Commentary on the draft revised guideline on the environmental risk assessment of medicinal products for human use

Applicants for marketing authorisation for human medicinal products in the European Union must submit an environmental risk assessment which is assessed by assessors from the national competent authorities. The EMA guideline on the environmental risk assessment of medicinal products for human use came into effect on 1 December 2006. After 12 years’ experience with the guideline, the EMA has released for public consultation a draft revision of the guideline. The revision proposes significant substantive and structural changes to the guideline. The major changes proposed in the revision are outlined together with the rationale for the changes and the expected impact on stakeholders.

The Phase II risk assessment starts with studies on physico-chemical properties, and on the 130 environmental fate and ecotoxicological effects of the active substance. For some groups of 131 substances, a tailored risk assessment strategy should be followed that addresses their specific 132 mechanism of action (section 4.3). In Tier A, the PEC is compared to an acceptable environmental 133 concentration, the Predicted No Effect Concentration (PNEC). When a risk is identified in Tier A, a Tier 134 B assessment with PEC refinement and if warranted further effect studies should be performed. 135 The studies that should be performed in Phase II Tier A on physico-chemical characteristics, fate and 136 ecotoxicity are described in section 4.2.1. The requirement for a risk assessment for certain 137 environmental compartments (soil and groundwater) depends on whether trigger values are met by 138 the outcome of these studies. Information on data search and evaluation is provided in section 6. 139 The Phase II risk assessment for the surface water compartment including options for risk refinement 140 is described in section 4.2.3. Sections 4.2.4. -4.2.7. give guidance on Phase II risk assessment and 141 risk refinement for sediment, functioning of sewage treatment plants (STP), soil and groundwater, 142 respectively. The assessment of risk to predators eating contaminated prey (secondary poisoning) is 143 described in section 4.2.8. 144 145 The PBT (Persistent, Bioaccumulative and Toxic) assessment concerns the identification of certain 146 intrinsic properties of the active substance. These properties make the long-term risks to the 147 environment unpredictable; hence environmental exposure should be prevented as much as possible. 148

PBT assessment
As the PBT assessment concerns intrinsic properties of the active substance subsequent exposure is 149 not considered. The assessment of PBT and vPvB properties is described in section 5. Compounds 150 entering the screening phase (section 5.1) are identified in the first part of the decision tree (Question 151 1-3). Depending on the outcome of the screening phase, a definitive assessment may be required. 152 (section 5.2). 153 In exceptional cases for substances which do not meet the trigger for PBT assessment (log Kow > 4.5) 154 an assessment of PBT/vPvB properties may be required. This will be the case if the results obtained in 155 Phase II of the risk assessment demonstrate that the B-and T-criteria are met, or if the vB-criteria is 156 met (see Table 16). 157

Finalization of risk and PBT assessment
158 When a risk is identified and/or a substance is classified as PBT/vPvB, this information should be 159 included in the SmPC and risk mitigation measures should be discussed. These are described in section 160

161
The structure of the risk assessment report is described in section 8. 162 163 The ERA should be performed for the environmentally relevant chemical species, which in most cases 164 is the parent compound. 165

166
The ERA is based on a 'total residue approach', i.e. the assumption that the active substance is 167 completely excreted as parent substance without metabolism or assuming that metabolites have 168 similar or lower toxicity than that of the parent substance. 169 Metabolism of the active substance may be taken into account in Phase II, see section 4.2.3.2. 170 For a prodrug, the most environmentally relevant substance will generally be the pharmacologically 171 active metabolite. However, there may be instances where a prodrug is incompletely converted to the 172 active (<50%), or excreted largely (>50%) intact or via metabolic pathways that do not generate the 173 active moiety. In these cases, the selection of the environmentally relevant chemical species should be 174 justified. In some cases, assessment of both prodrug and active may be necessary. 175 For fixed combination products, the ERA is performed separately for each compound within the 176 product. 177

200
This section presents guidance on how to conduct the Phase I risk assessment. The potential for 201 environmental exposure is assessed based on the nature of the active substance and the intended use. 202 In Phase I, products that require a more extensive Phase II risk assessment -either standard or 203 tailored -are identified. It is assumed that active substances with limited use and/or limited 204 environmental exposure will have limited environmental effects, and thus the risk assessment will stop 205 in Phase I. 206 The Phase I risk assessment consists of a decision tree (Figure 2). The questions in the decision tree 207 are described in detail below Figure 2. The outcome of Phase I may be that the risk assessment stops, 208 or that a Phase II risk assessment is required. When at least one of the Phase I criteria to stop the risk 209 Q3a: Is the active substance a non-natural peptide/protein? 256 Peptides and proteins that have been structurally modified using non-natural amino acids to increase 257 biostability are considered non-natural. 258 Protein-drug conjugates including natural proteins do not belong to this group and would require 259 standard assessment of the non-protein-moiety. 260 Q3b: Is the non-natural peptide/protein readily biodegradable? 261 For non-natural peptides/proteins, an additional screening step should be performed to demonstrate 262 that they will be quickly degraded in the environment and will not enter the STP. 263 When the non-natural peptide/protein is demonstrated to be excreted in amounts < 10% of the dose, 264 or shown to be readily biodegradable in an OECD 301 test, the ERA stops. If the PEC SW value is < 0.01 µg/L and no other environmental concerns are apparent, it is assumed 284 that the medicinal product is unlikely to represent a risk for the environment following its prescribed 285 usage in patients and no further risk assessment is required. 286 Q6: Is the refined PEC SW ≥ 0.01 µg/L? 287 PEC SW may be refined by refining the F PEN value based on prevalence data and/or based on the 288 treatment regimen. For medicinal products, which can be used for more than one indication, the 289 calculation of refined PEC SW should take into account all designated indications for the product. The 290 total PEC SW is the sum of the PEC SW for each indication, which should be calculated using the maximum 291 prescribed dose for each indication. The other default values representing a realistic worst case 292 environmental exposure scenario should not be replaced by other data. If the refined PEC SW value is < 293 0.01 μg/L, and no other environmental concerns are apparent (e.g. the compound is a potential EAS or 294 paraciticide), it is assumed that the medicinal product is unlikely to represent a risk for the 295 environment following its prescribed usage in patients and no further risk assessment is required. 296 Prevalence: The F PEN can be refined by submitting European disease prevalence data for the sought 297 indication(s). Such data should be published by a reliable and independent source, e.g. a peer-298 reviewed scientific journal or the World Health Organization (WHO) (e.g., the International Agency for 299 Research on Cancer (IARC)). It is assumed that 100% of the patient population is taking the medicinal 300 product for the relevant disease(s) daily and thus the F pen reflects the prevalence of the disease. If 301 regional differences exist, the F PEN should be calculated for the member state or region with the highest 302 prevalence of the disease. This member state should be one of the member states included in the 303 authorisation procedure. Prevalence data at subnational level (i.e. for regions smaller than a country) 304 can also be used in the risk assessment, provided they are of good quality as described above and 305 justification for use in the risk assessment is provided. Prevalence data should be as recent as 306 possible, preferably not older than 5 years. The use of older data should be justified. For orphan drug 307 submissions, the F PEN can be refined based on the prevalence for which the medicinal orphan drug 308 designation was based, as adopted by the Committee for Orphan Medicinal Product (COMP). One year 309 prevalence data should be used unless other prevalence data (e.g. multiple year prevalence, lifetime 310 prevalence or incidence if appropriate) can be justified considering epidemiologic and posology data 311 available for the supported indication. 312 Treatment regimen: The F PEN may be refined taking the worst-case treatment period (t TREATMENT ) and 313 worst-case number of treatment repetitions per year (n TREATMENT ) into consideration. This is easily done 314 for products intended for single use (e.g. during surgery, diagnostics, etc.) or other products with a 315 well-defined treatment regimen. For example, an anti-cancer drug administered for five days in 316 monthly cycles, t TREATMENT equals 5 days and n TREATMENT  The following approach may be used for the refinement of F PEN by prevalence data and /or by 327 If the PEC SW value based on a refined F PEN is < 0.01 μg/L, and no other environmental concerns are 336 apparent, it is assumed that the medicinal product is unlikely to represent a risk for the environment 337 following its prescribed usage in patients and no further risk assessment is required. 338 Q7: Does the active substance have a specific toxicity profile? 339 A tailored risk assessment is needed for compounds with a specific mode of action (e.g., endocrine 340 active substances, antibiotics), see section 4.3. 341 determination of some of these properties is therefore mandatory for the assessment.

Octanol/water partitioning coefficient (Kow) 367
The octanol/water partitioning coefficient, Kow, should be determined experimentally using the shake-368 flask method (OECD 107) or the slow-stirring method (OECD 123). A calculated value is generally not 369 acceptable. The results from the HPLC screening method (OECD 117) may only be used for indicative 370 purposes, e.g. for compounds, which are highly soluble and have a predicted log Kow < 1 at all 371 environmentally relevant pH values. 372 For compounds with log Kow > 4, the shake-flask method cannot be used and only the slow -stirring 373 method is acceptable. This range of applicability is based on OECD guidelines 123 and 107. 374 For dissociating compounds, an ion-corrected log Dow for the neutral molecule should be reported 375 together with the respective pKa value(s). The ion-corrected Dow is equal to Kow. 376 Log Dow values should be determined as a function of pH covering an environmentally relevant pH-377 range (at least 3 pH values ranging from pH 5 to 9) e.g. by measuring the pH-lipophilicity profile 378 (log D as function of pH). If the Dow value (for dissociating substances) at any pH value between pH 5 379 and pH 9 meets the trigger values for assessment of secondary poisoning (log Kow ≥ 3) or PBT 380 assessment (log Kow > 4.5), further assessment is required (see Section 4.2.8 and 5). 381

Dissociation constant 382
The dissociation constant should be determined for dissociating compounds. The results of this study 383 are used to verify exposure concentrations in fate and ecotoxicity tests. Additionally, the information is 384 required to determine the octanol/water partitioning coefficient. 385

Fate studies 386
Along with mandatory studies on physico-chemical properties, mandatory fate studies should be 387 included in the ERA in order to evaluate the fate and predict the environmental exposure of the 388 medicinal product. These mandatory studies are listed in Table 1. 389

Sorption to soil and sludge 390
Adsorption/desorption studies generate essential information on the mobility of the active substance 391 and its distribution in the soil and water compartments. This is a complex process depending on many 392 factors including chemical properties, characteristics of the soil and climatic factors. Therefore, 393 different sludge and soil types should be used in order to cover as widely as possible the interactions of 394 the active ingredient with sludge and soils. 395 A study according to OECD 106 using 2 types of sludge and 3 soil types, differing in organic carbon 396 content, and soil texture is preferred. The results are used to evaluate the requirement for soil and 397 groundwater assessment (section 4.2.2) and to perform PEC calculations for soil and sediment in 398 Phase II Tier A. In Phase II Tier B, adsorption data for at least 2 types of sludge, preferably from two 399 different STPs are necessary for PEC SW refinement (SimpleTreat modelling, section 4.2.3.2). Adsorption 400 data for at least 3 soils are needed for equilibrium partitioning calculations in the sediment risk 401 assessment (Section 4.2.4) and refinement of PEC GW in Tier B (section 4.2.6.2). An overview of Phase 402 II risk assessment steps where adsorption data are needed is listed in Table 2 below. 403 The targeted endpoint for adsorption studies should be the distribution coefficient (Kd), defined as the 404 ratio between the content of the substance in the soil/sludge phase and the mass concentration of the 405 substance in the aqueous solution, under the test conditions, when adsorption equilibrium is reached. 406 The organic carbon normalized adsorption coefficient (Koc) relates the distribution coefficient Kd to the 407 organic carbon content of the soil sample. 408 with radiolabelling provides the most robust information. 417

Ready biodegradability 418
The readily biodegradability of a substance should be determined according to OECD 301. The 419 microbial community should not be pre-exposed to the test compound in this test, and addition of 420 more inoculum is not allowed. OECD 301 can be waived if OECD 314 B (for PEC refinement in Phase II 421 Tier B) or OECD 308 (for PBT assessment or PEC refinement for groundwater) is performed. The 422 results of OECD 301 are used for triggering soil and groundwater assessment and in the Simple Treat 423 calculation. Substances classified as not readily biodegradable are considered potentially persistent. 424

Ecotoxicity studies 425
To determine the aquatic ecotoxicity, chronic ecotoxicity data i.e. No Observed Effect Concentration 426 (NOEC) or 10% effect concentration (EC10) for species from three trophic levels are required (See 427 Active substances with high affinity for organic carbon have a greater likelihood of accumulating in 468 sludge and ending up in the soil, unless the active substance is readily biodegradable. However, 469 substances with lower adsorption affinity may also be present in sludge at high concentrations, when 470 the release to sewage treatment plants is high. Hence, the final exposure of soil organisms depends on 471 both main parameters, i.e. the properties of the pharmaceutical (Koc value) and the total release to 472 the wastewater flow, which again depends on the dose and the fraction of a population receiving the 473 active substance during a given time. The PEC SW calculated in Phase I, reflects directly these 474 parameters, as it disregards processes such as biodegradation or retention of the active substance in 475 the STP. Hence, the PEC SW is used in combination with Koc to trigger assessment for the soil 476 compartment, see Table 3 and section 4.2.6. 477 STPs is assumed. When the PEC/PNEC ratio is ≥ 1, a risk to the aquatic compartment as a whole (not 491 a particular sensitive group of species) is indicated. If a risk is identified in Phase II Tier A, a refined 492 assessment may be performed in Phase II Tier B. 493

Phase II Tier A assessment for surface water 494
Exposure assessment for surface water 495 The final PEC SW as calculated in Phase I should be used (see Eq. 1-3). 496 Effect assessment for surface water 497 To derive a PNEC, chronic ecotoxicity data for species from at least three trophic levels (algae, Daphnia 498 and fish) are required, as described in section 4.2.1. 499 The PNEC SW is calculated by applying an assessment factor (AF) of 10 to the lowest EC10 or NOEC 500 value from the aquatic test species. The AF is an expression of the degree of uncertainty in the 501 extrapolation from a limited number of test species to complex ecosystems in the actual environment 502 and accounts for, inter-species variations in sensitivity, intra-species variability and laboratory data to 503 field impact extrapolation. 504

511
If the surface water RQ is < 1, then further testing in surface water is not required and it can be 512 concluded that the active substance is unlikely to represent a risk to surface water. 513 If the surface water RQ is ≥1, a Tier B assessment is required. 514

Phase II Tier B assessment for surface water 515
When a risk is established in Tier A, the PEC SW may be refined using one or more of the options below: 516

•
Fpen, if not refined in Phase I Tier A. For more information, see Q6 in section 4.1. 517

•
Metabolism 519 • Potential removal in the STP. 520 Refinement of PEC SW using consumption data 521 At the renewal of a marketing authorisation for a medicinal product, consumption data on the active 522 substance may be used to refine F PEN (equation 5) and the PEC SW , with the possibility of a 523 consequential impact on the conclusion of the previous ERA. The data used should come from a reliable 524 and publicly available source and demonstrate a stable consumption over the last 3 or more years. A 525 market share of 100% is always assumed. If regional differences exist, data from the member state 526 with the highest calculated F PEN should be used. 527 Parameters used in Eq. 5: 530 [inh] 531 Refinement of PEC SW using metabolism data 532 If a potential risk for the medicinal product to the environment has been identified based on the total 533 residue approach, then the total residue approach may be abandoned and the risk may be refined by 534 subtracting the fractions of metabolites. If the total residue approach is abandoned, a full Phase II risk 535 assessment is required for each metabolite constituting ≥10% of the administered dose. The PEC is 536 then calculated separately for the parent compound and these metabolites and all resulting PEC/PNEC 537 ratios are summed for the evaluation of environmental risk of the product. If it is not possible to 538 perform the ERA for the metabolites excreted in fractions ≥ 10% of the dose, the total residue 539 approach should be used. If a risk is identified and it is not possible to refine the risk by testing the 540 metabolites, the ERA should be concluded with the statement that the use of the product is expected 541 to result in a risk to the environmental compartment(s) concerned. 542 The following approach may be used for this refinement: 543

545
Parameters used in Eq. 6: 546 Adsorption of the active substance to sewage sludge in STPs, using the data from the estimation of 553 the adsorption coefficient (OECD 106) 554 • Test for ready biodegradability in the STP (OECD 301)/measured removal rates using the OECD 555 314 B study. 556 For local scale assessments, it is assumed that one point source is releasing its wastewater to one STP.   Table 2 FACTOR Factor taking the adsorption to suspended matter into account If the RQ for surface water is < 1, it may be anticipated that the active substance in the medicinal 624 product will not pose a risk to the aquatic environment. 625 When a risk to the surface water ecosystem cannot be excluded, the applicant should propose 626 adequate precautionary and safety measures to protect surface water ecosystems (see also section 7). 627

628
For the sediment risk assessment, PEC SED is derived from PEC SW as calculated in phase I (see equation 629 1-3) using equilibrium partitioning (EqP) between water and sediment consisting of freshly deposited 630 suspended matter. A PNEC SED is derived using tests with sediment dwelling organisms. Both PEC and 631 PNEC should be based on sediment with equal (normalized) organic carbon content and on a dry 632 weight basis. 633  The fraction bound residue that may have been determined in fate studies, may not be subtracted 665 from the PEC SED . 666

Effect assessment for sediment 667
To determine a PNEC SED , a minimum of one study with sediment dwelling organisms should be 668 performed using a sediment-water test system (Table 7). In general, tests using a spiked sediment 669 procedure are preferred. However, if the characteristics of the test substance make it impossible to 670 spike sediment in a reliable manner (e.g. high water solubility, low binding affinity to sediment) it may 671 be more appropriate to use the spiked water procedure. 672 For ionisable compounds, care should be taken that testing is performed at an environmentally 673 relevant pH (5-9). For these compounds, a tailor-made approach may be followed if it can be 674 substantiated and is well reported. 675 . .
If the risk quotient is ≥ 1, risk refinement may be performed in Phase II -Tier B. 694

Phase II Tier B assessment for sediment 695
If a risk is identified in Tier A, refinement of PEC SW (see section 4.2.3.2) may also be used for Tier B 696 sediment assessment. If a risk to sediment organisms still cannot be excluded, the applicant should 697 propose adequate precautionary and safety measures to protect sediment ecosystems (see also 698 section 7). 699

700
The functioning of STPs is essential for good water quality management. Substances with anti-701 microbial activity may affect microbial communities. The microbial community most likely exposed to 702 the highest concentrations of the substance(s) is the activated sludge community. In order to evaluate 703 the anti-microbial effects of anti-microbial substances, the activated sludge respiration inhibition test 704 (OECD 209) should be used. 705

Phase II Tier A assessment for STP 706
Exposure assessment for STPs 707 To determine the risk for STPs, PEC SW as calculated in phase I (see Eq. 1-3) should be recalculated into 708 a PEC STP . This is achieved by multiplying the PEC SW with a factor of 10, as there is no dilution of 709 effluent with surface water. 710 Effect assessment for STP 711 The PNEC is based on the respiration inhibition test for activated sludge (OECD 209), by applying an 712 assessment factor of 10 to the EC10 or NOEC value. 713

720
When the risk quotient is ≥ 1, risk refinement options as described for surface water may be used in 721 Phase II Tier B. 722

Phase II Tier B assessment for STP 723
The exposure concentration in the aeration tank of the SimpleTreat model (PEC AERATION TANK ) should be 724 used to refine the risk quotient for microorganisms. PEC AERATION TANK is equal to Clocal EFF , see also

729
Entry into the groundwater is considered to be via bank filtration, except for substances with an 730 average Koc >10,000 L kg -1 or for substances that are readily biodegradable. It is assumed that the 731 exposure of groundwater via sewage sludge incorporated into soil can be disregarded with reference to 732 the high sorption affinity of these active substances to the soil. 733

Phase II Tier A assessment for groundwater 734
Exposure assessment for groundwater 735 The groundwater PEC (PEC GW ) is based on the PEC SW as calculated in phase I (see eq. 1-3) and is 736 estimated by a simple equation. 737 738 = 0.25 × .

739
Effect assessment for groundwater 740 The PNEC GW is based on the PNEC SW (see 4.2.3.1) and an additional assessment factor. Groundwater 741 ecosystems are fundamentally different to surface water ecosystems and therefore may be more 742 vulnerable as they lack the ability to recover from perturbations. Consequently, an additional 743 assessment factor of 10 should be applied to extrapolate the PNEC GW from the PNEC SW (Eq. 22 below). 744 745 = 10 .

Risk characterization 747
The risk quotient (RQ) for the groundwater compartment is determined using the PNEC for 748 groundwater (equation 23). 749 750 Guideline on the environmental risk assessment of medicinal products for human use EMEA/CHMP/SWP/4447/00 Rev. 1 Page 29/48 = .

751
If the risk quotient is ≥ 1, risk refinement options should be used in Phase II Tier B as described 752 below. 753

Phase II Tier B assessment for groundwater 754
If the RQ GW is ≥1, further evaluation is needed in Tier B using one or more of the options below. 755

•
Groundwater modelling for a realistic worst case scenario according to SiMBaFi -a bank filtration 757 simulation model. The model and a detailed description can be downloaded here: 758 www.uba.de/simbafi 759 The following parameters are needed: 760 • PEC SW-REFINED as described in section 4.2.3.2. 761

•
Adsorption of the active substance to soil derived from batch equilibrium test (OECD 106). SiMBaFi 762 requires the non -oc-normalized Kd or Kf -value (Kf -Freundlich adsorption coefficient) as input. 763 The lowest Kd/Kf derived from soil should be used (n=3). If 4 or more soils are available the 764 geometric mean may be used. Kd derived from sludge cannot be used. 765 • Degradation as DT 50 value derived from an OECD 308 study (total system, calculated using single 766 first order kinetics, normalised to 12°C, highest value of 2 test systems). 767 For the calculation of the PEC GW the "realistic worst case" determined in SiMBaFi should be used, i.e. a 770 groundwater flow time of 5 days between the surface water and the groundwater well. For calculation 771 four steps are needed as described below: 772 Calculation of retardation: 773 .

785
As the percentage of bank filtrate at the production well is assumed to be 100 % the resulting PEC GW 786 equals the calculated concentration in the production well (eq. 27). 787 789 Parameters used in Eq. 24-27: 790 Koc value of 10 000 L kg -1 indicates. 800 To determine a possible risk to the soil compartment, the PEC SOIL is compared to the PNEC SOIL . This 801 PNEC SOIL is derived using experimental long-term ecotoxicity data for soil microorganisms, soil dwelling 802 invertebrates and plant species (Table 11). Since sludge associated active pharmaceutical residues 803 may be available in soil compartment for a long time, short-term effect tests are inappropriate for risk 804 assessment. When the PEC/PNEC ratio is ≥ 1, a risk to the entire soil compartment (not a particular 805 sensitive group of species) is indicated. If a risk is identified in Phase II Tier A, a refined assessment 806 may be performed in Phase II Tier B. 807

Phase II Tier A assessment for soil 808
Tier A Exposure assessment for soil 809 The Tier A exposure assessment considers sludge application as the major entry path for the active 810 substance to be released to the soil environment. In a first step, the initial concentration in soil after 811 the first application is calculated using the predicted concentration of the active substance in sludge. 812 For substances which accumulate and are not easily degraded, the concentration in soil after repeated 813 sludge application should also be assessed. In order to consider the biodegradation of the active 814 substance in soil in between sludge applications a study on degradation in soil (OECD 307) is required. 815     Table 11). The long-term toxicity to soil organisms should be assessed as active substances in soils 865 may persist for a long time, or accumulation of the substance may occur when sludge is applied over 866 consecutive years. The PNEC soil is calculated by applying an assessment factor (AF) of 10 to the lowest 867 EC10 or NOEC value from the soil test species. 868 . .

883
If the risk quotient is ≥ 1, the risk assessment proceeds to Phase II -Tier B. 884

Phase II Tier B Assessment for soil 885
Tier B Exposure assessment for soil 886 If a risk for soil organisms has been identified in Tier A, it is possible to refine the emission rate to 887 influent wastewater by using consumption data and metabolism data as performed in Tier B for surface 888 water (see 4.2.3.2). 889 The refined emission rate to influent wastewater is used to recalculate the sludge concentration C SLUDGE 890 and the relevant PEC SOIL , as described above for Tier A. 891

Tier B Effect Assessment for soil 892
If the RQ SOIL from nitrogen transformation in Tier A is still ≥1, further evaluation of the PNEC may be 893 possible in Tier B by extending the microorganisms Nitrogen Transformation Test (OECD 216) to 100 894 days (Table 12). 895 * An assessment factor is not relevant to this endpoint -when the difference in rates of nitrate formation between the lower 897 treatment (i.e., the maximum PEC) and control is ≤ 25% at any sampling time before day 100, the substance can be evaluated as 898 having no long-term influence on nitrogen transformation in soils.

Risk characterisation 900
The refined RQ SOIL should be recalculated using the refined PEC SOIL  case, it should also be checked whether the T-criterion (Table 16) is fulfilled. In this case, the P-920 criteria (Table 16) should be also assessed, either by using the study on degradation in soil (if soil 921 assessment was triggered) or by performing an aquatic degradation study (OECD 308 or 309). In case 922 of a BCF-value > 5000, degradation should be assessed using the vP criteria (Table 16). 923

Bioconcentration factor 924
The BCF is determined in fish using the OECD 305 test guideline (these results may also be used for 925 the PBT assessment, see section 5.2). Aqueous exposure is the preferred methodology when 926 technically feasible because dietary exposure yields a biomagnification factor (BMF) rather than a BCF, 927 which then should be estimated from the depuration rate constant. The kinetic calculation of BCF 928 (based on uptake and elimination rates and taking dilution due to fish growth into account) is preferred 929 over the steady state calculation (based on concentrations in fish and water) and BCF values should be 930 normalized to 5% lipid content. A minimized study design is also described in OECD 305 but this may 931 only be used for screening purposes. It may not be used to determine an accurate BCF value because 932 it cannot be determined whether steady state is reached (see OECD guidance document No. 264, 2017 933 for additional information). 934 approach, when the PNEC SW, SECPOIS is higher than the PEC SW , a risk due to secondary poisoning is 948 identified. 949 Alternatively, the risk of secondary poisoning for predators in the aquatic food chain may be calculated 950 as the ratio of the concentration of the contaminant in the predator's food (PEC BIOTA ) and the no-effect-951 concentration for the oral intake (PNEC BIOTA ). If this risk quotient is ≥1, a risk of secondary poisoning is 952 identified. PEC BIOTA is then derived from PEC SW multiplied by BCF FISH (experimental data) and BMF 953 For certain groups of active substances, a tailored assessment is required for the aquatic compartment 960 due to their specific mode of action. This concerns compounds for which the action limit does not 961 apply, such as endocrine active substances (see section 4.3.2), but may also concern compounds for 962 which the action limit applies, such as antibiotics (see section 4.3.1). 963 For all active substances that require a tailored risk assessment, an ERA Phase II assessment is 964 required for all compartments, including fate aspects. For the aquatic compartment, OECD ecotoxicity 965 tests are available for a number of species that may replace standard test species, depending on the 966 mode of action. For soil and sediment, tailoring with regard to the choice of test species is often not 967 possible. 968

969
For active substances with an antibacterial mode of action, and no other known pharmacological 970 targets, a targeted effect assessment should be performed for the aquatic compartment. Scientific 971 knowledge and empirical data demonstrate that a tailored risk assessment focused on the effects on 972 in any of the following situations: (a) the degradation half-life in marine water is higher than 60 days; (b) the degradation half-life in fresh or estuarine water is higher than 40 days; (c) the degradation half-life in marine sediment is higher than 180 days; (d) the degradation half-life in fresh or estuarine water sediment is higher than 120 days; (e) the degradation half-life in soil is higher than 120 days.
A substance fulfils the "very persistent" criterion (vP) in any of the following situations:

1116
The REACH guidance on PBT assessment should be followed as much as possible, and deviations 1117 should be scientifically justified. It should be noted that for the REACH PBT assessment a tiered 1118 approach is followed, since REACH chemicals do not necessarily contain all required information in the 1119 2 Substances known to have carcinogenic potential for humans (epidemiological and/or animal data) 3