Consumption of low-moderate level arsenic contaminated water does not increase spontaneous pregnancy loss: a case control study

Background: Previous work suggests an increased risk for spontaneous pregnancy loss linked to high levels of inorganic arsenic (iAs) in drinking water sources (> 10 mu g/L). However, there has been little focus to date on the impact of low-moderate levels of iAs in drinking water (< 10 mu g/L). To address this data gap we conducted a hospital-based case-control study in Timis County, Romania. Methods: We recruited women with incident spontaneous pregnancy loss of 5-20 weeks completed gestation as cases (n = 150), and women with ongoing pregnancies matched by gestational age (+/- 1 week) as controls (n = 150). Participants completed a physician-administered questionnaire and we collected water samples from residential drinking sources. We reconstructed residential drinking water exposure histories using questionnaire data weighted by iAs determined using hydride generation-atomic absorption spectrometry (HG-AAS). Logistic regression models were used to generate odds ratios (OR) and 95% confidence intervals (CI) for associations between iAs exposure and loss, conditioned on gestational age and adjusted for maternal age, cigarette smoking, education and prenatal vitamin use. We explored potential interactions in a second set of models. Results: Drinking water arsenic concentrations ranged from 0.0 to 175.1 mu g/L, with median 0.4 mu g/L and 90th%tile 9.4 mu g/L. There were no statistically significant associations between loss and average or peak drinking water iAs concentrations (OR 0.98, 95% CI 0.96-1.01), or for daily iAs intake (OR 1.00, 95% CI 0.98-1.02). We detected modest evidence for an interaction between average iAs concentration and cigarette smoking during pregnancy (P = 0.057) and for daily iAs exposure and prenatal vitamin use (P = 0.085). Conclusions: These results suggest no increased risk for spontaneous pregnancy loss in association with low to moderate level drinking water iAs exposure. Though imprecise, our data also raise the possibility for increased risk among cigarette smokers. Given the low exposures overall, these data should reassure pregnant women and policy makers with regard to the potential effect of drinking water iAs on early pregnancy, though a larger more definitive study to investigate the potential risk increase in conjunction with cigarette smoking is merited.