(a) Natural occurrence

PFOA is not known to occur naturally.

(b) Air

Although PFOA is not routinely monitored in air, sporadic measurements have been reported. Fromme et al. (2009) reviewed the literature and reported site mean concentrations of PFOA in air ranging from 1.4 to 552 pg/m³ from 11 rural and urban outdoor sampling sites in Japan, Canada, the United Kingdom, Norway, Ireland, and the USA; the highest measurements were from urban locations or adjacent to busy roads. PFOA and 8:2 FTOH have been found in remote Arctic areas far from known sources, suggesting long-range aerial transport. Concentrations of PFOA ranging from 0.012 to 0.147 ng/L were reported in polar ice caps in the High Arctic in 2006 (Environment Canada, 2012).

(c) Water

Samples from potable water supplies without known point sources of perfluorooctanoate contamination typically contain perfluorooctanoate at < 1 ng/L, or at levels below the detection limit (Fromme et al., 2009). However, higher concentrations in drinking-water have been reported for some locations. For example, Kim et al. (2011b) reported average concentration of perfluorooctanoate of 5.4 ng/L, and a maximum concentration of 33 ng/L, for 15 tap-water samples collected in 8 cities in the Republic of Korea. Surface water from Boulder basin of Lake Mead, the Hoover dam, and the lower Colorado River in the USA had average concentrations of perfluorooctanoate that were below the method reporting limit of 5 ng/L; however, samples affected by run-off from municipal wastewater treatment facilities had average concentrations of perfluorooctanoate ranging from 26 to 120 ng/L (Quiñones & Snyder, 2009).

Concentrations of perfluorooctanoate in water were measured in six public-water districts and for selected private wells in West Virginia, USA; these concentrations differed substantially by water district, varying by about three orders of magnitude (Fig. 1.2) (Shin et al., 2011a). Perfluorooctanoate has also been measured at concentrations exceeding 1 ng/L in many of more than 8000 samples of surface water and groundwater collected in the region surrounding a large fluoropolymer-production facility in West Virginia, USA, probably due to direct emissions to the Ohio River, the air, and long-term transport through the vadose zone (DuPont, 2010). The highest off-site environmental concentrations of PFOA were predicted to occur about 1 mile [1.6 km] away from the production facility, and average concentrations in drinking-water ranged from < 0.05 to 10.1 µg/L in 2002–2004 (Paustenbach et al., 2007).

Fig. 1.2

Measured and modelled concentrations of perfluorooctanoic acid (PFOA) in water for the six public water districts in the C8 Health Project/C8 Science Panel studies, USA

(d) Food

PFOA may be found in food due to contamination of plants and animals, and/or via transfer from food-packaging materials. Trudel et al. (2008) summarized several studies reporting measurements of PFOA in food in North America and Europe. Among the food categories, snacks and potatoes were reported to have the highest concentrations of PFOA (up to 3 ng/g wet weight), followed by packaged cereal products, meat, and North American fish/shellfish (up to 0.5, 1.0, and 2.0 ng/g, respectively). A list of measurements of PFOA concentrations in various foods is provided in Table 1.3.

Table 1.3

Concentrations of perfluorooctanoic acid (PFOA) in food and water.

(e) Dust

Trudel et al. (2008) and Fromme et al. (2009) reviewed the literature and estimated the concentrations of PFOA in dust in typical indoor environments as 100 ng/g and 19.72 ng/g, respectively.

Several studies suggested that the potential contribution of dust ingestion to exposure was higher than previously estimated. For example, one study in the USA reported a median concentration of PFOA in dust of 142 ng/g, and a 95th percentile of 1200 ng/g in dust collected at 102 homes and 10 day-care centres in Ohio and North Carolina, USA, in 2000–2001 (Strynar & Lindstrom, 2008). A study of 102 homes in Vancouver, Canada, reported median concentrations of 30 ng/g for PFOA in dust, 63 ng/g for 8:2 FTOH, and 1362 ng/g for the sum of polyfluoroalkyl phosphoric acid diesters containing at least one 8:2 polyfluoroalkyl group, suggesting that these potential precursors may contribute substantially to the body burden of PFOA if efficiently metabolized in the human body (De Silva et al., 2012).

(a) Serum concentrations

Human exposure to PFOA has often been assessed using measured or predicted concentrations of perfluorooctanoate in serum or plasma (Eriksen et al., 2009; Fromme et al., 2009; Bonefeld-Jorgensen et al., 2011; Barry et al., 2013; Vieira et al., 2013a; Hardell et al., 2014). The pharmacokinetics of PFOA differ widely between species, with short half-lives and strong sex differences in rats, but a half-life of 2.3–3.5 years and no observed sex differences in humans (Olsen et al., 2007; Bartell et al., 2010). The Canadian Health Measures Survey reported that the median and geometric mean plasma concentrations of perfluorooctanoate among Canadians aged 20–79 years in 2007–2009 were both 2.5 μg/L, and the 95th percentile was 5.5 μg/L (Environment Canada, 2012).

The California Environmental Contaminant Biomonitoring Program reported median serum measurements of perfluorooctanoate of 2.49 μg/L for 1337 teachers and school administrators in 2011–2014, and 0.474 μg/L for 77 pregnant women in 2010–2011 (California Department of Public Health, 2014). Yeung et al. (2013) reported a median serum PFOA concentration of 2.34 µg/L among 25 Australian liver donors in 2007–2009, noting a substantial decline in serum PFOA compared with previous reports of pooled Australian samples in 2002–2003 [7.6 μg/L] (Kärrman et al., 2006) and 2006–2007 [6.4 μg/L] (Toms et al., 2009). In a study of 413 pregnant and nursing women in Sweden, serum PFOA concentrations declined by an average of 3.1% per year (95% CI, 1.8–4.4%) from 1996–2010 (Glynn et al., 2012).

Geometric mean serum concentrations of perfluorooctanoate in the USA population based on serum measurements from the National Health and Nutrition Examination Survey were 5.2 μg/L, 3.9 μg/L, 3.9 μg/L, and 4.1 μg/L in 1999–2000, 2003–2004, 2005–2006, and 2007–2008, respectively, with similar concentrations in different age groups, but slightly higher concentrations in males than females (Calafat et al., 2007a; Kato et al., 2011). The 95th percentile of serum perfluorooctanoate concentrations did not exceed 12 μg/L in any of those years (Kato et al., 2011). Pooled samples from 3802 Australian residents in 2002–2003 yielded a mean perfluorooctanoate serum concentration of 7.6 μg/L (Kärrman et al., 2006) – a value roughly consistent with the geometric mean in the USA, considering that these measurements were positively skewed. Smaller studies of general populations in Europe, Asia, and the USA for samples collected in 1989–2006 have produced similar findings, with reported average concentrations ranging from 1.6 to 11.6 μg/L (Fromme et al., 2009).

Several studies of serum measurements of perfluorooctanoate are available for stored samples collected before the 1990s. Olsen et al. (2005) reported a median serum concentration of perfluorooctanoate of 2.3 μg/L for 178 blood samples collected in Maryland, USA, in 1974, and Harada et al. (2004) reported a geometric mean serum concentration of perfluorooctanoate of 0.2 μg/L for 39 blood samples collected from females in Miyagi, Japan, in 1977. Haug et al. (2009) reported serum concentrations of perfluorooctanoate in samples from a biobank of hospital patients in Norway, pooling samples by year (n > 19 for most years) for the period 1977–2006. Perfluorooctanoate serum concentrations in this study rose from 0.58 μg/L in 1977 to 1.3 μg/L in 1980, 3.3 μg/L in 1990, and 4.5 μg/L in 2000, falling to 2.7 μg/L in 2006.

Mean concentrations of perfluorooctanoate were measured in 258 samples of blood, serum, or plasma collected from men between 2000 and 2004 in the USA (Michigan) (5.7 μg/L; < 3–14.7 μg/L), Colombia (6.2 μg/L; 3.9–12.2 μg/L), Brazil (< 20 μg/L), Belgium (5.0 μg/L; 1.1–13 μg/L), Italy (< 3 μg/L), Poland (20.5 μg/L; 11–40 μg/L), India (3.5 μg/L; < 3–3.5 μg/L), Malaysia (< 10 μg/L), suggesting the presence of specific sources of PFOA in this country (Kannan et al., 2004). Relatively higher concentrations of PFOA were reported in the Republic of Korea (35.5 μg/L; < 15–71.4 μg/L) (Kannan et al., 2004).

Overall, the published data suggested that serum PFOA concentrations in the general population increased over time until about 2000, and have remained constant or decreased since that time.

Higher serum concentrations of perfluorooctanoate have been reported in general populations near production facilities and other known exposure sources. For example, the geometric mean serum concentration of perfluorooctanoate in 2005–2006 among 69030 residents living near a production facility in West Virginia, USA, was 32.9 μg/L (standard deviation, 241 μg/L). Exposures in that community varied substantially across six water districts; the mean serum concentration of PFOA was about 16 µg/L in the two water districts with the lowest water concentrations of PFOA, and 228 µg/L in the water district with the highest concentrations (Frisbee et al., 2009). A study of 641 residents of Arnsberg, Germany, in 2006 reported mean serum concentrations of perfluorooctanoate of 24.6, 26.7, and 28.5 μg/L in children, mothers, and men, respectively, due to surface water contamination from upstream agricultural use of soil conditioner mingled with industrial waste (Hölzer et al., 2008).

(b) Breast milk

PFOA has been measured in breast milk; these data are presented in Table 1.4.

Table 1.4

Concentrations of perfluorooctanoic acid (PFOA) in human breast milk.

In North Rhine-Westphalia, Germany, more than half of the samples analysed in 2007 (n = 203) contained PFOA; concentrations up to 610 ng/L have been reported (Bernsmann & Fürst, 2008). In China, PFOA was measured in 100% of the breast milk samples analysed (n = 19) in 2004 (So et al., 2006).

(c) Exposure sources

As PFOA and its precursors are not routinely or systematically monitored in air, water, dust, food, or drinking-water, the relative contributions of exposure sources in the general population are not well understood. Published studies of exposure have relied on synthesis of environmental measurements collected at varying times and places, often in different countries. These studies comprise the best available data, but are typically based on convenience samples in one or few locations, and may not be representative of regional, national, or global exposures.

One such study has estimated that diet (including transfer of PFOA from food packaging) contributes 99% of total exposure to PFOA for adults in the general population in “western” countries, with negligible contributions from inhalation and ingestion of house dust and drinking-water (Fromme et al., 2009). Estimated adult mean PFOA intakes via indoor air, outdoor air, house dust, diet, and drinking-water were 0.053, 0.076, 0.986, 169, and 1.3 ng/day, respectively. The estimated dietary contribution was based on PFOA measurements in a 7-day duplicate-diet study of 31 participants aged 16–45 years in Germany (Fromme et al., 2007); estimated inhalation contributions were based on PFOA measurements at one indoor and four outdoor sites in Europe (Barber et al., 2007; the estimated house-dust ingestion contribution was based on measurements of PFOA from 67 homes in Ottawa, Canada, in the winter (Kubwabo et al., 2005), and the estimated contribution of drinking-water ingestion was based on river-water samples from the Rhine and its tributaries in Germany (Skutlarek et al., 2006). The contribution of house dust to PFOA exposure was based on a conservative estimate of 5% conversion of 8:2 FTOH to PFOA. Other precursor concentrations may actually exceed those of PFOA and 8:2 FTOH in house dust, but the extent of precursor metabolism in humans is unclear (De Silva et al., 2012).

Trudel et al. (2008) estimated that ingestion of food and house dust contributed > 90% of exposure to PFOA in adults, noting that PFOA-treated carpets and ingestion of dust may account for a larger proportion of exposure among children than adults. Typical uptake doses of PFOA for infants, toddlers, children, and teenagers/adults were estimated at 9.8, 7.6, 5.0, and 2.5–3.1 ng/kg body weight (bw) per day in North America and 6.0, 7.6, 6.7, and 2.8–4.1 ng/kg bw per day in Europe, respectively, based primarily on food concentrations of PFOA from data extracted from four previous studies covering 17 food categories with 1–131 samples each (for most food categories, measurements from Europe and North America were combined due to small sample sizes) and house-dust concentrations of PFOA from data from three small studies in Canada, Japan, and the USA (Moriwaki et al., 2003; Costner et al., 2005; Kubwabo et al., 2005). Infants may be exposed primarily through mother’s milk (So et al., 2006; Kärrman et al., 2010; Kim et al., 2011a), for which the estimated perfluorooctanoate concentration was reported as 0.1 ng/g (Trudel et al., 2008).

However, drinking-water may have a larger contribution to exposure to PFOA in some populations. For example, Kim et al. (2011b) estimated that drinking-water ingestion contributes 30% of total exposure to PFOA in the Republic of Korea, where urban water supplies are often contaminated.

Drinking-water is thought to have been the predominant source of intake of PFOA for a highly exposed population near a production facility in West Virginia, USA, studied by the C8 Science Panel (Barry et al., 2013; Vieira et al., 2013a; Steenland et al., 2014), where both surface water and groundwater were contaminated by water and air emissions from the facility (Shin et al., 2011a).

Although residual amounts of PFOA (4–75 ng/g) and 8:2 FTOH are contained in non-stick cookware and can be released to the gas phase in small quantities when heated to normal cooking temperature, their contribution to exposure tends to decline with repeated use and is believed to be negligible compared with other exposure sources (Fromme et al., 2009). One study of four non-stick cookware items (with three samples each) reported emission rates of 19–287 and 42–625 pg/cm², for PFOA and 8:2 FTOH respectively, upon first heating after purchase; concentrations of gas-phase PFOA were shown to decrease after repeated use (four times) for some cookware brands, but not for others. PFOA may be off-gassed at different rates from non-stick coatings, depending on how the non-stick coating was prepared and applied (Sinclair et al., 2007).