Environmental risk assessment for excipients from

1 downloads 0 Views 321KB Size Report
Sep 26, 2015 - Many different excipients are used in galenical pharmaceutical production, in addition to the active .... Kaiseraugst for the year 2013 and the first six months of 2014 were ...... umentum:ProductBase_EU:09007af8800fcbff.pdf〉 (accessed 03.03.15.). .... Handbook of Environmental Data on Organic Chemicals,.
Sustainable Chemistry and Pharmacy 1 (2015) 28–35

Contents lists available at ScienceDirect

Sustainable Chemistry and Pharmacy journal homepage: www.elsevier.com/locate/scp

Environmental risk assessment for excipients from galenical pharmaceutical production in wastewater and receiving water Kim Catharina Wirz, Martin Studer, Jürg Oliver Straub n F.Hoffmann‐La Roche Ltd,, Group SHE, CH–4070 Basle Switzerland

art ic l e i nf o

a b s t r a c t

Article history: Received 20 March 2015 Received in revised form 6 July 2015 Accepted 24 August 2015 Available online 26 September 2015

Many different excipients are used in galenical pharmaceutical production, in addition to the active pharmaceutical ingredients. Excipients are little investigated regarding their environmental fate and impact, even though some of them are used in appreciable quantities. For 35 excipients used in galenical production at Roche Basle and Roche Kaiseraugst, both in Switzerland, in the years 2013 and 2014, the environmentally relevant properties were collated. A predicted environmental concentration (PEC) was calculated for the wastewater treatment plants (WWTPs) and the receiving water, the River Rhine for both sites, based on maximum daily losses of the single excipients to wastewater, derived by mass balance, and the site-specific dilution factor. Predicted no effect concentrations (PNECs) were derived for the WWTPs and the River Rhine. PECs and PNECs were compared for the WWTPs and the receiving water, in an environmental risk assessment. Additionally, to simulate a worst case scenario, certain galenical productions where given excipients are used in the highest amounts were assumed to take place in parallel on the same day, resulting in theoretical maximum excipient losses to wastewater. All PEC/ PNEC risk characterisation ratios for the excipients currently used by Roche in Switzerland are well below 1 throughout. Together with the fact that based on biodegradability data many excipients will be removed in the WWTP, this indicates that the excipients currently used do not present a risk to the environment. & 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Pharmaceutical excipients Galenical production Environmental risk assessment Wastewater treatment plant Receiving waters

1. Introduction Galenical production of medicines needs excipients for various purposes, e.g., to stabilise the active pharmaceutical ingredient (API), lend bulk and colour to tablets, assist in breaking these up in the stomach, control the time course of API release from the formulation, disguise a bitter taste of the pure API, or other functions (Swarbrick, 2015). The United States Pharmacopeia and National Formulary (United States Pharmacopeia, 2014) lists 48 functional categories for excipients, from Adhesive over

Abbreviations: API, Active pharmaceutical ingredient; BOD5, Biochemical oxygen demand in 5 days; EC50, 50% effect concentration; ERA, Environmental risk assessment; EU, European Union; NA, Not available; NOEC, No observed effect concentration; NRT, New Roche test; OECD, Organisation for Economic Co-operation and Development; PEC, Predicted environmental concentration; PNEC, Predicted no effect concentration; Q347, Minimum residual (5th percentile) river flow; QSAR, Quantitative structure–activity relationship; RA, Read-across; TOC, Total organic carbon; WWTP, Wastewater treatment plant n Corresponding to: F.Hoffmann-La Roche Ltd, Group SHE, 665/14, CH–4070 Basle, Switzerland, Fax þ 41 616881920. E-mail address: [email protected] (J.O. Straub).

Antimicrobial Preservative, Bulking Agent and many others, to Wetting/Solubilizing Agent. While in the USA and the European Union (EU), new APIs need an environmental risk assessment (ERA) for registration, there is no such requirement for excipients (even though an older EU ERA draft did mention excipients (Straub, 2002)). Excipients are being investigated regarding toxicological endpoints and quality standards (Kemsley, 2014), but not for environmental risks. Only one publication so far, by Silva et al. (2014), comparing algal toxicities of fluoxetine in drug products of differing compositions, i.e., containing varying excipients, suggested excipients to be responsible for manifestly different toxicities. However, some excipients are being used in comparatively high amounts in galenical production of medicines. Therefore, we collated data and performed a site-specific ERA for the wastewater treatment plants (WWTPs) and receiving water for the excipients used in two galenical production sites of F. Hoffmann‐La Roche Ltd (Roche) in Switzerland. To our knowledge, the present contribution is the first to investigate the potential environmental risks arising from losses of excipients from galenical production to wastewater.

http://dx.doi.org/10.1016/j.scp.2015.08.004 2352-5541/& 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

K.C. Wirz et al. / Sustainable Chemistry and Pharmacy 1 (2015) 28–35

2. Materials and methods 2.1. Excipient use amounts An initial list of excipients was prepared by collating excipient information from current or former, solid or liquid Roche drug products worldwide, based on unpublished company-internal documentation. Galenical production records at Roche Basle and Kaiseraugst for the year 2013 and the first six months of 2014 were then consulted to identify those excipients actually being used at those two sites. Maximum theoretical losses to wastewater were calculated by mass balances for the single productions as described by Hoerger et al. (2009). In brief, the final products and all documented losses were substracted from the total of active ingredients plus excipients, to result in worst-case losses, all of which were assumed to be drained to wastewater in the first approximation. Roche has two galenical production sites in Switzerland, viz. Roche Basle and Roche Kaiseraugst, both situated on the River Rhine, with Kaiseraugst approximately 13 km upriver from Basle. Production of solid medicinal products (‘solida’; coated or uncoated tablets, capsules, dry granulates (Hoerger et al., 2009)) in general needs more excipients than the formulation of liquid drugs (‘parenteralia’), e.g., due to the necessity for bulking or disintegrating agents, colourants and lacquers. In contrast, solida may be produced at up to 1000 kg per batch, with no or only little liquid (water or alcohol) added for technical reasons. Solida are mixed in large mixing units, in which at the end of the process a thin layer of caked mixture of APIs and excipients may remain (see Hoerger et al. (2009)). For cleaning, this cake is normally scraped out manually and disposed of as special waste by incineration, while the subsequent aqueous rinses from cleaning are discharged to the WWTP. In some special cases, due to high ecotoxicity of the API or in case of the API being categorized as ‘high-potency’ based on toxicological or pharmacological reasons, the first rinse (containing most of the active substance remaining in the mixer) will be separated and incinerated as well. Also, samples for quality control, air filters, tablet breakage or mixtures and products not meeting specifications will be incinerated. Substracting all those categories disposed of by incineration from the total of APIs and excipients weighed in, results in the worst case estimate of losses to wastewater from solida production (Hoerger et al., 2009). Parenteralia for Roche nearly exclusively means biologics, i.e., monoclonal antibodies and other biotechnologically produced protein APIs, usually of very high molecular mass 450,000 Da. As doses are low, galenical batch sizes are comparatively small, up to 100 L per batch for the two sites, moreover, most of the total consists of ultrapure water for formulation. Excipients are mainly organic buffers (e.g., arginine, glutamine, glycine, histidine), often with a sugar (e.g., sucrose, trehalose) and a detergent (e.g., polysorbate 80); as an example, see the SDS for MabThera SC (Safety Data Sheet for MabThera SC, 2014). Parenteralia are mostly filled into glass vials, which are then sealed, or into syringes. These containers are inspected visually for cloudiness (signifying either undissolved solids or microbial growth) and all units not passing are collected for incineration; the same holds for the bags with residuals of ready mixed solutions or mixtures not meeting specifications. Only liquid losses during the mixing or the filling process, including possible breakages, are discharged to wastewater. Substracting the incinerated and produced amounts from the initial total results in the worst-case estimate of losses from parenteralia production. 2.2. WWTP and receiving water data, predicted environmental concentrations Basic data for daily wastewater fluxes and treatment for the

29

two WWTPs serving Roche Kaiseraugst and Basle were searched for as well as flow data for the receiving water, which for both WWTPs is the River Rhine. Predicted environmental concentrations (PECs) for the WWTPs were calculated by dividing the worstcase loss of excipients in grams per day by the average total wastewater treated daily in the respective WWTP. Worst-case PECs for the River Rhine were calculated by dividing the loss in g/d by the minimum residual flow or Q347 in Swiss terminology (the long-term lower 5th percentile flow (Swiss Federal Office for the Environment (BAFU), 2015c)) of the River Rhine in a day, i.e., without factoring in removal in the WWTPs through biodegradation or adsorption, hence in this first approximation the WWTPs do not influence the surface water PEC. In addition, as a worst-case scenario, it was assumed that those productions with the highest amounts of the same excipients would take place on the same day, which leads to an inordinately high PEC for those substances. Roche Kaiseraugst discharges the wastewater from galenical production to WWTP Rhein in Pratteln, which treats approximately 15,800 m3 wastewater per day with an overall removal efficiency of 98% regarding biochemical oxygen demand in 5 days (BOD5) and of 90% regarding total organic carbon (TOC) (WWTP Rhein, 2015). Basle galenical production discharges to WWTP Basle, which treats approximately 93,100 m3/d with a removal efficiency of 92.9% regarding BOD5 and 93.0% regarding TOC [data for 2013, Ref. Pro Rheno Betriebs AG, 2013]. The River Rhine is the receiving water for both WWTP effluents. It has a long-term (1891–2013) average flow at the measuring station Rheinhalle Basle of 1051 m3/s (Swiss Federal Office for the Environment (BAFU), 2014) and a minimum residual Q347 flow of 459 m3/s (Swiss Federal Office for the Environment (BAFU), 2014). The only two significant surface water inflows into the Rhine between the two WWTP effluents are the smaller River Birs, which has an average flow of 14.6 m3/s and a monthly minimum of 8.0 m3/s (Swiss Federal Office for the Environment (BAFU), 2015a), and the River Wiese with an average of 9.6 m3/s and a monthly minimum of 4.9 m3/s (Swiss Federal Office for the Environment (BAFU), 2015b). In a low-flow situation for the River Rhine, also the tributaries will carry little water; hence, the contributions from the Birs and Wiese do not significantly change the Rhine flow and will be disregarded for upstream (WWTP Rhein effluent) surface water PEC derivation. 2.3. Environmental data for excipients Environmentally relevant substance data for these substances were researched in company-internal documentation Roche Safety Data Sheets (SDS) database and supplier SDS collection; older hardcopy collections like Verschueren (1996)) as well as in online databases and sources ChemIDPlus, Ref. ChemIDPlus database, 2015; Detergents Ingredients Database, 2014; European Chemicals Agency (ECHA), 2015; Hazardous Substances Data Base (HSDB), 2015; ECOTOX Database, 2015; Environmental Fate Data Base (EFDB), 2015 and on Scholar, 2015. The excipients were usually identified by their Chemical Abstracts Services (CAS) numbers, which was used for further data search. Search criteria included molecular mass, physico-chemical data, ready (OECD 301) (Organisation for Economic Co-Operation and Development (OECD), 2015) or inherent (OECD 302) or model WWTP (OECD 303) biodegradability, algal growth inhibition (OECD 201 test or equivalent), acute (OECD 202) or chronic (OECD 211) daphnid toxicity, acute (OECD 203) or subchronic (OECD 210) to chronic fish toxicity and activated sludge respiration inhibition (OECD 209) (Organisation for Economic Co-Operation and Development (OECD), 2015), where available. In the first step of the ERA, only experimental biological data were used to fill, as completely as possible, a spreadsheet with the relevant informations on the excipients.

Excipient

CAS no.

Wastewater treatment plant

Receiving water (River Rhine)

PEC, mg/L PNEC, mg/L PNEC Ref.

PEC/PNEC ratio PEC, ng/L 5.70E 04 2.60E  03

9005-25-8 64-1-5

3.79E þ1 3.08E þ1

6.70E þ 4 1.20E þ 4

D-Mannitol Calcium phosphate

69-65-8 7757-93-9

2.12E þ 1 2.05E þ1

7.58E þ 4 2.40E þ2

Cellulose Lactose

1.92E þ1 1.63E þ1

Crospovidone Polyvinyl-pyrrolidone

9004-34-6 10039-266 9003-39-8 9003-39-8

Sucrose Talcum

PNEC, ng/L

PNEC Ref.

PEC/PNEC ratio

8.90E þ 01 7.24Eþ 01

1.00E þ 05 9.60E þ05

(Detergents Ingredients Database, 2014) (European Chemicals Agency (ECHA), 2015)

8.90E  04 7.54E  05

4.51E  04 (4.73E  7) 3.00E  04 2.70E 04

NA

2.79E  04 8.50E  02

4.97E þ01 4.82E þ 01

6.70E þ 4 7.59E þ 4

(Zahn and Wellens, 1980) (European Chemicals Agency (ECHA), 2015) [NRT] RA (European Chemicals Agency (ECHA), 2015) (van der et al., 1994) [NRT]

2.90E  04 2.15E  04

4.51E þ01 3.83E þ 01

(1.00E þ 06) (European Chemicals Agency, 2006) 6.00E þ 04 RA (European Chemicals Agency (ECHA), 2015) 1.00E þ 05 (Detergents Ingredients Database, 2014) (8.10E þ 7) (US EPA, 2015)

1.07Eþ1 9.67E þ 0

3.00E þ 3 2.00E þ 5

[NRT] (BASF, 2015)

3.58E  3 4.80E  05

2.52E þ 01 2.27E þ01

8.40E þ04 8.40E þ04

57-50-1

7.59E þ0

1.00E þ6

7.60E  06

6.77E þ 0

1.00E þ4

6.80E  04

1.59E þ 01

NA

5.26E þ0 4.83E þ0

3.00E þ 3 1.00E þ4

1.75E  3 4.80E  04

1.24Eþ 01 1.13E þ 01

NA 6.00E þ 04

Carboxmethyl cellulose

14807-966 9063-38-1 13463-677 9000-11-7

3.54E þ0

1.00E þ5

3.50E  05

8.32E þ 00 5.00E þ 05

Keltose Magnesium stearate

9005-31-6 557-04-0

3.11E þ 0 2.90E þ0

6.50E þ4 8.83E þ4

4.79E  05 3.30E  05

7.31E þ 00 NA 6.81E þ 00 9.00E þ 02

Sorbitol Hydroxypropyl-methyl cellulose Sodium stearyl fumarate

50-70-4 9004-65-3 4070-80-8

2.69E þ0 2.69E þ0 1.40E þ0

1.00E þ6 6.70E þ 4 3.00E þ 3

Sodium phosphate

7558-80-7

1.07Eþ0

2.40E þ2

Benzyl alcohol

100-51-6

9.46E  1

3.90E þ2

Sodium chloride

7647-14-5

7.73E  1

5.00E þ %

Ethylcellulose

9004-57-3

7.52E  1

1.00E þ5

2-Methyl-oxirane, co-polymer with oxirane Stearic acid

9003-11-6

7.52E  1

4.00E þ 5

(Heukelekian and Rand, 1955; Gonzalez et al., 1972) (European Chemicals Agency (ECHA), 2015) [NRT] (European Chemicals Agency (ECHA), 2015) (Detergents Ingredients Database, 2014) [NRT] RA (European Chemicals Agency (ECHA), 2015) (Gonzalez et al., 1972; Straub, 2014) RA (van der et al., 1994) RA (European Chemicals Agency (ECHA), 2015) RA (European Chemicals Agency (ECHA), 2015) (European Chemicals Agency (ECHA), 2015) (European Chemicals Agency (ECHA), 2015) (Detergents Ingredients Database, 2014) (Straub, 2014)

RA (BASF, 2015; Dvořáková et al., 1999) (BASF, 2015; Dvořáková et al., 1999; Heukelekian and Rand, 1955) 3.03E þ 00 (1.00E þ 06) (European Chemicals Agency, 2006)

57-11-4

7.52E  1

8.83E þ4

Sodium dodecyl sulphate

151-21-3

6.45E  1

1.35Eþ3

Sodium citrate

1545-80-1

6.31E  1

1.00E þ6

Silicic acid Polyethylene glycol

4.08E  1 3.22E  1

NA 1.00E þ5

L-Histidine Sodium acetate

7631-86-9 25322-683 71-00-1 6131-90-4

1.99E  1 1.85E  1

1.00E þ4 1.15E þ 5

Succinic acid

110-15-6

1.77E 1

3.00E þ 4

Sodium starch glycolate Titanium dioxide

(European Chemicals Agency (ECHA), 2015) (European Chemicals Agency (ECHA), 2015) RA (European Chemicals Agency (ECHA), 2015) (Clariant, 2011) (Helfgott et al., 1977) RA (European Chemicals Agency (ECHA), 2015) (European Chemicals Agency (ECHA), 2015)

(4.97E  05) 8.03E  04

(3.03E  06)

(European Chemicals Agency (ECHA), 2015)

NA 1.89E  04

(Detergents Ingredients Database, 2014)

1.66E  05 NA 7.56E  03

2.45E  03

RA (European Chemicals Agency (ECHA), 2015) 6.32E þ 00 (1.00E þ 06) (European Chemicals Agency, 2006) 6.31E þ 00 3.76E þ04 (Detergents Ingredients Database, 2014) 3.28E þ 00 2.00E þ 03 RA (European Chemicals Agency (ECHA), 2015) 2.52E þ 00 6.00E þ 04 RA (European Chemicals Agency (ECHA), 2015) 2.22E þ 00 1.00E þ 04 (European Chemicals Agency (ECHA), 2015)

2.22E  04

1.50E  06

1.81E þ 00

3.72E þ 07

(European Chemicals Agency (ECHA), 2015)

4.88E  08

7.50E  06

1.77Eþ 00

1.00E þ 05

RA (Detergents Ingredients Database, 2014)

1.77E 05

1.90E  06

1.77Eþ 00

1.00E þ 05

(Clariant, 2011)

1.77E 05

8.52E  06

1.77Eþ 00

9.00E þ 02

(European Chemicals Agency (ECHA), 2015)

1.96E  03

4.77E 04

1.51E þ 00

4.00E þ 03

(European Chemicals Agency (ECHA), 2015)

3.78E  04

6.30E  07

1.48E þ 00

1.60E þ06

9.26E  07

NA 3.20E  06

9.58E  01 4.22E þ04 7.56E  01 1.00E þ 05

RA (European Chemicals Agency (ECHA), 2015) (European Chemicals Agency (ECHA), 2015) (Chemie, 2014)

1.99E  05 1.60E  06

4.67E 01 4.33E 01

(4.35E þ 4) 1.00E þ 05

5.90E  06

7.06E  02

4.07E þ 04

2.70E 06 4.00E  05 4.70E 04 4.50E  03

(US EPA, 2015) RA (European Chemicals Agency (ECHA), 2015) (European Chemicals Agency (ECHA), 2015)

(6.32E  06) 1.68E  04 1.64E  03 4.20E  05

2.27E  05 7.56E  06 (1.07E 5) 4.33E 06 1.73E  06

K.C. Wirz et al. / Sustainable Chemistry and Pharmacy 1 (2015) 28–35

Starch Ethanol

30

Table 1 Excipients used at Roche Basle and Kaiseraugst in 2013/2014 ordered by WWTP PEC, with their maximum PECs for single productions for the respective WWTP and the River Rhine, the corresponding PNECs with the references used for their derivation, and the PEC/PNEC risk characterisation ratios.

6.23E  06 (ECOTOX Database, 2015; Häner, 2006) 5.04E  02 8.10E þ03

1.61E þ 06 7.56E  02 2.8E  07

1.8E  06

4.70E 08

1.45E  07

RA (European Chemicals Agency (ECHA), 2015) (European Chemicals Agency (ECHA), 2015) 1.18E þ06 1.71E  01 4.5E  06 1.62E þ 4

1.15E þ 5

1.22E þ 4

7.28E  2

3.22E  2

2.15E  2

631-61-8

64-19-7

577-11-7

Acetic acid

Sodium dioctyl sulfosuccinate

RA (European Chemicals Agency (ECHA), 2015) (European Chemicals Agency (ECHA), 2015) (European Chemicals Agency (ECHA), 2015)

1.77E 01 1.00E þ3 7.52E  2

Polyoxyethylene–sorbitan– monolaurate Ammonium acetate

31

Quantitative structure–activity relationship (QSAR)-modelled data using EPISuite (US EPA, 2015) for acute ecotoxicity were only used in a second step for insufficiently documented, simple compounds; however, no QSAR data were derived for polymers. 2.4. Predicted no-effect concentrations

9005-64-5

(Clariant, 2011)

7.50E  05

6.30E þ04

(Clariant, 2011)

2.80E  06

K.C. Wirz et al. / Sustainable Chemistry and Pharmacy 1 (2015) 28–35

This spreadsheet served as a basis for a general overview, but also for deriving predicted no-effect concentrations (PNECs) for WWTPs and receiving (fresh) waters. PNECs were calculated according to the EU Technical Guidance Document (European Commission, 2003) as follows. For WWTP PNECs, the activated sludge respiration inhibition (OECD 209) (Organisation for Economic CoOperation and Development (OECD), 2015) 50% effect concentration (EC50) was divided by an assessment factor of 100 to arrive at the PNEC; where a no observed effect concentrations (NOEC) was available, this was divided by an assessment factor of 10; the NOEC-based PNEC was given preference over the EC50-based. Alternatively, if no activated sludge respiration inhibition test data were available, but confirmation of ready or inherent biodegradability instead, one-tenth of the biodegradation test concentration was used as a surrogate WWTP PNEC. In a few select cases, where no biodegradation or activated sludge respiration inhibition information was found for common excipients used in higher amounts, inherent biodegradation tests were performed in the Roche Basle environmental laboratory according to internal protocols, which correspond to a Zahn–Wellens test (based on OECD 302B) or a manometric respirometry test (OECD 302C) (Organisation for Economic Co-Operation and Development (OECD), 2015). The Zahn–Wellens tests were usually run for 14 days with a test substance concentration corresponding to 300 mg dissolved organic carbon/L and a 1:1 mix of rinsed activated sludge from an industrial and a predominantly municipal WWTP at 1000 mg dry weight/L. If the test substance was inherently degradable in the Zahn Wellens test, it was assumed that the high test substance concentration was not inhibitory to activated sludge. The respirometry tests were run for 28 days with a test substance concentration of 30 mg/L and the same mixed sludge as above at 100 mg dry weight/L. In addition, for the respirometry tests an inhibition/toxicity control was run in parallel with 30 mg test substance/L plus 30 mg sodium benzoate/L; if the biodegradation of sodium benzoate progressed normally in the presence of the test substance, the test substance was assumed not to inhibit the activated sludge. For freshwater PNECs, in the case of three available EC50s from algae and daphnia and fish, the lowest EC50 was divided by an assessment factor of 1000; in case of two (sub)chronic NOECs from two of the three organism groups, the lowest was divided by 50; in case of three NOECs (one from each groups), the lowest was divided by 10. PNECs derived from chronic toxicity results were given preference over acute-based PNECs. In a few cases where insufficient data were available, read-across from closely related substances was used for PNECs, e.g., for acetates from acetic acid, or for ethylcellulose from cellulose. 2.5. Environmental risk assessment The risk characterisation for the two WWTPs and the River Rhine was made by dividing the PECs by the PNECs for the single excipients. PEC/PNEC ratios o1 suggest no significant risk while PEC/PNEC ratios Z1 mean potential risk (European Commission, 2003).

32

K.C. Wirz et al. / Sustainable Chemistry and Pharmacy 1 (2015) 28–35

3. Results 3.1. Excipients used at Roche Basle and Kaiseraugst 2013/2014 In 2013 and the first half of 2014, a total of 15 different drug products were formulated at Roche Basel and Kaiseraugst, several more than once. While Basle produces both solid and liquid dosage forms, Kaiseraugst formulates only parenteralia. Eight of these 15 products were biologics (i.e., parenteralia), the remaining seven were chemically synthesised, small-molecule APIs (i.e., solida). For these galenical productions, 35 different excipients out of the general list above were used at both sites (Table 1). Values in brackets are derived from QSAR ecotoxicity data; in case of D-mannitol, sorbitol and sucrose, an EU reference (European Chemicals Agency, 2006) used QSAR data; for the other substances in brackets, EPISuite (US EPA, 2015) QSAR data were developed to fill in the blanks. Minor discrepancies between PECs, PNECs and PEC/PNEC ratios are due to rounding errors. NA ¼not available; [NRT] ¼new Roche internal tests; RA ¼read-across. 3.2. WWTP and river Rhine PECs 3.2.1. WWTP PECs The WWTP PECs range from 37.9 mg/L for starch to 0.0225 mg/L for sodium dioctyl sulfosuccinate, spanning three orders of magnitude (Table 1); both extremes refer to the Basle WWTP. Also high on the list with WWTP PECs 4 5 mg/L are ethanol (PEC¼ 30.8 mg/ L), mannitol (21.2 mg/L), calcium phosphate (20.5 mg/L), cellulose (19.2 mg/L), lactose (16.3 mg/L), crospovidone (10.7 mg/L), polyvinylpyrrolidone (9.67 mg/L), sucrose (7.59 mg/L), talcum (6.77 mg/L) and sodium starch glycolate (5.26 mg/L). The WWTP PECs show that the highest use excipients are starch and certain sugars, ethanol, cellulose, two polyvinylpyrrolidones and inorganic talcum, while pH regulators/buffers and detergents on the whole are at the lower end. With the exception of sucrose all of these major excipients are exclusively used in solida production. This confirms that (apart from ethanol) bulking agents, diluents, disintegrating aids and lubricants show the biggest masses, as would have been expected. Ethanol in solida is used for wetting dry mixtures and will evaporate during the granulating and tablet-pressing processes or during tablet storage. 3.2.2. River Rhine PECs The worst-case crude PECs for the River Rhine, calculated without removal in a WWTP through either biodegradability or sorption and assuming long-term Q347 low flow conditions, show the same distribution, ranging from starch (89.0 ng/L) to sodium dioctyl sulfosuccinate with 50.4 pg/L (Table 1). The same 11 excipients as above, plus titanium dioxide, have PECs 410 ng/L while 9 compounds have PECs o1 ng/L. The PECs derived here are worst-case in the sense that both all unaccounted-for amounts are assumed to be discharged to wastewater, that no biodegradation in the WWTP takes place for any of them and that the receiving water dilution is the Q347 low flow. Regarding losses from production, those solida productions involving so-called ‘high-potency’ APIs have the first rinse of the mixers separated and sent to incineration as a Standard Operating Procedure; the same applies to a few other solida productions of non-biodegradable APIs formulated in higher amounts [Rocheinternal data]. Hence, a major part of residual APIs and excipients from certain productions will not be discharged to wastewater. The assumption of no removal in WWTPs is not supported by the available data, either. Out of the organic high WWTP PEC compounds (PEC 4 5 mg/L), starch is readily biodegradable (Zahn and Wellens, 1980), ethanol is readily biodegradable beside being stripped during aeration [Roche-internal data], D-mannitol and

lactose are well inherently degradable in new Roche-internal tests, cellulose is readily biodegradable (van der et al., 1994), sucrose is readily biodegradable (Gonzalez et al., 1972) or at least ‘biodegradable’ (Straub, 2014), and sodium starch glycolate is ‘biodegradable’ (Straub, 2014) and proved to be inherently biodegradable in a new Roche-internal test. While polyvinylpyrrolidone is listed as inherently biodegradable in the EU Detergents Ingredients Database (Detergents Ingredients Database, 2014), both crospovidone and polyvinylpyrrolidone were not inherently biodegradable in the new Roche-internal tests. Last, the 1891–2013 average Q347 (2013 being the most recent Q347 available) River Rhine flow of 459 m3/s (Swiss Federal Office for the Environment (BAFU), 2014) is a very conservative assumption compared to the average flow of the whole 1891–2013 period of 1051 m3/s (Swiss Federal Office for the Environment (BAFU), 2014). All three arguments suggest that the WWTP and particularly the River Rhine PECs are decidedly on the high side. Hence, the highest WWTP PEC of 37.9 mg/L and the corresponding River Rhine PEC (without removal) of 89.0 ng/L, both for starch, are clearly worst-case figures. Using the long-term average flow of 1051 m3/s (Swiss Federal Office for the Environment (BAFU), 2014) for dilution, even without considering removal in the WWTPs, the highest River Rhine PEC for starch decreases to 38.9 ng/L. In parallel, only seven out of the original 12 excipients remain with a PEC 410 ng/L, while there are 16 (formerly nine) excipients with PECs o1 ng/L (data not shown). Out of these seven excipients, one is inorganic (calcium phosphate), while the removal-refined PECs of five (starch, ethanol, D-mannitol, cellulose and lactose) are predicted to decrease below 10 ng/L, and only crospovidone, being non-degradable, would maintain its average-flow PEC of 11.0 ng/L (data not shown). 3.3. Excipient PNECs 3.3.1. WWTP PNECs All excipients from galenical productions show WWTP PNECs ranging from 240 mg/L (calcium phosphate, read-across from phosphoric acid (European Chemicals Agency (ECHA), 2015) to 1,000,000 mg/L (sodium citrate, read-across from citric acid (European Chemicals Agency (ECHA), 2015 Sorbitol and sucrose both have their PNEC based on Straub (2014), while the PNEC for sodium stearyl fumarate is read-across from the lower PNEC of stearyl alcohol and fumaric acid (European Chemicals Agency (ECHA), 2015). Seven (starch, ethanol, mannitol, cellulose, lactose, L-histidine, sodium starch glycolate) of the remaining excipients are biodegradable [25, 27, 28, 31–34, 36, new Roche-internal tests], hence one-tenth of the biodegradability testing concentration was used as a substitute WWTP PNEC (Table 1). Crospovidone was not degradable in a new Roche-internal manometric respirometry test; however, there was no inhibition of co-substrate biodegradation, hence 3 mg crospovidone/L was used as a WWTP PNEC (Table 1). Hydroxypropylmethylcellulose is listed as non-biodegradable but showing 100% removal in the Detergents Ingredients Database (Detergents Ingredients Database, 2014); this is interpreted to mean that the substance was not toxic to activated sludge, but adsorbed to it, hence the PNEC from cellulose (67,000 mg/L (van der et al., 1994)) was read-across. No direct or read-across WWTP PNEC could be derived for silicic acid; however, in the crystal form used silicic acid is an inert substance and is not expected to show toxicity to bacteria. 3.3.2. Surface water PNECs The freshwater PNECs of the excipients range from 37,200 mg/L for sodium chloride (European Chemicals Agency (ECHA), 2015) to 0.9 mg/L for stearic acid (European Chemicals Agency (ECHA), 2015). In the following cases, the PNECs were derived by read-

K.C. Wirz et al. / Sustainable Chemistry and Pharmacy 1 (2015) 28–35

across: calcium phosphate and sodium phosphate from phosphoric acid, crospovidone from polyvinylpyrrolidone, magnesium stearate from stearic acid, sodium acetate from acetic acid, and ethylcellulose from cellulose. For hydroxypropylmethyl cellulose, the PNEC of 0.037 mg/L from the EU Detergents Ingredients Database (Detergents Ingredients Database, 2014) was originally derived from only two ecotoxicity data using an increased assessment factor of 5000; note that reading across from cellulose would give a PNEC of 0.1 mg/L, nearly three times higher. The PNECs for D-mannitol, sorbitol and sucrose were derived from an EU document (European Chemicals Agency, 2006) that used QSAR ecotoxicity data, see Section 4. For lactose and L-histidine, the lowest acute QSAR ecotoxicity data for algae, daphnia and fish predicted by EPISuite (US EPA, 2015) were used to derive a substitute PNEC. Thus, three excipients out of the 35 have no PNEC due to the unavailability of sufficient ecotoxicity data or the high uncertainty of modelling them, the inorganic magnesium silicate talcum and the organic polymers sodium starch glycolate and keltose. Surface water PNECs could be derived for 32 of the 35 excipients, but not for the remaining three (in order of PEC, talcum, sodium starch glycolate and keltose), due to lack of available ecotoxicity data. While none of those three data-poor excipients are suspected to exert any specific toxicity or to be a risk for receiving waters, it still means that no formal ERA is possible but that an assessment based on plausibility considerations must be developed. Talcum is a nearly insoluble inorganic magnesium silicate, sodium starch glycolate is a naturally extracted product from cereals or potatoes, and keltose is the ammonium salt of alginic acid, a linear block co-polymer of guluronic acid and mannuronic acid of natural algal origin. Hence, all organic compounds are natural or quasi natural products. A QSAR-based PNEC was used for sucrose, D-mannitol and sorbitol (European Chemicals Agency, 2006). In algae, low-molecular-weight carbohydrates including various common natural sugars and sugar alcohols (beside starch) are biosynthesized and function both as organic carbon and energy store, solutes, oxygen scavengers and building blocks for larger cell wall carbohydrates (Raven and Beardall, 2012), with mannitol being specifically mentioned as an example for the kelp Laminaria. To use the sugars, algae have glycolytic enzymes that would allow them to metabolise many different kinds of simple sugars and sugar alcohols. In Daphnia pulex, short-term exposure to lactose caused a reversible change in heart rate; however, the NOEC was a high 20 mM or 6.85 g/L (Campbell et al., 2004); in the same paper, sucrose and galactose had a NOEC of 100 mM each, corresponding to 34.2 g/L and 18.0 g/L, respectively. Fish, as vertebrates, are obviously closer to mammals, even though there is evidence that fish on the whole are relatively insulin- and glucose-resistant [reviewed in Hemre et al. (2002)]. Overall, natural sugar and sugar alcohol compounds are not regarded as inordinately toxic to algae, daphnia or fish. Indeed, the European Chemicals Agency has had the substances for Annex IV of the REACH chemicals legislation reviewed (European Chemicals Agency, 2006); Annex IV contains those substances where ‘sufficient information is known about these substances that they are considered to cause minimum risk because of their intrinsic properties’ (European Commission, 2015). In this review document (European Chemicals Agency, 2006), sucrose and the sugar alcohols mannitol and sorbitol are considered to have sufficient information for inclusion. The document refers to QSAR data for both ready biodegradability and low ecotoxicity, the latter resulting in surface water high PNECs 41 g/L for all three compounds. Similarly, EPISuite QSAR modelling (US EPA, 2015) was used for the simple organics lactose and L-histidine. However, for the inorganic talcum and the organic polymers sodium starch glycolate

33

and keltose, no useful PNEC could be derived. Their respective worst-case surface water PECs in the River Rhine are 15.9, 12.4 and 7.3 ng/L, moreover, both organic polymers proved to be inherently biodegradable in the new Roche-internal tests, decreasing their surface water PECs. All three excipients would need to exert an inordinately high, specific ecotoxicity to have a PNEC of a magnitude of approximately 10 ng/L; this is not expected. 3.4. PEC/PNEC risk characterisation ratios All PEC/PNEC quotients that can be derived for WWTPs and the River Rhine are o0 (Table 1). This means that for these excipients there is no reason to assume any significant risk either to WWTPs or the the receiving water. The highest WWTP risk characterisation ratio of 0.085 is for calcium phosphate, not because of an inordinately high WWTP PEC (20.5 mg/L) but rather because the WWTP PNEC was derived from phosphoric acid by read-across. The other WWTP risk quotients decrease to 2.80E  7 for acetic acid. One substance, silicic acid, has no basic data at all to derive a WWTP PNEC, thus preventing WWTP risk assessment. For the surface waters, the highest risk characterisation ratio is 0.00756 for magnesium stearate; again, this was read-across from stearic acid, which had low ecotoxicity NOECs due to very low water solubility (European Chemicals Agency (ECHA), 2015). The other surface water risk quotients decrease to 4.88E  8 for sodium chloride. Due to lack of experimental or other published data to base a PNEC on, no risk quotient can be derived for talcum, sodium starch glycolate and keltose. 3.5. Risk characterisation ratios for parallel productions In the theoretical worst case of three parallel productions with different APIs, selected for the maximum amounts of excipients, the highest compound losses would be for lactose (WWTP PEC 59.5 mg/L, River Rhine PEC 140 ng/L), cellulose (WWTP 51.0 mg/L, Rhine 120 ng/L) and starch (WWTP 45.6 mg/L, Rhine 107 ng/L); again, the Rhine PECs were derived without considering biodegradation or adsorption in the WWTP. The corresponding risk characterisation ratios for lactose are 0.020 in the WWTP and 4.73E  7 (based on a QSAR-derived PNEC) in the Rhine; for cellulose 0.00076 (WWTP) and 0.0012 (Rhine), and for starch 0.0068 (WWTP) and 0.0011 (Rhine). Based on the available data, even with three comparatively high-loss productions coinciding, no significant risk to the WWTPs or to the River Rhine is apparent. One minor caveat remains for lactose, where no experimental ecotoxicity data are available, but only QSARs, to derive a solid surface water PNEC. 3.6. A broader view Out of the 35 compounds, three are insufficiently documented for surface water PNEC derivation and initial ERA. Extending the data search to a bigger set of 184 former or current excipients used in Roche drug products from Swiss or other production sites (data not shown), presents a bleaker picture. The biggest single categories are sugars and sugar alcohols, detergents and buffers. Sufficient ecotoxicity data allow the derivation of at least an acutebased PNEC for nearly all buffers and detergents and most alcohols, amines, solvents, silicates and inorganic salts. However, these are lacking for two thirds of the sugar and sugar alcohol groups, also, particularly amino acids, peptides, celluloses and oils/fats are poorly documented. As natural substances the latter are not expected to exert significant toxicity, but still, sufficient environmental data for a PNEC were located for only about half of the 184 excipients, all others lack data.

34

K.C. Wirz et al. / Sustainable Chemistry and Pharmacy 1 (2015) 28–35

4. Discussion All risk quotients that can be derived for single excipients have a value well below 1, both for the WWTPs and the River Rhine. The same holds for the assumption of parallel productions. Even summing up all the risk characterisation ratios (except for those substances where no PNEC can be derived) results in a compound risk quotient of still o1 for the WWTPs and the River Rhine. This strongly suggests that excipients lost to wastewater from galenical production at the two Roche sites in Switzerland do not present a risk to the environment. However, pharmaceuticals are not only formulated by one company in Switzerland. This means that the number and data availability of the excipients, but also the parameters of the WWTPs and the receiving waters, may be quite different for other products and countries. As stated above, sufficient environmental data for a PNEC were located for just about half of the extended dataset of 184 excipients. While these compounds should not be a concern from a toxicological point of view, and with some probability also from an ecotoxicological one, the basic data situation is pretty thin. Obviously, more basic environmental data on excipients would improve this situation and, indeed, since the beginning of the European Union Chemicals Programme REACH, many more relevant data have become publicly available through the REACH internet database. Still, even those surface water PNECs that could be derived are not all of the same quality, in the sense that they rely on different basic data. Most PNECs are based three acute ecotoxicity tests with an assessment factor of 1000 applied to the lowest EC50, some on two chronic NOECs with an assessment factor of 50, a few on three chronic NOECs and an assessment factor of 10, while for some excipients either a higher assessment factor is included due to insufficient tests available, or read-across data from related compounds, or QSAR-modelled ecotoxicity data; finally, for a few, no PNEC could be derived. Thus, there is potentially a significant variability in the quality of PNECs and therefore a range of uncertainty, which may limit their direct comparability. The relationship between acute (EC50, LC50) values and chronic NOEC or LOEC values for the same group of organisms has been investigated by several groups over time, e.g., Roex et al., (2000) and Kienzler et al., (2015). In particular for non-polar narcotics and for molar concentrations of the test substance, these authors consistently found overall correlations of the type log(chronic NOEC or LOEC)¼factor  log(EC50 or LC50) – constant, with the factor defining the slope of the regression generally being smaller than 1. Such a slope means that for the same group, on the whole, the relationship between chronic and acute endpoints is less than 10, plus or minus the constant. Similarly, Ahlers et al. (2006) found median acute-to-chronic ratios up to 12.7; however, single ratios for specific substances may be much higher. For unspecifically acting, narcotic substances; however, an acute-based PNEC should be reasonably close to the chronic-based PNEC. For those PNECs derived by read-across or QSAR, again for narcotic substances, ECETOC (ECETOC, 2012) concluded that such derivations are acceptable, moreover, REACH also accepts QSARdeveloped ecotoxicity data for unproblematic substances. As excipients, in view of their functions, are not reactive substances on the whole, but rather acting by unspecific narcosis, the uncertainty arising from comparing acute- and chronic-based PNECs should therefore be reasonably low. However, in view of the disparate database for the excipients, some inherent uncertainty in the PNECs should be acknowledged, even if it cannot be quantified. Assuming reasonable wastewater treatment everywhere, another major difference may be in the local receiving water. The River Rhine is quite a large water body with a correspondingly high dilution factor, but this is not necessarily the case for other sites and receiving waters. Within the Roche Group, average receiving water dilution factors of galenical site WWTP effluents vary

considerably, over approximately three orders of magnitude [Roche unpublished data]. This means that the PECs for the same (theoretical) amount of excipient lost to wastewater at different sites would also vary in the same dimension – and the risk characterisation ration might increase by the same factor. Based on the current set of 35 excipients, stearic acid and its salts or esters might attain surface water risk characterisation ratios of around 1 based on the present data. Hence, the biodegradability (stearic acid is readily biodegradable (European Chemicals Agency (ECHA), 2015) and predicted removal in a WWTP as well as the solubility and potential PNEC limitations deriving from this, should be taken into account in a refined ERA. On a broader scale, the unknown situation at other production sites can only be addressed, and potential risks from excipient losses during galenical production can only be investigated, by site-specific ERA, in particular for those production sites that have a comparatively low receiving water dilution factor.

5. Conclusions An initial ERA was performed for excipient losses from solida and parenteralia galenical production at Roche Basel and Kaiseraugst in Switzerland, which resulted in the overall conclusion of no significant risk, both for single excipients, all excipients added together and for parallel productions. However, for a larger set of excipients, the environmental basic data situation for common compounds was revealed to be unsatisfactory. For the above sites there is no indication of potential negative effects from galenical production: the excipients pose no significant risk as detailed in this work, APIs for solida were risk-assessed some years ago without highlighting environmental problems (Hoerger et al., 2009) and there is good evidence that biologics active ingredients do not present a risk, either (Straub, 2010). Still, in particular for sites with comparatively low receiving water dilution factors, or for other, insufficiently documented excipients, a site-specific ERA would clarify the situation.

Conflict of interest All three authors are full-time employees of F.Hoffmann‐La Roche Ltd in Basle. The data collection and initial ERA was made by KCW during a 6-month work placement in 2014 that was supervised by MS and JOS.

Acknowledgements Many thanks to C Kan und M Pfaundler, BAFU (Swiss Federal Office for the Environment, Berne) for help with hydrology data. The support of the following colleagues at Roche is much appreciated: B Dörr, G Guttenberg, S Leber and S Tuncay (Basle), A Sterchi (Kaiseraugst) and S Degen (Mannheim, Germany). Special thanks to C Acklin and A Minniti of the Roche Basle Environmental Labs for competent internal biodegradation tests with selected excipients. The constructive criticism of two anonymous reviewers has improved the manuscript and is gratefully acknowledged.

References Abwasserreinigungsanlage Rhein (WWTP Rhein), 2015. 〈http://www.ararhein.ch/〉 (accessed 18.02.15.). Ahlers, J., Riedhammer, C., Vogliano, M., Ebert, R.-U., Kühne, R., Schüürmann, G., 2006. Acute to chronic ratios in aquatic toxicity – variation across trophic levels and relationship with chemical structure. Environ. Toxicol. Chem. 25, 2937–2945. BASF, 2015. Safety Data Sheet for Kollidon s30, 2015/01/07, BASF Co., Florham Park,

K.C. Wirz et al. / Sustainable Chemistry and Pharmacy 1 (2015) 28–35 NJ, USA. 〈http://worldaccount.basf.com/wa/NAFTA  en_US/Catalog/Pharma/ doc4/BASF/PRD/30403404/Material%20Safety%20Data%20Sheet-US-EN.pdf?ti tle ¼&asset_type¼ msds/pdf&language ¼EN&validArea¼ US&urn ¼urn:doc umentum:ProductBase_EU:09007af8800fcbff.pdf〉 (accessed 03.03.15.). Campbell, A.K., Wann, K.T., Matthews, S.B., 2004. Lactose causes heart arrhythmia in the water flea Daphnia pulex. Comp. Biochem. Physiol. B139, 225–234. ChemIDPlus database, 2015. Toxnet Toxicology Data Network, US National Library of Medicine, Bethesda MD. 〈http://chem.sis.nlm.nih.gov/chemidplus/che midheavy.jsp〉 (accessed 18.02.15.). Clariant, 2011. Material Safety Data Sheet for Polyglycol 6000 PF, 10/22/2011, Clariant Co., Charlotte, NC, USA. 〈http://www.hbchemical.com/wp-content/up loads/2014/04/PEG-6000-PF.pdf〉. (accessed 03.03.15.). Detergents Ingredients Database, 2014. Part A, List of ingredients, version 2014.1. 〈http://ec.europa.eu/environment/ecolabel/documents/did_list/didlist_part_a_ en.pdf〉 (accessed 19.02.15.). Dvořáková, D., Rojíčková-Padrtová, R., Maršálek, B., 1999. Is the replacement of Scenedesmus quadricauda L. by Selenastrum capricornutum L. in toxicity tests reasonable? Pol. Arch. Hydrobiol. 46, 345–352. ECETOC, 2012. Category approaches, Read-across, (Q)SAR, ECETOC Technical Report 116, 183 pp. Available from www.ecetoc.org. ECOTOX Database, 2015. US Environmental Protection Agency, Washington DC. 〈http://cfpub.epa.gov/ecotox/quick_query.htm〉. (accessed 18.02.15.). Environmental Fate Data Base (EFDB), 2015. Syracuse Research Corporation, North Syracuse NY, on behalf of the US Environmental Protection Agency. 〈http://esc. syrres.com/fatepointer/search.asp〉 (accessed 18.02.15.). European Chemicals Agency (ECHA), 2015. Helsinki, Finland, 〈http://www.echa. europa.eu/〉 (accessed 18.02.15.). European Chemicals Agency, 2015. Review of Annex IV of Regulation (EC) no. 1907/ 2006 (REACH), Evaluation of existing entries in Annex IV, Prepared by DHI, Milieu Ltd. and Protection Through Knowledge (PKT) Ltd., pp. 1–159, 〈http://ec. europa.eu/environment/chemicals/reach/pdf/6b_appendix_2.pdf〉 (accessed 02.03.15.). European Commission, 2003. Technical Guidance Document on Risk Assessment in Support of Commission Directive 93/67/EEC on Risk Assessment for New Notified Substances, Commission Regulation (EC) No 1488/94 on Risk Assessment for Existing Substances, Directive 98/8/EC of the European Parliament and of the Council concerning the Placing of Biocidal Products on the Market. European Union: Brussels, Belgium, 2003. 〈http://ihcp.jrc.ec.europa.eu/our_activ ities/health-env/risk_assessment_of_Biocides/doc/tgd/〉 (accessed 13.11.14.). European Commission, Environment, Commission Review of REACH Annexes, 2015. 〈http://ec.europa.eu/environment/chemicals/reach/reviews_en.htm〉 (accessed 02.03.15.). Gonzalez, C.F., Taber, W.A., Zeitoun, M.A., 1972. Biodegradation of ethylene glycol by a salt-requiring bacterium. Appl. Microbiol. 24, 911–919. GoogleScholar, 2015. http://scholar.google.com/ (accessed 18.02.15.). Häner, A., 2006. Dioctylsulfosuccinate, sodium salt; freshwater algal growth inhibition test with Desmodesmus subspicatus, BMG Engineering GmbH, Schlieren, Switzerland, test performed on behalf of F.Hoffmann-La Roche. Hazardous Substances Data Base (HSDB), 2015. Toxnet Toxicology Data Network, US National Library of Medicine, Bethesda, MD. 〈http://www.toxnet.nlm.nih.gov/ cgi-bin/sis/htmlgen?HSDB〉 (accessed 18.02.15.). HebroChemie, 2014. Sicherheitsdatenblatt F134–B21 hebrosgrind 1400 B, 12.09.2014, Hebro Chemie, Mönchengladbach, Germany. 〈http://www.hebrochemie.de/sdb/DE/DE/540142170.DE.DE.pdf〉 (accessed 03.03.15.). Helfgott, T.B., Hart, F.L., Bedard, R.G., 1977. An index of refractory organics. Office of Research and Development, US Environmental Protection Agency, Ada, OK, USA, Report EPA–600/2–77–174, 1977. Hemre, G.-I., Mommsen, T.P., Krogdahl, Å., 2002. Carbohydrates in fish nutrition: effects on growth, glucose metabolism and hepatic enzymes. Aquac. Nutr. 8, 175–194. Heukelekian, H., Rand, M.C., 1955. Biochemical oxygen demand of pure organic compounds. Sew. Ind. Wastes 27, 1040–1053. Hoerger, C.C., Dörr, B., Schlienger, C., Straub, J.O., 2009. Environmental risk assessment for the galenical formulation of solid medicinal products at Roche Basle, Switzerland. Integr. Environ. Assess. Manag. 5, 331–337. Raven, J.A., Beardall, J., 2012. Carbohydrate metabolism and respiration in algae. In:

35

Larkum, A.W.D., Douglas, S.E., Raven, J.A. (Eds.), Photosynthesis in Algae: Advances in Photosynthesis and Respiration vol. 14. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 205–224, chapter 10. Swarbrick, J., 2015. Encyclopedia of pharmaceutical technology, Third ed. Informa Healthcare, London, UK, Subscription only e-book. Kemsley, J., 2014. Triclosan under the microscope. Chem. Eng. News 92 (25), 10–13. Kienzler, A., Halder, M., Worth, A. 2015. Scientific approaches to support waiving of chronic fish tests: possible use of acute-to-chronic relationships and interspecies correlations. Poster. In: Proceedings of the SETAC Europe 29th Annual Meeting Barcelona. May 3rd–7th. Organisation for Economic Co-Operation and Development (OECD), 2015. OECD Test Guidelines, OECD, Paris, France. 〈http://www.oecd-ilibrary.org/〉 (accessed 18.02.15.). Pro Rheno Betriebs AG (WWTP Basle), Jahresbericht, 2013. 〈http://www.prorheno. ch/__temp/ProRheno_Jahresbericht2013Versand_rev2_definitiv.pdf〉 (accessed 18.02.15.). Roex, E.W.M., van Geestel, C.-A.-M.-, van Wezel, A.P., van Straalen, N.M., 2000. Ratios between acute toxicity and effects on population growth rates in relation to toxicant mode of action. Environ. Toxicol. Chem. 19, 685–693. Safety Data Sheet for MabThera SC, F.Hoffmann-La Roche Ltd., Basle, Switzerland, 2014. 〈http://www.roche.com/sustainability/for_communities_and_environ ment/environment/safety_data_sheets-row.htm, search term¼ MabThera SC〉 (accessed 19.02.15.). Silva, A., Santos, L.H.M.L.M., Delerue-Matos, C., Figueiredo, S.A., 2014. Impact of excipients in the chronic toxicity of fluoxetine on the alga Chlorella vulgaris. Environ. Technol. 35 (24), 3124–3129. http://dx.doi.org/10.1080/09593330. 2014.932438. Straub, J.O., 2002. Environmental risk assessment for new human pharmaceuticals in the European Union according to the draft guideline/discussion paper of January 2001. Toxicol. Lett. 135, 231–237. Straub, J.O., 2010. Protein and peptide therapeuticals: an example of ‘benign by nature’ active pharmaceutical ingredients. In: Kümmerer, K., Hempel, M. (Eds.), Green and Sustainable Pharmacy. Springer, Heidelberg, Germany, pp. 127–134. Straub, J.O., Shearer, R., Studer, M., 2014. Rational selection of alternative, environmentally compatible surfactants for biotechnological production of pharmaceuticals – a step toward Green Biotechnology. Environ. Toxicol. Chem. 33, 2140–2146. Swiss Federal Office for the Environment (BAFU), 2015a. Hydrodaten Birs (Münchenstein), BAFU, Berne, Switzerland. 〈http://www.hydrodaten.admin.ch/ lhg/sdi/jahrestabellen/2106Q_14.pdf〉 (accessed 18.02.15.). Swiss Federal Office for the Environment (BAFU), 2014. Hydrodaten Rhein Rheinhalle (Basel), BAFU, Berne, Switzerland. 〈http://www.hydrodaten.admin.ch/lhg/ sdi/jahrestabellen/2289Q_13.pdf〉 (accessed 18.02.15.). Swiss Federal Office for the Environment (BAFU), 2015b. Hydrodaten Wiese (Basel), BAFU, Berne, Switzerland, 〈http://www.hydrodaten.admin.ch/lhg/sdi/jahresta bellen/2199Q_14.pdf〉 (accessed 18.02.15.). Swiss Federal Office for the Environment (BAFU), 2015c. Information systems and methods: Q347 flow rate. 〈http://www.bafu.admin.ch/hydrologie/01835/02108/ index.html?lang ¼en〉 (accessed 27.02.15.). United States Pharmacopeia, 2014. USP and NF excipients, listed by functional category; 1st supplement to USP 37–NF 32. 〈http://www.uspnf.com/uspnf/pub/ index?usp ¼ 37&nf¼ 32&s ¼1&officialOn ¼August%201,%202014〉 (accessed 08.09.14.). US EPA, 2015. Estimation Programs Interface Suite™ (EPISuite) for Microsofts Windows, v 4.11, United States Environmental Protection Agency, Washington, DC, USA. 〈http://www.epa.gov/opptintr/exposure/pubs/episuite.htm〉 (accessed 03.03.15.). van der, M., M, Zee, Sijtsma, L., Tan, G.B., Tournois, H., de Wit, D., 1994. Assessment of biodegradation of water insoluble polymeric materials in aerobic and anaerobic aquatic environments. Chemosphere 28, 1757–1771. Verschueren, K., 1996. Handbook of Environmental Data on Organic Chemicals, Third ed. Van Nostrand Reinhold, New York. Zahn, R., Wellens, H., 1980. Prüfung der biologischen Abbaubarkeit im Standversuch – weitere Erfahrungen und neue Einsatzmöglichkeiten. Z. Wasser Abwasser Forsch. 13, 1–7.

Suggest Documents