Vitamin B-6 Intake Is Modestly Associated with Arsenic Methylation in Uruguayan Children with Low-Level Arsenic Exposure

ABSTRACT

Background

Detoxification of inorganic arsenic (iAs) occurs when it methylates to form monomethylarsonic acid (MMA) and dimethylarsinic acid (DMA). Lower proportions of urinary iAs and MMA, and higher proportions of DMA indicate efficient methylation. The role of B-vitamins in iAs methylation in children with low-level arsenic exposure is understudied.

Objectives

Our study objective was to assess the association between B-vitamin intake and iAs methylation in children with low-level arsenic exposure (<50 µg/L in water; urinary arsenic 5–50 µg/L).

Methods

We conducted a cross-sectional study in 290 ∼7-y-old children in Montevideo. Intake of thiamin, riboflavin, niacin, vitamin B-6, and vitamin B-12 was calculated by averaging 2 nonconsecutive 24-h recalls. Total urinary arsenic concentration was measured as the sum of urinary iAs, MMA, and DMA, and adjusted for urinary specific gravity; iAs methylation was measured as urinary percentage As, percentage MMA, and percentage DMA. Arsenic concentrations from household water sources were assessed. Linear regressions tested the relationships between individual energy-adjusted B-vitamins and iAs methylation.

Results

Median (range) arsenic concentrations in urine and water were 9.9 (2.2–48.7) and 0.45 (0.1–18.9) µg/L, respectively. The median (range) of urinary percentage iAs, percentage MMA, and percentage DMA was 10.6% (0.0–33.8), 9.7% (2.6–24.8), and 79.1% (58.5–95.4), respectively. The median (range) intake levels of thiamin, riboflavin, niacin, and vitamin B-6 were 0.81 (0.19–2.56), 1.0 (0.30–2.24), 8.6 (3.5–23.3), and 0.67 (0.25–1.73) mg/1000 kcal, respectively, whereas those of folate and vitamin B-12 were 216 (75–466) and 1.7 (0.34–8.3) µg/1000 kcal, respectively. Vitamin B-6 intake was inversely associated with urinary percentage MMA (β = −1.60; 95% CI: −3.07, −0.15). No other statistically significant associations were observed.

Conclusions

Although vitamin B-6 intake was inversely associated with urinary percentage MMA, our findings suggest limited support for a relation between B-vitamin intake and iAs methylation in children exposed to low-level arsenic.

Introduction

Inorganic arsenic (iAs) is the most common form of arsenic in nature. It is highly toxic and is classified as a group 1 human carcinogen (1, 2). Adverse health outcomes related to iAs exposure are well studied in adults and children at both high (>50 µg/L in water) (3–6) and low (<50 µg/L in water, urinary arsenic ∼5–50 µg/L) (7–9) exposure levels. In the human body, iAs is methylated by S-adenosylmethionine, a product of a 1-carbon metabolism cycle. The degree of iAs toxicity is a function of the dose of exposure and the individual's capacity to detoxify iAs through methylation. Two methylation cycles occur in the liver, resulting in the production of monomethylarsonic acid (MMA) and dimethylarsinic acid (DMA) (10). The proportions of iAs, MMA, and DMA excreted in urine reflect the individual's capacity to convert iAs to MMA and then DMA (11). Higher proportions of iAs and MMA and a lower proportion of DMA in the urine are regarded as susceptibility markers for arsenic-related health outcomes (11–15).

Methylation capacity is influenced to some extent by a number of factors, such as age, sex, pregnancy status, smoking status, and ethnicity (16). Nutrients also play a role in this regard (16). Folate acts as a methyl donor, and riboflavin, vitamin B-6, and vitamin B-12 as cofactors within the 1-carbon cycle (17, 18). Dietary intake and serum concentrations of several B-vitamins, including thiamin, riboflavin, vitamin B-6, folate, and vitamin B-12, appear to be associated with efficient methylation of iAs in observational studies (17, 19–24). For instance, in adults in New Hampshire, the weighted sum of the intake of 6 different B-vitamins (thiamin, riboflavin, niacin, vitamin B-6, folate, and vitamin B-12) from foods had an inverse association with urinary percentage MMA (β = −1.03; 95% CI: −1.91, −0.15) (21). A double-blinded, placebo-controlled trial in Bangladeshi adults assessing the effect of folic acid supplementation on iAs methylation showed that supplementation at a dose of 400 µg/d for 12 wk was associated with a 13% reduction (P < 0.0001) in urinary percentage MMA in the intervention group, compared with a 10% reduction in the placebo group (25). In the same participants, MMA concentrations in blood decreased by 22.24% in the intervention group, compared with 1.24% in the placebo group (P < 0.0001) (26).

A number of unanswered questions remain about the role of B-vitamins in iAs methylation. First, there is only 1 study in children, and that focused on plasma folate, cobalamin, cysteine, and homocysteine in relation to iAs methylation (17). Second, methylation is thought to differ somewhat between adults and children, possibly because factors that reduce methylation efficiency in adults, such as smoking and reduced liver function, increase with age (27). Because children appear to have more efficient methylation than adults, findings from studies in adults cannot be extrapolated to children. Third, the majority of the literature is from geographical areas that frequently have high levels of arsenic exposure (>50 µg/L in water) (17, 19, 25, 26). In fact, the only study on arsenic methylation in children was also conducted in a high-exposure area (mean urinary arsenic concentrations = 85.0 ± 69.70 µg/L) (17). Together, this points to a dearth of evidence on the associations of B-vitamins with iAs methylation in children when exposures are relatively low. There are some studies on B-vitamins and iAs methylation, largely in adults exposed to low-level arsenic (21, 23, 24, 28), but none specifically in children. Finally, because most studies relating B-vitamins to iAs methylation were conducted in malnourished groups, the role of nutrients in the detoxification of iAs in well-nourished populations is unclear. Because low-level arsenic exposure is associated with adverse health outcomes such as diabetes, abnormal blood pressure, inflammation, and skin diseases (7–9, 29), understanding the relation between B-vitamins and iAs methylation in populations with low-level exposure and more optimal nutritional status has the potential to inform dietary guidelines for these groups.

The aim of our cross-sectional study was to investigate the relation between intake of thiamin, riboflavin, niacin, vitamin B-6, and vitamin B-12 and iAs methylation efficiency in children. We previously examined the association between specific food groups, diet patterns, dietary folate intake (but not the intake of other B-vitamins), and iAs methylation in the same children (30). Because B-vitamins other than folate are involved in the methylation of iAs, our study provides much needed evidence on the role of B-vitamins in the detoxification of iAs in pediatric populations exposed to low levels of arsenic. We measured urinary percentage iAs, percentage MMA, and percentage DMA in ∼7-y-old children exposed to low levels of arsenic through food and drinking water in Montevideo, Uruguay. We hypothesized that B-vitamin intake would be associated with lower urinary percentage iAs and percentage MMA, and higher percentage DMA.

Methods

Study setting and participant recruitment

This study, called "Salud Ambiental Montevideo," was conducted in private elementary schools in various neighborhoods of Montevideo, which were considered to have potential sources of metal exposures based on previous studies and knowledge of local physicians (31–33). In this study, 357 children aged ∼7 y and their mothers were enrolled. The details of recruitment can be found in a previous publication (34). After excluding the participants with missing data, a complete case sample of 290 children, defined as having complete data on the exposures, outcomes, and covariates of interest, was used for the present analysis. The research protocol was approved by the Institutional Review Boards at the Catholic University of Uruguay, University of the Republic of Uruguay, Pennsylvania State University, and the State University of New York at Buffalo. Data collection began only after informed consent was obtained from the participants.

Measurements

Urinary arsenic measurement

Total urinary arsenic concentration was measured as the sum of urinary iAs, MMA, and DMA. Methylation of iAs was assessed as the proportions of urinary percentage iAs, percentage MMA, and percentage DMA. First morning void samples were collected by the participants in cups that were rinsed with 10% nitric acid and deionized water, and were given to them beforehand. Urinary arsenic concentration was assessed using HPLC-hydride generation (HG)-inductively coupled plasma mass spectrometry (ICP-MS) (HG selects inorganic arsenic and its methylated metabolites into the ICP-MS) at the Karolinska Institute, Stockholm, Sweden. Details of this method have been described previously (34, 35). The limit of detection was 0.1 μg/L for inorganic arsenic(III) and MMA, 0.2 μg/L for DMA, and 0.3–0.5 μg/L for inorganic arsenic(V). The intra- and interassay CVs were ∼4%. Seven urine samples (2.1%) were below the limit of detection for inorganic arsenic(III), and 26 (7.9%) were below the limit of detection for inorganic arsenic(V). The measured values were used in statistical analyses. Total urinary arsenic concentrations were adjusted for specific gravity to account for the hydration status of participants.

Water arsenic measurement

Water samples from the kitchen taps or water storage containers were collected as part of the home visit by the study staff. This water was passed through a 0.45-mm filter (VWR International) into a plastic bottle that was previously rinsed with 10% nitric acid and deionized water. Water arsenic concentrations were measured at the Pennsylvania State University Materials Characterization Laboratory using ICP-MS with collision cell technology (XSERIES2; Thermo Scientific). The limit of detection for arsenic was 0.03 mg/L.

B-vitamin intake

Two 24-h dietary recalls were conducted by nutritionists with the mother or a caregiver familiar with the child's diet. The child was also present and contributed to these recalls. The first recall was conducted in person at the school and the second was conducted over the phone ≥2 wk later. Information about the name of the meals, time and place of consumption, amounts consumed, preparation methods, recipe ingredients, brand names of commercial products, and use of vitamin and mineral supplements was obtained. Neutral probing questions and visual aids were used to elicit accurate responses. All foods were assigned a unique code and entered, along with the amounts consumed, into a database that contained the nutrient composition of typical Uruguayan foods and preparations. The amounts of thiamin, riboflavin, niacin, vitamin B-6, and vitamin B-12 present in each food were calculated using either the Uruguayan nutrient database or the USDA National Nutrient Database for Standard Reference, Release 28 (Version Current: September 2015) for the foods that were not listed in the Uruguayan database (36–39). Only “unenriched” foods from the USDA database were used in these calculations because foods in Uruguay are not fortified with B-vitamins other than folate. The total dietary intake of the B-vitamins was derived from the amounts of foods reported by participants, based on the average of the 2 dietary recalls. Each B-vitamin intake estimated in this way was adjusted for total energy intake using the nutrient density energy adjustment method and expressed as intake/1000 kcal/d.

Anthropometry

Children's height and weight were measured by trained nurses using a portable stadiometer (Seca 214; Shorr Productions) and a digital scale (Seca 872; Shorr Productions), respectively.

Statistical analyses

All analyses were conducted in SAS version 9.4 (SAS Institute Inc) in a sample of 290 participants, defined as having complete data on the exposures, outcomes, and covariates of interest. Descriptive analyses included calculating the means ± SD and medians (range) of water arsenic, total urinary arsenic, percentage iAs, percentage MMA, percentage DMA, and intakes of B-vitamins. Data from the NHANES, cycles 2009–2014, were also analyzed to compare B-vitamin intakes obtained from one 24-h diet recall in participants of comparable ages. Means ± SD of percentage iAs, percentage MMA, and percentage DMA by socioeconomic, anthropometric, and biochemical characteristics of participants were also calculated.

Linear regression analyses were conducted to assess the associations between dietary intake of individual energy-adjusted B-vitamins (thiamin, riboflavin, niacin, vitamin B-6, vitamin B-12) and the proportions of urinary metabolites of iAs (percentage iAs, percentage MMA, percentage DMA). Models were adjusted for age, sex, BMI, total urinary arsenic (because higher total exposure could result in higher levels of the methylated species, independently of methylation efficiency), and rice intake (because rice is a source of MMA and DMA, which would contribute to urinary MMA and DMA concentrations independently of methylation efficiency). Although seafood is also a source of DMA (40), models were not adjusted for seafood intake because only 22 participants reported consuming any seafood. A principal component analysis was conducted using all B-vitamins to assess the association between B-vitamins as a group and iAs methylation. The principal component analysis identified a single factor, with loadings ranging from 0.22 to 0.26 on all the B-vitamins. This factor explained 68% of the variance. As a next step, a factor score was calculated based on these loadings, referred to hereafter as the “B-vitamin index.” This B-vitamin index was entered into linear models as an independent variable with the urinary metabolites of iAs (percentage iAs, percentage MMA, percentage DMA) as dependent variables.

Results

The means ± SD and medians (range) of total urinary arsenic, water arsenic, and proportions of urinary metabolites of iAs (percentage iAs, percentage MMA, percentage DMA) are shown in Table 1. The median (range) for water arsenic concentration was 0.45 µg/L (0.1–18.9 µg/L), and for urinary arsenic concentration was 9.9 µg/L (2.2–48.7 µg/L). Table 2 presents the medians (range) of B-vitamin intake levels in the study participants, along with the intake levels of US children aged 5–8 y participating in the 2009–2014 NHANES, for comparison. Also presented is the RDA for the B-vitamins (41). The B-vitamin intake levels in study participants were comparable to those of NHANES participants, particularly for thiamin and riboflavin. The median intakes of niacin, vitamin B-6, and vitamin B-12 were slightly lower in Uruguayan children. The upper range of intake for all the B-vitamins was higher in NHANES participants. Comparisons with the RDA reveal an overall healthy intake of B-vitamins in Uruguayan study participants (41).

TABLE 1.

Total urinary arsenic, water arsenic, and proportions of urinary metabolites of inorganic arsenic in participants of the Salud Ambiental Montevideo study (n = 290)¹

Variable	Value2	Value3
Total urinary arsenic,4 µg/L	12.1 ± 7.6	9.9 (2.2–48.7)
Water arsenic, µg/L	0.64 ± 1.5	0.45 (0.1–18.9)
% iAs	11.5 ± 5.8	10.6 (0.0–33.8)
%MMA	9.8 ± 3.6	9.7 (2.6–24.8)
%DMA	78.8 ± 7.4	79.1 (58.5–95.4)

¹DMA, dimethylarsinic acid; iAs, inorganic arsenic; MMA, monomethylarsonic acid.

²Values presented as mean ± SD.

³Values presented as median (range).

⁴Urinary arsenic measured as the sum of iAs, MMA, and DMA, then adjusted for urinary specific gravity.

TABLE 2.

B-vitamin intake of the Salud Ambiental Montevideo study participants (n = 290), and that of 5–8-y-old NHANES 2009–2014 participants (n = 2318)

Variable	Salud Ambiental Montevideo study participants1	NHANES participants1,2	RDA3 (41)
Thiamin, mg/1000 kcal	0.81 (0.19–2.56)	0.80 (0.22–2.90)	0.6 mg
Riboflavin, mg/1000 kcal	1.0 (0.30–2.24)	1.01 (0.19–3.68)	0.6 mg
Niacin, mg/1000 kcal	8.6 (3.5–23.3)	10.36 (1.39–55)	8.0 mg, niacin equivalents
Vitamin B-6, mg/1000 kcal	0.67 (0.25–1.73)	0.84 (0.12–6.74)	0.6 mg
Vitamin B-12, µg/1000 kcal	1.7 (0.34–8.3)	2.35 (0.08–12.38)	1.2 µg

¹Values presented as median (range).

²Calculated by authors using the publicly available NHANES data from survey cycles 2009–2014.

³The RDAs presented are for children aged 4–8 y.

Table 3 shows the means ± SD of proportions of urinary metabolites of iAs by socioeconomic, anthropometric, and biochemical characteristics of the participants, split at the median. Children aged ≥81 mo had a lower urinary percentage MMA and higher urinary percentage DMA, indicating more efficient methylation compared with younger participants. Similarly, participants with a weight-for-age z-score ≥0.68 had a lower urinary percentage MMA and higher urinary percentage DMA, indicating more efficient methylation. Urinary percentage iAs and percentage MMA were lower in those with total urinary arsenic ≥9.9 µg/L, compared with those with total urinary arsenic <9.9 µg/L. Conversely, urinary percentage DMA was higher in those with total urinary arsenic ≥9.9 µg/L. Table 4 presents the associations of intake of individual B-vitamins and the B-vitamin index with the proportions of urinary metabolites of iAs. Vitamin B-6 intake was inversely associated with percentage MMA (β = −1.60; 95% CI: −3.07, −0.14).

TABLE 3.

Proportions of urinary metabolites of inorganic arsenic by socioeconomic, anthropometric, and biochemical characteristics of the Salud Ambiental Montevideo study participants (n = 290)¹

Variables	n (%)	%iAs1	%MMA2	%DMA2
All participants	290 (100)	11.5 ± 5.8	9.8 ± 3.6	78.7 ± 7.4
Age, mo
<81	132 (45.5)	12.2 ± 6.5	10.3 ± 3.7	77.5 ± 7.6
≥81	158 (54.5)	10.9 ± 5.2	9.4 ± 3.4	79.7 ± 7.1
P value2		0.07	0.04	0.01
Sex
Girls	131 (45.2)	12.0 ± 5.7	9.6 ± 3.6	78.4 ± 7.5
Boys	159 (54.8)	11.1 ± 5.9	10.0 ± 3.6	79.0 ± 7.3
P value3		0.17	0.45	0.49
Height-for-age z-score
<0.37	145 (50.0)	11.6 ± 5.9	10.1 ± 3.8	78.3 ± 7.3
≥0.37	145 (50.0)	11.4 ± 5.7	9.5 ± 3.4	79.1 ± 7.5
P value3		0.81	0.16	0.39
Weight-for-age z-score
<0.68	143 (49.3)	11.9 ± 5.8	10.3 ± 3.7	78.0 ± 7.3
≥0.68	147 (50.7)	11.1 ± 5.8	9.4 ± 3.4	79.6 ± 7.4
P value3		0.24	0.03	0.05
Total urinary arsenic4, µg/L
<9.9	145 (50.0)	13.6 ± 6.2	10.5 ± 3.5	75.9 ± 6.8
≥9.9	145 (50.0)	9.3 ± 4.4	9.1 ± 3.5	81.5 ± 6.8
P value2		<0.0001	<0.001	<0.0001

¹DMA, dimethylarsinic acid; iAs, inorganic arsenic; MMA, monomethylarsonic acid.

²Values presented as mean ± SD.

³Obtained from independent t tests.

⁴Measured as the sum of iAs, MMA, and DMA, then adjusted for urinary specific gravity.

TABLE 4.

Associations between B-vitamin intake and the proportions of urinary metabolites of inorganic arsenic in the Salud Ambiental Montevideo study participants (n = 290)¹

	%iAs		%MMA		%DMA
	β (95% CI)	β2 (95% CI)	β (95% CI)	β2 (95% CI)	β (95% CI)	β2 (95% CI)
Thiamin, mg/1000 kcal	1.25 (−0.85, 3.35)	1.21 (−0.76, 3.18)	0.54 (−0.75, 1.83)	0.05 (−1.19, 1.29)	−1.80 (−4.47, 0.88)	−1.27 (−3.68, 1.13)
Riboflavin, mg/1000 kcal	0.79 (−1.46, 3.04)	0.91 (−1.15, 2.98)	−0.06 (−1.45, 1.32)	−0.23 (−1.53, 1.07)	−0.72 (−3.58, 2.14)	−0.69 (−3.21, 1.84)
Niacin, mg/1000 kcal	−0.06 (−0.28, 0.16)	0.01 (−0.20, 0.21)	−0.07 (−0.20, 0.07)	−0.06 (−0.19, 0.07)	0.12 (−0.16, 0.41)	0.05 (−0.20, 0.30)
Vitamin B-6, mg/1000 kcal	−0.30 (−2.87, 2.28)	−0.09 (−2.43, 2.26)	−1.66 (−3.23, −0.09)	−1.60 (−3.07, −0.14)	1.94 (−1.32, 5.21)	1.68 (−1.18, 4.54)
Vitamin B-12, µg/1000 kcal	0.46 (−0.39, 1.31)	0.47 (−0.31, 1.25)	−0.12 (−0.62, 0.37)	−0.17 (−0.66, 0.32)	−0.34 (−1.42, 0.75)	−0.29 (−1.25, 0.66)
B-vitamin index	0.17 (−0.50, 0.85)	0.24 (−0.38, 0.86)	−0.13 (−0.55, 0.28)	−0.19 (−0.57, 0.20)	−0.04 (−0.90, 0.81)	−0.06 (−0.82, 0.70)

¹DMA, dimethylarsinic acid; iAs, inorganic arsenic; MMA, monomethylarsonic acid.

²Models adjusted for age, sex, BMI, specific gravity–adjusted sum of iAs metabolites, and rice intake.

Discussion

We found that in 290 Uruguayan schoolchildren exposed to low levels of arsenic, water arsenic concentrations were low, but food could be an important exposure source. We found evidence of an association between dietary intake of vitamin B-6 and efficient arsenic methylation; however, the point estimate was small. These findings suggest that, on average, the consumption of additional dietary sources or supplements of vitamin B-6 would have limited impact on arsenic methylation in relatively well-nourished children with low-level arsenic exposure.

Few studies have investigated the role of B-vitamins and iAs methylation in populations exposed to low concentrations of arsenic (<50 µg/L in water, urinary arsenic ∼5–50 µg/L), and in relatively well-nourished populations (21, 23, 24, 28). Furthermore, there is only 1 study in children, but their exposure to arsenic was high (mean urinary arsenic = 85.0 ± 69.70 µg/L) (17). In the present study, the range of urinary arsenic concentrations was 2.2–48.7 µg/L. Typically, the proportion of iAs in urine is 10–30%, MMA 10–20%, and DMA 60–80% (42). Children are thought to have somewhat better iAs methylation efficiency than adults (27). The relatively low proportions of iAs (median 10.6%) and MMA (9.7%), and higher proportion of DMA (79.1%) in our study are consistent with better methylation efficiency in children. Of note, within the low urinary concentrations we observed (2.2–48.7 µg/L), children with higher exposures had more methylated arsenic in their urine than children with lower exposures. This unexpected finding could be because when water arsenic concentrations are low, food often becomes an important exposure source, and many foods such as rice and seafood are major sources of DMA (40, 43, 44).

Based on the role of B-vitamins in the 1-carbon cycle and arsenic metabolism, we hypothesized that B-vitamin intake would be inversely related to urinary percentage iAs and percentage MMA, and positively associated with percentage DMA. Metabolism of iAs in the human body occurs through the 1-carbon cycle, wherein S-adenosylmethionine, produced from methionine, is required in the conversion of iAs to its methylated species (45). S-adenosylhomocysteine is produced as a result of this reaction, and in turn is converted to homocysteine. Homocysteine is either remethylated to methionine, or is directed to the transsulfuration pathway. Folate is reduced to dihydrofolate and tetrahydrofolate (THF). THF then uses a methyl group from serine to form glycine and 5,10-methylene-THF. An enzyme complex containing pyridoxal 5′-phosphate (a form of vitamin B-6)-dependent glycine decarboxylase is required in this reaction (46). 5-Methyl-THF is then formed from the reduction of 5,10-methylene-THF wherein riboflavin acts as a cofactor; 5-methyl-THF transfers its methyl group to homocysteine to form methionine. Vitamin B-12 acts as a cofactor in this reaction. Methionine then produces S-adenosylmethionine. Vitamin B-6–dependent enzymes also catalyze the conversion of homocysteine to cysteine, which further leads to the synthesis of glutathione, required for the reduction of the pentavalent arsenic metabolites to the trivalent forms, which accept a methyl group to form MMA and DMA (47). Thus, riboflavin, vitamin B-6, folate, and vitamin B-12 play important roles in the methylation of iAs (45, 48–50).

Our findings on the association between vitamin B-6 intake and urinary percentage MMA are consistent with other reports, all in adults (21, 23, 28). In adults from New Hampshire, a weighted sum of dietary intake of B-vitamins had an inverse association with urinary percentage MMA; thiamin, vitamin B-6, and vitamin B-12 were the main contributors to this association (21). In the Strong Heart Study, vitamin B-6 intake was associated with modestly reduced percentage iAs and percentage MMA, and increased percentage DMA, indicating an efficient second methylation cycle (23). A study using the 2003–2004 NHANES data showed that dietary intake of vitamin B-6 was associated with lower urinary percentage iAs (28). Vitamin B-6 intake was inversely associated with urinary percentage MMA in our study. Supplemental Table 1 presents the post hoc power calculation if the sample size was increased to 500 or 1000 participants, and is meant to aid in the interpretation of the observed results. These analyses suggest that a higher sample size would have given us more power to detect an association of vitamin B-6 with urinary percentage MMA, but not with percentage iAs or percentage DMA. This indication that the observed results might not be due to chance increases confidence in our findings. The top sources of vitamin B-6 in our study included bananas, meats, and meat-based dishes; in a previous study, we reported that meat intake was associated with higher urinary percentage DMA in our population (30).

Although thiamin and niacin are not directly involved in the 1-carbon metabolism cycle, they have been inversely associated with arsenic-induced health outcomes in arsenic-contaminated regions (51, 52). Thiamin has been shown to reduce arsenic concentration in tissues of rats (53). Further investigation is needed to understand the role of thiamin and niacin in arsenic toxicity in humans. We found no association between thiamin intake and urinary metabolites of iAs, consistent with another study in US adults (low-exposure area) (24). Conversely, thiamin contributed markedly to the inverse association between the weighted sum of B-vitamins and urinary percentage MMA in adults with low-level arsenic exposure in New Hampshire (21). Our null findings for niacin are also consistent with 2 other studies in low- to moderate-exposure areas, 1 of which reported largely adequate vitamin intake levels (13, 22). Our power analysis (Supplemental Table 1) indicates that increasing the sample size would be unlikely to yield statistically significant findings for thiamin and niacin.

We found no association between riboflavin intake and iAs methylation, similar to studies from the United States and Mexico, which included adult participants with generally adequate riboflavin intakes (13, 22). There is inconsistent evidence on the link between vitamin B-12 intake and iAs methylation. It was associated with lower urinary percentage iAs and higher percentage MMA in Bangladeshi adults (8% men and 13% women with vitamin B-12 deficiency) (54), and with lower percentage iAs in Mexican women (73% with intake below recommended level) (22). Our findings are consistent with studies conducted in populations with adequate vitamin B-12 intake or status, measured in terms of plasma cobalamin concentrations, which indicate no association between vitamin B-12 and iAs methylation for both low and high levels of arsenic exposure (17, 19, 23, 55).

Participants in our study appeared to have largely adequate B-vitamin intakes; however, interpretation of dietary adequacy based on two 24-h dietary recalls should be done with caution, because these might not reflect habitual intakes, especially for infrequently eaten foods that could be rich sources of some micronutrients (e.g., leafy greens). Furthermore, sources of inter- and intraindividual error across repeated dietary recalls should be taken into account for accurate estimates of adequacy. Foods can be sources of arsenic exposure, particularly DMA; higher urinary DMA could result from the intake of certain foods independent of iAs methylation efficiency. For instance, rice is a well-studied source of DMA (43, 44, 56). Rice intake was associated with higher urinary percentage DMA in our study population (30). Seafood is a predominant source of organic arsenicals; fish usually contain arsenobetaine, which is a nontoxic compound excreted from the human body without undergoing any chemical changes (40). However, other organic arsenicals such as arsenosugars or arsenolipids, which are present in mollusks and seaweed, are metabolized into several chemicals in the human body, including DMA (57, 58). Thus, urinary percentage DMA reflects exposure to iAs, methylation efficiency, and seafood consumption. In our study, seafood intake was extremely low; only 22 participants reported any intake over the 2 d of recall.

Our findings need to be interpreted in light of certain limitations. We had an overall participation rate of 53%, which could have resulted in selection bias if participation was associated with both the exposure and outcome. However, arsenic exposure is not assessed on a regular basis in Uruguay, so participants are likely to be unaware of their arsenic exposure levels. Thus, participation decisions are unlikely to be related to our study outcomes. Because the participation rate varied between schools, we were able to compare a number of variables of interest between schools with low (≤50%) and high (>50%) participation rates. Children from the 2 groups of schools did not differ on variables representing study covariates, exposures, or the outcome of interest (Supplemental Table 2). Therefore, we believe that participation rates did not affect our findings. We assessed arsenic exposure and methylation efficiency based on 1 spot urine sample. Urinary arsenic is a marker of recent exposure, and might not represent typical exposure. Similarly, nutrient intakes were calculated from two 24-h dietary recalls, which might not represent the usual dietary intake over a long time period or seasonal differences in food consumption. However, data collectors provided several cues and/or visual aids to reduce errors in recall, and data from the 2 recalls were averaged. A further limitation is that we assessed dietary intake of B-vitamins using 2 different databases, relying on the USDA database when information was not available in the Uruguayan database. This might have resulted in measurement error, despite our efforts to match the foods closely. In addition, for mixed dishes, we relied on commonly used recipes, whereas the ingredients and their amounts could have differed among households. Among our study strengths are the use of repeat diet recalls, and the inclusion of several B-vitamins (thiamin, riboflavin, niacin, vitamin B-6, and vitamin B-12), which expands the existing literature, as well as the use of a standard and well-accepted method of assessing urinary arsenic and its methylated species. Additionally, urine samples were collected for arsenic assessment on the day of the first dietary recall. Urinary arsenic, a short-term marker of exposure, was thus well matched to the dietary assessments and likely reflected dietary sources of arsenic, as well as the associations of B-vitamins with methylation. Finally, we adjusted our models for factors known to influence methylation capacity in children (age, BMI) and the presence of methylated arsenic species in urine (rice intake).

In conclusion, our findings suggest limited support for a relation between B-vitamin intake and iAs methylation in children exposed to low-level arsenic.

Notes

This work was supported by the National Institutes of Health, and the Fogarty International Center (ES019949, PI: KK; and ES016523, PI: KK).

Author disclosures: The authors report no conflicts of interests.

Supplemental Tables 1 and 2 are available from the “supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/.

Abbreviations used: DMA, dimethylarsinic acid; iAs, inorganic arsenic; HG, hydride generation; ICP-MS, inductively coupled plasma mass spectrometry; MMA, monomethylarsonic acid; THF, tetrahydrofolate.