Ecological screening assessment report on perfluorooctane sulfonate, salts and precursors: chapter 5


5. Risk Quotient Analyses

Risk quotient analyses, integrating known or potential exposures with known or potential adverse environmental effects, were performed for PFOS. An analysis of exposure pathways and subsequent identification of sensitive receptors were used to select environmental assessment endpoints (e.g., reduced body weight gain, increased offspring mortality, reductions in development, adverse histopathological effects in mammals etc.). For each endpoint, an EEV was selected based on empirical data from monitoring studies. Monitoring data from the Canadian environment were used preferentially for EEVs, but data from US or other countries were considered to supplement available Canadian data and as an indication of potential exposure to this persistent and bioaccumulative substance. EEVs usually represented protective scenarios, as an indication of the potential for these substances to reach concentrations of concern and to identify areas where those concerns would be most likely. An Estimated No-Effects Value (ENEV) was determined by dividing a CTV by an application factor. CTVs typically represented the lowest ecotoxicity value from an available and acceptable data set. Given the physical-chemical nature of PFOS and its salts, preference was generally given for chronic toxicity data, as long-term exposure was a concern. Where these data were not available, acute toxicity data were used. Application factors were derived using a multiplicative approach, which uses 10-fold factors (unless case-specific factors can be estimated) to account for various sources of uncertainty associated with making extrapolations and inferences related to the following: intra- and interspecies variations, differentially sensitive biological endpoints; laboratory to field impact extrapolation, extrapolation from single-species tests to ecosystems and extrapolation from low effect level to chronic no effect level.

Although risk quotients may be used to indicate potential to cause environmental harm for persistent and bioaccumulative substances, risks are likely to be underestimated using traditional quotient approaches. For example, if releases of a persistent substance have continued or increased in recent years, but maximum steady state concentrations have not yet been achieved in the environment, measured EEVs may underestimate possible exposure. In addition, ENEVs may underestimate potential for long term impacts of persistent and bioaccumulative substances since maximum concentrations are often not reached in the tissues of laboratory organisms, because toxicity test durations are insufficient to achieve steady state. Risk quotients derived for PFOS and its precursors are summarized in Appendices Appendix 33, Appendix 44, and Appendix 55.

Mammalian Wildlife

In Canada, the highest mean PFOS concentrations in wildlife were reported in a study of polar bears from 7 circumpolar locations. The highest Canadian concentrations were found in polar bear from South Hudson Bay (range 2000-3770 µg.kg-1 ww liver, mean 2730 µg.kg-1 ww liver) (Smithwick et al. 2005b). Concentrations in Canadian Arctic polar bear are among the highest in polar bears worldwide but the exposure concentrations are not considered an anomaly given similar concentrations in polar bears in other North America and European Arctic locations and higher concentrations in other wildlife globally (e.g., fish in Japan and the Netherlands). Given the relatively small sample size, which suggests further sampling could identify higher concentrations, and the fact that the species is a top level predator, the maximum exposure concentration in Canadian polar bear liver was considered appropriate for use in the risk quotient calculation.

In the assessment of risk to Canadian wildlife, the exposure concentration of 3770 µg.kg-1 ww liver from the Canadian South Hudson Bay polar bear was used as the EEV for wildlife. The CTV for mammalian wildlife was selected from a 2-year dietary rat study in which histopathological effects in the liver were seen in males and females at intakes as low as 0.06-0.23 mg.kg-1 bw per day and 0.07-0.21 mg.kg-1 bw per day, respectively (Covance Laboratories Inc. 2002). Average values were determined for males and females, to establish corresponding LOELs of 40.8 µg.g-1 in liver with an application factor of 100 to give an ENEV of 0.408 ug.g-1 liver. A risk quotient of 9.2 was, therefore, calculated using the maximum exposure concentration of 3770 µg.kg-1 ww liver from the South Hudson polar bear (Appendix 5). However, other risk quotients were also calculated given the range of toxicological endpoints but with the same maximum exposure concentration of 3770 µg.kg-1 ww liver from South Hudson Bay polar bear. These risk quotients were consistently above 1 ranging from 2.1 to 19 (Appendix 4). While the maximum concentration of PFOS in South Hudson polar bears is 3770 ug.kg-1 ww liver and risk quotients are calculated using this value, the mean concentration from this population is 2730 ug.kg-1 ww liver and risk quotients would, therefore, be approximately 27 % lower but within the same order of magnitude as risk quotients calculated using the maximum concentration. It is also noted that mean PFOS concentration in polar bear liver from 3 other Canadian locations (High Arctic, Northwest Territories, and South Baffin Island) ranged from 1170 - 1390 µg.kg-1. Risk quotients derived using the 2 year rat CTV results and mean concentrations resulted in quotients ranging from 2.9 to 3.4.

It is noted that the maximum concentration in East Greenland polar bear was higher (6340 µg.kg-1) than the South Hudson Bay bears and a risk quotient calculated from this value using the same application factors would yield quotient of 15.4. Using the highest tissue concentration (4870 µg.kg-1 liver) found in mink in the Midwestern United States would yield a risk quotient (11.9) of the same order of magnitude, which could also be considered relevant to Canadian wildlife in mid-latitudes. The risk quotient analysis indicates that the greatest potential risk from PFOS in the environment occurs in higher trophic level mammals.

Pelagic Organisms

A recent study by Macdonald et al. (2004) reported a 10 day NOEC of 0.0491 mg.L-1 for the growth and survival of the aquatic midge (Chironomus tentans). As such, the 10-day NOEC from the MacDonald et al (2004) study was chosen as the most appropriate CTV. An application factor of 10 was applied to account for lab to field variations and an application of 10 was applied to convert an acute endpoint to a chronic endpoint resulting in an ENEV of 0.491 µg.L-1. The EEV chosen for Canadian waters is the highest value measured in the Boulanger et al. (2004) study (121 ng.L-1 measured in Lake Ontario). The risk quotient is calculated as follows: 0.121 / 0.491 = 0.25 (Appendix 5).

Avian

For avian species, the CTV is based on the effects observed for male mallards in the 10 ppm (feed concentration) treatment group based on 21 weeks of exposure. At this dose, the level of PFOS in serum and liver were 87.3 µg.mL-1 and 60.9 µg.g-1 ww liver, respectively. Given the uncertainty in using this value (in the absence of a NOAEL) and that the effects in males (reduced testis size and effects on spermatogenesis) may have occurred before the end of study when the PFOS liver concentration was measured, an application factor of 10 is used to account for laboratory to field extrapolation and interspecies variability and an additional application factor of 10 is used to extrapolate from the observed effect level to a NOAEL. Therefore, the estimated no-effect value (ENEV) for PFOS in birds is 0.87 µg.mL-1 serum and 0.609 µg.g-1 liver. Risk quotients using these ENEVs are compared to EEV for a number of avian species that are native to Canada, including many piscivorous birds and migratory species (see Appendix 3). The range of risk quotients are either above or approaching one which indicate potential for harm at concentrations observed in native species, including migratory species.

Page details

Date modified: