Results
A total of 215 samples from the Cohort 1 were analyzed for 64 targeted PFAS and EOF, while 20 samples from Cohort 2 were analyzed for 25 targeted PFAS, fluoxetine, seproxetine, and EOF; fewer PFAS were analyzed in Cohort 2 due to limited amount of volume of the samples to reach lower detection limit of other PFAS. The samples from this cohort were only focused on some legacy PFAS and GenX. Among the 215 samples from Cohort 1, 143 from T1 and 72 were paired samples from the same individuals at T2, which were collected 2-3 weeks apart.
Concentrations of target PFAS and EOF
- Cohort 1- samples from the EuroMix study. Of the 64 PFAS analyzed, 20 showed detectable concentrations. Seven PFAS were detected in all 215 samples, with the highest median concentrations in the following order: TFA (6.75 ng/mL), PFOS (2.54 ng/mL), PFHxS (0.84 ng/mL), PFOA (0.75 ng/mL), PFNA (0.29 ng/mL), PFDA (0.11 ng/mL), and PFHpS (0.06 ng/mL). The reported PFAS concentrations were lower than those in an earlier study reporting PFAS in the same individuals.
- Cohort 2 - Samples were purchased commercially. These samples were first extracted at pH 4 and then pH 11; the extracts from pH 4 and pH 11 were analyzed using LC-MS separately. At pH 4, 14 of the 25 PFAS analysed had detectable concentrations. Seven PFAS showed detection frequency over 80% with the median concentrations in the following order: TFA (15.1 ng/mL), followed by PFOS (1.82 ng/mL), PFOA (0.67 ng/mL), PFHxS (0.51 ng/mL), PFNA (0.19 ng/mL), PFDA (0.08 ng/mL), and PFUnDA (0.03 ng/mL). No significant differences were observed between those taking medications and those who did not (Mann-Whitney U test, p>0.05), except for TFA. Detectable concentrations of fluoxetine and seproxetine were only observed in the subjects who had taken fluoxetine ranging from 0.03–6.95 ng/mL and <MDL – 0.48 ng/mL, respectively. On the other hand, at pH 11, no detectable concentrations of target PFAS were observed. However, at pH 11 higher concentrations of fluoxetine and seproxetine were observed in the subjects who had taken fluoxetine ranging 1.26–218 ng/mL and 16.0–92.5 ng/mL, respectively. Regarding EOF analysis, elevated concentrations of EOF were observed in the subjects where they had taken fluoxetine when compared to the control group (4 out of 10 subjects without taking fluoxetine had detectable EOF concentrations). At pH 4, the fluoxetine group showed the median of 140 ng F/mL versus the control group of <8 ng F/mL; significant higher EOF concentrations were observed in the fluoxetine group (Mann-Whitney U test, p<0.05, Figure 1). Two elevated EOF concentrations in the control group were due to significant levels of 6:2 FTSA and PFHxS in the samples, respectively. Detectable EOF concentrations were only observed in the fluoxetine group at pH 11 (median: 28.4 ng F/mL).
Mass balance analysis of extractable organofluorine (EOF)
Mass balance analysis of EOF revealed that the target PFAS accounted for varying proportions of EOF in the samples, ranging from 5.8 % to being fully accountable in the EuroMix samples, as well as in samples from subjects who either had or had not taken fluoxetine or other fluorinated drugs (Figure 2). An important finding from this investigation is that the contribution of target PFAS drops considerably if TFA is not included in the mass balance analysis from median of 62% to 29%, indicating that TFA is the most abundant PFAS in this study. In Cohort 2, the mass balance drops from 14% to 2% among those who had taken fluoxetine, and from 58% to 27% among those who had not taken fluoxetine when TFA is excluded in the mass balance analysis. For the samples treated at pH 11, the only detectable compounds were either fluoxetine or/and seproxetine from the subjects who had taken fluoxetine.
Figure 1. EOF concentrations (ng F/mL) in subjects without taking fluoxetine* and subjects taking fluoxetine of cohort 2. (*When EOF is below MDL, MDL was used; significant higher EOF concentration was observed in the fluoxetine group)
min | max | median | average | SD | detection frequency | |
PFHpA | 0.02 | 0.10 | 51 | |||
PFOA | 0.18 | 3.99 | 0.75 | 0.84 | 0.45 | 100 |
PFNA | 0.08 | 1.41 | 0.30 | 0.35 | 0.21 | 100 |
PFDA | 0.03 | 0.51 | 0.12 | 0.14 | 0.09 | 100 |
PFUnDA | 0.01 | 0.57 | 0.11 | 0.14 | 0.11 | 99 |
PFDoDA | 0.02 | 0.07 | 27 | |||
PFTrDA | 0.01 | 0.19 | 72 | |||
PFTDA | 0.02 | 0.03 | 6 | |||
PFBS | 0.010 | 0.03 | 8 | |||
PFHxS | 0.26 | 2.721 | 0.84 | 0.90 | 0.41 | 100 |
PFHpS | 0.01 | 0.37 | 0.06 | 0.08 | 0.06 | 100 |
L-PFOS | 0.30 | 8.36 | 1.65 | 2.09 | 1.61 | 100 |
BrPFOS | 0.17 | 5.13 | 0.89 | 1.25 | 0.95 | 100 |
6:2 FTSA | 0.01 | 0.12 | 2 | |||
8:2 FTSA | 0.01 | 0.08 | 46 | |||
TFA | 2.69 | 16.0 | 6.75 | 6.91 | 2.34 | 100 |
PFPrA | 0.18 | 0.59 | 22 | |||
FOSAA | 0.02 | 0.05 | 4 | |||
MeFOSAA | 0.01 | 0.22 | 22 | |||
EtFOSAA | 0.02 | 0.03 | 1 | |||
FOSA | 0.01 | 0.06 | 0.02 | 0.02 | 0.01 | 80 |
Table 1. Target* PFAS concentrations (ng/mL) and detection frequency (%) in samples from Cohort 1, the EuroMix study
*Reported concentrations are not recovery-corrected. Recovery-corrected concentrations, except for TFA and PFPrA, are provided in Thepaut et al. 2021.
Thépaut, E.; Dirven, H. a. a. M.; Haug, L. S.; Lindeman, B.; Poothong, S.; Andreassen, M.; Hjertholm, H.; Husøy, T. Per- and Polyfluoroalkyl Substances in Serum and Associations with Food Consumption and Use of Personal Care Products in the Norwegian Biomonitoring Study from the EU Project EuroMix. Environ. Res. 2021, 195, 110795. https://doi.org/10.1016/j.envres.2021.110795.
Figure 2. Percentage of target PFAS to EOF (%) in the sera samples
Note: When EOF was below MDL, value of MDL was used for EOF for calculating the percentage of PFAS to EOF. Since no TFA was detected in the fluoxetine group extracted at pH11, the graph is not shown.
Temporal changes of EOF and TFA in Cohort 1 between T1 and T2 (2–3 weeks between sample collections)
Table 2 summarizes the results of temporal analysis of EOF between paired samples collected with a 2 to 3 weeks time interval. Almost 50% of the samples did not show any significant changes (less than 25%) in EOF between T1 and T2; 31% showed a significant decrease with a maximum decrease of EOF of 54%; while only 5% showed significant increase with the maximum increase of 26%. In 15% of the cases, the EOF at T1 and T2 were below MDLs, for which we cannot tell if there would be any significant changes (25%) in EOF as undetermined. Eight pairs of samples showed 25% of increase of TFA concentration, while only 1 pair showed a decrease of 25% of TFA concentration. Only 3 pairs showed significant changes in both EOF and TFA.
Description | % of the pair samples (n=59) |
Undetermined (both values between T1 and T2 were below MDL) | 15 |
No difference (differences between T1 and T2 were below 25%) | 49 |
Increase (25% increases between T1 and T2) | 5 |
Decrease (min: 25% and max 54%) (25% decreases between T1 and T2) | 31 |
Table 2. Changes in EOF between T1 and T2 (sample collection 2–3 weeks apart)
Investigation on factors that may be associated with the proportion of unknown fluorinated chemicals in blood
No dietary habits or use of personal care products (p=0.139) were associated with the proportion of unknown fluorinated chemicals in samples T1. However, the proportion of unknown fluorinated chemicals in samples from T2 were associated with consumption of Potato/vegetables (p=0.0287).