Relatively few published studies have assessed the microbiological or chemical quality of bottled water in the USA. We purchased all single and multipackage bottled water sold in grocery and retail stores in one county in Virginia. 107 samples, across 26 brands, were tested for pH, conductivity, dissolved oxygen, total coliform, E. coli, nitrate, sulfate, free chlorine, total trihalomethanes (THMs, a component of disinfection byproducts), fluoride, and heavy metals. 64.5% (n = 69) of samples were purified water and 35.5% (n = 38) were spring water. Contaminant concentrations were higher overall in spring water samples. Total coliforms and E. coli were not detected in any samples. 1.4% (n = 1) of purified bottled water samples exceeded regulatory standards for THM (80 ppb), and 7.2% (n = 5) and 10.5% (n = 4) of purified and spring samples exceeded 1/2 the THM standard, respectively. No spring water samples exceeded standards for heavy metals, but one purified bottled water sample exceeded the standard for cadmium. We observed considerable within-brand variability in concentrations for a number of parameters including THMs, calcium, silicon, and strontium. This is the first study we are aware of to analyze contaminant variability within bottled water brands. Our results indicate that monitoring and reporting requirements for bottled water should be improved.

Abstract/Summary: We characterized concentrations of trihalomethanes (THMs), a measure of disinfection byproducts (DBPs), in tap water samples collected from households with utility-supplied water in two rural counties in Appalachian Virginia, and assessed associations with pH, free chlorine, and metal ions which can impact THM formation. Free chlorine concentrations in all samples (n = 27 homes) complied with EPA drinking water guidelines, though 7% (n = 2) of first draw samples and 11% (n = 3) of 5-min flushed-tap water samples exceeded the US Safe Drinking Water Act (SDWA) maximum contaminant level (MCL) for THM (80 ppb). Regression analyses showed that free chlorine and pH were positively associated with the formation of THM levels above SDWA MCLs (OR = 1.04, p = 0.97 and OR = 1.74, p = 0.79, respectively), while temperature was negatively associated (OR = 0.78, p = 0.38). Of the eight utilities serving study households, samples from water served by three different utilities exceeded the EPA MCL for THM. Overall, these findings do not indicate substantial exposures to DBPs for rural households with utility-supplied water in this region of southwest Virginia. However, given the observed variability in THM concentrations between and across utilities, and established adverse health impacts associated with chronic and acute DBP exposure, more research on DBPs in rural Central Appalachia is warranted.

Abstract/Summary: Consumption of unsafe drinking water is associated with a substantial burden of disease globally. In the US, ~1.8 million people in rural areas lack reliable access to safe drinking water. Our objective was to characterize and assess household-level water sources, water quality, and associated health outcomes in Central Appalachia. We collected survey data and water samples (tap, source, and bottled water) from consenting households in a small rural community without utility-supplied water in southwest Virginia. Water samples were analyzed for physicochemical parameters, total coliforms, E. coli, nitrate, sulfate, metals (e.g., arsenic, cadmium, lead), and 30+ enteric pathogens. Among the 69% (n = 9) of households that participated, all had piped well water, though 67% (n = 6) used bottled water as their primary drinking water source. Total coliforms were detected in water samples from 44.4% (n = 4) of homes, E. coli in one home, and enteric pathogens (Aeromonas, Campylobacter, Enterobacter) in 33% (n = 3) of homes. Tap water samples from 11% (n = 1) of homes exceeded the EPA MCL for nitrate, and 33% (n = 3) exceeded the EPA SMCL for iron. Among the 19 individuals residing in study households, reported diarrhea was 25% more likely in homes with measured E. coli and/or specific pathogens (risk ratio = 1.25, cluster-robust standard error = 1.64, p = 0.865). Although our sample size was small, our findings suggest that a considerable number of lower-income residents without utility-supplied water in rural areas of southwest Virginia may be exposed to microbiological and/or chemical contaminants in their water, and many, if not most, rely on bottled water as their primary source of drinking water.