Predictors of Vitamin B₆ and Folate Concentrations in Older Persons: The InCHIANTI Study

Abstract

Background

Low dietary intake and low serum concentrations of vitamin B₆ and/or folate are associated with increased risk of vascular events, possibly because of their association with inflammation, which plays a crucial role in the pathogenesis of cardiovascular diseases.

Methods

Using data from 1320 participants in the population-based InCHIANTI study (586 men and 734 women; median age, 69 years; range, 21–102 years) for whom complete data on folate, vitamin B₆, inflammatory markers, 5,10-methylenetetrahydrofolate reductase (MTHFR) C677T sequence variant, and important covariates were available, we evaluated the association of inflammatory markers with circulating concentrations of vitamin B₆ and folate, independently of dietary vitamin intake, circulating vitamin concentrations, and MTHFR C677T sequence variant.

Results

According to multiple linear regression analysis, C-reactive protein and interleukin-6 receptor were strongly and negatively associated with circulating vitamin B₆ but not with folate concentrations, independent of age, sex, serum creatinine, serum albumin, total energy intake, smoking history, dietary nutrient intake, and circulating homocysteine and vitamin concentrations. Serum folate concentrations were related to MTHFR 677 TT genotype in persons with folate intake in the lowest tertile (<221.2 μg/day). Vitamin C and retinol intakes were strongly and positively associated with serum folate concentrations independent of age, sex, serum creatinine, serum albumin, total energy intake, smoking history, homocysteine plasma concentrations, dietary nutrient intakes, serum vitamin B₆ and vitamin B₁₂ concentrations, and MTHFR C677T sequence variant.

Conclusions

Low serum vitamin B₆, but not serum folate, concentrations are independent correlates of the proinflammatory state, and both are influenced by antioxidant reserves.

Prospective and cross-sectional studies have consistently demonstrated that low dietary intake and low serum concentrations of folate and/or vitamin B₆ (1–7) are independent risk factors for vascular events. However, the mechanism of this association is unknown.

Evidence has accumulated on the crucial role of inflammation in the pathogenesis of atherosclerosis (8, 9). High C-reactive protein (CRP)⁷ concentrations are widely recognized as a strong independent risk factor for future cardiovascular events (10–12). Longitudinal studies have demonstrated that high interleukin-6 (IL-6) serum concentrations are associated with clinical progression of unstable angina and increased risk of myocardial infarction (13–15) and other cardiovascular events. IL-6 has greater predictive value than CRP as a cardiovascular risk factor (16).

Using data from the Framingham Heart Study, Friso et al. (17) demonstrated a strong relationship between impaired vitamin B₆ status and a proinflammatory state, as indicated by CRP concentrations independent of homocysteine (Hcy) concentrations. These findings suggested that vitamin B₆ may be implicated in some antiinflammatory mechanism, perhaps acting as a cofactor. In animal models, activated macrophages overexpress high-affinity folate receptors (18), suggesting that during an inflammatory state vitamin B₆ may be mobilized from the liver and peripheral tissues to inflammation sites (19) and raising the interesting hypothesis that a prolonged proinflammatory state may lead to vitamin depletion, which perhaps contributes to a sustained chronic inflammatory response. Thus, relatively low concentrations of folate and vitamin B₆ in older persons may be caused by the chronic mild proinflammatory state that characteristically affects this age group.

Using data from a large representative sample enrolled in the InCHIANTI study, we tested the hypothesis that the inflammatory state is associated with low vitamin B₆ and folate concentrations independent of dietary intake and other potential confounders, including 5,10-methyl-enetetrahydrofolate reductase (MTHFR)⁸ C677T sequence variant.

Materials and Methods

Data are from the population-based InCHIANTI study, an epidemiologic study to evaluate factors affecting mobility in older persons living in the Chianti geographic area (Tuscany, Italy). The details of the study have been described elsewhere (20). Briefly, in August 1998, 1616 persons 21–102 years of age were selected from the population registry of Greve in Chianti (a rural area; 11 709 inhabitants, with 19.3% of the population 65 years or older) and Bagno a Ripoli (Antella village, near Florence; 4704 inhabitants, with 20.3% of the population 65 years or older). The participation rate was 90% (1453 of 1616).

All participants gave informed consent, and the study was approved by the Ethics Committee of the Italian National Institute of Research and Care of Aging. Blood samples were collected in the morning after the participants had been fasting for at least 8 h and sitting for at least 15 min. The analysis reported here is based on 1320 participants for whom complete data on folate, vitamin B₆, inflammatory markers, MTHFR C677T sequence variant, and important covariates were available.

Assays

We used commercial enzymatic tests (Roche Diagnostics, GmbH) to measure serum total cholesterol, HDL-cholesterol, and triglyceride concentrations. Serum LDL-cholesterol concentrations were calculated by the Friedewald formula. The interassay CVs were <3.8% for total cholesterol, <5% for HDL-cholesterol, and <2.5% for triglycerides. Lipid values were expressed according to the risk categories suggested by the National Heart, Lung, and Blood Institute Expert Panel on Detection, Evaluation. and Treatment of High Blood Cholesterol in Adults (21).

Folate and vitamins B₆ and B₁₂ were measured in serum samples obtained from blood collected in evacuated tubes without anticoagulant, centrifuged at 2000g for 10 min, and stored at −80 °C.

We measured vitamin B₆ (pyridoxal-5-phosphate) with a commercial HPLC assay with fluorescence detection (Immun Diagnostik) and serum folate and vitamin B₁₂ with a radioligand-binding assay (SimulTrac-SNB Radio-assay; ICN Pharmaceuticals). The minimum detectable concentrations were 0.2 μg/L for vitamin B₆, 0.6 μg/L for folate, and 75 ng/L for vitamin B₁₂. The intraassay CVs were 2.8% for vitamin B₆, 4.1% for folate, and 11% for vitamin B₁₂, and the interassay CVs were 4.1% for vitamin B₆, 7.1% for folate, and 12% for vitamin B₁₂.

We measured circulating concentrations of inflammatory markers with commercial assays as described previously (22). Plasma Hcy concentrations were measured by a fluorometric polarized immunoassay method (IMX; Abbott Laboratories).

MTHFR C677T Sequence Variant Analysis

We extracted genomic DNA from peripheral blood samples with a commercial reagent set (FlexiGene DNA reagent set; Qiagen GmbH). Screening for C677T MTHFR sequence variants was carried out on genomic DNA by microelectronic array technology (Nanochip Molecular Biology Workstation; Nanogen) (23). Briefly, genomic DNA was amplified with forward and reverse PCR primers (forward, 5′-biotin-TGA AGG AGA AGG TGT CTG CGG GA-3′; reverse, 5′-AGG ACG GTG CGG TGA GAG TG-3′). In addition, we designed 2 reporter oligonucleotides—one labeled with Cy3, specific for the wild-type nucleotide, and the other, labeled with Cy5, specific for the mutant nucleotide T (wild-type reporter, 5′-Cy3-TGA TGA AAT CGG-3′; mutant reporter 5′-Cy5-ATG ATG AAA TCGA-3′)—and 1 stabilizer oligonucleotide (5′-CTC CCG CAG ACA CCT TCT CCT TCA-3′). PCR products were desalted with the NucleoFast (96-well PCR plates; Macherey-Nagel) and electronically addressed to specific pads on the chip (NanoChip cartridge). After the stabilizer and reporter oligonucleotides were hybridized to the chip, the cartridge was placed in the reader, and the temperature was increased to 36 °C for discrimination. A fluorescence scan was performed, and the software directly assigned a genotype to each sample.

Covariates

Average daily intakes of energy (kcal), alcohol (<30 vs >30 g/day), and vitamins were estimated by administration of the European Prospective Investigation into Cancer and Nutrition (EPIC) food frequency questionnaire. The EPIC food frequency questionnaire provides a detailed assessment of food consumption during the previous year through a large number of structured and precoded questions. Originally, the questionnaire was designed to be self-administered. However, in a pilot study we discovered that this method of administration provides ambiguous results for older adults because they are likely to misunderstand the questions. Thus, in the InCHIANTI study, the EPIC questionnaire was administered by the interviewers. The information provided by the questionnaire was transformed to average daily intake of macro- and micronutrients, including vitamins, by custom software that uses the table of food composition for epidemiologic study in Italy, edited by the European Institute of Oncology in 1998 (24, 25), for reference.

Smoking status was assessed by self-reported data. Pack-years, a measure of smoking exposure that combines intensity and duration, was calculated as (packs smoked per day) × (years of smoking).

Statistical Analysis

We performed all analyses with the SPSS statistical package, Ver. 11.5. P values <0.05 were considered statistically significant.

Log-transformed values for serum vitamin B₆, folate, serum inflammatory markers, nutrient intake, and the square root of alcohol intake were used in the analyses and back-transformed for data presentation.

We calculated the MTHFR C677T genotype distribution across serum folate tertiles (≤4.99, 5.0–8.16, and >8.16 nmol/L). Differences among serum folate tertiles were tested by general linear models adjusted for sex, age, serum creatinine, serum albumin, total energy intake, and smoking history.

Adjusted mean folate values according to MTHFR C677T genotype were estimated from a general linear model adjusted for age, sex, serum creatinine, serum albumin, smoking history, and total energy intake. We used multiple linear regression models to test the independent associations of inflammatory markers, nutrient intake, and other vitamins with vitamin B₆ and folate. The model predicting folate also included the MTHFR genotypes as potential predictors. Variables found not to be associated with the outcome were removed from the final, most rigorous, regression model through a backward selection algorithm.

We studied the role of Hcy in the relationship between inflammation and vitamin B₆ and serum folate concentrations by adding Hcy as a covariate in the final, most rigorous model predicting serum vitamin B₆ and folate. The b coefficient estimated in the linear regression analysis indicates the expected mean change in logarithmically transformed vitamin B₆ or folate associated with 1-unit changes in the independent variables. Therefore, for independent variables that are logarithmically transformed, a 1% change corresponds to a b% change in the dependent variable.

Results

Characteristics of the Study Group

The characteristics of the InCHIANTI participants are shown in Table 1. The study population in the analysis presented here included 734 females (55.6%) and 586 males (44.4%).

Table 1.

Characteristics of the study population (n = 1320).

Mean (SD) age, years	68.9 (15.6)
Age distribution, n (%)
<65 years	281 (21.3)
65–74 years	567 (43.0)
75–84 years	333 (25.2)
≥85 years	139 (10.5)
Sex, M/F	586/734
Mean (SD) serum creatinine, mg/L	9.1 (1.9)
Smoking habits, n (%)
Current smoker	201 (15.2)
Former smoker	363 (27.5)
Never smoked	756 (57.3)
Mediana daily nutrient intake
Energy intake, kcal/day	1903 (1104–3121)
Alcohol, g/day	7.95 (0.0–54.5)
Vitamin B6, mg/day	1.65 (0.97–2.63)
Folate, μg/day	254.4 (142.7–408.2)
Thiamine, mg/day	0.90 (0.53–1.37)
Riboflavin, mg/day	1.26 (0.76–2.04)
Niacin, mg/day	16.46 (9.26–26.10)
Retinol, μg/day	696.6 (300.0–2007.4)
β-Carotene, mg/day	1.97 (0.83–4.77)
Vitamin C, mg/day	107.1 (44.9–206.7)
Vitamin E, mg/day	6.17 (3.34–10.29)
Mediana Hcy, μmol/L	13.8 (8.7–27.3)
Mediana circulating serum vitamin concentrations
Vitamin B6, nmol/L	25.22 (7.22–65.39)
Folate, nmol/L	6.58 (2.72–15.44)
Vitamin B12, pmol/L	284.9 (121.8–794.3)
α-Tocopherol, μmol/L	29.01 (18.03–43.70)
γ-Tocopherol, μmol/L	1.29 (0.63–2.63)
Mediana concentrations of inflammatory markers
CRP, mg/L	2.38 (0.42–16.70)
IL-6, ng/L	1.27 (0.36–6.43)
IL-6r, μg/L	91.9 (33.8–212.0)
IL-1 receptor agonist, ng/L	130.4 (56.1–309.5)
Leukocyte count, 106 cells/μL	5.99 (4.10–9.03)
MTHFR C677T sequence variant, n (%)
CC	308 (23.33)
CT	711 (53.86)
TT	301 (22.81)

^aValues in parentheses are the 5th–95th percentile.

Correlates of Serum Vitamin B₆ Concentrations

In the multiple linear regression analysis, CRP and IL-6 receptor (IL-6r) were strongly and negatively associated with circulating vitamin B₆ concentrations independent of age, sex, serum creatinine, serum albumin, total energy intake, smoking habits, dietary nutrient intakes, and circulating serum vitamin concentrations (Table 2, model 1). After the inclusion of Hcy in the model (model 2), CRP, and soluble IL-6r (sIL-6r) remained significantly associated with vitamin B₆ concentrations as well as age, alcohol and β-carotene intakes, and serum folate, vitamin B₁₂. and albumin concentrations (Table 2, model 2).

Table 2.

Multiple linear regression models relating vitamin intake and markers of inflammation to circulating vitamin B₆ concentrations.^a

	Log (vitamin B6 in nmol/L)
	Model 1b		Model 2c
Independent variables	b (SE)	P	b (SE)	P
Age (years × 10)	−0.042 (0.006)	<0.0001	−0.036 (0.006)	<0.0001
Square root of alcohol intake	0.012 (0.004)	0.001	0.013 (0.004)	0.001
Log(β-carotene intake in μg/day)	0.194 (0.048)	<0.0001	0.191 (0.048)	<0.0001
Log(vitamin B6 intake in mg/day)	0.256 (0.126)	0.042	0.220 (0.126)	0.079
Log(α-tocopherol in μmol/L)	0.296 (0.071)	<0.0001	0.258 (0.071)	<0.0001
Log(folate in nmol/L)	0.132 (0.036)	<0.0001	0.108 (0.037)	0.004
Log(vitamin B12 in pmol/L)	0.109 (0.035)	0.002	0.083 (0.036)	0.022
Albumin (mg/dL)	0.180 (0.029)	<0.0001	0.179 (0.029)	<0.0001
Log(CRP in mg/L)	−0.116 (0.020)	<0.0001	−0.123 (0.018)	<0.0001
Log(IL-6r in ng/L)	−0.119 (0.036)	0.001	−0.117 (0.036)	0.001

^aIn the multiple linear regression analysis, circulating vitamin B₆ (serum concentrations) was used as the dependent variable and age, sex, serum creatinine, serum albumin, smoking habits, total energy intake, inflammatory markers, nutrient intakes, plasma Hcy concentrations, and serum vitamin concentrations were used as covariates. In the final regression model, all variables not significantly and independently associated with serum vitamin B₆ concentrations were removed through backward selection method.

^bModel 1 included age; sex; serum creatinine and albumin; total energy intake; smoking habits; folate, vitamin B₆, retinol, β-carotene, vitamin C, vitamin E, and alcohol intakes; circulating concentrations of vitamin B₁₂, folate, α-tocopherol, γ-tocopherol, CRP, IL-6, IL-1 receptor agonist, and IL-6r; and leukocyte count as covariates.

^cModel 2 included the same covariates as model 1 plus plasma Hcy concentrations.

Correlates of Serum Folate Concentrations

When we adjusted the model for age, sex, serum creatinine, serum albumin, total energy intake, and smoking habits, the prevalence of the MTHFR TT genotype was significantly (P < 0.001) higher in the lowest than in the highest tertile of circulating serum folate (Table 3). Serum folate concentrations were lower in participants carrying the MTHFR 677 TT genotype [median (5th–95th percentiles), 5.71 (5.35–6.10) nmol/L] than in participants with the CT [6.59 (6.32–6.87) nmol/L] or CC [6.95 (6.52–7.41) nmol/L] genotype. The percentage of participants with low serum folate concentrations (<5 nmol/L) was significantly higher among those carrying the TT genotype (42.9%) than among those carrying the CC or CT genotype (28.9% and 31.7%, respectively; P < 0.01).

Table 3.

MTHFR C677T genotype distribution according to tertiles of circulating folate.^a

	MTHFR C677T genotype, % (95% confidence interval)
Folate, nmol/L	CC	CT	TTb
≤4.99 (n = 430)	19.7 (19.4–19.9)	51.1 (50.9–51.4)	29.2 (29.0–29.4)
5.0–8.16 (n = 439)	23.9 (23.7–24.1)	53.0 (52.8–53.2)	23.1 (22.9–23.3)
38.16 (n = 451)	25.5 (25.2–25.8)	58.0 (57.7–58.3)	16.5 (16.4–16.6)

^aPercentages, 95% confidence intervals, and P values are adjusted for age, sex, serum creatinine, serum albumin, smoking habits, and total energy intake.

^bPrevalence of the MTHFR TT genotype significantly (P < 0.001) higher in the lowest than in the highest tertile of circulating serum folate.

Serum folate concentrations were related to MTHFR C677T sequence variant in participants with folate intake in the lowest tertile (<221.2 μg/day), but not in participants with folate intake in the highest tertiles (221.2–292.2 and >292.2 μg/day). In particular, in participants in the lowest tertile of folate intake and carrying the TT genotype, serum folate concentrations were significantly lower than those found in participants with the CT or CC genotype (Fig. 1).

Fig. 1. Serum folate concentration according to MTHFR C677T genotype and tertiles of folate intake.

Adjusted mean folate values according to MTHFR C677T genotype were estimated from linear models adjusted for age, sex, serum creatinine, serum albumin, smoking habits, and total energy intake. MTHFR genotypes: □, CC; , CT; ■, TT. NS, not significant.

In the multiple linear regression analysis, after adjustment for several confounders (age, sex, serum creatinine, serum albumin, total energy intake, smoking habits, dietary nutrient intakes, and circulating vitamin concentrations), vitamin C and retinol intakes and circulating serum concentrations of α-tocopherol were significantly and positively associated with circulating serum folate concentrations (Table 4, model 1). Smoking history as well as circulating γ-tocopherol concentrations and homozygosity for the MTHFR C677T sequence variant (TT genotype vs CT + CC genotypes) were significantly and inversely associated with circulating folate concentrations (Table 4, model 1). Serum concentrations of CRP, IL-6, IL-6r, IL-1 receptor antagonist, and albumin and the leukocyte count were not significantly associated with serum folate concentrations [b (SE) log(CRP), 0.028 (0.035); log(IL-6), −0.011 (0.022); log(IL-6r), −0.019 (0.030); log(IL-1 receptor agonist), 0.019 (0.030); albumin, −0.034 (0.024); log(leukocyte count), −0.087 (0.069)]. After the inclusion of plasma Hcy concentrations in the model (model 2), vitamin C and retinol intakes, circulating serum concentrations of vitamin B₆ and B₁₂, sex, and smoking habits remained significant predictors of serum folate concentrations (Table 4, model 2).

Table 4.

Multiple linear regression models relating vitamin intake and inflammation to circulating folate concentrations.^a

	Log(folate in nmol/L)
	Model 1b		Model 2c
Independent variables	b (SE)	P	b (SE)	P
Sex (women vs men)	0.038 (0.014)	0.005	0.014 (0.0054)	0.007
Smoking habits (current smoker vs ex-smoker and never smoked)	−0.060 (0.018)	0.001	−0.049 (0.018)	0.007
Log(retinol in μg/day)	0.031 (0.018)	0.001	0.077 (0.034)	0.021
Log(vitamin C intake in mg/day)	0.244 (0.048)	<0.0001	0.255 (0.053)	<0.0001
Log(α-tocopherol in μmol/L)	0.169 (0.062)	0.006	0.088 (0.060)	0.144
Log(γ-tocopherol in μmol/L)	−0.079 (0.038)	0.035	−0.038 (0.033)	0.246
Log(vitamin B6 in nmol/L)	0.074 (0.021)	<0.0001	0.053 (0.021)	<0.013
Log(vitamin B12 in pmol/L)	0.195 (0.027)	<0.0001	0.130 (0.028)	<0.0001
MTHFR (TT vs CT+TT)	−0.041 (0.016)	0.009	−0.024 (0.015)	0.116
Log(Hcy in μmol/L)			−0.519 (0.052)	<0.0001

^aIn the multiple linear regression analysis, circulating folate (serum concentrations) was used as the dependent variable and age, sex, serum creatinine, serum albumin, smoking habits, total energy intake, inflammatory markers, nutrient intakes, Hcy, and circulating vitamin concentrations were used as covariates. In the final regression model, all variables not significantly and independently associated with circulating folate were removed through backward selection method.

^bModel 1 included age; sex; serum creatinine; total energy intake; smoking habits; folate, vitamin B₆, thiamine, retinol, β-carotene, vitamin C, and vitamin E intakes; circulating concentrations of vitamin B₆, vitamin B₁₂, and α- and γ-tocopherol; and MTHFR sequence variant (TT genotype vs CT plus CC genotypes) as the covariates.

^cModel 2 included the same covariates as model 1 plus plasma Hcy concentrations.

Discussion

Our results demonstrate that in the elderly, inflammatory markers such as CRP, IL-6r, and serum antioxidant vitamin concentrations are correlates of serum vitamin B₆ concentrations independent of dietary intake, smoking history, serum albumin, and plasma Hcy concentrations and that antioxidant vitamin intakes, but not inflammatory markers, are significant correlates of serum folate concentrations independent of relevant covariates, including folate intake and smoking history. Furthermore, this large study extends, in elderly adults, the observation that the TT genotype of the MTHFR C677T sequence variant is significantly and independently associated with low concentrations of serum folate, particularly in participants with low folate intake.

Our findings support the notion already suggested by the Framingham study of a relationship between vitamin B₆ and inflammation, and provide an original contribution to this literature by showing that such a relationship is not limited to CRP but also involves the sIL-6r. In addition, our data are in accordance with recent evidence of a link between inflammation and vitamin B₆ not only in patients affected by inflammatory diseases such as rheumatoid arthritis and inflammatory bowel disease (19, 26), but also in stroke patients and in controls (27). Moreover, our findings are at variance with those observed in a subset of the ARIC study, in which leukocyte count but not CRP was a correlate of serum vitamin B₆ (28). Changes in sIL-6r concentrations have been determined in numerous clinical disorders (29), but the biological significance of sIL-6r within these pathologies remains unclear.

The mechanism by which inflammation is associated with low vitamin B₆ concentrations has not been completely elucidated. It has been demonstrated that during a chronic inflammatory state, circulating and hepatic concentrations of vitamin B₆ tend to decrease (19). Because vitamin B₆ is involved in the repair and synthesis of nucleic acid and proteins, low vitamin B₆ concentrations may reflect the increased consumption of this coenzyme in the accelerated synthesis of cytokines and antibodies and in accelerated cell proliferation.

Interestingly, we found that low vitamin B₆ concentrations were independently associated with low β-carotene intake and α-tocopherol concentrations, suggesting that an imbalance in the oxidative stress and antioxidant mechanisms may contribute to the decreases in vitamin B₆ concentrations in inflammatory states.

We previously demonstrated that inflammation is strongly related to Hcy (22), and it is well known that vitamin B₆ affects the serum concentration of Hcy. Thus, it was reasonable to hypothesize that the relationship between vitamin B₆ and inflammatory state could be mediated by Hcy. However, the inclusion of Hcy in the linear regression model did not substantially attenuate the relationship between inflammatory markers and vitamin B₆.

In our study, vitamin C intake and α-tocopherol concentrations were strongly associated with serum folate concentrations, independent of age, sex, smoking history, and folate intake, suggesting that noncompensated oxidative stress may contribute to the decrease in folate as well as serum vitamin B₆ concentrations. The antioxidant molecule vitamin C is known to influence folate synthesis by enhancing formation of the polyglutamate derivatives from tetrahydrofolic acid and enhancing the transformation of folic acid into tetrahydrofolic acid (30). Interestingly, in Italian blood donors who were smokers, supplementation with folate plus vitamin C led to greater increases in serum and erythrocyte folate than did supplementation with folate alone (31).

By showing that MTHFR C677T sequence variant is an independent predictor of serum folate in older persons, especially those with low folate intake, our study confirms similar observations obtained in middle-aged European and North American populations (32–34).

In addition to the well-known influence on Hcy concentrations, folate may influence genomic DNA methylation reactions, particularly in participants with low serum folate concentrations and persons homozygous for the MTHFR C677T TT sequence variant (33).

The finding of the present investigation, demonstrating that low vitamin B₆ is an independent correlate of proinflammatory state, may appear not to be consistent with the results of recent clinical trials showing no effect of folic acid, vitamin B₆, and vitamin B₁₂ supplementation in reducing cardiovascular events (35, 36). Nevertheless, all of the interventional trials confirmed the role of Hcy as marker of an increased risk of vascular events.

A limitation of this study is represented by the nature of our study (cross-sectional study), which did not allow us to determine the causative role of inflammation in inducing low circulating vitamin B₆ concentrations.

In conclusion, we demonstrated that CRP and sIL-6r, 2 important biomarkers of inflammation, are independent correlates of vitamin B₆ concentrations. Antioxidant vitamins, but not inflammatory markers, were independent correlates of serum folate. The authors of future studies looking at vitamins as modifiable risk factors for health outcomes must consider inflammation related to vitamin B₆ and oxidative stress related to vitamin B₆ and folate as potential confounders in the analyses.