PLoS ONE (Jan 2022)

Arsenic exposure and respiratory outcomes during childhood in the INMA study.

  • Antonio J Signes-Pastor,
  • Susana Díaz-Coto,
  • Pablo Martinez-Camblor,
  • Manus Carey,
  • Raquel Soler-Blasco,
  • Miguel García-Villarino,
  • Ana Fernández-Somoano,
  • Jordi Julvez,
  • Paula Carrasco,
  • Aitana Lertxundi,
  • Loreto Santa Marina,
  • Maribel Casas,
  • Andrew A Meharg,
  • Margaret R Karagas,
  • Jesús Vioque-Lopez

DOI
https://doi.org/10.1371/journal.pone.0274215
Journal volume & issue
Vol. 17, no. 9
p. e0274215

Abstract

Read online

Ingested inorganic arsenic (iAs) is a human carcinogen that is also linked to other adverse health effects, such as respiratory outcomes. Yet, among populations consuming low-arsenic drinking water, the impact of iAs exposure on childhood respiratory health is still uncertain. For a Spanish child study cohort (INfancia y Medio Ambiente-INMA), low-arsenic drinking water is usually available and ingestion of iAs from food is considered the major source of exposure. Here, we explored the association between iAs exposure and children's respiratory outcomes assessed at 4 and 7 years of age (n = 400). The summation of 4-year-old children's urinary iAs, monomethylarsonic acid (MMA), and dimethylarsinic acid (DMA) was used as a biomarker of iAs exposure (∑As) (median of 4.92 μg/L). Children's occurrence of asthma, eczema, sneeze, wheeze, and medication for asthma and wheeze at each assessment time point (i.e., 4- and 7-year) was assessed with maternal interviewer-led questionnaires. Crude and adjusted Poisson regression models using Generalized Estimating Equation (GEE) were performed to account for the association between natural logarithm transformed (ln) urinary ∑As in μg/L at 4 years and repeated assessments of respiratory symptoms at 4 and 7 years of age. The covariates included in the models were child sex, maternal smoking status, maternal level of education, sub-cohort, and children's consumption of vegetables, fruits, and fish/seafood. The GEE-splines function using Poisson regression showed an increased trend of the overall expected counts of respiratory symptoms with high urinary ∑As. The adjusted expected counts (95% confidence intervals) at ln-transformed urinary ∑As 1.57 (average concentration) and 4.00 (99th percentile concentration) were 0.63 (0.36, 1.10) and 1.33 (0.61, 2.89), respectively. These exploratory findings suggest that even relatively low-iAs exposure levels, relevant to the Spanish and other populations, may relate to an increased number of respiratory symptoms during childhood.