C as reference) in model 2 (Table 4). This estimate was essentially the same when carried out in the 4 age and sex strata (data not shown).

Including coffee consumption, physical exercise, body mass index, or self-reported cardiovascular disease in the logistic regression analyses did not significantly influence the estimates. We observed no effect modification by age or sex or by use of B vitamins, tranquilizers, or antidepressants.

There were no significant associations between levels of plasma folate, vitamin B₁₂, or tHcy or MTHFR 677C→T polymorphism and comorbid anxiety disorder and depression. Nonusers of B-vitamin supplements were not more prone to anxiety or depression than users.

Elevating the cutoff for both anxiety disorder and depression to HADS-A score of 11 or more and HADS-D score of 11 or more, respectively, resulted in only minor changes in the estimates. The relationship with tHcy was weaker for anxiety disorder and stronger for depression; neither of them was significant. The OR for anxiety disorder and depression associated with folate or vitamin B₁₂ was unchanged, except for the relationship between vitamin B₁₂ and depression, which became significant at the lower vitamin B₁₂ level with an OR (95% CI) of 2.33 (1.05-5.21) in model 1 and 2.39 (1.07-5.36) in model 2. Significant effects of MTHFR T/T genotype were not obtained for anxiety disorder, but they were for depression. The OR (95% CI) increased to 2.65 (1.16-6.08) in model 1 and 2.75 (1.20-6.32) in model 2.

We investigated the association between anxiety disorders and depression and key components of the 1-carbon metabolism in a cohort of 5948 subjects. The strongest relationship was observed between the T/T MTHFR genotype and depression, and the association was present for both cutoff levels of depression. Associations were observed between tHcy and depression, lowest level of vitamin B₁₂ (<230.0 pmol/L [<312 pg/mL]) and depression with high cutoff (HADS-D score ≥11), and in middle aged women, between depression and folate. Only a weak relationship or no relationship was seen between anxiety disorder and tHcy, folate, or vitamin B₁₂ level or MTHFR genotype.

Although the cross-sectional designof this study allows association between parameters to be assessed, causality cannot be determined. Because anxiety disorders and depression may influence dietary habits, it is conceivable that these disorders may alter homocysteine and B-vitamin status through dietary changes. Among the parameters studied, only MTHFR genotype is a nonmodifiable trait.

Published studies on folate status and depression, most of which are case-control studies, demonstrate stronger relationships⁵⁶ than documented in the present study. Such case-control studies are more prone to selection bias than a population-based cross-sectional study. Although the primary participation rate of the present study was relatively high (77%), it might have been affected by mental disorders themselves, since persons with anxiety disorders and depression might be less likely to participate in the study.

The plasma levels of folate in the frozen samples were lower than would be expected in a fresh sample, in contrast to the plasma levels of vitamin B₁₂ and tHcy. Others have found similar deviations of plasma folate levels in frozen samples.⁵⁷ Nevertheless, plasma folate level was correlated with tHcy level (Spearman coefficient r = −0.40, P<.001), as reported elsewhere.⁵⁵

Serum (or plasma) folate level has been considered a less reliable measure of folate status than red blood cell folate content.⁵⁸ Hence, if red blood cell folate level had been measured, one would expect a stronger association with depression compared with our findings. However, because of substantial intermethod variation, red blood cell folate level has its shortcomings as a measure of folate deficiency as well.⁵⁹

The HADS is a dimensional rating scale, and categorical diagnoses have to be based on research-proved cutoff levels. Optimal cutoff levels for anxiety disorders and depressive disorders are at scores of 8 or more for both subscales, resulting in sensitivities and specificities of approximately 0.80 for both HADS-A and HADS-D.⁴⁸ Higher specificity is usually obtained by raising the cutoff level, which was demonstrated by the increased OR for depression in the MTHFR polymorphism, tHcy level, and vitamin B₁₂ level.

The observed association between the MTHFR T/T genotype and depression is in accordance with results from one case-control study,³⁶ while another case-control study did not confirm such an association.³⁷ These somewhat inconsistent results of smaller studies could be explained by low statistical power due to the limited number of cases (n = 71 and 32, respectively), combined with the low frequency (10%) of MTHFR T/T homozygosity.⁸

We found that depression was related to tHcy, folate, and vitamin B₁₂ levels, but the associations were weak and/or present only in certain subgroups. This contrasts somewhat with the findings in clinical studies of impaired folate status in depressed patients,⁵⁶ and one report on the high frequency of elevated tHcy levels in depression.¹⁸ The relatively weak associations found in the present work could be explained by better folate status and less severe depression of the subjects recruited from a general population as compared with hospitalized depressed patients.

Some antidepressant effect by folic acid supplementation has been reported in randomized clinical trials.^{26 - 29} The largest (n = 127) of these trials²⁶ showed, however, a significant beneficial effect only in women. The lack of significant results in men could be due to the small sample size or an insufficient dosage. In contrast, in our study, use of vitamin B supplements was not associated with a lower prevalence of depression. Notably, most of the B-vitamin supplements in Norway at the time of recruitment did not contain folic acid, or only in relatively low doses of 100 µg.

Associations between folate, vitamin B₁₂, and tHcy levels or the MTHFR 677C→T polymorphism and anxiety disorders have seldom been reported so far; this may be due to reporting or publication bias of negative findings. We found only a weak, nonsignificant relationship between plasma tHcy level and anxiety disorder, and no associations between any of the index variables and comorbid depression and anxiety disorder. Hence, our data suggest that impaired 1-carbon metabolism is related to the subgroup of depression without comorbid anxiety disorder. One possible explanation is that they have a different cause. Further subgrouping to obtain more specific categories, eg, unipolar vs bipolar depression and lifetime or recurrent depression, was not possible when HADS was used in this cross-sectional design. Such analyses of subgroups should be carried out in future studies.

In summary, results from this large population-based study suggest a role of impaired 1-carbon metabolism in depression without comorbid anxiety disorder. This conclusion is supported by increased risk among subjects with the MTHFR T/T genotype, which has an effect on folate distribution and thereby tHcy level. The observation that the T/T genotype confers increased risk suggests that altered B-vitamin status may be a risk factor for, rather than the result of, depression. However, our results are preliminary and there is a need for prospective studies. Furthermore, a dose-response relationship between B vitamins, metabolic markers, and clearly defined subgroups of depression should be investigated. More adequately sized double-blind randomized trials are warranted as well.

Corresponding author: Ingvar Bjelland, MD, Department of Public Health and Primary Health Care, University of Bergen, Armauer Hansen Building, Haukeland Hospital, N-5021 Bergen, Norway (e-mail: ingvar.bjelland@uib.no).

Submitted for publication September 17, 2002; final revision received January 8, 2003; accepted January 9, 2003.

This study was supported by the Norwegian Research Council, Oslo. Dr Bjelland has research grants from H. Lundbeck Ltd (Norway), Lysaker, and Upjohn Ltd (Norway), Oslo.

We thank Alv A. Dahl, MD, PhD, for his advice and encouragement in writing and revising the drafts of this article.