Persistence and Bioaccumulation Potential

The information below was considered in evaluating whether PFOA meets the criteria for persistence and bioaccumulation as defined under the Persistence and Bioaccumulation Regulations of CEPA 1999 (Canada 2000a). Persistence criteria are half-lives of greater than or equal to 2, 182, 365 or 182 days for air, water, sediment and soil, respectively. Bioaccumulation criteria are bioaccumulation factors (BAFs) or bioconcentration factors (BCFs) of greater than or equal to 5000 or a log octanol–water partition coefficient (K_ow) of greater than or equal to 5.0.

Available data indicate that PFOA does not significantly photodegrade under relevant environmental conditions (Todd 1979; Nubbeet al. 1995; Scrano et al. 1999; Hatfield 2001; Hori et al. 2004, 2005, 2008), hydrolyze (Ellis et al. 2004b) nor undergo abiotic or biotic degradation (Reiner 1978; 3M Company 1979, 1980b, 1985b; Pace Analytical 1997; Oakeset al.2004; Moriwaki et al. 2005; Cheng et al. 2008). In summary, PFOA has been shown to be persistent (Table 4), with studies indicating no abiotic or biotic degradation in the environment under relevant environmental conditions.

Table 4. Summary of Persistence Data

Half-life values for PFOA and its salts are estimated and remain highly speculative owing to the short study periods. The atmospheric lifetime of PFOA has been predicted to be 130 days (Hurley et al. 2004). Franklin (2002) calculated an atmospheric lifetime of PFOA to be in the order of days when it was emitted from a ground source, and therefore likely not subject to long-range transport. However, if PFOA is produced from an atmospheric source (i.e., via precursors) and if the major loss mechanism is wet or dry deposition, then it may have a lifetime of 20–30 days before deposition (Ellis et al. 2004b). This would be sufficient time to allow transport over many thousands of kilometres, implying a long-range transport mechanism. Thepresence of PFOA in the Canadian Arctic also may provide evidence for the long-range transport of either PFOA (e.g., via ocean currents) (Caliebe et al. 2004; Yamashita et al. 2005) or volatile precursors to PFOA through the atmosphere (Stock et al. 2007). A suggested hypothesis for the presence of PFOA in biota in remote regions is that a precursor (e.g., FTOHs) is emitted to the atmosphere and ultimately degrades to yield PFOA through biotic and abiotic degradation. Ellis et al. (2004a) showed that the atmospheric lifetime of short-chain FTOHs, as determined by their reaction with hydroxyl radicals, was approximately 20 days. Piekarz et al. (2007) estimated that atmospheric residence times of 6:2 FTOH, 8:2 FTOH and 10:2 FTOH were 50, 80 and 70 days, respectively.

Shoeib et al. (2006) collected air samples during a crossing of the North Atlantic and the Canadian Arctic Archipelago in July 2005 to investigate concentrations of FTOHs. The highest concentrations of FTOHs were for 8:2 FTOH (5.8–26 pg/m³), followed by 10:2 FTOH (1.9–17 pg/m³) and then 6:2 FTOH (ND–6.0 pg/m³). Ju et al. (2008) measured PFOA in the sea microlayer (air/water interface) and subsurface seawater near the Dalian coastal waters in China. PFOA concentrations in the sea microlayer ranged from 0.26 to 1.19 ng/L. In the subsurface water, PFOA concentrations ranged from 0.17 to 0.67 ng/L. The detection of PFOA in oceanic waters suggests another potential mechanism for its long-range transport to remote locations such as the Canadian Arctic.

Mass calculations for marine transfer of PFO to the Arctic resulted in a flux between 2 and 12 tonnes per year (Prevedouros et al. 2006). Armitage et al. (2006) estimated a net PFO flux of between 8 and 23 tonnes per year to the Arctic. However, as stated previously, based on the modelling assumptions made regarding the pK_a, pH and PFOA/PFO speciation, partitioning processes related to the neutral acid may be ignored, and this may result in an underestimation of the concentrations of PFOA (Burns et al. 2008; Goss 2008). PFOA was measured in polar ice caps from three areas in the High Arctic (Melville ice cap, Northwest Territories; Agassiz ice cap, Nunavut; and Devon ice cap, Nunavut) (Young et al. 2007). PFOA concentrations ranged from 0.012 to 0.147 ng/L, suggesting that contamination may be a result of atmospheric input. Between 1996 and 2005, there was no significant trend of PFOA concentrations (regression analysis, p = 0.140) (Young et al. 2007). Fluxes were calculated using the density corrected concentration, multiplied by the yearly accumulation. Fluxes calculated to each of the ice caps were multiplied by the area of the Arctic to yield a flux of PFOA to the area north of 65°N. These fluxes are estimates and may not be representative of actual deposition in this region due to wide variations in precipitation rates. PFOA showed a flux ranging between 114–587 kg/year in 2005 (Young et al. 2007).

The log K_ow has been calculated or modelled for PFOA and its various salts. A modelled value has been determined to be 5 ± 0.5 (Jasinski et al. 2009). However, the K_ow is a problematic parameter for ionized surfactants because of their tendency to aggregate at the interface of a liquid–liquid system. Therefore, the K_ow should not be used to model bioaccumulation of PFAs. Instead, more emphasis should be placed on the results of experimental studies. In most PFOA bioaccumulation studies examined, BCFs and BAFs are below the “5000” bioaccumulation criterion stipulated in the Persistence and Bioaccumulation Regulations of CEPA 1999 (Canada 2000a). However, the criteria upon which these regulations are based were developed over 10 years ago, as outlined in the 1995 Toxic Substances Management Policy (Canada 1995a) and the Toxic Substances Management Policy: Persistence and Bioaccumulation Criteria (Canada 1995b). The criteria focused specifically on aquatic organisms (fish) and on neutral organic compounds, which, unlike perfluorinated substances, preferentially partition to lipids. PFOA is an ionized substance that primarily binds to proteins (Han et al. 2004) and preferentially partitions to liver and blood and kidney tissues. The numeric criteria for bioaccumulation, outlined in the Persistence and Bioaccumulation Regulations of CEPA 1999, are based on bioaccumulation data for aquatic species (fish) only and for substances that preferentially partition to lipids. As a result, the criteria do not account for the bioaccumulation of PFOA that is preferentially partitioning in the proteins of liver, blood and kidney in terrestrial and marine mammals. Notably, the Toxic Substances Management Policy (Canada 1995a) states that expert opinion and a weight of evidence approach are also very important when considering how to interpret and apply the criteria.

PFOA is absorbed in juvenile rainbow trout (Oncorhynchus mykiss), with a BCF of 4.0 (Martin et al. 2004b). Dietary exposure to PFOA in the same species did not result in PFOA bioaccumulation (BAF = 0.038) (Martin et al. 2004b). Carp (Cyprinus carpio) were exposed to two concentrations of PFOA (5 and 50 µg/L) for 28 days in a flow-through system (Kurume Laboratory 2001). The chemical was prepared for exposure by using a dispersant made up of hydrogenated castor oil and mixed in acetone. Water quality was monitored daily for the duration of the experiment. BCFs ranged from 3.1 at the low concentration to <5.1–9.1 at the high concentration, indicating low bioaccumulation (Kurume Laboratory 2001). Although experiments with fish and other aquatic species provide evidence that PFOA is not highly bioaccumulative, these results should not be extrapolated to other animals. Fish gills may provide an additional mode of elimination and uptake that air-breathing organisms such as birds, terrestrial organisms and marine mammals do not possess (Kelly et al. 2004). The high water solubility of PFOA causes its tendency to escape from the gills into water to be relatively high, whereas the tendency of PFOA to escape to air across the alveolar membrane of the lung would be relatively low because of its low vapour pressure and negative charge. For example, the half-life of PFOA in fish was 3 days (Martin et al. 2003a), whereas the half-life in male rats was 11 days (Ylinen and Auriola 1990); in humans, it was 4.37 years (KudoandKawashima 2003) and 3.8 years (Olsen et al. 2007).

Two species of wild turtles (red-eared slider, Trachemys scripta elegans; and Reeves’ turtle, Chinemys reevesii) were examined for BCFs. In Japan, these turtles occupy the highest trophic level of the food chain in the river ecological system and have small territories (Morikawa et al. 2006). PFOA concentrations in surface water ranged from 16.7 to 87 100 ng/L, and PFOA was observed in almost all serum samples (91 of 94), where concentrations ranged from <200 to 870 000 ng/L. The calculated BCFs ranged from 0.8 to 15.8 (Morikawa et al. 2006). The authors noted that the reported BCFsdecreased as PFOA concentrations in surface water increased, suggesting that the absorption of PFOA from the gut might be a saturable process. However, it should be noted that these BCFs are actually BAFs, as the wild turtles’ exposure to PFOA was probably not limited to surface water only.

Kannan et al. (2005a) measured PFOA concentrations in a benthic food chain of the Great Lakes and found that despite relatively high concentrations of PFOA in water, it was not detected in invertebrates or fish. Preliminary (unpublished) results from one study showed a biomagnification factor (BMF) of 8 from ringed seal (Pusa hispida) liver to polar bear (Ursus maritimus) liver (Butt and Smithwick 2004). In the pelagic aquatic food web of Lake Ontario, PFOA concentrations did not increase with increasing trophic level (as determined by stable isotopes of nitrogen) (Martin et al. 2004b). In fact, PFOA concentrations were lower in the top-predator lake trout (Salvelinus namaycush) than in the invertebrate opossum shrimp (Mysis relicta) (Table 5). For example, a trophic magnification factor (TMF) for PFOA was calculated as 0.37 for the slimy sculpin (Cottus cognatus)–burrowing amphipod (Diporeia hoyi)relationship; for the upper end of the food web (Mysis relicta–alewife [Alosa pseudoharengus]–rainbow smelt [Osmerus mordax]–lake trout), the calculated TMF was 0.58 (Martin et al. 2004b). The different ends of the aquatic food web could indicate differences in benthic and pelagic natures of the relationships. TMFs less than 1 indicate no biomagnification. In another study, the trophic level–corrected BMFs were calculated for several Arctic biota (Tomy et al. 2004). Tomy et al. (2004) determined TMFs for the entire food web based on the relationship between d¹⁵ N and contaminant concentration. Trophic level was determined relative to the clam, which was assumed to have a trophic level of 2 (i.e., primary herbivore). For each individual sample of zooplankton, fish and marine mammal, trophic level was determined using the following relationship:

where TL consumer is the trophic level of the organism and 3.8 is the isotopic enrichment factor. The second method determined biomagnification factors (BMF_TL) for individual species corrected for trophic level:

where [predator] and [prey] are the wet weight (ww) concentrations of analyte in the predator and prey species, respectively, and TL is the trophic level based on d¹⁵ N for the predator and prey. The resulting BMFs for PFOA were often greater than 1, suggesting a potential for PFOA to biomagnify for the clam (Mya truncata; Serripes groenlandica)–walrus (Odobenus rosmarus), Arctic cod (Boreogadus saida)–narwhal (Monodon monoceros) and cod–beluga (Delphinapterus leucas) food chain (Table 5). These results may reflect food web differences and perhaps not the bioaccumulation potential for PFOA (e.g., one food web had fish as a top predator and the other had a mammal).

Martin et al.(2004a) found that polar bears, which occupy the highest trophic level in the Canadian Arctic, have higher levels of PFOA than all other Arctic organisms examined. Butt et al. (2008) determined regionally-based ringed seal/polar bear BMF values for PFOA that ranged from 45 – 125. These regionally based BMFs were calculated by grouping ringed seal populations to corresponding similarly located polar bear populations in the Canadian Arctic. Tomy et al. (2009) determined PFOA TMFs for a marine food web in the western Canadian Arctic (Hendrickson Island and Holman Island) comprising of the Beaufort Sea beluga whale (Delphinapterus leucas), ringed seal (Phoca hispida), Arctic cod (Boreogadus saida), Pacific herring (Clupea pallasi), Arctic cisco (Coregonus autumnalis), a pelagic amphipod (Themisto libellula), and a Arctic copepod (Calanus hyperboreus). TMFs ranged from 0.1 (ringed seal/Arctic cod) to 2.2 (Arctic cod/Calanus hyperboreus).

Houde et al. (2005) reported BMFs of 1.8–13 using whole-prey homogenates and whole-body bottlenose dolphin (Tursiops truncatus) conjugate base concentrations in the bottlenose dolphin food web at Charleston, South Carolina. Kelly et al. (2009) found a TMF of 3.28 over the Canadian Arctic (Hudson Bay region) marine food web (macroalgae, bivalves, fish, seaduck and beluga whale). van den Heuvel-Greve et al. (2009) found BMFs of 3.8 and 23 for the benthic and the pelagic food webs with Harbour seals (Phoca vitulina) as the apex predator, respectively. BMFs for other species in the Westerschelde, Netherlands estuary ranged from 0.03 to 31.

Table 5. Summary of Bioaccumulation Data

Bioaccumulation (BCFs, BAFs, BMFs) may indicate either direct toxicity in organisms that have accumulated PFOA or indirect toxicity in organisms that consume prey containing PFOA (via food chain transfer). The minimum concentration of a substance in an organism that will cause an adverse effect (the critical body burden) is used to determine the potential to cause direct toxicity. From a physiological perspective, it is the concentration of a substance at the site of toxic action within the organism that determines whether a response is observed, regardless of the external concentration. In the case of PFOA, the site of toxic action is often considered to be the liver. However, when the potential for toxicity in consumer organisms is being determined, it is the concentration in the whole body of a prey item that is of interest, since the prey is often completely consumed by the predator (including individual tissues and organs, such as the liver and blood). Conder et al. (2008) noted that the concentration of perfluorinated acids on a whole body mass basis has been estimated to be 2–10 times lower than the concentrations of perfluorinated acids in blood and liver of trout. Although, the bioaccumulation potential of PFOA may be low in fish, the presence of detectable concentrations in higher trophic levels (polar bear, caribou, walrus) has generated concerns regarding the biomagnification potential of PFCAs, including PFOA, in food webs (Conder et al. 2008). Since perfluorinated substances partition to liver and blood, most field measurements for these substances have been performed on those individual organs and tissues. This is especially true for organisms at the higher trophic levels (e.g., polar bear), where whole-body analysis is not feasible for ethical reasons and due to the challenging logistics with respect to sampling and laboratory constraints. While it is feasible to measure whole-body BAFs on smaller species at lower trophic levels, the lower trophic status of the organism means that the estimated overall BAFs for perfluorinated substances may be underestimated. Thus, from a toxicological perspective, BCFs, BAFs and BMFs based on concentrations in individual organs, such as the liver, may be more relevant when the potential for direct organ-specific toxicity (i.e., liver toxicity) is being predicted. BCFs and particularly BMFs based on concentrations in whole organisms may provide a useful measure of overall potential for transfer up the food chain. As such, it is considered that PFOA may not be bioaccumulative in aquatic species but may be considered to bioaccumulate and biomagnify in terrestrial and marine mammals.