Reaction mass of 4,4'-isopropylidenediphenol and phenol

Toxicological Summary

Workers - Hazard via inhalation route

Systemic effects

Long term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

2 mg/m³

Most sensitive endpoint:

repeated dose toxicity

Route of original study:

By inhalation

DNEL related information

DNEL derivation method:

other: Recent SCOEL Recommendation 2014

Modified dose descriptor starting point:

NOAEC

Value:

10 mg/m³

Explanation for the modification of the dose descriptor starting point:

Subchronic inhalation study

Acute/short term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

2 mg/m³

Most sensitive endpoint:

repeated dose toxicity

Route of original study:

By inhalation

DNEL related information

DNEL derivation method:

other: Recent SCOEL Recommendation 2014

DNEL extrapolated from long term DNEL

Local effects

Long term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

2 mg/m³

Most sensitive endpoint:

irritation (respiratory tract)

DNEL related information

DNEL derivation method:

other: Recent SCOEL Recommendation 2014

Acute/short term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

2 mg/m³

Most sensitive endpoint:

irritation (respiratory tract)

DNEL related information

DNEL derivation method:

other: Recent SCOEL Recommendation 2014

DNEL extrapolated from long term DNEL

Workers - Hazard via dermal route

Systemic effects

Long term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

66 µg/kg bw/day

Most sensitive endpoint:

repeated dose toxicity

Route of original study:

Oral

DNEL related information

DNEL derivation method:

ECHA REACH Guidance

Overall assessment factor (AF):

75

Modified dose descriptor starting point:

BMDL10

Value:

387 µg/kg bw/day

AF for dose response relationship:

1

Justification:

See discussion

AF for differences in duration of exposure:

1

Justification:

Generation studies up to chronic studies are available.

AF for interspecies differences (allometric scaling):

1

Justification:

Kinetic data are available in mice, rats and monkeys. Allometric scaling is considered based on AUC values and taken into account to drive the dose descriptor starting point. See discussion for detailed information.

AF for other interspecies differences:

2.5

Justification:

Default factor

AF for intraspecies differences:

5

Justification:

Default factor

AF for the quality of the whole database:

1

Justification:

See discussion

AF for remaining uncertainties:

6

Justification:

Factor applied by EFSA (2015) to derive oral TDI – see discussion for more details

Acute/short term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

66 µg/kg bw/day

Most sensitive endpoint:

repeated dose toxicity

Route of original study:

Oral

DNEL related information

DNEL derivation method:

ECHA REACH Guidance

DNEL extrapolated from long term DNEL

Local effects

Long term exposure

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

sensitisation (skin)

Acute/short term exposure

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

sensitisation (skin)

Workers - Hazard for the eyes

Local effects

Hazard assessment conclusion:

medium hazard (no threshold derived)

Additional information - workers

No mammalian toxicity information is available on the target substance itself, as its constituents phenol and BPA are well characterised and any testing of the reaction mass prohibits itself because of the toxicity of the constituents.

While the reaction mass consists of up to 25% (sum) impurities, no toxicological data are available to evaluate the hazard profile of the impurities. Due to the relatively high hazard posed by phenol, which drives strict exposure control measures, and the lower concentration of the impurities, it is highly unlikely that the impurities would have a relevant impact on the human health hazard profile or the work place risk management measures.

Please note that due to the extensive datasets available for BPA and phenol, the constituents of the reaction mass, we refrained from including all available data on both substances into this dataset, but used the key and weight of evidence information of phenol and BPA to document the hazard profile of the reaction mass.

Additivity of the effects of phenol and BPA is unlikely as both components have different toxicological profiles. Nethertheless, the effect levels derived for the components will not be adapted to the lower concentration in the reaction mass, but employed directly for hazard characterisation for the sake of simplicity.

The DNEL derivation for the reaction mass is

based on the IOEL for phenol set by the SCOEL in 2003, as no new data which would change the assessment by the SCOEL became available since then. The STEL of phenol was not adopted for the reaction mass, as the STEL of phenol (16 mg/m3) would exceed the OEL for BPA (10 mg/m3) for which no STEL was defined, as the OEL is based on respiratory irritation effects. Therefore, no separate inhalation short term DNEL was defined for the reaction mass, but the long-term DNEL will cover the endpoint.

 

 

Inhalation:

The systemic and local inhalation long-term DNEL are taken from SCOEL recommendation dated June 2014 (SCOEL/SUM/113). SCOEL did not define a STEL value. As a conservative approach short-term and long-term DNEL will be set at the same value; 2 mg/m3.

 

Dermal:

Local effects:

Bisphenol A is classified as Skin Sens. 1 according to Annex VI of Regulation (EC) No 1272/2008 and allocated the high hazard band. The EU-RAR update 2008 concluded that sensitisation is unlikely at lower concentrations but cannot be excluded at high concentrations:"Overall the new information does not confirm the previously reported evidence of a skin sensitisation potential of Bisphenol A. While the data do not exclude a skin sensitising activity of Bisphenol A at high concentrations (> 30%), there is no evidence that this is a concern for workers in current Bisphenol A manufacturing plants (such workers are believed to represent the group most likely to be exposed to Bisphenol A dust)."

 

Systemic effects:

Starting point for systemic dermal DNEL calculation:

A BMDL(10) of 8960 μg/kg/day was calculated by EFSA (2015) for changes in the mean relative kidney weight in a two generation toxicity study in mice. This value is taken by EFSA as a starting point for their TDI calculation and also taken as a starting point for the DNEL calculation.

 

Correction of the starting point:

Time extrapolation is taken into account because animals were dosed 7 days per week and workers are potentially exposed 5 days per week.

 

The corrected dermal starting point takes into account differences in absorption between oral studies in animals and dermal absorption in humans.

 

AUC mice orally dosed with 100 µg/kg:

EFSA (2015) calculated the AUC of adult mice orally dosed with 100 µg/kg to be 0.244 (nmol x h/l). There is a high uncertainty concerning the data reported in this particular study (Doerge et al., 2011). The authors reported that levels of unconjugated Bisphenol A that were above the detection limit were observed only at the earliest three time points, and only in one or two samples out of the twelve determinations at each time. EFSA calculated different scenarios and derived lower-bound (LB), middle-bound (MB), and upper-bound (UB) estimates of the AUC with values ranging from 0.244 – 1.257 (nmol x h/l).

 

Based on above mentioned limitations of this study an alternative approach was conducted to derive AUC values for mice applying species scaling.

 

In general AUC is correlated to systemic clearance (CL) via the following equation: AUC = dose/CL

CL is correlated to body weight (BW) for many compounds according to the following equation:

CL = a x BW^b

 

Studies with relevant Toxicokinetic data are summarised below:

Reference	Species	Dose (mg/kg)	AUCFree(0--‐∞) (nmol/hr/L)
Doerge et al., 2011. Toxicol. Lett. 207, 298–305	mouse	0.1	0.1
Doerge et al., 2010. Toxicol. Appl. Pharmacol. 247, 158–65	rat	0.1	2.6
Doerge et al., 2010. Toxicol. Appl. Pharmacol. 248, 1–11.	monkey	0.1	1.5
Gayrard et al., 2013. Environ. Health Perspect. 121, 951–6	dog	20	1577
Sieli et al., 2011. Environ. Health Perspect. 119, 1260–5	mouse	20	919
Tominaga et al., 2006. Toxicology 226, 208–17.	Rat     Monkey     chimpanzee	10 100   10 100   10	49 2431   189 1596   148
Taylor et al., 2011. Environ. Health Perspect. 119, 422–30	Mouse     monkey	0.4 100   0.4	170 1.3e4   59
Teeguarden et al., 2015. Toxicol. Appl. Pharmacol. 288, 131-142	human	0.03	2.5
Thayer et al., 2015. Environ. Int. 83, 107–115.	human	0.1	23

 

Fitting the power function equation to the AUC for unconjugated Bisphenol A in humans, monkeys, mice, chimpanzees, dogs and rats yields values of 35.6 for “a” and 0.92 for “b” with good agreement across species (r²= 0.95 excluding the mice data by Doerge et al., 2011). See Figure 1 and 2 in the attached documentation for more details. Based on this evaluation the predicted AUC for mice (dosed with 100 µg/kg) is 2.9 nmol x h/l.

 

This predicted values is taken to derive a corrected dermal starting point for DNEL derivation instead of the value derived by EFSA (2015) since 1) high uncertainty concerning the data reported in Doerge et al., 2011 is recognized by EFSA and 2) additional human data became available after the EFSA evaluation in 2015 (Teeguarden 2015, Thayer 2015) which fit the species scaling approach.

 

Calculated human dermal AUC (100 µg/kg):

PB-PK models for humans after dermal exposure are reported by Mielke et al (2011. Toxicology Letters. 204, 190-198) and Fischer et al (2011. Toxicology and Applied Pharmacology. 257. 122-136). Mielke et al reported an AUC of 697 (pg/ml x h) after dosing with 0.97 µg/kg/day (taking into account 100% absorption through the skin). Scaling this AUC to a dose of 100 µg/kg results in an AUC of 314 (nmol x h/l). Similar data are reported by Fisher (2011)The relationship between the two PBPK models for dermal AUCs is: AUCdermal,Fisher/Yang = 0.94 x AUCdermal,Mielke (see p. 585 of PART II in EFSA 2015).

A calculated dermal AUC value of 314 (nmol x h/l) is taken to derive a corrected dermal starting point for DNEL derivation.

 

Dermal absorption rate in humans.

To derive a dermal absorption rate an in vitro dermal absorption and metabolism study was conducted upon request by ECHA (Toner et al 2015). A summary of the mean results is provided in the following table:

 

Target Concentration (mg/L)	300	60	12	2.4
	(% Applied Dose)
Total Dislodgeable Dose	72.34±5.64	70.91±6.20	71.95±7.98	71.57±9.11
WholeStratum Corneum	10.25±5.44	9.25±4.30	7.31±3.33	7.70±4.92
Total Unabsorbed Dose	82.61±8.37	80.31±6.92	79.33±9.97	79.37±9.91
Epidermis	10.66±6.40	10.45±5.73	10.38±5.36	11.91±4.86
Dermis	3.28±2.44	3.97±1.99	6.19±4.28	4.51±3.73
Total Absorbed Dose	1.98±1.42	1.68±1.20	2.72±1.95	3.62±1.69
Dermal Delivery	15.92±8.14	16.10±7.01	19.28±8.54	20.04±6.24
Mass Balance	98.53±1.99	96.41±1.45	98.62±2.18	99.40±6.54

Total Dislodgeable Dose = skin wash + tissue swabs + pipette tips + donor chamber wash

Unabsorbed Dose = dislodgeable dose + wholestratum corneum (all tape strips+ unexposed skin

Absorbed Dose = receptor fluid + receptor chamber wash + receptor rinse

Dermal Delivery = epidermis + dermis + absorbed dose

Mass balance = dermal delivery + unabsorbed dose

 

Overall, the majority of the applied radioactivity was associated with epidermis samples (10.66%) compared to dermis (3.28%) and receptor fluid (1.98%) samples.

 

Applying the EFSA Guidance on Dermal Absorption (EFSA Journal 2012. 10, 2665) or SCCS Basic Criteria for the in vitro assessment of dermal absorption of cosmetic ingredients (SCCS/1358/1) the potentially bioavailable portion of Bisphenol A is ca. 30% over the whole dose range tested.

 

EFSA guidance (data provided as % applied dose)

Test Preparation	Mean Potentially Absorbable Dose	Mean+1SD	SD>25% of mean
300 mg/L	23.70	31.84	Yes
60 mg/L	23.09	29.13	Yes
12 mg/L	24.80	31.79	Yes
2.4 mg/L	25.42	30.79	No

Bold values denotes value to be used according to the EFSA Guidance. (For more detail see chapter Toxicokinetics).

 

SCCS criteria (data provided as % applied dose)

Test Preparation	Mean dermal delivery	Mean+1SD	Mean+2SD
300 mg/L	15.92	24.06	32.20
60 mg/L	16.10	23.11	30.12
12 mg/L	19.28	27.82	36.36
2.4 mg/L	20.04	26.28	31.59

No clear guidance is given on guideline to define if 1 or 2 SD should be used.

 

No metabolism was observed in any of the epidermis samples, however limited levels of metabolism were observed in dermis and receptor fluid samples (0-14%) with formation of Bisphenol A-glucuronide and Bisphenol A-sulfate identified in supernatant from incubation of viable skin disks for 24 h (<25%). Metabolites with retention consistent with Bisphenol A‑glucuronide and Bisphenol A-sulfate, and also more polar components, were identified. It might be assumed, but is not analytically verified, that these polar compounds are mixed sulfate/glucuronide bis‑conjugate Bisphenol A metabolites. It can be concluded qualitatively that fresh human skin has some in vitro metabolic capacity but further experiments may be necessary to optimize the experimental conditions to quantify that metabolism.

Overall,a potentially bioavailable portion of Bisphenol A of 30% is taken to derive a corrected dermal starting point for DNEL derivation. As a conservative approach no metabolism was taken into account. As outlined in the chapter "Toxicokinetics", taking into account the 10 % dermal penetration taken by EFSA for their dermal risk assessment based on the ca. 10% Bisphenol A found in the receptor fluid in the Demierre et al. (2012) study a significantly lower penetration value into the receptor fluid was observed in this in vitro dermal penetration study using viable skin (1.7 – 3.6 %). The approach mentioned in the study request by ECHA to consider the dermis, epidermis (with stratum corneum) and receptor fluid for further calculations lead to substantially higher potential penetration values compared to the EFSA 2015 evaluation of Bisphenol A.

 

Comparison of calculated AUC human dermal dosed with 100 µg/kg with experimental data indicate that the derived values are conservative.

Taking into account the above mentioned calculated AUC value of 314 (nmol x h/l) and 30% dermal absorption and no metabolism an internal AUC of 314 x 0.3 = 94.2 nmol x h/l can be calculated.

Recent information on a human toxicokinetic study is available (Thayer KA, Doerge DR, Hunt D, Schurman S, Twaddle NC, Churchwell MI, Garantziotis S, Kissling GE, Easterling MR, Bucher JR and Birnbaum LS (2014a) Pharmacokinetics of Bisphenol A in Humans Following Dermal Administration. Preliminary results. Submitted information via public consultation on the restriction proposal to ECHA. Comment reference number 1043. Public comments are available athttp://echa.europa.eu/restrictions-under-consideration/-/substance-rev/1894/term) 

The authors report data on a 3-day pilot study of 2 men and 2 women treated with 100 μg/kg bw of deuterated Bisphenol A (d-BPA) by dermal (forearm) administration in a suspension using 0.3% carboxymethylcellulose (CMC) as a vehicle housed under a Hill Top Chamber. These subjects also participated in a recently completed oral PK study using the administered dose level (Thayer et al., 2015. Environ. Int. 83, 107–115). The use of d-BPA allowed administered d-BPA to be distinguished from background native (unlabelled) Bisphenol A. The following results are reported: “The AUCs for total and aglycone d-BPA following dermal administration were compared in the same subjects after oral dosing with the same administered dose (100 μg/kg bw).The AUC following dermal administration was 0.002-1.2% of the AUC following oral administration for total d-BPA and 0.05-43% for aglycone d-BPA.“

 

Comparing the oral AUC of 23 nmol x hr/L reported for dosing with 100 µg/kg within the same volunteers by Thayer et al (2015) and the calculated dermal AUC of 94.2 nmol x h/l (adjusted to 100 µg/kg and taking into account 30% adsorption) it can be concluded that the calculated dermal AUC is higher than the oral AUC and, consequently, should be regarded as very conservative.

 

Calculation of corrected starting point:

Starting point: BMDL(10) =8 960 μg/kg/day

Time correction 7/5

AUC oral mice (100 µg/kg) = 2.98 nmol x h/l

AUCdermal, human (100 μg/kg) = 7.51 nmol x h/l
Corrected starting point = 8960 μg/kg bw/day x 7/5 x 2.98/7.51 = 4978 μg/kg bw/day

 

Assessment factors:

Interspecies differences (toxicokinetics): 1
Differences in toxicokinetics after oral dosing in mice and potential dermal human exposure is taken into account to derive the corrected starting point.

Interspecies differences (additional uncertainty): 2.5
2.5 is the REACH guidance default factor. This factor is also applied by EFSA (2015).

Intraspecies factor (worker): 5
5 is the REACH guidance default factor. This factor should be considered conservative based a recent PB-PK model reported by Yang et al., (2015. Toxicol. Appl. Pharmacol.). The authors conclude concerning the interspecies factor in the general population:“In this study, the recalibrated human Bisphenol A PBPK model was used to estimate the inter-individual variability of internal dose metrics of Bisphenol A for the general population based on the estimated daily intake of Bisphenol A in the United States (FDA 2014b; Lakind and Naiman 2008). Model predicted peak serum Bisphenol A levels fell within the range of pM, with 95% of human variability ranged within an order of magnitude, suggesting that an uncertainty factor of less than 10 would be reasonable to account for the inter-individual variability in pharmacokinetics.”

Additional uncertainty: 6

EFSA applied an additional factor of 6 for the t-TDI derivation for the general population. EFSA defined a HED of 609 μg/kg/day based on the BMDL10 of 8960 μg/kg/day for kidney weight in the mice 2-generation study (Tyl et al., 2008) and toxicokinetic differences in mice and humans. EFSA evaluated all available data in a weight-of-evidence approach and considered all other endpoints, with one exception, to be “less than likely” in their hazard evaluation.
EFSA (2015) assigned a likelihood level of “likely” to Bisphenol A induced proliferative changes in the mammary gland. No BMDL10 could be calculated for mammary gland effects based on the Delclos (2014) study. As outlined in the chapter “Carcinogenicity” the authors of this study and independent pathologists concluded: „Taking the incidences, the statistical testing results, and all pathologists and study authors opinions together, the authors of the NTP report (Gu and Mitkus, 2013), concluded that the evidence for duct hyperplasia in the mammary gland of females on either PND 21 or PND 90 was weak. They considered it an equivocal finding that may be the reflection of normal variability and/or a reflection of limits in tissue processing. Bisphenol A did not cause duct hyperplasia in the mammary glands of male rats, while conversely the reference estrogen EE2 induced hyperplasia in the male but not the female mammary gland.”
EFSA performed an uncertainty evaluation and defined the dose range which approached “likely” in the (HED) to be 100–1000 μg/kg bw per day. This dose range covers the HED of 609 μg/kg/day derived for kidney toxicity and used as a starting point for the t-TDI derivation for the general population. EFSA defined an uncertainty factor of 6 for the general population to cover the lower border of the 10-fold dose range which approached “likely”.

As a conservative approach we included this uncertainty factor of 6 for the DNEL derivation of the general population and workers, however we consider it as overconservative as no health effect is attributed to this assessment factor of 6 and the induced proliferative changes in the mammary gland identified by EFSA in 2015 were considered as an equivocal finding that may be the reflection of normal variability by the NCTR/NTP study authors. In line, also in the largest-ever scientific investigation of Bisphenol A, the CLARITY Core study, no consistent effect on proliferative changes in the mammary gland were observed, which supports this conclusion.

The overall assessment factor considered for worker dermal DNEL derivation is: 5 x 2.5 x 6 = 75

 

 

General Population - Hazard via inhalation route

Systemic effects

Long term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

1 mg/m³

Most sensitive endpoint:

repeated dose toxicity

Route of original study:

By inhalation

DNEL related information

DNEL derivation method:

other: based on SCOEL Recommendation 2014

Modified dose descriptor starting point:

NOAEC

Value:

10 mg/m³

Explanation for the modification of the dose descriptor starting point:

subchronic inhalation study

Acute/short term exposure

Hazard assessment conclusion:

hazard unknown but no further hazard information necessary as no exposure expected

DNEL related information

Local effects

Long term exposure

Hazard assessment conclusion:

hazard unknown but no further hazard information necessary as no exposure expected

Acute/short term exposure

Hazard assessment conclusion:

hazard unknown but no further hazard information necessary as no exposure expected

DNEL related information

General Population - Hazard via dermal route

Systemic effects

Long term exposure

Hazard assessment conclusion:

hazard unknown but no further hazard information necessary as no exposure expected

Acute/short term exposure

Hazard assessment conclusion:

hazard unknown but no further hazard information necessary as no exposure expected

DNEL related information

Local effects

Long term exposure

Hazard assessment conclusion:

hazard unknown but no further hazard information necessary as no exposure expected

Acute/short term exposure

Hazard assessment conclusion:

hazard unknown but no further hazard information necessary as no exposure expected

General Population - Hazard via oral route

Systemic effects

Long term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

53 µg/kg bw/day

Most sensitive endpoint:

repeated dose toxicity

Route of original study:

Oral

DNEL related information

DNEL derivation method:

ECHA REACH Guidance

Overall assessment factor (AF):

150

Modified dose descriptor starting point:

other: human equivalent dose (HED)

AF for dose response relationship:

1

Justification:

EFSA 2015

AF for differences in duration of exposure:

1

Justification:

EFSA 2015

AF for interspecies differences (allometric scaling):

1

Justification:

Kinetic data is available in a variety of species. Allometric scaling is considered based on AUC values and taken into account to drive the dose descriptor starting point. See discussion for detailed information.

AF for other interspecies differences:

2.5

Justification:

EFSA 2015

AF for intraspecies differences:

10

Justification:

EFSA 2015

AF for the quality of the whole database:

1

Justification:

EFSA 2015

AF for remaining uncertainties:

6

Justification:

A BMDL10 of 8960 μg/kg/day was calculated by EFSA (2015) for changes in the mean relative kidney weight in a two generation toxicity study in mice. This value is taken by EFSA as a starting point for their TDI calculation. The same starting point is taken for systemic DNEL calculation. For extrapolation from animal to human EFSA (2015) used the human equivalent dose concept. This approach is also used for the systemic DNEL calculation. To this, an AUC of 2.98 nmol x h/L for mice orally dosed with 100 μg/kg and an AUC of 3.36 nmol x h/L for human orally dosed with 100 μg/kg are calculated based on animal scaling (see discussion for more details).

Acute/short term exposure

Hazard assessment conclusion:

hazard unknown but no further hazard information necessary as no exposure expected

DNEL related information

General Population - Hazard for the eyes

Local effects

Hazard assessment conclusion:

hazard unknown but no further hazard information necessary as no exposure expected

Additional information - General Population

Inhalation:
An inhalation DNEL(worker) of 2 mg/m3 is taken based on the SCOEL recommendation for an OEL of 2 mg/m3 for worker covering local and potential systemic effects (SCOEL/SUM/113; June 2014). According to the REACH Guidance documents there is a factor of 2 between the intraspecies differences for worker (5) and intraspecies variability for the general population (10). This factor is applied to the worker DNEL of 2 mg/m3 and results in a DNEL for the general population of 1 mg/m3.

As a conservative approach short-term and long-term DNEL will be set at the same value; 1 mg/m3.

This value of 1 mg/m3 based on the SCOEL recommendation is supported by a supplemental derivation of the inhalation DNEL (general population) according to ECHA guidance (see below).

Supplemental Derivation of the DNELinhalative, systemic for general population:
As Starting point a NOAEC based on systemic toxicity is derived from a sub chronic inhalation toxicity study (Nitschke et al., 1988). In accordance with the recent recommendation of SCOEL (2014) no evidence for systemic toxicity was observed in this study and therefore, the highest tested concentration of 150 mg/m3 is chosen as NOAEC for systemic toxicity. Time extrapolation needs to be taken into account as for the general population an exposure of 24 hours per day and 7 days per week is assumed, however in Nitschke et al., 1988 treatment was conducted 6 hours per day and 5 days per week. Interspecies differences allometric scaling is already included in this inhalation study and in the correction of exposure a default value of 2.5 is required according to ECHA Guidance. Moreover, a default factor of 10 is chosen for intraspecies differences, a factor of 1 for Dose-response/endpoint-specific/severity issue and for Quality of Database. Based on the available literature there is no evidence for i) systemic toxicity at concentrations already leading to local effects in the nasal cavity and ii) time-dependency of any observed effect based on available subacute and subchronic studies. Therefore, a factor of 1 for differences in duration of exposure is determined.

Calculation of corrected starting point:
Starting point: NOAEC (systemic toxicity)= 150 mg/m3
Time correction: 6h/24h x 5d/7d
Corrected starting point= 150 mg/m3 x 6h/24h x 5d/7d= 26,8 mg/m3

Assessment Factor                                                     Value
Interspecies differences                                               2.5
Intraspecies differences                                               10
Dose-response/endpoint specific/severity issue              1
Quality of whole database                                            1
Differences in duration of exposure                              1
Overall Assessment Factor                                           25

Supplemental DNELinhalative, systemic= 26,8/25 = 1,08 mg/m3≈ 1mg/m3
This shows that a derivation of the DNELsystemic, inhalative for general population according to ECHA guidance results in the same value as the derivation based on the SCOEL recommendation in 2014.

 

Oral:
Starting point for systemic oral DNEL calculation:

A BMDL10 of 8960 μg/kg/day was calculated by EFSA (2015) for changes in the mean relative kidney weight in a two generation toxicity study in mice. This value is taken by EFSA as a starting point for their TDI calculation and is also taken as a starting point for the DNEL calculation.

Correction of the starting point:

EFSA (2015) used the so called Human Equivalent Dose concept for extrapolation from animal experiment to a human equivalent dose. This concept is also used for the derivation of the DNELoral for general population.

Derivation AUCmouse, oral:
The AUCmouse, oral value was calculated according to a recent species scaling approach (Summit Toxicology, 2020 – Update to Poet & Hays, 2018). A general instruction how to derive the AUCmouse, oral according to this approach is described above (see additional information DNELdermal for general population).

Derivation AUChuman, oral:

EFSA (2015) simulated the AUC (Area under the Curve) of human orally dosed with 100 μg/kg to be 3.6 (nmol x h/l) from a PBPK model based on life stage-specific kinetic data in monkey and rat, which are scaled up to the human situation (Yang et al., 2013). In the meantime, human exposure data became available from two recently published studies (Thayer et al., 2015; Teeguarden et al., 2015).
Relying on one particular study would directly lead to the acceptance of variabilities in the experimental setup, the dosing and the sampling procedures. Instead the already addressed species scaling approach is preferred as by analyzing and utilizing multiple toxicokinetic data derived under different experimental conditions and over a wide range of doses/species increased confidence can be provided (Summit Toxicology, 2020 – Update to Poet & Hays, 2018). A general instruction how to derive an species specific AUC according to this approach is described above (see additional information DNELdermal for general population). For the derivation of an AUChuman, oral a bodyweight of 70kg and a dose of 100 μg/kg (=438 nmol/kg) are assumed.

CLhuman, oral [L/h*kg]= 138.8*70-0.015= 130.2 L/h*kg

AUChuman, oral [nmol*h/L]= 438/130.2= 3.36 nmol*h/L

This predicted AUC value is taken to derive a HED for DNEL derivation instead of the value derived by EFSA (2015).

Calculation of Human Equivalent Dose (HED):
Starting point: BMDL10 =8960 μg/kg bw/day
AUCoral, mice (100 μg/kg) = 2.98 nmol x h/l
AUCdermal, human (100 μg/kg) = 3.36 nmol x h/l
Human Equivalent Dose Factor (HEDf)= 2.98 nmol*L/3.36nmol*L= 0.9

HED = 8 960 μg/kg bw/day x 2.98/3.36 = 7947 μg/kg bw/day

Assessment factors:
Interspecies differences (toxicokinetics): 1
Differences in toxicokinetics after oral dosing in mice and potential oral human exposure is taken into account to derive a Human Equivalent Dose.

Interspecies differences (additional uncertainty): 2.5

2.5 is the REACH guidance default factor. This factor is also applied by EFSA (2015).

Intraspecies actor (general population): 10

10 is the REACH guidance default factor. This factor should be considered conservative based a recent PB-PK model reported by Yang et al., (2015. Toxicol. Appl. Pharmacol.). The authors conclude concerning the interspecies factor in the general population: „In this study, the recalibrated human Bisphenol A PBPK model was used to estimate the inter-individual variability of internal dose metrics of Bisphenol A for the general population based on the estimated daily intake of Bisphenol A in the United States (FDA 2014b; Lakind and Naiman 2008). Model predicted peak serum Bisphenol A levels fell within the range of pM, with 95% of human variability ranged within an order of magnitude, suggesting that an uncertainty factor of less than 10 would be reasonable to account for the inter-individual variability in pharmacokinetics.”

Additional uncertainty: 6

EFSA applied an additional factor of 6 for the t-TDI derivation for the general population. EFSA defined a HED of 609 μg/kg/day based on the BMDL10 of 8960 μg/kg/day for kidney weight in the mice 2-generation study (Tyl et al.,2008) and toxicokinetic differences in mice and humans. EFSA evaluated all available data in a weight-of-evidence approach and considered all other endpoints, with one exception, to be “less than likely” in their hazard evaluation.

EFSA (2015) assigned a likelihood level of “likely” to Bisphenol A induced proliferative changes in the mammary gland. No BMDL10 could be calculated for mammary gland effects based on the Delclos (2014) study. As outlined in the chapter “Carcinogenicity” the authors of this study and independent pathologists concluded: “Taking the incidences, the statistical testing results, and all pathologists and study authors opinions together, the authors of the NTP report (Gu and Mitkus, 2013), concluded that the evidence for duct hyperplasia in the mammary gland of females on either PND 21 or PND 90 was weak. They considered it an equivocal finding that may be the reflection of normal variability and/or a reflection of limits in tissue processing. Bisphenol A did not cause duct hyperplasia in the mammary glands of male rats, while conversely the reference estrogen EE2 induced hyperplasia in the male but not the female mammary gland.”
EFSA performed an uncertainty evaluation and defined the dose range which approached “likely” in the (HED) to be 100–1000 μg/kg bw per day. This dose range covers the HED of 609 μg/kg/day derived for kidney toxicity and used as a starting point for the t-TDI derivation for the general population. EFSA defined an uncertainty factor of 6 for the general population to cover the lower border of the 10-fold dose range which approached “likely”.

As an conservative approach we included this uncertainty factor of 6 for the DNEL derivation of the general population and workers, however we consider it as overconservative as no health effect is attributed to this assessment factor of 6 and the induced proliferative changes in the mammary gland identified by EFSA in 2015 were considered as an equivocal finding that may be the reflection of normal variability by the NCTR/NTP study authors. In line, also in the largest-ever scientific investigation of Bisphenol A, the CLARITY Core study, no consistent effect on proliferative changes in the mammary gland were observed, which supports this conclusion.

The overall assessment factor considered for general population dermal DNEL derivation is: 10 x 2.5 x 6 = 150