Associations between Diet and Toenail Arsenic Concentration among Pregnant Women in Bangladesh: A Prospective Study

Abstract

This prospective study evaluated the relationship between long-term dietary habits and total arsenic (As) concentration in toenail clippings in a cohort of 1616 pregnant women in the Bangladeshi administrative regions of Sirajdikhan and Pabna Sadar. Diet was assessed at Gestation Week 28 and at Postpartum Month 1, using a locally-validated dish-based semi-quantitative food-frequency questionnaire. Toenail As concentration was analyzed by microwave-assisted acid digestion and inductively coupled plasma mass spectrometry. Associations between natural log-transformed consumption of individual food items and temporally matched natural log-transformed toenail As concentration were quantified using general linear models that accounted for As concentration in the primary drinking water source and other potential confounders. The analysis was stratified by As in drinking water (≤50 μg/L versus >50 μg/L) and the time of dietary assessment (Gestation Week 28 versus Postpartum Week 1). Interestingly, toenail As was not significantly associated with consumption of plain rice as hypothesized. However, toenail As was positively associated with consumption of several vegetable, fish and meat items and was negatively associated with consumption of rice, cereal, fruits, and milk based food items. Further studies in pregnant women are needed to compare As metabolism at different levels of As exposure and the interaction between dietary composition and As absorption.

Keywords: Bangladesh; arsenic exposure; dietary assessment; food frequency questionnaire; pregnancy.

Figure 1

Study flow chart.

Figure 2

Associations between natural log-transformed food intake (g/day) and natural log-transformed toenail As concentration (μg/g) measured by the first and the second dietary assessment (FFQ1 and FFQ2, respectively) among women exposing to drinking water As concentration of >50 μg/L and ≤50 μg/L in Bangladesh. Effect size is the slope coefficients ($\widehat{\mathsf{\beta }}$) for each dietary item and has the unit of ln((toenail As concentrations, μg/g)·(g/day)⁻¹). “C” stands for crude model adjusted for water As concentration only; “A” stands for adjusted model adjusted for water As level, sex, smoking in the living environment, chewing betel nut, BMI, daily water intake, daily energy intake, and education level. Q-value accounting for multiple comparisons using the false discovery rate (FDR = 0.05) method. (Quantitative values presented in the Supplementary Materials, Tables S1 and S2).