Risk assessment in recreational water

HUGO R. POMA; VERÓNICA B. RAJAL

Salto

Simposio; II Simposio Latinoamericano de Virología Ambiental; 2013

The aims of this work were to characterize the water quality of Arenales River, in order to identify changes in time and space and thus design a better management strategy in the future and to carry out a quantitative risk assessment of acquiring an infectious disease caused by enteric viruses in recreational waters in the province of Salta. Incidences of diarrheal and parasitic disease in the human population were analyzed in the main city of the Province of Salta using Geographic Information Systems, and the results suggested a potential influence of Arenales River. To establish the critical points for systematic monitoring, the river section that crosses the city was assessed and eleven locations (five on the river and six on discharges) were selected along the river. Physicochemical (temperature, pH, dissolved oxygen, conductivity, salinity, and turbidity, using a multiparametric probe) and microbiological variables including bacteria (total and thermotolerant coliforms, E. coli, and Enterococcus, by colony forming units count), parasites (by microscopy), and viruses (adenovirus [AdV], enterovirus [EV], and norovirus [NoV] by real-time PCR) were analyzed monthly for thirteen months. As the amount of viruses and parasites in environmental samples is usually low we used ultrafiltration to concentrate pathogens from 20 L of water to 50-70 mL. All samples were spiked with a known amount of the bacteriophage PP7 as internal control of the process to evaluate the filtration recovery, to calculate the sample limit of detection, and to assess enzymatic inhibition during PCR. Multivariate statistical techniques were used to evaluate correlations between physicochemical variables, microbial indicators and the pathogens detected, and to design a rational monitoring scheme for the optimum use of limited resources. Finally, quantitative microbial risk assessment (QMRA) was conducted using the densities of enteric viruses obtained from the river to calculate the risk of contracting an infectious disease caused by these microorganisms in water. Three scenarios were considered: 1) ingestion by bathing adults (BA), 2) ingestion by bathing children (BC), and 3) inhalation due to secondary contact (SC). Turbidity, dissolved oxygen, and conductivity were the physicochemical variables that were the most impacted by the seasonal effect. During the rainy season (summer, from November to April) the turbidity increased remarkably as a result of the turbulence of the river, carrying sediments from the bed. During the dry season (winter, May to October) dissolved oxygen decreased and conductivity increased mainly at the last two monitoring sites, showing the deterioration of water quality. Most of the samples analyzed exceeded the limits established by international guidelines for recreational waters for all the bacteria evaluated (Argentina does not have legislation for recreational waters). In particular, the last monitoring locations showed levels of total and thermotolerant coliforms similar to those in a raw wastewater discharge (identified at one of the locations). At least three different types of parasites were identified in each sample analyzed; AdV, EV, and NoV were detected in 45, 9, and 6% of the samples analyzed, respectively. The resulting large and complex data matrix was analyzed by multivariate analysis tools to understand the nature of contamination, to interpret the variability of data, and to reduce the number of monitoring sites and of variables to be measured, without significant loss of information. No significant correlations were found between indicator bacteria and viruses and parasites. However, AdV emerged as a strong candidate indicator of viral pollution while Microsporidium predicted the presence of parasites. For the QMRA, a lognormal distribution was found to fit better the experimental results obtained for AdV, while a gamma distribution fit the results from EV and NoV. Those distributions were used to evaluate the individual exposure rate and the dose for all the scenarios studied. An exponential dose-response model was adopted for AdV, while a Beta-Poisson distribution was used for EV and NoV. The calculated risks of AdV and NoV infection were high (especially for NoV due to the low infectious dose) and exceeded by far the limit accepted by the USEPA for gastroenteritis (0.8%), especially for primary contact recreation. In contrast, the average risk for EV infection was low for all scenarios studied (Table 1). We concluded that the Arenales river was highly contaminated with pathogens in the section studied. There was a considerable risk of adverse health effects caused by exposure to norovirus and adenovirus, while the risk was low to moderate for enterovirus. The tools and strategies developed in this work can be applied to any other water body for monitoring and environmental management of water resources and for epidemiological surveillance in public health. Site-specific QMRA is a promising tool to generate helpful information for public health management, given the great complexity and costs associated with epidemiological studies.