Abstract

The authors performed a meta-analysis of studies examining the association between polymorphisms in the 5,10-methylenetetrahydrofolate reductase (MTHFR) gene, including MTHFR C677T and A1298C, and common psychiatric disorders, including unipolar depression, anxiety disorders, bipolar disorder, and schizophrenia. The primary comparison was between homozygote variants and the wild type for MTHFR C677T and A1298C. For unipolar depression and the MTHFR C677T polymorphism, the fixed-effects odds ratio for homozygote variants (TT) versus the wild type (CC) was 1.36 (95% confidence interval (CI): 1.11, 1.67), with no residual between-study heterogeneity (I2 = 0%)—based on 1,280 cases and 10,429 controls. For schizophrenia and MTHFR C677T, the fixed-effects odds ratio for TT versus CC was 1.44 (95% CI: 1.21, 1.70), with low heterogeneity (I2 = 42%)—based on 2,762 cases and 3,363 controls. For bipolar disorder and MTHFR C677T, the fixed-effects odds ratio for TT versus CC was 1.82 (95% CI: 1.22, 2.70), with low heterogeneity (I2 = 42%)—based on 550 cases and 1,098 controls. These results were robust to various sensitively analyses. This meta-analysis demonstrates an association between the MTHFR C677T variant and depression, schizophrenia, and bipolar disorder, raising the possibility of the use of folate in treatment and prevention.

GENE

The 5,10-methylenetetrahydrofolate reductase (MTHFR) gene is located at the end of the short arm of chromosome 1 (1p36.3). The enzyme plays a central role in folate metabolism by irreversibly converting 5,10-methylenetetrahydrofolate to 5-methylenetetrahydrofolate, the predominant circulating form of folate. 5-Methylenetetrahydrofolate donates a methyl group to homocysteine in the generation of S-adenosylmethionine, a major source of methyl groups in the brain (1).

GENE VARIANTS

Two common single nucleotide polymorphisms in MTHFR have been reported, a C→T transition at nucleotide 677 in exon 4 and an A→C transversion in exon 7 at position 1298. Both of these polymorphisms are functional and result in diminished enzyme activity. For the C677T polymorphism, homozygote variants have 30 percent enzyme activity in comparison with homozygotes for the wild-type C allele, while heterozygotes retain 65 percent of wild-type MTHFR enzyme activity (2). The consequences of the C677T polymorphism have been demonstrated in population studies, where lower levels of red blood cell folate, plasma folate, and vitamin B12 have been reported among nondiseased persons with the 677 TT genotype in comparison with persons with other genotypes (3). The 1298 polymorphism has been less extensively studied; however, it is known that persons with the 1298 CC genotype have approximately 60 percent of the enzyme activity of those with the common AA genotype (4).

The frequency of the C677T allele is subject to considerable ethnic and geographic variation, with a high allele frequency being reported in California Hispanics and a low allele frequency in US Blacks. There is also marked variation in the frequency of C677T homozygote variants between populations. The highest frequency (>20 percent) is found among US Hispanics, Colombians, and Amerindians in Brazil; conversely, in Black populations, less than 2 percent have the variant genotype (5). Among White populations in Europe, North America, and Australia, the frequency ranges from 8 percent to 20 percent, although interestingly in Europe there seems to be a drift in the occurrence of the homozygote variant from north to south. The prevalence of the A1298C homozygote variant genotype ranges from 7 percent to 12 percent in White populations from North America and Europe. Lower frequencies have been reported in Hispanics (4–5 percent), Chinese (1–4 percent), and Asian populations (1–4 percent). (For more information, see the previously published Human Genome Epidemiology (HuGE) reviews by Botto and Yang (5), Robein and Ulrich (6), and Sharp and Little (7).)

DISEASES

The diseases covered in this review are the common functional psychiatric disorders: major unipolar depression, anxiety disorders, bipolar affective disorder, and schizophrenia (8). Organic disorders and disorders of addiction are not covered in this review. Common functional psychiatric disorders contribute 15 percent of the global burden of disease (9), and each has been shown to have both a genetic etiology and an environmental etiology (10, 11).

Depression is the most common psychiatric disorder, with a 12-month incidence of 6–12 percent (12, 13). It is characterized by low mood and biologic disturbances of sleep, volition, and physical and mental activity (14).

Anxiety disorders are common, with an annual incidence of 5 percent (15). They are characterized by excessive, inappropriate, and disabling anxiety, accompanied by autonomic symptoms and catastrophic thoughts (14).

Schizophrenia is the most common functional psychotic illness, with a lifetime incidence of 1–2 percent (14). It is characterized by disorders of thought and volition and by abnormal perceptions (hallucinations) and beliefs (delusions) (14).

Bipolar disorder (manic depressive illness) is much less common than unipolar depression, with an annual incidence of 0.5 percent (15). It is characterized by alternating and severe disorders of mood (elation and depression), with accompanying shifts in energy and volition (14).

Functional psychiatric disorders, including depression (16), schizophrenia (17), and bipolar disorder (18), have been linked to low folate levels (19) and defective folate metabolism (16). The most extensively studied link is between a low folate level and depression, where an association has been demonstrated and a therapeutic role for folate has been proposed (16). However, evidence of association comes from either case-control studies, which do not control for important confounding factors (20, 21), or cross-sectional studies, where the direction of causality is impossible to establish (22). In contrast, associations with functional MTHFR polymorphisms are less susceptible to confounding or reverse causality.

We carried out a meta-analysis to examine whether genetic polymorphisms in MTHFR were associated with common psychiatric disorders.

MATERIALS AND METHODS

We conducted a systematic review and meta-analysis in accordance with the guidelines provided by the Human Genome Epidemiology Network (HuGE Net) (23–25).

Inclusion criteria

We included all studies which examined an association between genetic polymorphisms of MTHFR and a functional psychiatric disorder. Nonfamilial cross-sectional, case-control, and cohort studies were included. The presence of a psychiatric disorder was established by means of any standardized method, including a validated diagnostic instrument or a standardized diagnostic interview (8, 14). Studies using self-reported diagnoses were included, but the robustness of all analyses was tested to assess their impact.

Search strategy

The following databases were searched from their inception to May 15, 2006, with no language restrictions: MEDLINE, EMBASE, PsycINFO, BIOSIS, HuGE Net, Genetics Abstracts, and Science Direct. We used exploded subject headings, synonyms, and free text terms for folate, MTHFR, MTHFR polymorphisms (C677T and A1298C), and psychiatric disorders. Reference lists were also scrutinized for relevant studies.

Data extraction

Two investigators independently extracted the following data from each publication: author; country of origin; selection and characteristics of cases and controls; demographic information; racial descent of the study population; numbers of eligible and genotyped cases and controls; and numbers of cases and controls for each MTHFR genotype. Particular care was taken to avoid “double counting” of evidence in cases where papers had the same authors. We also examined matching and blinding to the clinical status of the subjects. We sought assistance from authors in situations where data were missing, ambiguous, or incomplete.

Meta-analysis

Our primary analysis for MTHFR C677T compared the homozygote variant with the homozygote wild type (TT vs. CC) for patients versus controls. We also examined C-versus-T allele frequency and contrasted persons who were homozygous wild-type (CC) with heterozygotes (CT). In addition, we tested for deviations from Hardy-Weinberg equilibrium in control populations, using the exact method (26).

We combined odds ratios using both fixed-effects and random-effects models (27). In subgroup analyses, we estimated race-specific odds ratios for each allele contrast. We also performed cumulative meta-analysis (28) to evaluate whether the association changed over time. We estimated between-study heterogeneity using the I2 statistic (29). As a guide, I2 values of 25 percent may be considered “low,” values of 50 percent may be considered “moderate,” and values of 75 percent may be considered “high” (29).

Genetic association studies may be especially susceptible to the selective publication of positive findings on gene associations (30). To test for publication bias, we applied funnel plot analysis and a regression test (31). We evaluated whether the summary results were different when the analyses were limited to studies with rigorous selection of cases (those that confirmed the diagnosis using a standardized interview) or when we excluded studies that were in Hardy-Weinberg disequilibrium (30). We also adjusted our meta-analyses for control group Hardy-Weinberg disequilibrium according to a recently introduced method (32).

Where analyses were stratified by race, quality, disequilibrium, or diagnostic method, we tested for relations between study-level variables and relations between odds ratios by means of logistic meta-regression (33), using a permutation test (with 1,000 Monte Carlo simulations) to calculate p values and to reduce the chance of spurious false-positive findings (34). The amount of heterogeneity explained by the use of predictive covariates was examined by reductions in the I2 inconsistency statistic.

All analyses were conducted in Stata, version 8 (Stata Corporation, College Station, Texas), using the metan, metabias, metareg, metacum, and genhw commands.

META-ANALYSIS RESULTS

Depression

Searches identified 10 studies (11,709 participants; 1,280 cases of depression, 10,429 controls) meeting our inclusion criteria. All studies examined the MTHFR C677T polymorphism (22, 35–43), and one study (38) also examined the MTHFR A1298C polymorphism.

Three studies (22, 41, 43) were cross-sectional population-based studies, and seven studies (22, 35–42) used a case-control design. All case-control studies used clinician-diagnosed cases of major unipolar depression according to accepted criteria and unrelated control subjects. Three studies screened controls for either current or previous depression (36, 39, 42). One cross-sectional study (43) used patient self-report and current use of antidepressants as a method of case identification, in combination with self-report on a nonspecific depression/anxiety question on the EuroQol instrument (44). All studies employed polymerase chain reaction methods in genotyping (see tables 1 and 2 for study details).

TABLE 1

Racial origin and case/control characteristics of participants in studies of methylenetetrahydrofolate reductase (MTHFR) genetic polymorphisms

First author and year (ref. no.) Country Racial descent Cases Controls/comparators 
Depression 
Tan 2004 (36Singapore (Chinese) Asian DSM-IV* (14) major depression. No other psychiatric comorbidity (n = 88). Racially matched unrelated hospital employees with no psychiatric history (n = 120) 
Chen 2005 (37Taiwan Asian DSM-IV diagnosis of major depression made by Geriatric Mental State—late onset (>60 years). No other psychiatric comorbidity (n = 39). Age-matched unrelated controls (n = 20) 
Reif 2005 (38Germany European Consecutive female inpatients with ICD-10* (8) major depression (n = 45) Nonmatched healthy blood donors from the same region. Not screened for depression (n = 284). 
Hickie 2001 (39Australia European DSM-IV major depression. Mean age 55 years (n = 78). Nonmatched healthy unrelated controls recruited by newspaper and screened for current and previous depression (n = 22) 
Arinami 1997 (40Japan Asian DSM-III* (68) depression. Persons with heart disease excluded. Mean age 54 years (n = 32). Nonmatched unrelated controls. Mean age 49 years. Screening for previous history unclear (n = 419). 
Kunugi 1998 (35Japan Asian DSM-IV major depression diagnosed by two psychiatrists. Mean age 55 years (n = 71). Nonmatched healthy controls. Not formally screened for psychiatric disorder. Mean age 31 years (n = 258). 
Almeida 2005 (41Australia European Beck Depression Inventory score ≥13 (n = 42) Nonmatched unrelated controls with Beck Depression Inventory score <13. No screening for previous history (n = 198). 
Bjelland 2003 (22Norway European HADS* depression score ≥8 (n = 243) Nonmatched unrelated controls with HADS depression score <8. No screening for previous history (n = 4,806). 
Kelly 2004 (42Northern Ireland European ICD-10 depression diagnosed by clinician. Mean age 48 years (n = 100). Age- and sex-matched unrelated controls screened for previous and current depression (n = 89) 
Lewis 2006 (43England European Women with self-reported history of ever having been diagnosed with depression (n = 545) Women with no self-report of depression (n = 2,942) 
Schizophrenia 
Muntjewerff 2003 (45Netherlands European DSM-IV schizophrenia diagnosed by clinician. Males and females. Mean age 39 years (n = 35). Unrelated healthy hospital employee controls. Mean age 36 years (n = 104). 
Tan 2004 (36Singapore (Chinese) Asian DSM-IV schizophrenia diagnosed by clinician. No other psychiatric comorbidity (n = 236). Racially matched unrelated hospital employees with no psychiatric history (n = 120) 
Joober 2000 (46Canada European DSM-IV schizophrenia diagnosed by clinician. No other psychiatric comorbidity (n = 90). Racially matched unrelated controls with no psychiatric history (n = 105) 
Sazci 2003 (47Turkey European DSM-IV schizophrenia diagnosed by clinician (n = 130) Age-, sex-, and racially matched unrelated healthy controls with no history of schizophrenia over three generations (n = 226) 
Sazci 2005 (48Turkey European DSM-IV schizophrenia diagnosed by clinician (n = 297) Age-, sex-, and racially matched unrelated healthy controls (n = 341) 
Virgos 1999 (49Spain European DSM-IV schizophrenia diagnosed by clinician (n = 210) Nonmatched healthy controls (n = 218) 
Muntjewerff 2005 (50Netherlands European DSM-IV schizophrenia diagnosed by clinician. Males and females. Mean age 39 years (n = 254). Unrelated healthy hospital employee controls. Mean age 36 years (n = 414). 
Yu 2004 (51China Asian DSM-IV schizophrenia diagnosed by clinician (n = 230) Nonmatched healthy controls (n = 251) 
Yu 2004 (51Scotland European DSM-IV schizophrenia diagnosed by clinician (n = 426) Nonmatched healthy controls (n = 628) 
Kunugi 1998 (35Japan Asian DSM-IV schizophrenia diagnosed by two psychiatrists. Mean age 48 years (n = 343). Nonmatched healthy controls. Not formally screened for psychiatric disorder. Mean age 31 years (n = 258). 
Arinami 1997 (40Japan Asian DSM-III schizophrenia. Persons with heart disease excluded. Mean age 45 years (n = 297). Nonmatched unrelated controls. Mean age 49 years. Screening for previous history unclear (n = 419). 
Kempisty 2006 (52Poland European DSM-IV schizophrenia diagnosed by two psychiatrists. Mean ages 30 years and 33 years in males and females, respectively (n = 200). Nonmatched unrelated controls (blood donors). Mean ages 45 years and 40 years in males and females, respectively. No screening for psychiatric disorder (n = 300). 
Bipolar disorder 
Kunugi 1998 (35Japan Asian DSM-IV bipolar disorder diagnosed by two psychiatrists. Mean age 48 years (n = 143). Nonmatched healthy controls. Not formally screened for psychiatric disorder. Mean age 31 years (n = 258). 
Arinami 1997 (40Japan Asian DSM-III bipolar disorder. Persons with heart disease excluded. Mean age 54 years (n = 40). Nonmatched unrelated controls. Mean age 49 years. Screening for previous history unclear (n = 419). 
Tan 2004 (36Singapore (Chinese) Asian DSM-IV bipolar disorder diagnosed by clinician. No psychiatric comorbidity (n = 167). Racially matched unrelated hospital employees with no psychiatric history (n = 120) 
Kempisty 2006 (52Poland European DSM-IV bipolar disorder diagnosed by two psychiatrists. Mean ages 44 years and 46 years in males and females, respectively (n = 200). Nonmatched unrelated controls (blood donors). Mean ages 45 years and 40 years in males and females, respectively. No screening for psychiatric disorder (n = 300). 
Anxiety 
Bjelland 2003 (22Norway European HADS anxiety score ≥8 (n = 621) Nonmatched unrelated controls with HADS anxiety score <8. No screening for previous history (n = 6,185). 
First author and year (ref. no.) Country Racial descent Cases Controls/comparators 
Depression 
Tan 2004 (36Singapore (Chinese) Asian DSM-IV* (14) major depression. No other psychiatric comorbidity (n = 88). Racially matched unrelated hospital employees with no psychiatric history (n = 120) 
Chen 2005 (37Taiwan Asian DSM-IV diagnosis of major depression made by Geriatric Mental State—late onset (>60 years). No other psychiatric comorbidity (n = 39). Age-matched unrelated controls (n = 20) 
Reif 2005 (38Germany European Consecutive female inpatients with ICD-10* (8) major depression (n = 45) Nonmatched healthy blood donors from the same region. Not screened for depression (n = 284). 
Hickie 2001 (39Australia European DSM-IV major depression. Mean age 55 years (n = 78). Nonmatched healthy unrelated controls recruited by newspaper and screened for current and previous depression (n = 22) 
Arinami 1997 (40Japan Asian DSM-III* (68) depression. Persons with heart disease excluded. Mean age 54 years (n = 32). Nonmatched unrelated controls. Mean age 49 years. Screening for previous history unclear (n = 419). 
Kunugi 1998 (35Japan Asian DSM-IV major depression diagnosed by two psychiatrists. Mean age 55 years (n = 71). Nonmatched healthy controls. Not formally screened for psychiatric disorder. Mean age 31 years (n = 258). 
Almeida 2005 (41Australia European Beck Depression Inventory score ≥13 (n = 42) Nonmatched unrelated controls with Beck Depression Inventory score <13. No screening for previous history (n = 198). 
Bjelland 2003 (22Norway European HADS* depression score ≥8 (n = 243) Nonmatched unrelated controls with HADS depression score <8. No screening for previous history (n = 4,806). 
Kelly 2004 (42Northern Ireland European ICD-10 depression diagnosed by clinician. Mean age 48 years (n = 100). Age- and sex-matched unrelated controls screened for previous and current depression (n = 89) 
Lewis 2006 (43England European Women with self-reported history of ever having been diagnosed with depression (n = 545) Women with no self-report of depression (n = 2,942) 
Schizophrenia 
Muntjewerff 2003 (45Netherlands European DSM-IV schizophrenia diagnosed by clinician. Males and females. Mean age 39 years (n = 35). Unrelated healthy hospital employee controls. Mean age 36 years (n = 104). 
Tan 2004 (36Singapore (Chinese) Asian DSM-IV schizophrenia diagnosed by clinician. No other psychiatric comorbidity (n = 236). Racially matched unrelated hospital employees with no psychiatric history (n = 120) 
Joober 2000 (46Canada European DSM-IV schizophrenia diagnosed by clinician. No other psychiatric comorbidity (n = 90). Racially matched unrelated controls with no psychiatric history (n = 105) 
Sazci 2003 (47Turkey European DSM-IV schizophrenia diagnosed by clinician (n = 130) Age-, sex-, and racially matched unrelated healthy controls with no history of schizophrenia over three generations (n = 226) 
Sazci 2005 (48Turkey European DSM-IV schizophrenia diagnosed by clinician (n = 297) Age-, sex-, and racially matched unrelated healthy controls (n = 341) 
Virgos 1999 (49Spain European DSM-IV schizophrenia diagnosed by clinician (n = 210) Nonmatched healthy controls (n = 218) 
Muntjewerff 2005 (50Netherlands European DSM-IV schizophrenia diagnosed by clinician. Males and females. Mean age 39 years (n = 254). Unrelated healthy hospital employee controls. Mean age 36 years (n = 414). 
Yu 2004 (51China Asian DSM-IV schizophrenia diagnosed by clinician (n = 230) Nonmatched healthy controls (n = 251) 
Yu 2004 (51Scotland European DSM-IV schizophrenia diagnosed by clinician (n = 426) Nonmatched healthy controls (n = 628) 
Kunugi 1998 (35Japan Asian DSM-IV schizophrenia diagnosed by two psychiatrists. Mean age 48 years (n = 343). Nonmatched healthy controls. Not formally screened for psychiatric disorder. Mean age 31 years (n = 258). 
Arinami 1997 (40Japan Asian DSM-III schizophrenia. Persons with heart disease excluded. Mean age 45 years (n = 297). Nonmatched unrelated controls. Mean age 49 years. Screening for previous history unclear (n = 419). 
Kempisty 2006 (52Poland European DSM-IV schizophrenia diagnosed by two psychiatrists. Mean ages 30 years and 33 years in males and females, respectively (n = 200). Nonmatched unrelated controls (blood donors). Mean ages 45 years and 40 years in males and females, respectively. No screening for psychiatric disorder (n = 300). 
Bipolar disorder 
Kunugi 1998 (35Japan Asian DSM-IV bipolar disorder diagnosed by two psychiatrists. Mean age 48 years (n = 143). Nonmatched healthy controls. Not formally screened for psychiatric disorder. Mean age 31 years (n = 258). 
Arinami 1997 (40Japan Asian DSM-III bipolar disorder. Persons with heart disease excluded. Mean age 54 years (n = 40). Nonmatched unrelated controls. Mean age 49 years. Screening for previous history unclear (n = 419). 
Tan 2004 (36Singapore (Chinese) Asian DSM-IV bipolar disorder diagnosed by clinician. No psychiatric comorbidity (n = 167). Racially matched unrelated hospital employees with no psychiatric history (n = 120) 
Kempisty 2006 (52Poland European DSM-IV bipolar disorder diagnosed by two psychiatrists. Mean ages 44 years and 46 years in males and females, respectively (n = 200). Nonmatched unrelated controls (blood donors). Mean ages 45 years and 40 years in males and females, respectively. No screening for psychiatric disorder (n = 300). 
Anxiety 
Bjelland 2003 (22Norway European HADS anxiety score ≥8 (n = 621) Nonmatched unrelated controls with HADS anxiety score <8. No screening for previous history (n = 6,185). 
*

DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; ICD-10, International Classification of Diseases, Tenth Revision; DSM-III, Diagnostic and Statistical Manual of Mental Disorders, Third Edition; HADS, Hospital Anxiety and Depression Scale.

TABLE 2

Distribution of C677T genotypes in association studies of methylenetetrahydrofolate reductase (MTHFR) genetic polymorphisms and susceptibility to psychiatric disorders

First author and year (ref. no.) CC CT TT T allele Hardy-Weinberg p value* 
No. of cases No. of controls No. of cases No. of controls No. of cases No. of controls No. of cases No. of controls 
Depression 
Tan 2004 (3649 55.7 80 66.1 34 38.6 34 28.1 5.7 5.8 44 25.0 47 19.6 0.16 
Chen 2005 (3722 56.4 11 55.0 15 38.5 45.0 5.1 0.0 19 24.4 22.5 0.53 
Reif 2005 (3823 50.0 75 42.6 17 37.0 80 45.5 13.0 21 11.9 29 31.5 122 34.7 1.00 
Hickie 2001 (3933 44.0 12 54.5 33 44.0 40.9 12.0 4.5 34.0 25.0 1.00 
Arinami 1997 (4028.1 154 36.8 14 43.8 214 51.1 28.1 51 12.2 32 50.0 316 37.7 0.10 
Almeida 2005 (4113 31.0 85 42.9 26 61.9 87 43.9 7.1 26 13.1 32 38.1 139 35.1 0.64 
Bjelland 2003 (22127 52.5 3,381 49.7 85 35.1 2,864 42.1 30 12.4 561 8.2 145 30.0 3,986 29.3 0.19 
Kelly 2004 (4230 30.0 40 44.9 56 56.0 37 41.6 14 14.0 12 13.5 84 42.0 61 34.3 0.48 
Kunugi 1998 (3530 42.3 95 36.8 31 43.7 129 50.0 10 14.1 34 13.2 51 35.9 197 38.2 0.43 
Lewis 2006 (43221 40.6 1,344 45.7 251 46.1 1,269 43.1 73 13.4 329 11.2 397 36.4 1,927 32.7 0.26 
Schizophrenia 
Muntjewerff 2003 (4515 44.1 45 46.4 16 47.1 35 36.1 8.8 17 17.5 22 32.3 69 35.6 0.05 
Tan 2004 (36136 57.6 80 66.1 84 35.6 34 28.1 16 6.8 5.8 116 24.6 48 19.8 0.25 
Joober 2000 (4630 28.6 41 45.6 52 49.5 36 40.0 23 21.9 13 14.4 98 46.7 62 34.4 0.35 
Sazci 2003 (4759 45.4 106 46.9 49 37.7 103 45.6 22 16.9 17 7.5 93 35.8 137 30.3 0.27 
Sazci 2005 (48144 48.5 161 47.2 115 38.7 156 45.7 38 12.8 24 7.0 191 32.2 204 29.9 0.47 
Virgos 1999 (4981 38.6 79 36.2 98 46.7 106 48.6 31 14.8 33 15.1 160 38.0 172 39.4 0.89 
Muntjewerff 2005 (50112 44.1 212 51.2 111 43.7 166 40.1 31 12.2 36 8.7 173 34.1 238 28.7 0.72 
Yu 2004 (51199 46.7 306 48.7 186 43.7 260 41.4 41 9.6 62 9.9 268 31.5 384 30.6 0.57 
Yu 2004 (5191 39.6 85 33.9 96 41.7 126 50.2 43 18.7 40 15.9 182 39.6 206 41.0 0.60 
Kunugi 1998 (35121 35.3 95 36.8 168 49.0 129 50.0 54 15.7 34 13.2 276 40.2 197 38.2 0.43 
Arinami 1997 (4096 32.3 154 36.8 138 46.5 214 51.1 63 21.2 51 12.2 264 44.4 316 37.7 0.10 
Kempisty 2006 (52113 56.5 210 70.0 68 34.0 79 26.3 19 9.5 11 3.7 106 26.5 101 16.8 0.30 
Bipolar disorder 
Kunugi 1998 (3541 28.7 95 36.8 74 51.7 129 50.0 28 19.6 34 13.2 130 45.5 197 38.2 0.43 
Arinami 1997 (4015 37.5 154 36.8 20 50.0 214 51.1 12.5 51 12.2 30 37.5 316 31.7 0.10 
Tan 2004 (3699 59.3 80 66.1 60 35.9 34 28.1 4.8 5.8 76 22.7 48 19.8 0.25 
Kempisty 2006 (52108 54.0 210 70.0 73 36.5 79 26.3 19 9.5 11 3.7 111 27.5 101 16.8 0.30 
Anxiety 
Bjelland 2003 (22308 49.6 3,073 49.7 263 42.4 2,601 42.1 50 8.1 511 8.3 363 29.2 3,623 29.3 0.19 
First author and year (ref. no.) CC CT TT T allele Hardy-Weinberg p value* 
No. of cases No. of controls No. of cases No. of controls No. of cases No. of controls No. of cases No. of controls 
Depression 
Tan 2004 (3649 55.7 80 66.1 34 38.6 34 28.1 5.7 5.8 44 25.0 47 19.6 0.16 
Chen 2005 (3722 56.4 11 55.0 15 38.5 45.0 5.1 0.0 19 24.4 22.5 0.53 
Reif 2005 (3823 50.0 75 42.6 17 37.0 80 45.5 13.0 21 11.9 29 31.5 122 34.7 1.00 
Hickie 2001 (3933 44.0 12 54.5 33 44.0 40.9 12.0 4.5 34.0 25.0 1.00 
Arinami 1997 (4028.1 154 36.8 14 43.8 214 51.1 28.1 51 12.2 32 50.0 316 37.7 0.10 
Almeida 2005 (4113 31.0 85 42.9 26 61.9 87 43.9 7.1 26 13.1 32 38.1 139 35.1 0.64 
Bjelland 2003 (22127 52.5 3,381 49.7 85 35.1 2,864 42.1 30 12.4 561 8.2 145 30.0 3,986 29.3 0.19 
Kelly 2004 (4230 30.0 40 44.9 56 56.0 37 41.6 14 14.0 12 13.5 84 42.0 61 34.3 0.48 
Kunugi 1998 (3530 42.3 95 36.8 31 43.7 129 50.0 10 14.1 34 13.2 51 35.9 197 38.2 0.43 
Lewis 2006 (43221 40.6 1,344 45.7 251 46.1 1,269 43.1 73 13.4 329 11.2 397 36.4 1,927 32.7 0.26 
Schizophrenia 
Muntjewerff 2003 (4515 44.1 45 46.4 16 47.1 35 36.1 8.8 17 17.5 22 32.3 69 35.6 0.05 
Tan 2004 (36136 57.6 80 66.1 84 35.6 34 28.1 16 6.8 5.8 116 24.6 48 19.8 0.25 
Joober 2000 (4630 28.6 41 45.6 52 49.5 36 40.0 23 21.9 13 14.4 98 46.7 62 34.4 0.35 
Sazci 2003 (4759 45.4 106 46.9 49 37.7 103 45.6 22 16.9 17 7.5 93 35.8 137 30.3 0.27 
Sazci 2005 (48144 48.5 161 47.2 115 38.7 156 45.7 38 12.8 24 7.0 191 32.2 204 29.9 0.47 
Virgos 1999 (4981 38.6 79 36.2 98 46.7 106 48.6 31 14.8 33 15.1 160 38.0 172 39.4 0.89 
Muntjewerff 2005 (50112 44.1 212 51.2 111 43.7 166 40.1 31 12.2 36 8.7 173 34.1 238 28.7 0.72 
Yu 2004 (51199 46.7 306 48.7 186 43.7 260 41.4 41 9.6 62 9.9 268 31.5 384 30.6 0.57 
Yu 2004 (5191 39.6 85 33.9 96 41.7 126 50.2 43 18.7 40 15.9 182 39.6 206 41.0 0.60 
Kunugi 1998 (35121 35.3 95 36.8 168 49.0 129 50.0 54 15.7 34 13.2 276 40.2 197 38.2 0.43 
Arinami 1997 (4096 32.3 154 36.8 138 46.5 214 51.1 63 21.2 51 12.2 264 44.4 316 37.7 0.10 
Kempisty 2006 (52113 56.5 210 70.0 68 34.0 79 26.3 19 9.5 11 3.7 106 26.5 101 16.8 0.30 
Bipolar disorder 
Kunugi 1998 (3541 28.7 95 36.8 74 51.7 129 50.0 28 19.6 34 13.2 130 45.5 197 38.2 0.43 
Arinami 1997 (4015 37.5 154 36.8 20 50.0 214 51.1 12.5 51 12.2 30 37.5 316 31.7 0.10 
Tan 2004 (3699 59.3 80 66.1 60 35.9 34 28.1 4.8 5.8 76 22.7 48 19.8 0.25 
Kempisty 2006 (52108 54.0 210 70.0 73 36.5 79 26.3 19 9.5 11 3.7 111 27.5 101 16.8 0.30 
Anxiety 
Bjelland 2003 (22308 49.6 3,073 49.7 263 42.4 2,601 42.1 50 8.1 511 8.3 363 29.2 3,623 29.3 0.19 
*

p value from test for Hardy-Weinberg equilibrium.

MTHFR C677T meta-analysis.

Ten studies allowed us to make all planned comparisons of alleles and genotypes (following the inclusion of unpublished data provided by the authors of two studies (39, 41)). With our primary analysis, there was an increased risk of depression among persons with the homozygote variant TT genotype, with no discernible statistical between-study heterogeneity (fixed-effects odds ratio (OR) ORTTvCC = 1.36, 95 percent confidence interval (CI): 1.11, 1.67; I2 = 0 percent) (table 3 and figure 1). We found a smaller magnitude of association for heterozygotes that was of borderline significance and subject to greater between-study heterogeneity (fixed-effects ORCTvCC = 1.10, 95 percent CI: 0.96, 1.25; I2 = 49 percent). Depressed subjects showed a significantly increased frequency of the T allele (fixed-effects ORTvC = 1.14, 95 percent CI: 1.04, 1.26; I2 = 0 percent). A cumulative meta-analysis by year of publication showed that the association between depression and MTHFR C677T has remained significant over time, but the magnitude of association has been reduced with the emergence of more recent studies (figure 2).

TABLE 3

Summary odd ratios from meta-analysis of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism and risk of psychiatric disorders

Comparison and racial descent No. of studies Total sample size No. of cases No. of controls ORfixed95% CI* I2 (%) ORrandom 95% CI ORfixed adjusted for Hardy-Weinberg equilibrium 95% CI Small-study bias (31) p value 
Depression 
T versus C 10 11,709 1,280 10,429 1.14 1.04, 1.26 1.14 1.04, 1.26 1.15 1.04, 1.26 0.62 
    Asian 1,047 230 817 1.19 0.93, 1.52 22 1.21 0.90, 1.61 1.18 0.93, 1.51  
    European 10,662 1,050 9,612 1.14 1.03, 1.26 1.14 1.03, 1.26 1.14 1.03, 1.26  
TT versus CC 10 11,709 1,280 10,429 1.36 1.11, 1.67 1.36 1.11, 1.67 1.37 1.12, 1.68 0.79 
    Asian 1,047 230 817 1.43 0.84, 2.44 16 1.48 0.80, 2.75 1.38 0.81, 2.36  
    European 10,662 1,050 9,612 1.35 1.08, 1.68 1.35 1.08, 1.68 1.37 1.10, 1.71  
CT versus CC 10 11,709 1,280 10,429 1.10 0.96, 1.25 49 1.13 0.89, 1.43 1.09 0.95, 1.24 0.73 
    Asian 1,047 230 817 1.08 0.76, 1.53 21 1.08 0.72, 1.62 1.13 0.79, 1.60  
    European 10,662 1,050 9,612 1.10 0.95, 1.27 64 1.16 0.84, 1.60 1.08 0.93, 1.25  
Schizophrenia 
T versus C 12 6,125 2,762 3,363 1.17 1.08, 1.26 46 1.18 1.06, 1.32 1.17 1.08, 1.26 0.47 
    Asian 2,155 1,106 1,049 1.15 1.01, 1.31 35 1.15 0.98, 1.35 1.15 1.01, 1.31  
    European 3,970 1,656 2,314 1.18 1.07, 1.30 57 1.21 1.04, 1.41 1.18 1.07, 1.30  
TT versus CC 12 6,125 2,762 3,363 1.44 1.21, 1.70 42 1.47 1.16, 1.86 1.41 1.19, 1.67 0.98 
    Asian 2,155 1,106 1,049 1.40 1.07, 1.83 26 1.38 1.00, 1.91 1.35 1.04, 1.76  
    European 3,970 1,656 2,314 1.46 1.18, 1.82 53 1.54 1.10, 2.16 1.45 1.17, 1.80  
CT versus CC 12 6,125 2,762 3,363 1.07 0.96, 1.20 41 1.08 0.93, 1.26 1.09 0.98, 1.22 0.36 
    Asian 2,155 1,106 1,049 1.00 0.83, 1.21 42 1.00 0.78, 1.29 1.07 0.88, 1.29  
    European 3,970 1,656 2,314 1.11 0.97, 1.27 44 1.13 0.93, 1.37 1.11 0.96, 1.27  
Bipolar disorder 
T versus C 1,648 550 1,098 1.41 1.19, 1.68 54 1.37 1.05, 1.78 1.41 1.18, 1.68 NA* 
    Asian 1,148 350 798 1.23 0.99, 1.51 1.23 0.99, 1.52 1.23 0.99, 1.52  
    European 500 200 300 1.90 1.40, 2.58  1.90 1.40, 2.56 1.88 1.39, 2.56  
TT versus CC 1,648 550 1,098 1.82 1.22, 2.70 42 1.72 0.99, 3.01 1.90 1.27, 2.84 NA 
    Asian 1,148 350 798 1.44 0.90, 2.31 1.45 0.90, 2.33 1.46 0.90, 2.34  
    European 500 200 300 3.36 1.54, 7.31  3.36 1.54, 7.31 4.07 1.78, 9.29  
CT versus CC 1,648 550 1,098 1.45 1.14, 1.86 1.46 1.14, 1.86 1.44 1.13, 1.83 NA 
    Asian 1,148 350 798 1.28 0.94, 1.74 1.28 0.94, 1.74 1.31 0.96, 1.78  
    European 500 200 300 1.80 1.21, 2.66  1.80 1.21, 2.66 1.67 1.13, 2.47  
Anxiety disorders 
T versus C 6,806 621 6,185 1.00 0.88, 1.13  1.00 0.88, 1.13 1.00 0.88, 1.13 NA 
TT versus CC 6,806 621 6,185 0.98 0.71, 1.34  0.98 0.71, 1.34 0.95 0.69, 1.29 NA 
CT versus CC 6,806 621 6,185 1.01 0.85, 1.20  1.01 0.85, 1.20 1.03 0.87, 1.23 NA 
Comparison and racial descent No. of studies Total sample size No. of cases No. of controls ORfixed95% CI* I2 (%) ORrandom 95% CI ORfixed adjusted for Hardy-Weinberg equilibrium 95% CI Small-study bias (31) p value 
Depression 
T versus C 10 11,709 1,280 10,429 1.14 1.04, 1.26 1.14 1.04, 1.26 1.15 1.04, 1.26 0.62 
    Asian 1,047 230 817 1.19 0.93, 1.52 22 1.21 0.90, 1.61 1.18 0.93, 1.51  
    European 10,662 1,050 9,612 1.14 1.03, 1.26 1.14 1.03, 1.26 1.14 1.03, 1.26  
TT versus CC 10 11,709 1,280 10,429 1.36 1.11, 1.67 1.36 1.11, 1.67 1.37 1.12, 1.68 0.79 
    Asian 1,047 230 817 1.43 0.84, 2.44 16 1.48 0.80, 2.75 1.38 0.81, 2.36  
    European 10,662 1,050 9,612 1.35 1.08, 1.68 1.35 1.08, 1.68 1.37 1.10, 1.71  
CT versus CC 10 11,709 1,280 10,429 1.10 0.96, 1.25 49 1.13 0.89, 1.43 1.09 0.95, 1.24 0.73 
    Asian 1,047 230 817 1.08 0.76, 1.53 21 1.08 0.72, 1.62 1.13 0.79, 1.60  
    European 10,662 1,050 9,612 1.10 0.95, 1.27 64 1.16 0.84, 1.60 1.08 0.93, 1.25  
Schizophrenia 
T versus C 12 6,125 2,762 3,363 1.17 1.08, 1.26 46 1.18 1.06, 1.32 1.17 1.08, 1.26 0.47 
    Asian 2,155 1,106 1,049 1.15 1.01, 1.31 35 1.15 0.98, 1.35 1.15 1.01, 1.31  
    European 3,970 1,656 2,314 1.18 1.07, 1.30 57 1.21 1.04, 1.41 1.18 1.07, 1.30  
TT versus CC 12 6,125 2,762 3,363 1.44 1.21, 1.70 42 1.47 1.16, 1.86 1.41 1.19, 1.67 0.98 
    Asian 2,155 1,106 1,049 1.40 1.07, 1.83 26 1.38 1.00, 1.91 1.35 1.04, 1.76  
    European 3,970 1,656 2,314 1.46 1.18, 1.82 53 1.54 1.10, 2.16 1.45 1.17, 1.80  
CT versus CC 12 6,125 2,762 3,363 1.07 0.96, 1.20 41 1.08 0.93, 1.26 1.09 0.98, 1.22 0.36 
    Asian 2,155 1,106 1,049 1.00 0.83, 1.21 42 1.00 0.78, 1.29 1.07 0.88, 1.29  
    European 3,970 1,656 2,314 1.11 0.97, 1.27 44 1.13 0.93, 1.37 1.11 0.96, 1.27  
Bipolar disorder 
T versus C 1,648 550 1,098 1.41 1.19, 1.68 54 1.37 1.05, 1.78 1.41 1.18, 1.68 NA* 
    Asian 1,148 350 798 1.23 0.99, 1.51 1.23 0.99, 1.52 1.23 0.99, 1.52  
    European 500 200 300 1.90 1.40, 2.58  1.90 1.40, 2.56 1.88 1.39, 2.56  
TT versus CC 1,648 550 1,098 1.82 1.22, 2.70 42 1.72 0.99, 3.01 1.90 1.27, 2.84 NA 
    Asian 1,148 350 798 1.44 0.90, 2.31 1.45 0.90, 2.33 1.46 0.90, 2.34  
    European 500 200 300 3.36 1.54, 7.31  3.36 1.54, 7.31 4.07 1.78, 9.29  
CT versus CC 1,648 550 1,098 1.45 1.14, 1.86 1.46 1.14, 1.86 1.44 1.13, 1.83 NA 
    Asian 1,148 350 798 1.28 0.94, 1.74 1.28 0.94, 1.74 1.31 0.96, 1.78  
    European 500 200 300 1.80 1.21, 2.66  1.80 1.21, 2.66 1.67 1.13, 2.47  
Anxiety disorders 
T versus C 6,806 621 6,185 1.00 0.88, 1.13  1.00 0.88, 1.13 1.00 0.88, 1.13 NA 
TT versus CC 6,806 621 6,185 0.98 0.71, 1.34  0.98 0.71, 1.34 0.95 0.69, 1.29 NA 
CT versus CC 6,806 621 6,185 1.01 0.85, 1.20  1.01 0.85, 1.20 1.03 0.87, 1.23 NA 
*

OR, odds ratio; CI, confidence interval; NA, not applicable.

FIGURE 1

Results from fixed-effects meta-analysis of homozygous (TT vs. CC) genotypes of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism and psychiatric disorders. Reference numbers are given in parentheses. For study details and locations, see tables 1 and 2.

FIGURE 1

Results from fixed-effects meta-analysis of homozygous (TT vs. CC) genotypes of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism and psychiatric disorders. Reference numbers are given in parentheses. For study details and locations, see tables 1 and 2.

FIGURE 2

Results from cumulative fixed-effects meta-analysis of homozygous (TT vs. CC) genotypes of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism in depressed and nondepressed populations, by year of publication. Reference numbers are given in parentheses. For study details and locations, see tables 1 and 2.

FIGURE 2

Results from cumulative fixed-effects meta-analysis of homozygous (TT vs. CC) genotypes of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism in depressed and nondepressed populations, by year of publication. Reference numbers are given in parentheses. For study details and locations, see tables 1 and 2.

Sensitivity analyses.

Stratification by ethnic subtype showed that there was little variation between Asian and European populations and no reduction in overall heterogeneity (for European studies, ORTTvCC = 1.35, 95 percent CI: 1.08, 1.68 (I2 = 0 percent); for Asian studies, ORTTvCC = 1.43, 95 percent CI: 0.84, 2.44 (I2 = 16 percent) (table 3); logistic meta-regression beta coefficient = 0.076, p = 0.81). No studies were in Hardy-Weinberg disequilibrium. The one study which did not employ a standardized diagnostic interview or validated screening test (43) was also the largest and most influential study. The overall homozygote meta-analysis was robust to the inclusion and exclusion of this study (with the study, ORTTvCC = 1.36, 95 percent CI: 1.11, 1.67; without the study, ORTTvCC = 1.37, 95 percent CI: 1.03, 1.82). All meta-analysis results were stable when we adjusted for Hardy-Weinberg disequilibrium (table 3). Funnel plot analysis indicated no evidence of publication bias (table 3 and figure 3).

FIGURE 3

Funnel plot of effect size versus inverse standard error for homozygous (TT vs. CC) genotypes of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism in depressed and control populations.

FIGURE 3

Funnel plot of effect size versus inverse standard error for homozygous (TT vs. CC) genotypes of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism in depressed and control populations.

MTHFR A1298C polymorphism.

One European study (38) reported a positive association with MTHFR A1298C when homozygote variants were compared with the homozygote wild type (CC vs. AA) (ORCCvAA = 3.25, 95 percent CI: 1.19, 8.88). No association with depression was observed for heterozygotes compared with the homozygous wild-type group (ORCCvAC = 1.10, 95 percent CI: 0.84, 1.43). This study was in Hardy-Weinberg disequilibrium (p = 0.02) (table 4).

TABLE 4

Distribution of A1298C genotypes in association studies of methylenetetrahydrofolate reductase (MTHFR) genetic polymorphisms and susceptibility to psychiatric disorders

First author and year (ref. no.) CC CT TT T allele Hardy-Weinberg p value* 
No. of cases No. of controls No. of cases No. of controls No. of cases No. of controls No. of cases No. of controls 
Depression 
Reif 2005 (3816 34.8 75 40.8 21 45.7 96 52.2 19.6 13 7.1 39 54.9 122 44.8 0.02 
Schizophrenia 
Sazci 2003 (4757 43.8 114 50.4 59 45.4 93 41.2 14 10.8 19 8.4 87 43.3 131 36.5 1.00 
Sazci 2005 (48130 43.8 159 46.6 129 43.4 155 45.5 38 12.8 27 7.9 205 44.1 209 39.7 0.25 
First author and year (ref. no.) CC CT TT T allele Hardy-Weinberg p value* 
No. of cases No. of controls No. of cases No. of controls No. of cases No. of controls No. of cases No. of controls 
Depression 
Reif 2005 (3816 34.8 75 40.8 21 45.7 96 52.2 19.6 13 7.1 39 54.9 122 44.8 0.02 
Schizophrenia 
Sazci 2003 (4757 43.8 114 50.4 59 45.4 93 41.2 14 10.8 19 8.4 87 43.3 131 36.5 1.00 
Sazci 2005 (48130 43.8 159 46.6 129 43.4 155 45.5 38 12.8 27 7.9 205 44.1 209 39.7 0.25 
*

p value from test for Hardy-Weinberg equilibrium.

Schizophrenia

Our searches identified 12 studies (6,125 participants; 2,762 cases of schizophrenia, 3,363 controls) meeting our inclusion criteria. All of the studies examined the MTHFR C677T polymorphism (35, 36, 40, 45–52), and two studies also examined the MTHFR A1298C polymorphism (47, 48).

All studies used a case-control design, with standardized clinician-diagnosed cases of schizophrenia and unrelated control subjects. In three studies, investigators took explicit steps to match participants on the basis of age, sex, and race (36, 47, 48). These investigators also took explicit steps to exclude current or previous history of schizophrenia, in two cases verifying this over three generations (47, 48). Participation rates among cases and controls were not reported in any detail. Polymerase chain reaction methods were used in all genotyping (see tables 1 and 2 for study details).

MTHFR C677T meta-analysis.

Twelve studies allowed us to make all planned comparisons of alleles and genotypes. With our primary analysis, there was an increased risk of schizophrenia among homozygote variants (TT), with low statistical between-study heterogeneity (fixed-effects ORTTvCC = 1.44, 95 percent CI: 1.21, 1.70; I2 = 42 percent) (table 3 and figure 1). A nonsignificant association was observed for heterozygotes (fixed-effects ORCTvCC = 1.07, 95 percent CI: 0.96, 1.20; I2 = 41 percent). Subjects with schizophrenia showed a significantly increased frequency of the T allele (fixed-effects ORTvC = 1.17, 95 percent CI: 1.08, 1.26; I2 = 46 percent). A cumulative meta-analysis showed that a moderate and significant association between schizophrenia and MTHFR C677T has remained over time (figure 4).

FIGURE 4

Results from cumulative fixed-effects meta-analysis of homozygous (TT vs. CC) genotypes of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism in persons with schizophrenia and control populations, by year of publication. Reference numbers are given in parentheses. For study details and locations, see tables 1 and 2.

FIGURE 4

Results from cumulative fixed-effects meta-analysis of homozygous (TT vs. CC) genotypes of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism in persons with schizophrenia and control populations, by year of publication. Reference numbers are given in parentheses. For study details and locations, see tables 1 and 2.

Sensitivity analyses.

Stratification by ethnic subtype showed that there was little variation between Asian and European pooled populations (for European populations, ORTTvCC = 1.46, 95 percent CI: 1.18, 1.82 (I2 = 53 percent); for Asian populations, ORTTvCC = 1.40, 95 percent CI: 1.07, 1.83 (I2 = 26 percent) (table 3); logistic meta-regression beta coefficient = –0.02, p = 0.94). One study was in Hardy-Weinberg disequilibrium (45); removal of this study did not alter the overall result (for studies in Hardy-Weinberg equilibrium, ORTTvCC = 1.41, 95 percent CI: 1.19, 1.67 (I2 = 41 percent)). The overall meta-analysis results were stable when we adjusted for Hardy-Weinberg disequilibrium (table 3). There was no evidence of small-study bias or publication bias for any comparisons.

MTHFR A1298C meta-analysis.

From two studies (47, 48), there was an increased risk of schizophrenia among persons with the homozygote variant CC genotype, with no between-study heterogeneity (fixed effects ORCCvAA = 1.64, 95 percent CI: 1.05, 2.54; I2 = 0 percent). This association was not found for heterozygotes when they were compared with the homozygous wild-type group (fixed-effects ORACvAA = 1.10, 95 percent CI: 0.84, 1.43; I2 = 0 percent).

Bipolar disorder

Our searches identified four studies of the MTHFR C677T polymorphism and bipolar disorder (35, 36, 40, 52) (1,648 participants; 550 cases of bipolar disorder, 1,098 controls). All studies used a case-control design, with clinician-diagnosed cases of bipolar disorder and unrelated control subjects. Tan et al. (36) reported taking explicit steps to provide racial matching and to exclude current or previous history of bipolar disorder. All studies employed polymerase chain reaction methods in genotyping (tables 1 and 2).

MTHFR C677T meta-analysis.

There was an increased risk of bipolar disorder associated with the homozygote variant genotype (fixed-effects ORTTvCC = 1.82, 95 percent CI: 1.22, 2.70; I2 = 42 percent) (table 4 and figure 1). The frequency of the T allele was increased among persons with bipolar disorder (fixed-effects ORTvC = 1.41, 95 percent CI: 1.19, 1.68; I2 = 54 percent) (table 3).

Sensitivity analyses.

All studies but one evaluated persons of Asian origin, and all showed a positive trend. No studies were in Hardy-Weinberg disequilibrium. The overall meta-analysis results were stable when adjusted for Hardy-Weinberg disequilibrium. There were insufficient studies to construct funnel plots, and the presence of publication bias could not be excluded (table 3).

Anxiety disorders

One large (n = 6,806) population-based cross-sectional study (22) examined the risk of anxiety, using a validated self-report questionnaire. No association between MTHFR C677T and anxiety disorders was found (ORTTvCC = 0.98, 95 percent CI: 0.71, 1.34) (tables 1 and 3).

DISCUSSION

Main findings

To our knowledge, this is the first general overview of associations between MTHFR polymorphisms and psychiatric disorders. This is a rapidly evolving area of interest, and our meta-analysis represents the most up-to-date quantitative synthesis of available results from gene-psychiatric disorder association studies. In the future, this review will be updated as a mini-HuGE review in line with emerging evidence.

Our main finding was that the MTHFR C677T polymorphism is associated with major depression (OR = 1.36), schizophrenia (OR = 1.44), and bipolar disorder (OR = 1.82). The magnitudes of the associations were moderate but statistically significant. Emerging evidence of an association with the MTHFR A1298C polymorphism was also found. However, in line with other HuGE reviews, associations with this variant have been less extensively researched. Further research on disease associations with this MTHFR polymorphism is needed.

Our findings were based upon several gene-association studies and several thousand participants and were robust to each of the planned sensitivity analyses used. There was no evidence of publication bias, and we found low between-study heterogeneity. When we examined Hardy-Weinberg equilibrium, the majority of studies showed no deviation, and the overall results were robust to the exclusion of disequilibrium studies and statistical adjustment for control-group disequilibrium, using a recently introduced test (32). Of particular interest was the finding that our results were consistent across different racial groups, irrespective of allele frequency—a finding that is in line with those for other gene-disease associations (53).

Limitations of the epidemiologic evidence

When we examined the design and methodological aspects of these studies, we found that the majority employed a case-control design. Few investigators took explicit steps to provide racial matching or to exclude participants with a previous, current, or family history of the disorder. Participation rates for cases and controls were rarely reported. However, nonparticipation and selection bias are much less of a problem in gene-association studies than in classical epidemiologic designs, since nonparticipation is rarely related to genotype (54). Among large-scale population-level studies, cases were necessarily less rigorously defined. When we took steps to examine the robustness of our overall results to the inclusion and exclusion of these studies, we found that all significant gene-disease associations remained.

Possible mechanisms of gene-disorder association

Folate and MTHFR polymorphisms are related to neural tube defects (55) and have now been implicated in the pathogenesis of several diseases and disorders, including leukemia (6), colorectal cancer (7), cardiovascular disease (56), and other congenital abnormalities (5).

The 1-carbon cycle/folate metabolic pathway is complex and regulates not only nucleotide synthesis but also DNA methylation. 5-Methyltetrahydrofolate is the predominant circulating form of folate, and it donates a methyl group to homocysteine in the generation of S-adenosylmethionine, a major source of methyl groups in the brain (1). MTHFR is a critical component of the 1-carbon cycle, and the MTHFR polymorphisms C677T and A1298C affect both nucleotide synthesis and DNA methylation (57). This forms a plausible biologic explanation for potential associations between genetic variation in folate metabolism and both depression and schizophrenia (16). The MTHFR C677T polymorphism is associated with a reduction in the bioavailability of folate and folate metabolites and “mimics” low dietary folate intake (3).

Depression in association with reduced folate has been demonstrated within observational epidemiologic studies (22), but a causal link has been difficult to establish because of confounding and reverse causality (58). Estimates of association between low folate and depression range from an odds ratio of 12 in case-control studies measuring serum folate levels (59) to an odds ratio of 1.6 in a large cross-sectional study estimating dietary folate intake from food diaries (60). In contrast to traditional epidemiologic approaches, the choice of bioassay method, reverse causality, and confounding of genotype-depression associations are unlikely to influence the association between MTHFR polymorphisms and functional psychiatric disorders (54).

A possible gene-environment association between impaired folate metabolism and dietary folate might also exist, although we did not explicitly address this question. Three studies within this review (22, 37, 41) measured both MTHFR polymorphisms and folate levels. In case-control studies (37, 41), the results were inconclusive. In one cross-sectional study (22), the authors reported on the association of folate levels with MTHFR polymorphisms but did not explicitly examine gene-environment interaction. Similarly, there are several possible genes which also work in the folate/1-carbon pathway and which might interact with MTHFR polymorphisms (57). We did not set out to examine these interactions, nor did we find any studies which did so. Gene-environment and gene-gene interactions require examination in more detail; future updates of this HuGE review will explicitly address these issues.

Schizophrenia is increasingly considered to be a neurodevelopmental disorder (61), with in-utero exposures and epigenetic mechanisms such as DNA methylation being important in its etiology (62, 63). DNA methylation is a critical epigenetic modification of the genome that controls many biologic processes, including embryonic development, X-chromosome inactivation, imprinting, and gene expression. Incorrect methylation patterns can affect embryogenesis, leading to developmental malformations and embryonic death. Although these patterns are established during early life, they are not fixed, and gradual hypomethylation of the genome occurs in most tissues with age, together with aberrant hypermethylation of gene promoter regions. Thus, the correct development of DNA methylation patterns is important not only for early life but also for long-term health benefits, including neurologic disease susceptibility.

Methylation is genetically predetermined, either by imprinting or by inheritance of genes which influence methylation, such as MTHFR and other genes involved in the 1-carbon cycle (64). Methyl groups required for methylation are synthesized de novo or are supplied in the diet, primarily from folate. Thus, methylation may be modified by gene-exposure interactions occurring during development. This link between folate, folate metabolism, and DNA methylation therefore provides a plausible biologic mechanism for the observed association between MTHFR and schizophrenia.

Recently, Coppen and Bolander-Gouaille (16) recommended the use of folate in the treatment and prevention of depression as a population-level strategy. The results of the present review also raise the possibility of folate supplementation's playing a protective role in the development of schizophrenia, when taken during pregnancy. Positive results of population-level folate fortification strategies in preventing neural tube defects have been found (65), and studies examining the potential of folate in reducing schizophrenia risk are required. Mandatory food fortification is used in several countries (including the United States, Canada, and Australia) (66, 67). The cost-effectiveness and safety of this approach might be additionally supported by the demonstration of a further gene-environment association for folate, folate metabolism, and depression.

Abbreviations

    Abbreviations
  • CI

    confidence interval

  • HuGE

    Human Genome Epidemiology

  • MTHFR

    methylenetetrahydrofolate reductase

  • OR

    odds ratio

Conflict of interest: none declared.

References

1.
Cantoni
GL
S-Anenysylmethionine: a new intermediate formed enzymatically from 1-methionine and adenosine triphosphate
J Biol Chem
 , 
1953
, vol. 
204
 (pg. 
403
-
16
)
2.
Rozen
R
Molecular genetics of methylenetetrahydrofolate reductase deficiency
J Inherit Metab Dis
 , 
1996
, vol. 
19
 (pg. 
589
-
94
)
3.
Molloy
AM
Daly
S
Mills
JL
, et al.  . 
Thermolabile variant of 5,10-methylenetetrahydrofolate reductase associated with low red-cell folates: implications for folate intake recommendations
Lancet
 , 
1997
, vol. 
349
 (pg. 
1591
-
3
)
4.
Lievers
KJ
Boers
GH
Verhoef
P
, et al.  . 
A second common variant in the methylenetetrahydrofolate reductase (MTHFR) gene and its relationship to MTHFR enzyme activity, homocysteine, and cardiovascular disease risk
J Mol Med
 , 
2001
, vol. 
79
 (pg. 
522
-
8
)
5.
Botto
LD
Yang
Q
5,10-Methylenetetrahydrofolate reductase gene variants and congenital anomalies: a HuGE review
Am J Epidemiol
 , 
2000
, vol. 
151
 (pg. 
862
-
77
)
6.
Robien
K
Ulrich
CM
5,10-Methylenetetrahydrofolate reductase polymorphisms and leukemia risk: a HuGE minireview
Am J Epidemiol
 , 
2003
, vol. 
157
 (pg. 
571
-
82
)
7.
Sharp
L
Little
J
Polymorphisms in genes involved in folate metabolism and colorectal neoplasia: a HuGE review
Am J Epidemiol
 , 
2004
, vol. 
159
 (pg. 
423
-
43
)
8.
World Health Organization
ICD-10: international statistical classification of diseases and related health problems. Tenth Revision
 , 
1992
Geneva, Switzerland
World Health Organization
9.
Murray
CJ
Lopez
AD
The global burden of disease: a comprehensive assessment of mortality and disability from disease, injuries and risk factors in 1990
 , 
1996
Boston, MA
Harvard School of Public Health
10.
Leonardo
ED
Hen
R
Genetics of affective and anxiety disorders
Annu Rev Psychol
 , 
2006
, vol. 
57
 (pg. 
117
-
37
)
11.
Prathikanti
S
Weinberger
DR
Psychiatric genetics—the new era: genetic research and some clinical implications
Br Med Bull
 , 
2005
, vol. 
73–74
 (pg. 
107
-
22
)
12.
Kessler
RC
Berglund
P
Demler
O
, et al.  . 
The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R)
JAMA
 , 
2003
, vol. 
289
 (pg. 
3095
-
105
)
13.
Meltzer
HY
Gill
B
Petticrew
M
, et al.  . 
OPCS surveys of psychiatric morbidity in Great Britain
 , 
1995
London, United Kingdom
HMSO
14.
American Psychiatric Association
Diagnostic and statistical manual of mental disorders. Fourth Edition
 , 
1994
Washington, DC
American Psychiatric Association
15.
Regier
DA
Narrow
WE
Rae
DS
, et al.  . 
The de facto US mental and addictive disorders service system. Epidemiologic catchment area prospective 1-year prevalence rates of disorders and services
Arch Gen Psychiatry
 , 
1993
, vol. 
50
 (pg. 
85
-
94
)
16.
Coppen
A
Bolander-Gouaille
C
Treatment of depression: time to consider folic acid and vitamin B12
J Psychopharmacol
 , 
2005
, vol. 
19
 (pg. 
59
-
65
)
17.
Muntjewerff
JW
Blom
HJ
Aberrant folate status in schizophrenic patients: what is the evidence?
Prog Neuropsychopharmacol Biol Psychiatry
 , 
2005
, vol. 
29
 (pg. 
1133
-
9
)
18.
Hasanah
CI
Khan
UA
Musalmah
M
, et al.  . 
Reduced red-cell folate in mania
J Affect Disord
 , 
1997
, vol. 
46
 (pg. 
95
-
9
)
19.
Carney
MWP
Serum folate values in 423 psychiatric patients
BMJ
 , 
1967
, vol. 
4
 (pg. 
512
-
16
)
20.
Carney
MW
Chary
TK
Laundy
M
, et al.  . 
Red cell folate concentrations in psychiatric patients
J Affect Disord
 , 
1990
, vol. 
19
 (pg. 
207
-
13
)
21.
Lee
S
Wing
YK
Fong
S
A controlled study of folate levels in Chinese inpatients with major depression in Hong Kong
J Affect Disord
 , 
1998
, vol. 
49
 (pg. 
73
-
7
)
22.
Bjelland
I
Tell
GS
Vollset
SE
, et al.  . 
Folate, vitamin B12, homocysteine, and the MTHFR 677C→T polymorphism in anxiety and depression: The Hordaland Homocysteine Study
Arch Gen Psychiatry
 , 
2003
, vol. 
60
 (pg. 
618
-
26
)
23.
Little
J
Khoury
MJ
Bradley
L
, et al.  . 
The Human Genome Project is complete. How do we develop a handle for the pump? (Editorial)
Am J Epidemiol
 , 
2003
, vol. 
157
 (pg. 
667
-
73
)
24.
Ioannidis
JPA
Bernstein
J
Boffetta
P
, et al.  . 
A network of investigator networks in human genome epidemiology
Am J Epidemiol
 , 
2005
, vol. 
162
 (pg. 
302
-
4
)
25.
Little
J
Higgins
JPT
The HuGENet™ HuGE review handbook, version 1.0
 , 
2006
Ottawa, Ontario, Canada
HuGENet Canada Coordinating Centre
 
(http://www.hugenet.ca/). (Accessed May 19, 2006)
26.
Emigh
T
A comparison of tests for Hardy-Weinberg equilibrium
Biometrics
 , 
1980
, vol. 
36
 (pg. 
627
-
42
)
27.
Sutton
AJ
Abrams
KR
Jones
DR
, et al.  . 
Systematic reviews of trials and other studies
Health Technol Assess
 , 
1999
, vol. 
2
 (pg. 
1
-
276
)
28.
Lau
J
Antman
EM
Jimenez-Silva
J
, et al.  . 
Cumulative meta-analysis of therapeutic trials for myocardial infarction
N Engl J Med
 , 
1992
, vol. 
327
 (pg. 
248
-
54
)
29.
Higgins
JP
Thompson
SG
Deeks
JJ
, et al.  . 
Measuring inconsistency in meta-analyses
BMJ
 , 
2003
, vol. 
327
 (pg. 
557
-
60
)
30.
Salanti
G
Sanderson
S
Higgins
JP
Obstacles and opportunities in meta-analysis of genetic association studies
Genet Med
 , 
2005
, vol. 
7
 (pg. 
13
-
20
)
31.
Egger
M
Davey-Smith
G
Schneider
M
, et al.  . 
Bias in meta-analysis detected by a simple, graphical test
BMJ
 , 
1997
, vol. 
315
 (pg. 
629
-
34
)
32.
Trikalinos
TA
Salanti
G
Khoury
MJ
, et al.  . 
Impact of violations and deviations in Hardy-Weinberg equilibrium on postulated gene-disease associations
Am J Epidemiol
 , 
2006
, vol. 
163
 (pg. 
300
-
9
)
33.
Thompson
SG
Sharp
SJ
Explaining heterogeneity in meta-analysis: a comparison of methods
Stat Med
 , 
1999
, vol. 
18
 (pg. 
2693
-
708
)
34.
Higgins
JPT
Thompson
SG
Controlling the risk of spurious findings from meta-regression
Stat Med
 , 
2004
, vol. 
23
 (pg. 
1663
-
82
)
35.
Kunugi
H
Fukuda
R
Hattori
M
, et al.  . 
C677T polymorphism in methylenetetrahydrofolate reductase gene and psychoses
Mol Psychiatry
 , 
1998
, vol. 
3
 (pg. 
435
-
7
)
36.
Tan
EC
Chong
SA
Lim
LC
, et al.  . 
Genetic analysis of the thermolabile methylenetetrahydrofolate reductase variant in schizophrenia and mood disorders
Psychiatr Genet
 , 
2004
, vol. 
14
 (pg. 
227
-
31
)
37.
Chen
CS
Tsai
JC
Tsang
HY
, et al.  . 
Homocysteine levels, MTHFR C677T genotype, and MRI hyperintensities in late-onset major depressive disorder
Am J Geriatr Psychiatry
 , 
2005
, vol. 
13
 (pg. 
869
-
75
)
38.
Reif
A
Pfuhlmann
B
Lesch
KP
Homocysteinemia as well as methylenetetrahydrofolate reductase polymorphism are associated with affective psychoses
Prog Neuropsychopharmacol Biol Psychiatry
 , 
2005
, vol. 
29
 (pg. 
1162
-
8
)
39.
Hickie
I
Scott
E
Naismith
S
, et al.  . 
Late-onset depression: genetic, vascular and clinical contributions
Psychol Med
 , 
2001
, vol. 
31
 (pg. 
1403
-
12
)
40.
Arinami
T
Yamada
N
Yamakawa-Kobayashi
K
, et al.  . 
Methylenetetrahydrofolate reductase variant and schizophrenia/depression
Am J Med Genet
 , 
1997
, vol. 
74
 (pg. 
526
-
8
)
41.
Almeida
OP
Flicker
L
Lautenschlager
NT
, et al.  . 
Contribution of the MTHFR gene to the causal pathway for depression, anxiety and cognitive impairment in later life
Neurobiol Aging
 , 
2005
, vol. 
26
 (pg. 
251
-
7
)
42.
Kelly
CB
McDonnell
AP
Johnston
TG
, et al.  . 
The MTHFR C677T polymorphism is associated with depressive episodes in patients from Northern Ireland
J Psychopharmacol
 , 
2004
, vol. 
18
 (pg. 
567
-
71
)
43.
Lewis
SJ
Lawlor
DA
Davey Smith
G
, et al.  . 
The thermolabile variant of MTHFR is associated with depression in the British Women's Heart and Health Study and a meta-analysis
Mol Psychiatry
 , 
2006
, vol. 
11
 (pg. 
352
-
60
)
44.
The EuroQol Group. EuroQol
—a new facility for the measurement of health-related quality of life
Health Policy
 , 
1990
, vol. 
16
 (pg. 
199
-
208
)
45.
Muntjewerff
JW
van der Put
N
Eskes
T
, et al.  . 
Homocysteine metabolism and B-vitamins in schizophrenic patients: low plasma folate as a possible independent risk factor for schizophrenia
Psychiatry Res
 , 
2003
, vol. 
121
 (pg. 
1
-
9
)
46.
Joober
R
Benkelfat
C
Lal
S
, et al.  . 
Association between the methylenetetrahydrofolate reductase 677C→T missense mutation and schizophrenia
Mol Psychiatry
 , 
2000
, vol. 
5
 (pg. 
323
-
6
)
47.
Sazci
A
Ergul
E
Guzelhan
Y
, et al.  . 
Methylenetetrahydrofolate reductase gene polymorphisms in patients with schizophrenia
Brain Res Mol Brain Res
 , 
2003
, vol. 
117
 (pg. 
104
-
7
)
48.
Sazci
A
Ergul
E
Kucukali
I
, et al.  . 
Association of the C677T and A1298C polymorphisms of methylenetetrahydrofolate reductase gene with schizophrenia: association is significant in men but not in women
Prog Neuropsychopharmacol Biol Psychiatry
 , 
2005
, vol. 
29
 (pg. 
1113
-
23
)
49.
Virgos
C
Martorell
L
Simo
JM
, et al.  . 
Plasma homocysteine and the methylenetetrahydrofolate reductase C677T gene variant: lack of association with schizophrenia
Neuroreport
 , 
1999
, vol. 
10
 (pg. 
2035
-
8
)
50.
Muntjewerff
JW
Hoogendoorn
ML
Kahn
RS
, et al.  . 
Hyperhomocysteinemia, methylenetetrahydrofolate reductase 677TT genotype, and the risk for schizophrenia: a Dutch population based case-control study
Am J Med Genet B Neuropsychiatr Genet
 , 
2005
, vol. 
135
 (pg. 
69
-
72
)
51.
Yu
L
Li
T
Robertson
Z
, et al.  . 
No association between polymorphisms of methylenetetrahydrofolate reductase gene and schizophrenia in both Chinese and Scottish populations
Mol Psychiatry
 , 
2004
, vol. 
9
 (pg. 
1063
-
5
)
52.
Kempisty
B
Mostowska
A
Gorska
I
, et al.  . 
Association of 677C→T polymorphism of methylenetetrahydrofolate reductase (MTHFR) gene with bipolar disorder and schizophrenia
Neurosci Lett
 , 
2006
, vol. 
400
 (pg. 
267
-
71
)
53.
Ioannidis
JP
Ntzani
EE
Trikalinos
TA
‘Racial’ differences in genetic effects for complex diseases
Nat Genet
 , 
2004
, vol. 
36
 (pg. 
1312
-
13
)
54.
Davey Smith
G
Ebrahim
S
‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease?
Int J Epidemiol
 , 
2003
, vol. 
32
 (pg. 
1
-
22
)
55.
Kirke
PN
Mills
JL
Molloy
AM
, et al.  . 
Impact of the MTHFR C677T polymorphism on risk of neural tube defects: case-control study
BMJ
 , 
2004
, vol. 
328
 (pg. 
1535
-
6
)
56.
Lewis
SJ
Ebrahim
S
Davey Smith
G
Meta-analysis of MTHFR 677C→T polymorphism and coronary heart disease: does totality of evidence support causal role for homocysteine and preventive potential of folate?
BMJ
 , 
2005
, vol. 
331
 pg. 
1053
 
57.
Lucock
M
Folic acid: nutritional biochemistry, molecular biology, and role in disease processes
Mol Genet Metab
 , 
2000
, vol. 
71
 (pg. 
121
-
38
)
58.
Alpert
JE
Mischoulon
D
Nierenberg
AA
, et al.  . 
Nutrition and depression: focus on folate
Nutrition
 , 
2000
, vol. 
16
 (pg. 
544
-
6
)
59.
Abou-Saleh
MT
Coppen
A
Serum and red blood cell folate in depression
Acta Psychiatr Scand
 , 
1989
, vol. 
80
 (pg. 
78
-
82
)
60.
Tolmunen
T
Voutilainen
S
Hintikka
J
, et al.  . 
Dietary folate and depressive symptoms are associated in middle-aged Finnish men
J Nutr
 , 
2003
, vol. 
133
 (pg. 
3233
-
6
)
61.
Picker
JD
Coyle
JT
Do maternal folate and homocysteine levels play a role in neurodevelopmental processes that increase risk for schizophrenia?
Harv Rev Psychiatry
 , 
2005
, vol. 
13
 (pg. 
197
-
205
)
62.
Singh
SM
Murphy
B
O'Reilly
RL
Involvement of gene-diet/drug interaction in DNA methylation and its contribution to complex diseases: from cancer to schizophrenia
Clin Genet
 , 
2003
, vol. 
64
 (pg. 
451
-
60
)
63.
Singh
SM
McDonald
P
Murphy
B
, et al.  . 
Incidental neurodevelopmental episodes in the etiology of schizophrenia: an expanded model involving epigenetics and development
Clin Genet
 , 
2004
, vol. 
65
 (pg. 
435
-
40
)
64.
Bjornsson
HT
Fallin
MD
Feinberg
AP
An integrated epigenetic and genetic approach to common human disease
Trends Genet
 , 
2004
, vol. 
20
 (pg. 
350
-
8
)
65.
MRC Vitamin Study Research Group
Prevention of neural tube defects: results of the Medical Research Council Vitamin Study
Lancet
 , 
1991
, vol. 
338
 (pg. 
131
-
7
)
66.
Jacques
PF
Selhub
J
Bostom
AG
, et al.  . 
The effect of folic acid fortification on plasma folate and total homocysteine concentrations
N Engl J Med
 , 
1999
, vol. 
340
 (pg. 
1449
-
54
)
67.
Hickling
S
Hung
J
Knuiman
M
, et al.  . 
Impact of voluntary folate fortification on plasma homocysteine and serum folate in Australia from 1995 to 2001: a population based cohort study
J Epidemiol Community Health
 , 
2005
, vol. 
59
 (pg. 
371
-
6
)
68.
American Psychiatric Association
Diagnostic and statistical manual of mental disorders. Third Edition
 , 
1980
Washington, DC
American Psychiatric Association

Author notes

Editor's note:This article also appears on the website of the Human Genome Epidemiology Network (http://www.cdc.gov/genomics/hugenet/).