Letter to the Editor on Bil et al. 2021 “Risk Assessment of Per‐ and Polyfluoroalkyl Substance Mixtures: A Relative Potency Factor Approach”

Response to Letter to the Editor on Bil et al. 2021 “Risk Assessment of Per‐ and Polyfluoroalkyl Substance Mixtures: A Relative Potency Factor Approach”

“…On the contrary, we are of the opinion that it is important to take into account combined exposure to PFAS as much as possible at this very moment. We respond to the points raised by Rietjens et al (2022), and provide a rationale to substantiate why the use of RPFs in mixture risk assessment of PFAS is justified.…”

“…In the first and second points of their letter, Rietjens et al (2022) argue that differences in experimental setup may influence the RPFs, and it is noted that sorption behavior should be taken into account when addressing the risks of PFAS mixtures. In Bil et al (2021), we recognize that "the influence of other external factors such as interlaboratory variation may not be ruled out."…”

“…A meaningful comparison is feasible when both analyses follow the basic principles of dose-response analysis. However, the dose-response analysis provided by Rietjens et al (2022) does not comply with several important starting points for dose-response analysis of continuous data that are stipulated in authoritative guidances (European Food Safety Authority [EFSA], 2017; US Environmental Protection Agency, 2012; World Health Organization [WHO], 2021):…”

“…This is a serious flaw in doseresponse assessment, because the variation in the data is not taken into account. Consequently, it is difficult to assess how robust the RPFs by Rietjens et al (2022) are. 2.…”

“…A significance test for parallel curves indicated by homogenous model parameter d is highly sensitive to dose-group level errors that may occur in toxicological data, and therefore this test outcome is not directly interpretable (Slob & Setzer, 2014). In the study of Goodrum et al (2021), quoted by Rietjens et al (2022), the influence of dose-group errors is also overlooked. Visually, dose-response curves with one and the same value for parameter d provide a good description of the assessed dose-response data (Figure 2 in the present letter and Figure 1 in Bil et al, 2021).…”