antimicrobials in wastewater treatment27735/eth-27735-02.pdfmakrolid- und trimethoprim-frachten...

Research Collection

Doctoral Thesis

Occurrence and fate of sulfonamide and macrolideantimicrobials in wastewater treatment

Author(s): Göbel, Anke

Publication Date: 2004

Permanent Link: https://doi.org/10.3929/ethz-a-004930749

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.

ETH Library

https://doi.org/10.3929/ethz-a-004930749

http://rightsstatements.org/page/InC-NC/1.0/

https://www.research-collection.ethz.ch

https://www.research-collection.ethz.ch/terms-of-use

Diss.ETHNo. 15703

Occurrence and Fate of

Sulfonamide and Macrolide Antimicrobials

in Wastewater Treatment

A dissertation submitted to the

SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH

for the degree of

Doctor of Science

presented by

ANKE GÖBEL

Lebensmittelchemikerin

University of Bonn

born on March 14, 1975

citizen of Germany

accepted on recommendation of

Prof. Dr. Walter Giger, examiner

Dr. Christa S. MeArdell, co-examiner

Prof. Dr. Hansruedi Siegrist, co-examiner

PD Dr. habil. Thomas A. Ternes, co-examiner

Zürich, 2004

I

Table of Contents

Summary V

Zusammenfassung VII

Abbreviations IX

1 Introduction 13

1.1 Pharmaceuticals in the Environment 15

1.2 Selection of Antimicrobials 18

1.3 Research Framework 21

1.4 Scope of this Study 23

1.5 Literature cited 25

2 Analytical Method for Wastewater 29

2.1 Introduction 31

2.2 Experimental Section 34

2.2.1 Chemicals and Reagents 34

2.2.2 Internal Standards 35

2.2.3 Sample Collection and Preparation 36

2.2.4 Liquid Chromatography 37

2.2.5 Tandem Mass Spectrometry 37

2.2.6 Method Validation 39

2.2.7 Identification and Quantification 40

2.3 Results and Discussion 41

2.3.1 Method Development 41

2.3.2 Method Validation 45

2.3.3 Wastewater Application 48

2.4 Conclusions 50

2.5 Literature cited 51

3 Analytical Method for Sewage Sludge 55

3.1 Introduction 57

3.2 Experimental Section 58

3.2.1 Chemicals and Reagents 58

3.2.2 Sample Collection 60

3.2.3 Sample Preparation 60

II

3.2.4 Chemical Analysis

3.2.5 Method Development

3.2.6 Method Validation

3.3 Results and Discussion



3.3.3 Application to Sewage Sludge Sampl3.4 Conclusions

3.5 Acknowledgments

3.6 Literature cited

4 Occurrence in Wastewater Treatment

4.1 Introduction

4.2 Experimental Section

4.2.1 Wastewater Treatment Plants

4.2.2 Sample Collection

4.2.3 Analytical Methods

4.2.4 Estimation of Sorption Constants

4.2.5 Calculation of Loads


4.3.1 Occurrence in Wastewater Samples4.3.2 Daily Variations

4.3.3 Seasonal Differences

4.3.4 Occurrence in Sewage Sludge4.3.5 Sorption to Sewage Sludge

4.3.6 Mass Balances

4.4 Conclusions


5 Behavior in Wastewater Treatment

5.1 Introduction




5.2.3 Chemical Analysis

5.2.4 Calculated Elimination

m


5.3.1 Primary Treatment

5.3.2 Secondary Treatment

5.3.3 Solid Retention Time

5.3.4 Substrate Dependencies

5.3.5 Anaerobic Compartment

5.3.6 Wastewater Temperature

5.3.7 Hydraulic Retention Time

5.3.8 Sand Filtration

5.4 Conclusions


6 Ozonation

6.1 Introduction


6.2.1 Ozonation Pilot Plant

6.2.2 Wastewater Effluents

6.2.3 Spiking of Wastewater Effluents

6.2.4 Sample Collection and Chemical Analysis

6.2.5 Calculation of Relative Residual



6.3.2 Oxidation Efficiencies

6.3.3 Influence of the Wastewater Matrix

6.3.4 Comparison of Spiked and Non-Spiked Sampl

6.3.5 Conclusions

6.4 References cited

7 Conclusions and Outlook

Seite Leer /Blank leaf

V

Summary

Antimicrobials are used worldwide in human medicine for the treatment of

infections. Their environmental occurrence and fate is of particular interest, due

to the potential spread and maintenance of antibacterial resistance. In

Switzerland, the annual human consumption of sulfonamides and macrolides,

two major groups of antimicrobials, amounts to approximately 6 and 4 t/a,

respectively. After consumption, human-use pharmaceuticals mainly enter

municipal wastewater treatment as the unchanged substances or as human

metabolites, and can eventually reach receiving surface waters. Their behavior

in wastewater treatment, as well as the efficiencies of distinct treatment steps

and technologies are largely unknown.

In this study, the occurrence, behavior and fate of sulfonamides, macrolides and

trimethoprim in wastewater treatment was investigated. Field studies were

performed in full-scale and pilot wastewater treatment plants, focusing on

activated sludge treatment, a fixed-bed reactor and a membrane bioreactor,

operated at three different solid retention times (SRT). In addition, the ozonation

of wastewater effluents was investigated as a procedure to reduce the loads of

antimicrobials entering the aquatic environment.

Analytical methods were developed and validated for the determination of

sulfonamides, macrolides and trimethoprim in aqueous and solid samples.

Pressurized liquid extraction was used for the extraction of activated and

digested sewage sludge. Solid-phase extraction and reversed-phase liquid

chromatography was applied to diluted sludge extracts and water samples. For

the detection and quantitative determination of the analytes, tandem mass

spectrometry with electrospray positive ionization in the multiple reaction mode

was used. Limits of quantification ranged between 3 and 214 ng/L in primary

effluent and between 1 and 23 ng/L in secondary and tertiary effluent. For

activated sludge samples they varied between 3 and 41 u,g/kg dry weight.

Relative recoveries were generally above 80% and the combined measurement

uncertainty ranged between 2.4 and 16% for aqueous samples.

Mass balances were performed, including sorption to sewage sludge, to assess

the elimination of sulfonamides, macrolides and trimethoprim in wastewater

treatment. In accordance to consumption data, clarithromycin and sulfa¬

methoxazole were the most predominant macrolide and sulfonamide,

respectively, found in Swiss wastewater. In the case of sulfamethoxazole it

VI

proved to be crucial to include the amount present as the main human

metabolite, A^-acetylsulfamethoxazole. No significant elimination was observed

for all investigated compounds in primary treatment. During secondary

treatment, the elimination observed depended on the treatment technology

investigated. Overall, sorption to sludge is of minor importance for all

compounds investigated. The predicted sorption constants for activated sludge

ranged below 500 L/kg, which results in an elimination of below 10%, assuming

a sludge production of 0.2 g/L.

Different treatment technologies were investigated for secondary wastewater

treatment concerning the elimination of the selected compounds. Similar results

were obtained in two conventional activated sludge systems and a fixed-bed

reactor. While no significant removal was observed for trimethoprim,

sulfamethoxazole, including the amount present as A^-acetylsulfamethoxazole,was eliminated by about 60%. For the macrolides investigated the results varied

irregularly between sampling campaigns in the two conventional activated

sludge systems and the fixed-bed reactor, with an elimination of up to 55% in

one single case. Approximately 80% of the total sulfamethoxazole load was

eliminated in the membrane bioreactor, independently of the SRT. Macrolides

and trimethoprim were generally eliminated by 25-50% at a SRT of 16 ± 2 and

33 ± 3 days in the membrane bioreactor. Significantly higher elimination of up

to 90% was observed at a SRT of 60 - 80 days for these compounds. High SRT

and resulting low substrate loading therefore seem to have an influence on the

diversity of the microbial population and consequently on the multitude of

degradation pathways being expressed. Elimination in tertiary treatment was

only observed for macrolides and trimethoprim in one of the sand filters

investigated and appeared to be oxygen limited.

Ozonation proved to be efficient for removing sulfonamides, macrolides and

trimethoprim from wastewater effluents. An ozone dose of 2 mg/L resulted in a

reduction of by over 90%, showing no dependence on the amount of suspended

solids present.

The presented study provides a representative example for the investigation of

human-use pharmaceuticals in wastewater treatment, necessary for a more

profound environmental risk assessment of these compounds. The results from

different treatment conditions and technologies show the significance of various

parameters for the elimination of sulfonamides, macrolides and trimethoprim in

wastewater treatment.

VII

Zusammenfassung

Antimikrobielle Wirkstoffe werden weltweit in der Humanmedizin zur

Behandlung von Infektionskrankheiten verwendet. Ihr Vorkommen und

Verhalten in der Umwelt ist von besonderem Interesse, da ein Zusammenhang

mit der Verbreitung und dem Erhalt von Antibiotikaresistenzen nicht aus¬

geschlossen werden kann. Der jährliche Verbrauch an Makroliden und Sulfon¬

amiden, zwei in der Humanmedizin wichtigen Antibiotikaklassen, beläuft sich

in der Schweiz auf ungefähr 6 bzw. 4 Tonnen pro Jahr. Arzneimittel gelangen

nach der Einnahme durch den Patienten als unveränderte Substanz oder in Form

von Metaboliten in die Abwasserreinigung und können auf diesem Weg

schließlich auch in die Oberflächengewässer gelangen. Ihr Verhalten in der

Abwasserreinigung sowie die Eliminationsleistung einzelner Behandlungsstufen

und verschiedener Reinigungstechnologien ist weitgehend unbekannt.

Vorkommen, Verhalten und Schicksal von Sulfonamiden, Makroliden und

Trimethoprim in der Abwasserreinigung war daher das Thema dieser Arbeit.

Dazu wurden Feldstudien in bestehenden Kläranlagen sowie in Pilotanlagen

durchgeführt mit dem Ziel, Belebtschlammverfahren, Biofilter sowie einen

Membranbioreaktor bei drei verschiedenen Schlammaltern zu vergleichen.

Zusätzlich untersucht wurde die Ozonung von Abwasser, als ein mögliches

Verfahren zur weiteren Verringerung der Umweltbelastung durch Antibiotika.

Analytische Methoden für die quantitative Bestimmung in Abwasser und

Klärschlamm wurden entwickelt und validiert. Belebt- und Faulschlammproben

wurden unter erhöhtem Druck und erhöhter Temperatur extrahiert. Die

verdünnten Extrakte und Abwasserproben wurden mittels Festphasenextraktion

und Flüssigchromatographie analysiert. Für die Detektion der Analyten wurde

Tandemmassenspektrometrie mit positiver Elektrospray-Ionisation verwendet.

Die Bestimmungsgrenzen bewegten sich zwischen 3 und 214 ng/L für Zuläufe

und zwischen 1 und 23 ng/L für Abläufe. Im Belebtschlamm lagen sie zwischen

3 und 41 u.g/kg Trockengewicht. Relative Wiederfindungsraten waren im

Normalfall größer als 80%, und die kombinierte Messunsicherheit betrug 2.4 bis

16% für Abwasserproben.

Mit Hilfe von Massenbilanzen wurde die Elimination der ausgewählten

Substanzen in der Abwasserreinigung untersucht. In Übereinstimmung mit den

vorliegenden Verbrauchszahlen, wurden Sulfamethoxazol und Clarithromycin

als Hauptvertreter der Sulfonamide, bzw. Makrolide in schweizerischem

VIII

Abwasser identifiziert. Für Sulfamethoxazol stellte sich das Einbeziehen des

mengenmäßig wichtigsten menschlichen Metaboliten A^-Acetylsulfa-methoxazol als unerlässlich heraus.

Keine signifikante Elimination der untersuchten Substanzen konnte in der

Vorklärung festgestellt werden. Unterschiedliche Eliminationsleistungen wurden

für die verschiedenen biologischen Abwasserreinigungsverfahren

beobachtet. Für alle untersuchten Substanzen stellte sich die Sorption an

Schlamm als unbedeutend heraus. Für die abgeschätzten Sorptionskonstanten

von unter 500 L/kg Überschussschlamm, liegt diese unter 10% bei einer

Schlammproduktion von 0.2 g/L. Vergleichbare Eliminationen wurde in zwei

Belebtschlammsystemen und dem Biofilter ermittelt. Im Fall von Trimethoprim

konnte kein signifikanter Abbau festgestellt werden, während Sulfamethoxazol

inklusive A^-Acetylsulfamethoxazol zu ca. 60% eliminiert wurden. Für die

Makrolide wurden stark schwankende Eliminationsleistungen in den

verschiedenen Probenahmen für diese drei Systeme ermittelt, mit einzelnen

Höchstwerten von 55%. Im Membranbioreaktor wurde unabhängig vom

Schlammalter eine Elimination von ca. 80% für die Gesamtmenge an

Sulfamethoxazol beobachtet. Makrolid- und Trimethoprim-Frachten wurden

zwischen 25 und 50% im Membranbioreaktor reduziert bei Schlammaltern von

16 ± 2 und 33 ± 3 Tagen, während signifikant höhere Eliminationsleistungen

von bis zu 90% bei 60 - 80 Tagen für diese Substanzen beobachtet wurden. Ein

erhöhtes Schlammalter und die daraus resultierende reduzierte

Schlammbelastung scheinen einen Einfluss auf die mikrobielle Vielfalt und

damit auf die Anzahl der möglichen Abbauwege zu haben. Eine Elimination der

untersuchten Substanzen während der abschließenden Filtration wurde nur für

Makrolide und Trimethoprim und nur auf einem der untersuchten Sandfilter

beobachtet. Die Elimination scheint durch Sauerstoff limitiert zu sein.

Die Ozonung von Abwasser hat sich als wirkungsvolles Verfahren für die

Oxidation von Sulfonamiden, Makroliden und Trimethoprim herausgestellt.

Bereits bei Ozongehalten von 2 mg/L wurde eine Elimination von über 90%

festgestellt, unabhängig von der Menge an Schwebstoffen im Abwasser.

In dieser Arbeit wird exemplarisch die Untersuchung von Humanarzneimitteln

in der Abwasserreinigung gezeigt, deren Ergebnisse zu einer umfassenderen

Risikobeurteilung dieser Substanzen beiträgt. Weiterhin konnte der Einfluss

verschiedener Faktoren auf die Elimination von Sulfonamiden, Makroliden und

Trimethoprim in der Abwasserreinigung aufgezeigt werden.

IX

Abbreviations

1°EFFL primary effluent

2°EFFL secondary effluent

3°EFFL tertiary effluent

13C6SMZ sulfamethazine-phenyl-"CtfASE accelerated solvent extraction

AZI azithromycin

BOD5 biological oxygen demand in 5 days

v^/\Ï3 conventional activated sludge

CASRN chemical abstract services registry number

CLA clarithromycin

CMU combined measurement uncertainty

COD chemical oxygen demand

d4SDZ sulfadiazine-«/

d4SMX sulfamethoxazole-«/

d4STZ sulfathiazole-«/

d5N4AcSMX A^-acetylsulfamethoxazole-d5DOC dissolved organic carbon

dw dry weight

EAWAG Swiss Federal Institute of Environmental Science and

Technology

ERY erythromycin

ERY-H20 dehydro-erythromycin

ETH Swiss Federal Institute of Technology

FBR fixed-bed reactor

H NMR proton nuclear magnetic resonance spectroscopy

HPLC high performance liquid chromatography

HUMABRA acronym for the Swiss national research project (NRP 49)

entitled "Occurrence of human-use antibiotics and antibiotic

resistance in the aquatic environment"

JOS josamycin

Kd sorption constant

Koc organic carbon normalized sorption constant

LC liquid chromatography

LOD limit of detection

X

LOQ

m/z

MBR

MEC

MS

MS/MS

V

N4

N4AcSMX

N4AcSMZ

NRP

Ntot

PE

PLE

POSEIDON

PPCP

Ptot

R

RAW

ROX

RT

S/N

SD

SDZ

SF

SMR

SMX

SMZ

SPE

SPY

STZ

t/a

limit of quantification

mass-to-charge ratio

membrane bioreactor

measured environmental concentration

mass spectrometry

tandem mass spectrometry

nitrogen in the sulfonamide group of sulfonamides

nitrogen in the para-amino group of sulfonamides

A^-acetylsulfamethoxazole//-acetylsulfamethazinenational research project

total nitrogen concentration

population equivalent

pressurized liquid extraction

acronym for the European project entiteld "Assessment of

Technologies for the Removal of Pharmaceuticals and

Personal Care Products in Sewage and Drinking Water

Facilities to Improve the Indirect Potable Water Reuse"

(EVK1-CT-2000-00047)

pharmaceutical and personal care product

total phosphorus concentration

substituent

raw influent

roxithromycin

retention time

signal-to-noise ratio

standard deviation

sulfadiazine

sand filter

sulfamerazin

sulfamethoxazole

sulfamethazine

solid-phase extraction

sulfapyridine

sulfathiazole

tons per annum

XI

TRI trimethoprimTSS total amount of suspended solids

TYL tylosin

USE ultrasonic solvent extraction

WWTP wastewater treatment plant

WWTP-A municipal WWTP at Altenrhein, Switzerland

WWTP-K municipal WWTP at Kloten-Opfikon, Switzerland

WWTP-W municipal WWTP at Wiesbaden, Germany

1 W W 1 "W Vf* I #

^ i ris i i\ i<< rî % : l

Chapter 1

Introduction

1 V» 1 ,~=

8 1% ^i „'

Introduction 15

1.1 Pharmaceuticals in the Environment

The protection of water resources for drinking water production, agricultural

crops, recreational activities and natural reserves from contamination is one of

the main issues in a sustainable water policy. In ecosystems and drinking water,

the absence of contaminants is particularly required based on the precautionary

principle. Within the large group of contaminants identified until today, the

occurrence of pharmaceuticals in the environment is of special interest. Due to

their high biological activity, possible adverse effects may occur in the

environment, which are so far not covered by conventional ecotoxicological

testing procedures. In contrast to the large number of high volume chemicals

applied in agriculture, industry and households, a complete ban or even a

significant use reduction of pharmaceuticals can not be envisaged because of the

immense benefits of these substances for society.

Pharmaceuticals comprise a very broad and diverse spectrum of hundreds of

chemical substances, including prescription and over-the-counter therapeutic

drugs, diagnostic agents, and many others. In the last decade, the attention of

researchers, authorities and the public has been drawn to the occurrence and fate

of pharmaceuticals in the environment. This development is strongly correlated

to the simultaneous improvements in chemical analysis enabling the detection

and quantitative trace determination of these mostly highly polar and

hydrophilic contaminants. Therefore, pharmaceuticals are considered as

emerging contaminants, even though it must be assumed that they have been

present in wastewater and in ambient waters as long as they have been used.

In 1976, Garrison and co-workers firstly reported the presence of salicylic acid

and clofibric acid, the metabolite of Clofibrate, a lipid regulator, in domestic

wastewater.[1] Approximately one decade later, in 1985, a wide variety of

pharmaceuticals was detected by Richardson and Bowron in wastewater

treatment plant effluents and surface waters with estimated concentrations

ranging up to 1 u.g/L.l2J From the beginning of the 1990's the occurrence of

pharmaceuticals in the environment has gained more and more interest in

Europe, and later worldwide, starting with Stan and Linkenäger,[3] who found

clofibric acid in groundwater in Berlin, Germany. This led to comprehensive

monitoring studies, first performed by Ternes et al.[4] Until today, over 80

different pharmaceutical compounds have been measured in various samples

and the results are summarized and discussed in many articles and

16 Chapter 1

reviews (e.g. [4-13]). The main interest has been the occurrence and

concentration range of pharmaceuticals in mainly aqueous environmental

compartments. Further research, however, is necessary to better understand the

behavior and ultimate fate of pharmaceuticals in the environment after their

application.

A large number of pharmaceutical compounds is used today in human and

veterinary medicine, reaching the environment through their manufacture, use

and disposal. Figure 1.1 shows the main exposure routes of pharmaceuticals into

the environment.

After their medicinal application and excretion via urine and feces, human used

pharmaceuticals mainly enter municipal sewage treatment plants. In Switzerland

over 95% of the population is connected to sewer systems. However, direct

input of human used pharmaceuticals into natural waters is also possible during

rain events and by leaks in the sewers. The behavior and fate of pharmaceutical

compounds in wastewater treatment is mostly unknown. Their frequent

Figure 1.1 Principal routes ofenvironmental exposure to pharmaceuticals

veterinaryuse

fish

farming human

use

manure

run¬

off

over

flow

sewage

soil <H- + -h-

leaks

treatment

plant

-^

surface water <^>

disposal

municipalwaste

landfill

groundwater

drinking water

Introduction 17

detection in wastewater effluents and receiving surface waters (e.g. [4, 14-22]),

however, suggests an incomplete removal of many pharmaceuticals. Next to

chemical or biological transformation, sorption to sewage sludge can play a role

in the elimination of pharmaceuticals in wastewater treatment. The use of

sewage sludge as fertilizer on arable land may therefore be an alternative route

to the environment. Up to date, only little data is available on the occurrence of

human used pharmaceuticals in sewage sludge and soil.[23"27]

In veterinary medicine, pharmaceuticals are applied for therapeutic use and as

growth promoters in livestock production. Due to the ban of growth promoters

in many European countries, including Switzerland, the overall consumption has

decreased in recent years. They mainly reach the environment via animal

manure - through the direct urination or defecation on the fields or after the

application of stored manure to arable land. They may therefore reach

groundwater after soil passage, or surface waters due to field runoff during rain

events or through drainage systems. In the field of veterinary pharmaceuticals,

research activities have also increased, mainly focusing on the occurrence and

fate of veterinary pharmaceuticals, especially antimicrobials, in manure, soil and

surface waters (e.g. [28-31]) Through the use as feed additives in fish farming,

veterinary pharmaceuticals are also directly introduced into surface waters

(e.g. [32,33]).

After consumption, pharmaceutical substances are metabolized in the body to

different extents and are excreted only partly unchanged. Metabolism usually

consists of two phases (Figure 1.2). In a first phase pharmaceutical compounds

may undergo an oxidation, reduction or hydrolysis reaction, resulting in the

introduction of reactive functional groups. In a second phase the pharmaceutical

or its first phase metabolite can then be covalently bound to polar molecules,

e.g. acids, sulfates or sugars. As a conclusion, metabolism usually results in

compounds with a higher polarity. Being more hydrophilic than the original

compound, metabolites are more easily excreted from the body. The excreted

metabolites, exhibiting diverse biological activity, increase the already complex

mixture of pharmaceuticals. In some cases the predominant form of a

pharmaceutical in the environment may be its metabolite (e.g. clofibric acid).

Additionally, metabolites may be transformed back to the parent compound

under environmental conditions or during wastewater treatment

(e.g. deglucuronidation). However, metabolites are so far generally not included

18 Chapter 1

Figure 1.2 Metabolism ofpharmaceuticals

metabolite

e.g. oxidation,

reduction, hydrolysis

conjugation

e.g. glucuronidation,

acetylation

phase 1

i

pharmaceutical —

r

metabolite

phase II

in studies on the occurrence, behavior and fate of pharmaceuticals in the

environment, with a few exceptions (e.g. dehydro-erythromycin, clofibric acid

and salicylic acid). This can mainly be attributed to the usually high polarity of

the metabolites and the lack of reference substances, both rendering the

determination of these compounds difficult.

Overall, it can be stated, that residual concentrations of pharmaceuticals,

unchanged or partly transformed, are continuously discharged to the

environment. Although pharmaceuticals are usually not included in legal

regulation, e.g. the priority list of the European Water Framework Directive, the

precautionary principle implies an efficient removal of these potentially harmful

substances. In some cases, pharmaceuticals have also been detected in

groundwater and were generally connected to specific input sources, e.g.

wastewater, agricultural use, or landfill sites.l34"37] Occasionally pharmaceutical

concentrations in the lower nanogram per liter range were found in drinking

water, caused by the widespread occurrence of these compounds in the aquatic

environment.f38,39] Even though these concentrations are unlikely to cause

adverse health effects in humans, drinking water should be devoid of

anthropogenic contaminations based on the application of the precautionary

principles.

1.2 Selection of Antimicrobials

The presence of pharmaceuticals in the environment is of concern, since they are

designed to cause specific effects in humans or animals. Within this large group,

antimicrobials are of special concern due to the possible spread and maintenance

Introduction 19

of bacterial resistance. While resistant pathogens are mainly found in clinics,

little is known on the contribution of the low but continuously discharged

antimicrobial concentrations from wastewater treatment to the widespread

occurrence of antimicrobial resistance observed in the environment.[40"44J

Antimicrobials are used in large quantities in human and veterinary medicine,

mainly for the treatment of bacterial infections. The overall consumption of

antimicrobials amounted to ~90 tons in Switzerland in 1997 and -40% thereof

were applied in human medicine.145"47^ Domestic consumption, i.e. during

ambulant treatment, accounts for 60 to 80% of the total human consumption

making urban wastewater the main source. With -18 tons per annum (t/a),

ß-lactam antimicrobials represent the largest fraction of human used

antimicrobials. They include penicillins and cephalosporins and seem to be

hydrolyzed shortly after excretion. Following the class of ß-lactams, macrolides

(4.3 t/a), sulfonamides (5.7 t/a) and fluoroquinolones (3.9 t/a) are mainly used in

human medicine. While the occurrence and fate of fluoroquinolone

antimicrobials in the environment has already been intensively studied,[23'48"50]

little is know regarding sulfonamide and macrolide antimicrobials.

The sulfonamide antimicrobials are a class of synthetic compounds derived from

sulfanilamide, whose antibacterial activity was discovered in the early 1930> s by

Domagk and Tréfouel.[51] Their core structure is shown in Figure 1.3. The acid

dissociation constants of the two amino groups in the molecule range between

1.7 and 2.4 for the protonated/>-amino group (TV4) and between 5.0 and 8.5 for

Figure 1.3 Core structures ofsulfonamide (A) and 14-membered-ring macrolide

antimicrobials (B)

A)

N4

X

M /

Nl

o

-s-

IIo

R1

/

\

V, TV4 = numbering of the nitrogens

RI, R2 = varying substituents

P—R1

20 Chapter 1

the sulfonamide nitrogen (TV1) in commonly used representatives. VaryingN1 substituents results in a large variety of compounds with a wide range of

pharmacological properties. Substitution on the /?-amino group, however, leads

to inactive compounds, since sulfonamides act as competitive antagonist of

/»-aminobenzoic acid in bacterial folic acid synthesis. Sulfonamides are

prescribed against a wide variety of bacterial infections, especially in the case of

a potential hypersensitivity to penicillins. However, the widespread bacterial

resistance to these compounds limits their application spectrum today. In

monotherapy, sulfonamides lead to a bacteristatic effect, while a combination

with trimethoprim, a diaminopyridin derivative, results in a bactericidal effect.

Trimethoprim, almost exclusively used in combination with sulfonamides, also

interferes with the bacterial folic acid synthesis through inhibition of the

bacterial dihydrofolate reductase.[52] Sulfonamides are metabolized to a varying

extent in the human body (e.g. by A^-acetylation and hydroxylation) and are

subsequently excreted mainly via the urine. Figure 1.4 shows the main human

metabolite of sulfamethoxazole, the most predominant sulfonamide in human

medicine. Approximately 50% of the administered dose is excreted as the

inactive metabolite A^-acetylsulfamethoxazole.[53] Sulfapyridine, the other

sulfonamide important in human medicine, is not administered as antimicrobial

agent directly, but in form of sulfasalazine, which is mainly used in the

treatment of ulcerative colitis and rheumatoid arthritis.[54] In sulfasalazine,

sulfapyridine is linked to 5-aminosalicylic acid via an azo bridge, which is

cleaved in the colon (Figure 1.5).

Figure 1.4 Main human metabolite ofsulfamethoxazole

c

/o

/o s phase I

H2N—<v />—S—NH—& I HN—(v /)—S—NH—<\

O

-NH—C

sulfamethoxazole A^-acetylsulfamethoxazole(50% of the administered dose)

Introduction 21

Figure 1.5 Chemical structure ofsulfasalazine

HOOC

~0-JhH~0Macrolide antimicrobials are derived from natural sources and are therefore also

classified as antibiotics.[55] They consist of a macrocyclic lactone ring attached

to one or more amino or neutral sugars. The ring structure is usually substituted

by various functional groups, e.g. hydroxyl, methyl or methoxy groups. In

Figure 1.3 the core structure of 14-membered-ring macrolides is given,

representative for the most commonly used macrolides in human medicine -

erythromycin and clarithromycin. The acid dissociation constant of the

protonated tertiary amine in the attached sugar ranges between 8.7 and 9.2,

making macrolides weakly basic. Their mode of action is based on the inhibition

of bacterial protein synthesis by a reversible interaction with the bacterial 50S

ribosomal subunit. Macrolides are active against a variety of gram-positive

bacteria and are mainly applied in the treatment of respiratory tract infections.

After consumption, macrolides are excreted mainly via the feces. Metabolism,

including 7V-demethylation of the amino groups and hydroxylation of the ring, is

generally of minor importance and macrolides are predominately excreted

unchanged.[55] As an example the main human metabolites of clarithromycin, the

most commonly used macrolide in human medicine are given in Figure 1.6.

1.3 Research Framework

The occurrence and fate of organic micropollutants in the environment has

already been a focus of the research at EAWAG for a long time. Field studies in

wastewater treatment plants, groundwater, rivers and lakes, and hospital

effluents as well as controlled laboratory studies were performed for a large

variety of water pollutants. In the area of human-use antimicrobials, the first

investigations dealt with the fluoroquinolone ciprofloxacin, for which genotoxic

studies in wastewaters of hospitals were performed.[56] Subsequently, several

fluoroquinolone antimicrobial agents were intensively studied in a dissertation

by Eva Golet.[57] After the development of analytical methods for aqueous

22 Chapter 1

Figure 1.6 Main human metabolites ofclarithromycin

clarithromycin

^\CH3

IH0

/

14-OH-(R)-clarithromycin

(20% of the administered dose)

*~ H0'**.../ Ö HO .

CH, "\OH

14-OH-(R)-N-demethyl-clarithromycin(13% of the administered dose)

and solid samples,[23] the occurrence and fate of flouroquinolones were

investigated in rivers, wastewater treatment and sludge-treated soil.[48,50] The

occurrence of sulfonamide and macrolide antimicrobials has been investigated

in several rivers and lakes as well as wastewater treatment plant effluents by

McArdell et al., followed by a detailed study on the environmental behavior of

macrolides in the Glatt valley watershed.[20] Within HUMABRA, a project in the

framework of the national research program NRP 49 on antibiotic resistance of

the Swiss National Science Foundation, the question of a potential

environmental risk of antimicrobials occurring as trace contaminants concerning

resistance is addressed.[43] In particular, residual levels of human-use antibiotics

are determined in wastewater, hospital effluents and in ambient waters.

Correlations between the measured concentrations and resistant bacteria strains

are investigated.

This dissertation is closely related to the European project POSEIDON, in which

eight research groups from seven countries are involved.[58'59] It is focused on the

Introduction 23

assessment and improvement of technologies for the removal of pharmaceuticals

and personal care products (PPCPs) in wastewater and drinking water treatment

facilities. Within POSEIDON eleven different compounds belonging to various

PPCP classes were chosen according to the amount consumed, chemical

properties, possible adverse effects and the availability of reference compounds.

Sulfamethoxazole and roxithromycin were included as antimicrobials. The

overall goal of the project is to reduce the contamination of receiving waters,

groundwater and drinking water with PPCPs from treated wastewater by

planned and unplanned indirect reuse. This clearly illustrates the need of

integrating different scientific disciplines, e.g. chemists and engineers, to

address the questions raised by environmental PPCP contamination. In

particular, in the field of water pollution, an integrated approach is required to

assure the sustainability of water use and re-cycling for various purposes.

Parallel to this study, Felix Wettstein has investigated in his dissertation

nonylphenoxy acetic acid and other nonylphenol compounds in wastewater

treatment field studies similar to those presented here.f60] Results obtained for

this very different class of organic substances allow interesting comparisons

with the selected antimicrobials.

1.4 Scope of this Study

This thesis project aimed at investigating the occurrence and behavior of

sulfonamides, macrolides and trimethoprim in municipal wastewater treatment.

The pathway of selected constituents was traced from the raw influents to the

final effluents of wastewater treatment plants. Different treatment stages and

technologies were evaluated with respect to their elimination efficiencies for

selected antimicrobials. Within this dissertation the following topics and

questions have been thoroughly addressed:

Analytical methods

Reliable and specific analytical methods for many different types of sample

matrices are essential to investigate the occurrence and fate of pharmaceuticals

in the environment. One of the main goals of this work was to develop the

methods necessary to investigate the selected substances in wastewater

treatment. Methods were validated for the simultaneous trace determination of

24 Chapter 1

sulfonamides, macrolide and trimethoprim in various wastewaters (Chapter 2),

and in sewage sludge (Chapter 3). Pressurized liquid extraction (PLE) was

chosen for efficient extraction from solid samples. Enrichment and clean-up of

sludge extracts or wastewater samples was achieved using solid-phase extraction

on polymeric sorbent cartridges. Analyses were performed by liquid

chromatography on a reversed phase column coupled to electrospray positive

mass spectrometry. Tandem mass spectrometry in the multiple reaction mode

provided the sensitivity and selectivity necessary for the analyses of complex

environmental samples, such as raw wastewater and sludge extracts.

Study of municipal wastewater treatment

In the field studies performed, the occurrence of sulfonamides, macrolides and

trimethoprim in municipal wastewaters from the raw influent to the final

effluent was one of the main questions addressed. Additionally, the studies were

designed to elucidate possible eliminations. Conventional activated sludge

treatment, being the most commonly applied wastewater treatment technology,

was investigated in detail (Chapter 4 and 5). In addition to aqueous matrices,

concentrations of the selected analytes were also measured in activated and

digested sewage sludge. Average daily loads were determined as well as daily

variations of antimicrobial loads entering wastewater treatment. By performing

complete mass balances the overall removal of the analytes was determined in

mechanical and biological municipal wastewater treatment. Furthermore, the

efficiency of individual treatment steps for the removal of sulfonamides,

macrolides and trimethoprim was investigated in depth. By attributing the

observed eliminations to different extents to transformation and sorption

processes, respectively, new information on the sorption characteristics of these

compounds could be gained.

Evaluation of different treatment technologies

The influence of various factors on the elimination of micropollutants was

addressed by studying the fate of sulfonamides, macrolides and trimethoprim

also in newly developed treatment technologies. Field studies were performed

on a fixed-bed reactor and on a membrane bioreactor (Chapter 5). In the case of

the membrane bioreactor, three different solid retention times were investigated

Introduction 25

in weekly sampling campaigns. The observed eliminations in the different

treatment technologies are discussed in relation to various parameters, e.g.

temperature and solid retention time. In two sand filters, used for tertiary

treatment, removal efficiencies concerning sulfonamides, macrolides and

trimethoprim are compared.

Ozonation of wastewater effluents

Since the studied antimicrobials are not completely eliminated in wastewater

treatment, other measures are necessary to further reduce their loads entering

ambient waters. Ozonation of wastewater effluents was investigated as a

possible additional treatment step (Chapter 6). The removal efficiencies were

determined for several sulfonamides, macrolides and trimethoprim at ozone

doses below 5 mg/L. The influence of suspended solids in the wastewater matrix

was particularly investigated. Using three different wastewater effluents, the

impact ofpH variations on the oxidation process could also be addressed.


[I] Garrison, A. W.; Pope, J. D.; Allen, F. R. In Identification & Analysis of

organic pollutants in water, Keith, L. H., Ed.; Ann Arbor Science: Ann

Arbor, 1976; pp 517-566.

[2] Richardson, M. L.; Bowron, J. M. J. Pharm. Pharmacol. 1985, 37, 1-12.

[3] Stan, H. J.; Linkerhäger, M. Vom Wasser 1992, 79, 75-88.

[4] Ternes, T. A. Water Res. 1998, 32, 3245-3260.

[5] Stan, H. J.; Heberer, T. Analusis 1997, 25, M20-M23.

[6] Halling-Sorensen, B.; Nors Nielsen, S.; Lanzky, P. F.; Ingerslev, F.; Holten

Lützenhoft, H. C; J0rgensen, S. E. Chemosphere 1998, 36, 357-393.

[7] Daughton, C. D.; Ternes, T. A. Environ. Health Perspect. 1999, 107, 907-

938.

[8] Kümmerer, K. Chemosphere 2001, 45, 957-969.

[9] Heberer, T. Toxicol. Lett. 2002,131, 5-17.

[10] Snyder, S. A.; Westerhoff, P.; Yoon, Y.; Sedlak, D. L. Environ. Eng. Sei.

2003, 20, 449-469.

[II] Debska, J.; Kot-Wasik, A.; Namiesnik, J. Crit. Rev. Anal. Chem. 2004, 34,

51-67.

26 Chapter 1

[12] Rooklidge, S. J. Sei. Total Environ. 2004, 325, 1-13.

[13] Giger, W.; Alder, A. C; Golet, E. M.; Kohler, H.-P. E.; McArdell, C. S.;

Molnar, E.; Siegrist, H. R.; Suter, M. J.-F. Chimia 2003, 57, 485-491.

[14] Stumpf, M.; Ternes, T. A.; Haberer, K.; Seel, P.; Baumann, W. Vom

Wasser 1996, 55,291-303.

[15] Sacher, F.; Lochow, E.; Bethmann, D.; Brauch, H.-J. Vom Wasser 1998, 90,

233-243.

[16] Zuccato, E.; Calamari, D.; Natangelo, M.; Fanelli, R. Lancet 2000, 355,

1789-1790.

[17] Alder, A. C; McArdell, C. S.; Golet, E. M.; Ibric, S.; Molnar, E.; Nipales,

N. S.; Giger, W. In Pharmaceuticals and Personal Care Products in the

Environment: Scientific and Regulatory Issues; Daughton, C. G., Jones-

Lepp, T., Eds.; Symposium Series 791; American Chemical Society:

Washington, D.C., 2001; pp 56-69.

[18] Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaugg, S.

D.; Buxton, H. T. Environ. Sei. Technol. 2002, 36, 1202-1211.

[19] Yang, S.; Carlson, K. Water Res. 2003, 37, 4645-4656.

[20] McArdell, C. S.; Molnar, E.; Suter, M. J.-F.; Giger, W. Environ. Sei.

Technol. 2003, 37, 5479-5486.

[21] Metcalfe, C. D.; Miao, X.-S.; Koenig, B. G.; Struger, J. Environ. Toxicol.

Chem. 2003,22,2881-2889.

[22] Miao, X.-S.; Bishay, F.; Chen, M.; Metcalfe, C. D. Environ. Sei. Technol.

2004,55,3533-3541.

[23] Golet, E. M.; Strehler, A.; Aider, A. C; Giger, W. Anal. Chem. 2002, 74,

5455-5462.

[24] Küpper, T.; Berset, J. D.; Etter-Holzer, R.; Tarradellas, J. Chemosphere

2004,54,1111-1120.

[25] Ternes, T. A.; Herrmann, N.; Bonerz, M.; Knacker, T.; Siegrist, H.; Joss, A.

Water Res. 2004,35,4075-4084.

[26] Kreuzig, R.; Kullmer, C; Matthies, B.; Höltge, S.; Dieckmann, H. Fresen.

Environ. Bull. 2003,12, 550-558.

[27] Thiele-Bruhn, S. J. Plant Nutr. Soil Sei. 2003,166, 145-167.

[28] Haller, M. Y.; Müller, S. R.; McArdell, C. S.; Aider, A. C; Suter, M. J.-F.

J. Chromatogr., A 2002, 952, 111-120.

[29] Hamscher, G.; Sczesny, S.; Höper, H.; Nau, H. Anal. Chem. 2002, 74,

1509-1518.

Introduction 27

[30] Tolls, J. Environ. Sei. Technol. 2001, 35, 3397-3406.

[31] Stoob, K. "Fate and occurrence of veterinary sulfonamide antimicrobials in

the environment," PhD Thesis, ETH Zurich, in preparation.

[32] Björklund, H. V.; Rabergh, C. M. I.; Bylund, G. Aquaculture 1991, 97, 85-

97.

[33] Capone, D. G.; Weston, D. P.; Miller, J.; Shoemaker, C. Aquaculture 1996,

145, 55-75.

[34] Sacher, F.; Lange, F. T.; Brauch, H.-J.; Blankenhorn, I. J. Chromatogr., A

2001,935,199-210.

[35] Hirsch, R.; Ternes, T. A.; Haberer, K.; Kratz, K.-L. Sei. Total Environ.

1999,225,109-118.

[36] Holm, J. V.; Rügge, K.; Bjerg, P. L.; Christensen, T. H. Environ. Sei.

Technol. 1995, 29, 1415-1420.

[37] Ahel, M.; Jelicic, I. In Pharmaceuticals andpersonal care products ind the

environment - scientific and regulatory issues; Daughton, CD., Jones-

Lepp, T.L., Ed.; American Chemical Society, ACS Symposium Series 791,

2001; pp 100-115.

[38] Heberer, T.; Stan, H. J. Vom Wasser 1996, 86, 19-31.

[39] Bund/Länderausschuss für Chemikaliensicherheit (BLAC), "Arzneimittel in

der Umwelt - Auswertung der Untersuchungsergebnisse," Hamburg, 2003.

[40] Iwane, F.; Urase, T.; Yamamoto, K. Water Sei. & Technol. 2001, 43, 91-99.

[41]Bendt, T.; Pehl, B.; Gehrt, A.; Rolfs, C.-H. KA - Wasserwirtschaft,

Abwasser, Abfall 2002, 49, 49-56.

[42] Klare, T.; Konstabel, C; Badstübner, D.; Werner, G.; Witte, W. Int. J. Food

Microbiol. 2003, 55, 269-290.

[43] http://www.nrp49.ch/pages/.

[44] Husevag, B.; Lunestad, B. T.; Johannessen, B. J.; Enger, O.; Samuelsen, O.

B. J. Fish Dis. 1991,14, 631-640.

[45] Annual Report; Swiss Importers of Antibiotics (TSA): Berne, Switzerland,

1998.

[46] Pharmaceuticals Sold in Switzerland; Swiss Market Statistics, 1999.

[47] Antibiotics used in Veterinary Medicine; Swiss Federal Office for

Agriculture (BLW): Berne, Switzerland, 2001.

[48] Golet, E. M.; Alder, A. C; Giger, W. Environ. Sei. Technol. 2002, 36,

3645-3651.

25 Chapter 1

[49] Golet, E. M.; Alder, A. C; Hartmann, A.; Ternes, T. A.; Giger, W. Anal.

Chem. 2001,73, 3632-3638.

[50] Golet, E. M.; Xifra, I.; Siegrist, H. R.; Alder, A. C; Giger, W. Environ. Sei.

Technol. 2003, 37, 3243-3249.

[51] Vree, T. B.; Hekster, Y. A. Clinical pharmacokinetics ofsulfonamides and

their metabolites; Karger: Basel, 1987; Vol. 37.

[52] Poe, M. Science 1976,194, 533-535.

[53] Vree, T. B.; Hekster, Y. A. Pharmacokinetics of sulfonamides revisited;

Karger: Basel, New York, 1985; Vol. 34.

[54] Astbury, C; Dixon, J. S. J. Chromatogr. 1987, 414, 223-227.

[55] Bryskier, A. J.; Butzler, J.-P.; Neu, H. C; Tulkens, P. M. Macrolides;

Arnette Blackwell: Paris, 1993.

[56] Hartmann, A.; Alder, A. C; Koller, T.; Widmer, R. M. Environ. Toxicol.

Chem. 1998,77,377-382.

[57] Golet, E. M. "Environmental exposure assessment of fluoroquinolone

antibacterial agents in sewage, river water and soil.," PhD Thesis, No.

14690, ETH Zurich, 2002.

[58] http://www.eu-poseidon.com.

[59] Ternes, T. A. "Final Report ofPOSEIDON," 2004.

[60] Wettstein, F. "Auftreten und Verhalten von Nonylphenoxyessigsäure und

weiteren NonylphenolVerbindungen in der Abwasserreinigung," PhD

Thesis No. 15315, ETH Zurich, 2004.

Chapter 2

Analytical Method for Wastewater

An analytical method has been developed and validated for the simultaneous

trace determination of four macrolide antimicrobials, five sulfonamides, the

human metabolite A^-acetyl-sulfamethoxazole, and trimethoprim in wastewater.

The method was validated for tertiary, secondary, and - unlike in previously

published methods - also for primary effluents of municipal wastewater

treatment plants. This wide range of application is necessary to thoroughly

investigate the occurrence and fate of chemicals in wastewater treatment.

Wastewater samples were enriched by solid-phase extraction, followed by

reversed-phase liquid chromatography coupled to tandem mass spectrometry

using positive electrospray ionization. Recoveries from all sample matrices were

generally above 80%, and the combined measurement uncertainty varied

between 2.4 and 16%. Concentrations measured in tertiary effluents ranged

between 10 ng/L for roxithromycin and 423 ng/L for sulfamethoxazole.

Corresponding levels in primary effluents varied from 22 to 1 450 ng/L,

respectively. Trace amounts of these emerging contaminants reach ambient

waters, since all analytes were not fully eliminated during conventional

activated sludge treatment followed by sand flltration. In the case of

sulfamethoxazole, the amount present as human metabolite A^-acetyl-sulfamethoxazole had to be taken into account in order to correctly assess the

fate of sulfamethoxazole in wastewater treatment.

Göbel, A., McArdell, CS., Suter, M. J.-F., Giger, W.

Trace Determination of Macrolide and Sulfonamide Antimicrobials, a Human

Sulfonamide Metabolite, and Trimethoprim in Wastewater Using Liquid

Chromatography Coupled to Electrospray Tandem Mass Spectrometry

Analytical Chemistry, 2004, 76, 4756 - 4764.

Analytical Methodfor Wastewater 31

2.1 Introduction

Since 1997, interest in the occurrence and behavior of pharmaceuticals in the

aquatic environment has significantly increased.^1"71 One motivation for this

attention is the fact that these chemicals are designed to trigger specific

biological effects, and, hence, pose a potential threat for the aquatic

environment. In the case of antimicrobial agents (including both naturally and

synthetically derived compounds), the possible maintenance and spread of

bacterial resistance is a major point of concern. In 1997, the human consumption

of antimicrobials in Switzerland exceeded 30 tons per annum, (t/a).["

In

addition -55 t/a of antimicrobials were used in veterinary medicine. Macrolides

(4.3 t/a), sulfonamides (5.7 t/a), and fluoroquinolones (3.9 t/a) represent the

most important groups in human medicine, next to ß-lactams (17.5 t/a). The

latter include penicillins and cephalosporins and seem to be hydrolyzed shortly

after excretion.

In industrialized countries, most human used antimicrobials and other human

used pharmaceuticals reach the aquatic environment, unchanged or transformed,

mainly via discharge of effluents from municipal wastewater treatment plants

(WWTPs). The residual concentrations of these bioactive compounds in treated

effluents depend on their removal during wastewater treatment. They can

potentially pose a hazard for aquatic organisms if the removal is incomplete. In

addition, exposure via sewage sludge disposal on land could represent a hazard

for soil organisms.

Detailed knowledge of the behavior of antimicrobials in wastewater treatment

and the aquatic environment will help to achieve a reliable basis for

environmental risk assessment (e.g., by providing measured environmental

concentrations (MECs)). MECs can be used in environmental risk assessment

studies since they provide accurate indications of actual concentrations present

in environmental systems. Investigations on the occurrence and fate of

antimicrobial agents in various wastewater treatment steps can be exploited in

order to evaluate wastewater treatment technologies with respect to elimination

of specific contaminants. Reducing the release of residual pharmaceuticals into

the aquatic environment would presumably decrease any potential

environmental risks. By monitoring receiving surface waters as well as

wastewater treatment plants, locations of particular concern can be identified

and mitigated specifically.

32 Chapter 2

Table 2.1 Investigated compounds

compound acronym CASRN PKal/pKa2

sulfadiazine SDZ 68-35-9 64 [ii]/ 2.1 [12]

sulfathiazole STZ 72-14-0 7.2[11]/2.1[12]

sulfamethazine SMZ 57-68-1 7.4[11]/2.3[12]

sulfapyridine SPY 144-83-2 8.4[11J/2.6[13]

sulfamethoxazole SMX 723-46-6 5.7[1,]/1.8[12J

A^-acetylsulfamethoxazole N4AcSMX 5.0 [U1

trimethoprim TRI 738-70-5 7.2[14J

azithromycin AZI 83905-01-5 87[i5]/95[i5]

erythromycin ERY 114-07-8 8.8 [16]

clarithromycin CLA 81103-11-9 8.9[15]

roxithromycin ROX 80214-83-1 9.2[16]

tylosina TYL 1401-69-0 j i [15]

sulfamerazinea

SMR 127-79-7 7.0[,,]/2.2[12]

josamycina JOS 16846-24-5

Used as internal standard.

To reach the aims stated above, selective and sensitive analytical methods for

many different sample matrices are essential. Until now, published methods for

antimicrobial agents have focused on wastewater treatment plant effluents and

surface waters,tl7~231 with the exception of fluoroquinolones, which were studied

in detail by Golet et al.[2427] Analytical methods for wastewater matrices other

than final effluents including sludge extracts, however, are lacking. Another

important aspect that has not yet been sufficiently addressed is the presence of

human metabolites of antimicrobials in wastewaters. Sulfamethoxazole, for

example, is metabolized in the human body and -50% of the administered dose

is excreted as the inactive human metabolite A^-acetylsulfamethoxazole and

only 10% as the unchanged compound.[28] The retransformation of

A^-acetylsulfamethazine to the active sulfamethazine during the storage of

manure has already been shown by Berger et al., suggesting a similar cleavage

of A^-acetylated sulfonamides, for instance in wastewater treatment.[29]


Observed elimination rates may be biased, if the possible retransformation to the

active pharmaceutical is not considered. To the best of our knowledge, only one

study included A^-acetylsulfamethoxazole in the analysis of surface water and

WWTP effluents - indicating concentrations of up to 2 200 ng/L in WWTP

effluents.[30J Unfortunately, the state of treatment has not been reported. This

clearly shows the importance of considering the main human metabolite of

sulfonamides when assessing the occurrence and fate of sulfamethoxazole in

wastewater treatment.

In this article, we present a reliable analytical method for the trace determination

of the most important macrolide and sulfonamide antimicrobials in the various

aqueous compartments of a WWTP, including primary effluent. In addition, the

human metabolite A^-acetylsulfamethoxazole and trimethoprim - frequently

used as a synergist to sulfamethoxazole - were measured. Table 2.1 lists the

selected macrolides and sulfonamides; their respective chemical structures are

given in Figure 2.1 and 2.2. Using solid-phase extraction combined with liquid

chromatography tandem mass spectrometry (positive electrospray ionization)

concentrations down to the low nanogram per liter range can be determined. The

Figure 2.1 Chemical structures ofmacrolide antimicrobials

K2

Ri

erythromycin (ERY) H

clarithromycin (CLA) CH3

roxithromycin (ROX) H

R2

O

O

HO,

H3CT

>ivrt azithromycin

(AZI)

'v/^oh y'""*CH3

H3C, *xCH3 dehydro-erythromycin

(ERY-H20)

/

\ CH3

°~7^CZ2o>

\

34 Chapter 2

Figure 2.2 Chemical structures ofsulfonamide antimicrobials and trimethoprim

pK.2 pKJ

R2

/N

\

N»

N MH2

trimethoprim

(TRI)

sulfadiazine

(SDZ)

sulfathiazole

(STZ)

sulfamethazine

(SMZ)

sulfapyridine

(SPY)

sulfamethoxazole

(SMX)

A^-acetyl-sulfamethoxazole

(N4AcSMX)

Ri

H

H

H

R2

N=v

-o

-0

H

H

-o

COCH3

-tf

CH3

CH3

presented method is feasible to study the occurrence and fate of the selected

compounds in all compartments of a wastewater treatment plant as well as for

environmental monitoring studies. Preliminary results on the occurrence of

macrolides and sulfonamides in Swiss wastewater treatment plants are

presented.


2.2.1 Chemicals and Reagents

HPLC-grade methanol, acetonitrile, and water are purchased from Scharlau

(Barcelona, Spain). Analytical ethyl acetate, ammonia solution, 25% sulfuric

acid, sodium chloride, sodium hydroxide, ammonium acetate, and formic acid

were obtained from Merck (Darmstadt, Germany). Sulfamethazine,

sulfamethoxazole, sulfadiazine, and roxithromycin were purchased from Sigma-

Aldrich (Buchs, Switzerland). Sulfathiazole, sulfapyridine, trimethoprim,


tylosin, josamycin, and erythromycin were obtained from Fluka Chemicals

(Buchs, Switzerland), and sulfamerazine was from Riedel-de Haën (Seelze,

Germany). Sulfamethazine-phenyl-13C6 was purchased from Cambridge Isotope

Laboratories (Andover, MA) and sulfamethoxazole-*/, sulfadiazine-*/,

sulfathiazole-d4, and A^-acetylsulfamethoxazole-d6 were purchased from

Toronto Research Chemicals (North York, ON, Canada). Clarithromycin was

kindly supplied by Abbott (Wiesbaden, Germany) and azithromycin by Pfizer

(Zurich, Switzerland). Azithromycin is also available from Sigma-Aldrich

(Buchs, Switzerland). Standard solutions for dehydro-erythromycin were

prepared from erythromycin as described by McArdell et al. The acidic

solution was readjusted to pH 6 after 4 h using IM NaOH to ensure stability

during storage. A^-acetylsulfamethoxazole was synthesized by acetylation with

acetic acid anhydrate according to Neumann with a yield of 70%.[31] Identity and

purity was confirmed by LC/UV, LC/MS/MS, and H NMR analysis.

2.2.2 Internal Standards

Deuterated sulfonamide standards were commercially available in most cases.

Erythromycin-13C2 was tested as an internal standard for the macrolides, but

proved to be unsuitable due to the significant natural contribution to M + 2 from

unlabeled erythromycin (11.6%). Similar observations were described by

Vanderford and co-workers for erythromycin-13Ci.[23] The absence in water

samples of all internal standards used was confirmed by enriching a

representative samples from each matrix, to which no surrogate or instrumental

standard was added. No peaks could be detected at the retention times of the

used internal standards. Surrogate standards were added prior to enrichment to

assess possible losses during the analytical procedure. Instrumental standards

were added to the final extracts prior to measurement. The following substances

were used as surrogate standards: sulfamethazine-phenyl- C6 ( C6SMZ) for

SMZ, TRI, and SPY, sulfamethoxazole-^/4 (d4SMX) for SMX, sulfadiazine-*/4

(d4SDZ) for SDZ, sulfathiazole-t/4 (d4STZ) for STZ, V-acetyl-sulfamethoxazole-^ (d5N4AcSMX) for N4AcSMX, and tylosin (TYL) for

ERY-H20, AZI, CLA, and ROX. As instrumental standards sulfamerazine

(SMR) was used for all sulfonamides and TRI, and josamycin (JOS) for all

macrolides. While the surrogate standards were used for quantification, the

instrumental standards were used to check the instrument performance during

36 Chapter 2

measurement. Its peak area was monitored over the whole measurement series in

order to detect problems with, for example, instrument sensitivity or injection

volume. If the area of the instrumental standard decreases significantly (signal

reduction of > 20% within same matrix), the series was stopped and the

instrument cleaned.

2.2.3 Sample Collection and Preparation

Flow-proportional composite samples of the primary effluent (1°EFFL) after

mechanical treatment, the secondary effluent (2°EFFL) after biological

treatment, and the tertiary effluent (3°EFFL) after sand filtration were collected.

The samples were transferred into amber glass bottles and filtered as soon as

possible but no later than 6 h after sampling through 0.45-um cellulose nitrate

filters (Schleicher & Schuell). The filtered samples were directly extracted or

kept at -20 °C in half-filled amber glass bottles in horizontal position until

extraction. The sample volumes were 250 mL for 2°EFFL and 3°EFFL samples

and 50 mL for 1°EFFL samples. The latter were diluted with 150 mL of water

prior to extraction. After addition of 1 g of sodium chloride, the pH was adjusted

to 4 with sulfuric acid, and the surrogate standard (50 - 100 ng) was added.

Solid-phase extraction was performed on 6-mL Oasis HLB sorbent cartridges

(200 mg; Waters, Bergen op Zoom, The Netherlands) using a 12-fold vacuum

extraction box (J.T. Baker, Phillipsburg, NY). The sorbent material is a

copolymer of two monomers, TV-vinylpyrrolidone and divinylbenzol. The

cartridges were preconditioned with 2 x 1.5 mL of methanol-ethyl acetate (1:1),

2 x 1.5 mL of methanol containing 1% (v/v) ammonia, and 2 x 1.5 mL of water

adjusted to pH 4 with H2S04. The wastewater samples were percolated through

the cartridges at a flow rate of less than 5 mL/min. After percolation, the

cartridges were washed with 1.5 mL of water-methanol (95:5) and the eluent

was discarded. Subsequently, the cartridges were dried completely in a nitrogen

flow for 1 h. The analytes were then eluted with 2 x 1.5 mL of methanol-ethyl

acetate (1:1) and 2 x 1.5 mL of methanol containing 1% ammonia into 10-mL

graduated glass vessels. Eluates were reduced to -50 pL by a gentle flow of

nitrogen at room temperature. After the addition of the instrumental standard

(100 ng), the sample volume was adjusted to 0.5 mL with water. Final extracts

were stored in amber glass vials at -15 °C until analysis.


2.2.4 Liquid Chromatography

HPLC analyses were performed using a Rheos 2000 pump equipped with a

solvent degasser (Flux Instruments AG, Switzerland), a HTS Pal autosampler

(CTC Analytics, Zwingen, Switzerland), and a Jones Chromatography column

oven, Model 7956 (Omnilab AG, Mettmenstetten, Switzerland). Sample aliquots

of 20 pL were injected. Two analytical columns were tested for separation.

Initially, a 125 x 2 mm Nucleosil 100-5 C18HD end-capped column (Macherey-

Nagel, Dueren, Germany) equipped with a 8 x 2 mm precolumn containing the

same sorbent material was used (column 1). Gradient elution was performed

with water adjusted to pH 4.6 by acetic acid and acetonitrile, both containing

10 mM ammonium acetate. Later, a 150 x 2 mm YMC Pro Ci8, 120 Â, 3 urn

(Stagroma, Reinach, Switzerland) column equipped with a 10 x 2 mm

precolumn containing the same sorbent (column 2) was used. Optimal

separation was achieved using column 2 maintained at 30 °C and with a flow

rate of 0.15 mL/min. Solvent A was water acidified with 1% (v/v) formic acid,

resulting in a pH of 2.1, and solvent B was methanol acidified with 1% (v/v)

formic acid. The run (0.15 mL/min) started at 10% B for 5 min, followed by a 5-

min linear gradient to 15% B, a 5-min linear gradient to 40% B, and another 5-

min linear gradient to 45% B and was terminated by a 10~min linear gradient to

70% B. Afterward, the eluent was brought to 100% B in 2.5 min and the column

washed at a flow rate of 0.25 mL/min for 10 min. Initial conditions were

reestablished in 2.5 min, and the column was equilibrated for 10 min at a flow

rate of 0.25 mL/min prior to the next analysis. The total time per analysis was 55

min. Table 2.2 gives the retention times of the individual analytes. To prevent

sensitivity losses of the mass spectrometer, the eluate of the first 8 min and of

the last 20 min of the chromatographic run were bypassed and discarded.

2.2.5 Tandem Mass Spectrometry

A triple quadrupole mass spectrometer, TSQ Quantum Discovery (Thermo

Finnigan, San Jose, CA), equipped with electrospray ionization was used for

detection. Analyses were performed in the positive mode, with a spray voltage

of 3 500 V and an ion-transfer capillary temperature of 350°C. Nitrogen was

used as sheath gas (40 bar) and as auxiliary gas (10 bar), and argon as collision

gas (1.5 mTorr). Both mass analyzers were set to unit resolution.

38 Chapter 2

Table 2.2 Precursor ions, selected fragment ions, and retention times of the

measured compounds

analyte precursor ion product ions retention time

(m/z) ; (m/z) (min)

SDZ 251.06 156.01,

108.04 10.3

d4SDZa 255.08 160.01 112.04 10.0

STZ 256.02 156.01 108.04 12.7

d4STZa 260.05 160.01 112.04 12.4

SMZ 279.09 124.09 186.03 17.6

13C6SMZ a 285.09 124.09 186.03 17.6

SPY 250.07 156.01 184.09 12.6

SMX 254.06 156.01 108.04 20.4

d4SMXa 258.08 160.01 112.04 20.3

N4AcSMX 296.07 134.06 198.02 24.1

d5N4AcSMXa

301.10 139.06 203.02 24.0

TRI 291.15 123.07 275.14 17.1

AZI 375.26 591.40 158.12 21.1

ERY-H20 716.46 540.33 558.34 30.1

CLA 748.49 158.12 590.37 31.5

ROX 837.53 158.12 ; 679.41 31.6

TYLa 916.53 174.11.

772.45 29.1

SMRa 265.08 156.01 ; 172.02 14.9

JOSa 828.48 108.91

,174.11 31.1

Used as internal standard.

Usually, the protonated molecular ion ([M + H]+) of the compounds was

selected as precursor ion except for azithromycin, for which the doubly charged

molecular ion ([M + 2H] ) was chosen as precursor ion because of its greater

abundance under the given conditions. Detection was performed in multiple

reaction monitoring mode using the two most intense and specific fragment ions.

Table 2.2 lists the monitored transitions for the individual analytes. The


detection of the compounds was divided in time windows during the course of

the chromatographic run with a dwell time of 100 ms. Figure 2.3 shows a

chromatogram of a 1°EFFL sample. In the case of SDZ, STZ, and SMZ, which

were not present in the sample, the peaks obtained from the measurement of a

1°EFFL sample spiked with 25 ng prior to sample preparation are included

(dashed lines).

Figure 2.3 Total ion chromatogram (sum oftwo transitions) ofan extractfrom a

primary wastewater effluent (1 °EFFL)

SPY

SDZ»

ill

TRI RQX

SMZ'

ERY-H20

100

c

at

^1 Vin*J*

II

i "WpftDwWfI

STZ«

50

Ï%

a.

SMX

AZI JfMüU: lé JIM kuAJuiiw

^'»^«,

N4AcSMX

! I

10

Jj.tf| --T-

15

.))ê*i

CLA

I" I f I

20 25

Retention "lime (min)

JWjJUj^jU 1^ A

T J "T>-r y --, |

30 35

aThe peaks shown for SDZ, STZ, and SMZ correspond to a 1°EFFL sample spiked

with 25 ng of these compounds since they were not present in the unspikcd sample.


For the method validation flow-proportional composite samples from the

respective effluents of a municipal wastewater treatment plant (WWTP Kloten-

Opfikon) were taken. Breakthroughs were determined by extracting spiked

40 Chapter 2

wastewater samples (duplicate analyses) using two stacked cartridges. A

breakthrough on the first cartridge triggered an enrichment on the consecutive

cartridge, which was then eluted separately. For the 1°EFFL a 250-mL sample

with a spiked analyte concentration of 2 000 ng/L was used, and for the 3°EFFL

a 500-mL sample with a spiked analyte concentration of 5 000 ng/L was

extracted. Complete elution of the cartridges was verified by eluting cartridges

of spiked samples for a second time with 1.5 mL of acetone as a stronger

solvent. The acetone extract was then treated as a separate sample. Instrumental

limits of detection (LODs) and limits of quantification (LOQs) were calculated

on the basis of standard deviation of the repeated measurement (n = 10) of a

standard mixture (100 pg on column). The LOD is defined as 3 times, and the

LOQ as 10 times the standard deviation. If the resulting value for the LOQ was

below the linear range, the lower limit of the linear range was set as LOQ.

Sample-based LOD and LOQ were defined as concentrations in a sample matrix

resulting in peak areas with signal-to-noise ratios (S/N) of 3 and 10,

respectively. Since samples typically contained analytes in higher amounts, the

concentration corresponding to the defined S/N was determined by scaling

down, using the measured concentration and the assigned S/N of the peak -

assuming a linear correlation through zero. Instrumental precision of the

measurement was assessed using an average of 10 independent injections of 100

pg on column of a standard mixture. The precision of the entire method was

determined using four replicates of each matrix investigated, spiked with 50 ng

of analyte prior to extraction. It is indicated by the relative standard deviation of

the measured concentrations of native plus spiked analyte. For recovery studies

over the entire procedure, wastewater samples (duplicate analyses) were spiked

prior to extraction with surrogate standard and with 25 and 50 ng of analytes,

respectively. The calculated amount of antimicrobials minus the amount already

present before spiking was then divided by the spiked concentration.

2.2.7 Identification and Quantification

For each substance, two transitions of the precursor ion were monitored.

Together with the retention times, they were used to ensure correct peak

assignment and to evaluate peak purity. For instrumental and surrogate

standards, peak purity was tested using the area ratio of the two product ions

monitored. Their individual ratio was calculated as well as the mean ratio of all


samples and its relative standard deviation. The ratio in one sample was

compared to the mean ratio of all the samples measured in one series. The

variance had to be within the range of twice the standard deviation of the mean

sample ratio. The peak purity of the analytes was tested by calculating a

concentration (as described below) for both product ions measured. The

respective surrogate product ion was used. If no surrogate product ion resulting

from the same fragmentation reaction can be used, i.e., if no isotope labeled

surrogate standard is available, the sum of both product ions of the compound

assigned as surrogate standard was used for quantification to simplify the

procedure. The relative average deviation of the calculated concentrations from

the two product ions had to be less than 10%. Peaks not fulfilling the

requirements for peak purity were not used for quantification.

Quantification was performed using the ratio of the peak areas of the analytes

and of the surrogate standard. The sum of the two monitored product ions was

used. An external calibration curve, plotting ratio against concentration, was

obtained by diluting standards in HPLC water. A standard curve was acquired at

the beginning, at the end, and also in the middle of a measurement series. At

least five concentration points in the appropriate concentration range were used

for quantification.

Concentrations in the samples were calculated by comparing the peak area ratios

of the analytes and their assigned surrogate standards in the SPE extracts, to the

corresponding ratios in the standard solutions. These results were corrected with

the corresponding recovery rates obtained in the same matrix and sample batch

to provide accurate amounts. For routine determination, duplicate analyses of all

samples were performed. Procedural blanks, consisting of de-ionized water,

were analyzed with each set of 12 extractions as a control for laboratory

contamination. Additional instrumental blanks using de-ionized water were

checked with each calibration curve in order to uncover potential analytical

interferences.



The crucial parameters for enrichment, separation, and detection of the analytes

were identified and optimized. The pH of the sample proved to be the most

42 Chapter 2

influential variable during sample extraction. A critical impact on the retention

of the analytes on the cartridge material was observed, especially for

sulfonamides caused by their amino groups. Our enrichment tests between pH 2

and 6 revealed, as expected, highest recoveries at pH 4 for the sulfonamides,

while the recovery of the macrolides and trimethoprim showed no strong pH

dependence. This behavior can be explained by the charge state of the

sulfonamides at the particular pH values (Table 2.1).[1M6] With a compound

specific pKa of 5 - 8 for the sulfonamino groups (pKa 1) and a pKa of 2 - 2.5 for

the arylamin (pKa 2), the sulfonamides are positively charged at pH 2 and

negatively at a pH above 5. The interaction with the cartridge material is

strongest for analytes in uncharged forms occurring at a pH of -4 in the case of

the sulfonamides. The dilution of the 1°EFFL samples prior to enrichment

additionally increased signal intensity provided by the mass spectrometer for the

sulfonamides - in most cases by a factor of 2. For the macrolides and

trimethoprim, no significant improvement was observed.

While A^-acetylsulfamethoxazole is stable during sample preparation,

erythromycin present in the samples is completely transformed to dehydro-

erythromycin at pH 4. This is in agreement with the reported instability of

erythromycin under acidic conditions resulting in the formation of the inactive

dehydro-erythromycin.[16] Erythromycin was therefore assessed as the main

environmental metabolite, dehydro-erythromycin.

Tandem mass spectrometric conditions were optimized for each analyte and

internal standard through automated tuning procedures implemented in the

instrument software. Figure 2.4 shows the breakdown curves for A^-acetyl-

sulfamethoxazole and its four most intense fragments as a function of the

collision energy. As expected, the collision energy, which gives the most intense

signal, increases for the formation of smaller fragments. Tentative product ion

structures are given also. These structures have not been reported previously but

are in agreement with transitions known to be typical for sulfonamides. '

During LC/MS/MS measurement, matrix compounds can be deposited on the

instruments sample interface, especially on the ion-transfer capillary, and can

thus significantly reduce instrument sensitivity. The higher the sample volume

the more matrix will be introduced into the mass spectrometer within one run.

On the other hand, a high enrichment factor is desirable to achieve the low limits

of detection, which are necessary for the environmental analysis of

antimicrobials. The sample volume was optimized by using a variable splitting


Figure 2.4 Breakdown curves of N*-acetylsulfamethoxazole and proposed

product ion structures (absolute intensity 3.56e+0.5, collision pressure 1.5

mTorr)

IM

#

I

>

t8 4*-

«

1» 20

i 1 r

JO 40

Collision Energy (V)

o

-i r-=N

\ /©/

S -N.

[M+H]+m/z=296

H_

/H2N=/ N ©

O

v_/ \

m/z=65 m/z=108 m/z=134 m/z=198

if TT Chapter 2

device prior to the electrospray interface. For this experiment, higher sample

volumes were chosen. Therefore, 200 mL of 1°EFFL and 1 000 mL of 2°EFFL

and 3°EFFL, respectively, were enriched and measured in one series. The eluent

flow and split ratio were varied, so that the instrument remained sensitive

enough for the measurement of up to 30 samples of each matrix. The sample

volume used was then adjusted according to the split ratio. Subsequent samples

were measured without the additional splitting device that may pose as a source

for errors (for example, plugging of the capillary). The given sample volumes

therefore represent a compromise between method sensitivity and routine

analysis. Two columns were tested for the separation of the selected

antimicrobial agents. In both cases, a reversed-phase end-capped Ci8 column

was chosen - one belonging to the older (column 1) and one belonging to the

newer (column 2) generation of silica gels. On column 1, azithromycin produces

a peak with substantial tailing. To our knowledge, azithromycin has not been

included in analytical methods for environmental samples so far, likely also due

to analytical difficulties like this. The observed tailing on column 1 is probably

due to the interaction of the two basic functional groups with residual silanol

groups and metal impurities of the column material. In the case of the other

macrolides, which contain one amino group less than azithromycin in the

lactone moiety, this interaction was sufficiently suppressed by the addition of

ammonium acetate. On column 2, belonging to the new generation of silica gels

using a more efficient purification and end-capping procedure, however, good

separation was achieved with almost symmetrical peaks for all analytes. In

addition, the use of ammonium acetate in the eluent was no longer necessary.

This significantly increased the sensitivity of the method for sulfonamides,

which tend to form ammonium adducts during ionization.

In the case of some 1 °EFFL samples, the extracts needed to be diluted up to five

times in order to obtain good peak shapes for azithromycin, which seems to

form complexes with matrix compounds. For most analytes, the assumed loss of

sensitivity, due to dilution is partly compensated by the simultaneous reduction

of ion suppression, since signal intensities observed are reduced to a lesser

extent.



The developed method was validated for primary effluents after mechanical

treatment, secondary effluents after biological treatment, and tertiary effluents

after sand filtration. For breakthrough studies, samples representing unnaturally

high concentrations and high loads of sample matrix were enriched on two

stacked cartridges. No quantifiable amounts of the analytes could be detected on

the second cartridge for both sample matrices (1°EFFL and 3°EFFL). When

testing for complete elution, no quantifiable amounts of analytes could be

measured in the acetone eluates of already eluted cartridges. Thus, the analytes

are quantitatively enriched by one cartridge and exhaustively eluted by the

procedure described above.

For the standard curves good linearity was observed with correlation factors

typically above 0.99. The linear range of the measurement varied with the

analyte due to differences of the ionization efficiencies (Table 2.3). The

instrumental LOQ ranges between 16 and 100 pg of analyte on column. In the

case of the sample-based LOQ and LOD the range and the average of the

resulting values for each matrix from different samples are given in Table 2.3.

Since the LOD and LOQ in an individual sample can be higher or lower than the

average LOD and LOQ, all peaks with an assigned S/N greater than or equal to

3 and 10, respectively, are considered valid.

The instrumental precision of the method was addressed for various aspects, and

the following relative standard deviations were obtained: the retention time

ranged between 0.06 and 0.35% and the peak area between 1.3 and 9.2%. The

peak area ratios of analyte versus surrogate standard varied to a lesser extent in

most cases (between 1.3 and 7%), since the surrogate standard compensates for

analytical variability. The precision of the entire method (reproducibility) is

indicated by the standard deviation of multiple analyses and ranged between 0.5

and 15%. Detailed results are shown in Table 2.4.

Accuracies of the method were determined by relative recovery studies over the

entire procedure (Table 2.4). The resulting recoveries obtained in all matrices

investigated were generally above 80%, with the exception of TRI where they

ranged between 30 and 47%. For TRI, this was caused by the use of a nonideal

surrogate standard (13C6SMZ), but none better suited could be found.

Recoveries, and thereby LODs and LOQs, of the analytes vary between samples,

mainly due to varying matrix effects, if no isotopically labeled surrogate

Table2.3Linearrangesand

limitsofquantification

(LOQ)

samplebasedLOQ

a

(ng/

L)

linearrange

primaryeffluent

seco

ndar

yeffluent

tertiary

effluent

anal

yte

(pgoncolumn)

average

range

average

range

average

range

SDZ

20-

30000

68

60-77

11

6-32

75-11

STZ

20-

30000

214

194-236

21

15-30

16

10-22

SMZ

10-30000

42

35-48

97-12

94-17

SPY

50-

30000

96

51-150

14

8-31

96-19

SMX

10-30000

62

35-

104

12

8-17

11

6-15

N4AcSMX

100-30000

212

155-288

23

17-29

22

16-28

TRI

5-

2000

21

15-27

64-9

43-7

AZI

10-4000

74.6-9.9

32.0-3.4

21.6-2.6

ERY-H20

5-

4000

19

4.4-13

53.5-8

63.4-9.5

CLA

10-6000

42.3-5.7

21.3-2.9

21.3-3.1

ROX

10-4000

31.2-6.3

10.3-2.0

10.4-1.4

Concentrationestimatedfrommeasuredsamplesforasi

gnal

-to-

nois

eratioof10.

Table2.4Methodpr

ecis

ions

,accuracies,andcombinedmeasurementuncertainties

precisio

naccuracy

combinedmeasurement

relativeSD

a

(%),

nb=4

relativerecovery±SD

a

(%),

nb=4

unce

rtai

nty(%)

prim

ary

secondar

ytertiary

prim

ary

seco

ndar

ytertiary

primary

secondary

tertiary

anal

yte

effluent

effluent

effluent

effluent

effluent

effluent

effluent

effluent

effluent

SDZ

3.4

0.5

1.0

102±2

95±2

98±5

3.1

2.7

4.0

STZ

0.7

5.2

2.0

101±2

102±2

103±3

2.6

3.3

5.7

SMZ

1.1

1.6

2.1

98±1

98±3

98±6

2.4

3.9

4.9

SPY

4.3

4.5

2.2

108±4

101±5

106±1

6.1

3.8

13

SMX

3.0

0.4

1.4

101±3

105±10

105±6

5.2

3.6

3.4

N4AcSMX

0.7

2.4

4.2

91±4

100±7

93±6

3.8

4.4

5.7

TRI

8.5

2.6

12

47±3

30±12

35±10

10

8.0

13

AZI

2.4

5.7

11

83±7

85±4

86±10

9.4

11

16

ERY-H20

5.6

4.1

8.9

91±1

82±1

86±8

4.8

3.8

15

CLA

5.9

3.0

15

89±8

81±8

78±6

6.5

7.4

12

ROX

6.6

5.9

15

100±5

108±2

124±3

8.6

8.7

14

aSD=

standarddeviation.

bn=numberofsamples.

48 Chapter 2

standard is available. Correct quantification can still be ensured if recovery

studies are performed in each matrix and sample batch as was the case within

the work presented here. The combined measurement uncertainty was quantified

using data from the method validation as described in Example A4 of the

EURACHEM/CITAG Guide Quantifying Uncertainty in Analytical

Measurement.1341 Therefore, all uncertainty sources were identified and

quantified. The main contributions result from the repeatability of the

measurement, calculated from duplicate sample analysis, and its accuracy,

represented by recovery studies. The relative values of all uncertainty sources

are finally combined using statistical methods. The values for the combined

measurement uncertainty vary between 2.4 and 16% with the analyte and the

matrix investigated (Table 2.4).

2.3.3 Wastewater Application

The developed method was successfully applied to the analyses of wastewater

samples from two urban wastewater treatment plants in Switzerland: WWTP

Kloten-Opfikon, located near the international airport of Zurich (WWTP-K),

and WWTP Altenrhein, which is located in the canton St. Gall close to the

border with Austria (WWTP-A). In both cases, mechanically treated wastewater

(primary effluent) passes through conventional activated sludge treatment,

followed by secondary settling (secondary effluent). After biological treatment,

both treatment plants use sand flltration as a tertiary treatment step (tertiary

effluent). Table 2.5 shows the results obtained from duplicate analyses of 72-h

flow-proportional composite samples of the primary, secondary, and tertiary

effluents. Samples were taken in February 2003. With 54 100 and 40 000

inhabitant equivalents, the two investigated treatment plants are of similar size

and have comparable volumes of wastewater inflow. This is also reflected in the

similar concentration ranges found at each plant, with the exception of

azithromycin. The latter appears to be more frequently used in the catchment

area of WWTP-A. Sulfamethoxazole and clarithromycin were found to be the

most commonly used sulfonamides and macrolides, respectively. Sulfadiazine -

which is very rarely applied in human medicine - could not be quantified in any

of the samples, nor could sulfathiazole - which is almost exclusively used in

veterinary medicine.

Table2.5Sulfonamideandmacrolideconcentrationsmeasured

intwomunicipalwastewatertreatmentpl

ants

inSwitzerland

sampleconcentrationa

(ng/L), n

=2

WWTP-Kc

WWTP-A

e

prim

ary

secondary

tertiary

prim

ary

secondary

tertiary

analyte

effluent

effluent

effluent

effluent

effluent

effluent

SDZ

ndd

nqd

nq

nd

nd

nd

STZ

nd

nd

nd

nd

nd

nd

SMZ

nd

nd

nd

39

18

19

SPY

72

82

88

135

63

85

SMX

343

344

352

641

352

352

N4AcSMX

518

86

82

943

nq

71

TRI

168

170

81

110

86

68

AZI

86

110

85

224

129

255

ERY-H20

67

96

75

44

54

55

CLA

234

374

329

160

188

220

ROX

22

11

10

30

21

23

aConcentrationmeasured

in

filtered72hflowproportional

composite

sample.Averageofdu

plic

atedetermination.

bn=number

of

measurements.cWWTP-K=

Kloten-Opfikon(cantonZu

rich

),WWTP-A=Altenrhein(c

anto

nSt

.Gall).

dnd=not

detected=

sign

al-t

o-no

isebelow

3;nq=not

quantifiable=

sign

al-t

o-no

isebelow

10.

50 Chapter 2

Sulfamethazine - another sulfonamide used mainly for veterinary applications -

was only present in one of the treatment plants (WWTP-A). A^-acetyl-

sulfamethoxazole is typically present in high amounts in the primary effluents,

but only small amounts can be found in the tertiary effluents. If the amount of

sulfamethoxazole present as acetyl metabolite is neglected, the elimination of

sulfamethoxazole will be underestimated. Concentrations of the analytes in both

tertiary effluent range between 19 and 352 ng/L. This clearly shows that the

compounds investigated are not eliminated completely and reach receiving

surface waters. Compared to results obtained in Germany/20'35 the

concentrations found are in the same range but generally lower.

2.4 Conclusions

Solid-phase extraction Oasis HLB cartridges coupled with reversed-phase liquid

chromatography and tandem mass spectrometry were successfully applied for

the determination of selected sulfonamides and macrolides, in addition to

trimethoprim and the human sulfonamide metabolite A^-acetylsulfamethoxazole,in municipal wastewater. As a result of this method's applicability to wastewater

samples spanning the whole treatment process (including primary effluent

samples), it can be used to investigate the fate of these compounds through the

various steps of wastewater treatment. The resulting information can be used to

evaluate the performance of wastewater treatment procedures and to highlight

options for the optimization ofWWTPs with the aim of minimizing the input of

pharmaceuticals into ambient receiving waters. Additionally, by including N4-

acetylsulfamethoxazole - the main human metabolite of sulfamethoxazole - the

fate of the most commonly used sulfonamide in human medicine can be

investigated more thoroughly.

The presented method provides the necessary basis for a comprehensive study

on antimicrobials in wastewater treatment including alternative wastewater

technologies such as biofilter technology and membrane flltration.[36] Our

ongoing studies are also aimed at achieving complete mass balances of

antimicrobials in wastewater treatment plants, including sewage sludge

treatment steps. Preliminary results show that the method can easily be adapted

for the analyses of sewage sludge extracts. Applications to drinking water,

ambient waters, and hospital wastewaters also seem to be possible judging from

first measurements without major changes in the procedure. With this method,


we therefore present a powerful tool to fully assess the fate and occurrence of

macrolides and sulfonamides throughout their main pathways to and within the

aquatic environment.

Acknowledgments

Abbott GmbH (Wiesbaden, Germany) is acknowledged for supplying

clarithromycin and Pfizer AG (Zurich, Switzerland) for supplying azithromycin.

Partial financial support came from the EU project POSEIDON (EVK1-CT-

2000-00047)[37] and the EAWAG project on human-use antibiotics

(HUMABRA) within the framework of the National Research Program on

antibiotic resistance funded by the Swiss National Science Foundation.[38J We

thank Eva Molnar, Norriel Nipales, and René Schönenberger for their technical

assistance and advice. For helpful comments on the manuscript, we

acknowledge our colleagues Alfredo Alder, Michael Dodd, Stephan Müller, and

Krispin Stoob.


[I] Stan, H. J.; Heberer, T. Analusis 1997, 25, M20-M23.


[3] Halling-Serensen, B.; Nors Nielsen, S.; Lanzky, P. F.; Ingerslev, F.; Holten

Lützenhoft, H. C; torgensen, S. E. Chemosphere 1998, 36, 357-393.


938.






1998.




[II] Vree, T. B.; Hekster, Y. A. Clinical pharmacokinetics ofsulfonamides and


i

52 Chapter 2

[12] Lin, C.-E.; Chang, C.-C; Lin, W.-C. J. Chromatogr., A 1997, 768, 105-

112.

[13] Petz, M., Habilitation Thesis, Westfälische Wilhelms-Universität, Münster,

1986.

[14] Neuman, M. Antibiotika-Kompendium; Verlag Hans Huber: Bern, 1981.

[15] McFarland, J. W.; Berger, C. M.; Froshauer, S. A.; Hayashi, S. F.; Hecker,

S. J.; Jaynes, B. H.; Jefson, M. R.; Kamicker, B. J.; Lipinski, C. A.; Lundy,

K. M.; Reese, C. P.; Vu, C. B. J. Med. Chem. 1997, 40, 1340-1346.



[17] Hirsch, R.; Ternes, T. A.; Haberer, K.; Mehlich, A.; Ballwanz, F.; Kratz,

K.-L. J. Chromatogr., A 1998, 815, 213-223.


2001,938,199-210.

[19] Lindsey, M. E.; Meyer, M.; Thurman, E. M. Anal. Chem. 2001, 73, 4640-

4646.

[20] Hartig, C; Storm, T.; Jekel, M. J. Chromatogr., A 1999, 854, 163-173.


Technol. 2003, 37, 5479-5486.

[22] Giger, W.; Aider, A. C; Golet, E. M.; Kohler, H.-P. E.; McArdell, C. S.;


[23] Vanderford, B. J.; Pearson, R. A.; Rexing, D. J.; Snyder, S. Anal. Chem.

2003, 75, 6265-6274.

[24] Golet, E. M.; Aider, A. C; Hartmann, A.; Ternes, T. A.; Giger, W. Anal.

Chem. 2001, 73,3632-3638.


5455-5462.

[26] Golet, E. M.; Aider, A. C; Giger, W. Environ. Sei. Technol. 2002, 36,

3645-3651.

[27] Golet, E. M.; Xifra, 1.; Siegrist, H. R.; Aider, A. C; Giger, W. Environ. Sei.

Technol. 2003, 37, 3243-3249.

[28] Vree, T. B.; Hekster, Y. A. Pharmacokinetics of sulfonamides revisited;


[29] Berger, K.; Petersen, B.; Büning-Pfaue, H. Arch. Lebensmittelhyg. 1986,

37, 85-108.

[30] Hilton, M. J.; Thomas, K. V. J. Chromatogr., A 2003,1015, 129-141.


[31] Neumann, J. "Untersuchungen zur BioVerfügbarkeit und Pharmakokinetik

von Sulfasalazin und seinen Metaboliten," PhD Thesis, Free University of

Berlin, 1989.

[32] Volmer, D. Rapid Commun. Mass Spectrom. 1996, 10, 1615-1620.


J. Chromatogr., A 2002, 952, 111-120.

[34] http://www.measurementuncertainty.org/mu/guide/index.html, Quantifying

uncertainty in analytical measurement / prep, by the EURACHEM

Working Group on Uncertainty in Chemical Measurement (ISBN 0 948926

15 5).


1999,225,109-118.

[36] Göbel, A. PhD Thesis, ETHZurich, No. 15703 2004.

[37] http://www.eu-poseidon.com.

[38] http://www.nrp49.ch/pages/.

Seite Leer /Blank 1er?

Chapter 3

Analytical Method for Sewage Sludge

Pressurized liquid extraction (PSE) was optimized and validated for the

determination of sulfonamide and macrolide antimicrobials and trimethoprim in

sewage sludge samples. A mixture of water/methanol (50:50 v/v) was found as

the most efficient extraction solvent. A temperature of 100 °C and a pressure of

100 bar were chosen for extraction. Two cycles of 5 minutes each, efficiently

extracted at least 97% of all studied analytes from activated sludge. The limits of

quantification (S/N = 10) varied between 3 and 41 pg/kg dry weight (dw).

Additionally the influence of pH and analytical method on the absolute

recoveries was assessed. Of the investigated antimicrobials sulfapyridin,

Sulfamethoxazol, trimethoprim, azithromycin, clarithromycin, and

roxithromycin were detected in municipal sewage sludge samples.

Concentrations in activated sludge ranged up to 197 pg/kg dw. In comparison,

results obtained by ultrasonic solvent were generally lower in the case of

sulfonamides and in tendency lower for macrolides.

Göbel, A., Thomsen, A., McArdell, CS., Alder, A.C., Giger, W., Theiß, N.,

Löffler, D., Ternes, T.A.

Extraction ofSulfonamide and Macrolide Antimicrobialsfrom Sewage Sludge

submitted to Journal of Chromatography A

Analytical Methodfor Sewage Sludge 57

3.1 Introduction

Antimicrobial agents are widely used in human and veterinary medicine. The

overall human consumption of antimicrobials amounts to over 30 tons per

annum (t/a) in Switzerland and over 400 t/a in Germany - resulting in a similarnil

consumption of approximately 5 g per person and year in both countries."

Sulfonamides (16-21% of the total human consumption) and macrolides (9-

12%) belong to the most important groups of human used antimicrobials

following the ß lactams (50-60%).

Human used pharmaceuticals, including antimicrobial agents, are excreted,

unchanged or metabolized, from the patients' body. They therefore mainly reach

wastewater treatment plants (WWTPs) through household wastewater. The

occurrence and fate of pharmaceuticals in WWTPs and receiving surface waters

has hence been of increasing interest in recent years.[410] In the case of

antimicrobials this is also motivated by the possible maintenance and spread of

resistance caused by the constant input of low concentrations of antimicrobials.

They have been detected in WWTP effluent and receiving surface waters

illustrating the importance of WWTPs as point sources and the almost

ubiquitous presence of these emerging contaminants.[1M6] The occurrence of

macrolides and sulfonamides in WWTP effluents also indicates an incomplete

removal during conventional wastewater treatment. No distinction between

sorption and degradation can be made since the studies performed so far focus

on the fate and occurrence in the aqueous phase, except for fluorochinolones.

Golet et al. showed that sorption to sludge is the main removal route of the

highly polar fluorochinolones in wastewater treatment.ll7] This clearly illustrates

the need for analytical methods for sewage sludge when assessing the fate and

occurrence of contaminants in wastewater treatment. Methods published so far

for the determination of other antimicrobials in environmental biosolids focus on

the veterinary use and on the spread of contaminated manure onto soil.

Analytical methods and studies performed range from animal food products[18]

over manure[19"21]

to soil.[2228] Additionally river sediments [29] and meat from

production animals [30'31]were analyzed for sulfonamide and/or macrolides. A

review on part of the literature available can be found in [32]. In most cases the

compounds of interest were extracted from the samples by ultrasonic solvent

extraction (USE) or blending with a suitable solvent. USE thereby represents a

simple and relatively low priced approach. In a few cases, pressurized liquid

58 Chapter 3

extraction (PSE), also known as accelerated solvent extraction (ASE; Dionex),

was applied.[17,25] Using PSE the sample is extracted under high pressure and

high temperature to enhance solubility and mass transfer effects. ] Further

advantages of PSE are the minimal solvent usage and automation, which enables

the simultaneous extraction of a high number of samples.

In this study we aimed at developing a sensitive and reliable method for the

extraction of macrolides, sulfonamides and trimethoprim from activated and

digested sewage sludge (Figure 3.1). By comparing different extraction

approaches (PSE and USE) and the application of different analytical methods

in two different laboratories, an expanded validation of the method is achieved.

Results from the analysis of municipal activated and digested sludge samples

from Germany and Switzerland are given to show the applicability of the

methods presented.


3.2.1 Chemicals and Reagents

HPLC-grade methanol, acetonitrile, and water were purchased from Scharlau

(Barcelona, Spain). Analytical grade ethyl acetate, acetone, ammonia solution,

25% sulfuric acid, sodium chloride, sodium hydroxide, ammonium acetate, and

formic acid were obtained from Merck (Darmstadt, Germany).

Sulfamethazine, sulfamethoxazole, sulfadiazine, oleandomycin and

roxithromycin were purchased from Sigma-Aldrich (Buchs, Switzerland).

Sulfathiazole, sulfapyridine, trimethoprim, tylosin, josamycin, and erythromycin

were obtained from Fluka Chemicals (Buchs, Switzerland) and sulfamerazine

from Riedel-de Haën (Seelze, Germany). Sulfamethazine-phenyl-13C6 was

purchased from Cambridge Isotope Laboratories (Andover, MA, USA) and

sulfamethoxazole-*/, sulfadiazine-«/, sulfathiazole-d4 as well as A^-

acetylsulfamethoxazole-öf5 were purchased from Toronto Research Chemicals

(North York, ON, Canada). Clarithromycin was kindly supplied by Abbott

(Wiesbaden, Germany) and azithromycin by Pfizer (Zurich, Switzerland).

Azithromycin is also available from Sigma-Aldrich (Buchs, Switzerland).

Standard solutions for dehydro-erythromycin were prepared from erythromycin

as described by McArdell et al.[16] The acidic solution was readjusted to pH 6

after 4 h using IM NaOH to ensure stability during storage.


Figure 3.1 Chemical structures ofthe investigated sulfonamides, macrolides

and trimethoprim

P—Rl

""*o v^v N\

erythromycin (ERY)CASRN 114-07-8

roxithromycin (ROX)CASRN 80214-83-1

Ri

H

clarithromycin (CLA) CH3

CASRN 81103-11-9

R2

O

O

H/^/V-N/

\t

, .„.111*

'V/^OH \>

H3C^"*/''CH3

'"///,

azithromycin (AZI)

CASRN 83905-01-5

o^v^X \I \ CH3

0B-

^zS°~\

\dehydro-erythromycin

(ERY-H20)»««

VV v*'*«h'CH3

'"*/(

? HO

/N.

\ GH3

-0„

o^~~~^ZSw

H7N-

jj R3

\ nr\.

R.

N .NH2

trimethoprim (TRI)CASRN 738-70-5

sulfadiazine (SDZ)CASRN 68-35-9

sulfathiazole (STZ)CASRN 72-14-0

sulfamethazine (SMZ)CASRN 57-68-1

sulfapyridine (SPY)CASRN 144-83-2

sulfamethoxazole (SMX)CASRN 723-46-6

( /N—,J

CH3

N=<

CH3

N=\

-o

Tu-O

60 Chapter 3


Grab samples were taken from the end of the nitrification compartment at

different municipal WWTPs in Germany and Switzerland (activated sludge). All

plants consist of primary clarification and a denitrification - nitrification cascade

with an internal recirculation of sludge as secondary treatment. Phosphate

removal is performed by the addition of iron salts to different treatment steps. In

WWTP-W, located at Wiesbaden, Germany, serving 350 000 population

equivalents (PE), Fe(II)Cl2 is added to the final clarification. Simultaneous

precipitation with Fe3+ in secondary treatment is performed at WWTP-K,

located in Kloten-Opfikon, Switzerland, near the international airport of Zurich

(55 000 PE), and at WWTP-A, located in Altenrhein, Switzerland, close to the

border with Austria (40 000 PE). Additionally, a grab sample was collected from

outlet of the anaerobic, mesophilic digester at WWTP-K containing a mixture of

primary and secondary sludges (digested sludge).

Activated sludge samples were filtered through glass fiber filters (GF8,

Whatman) and the solid fraction was frozen. Digested sludge was directly frozen

without filtration. Samples were subsequently freeze-dried and finely ground in

a mortar. They were stored in amber glass bottles at -25 °C until analysis (up to

2 years). Consequently, the results obtained for activated sludge are given in

pg/kg dry weight (dw), while those for digested sludge, including the aqueous

phase, are given in pg/L. The concentration of solids in the freeze-dried digested

sludge was determined to be 17 ± 6 g/L.

3.2.3 Sample Preparation

For USE an aliquot (500 mg) of freeze-dried sludge was successively extracted

with 4 mL and 2 mL methanol and then two times with 2 mL acetone (Table

3.1). In each extraction step, the sample slurry was ultrasonicated for 5 min.

Surrogate standards (see 3.2.4) were spiked into the slurry of the first methanol

extraction before ultrasonication. The slurries were centrifuged at 3 000 rpm for

5 min after each extraction step and the supematants collected. The solvent of

the combined supematants was evaporated to a volume of -200 pL, which was

then diluted with 150 ml of local ground water for solid phase extraction as a

clean-up step.


Table 3.1 Extraction procedurefor sulfonamides, macrolides and trimethoprim

from activated sludge

pressurized liquid ultrasonic solvent

parameter extraction extraction

(PSE) (USE)

sample amount 200 mg 500 mg

solvent methanol : water methanol

(1:1, v:v) acetone

time 3 cycles of 5 min 4 times for 5 minutes

(preheating time 5 min) (4 mL methanol, 2 mL methanol,

2 mL acetone, 2 mL acetone)

temperature 100 °C -

pressure 100 bar -

flush 120% of cell volume

for all 3 cycles

-

nitrogen purge 60 sec -

For PSE samples of freeze-dried sludge were weighed (200 mg) and transferred

into 11-mL extraction cells (Dionex) partly filled with quartz sand. During

mixing, more sand was added until the cell was completely filled. For extraction

an automated Dionex ASE 200 accelerated solvent extractor (Sunnyvale, CA,

USA) equipped with a solvent controller was used. A methanol-water mixture

(50/50, v/v) proved to be optimal as extraction solvent. An extraction

temperature of 100 °C and an extraction pressure of 100 bar were chosen as

operating conditions (Table 3.1). Preheating time and static time were set to 5

minutes each. A total flush volume of 120% the cell volume and a purge time of

60 sec with nitrogen was used. The final extraction volume was ~ 22 mL with 3

extraction cycles for activated sludge and 2 for digested sludge. The PSE

extracts were completely transferred to 500 mL amber glass bottles by rinsing

the collection vial with -100 mL of de-ionized water in 3 steps. They were

further diluted with -350 mL de-ionized or local ground water to reduce the

methanol content of the sample for solid-phase extraction to below 5%.

Surrogate standard (see 3.2.4) was spiked directly on the sludge in the extraction

cell or in the PSE extract prior to dilution.

62 Chapter 3

The respective extracts of both extraction methods were adjusted to pH 4 with

sulfuric acid or directly enriched (pH -7). Solid-phase extraction was performed

on 6 mL Oasis HLB sorbent cartridges (200 mg) (Waters, Bergen op Zoom, The

Netherlands). Detailed information on sample preparation can be found in

Chapter 2.[34]

3.2.4 ChemicalAnalysis

Different methods were used at two different laboratories for the separation and

detection of sulfonamides, macrolides and trimethoprim in sludge extracts

(Figure 3.2), both based on the method published for aqueous wastewater

samples.[34'35J In method 1, separation was achieved using a 150 x 2 mm YMC

Pro CI8 column (120 Â, 3 pm, Stagroma, Reinach, Switzerland) and a mobile

phase of methanol-water containing 1% (v/v) formic acid. Gradient elution was

used at a flow rate of 150 pl/min. A triple quadrupole mass spectrometer, TSQ

Quantum Discovery (Thermo Finnigan, San Jose, CA, USA), equipped with

electrospray ionization was used for detection. A spray voltage of 3 500 V and a

capillary temperature of 350 °C were applied. Analyses were performed in the

positive multiple reaction mode using two transitions per analyte. An external

calibration curve in de-ionized water was used for quantification. Results were

corrected by relative recoveries over solid-phase extraction and measurement

determined in the same experimental series. Therefore the following substances

(100 ng) were added to the ASE extracts: d4SDZ, d4STZ, d4SMX, ,3C6SMZ as

surrogate standards for sulfonamides and trimethoprim and tylosin (TYL) as

surrogate standard for macrolides.

In method 2, separation was achieved on a 100 x 4.6 mm Chromolith

Performance RP-18e column at a flow rate of 400 pL/min and a total mn time of

50 min. Gradient elution was performed with solvent A (water containing 10%

acetonitrile and ammoniumacetat (lOmM)) and solvent B being a mixture of

80% acetonitrile and 20% solvent A. Initial conditions were set to 100% A.

After 10 min the percentage of B was increased to 26% within 5 min and to 38%

in the following 2 min. After 7 min of 38% B, the percentage of B amounts to

100% in a time span of 6 min, where it was kept for 4 min. Within 2 min initial

conditions were restored and mn for another 14 min. Detection was performed

using a triple quadmpole mass spectrometer, API 4000 (Applied Biosystems,

Foster City, CA, USA), equipped with electrospray ionization. An ion source


Figure 3.2 Scheme for the extraction and analysis ofsulfonamides, macrolides

and trimethoprim in sewage sludge

0.5 g activated or digested sludge

(freeze-dried)

0.2 g activated or digested sludge

(freeze-dried)

ultrasonic solvent extraction (USE)see Table 1

(addition of surrogate standards prior to extraction)

pressurized liquid extraction (PSE)see Table 1

(addition of surrogate standards after (method 1 ) or prior

(method 2) to extraction )

dilluted sludge extract

(pH 4 or pH 7)

solid phase extraction

Oasis HLB, 6 mL, 200 mg

elution: methanol/ethyl acetate/ammonia

evaporation to -50 uL by N2-stream(addition of instrumental standard)

Method 1

liquid chromatography

column: YMCproC18

gradient: water/methanol/

formic acid


(TSQ Quantum Discovery)

Method 2

liquid chromatography

column: Chromolith Performance

RP-18e

gradient: water/acetonitrile/

ammonium acetate


(API 4000)

voltage of 5 000 V and a temperature of 750 °C were applied, while the

declustering potential was compound dependent and ranged between 56 and

106 V. Analyses were performed in the positive multiple reaction mode using

two transitions per analyte. Quantification was performed using an internal

calibration curve in local ground water. Surrogate standard (100 ng) was added

prior to extraction. d4SMX was used as surrogate standard for all sulfonamides

investigated and oleandomycin (OLE) for all macrolides. No surrogate standard

was used for azithromycin and sulfapyridine, which were subsequently

quantified by comparing peak areas of the samples and the calibration. In the

case of sulfapyridine, results obtained were additionally corrected by the

respective absolute recoveries obtained from spiking experiments.

64 Chapter 3

In both methods, the SPE extracts were mostly diluted up to 10-fold with de-

ionized water prior to measurement. Filtration of the final extracts prior to

measurement, lead to significant losses of the analytes, especially in the case of

the macrolide antimicrobials.

3.2.5 MethodDevelopment

For PSE method development, activated sludge from WWTP-K was filtered and

the solid fraction was spiked with an aqueous solution raising the concentration

of analytes by approximately 400 pg/kg dw .The mixture was stirred manually

for Vi h and subsequently freeze-dried. This was considered to be the best

substitute for native sludge where the interaction between compounds and

sludge may be different due to aging effects. Spiking was necessary since not all

compounds investigated were present in the sludge sample taken. By varying

extracting conditions the following parameters were optimized by duplicate

analyses in the order given: extraction solvent (9 solvents and mixtures),

extraction temperature (60/80/100/150/200 °C), cycle time (1/3/5/10/20 min),

extraction pressure (60/80/100/120/150 bar) and sample amount (100/200/400

mg). Multiple sequential extraction (4x5 min, n = 2) of the same sludge sample

(activated and digested) was performed to ensure quantitative extraction.

Therefore the extracts of the individual cycles were collected separately. The

maximum extractable amount was defined as the sum of the amounts measured

in the four cycles. The amount recovered in each cycle was expressed as a

percentage of this sum (extraction yield). To assess the stability of the

compounds investigated during PSE extraction, quartz sand as inert matrix was

spiked with analytes (100 ng) and extracted (n = 2) as described.

In the case of the USE method, parameters generally suitable for the extraction

of sewage sludge were chosen (Table 3.1).[36] Exhaustive extraction under the

given conditions was tested by prolonged extraction of activated sludge with

acetone.


Accuracy was assessed by relative recovery studies using area ratios

(analyte/surrogate standard) for quantification. To evaluate the whole method,

freeze-dried activated sludge was spiked prior to extraction in the extraction cell


with analytes (50-100 ng) in methanol and surrogate standard and subsequently

analyzed (n = 2-3). For relative recoveries over solid-phase extraction and

measurement, activated sludge extracts were spiked with analytes (50-100 ng)

and surrogate standard prior to solid-phase extraction (n = 2). The calculated

amount of antimicrobials minus the amount already present before spiking (n =

2-3) was related to the spiked concentration. Absolute recoveries were obtained

using absolute areas instead of area ratios. The areas obtained in spiked

activated sludge (50-100 ng, prior to or after extraction) minus the areas

obtained in the respective non-spiked samples, were compared to the areas

obtained from an external standard with the same concentration as the spike.

Breakthrough of the analytes on the SPE cartridges was determined by the

enrichment of spiked activated sludge (400 pg/kg dw) in duplicate analyses

using two stacked cartridges. A breakthrough on the first cartridge triggered an

enrichment on the consecutive cartridge, which was then eluted separately.

Complete elution of the cartridges was verified by eluting cartridges for a

second time with 1.5 mL acetone as a stronger solvent (n = 2). The acetone

extract was then treated as a separate sample. The precision of the entire method

was determined by extracting replicates (n = 3-6) of spiked activated sludge (90-

500 pg/kg dw). It was defined as the relative standard deviation of the amount

measured. Limits of quantification (LOQ) were defined by two methods. In the

case of PSE, the LOQ was defined as concentrations in a sample matrix

resulting in signals with signal-to-noise (S/N) ratios of 10. The concentration

corresponding to the defined S/N was determined by scaling down, using the

measured concentration and the assigned S/N ratio of the peak - assuming a

linear correlation through zero. Results from several samples (n = 6) were used

to yield an average value. In the case of USE, the second lowest concentration in

the linear range of the internal calibration curve in local groundwater with a S/N

ratio exceeding 10 was used to estimate the LOQ.


3.3.1 MethodDevelopment

For PSE the effect of the different extraction parameters on the extraction

efficiency was evaluated to obtain optimal relative extraction conditions for

sulfonamides, macrolides and trimethoprim from activated sludge (Table 3.1).

66 Chapter 3

Various solvents and mixtures were tested first. Once the optimum solvent

mixture was determined other extraction parameters, such as extraction

temperature and pressure, cycle time, number of cycles and sample amount,

were investigated.

Extraction solvent

Table 3.2 shows the results obtained from using water, organic solvents and

various mixtures as extraction solvent. A total of 10 substances was

investigated, however, only the results of the compounds mainly found in

activated sludge samples are presented: sulfamethoxazole, sulfapyridine,

azithromycin, clarithromycin, roxithromycin and trimethoprim.

Lower extraction efficiencies were observed for all compounds investigated,

especially macrolides, when mixtures of methanol and other organic solvents

(aceton or acetonitrile, 1:1) were used. Water itself proved to be a good

extraction solvent for the sulfonamides but resulted in low extraction

efficiencies for macrolides. More trimethoprim seems to be extracted with

increasing amounts of methanol, whereas no significant influence on the

sulfonamides was observed. For macrolides the highest extraction efficiencies

were observed using a mixture of water and organic solvent at a ratio of 1:1.

This is in accordance with previous findings of Salvatore & Katz[37] that

reported increasing solubility of macrolides to a maximum with increasing

solvent polarity. Mixtures of water with organic solvents other than methanol

(1:1) showed similar results for most analytes but resulted in lower extraction

efficiencies for sulfapyridine and trimethoprim. Methanol-water at a ratio of 1:1

was finally chosen as extraction solvent representing the best compromise for all

compounds investigated. With a pKa of -9 macrolides are weak bases that are

positively charged at neutral pH. Since the surface of most particles in sewage

sludge are negatively charged[381 ionic interactions may play a role in the

sorption of macrolides to sewage sludge. Therefore the effect of the pH of the

chosen extraction solvent was investigated. No significant change in extraction

efficiency for any of the analytes was observed when the pH of the water used

was adjusted to 10 with sodium hydroxide. This may be caused by the buffer

capacity of the sludge or indicate that hydrophobic interactions are

predominantly responsible for the sorption of macrolides to activated sludge.


Table 3.2 Solvent influence on the extraction of sulfonamides, macrolides and

trimethoprimfrom activated sludgea

concentrationb

(pg/kg dw)

extraction solvent SPY SMX TRI AZI CLA ROX

methanol/acetone (1:1) 116 527 138 113 42 112

methanol/acetonitrile (1:1) 120 572 139 133 74 121

methanol 268 594 321 252 180 195

methanol/water (3:1) 282 635 295 260 219 251

methanol/water (1:1) 287 667 225 368 337 351

methanol/water (1:3) 289 663 217 103 339 369

water 291 667 228 33 211 231

water/acetone (1:1) 125 652 144 485 341 364

water/acctonitrile (1:1) 214 698 222 375 314 343

aSelected operating condition in bold letters.

Mean of duplicate analyses using pressurized liquid extraction. Extraction

parameters: 100 °C, 100 bar, 1 cycle of 10 minutes, 150% flush. Extracts adjusted to

pH 4 prior to solid phase extraction. Chemical analysis: Method 1.

Similar conclusions for the macrolide tylosin were made by Tolls,[28] when

investigating the sorption of veterinary pharmaceuticals in soil.

Extraction temperature and pressure

The effect of extraction temperature on the extraction efficiencies of the analytes

turned out to be less profound. An extraction temperature of 100 °C was selected

as operating condition. Slightly lower extraction efficiencies (10 - 20%) were

observed for all analytes at temperatures below 100 °C. However, if the

extraction temperature was increased above 100 °C, the extracted amounts

decreased drastically. Compared to the chosen extraction temperature, only 60 -

80% of most sulfonamides and trimethoprim were measured at an extraction

temperature of 200 °C. For sulfamethoxazole a reduction by 95% was observed

and by 60 - 90% for the macrolides investigated. These findings may be ascribed

to a thermal degradation of the analytes at temperatures above 100 °C.

Additionally, it was observed that increasingly darker extracts were obtained at

higher extraction temperatures, indicating a larger extraction of soluble organic

i

68 Chapter 3

matter. This resulted in problems during solid-phase extraction due to the

clogging of the cartridges. An identical effect was observed when increasing the

extraction pressure from 60 to 150 bar. However, no significant impact of

increasing extraction pressure was observed on the extraction efficiencies of the

compounds investigated (data not shown).

Cycle time and sample amount

A cycle time of 5 min resulted in maximum extraction efficiencies for almost all

compounds. However, the effect of the extraction time observed was low

(variations below 20%) for the investigated sulfonamides, macrolides and

trimethoprim (data not shown). An influence of the cycle time on the extraction

efficiencies may be expected due to the higher extraction temperature used in

PSE resulting in a reduction of the viscosity of the solvent. It may therefore

penetrate further into the sample matrix, a process also facilitated by the

increased pressure. The extraction efficiencies may furthermore be enhanced by

the swelling of the matrix while in contact with the solvent. These processes can

also be influenced by the ratio of sample matrix to extraction solvent. However,

no significant influence on the extraction efficiency of the analytes from varying

sample amounts was observed (data not shown).

Number of cycles

Multiple sequential extractions of the same sample (activated and digested

sludge) were performed to evaluate the ability of the method to quantitatively

extract sulfonamides, macrolides and trimethoprim from the matrices

investigated. For all analytes, except azithromycin, no significant amounts

(>2%) were recovered from activated or digested sludge after the first cycle. As

shown in Figure 3.3 approximately 90% of azithromycin was recovered from

activated sludge in the first cycle. Another 7% were recovered in the second

cycle, whereas the amounts present in the last two cycles were not quantifiable.

Therefore 3 cycles were performed in the analyses of activated sludge to assure

quantitative extraction. In the case of digested sludge 82% of azithromycin was

recovered in the first cycle and another 12% in the second cycle. Even though

small amounts could still be detected in the third (4%) and forth (2%) cycle, two

extraction cycles were chosen for the extraction of digested sludge. The slightly


Figure 3.3 Results for azithromycin from the multiple sequential extraction of

activated and digested sludgea

100

80

S 6092">,

c

o

"S 40

CD

CD

20

h-H

i

I

1

,

activated sludge

Ddigested sludge

*r12 3 4

number of extraction cycle

aError bars represent the range of duplicate analyses. Pressurized liquid extraction:

parameters see Table 3.1. Extracts adjusted to pH 4 prior to solid phase extraction.

Chemical analysis: Method 1.

bPercentage ofthe total amount extracted in the four cycles.

incomplete extraction of azithromycin was neglected since severe problems

were encountered in solid-phase extraction (clogging of the cartridges) and

measurement (bad peak shape) when more than 2 cycles were performed. These

findings indicate that the extraction efficiency of azithromycin varies with the

sample matrix. It has to be noted however, that complete method development

for PSE was performed only for activated sludge.

Also in the case of ultrasonic solvent extraction, exhaustive extraction of

activated sludge was achieved with the chosen parameters (Table 3.1), since no

significant amounts of analyte could be detected in the acetone extract of an

already extracted sample.

Thermal degradation

Since thermal degradation seems to occur at elevated temperatures and longer

extraction times, the stability of the analytes under the chosen extraction

70 Chapter 3

conditions for PSE was of potential concern. However, recoveries from spiked

quartz sand (n = 2) varied around 100% for all substances giving no evidence of

thermal instability. Deviating results were obtained for trimethoprim (150%) and

azithromycin (81 %) and are probably due to a different behavior of the analytes

and the respective surrogate standards (13C6SMZ and TYL) during solid phase

extraction.


Accuracy

The accuracy of the method, expressed by relative recoveries, is influenced by

different parameters, e.g. the suitability of the surrogate standard used or the

method applied for chemical analysis. For pressurized liquid extraction, solid

phase extraction at pH 4 and method 1 for separation and detection for example

the relative recovery ranged between 78 and 106% for the sulfonamides and

trimethoprim and between 91 and 142% for the macrolides (Table 3.3). In that

case no major differences were observed between relative recoveries over the

entire method (including extraction) and over solid-phase extraction and

measurement (excluding extraction). The results from both studies were

therefore combined. The small variations obtained when combining both,

illustrate the thermal stability of the compounds during extraction. Additionally,

they indicate that the analytes spiked on the freeze-dried activated sludge are

extracted quantitatively with the selected extraction conditions. Since spiked

analytes are not exposed to the same active sites as native pollutants this result

cannot be extrapolated to native activated sludge samples. However, quantitative

extraction of native sulfonamides, macrolides and trimethoprim was shown for

activated sludge with the developed method by performing multiple sequential

extraction experiments.

In the case of absolute recoveries no correction by using surrogate standards is

performed. They therefore mirror possible losses during extraction, sample

preparation and variations in measurement due to matrix effects. From the

results obtained during method development and validation it seems that matrix

effects, e.g. ion suppression, are the most important factor. Absolute recoveries

were determined using two different methods for chemical analysis (see 3.2.4),

but the same method for sample preparation. In both cases PSE with identical


Table 3.3 Relative and absolute recoveries for sulfonamides, macrolides and

trimethoprim in activated sludge using method 1 for chemical analysisa

retention relative recoveryb absolute recoveryc

time (%) (%)

compound (min) average

106

t %SD

7

average

63

%SD

SDZ 10.3 6

STZ 12.7 99 5 55 7

SMZ 17.6 97 5 64 17

SPY 12.6 79 5 64 8

SMX 20.4 100 3 64 3

TRI 17.1 78 3 51 4

AZI 21.1 91 10 29 7

ERY-H20 30.1 112 9 37 14

CLA 31.5 110 13 33 24

ROX 31.6 142 16 45 27

a

Pressurized liquid extraction: parameters see Table 3.1. Extracts adjusted to pH 4

prior to solid phase extraction. Chemical analysis: Method 1.

Relative recoveries were determined using area ratios of analyte to surrogate

standard. Average and relative standard deviation (%SD) combing results from

experiments with surrogate standard added prior to or after extraction (n = 4).cAbsolute recoveries were determined using areas. Average and relative standard

deviation (%SD) combing results from experiments with surrogate standard added

prior to or after extraction (n = 4).

parameters was used and the extracts were adjusted to pH 4 prior to SPE.

Results obtained for method 1 are given in Table 3.3, while those for method 2

are included in Table 3.4 (PSE, pH = 4). Similar absolute recoveries were

obtained with both methods for sulfonamides and trimethoprim. In the case of

the macrolides, significantly lower values, and therefore higher ion suppression,

were obtained for method 1 compared to method 2. This could be caused by a

different separation of matrix and analytes during liquid chromatography, i.e. by

the choice of column and gradient. Differences in separation are also mirrored

by the varying retention times of the compounds in the two methods.

Additionally, two different mass spectrometers were used, which may also

Table3.4Absoluterecoveriesforsu

lfon

amid

es,macrolidesandtrimethoprim

inactivatedsludge

usingmethod2forchemical

analyis

absoluterecovery

a

(%)

ASE

bUSE

c

PH==4

pH=--1

pH=4

pH=--

1

compound

RT

average

83

%SD

12

average

54

%SD

6

average

41

%SD

13

average

53

%SD

SDZ

8.6

11

SMX

20.4

37

19

41

416

16

62

7

TRI

20.0

47

744

325

10

31

8

SMX-J4

d20.4

37

15

44

416

11

62

8

CLA

33.4

74

21

90

555

859

15

ROX

33.9

91

33

88

373

10

76

8

OLEd

25.3

93

995

367

557

14

a

Absoluterecoveriesweredeterminedusing

areas.Averageand

relativestandarddeviation(%SD)

isgiven

(n=

3)Re

spec

tive

relative

recoveriescanbecalculatedfromtheabsoluterecoveryratiooftheanalyteand

itssurrogatestandard.

bPressurizedliquid

extraction(P

SE):

parameters

seeTable

3.1.

Extractsadjusted

topH

4priortoso

lid-

phas

eextraction(pH=

4)or

directly

enriched(pH=

7).Chemical

anal

ysis

:Method

2.Surrogatestandardaddedpr

iortoextraction.

0

Ultrasonicsolventextraction(USE):parameters

seeTable

3.1.Extractsadjusted

topH4

priortoso

lid-

phas

eextraction(pH=

4)or

directly

enriched

(pH=

7).Chemical

anal

ysis

:Method

2.Surrogatestandardaddedpr

iortoextraction.

dUsed

assurrogatestandard.


influence the ionization efficiency of macrolides in the samples. Especially, the

differences in temperature applied and the amount of in-source fragmentation

may lead to different ionization efficiencies for the two methods. Further on, the

absolute recoveries were obtained from the analysis of different activated sludge

samples, which also has an effect on the matrix present.

Additionally, the influence of the sample pH during solid phase extraction (SPE)

on the absolute recoveries was investigated. No distinct influence was observed

on the absolute recoveries for the investigated antimicrobials (Table 3.4). The

strong pH dependence of the sulfonamide interaction with the SPE cartridge, as

described for aqueous wastewater samples in Chapter 2, seems not to occur in

sewage sludge extracts. More or less comparable absolute recoveries were also

observed for the investigated compounds at both pH values independently of the

extraction method used. However, a significantly higher relative standard

deviation, of up to 33%, was observed if the pH of the sample was adjusted to 4

prior to SPE. This is caused by an increased clogging of the SPE cartridges at

the lower pH, which made the enrichment of the total sample volume in some

cases impossible.

A dilution of the samples prior to analysis lead to a decrease of matrix effects,

since the areas obtained in diluted samples were reduced to a lesser extent than

expected by the respective dilution factor. In method 1, for example, absolute

recoveries in undiluted samples were 26-50% lower than in 6-fold diluted

samples for sulfonamides. For macrolides and trimethoprim the reduction

ranged between 40 and 80% compared to diluted samples.

Breakthrough and complete elution

Due to the simultaneous extraction of significant amounts of soluble organic

matter during extraction of sewage sludge, breakthrough of the analytes from the

cartridges and complete elution from the cartridges were investigated. No

quantifiable amounts of the analytes could be detected on the second cartridge,

which was eluted separately. When testing for complete elution, also no

quantifiable amounts of analytes could be measured in the acetone eluates of

already eluted cartridges. Thus, the analytes are quantitatively enriched by one

cartridge and exhaustively eluted in the case of activated sludge extracts by the

procedure applied.

74 Chapter 3

Table 3.5 Limits of quantification for sulfonamides, macrolides and

trimethoprim in activated sludge

limit of quantification (pg/kg dw)

pressurized liquid extractiona ultrasonic solvent extraction

compound average range

SDZ 4 3-7 4

STZ 41 31-51 -

SMZ 16 12-20 4

SPY 29 21-36 4

SMX 15 10-23 4

TRI 14 9-17 10

AZI 3 2-4 40

ERY-H20 6 5-8 -

CLA 4 3-6 10

ROX 3 2-4 10

aConcentration estimated from measured samples (Method 1) for a signal-to-noise

ofl0(n=6).bDefined as the second lowest linear concentration of the internal calibration curve in

local groundwater (Method 2).

Precision

Precision was characterized as the relative standard deviation resulting from the

multiple determination of the analytes in activated sludge. It ranged between 2

and 8% for pressurized liquid extraction and between 7 and 20% for ultrasonic

solvent extraction. The higher values for USE are probably caused by a higher

amount of matrix extracted with the solvents used for ultrasonic solvent

extraction. Another reason may lay in the series of manual extraction steps

necessary compared to the fully automated extraction during PSE.

Limit of quantification

The limits of quantification for the analytes in activated sludge were defined

using two different approaches for pressurized liquid and ultrasonic solvent

extraction, respectively (Table 3.5). Overall, it ranges between 3 and 41 pg/kg


dw for the investigated antimicrobials. The differences observed result from a

combination of various factors. Next to the different approaches applied for the

estimation of the LOQ, the higher sample amount used in USE compared to PSE

plays a role. Additionally, differences in the methods used for separation and

detection have an influence, e.g. via peak shape and matrix effects. The results

clearly indicate that the limits of quantification given can only be considered

rough estimates. In routine analysis all peaks with a S/N above 10 were

therefore considered valid results.

3.3.3 Application to Sewage Sludge Samples

The developed methods were applied to selected activated and digested sludge

samples from different wastewater treatment plants in Germany and Switzerland

(Table 3.6). The results for the most commonly detected sulfonamides,

sulfapyridine and sulfamethoxazole, and macrolides, azithromycin,

clarithromycin and roxithromycin, are given. Additionally results for

trimethoprim, used almost exclusively in combination with sulfonamides, are

included. The occurrence of antimicrobials in activated sludge generally

correlates well with the respective aqueous phase.[n'34] Higher concentrations

were generally determined in German activated sludge samples (WWTP-W),

ranging up to 197 pg/kg dw for sulfapyridine, indicating a lower wastewater

dilution compared to Switzerland. A maximum concentration of 73 pg/kg dw

was found for sulfamethoxazole in Swiss samples (WWTP-K and WWTP-A). A

more detailed discussion on the occurrence of sulfonamides, macrolides and

trimethoprim in Swiss municipal wastewater treatment is given in Chapter 4.

Overall, similar results were obtained in activated and digested sludge using

pressurized liquid extraction, independently of the sample pH and the method

used for chemical analysis (Table 3.6). However, using ultrasonic solvent

extraction, the concentrations determined are generally lower for the

investigated sulfonamides and in tendency lower for the investigated macrolides.

This may be caused by the less radical extraction conditions, e.g. temperature

and pressure, compared to pressurized liquid extraction. Additionally, the

extraction conditions used for USE, especially the choice of solvent, were not

optimized particularly for the extraction of sulfonamide and macrolide

antimicrobials.

76 Chapter 3

Table 3.6 Concentrations of sulfonamides, macrolides and trimethoprim in

activated and digested sewage sludge from different wastewater treatment

plants in Germany (WWTP Wiesbaden) and Switzerland (WWTP Kloten-

Opfikon and WWTP Altenrhein)

concentrationa

SPY SMX TRI AZI CLA ROX

activated sludges ug/kg dwb

WWTP-W ASE + pH 4c 57 113 91 127 34 46

sample 1 ASE + pH 7d

51 100 87 158 41 61

USE + pH 7c

26 41 79 127 34 45

WWTP-W ASE + pH 4 197 41 107 151 27 131

sample 2 ASE + pH7 160 37 133 115 16 83

USE + pH 7 85 18 96 47 (9)e 50

WWTP-K ASE + pH 4 29 73 30 52 30 ndg

ASE + pH 7 24 51 (18)f (7)f 25 nd

USE + pH 7 nag 20 14 (21)f 12 nd

WWTP-A ASE + pH 4 (ll)f 60 21 56 63 nd

ASE + pH 7 nd 34 13 (5)' 32 nd

USE + pH 7 nd 27 nd 48 41 nd

digested sludges

ASE + pH 4

ug/Lh

WWTP-K 1.0 nd (0.1)f 2.3 0.8 nd

ASE + pH 7 0.8 nd nd 1.6 0.3 nd

USE + pH 7 1.2 nd nd 1.3 0.3 nd

a

Mean of duplicate analyses for pressurized liquid extraction (PSE) and single

analysis for ultrasonic solvent extraction (USE).b

Separation of solid and aqueous phase through flltration before freeze-drying.cPressurized liquid extraction: parameters see Table 3.1. Extracts adjusted to pH 4

prior to solid-phase extraction. Chemical analysis: Method 1.

dPressurized liquid extraction: parameters Table 3.1. Extracts not pH-adjusted prior to

solid-phase extraction. Chemical analysis: Method 2.

e

Ultrasonic solvent extraction: parameters Table 3.1. Extracts not pH-adjusted prior to

solid-phase extraction. Chemical analysis: Method 2.

fEstimated concentrations below the limit of quantification (S/N < 10).

ê nd: not detected (S/N < 3), na: not analyzedhNo separation of solid (15-18 g/L) and aqueous phase through filtration before

freeze-drying.


3.4 Conclusions

A robust and selective method for the pressurized liquid extraction of

sulfonamides, macrolides and trimethoprim, from sewage sludge was developed

and validated. Several extraction parameters were investigated and the

optimized procedure is summarized in Table 3.1. The method was successfully

applied to activated and digested sewage sludge. Even though comparable

results were obtained for different sample pHs, it is suggested to not adjust the

pH of the extracts prior to solid-phase extraction, to minimize the clogging of

the cartridges. The method presented can be used to investigate the occurrence

and fate of sulfonamides, macrolides and trimethoprim in wastewater treatment,

including the sorption to sewage sludge. Additionally, it may serve as the basis

for the determination of pharmaceuticals in general in sewage sludge and other

biosolids. Ultrasonic solvent extraction seems to be equally or slightly less

efficient for the extraction of macrolides and trimethoprim, while significantly

lower extraction efficiencies seem to result for sulfonamides compared to

pressurized liquid extraction.

3.5 Acknowledgments



Financial support came from the EU project POSEIDON (EVK1-2000-00047,

www.eu-poseidon.com) and the EAWAG project on human-use antibiotics

(HUMABRA, www.nrp49.ch/pages/) within the framework of the National

Research Program on antibiotic resistance funded by the Swiss National Science

Foundation.[40] We thank Elvira Keller, Niccolo Hartmann and Matthias Ruff for

their technical assistance and advice. For helpful comments on the manuscript

we acknowledge M. Suter and M. Ruff.



1998.


78 Chapter 3

[3] Bund/Länderausschuss für Chemikaliensicherheit (BLAC), "Arzneimittel

in der Umwelt - Auswertung der Untersuchungsergebnisse," Hamburg,

2003.

[4] Stan, H. L; Heberer, T. Analusis 1997, 25, M20-M23.



Lützenhoft, H. C; Jorgensen, S. E. Chemosphere 1998, 36, 357-393.


938.

[8] Kümmerer, K. Pharmaceuticals in the environment: Source, fate, effects

and risks; Springer: Berlin, Heidelberg, New York, 2001.


[10] Giger, W.; Alder, A. C; Golet, E. M.; Kohler, H.-P. E.; McArdell, C. S.;



1999,225,109-118.



[13] Sacher, F.; Lochow, E.; Bethmann, D.; Brauch, H.-J. Vom Wasser 1998,

90, 233-243.

[14] Alder, A. C; McArdell, C. S.; Golet, E. M.; Ibric, S.; Molnar, E.; Nipales,

N. S.; Giger, W. In Pharmaceuticals and Personal Care Products in the





3645-3651.


Technol. 2003, 37, 5479-5486.


5455-5462.

[18] Di Corcia, A.; Nazzari, M. J. Chromatogr., A 2002, 974, 53-89.


J. Chromatogr., A 2002, 952, 111-120.

[20] Pfeiffer, T.; Tuerk, J.; Bester, K.; Spiteller, M. Rapid Commun. Mass

Spectrom. 2002,16, 663-669.


[21] Loke, M.-L.; Tjornelund, J.; Halling-Sorensen, B. Chemosphere 2002, 48,

351-361.

[22

[23

[24

[25

[26

[27

[28

[29

[30

[31

[32

[33

[34

[35

[36

[37

[38

Fedeniuk, R. W.; Shand, P. J. J. Chromatogr. 1998, 812, 3-15.

Hamscher, G.; Sczesny, S.; Höper, H.; Nau, H. Anal. Chem. 2002, 74,

1509-1518.

Boxall, A. B. A.; Blackwell, P.; Cavallo, R.; Kay, P.; Tolls, J. Toxicol. Lett.

2002,131, 19-28.

Schlüsener, M. P.; Spiteller, M.; Bester, K. J. Chromatogr., A 2003, 1003,

21-28.

Hartmann, N.; Diploma Thesis, ETH Zurich, Switzerland, 2003.

Thiele-Bruhn, S. J. Plant Nutr. Soil Sei. 2003,166, 145-167.

Tolls, J. Environ. Sei. Technol. 2001, 35, 3397-3406.

Löffler, D.; Ternes, T. A. J. Chromatogr., A 2003,1021, 133-144.

Horie, M.; Takegami, H.; Toya, K.; Nakazawa, H. Anal. Chim. Acta 2003,

492, 187-197.

Van Eeckhout, N.; Castro Perez, J.; Van Peteghem, C. Rapid Commun.

Mass Spectrom. 2000,14, 2331-2338.

Diaz-Cruz, M. S.; Lopez de Aida, M. J.; Barcelo, D. Trends Anal. Chem.

2003,22,340-351.

Dean, J. R. Extraction methodsfor environmental analysis, 1998.

Göbel, A.; McArdell, C. S.; Suter, M. J.-F.; Giger, W. Anal. Chem. 2004,

76, 4756 - 4764.

Hirsch, R.; Ternes, T. A.; Haberer, K.; Mehlich, A.; Ballwanz, F.; Kratz,

K.-L. J. Chromatogr., A 1998, 815, 213-223.

Ternes, T. A.; Bonerz, M.; Hermann, N.; Keller, E.; Bago Lacida, B.;

Aider, A. C. submitted to Journal ofChromatography A 2004.

Salvatore, M. J.; Katz, S. E. J. AOACInt. 1993, 76, 952-956.

Carberry, J. B.; Englande, A. J. Sludge Characteristics and Behavior;

Martinus Nijhoff Publishers: Boston, The Hague, Dordrecht, Lancaster,

1983.

P*" /

Ö

Chapter 4

Occurrence in Wastewater Treatment

The occurrence of sulfonamide and macrolide antimicrobials, as well as

trimethoprim, was investigated in conventional activated sludge treatment.

Average daily loads in untreated wastewater correlated well with those estimated

from annual consumption data and pharmacokinetic behavior. Considerable

variations were found during a day and seasonal differences seem to occur for

the macrolides, probably caused by a higher consumption of these substances in

winter. The most predominant macrolide and sulfonamide antimicrobials were

clarithromycin and sulfamethoxazole, respectively. In the case of

sulfamethoxazole, the main human metabolite, A^-acetylsulfamethoxazole, was

included as an analyte, accounting for up to 86% of the total load in untreated

wastewater. The results obtained illustrate the importance of considering

retransformable substances, for example human metabolites, when investigating

the behavior and fate of pharmaceuticals. Average concentrations of

sulfapyridine, sulfamethoxazole, trimethoprim, azithromycin, and clarithromycin

in activated sludge ranged between 28 and 68 pg/kg dry weight. Overall the

sorption to activated sludge was shown to be low for the investigated

antimicrobials, with estimated sorption constants below 500 L/kg. Elimination in

activated sludge treatment was found to be incomplete for all investigated

compounds. In final effluents median concentrations for sulfamethoxazole were

290 ng/L and 240 ng/L for clarithromycin.

i

Göbel, A., Thomsen, A., McArdell, CS., Joss, A., Giger, W.

Occurrence ofSulfonamides, Macrolides and Trimethoprim in Activated Sludge

Treatment including Sorption to Sewage Sludge

Environmental Science and Technology, 2005, in press.

Occurrence 83

4.1 Introduction

The occurrence and potential adverse effects of pharmaceuticals in the aquatic

environment have been the object of increasing interest in recent years, indicated

by the growing number of scientific publications published, for example review

articles.[111] For human-use pharmaceuticals the principle entry route to ambient

surface waters, as unchanged compound or after transformation through human

metabolism, is via municipal wastewater treatment plants (WWTP).

Antimicrobial agents, which are used in human and veterinary medicine to

similar extents, are a particularly important pharmaceutical group due to the

possible spread and maintenance of bacterial resistance. The overall human

consumption of antimicrobials in Switzerland exceeded 30 tons per annum (t/a)

in 1997, with ß-lactam antibiotics (18 t/a) as the largest group, including

penicillins, cephalosporins, penems and other smaller sub-classes.[12"4]

Additional important groups in human medicine are the sulfonamides (6 t/a),

macrolides (4 t/a) and the fluoroquinolones (4 t/a). Domestic consumption, that

is through prescription, accounts for 60 to 80% of the total consumption of

human used pharmaceuticals, making urban wastewater the main input source

into the aquatic environment.

Macrolides, which are mainly active against gram-positive bacteria, inhibit

ribosomal protein synthesis.^151 Their main application is in the treatment of

upper and lower respiratory tract infections, especially as an alternative to

penicillins. They are generally not metabolized to a large extent but are mainly

excreted with bile and feces as the unchanged parent substance (Table 1).

Sulfonamides, which are active against a wide range of gram-positive and gram-

negative bacteria, function as competitive antagonists to ^-aminobenzoate in

bacterial folate synthesis.[1617] They are excreted via urine mainly in the form of

metabolites and partly as the unchanged active compound. One class of

important human metabolites is the acetylated forms of sulfonamides, for

example A^-acetylsulfamethoxazole, which accounts for approximately 50% of

the administered dose. The retransformation of A^-acetylsulfamethazine to the

active parent compound sulfamethazine during the storage of manure has been

reported by Berger et al. indicating a possible analogous behavior for other /V4-

acetylated sulfonamides.1183 Some of the excreted metabolites may therefore be

retransformed to the parent compound during wastewater treatment.

Trimethoprim is almost exclusively used in combination with sulfonamides, in a

Table

4.1Characteristicsofsu

lfon

amid

es,trimethoprim

andmacrolides

compound

acronym

CASRN

main

applicationa

humanconsumption

[-

(Switz

erla

nd,1999,

-kg/a]

[13K

%excretedunchanged

inhumanurine

/feces

c

sulfadiazine

SDZ

68-35-9

V+H

68

25%

/-[

17]

sulfathiazole

STZ

72-14-0

V-

50%/-[17]

sulfamethazine

SMZ

57-68-1

V-

10%/

-[17

]

sulf

apyr

idine

SPY

144-83-2

H836

15%/

-tI7

]

sulfamethoxazole

SMX

723-46-6

H2572

10%/

-[17

]

A^-acetyl

sulfamethoxazole

N4AcSMX

humanmetabolite

-

(50%

/-)

d[17]

trimethoprim

TRI

738-70-5

V+H

522

50%

/-[

47]

azithromycin

AZI

83905-01-5

H318

<10%/>65%[15]

erythromyc

inERY-H20

114-07-8

H218

4-20%

/40-50%

[15]

clarithromyc

inCLA

81103-11-9

H1743

20-30%/4-

11%[

15]

roxithromyci

nROX

80214-83-1

H149

8%

/55%

aV=

veterinary

medicine,H=humanmedicine.

bConsumptioncalculatedfromtheusageofsu

lfas

alaz

ine,

assuming100%transformationtosulfapyridineinthecolon.

c

Percentageoftheadministereddose.

dPercentageoftheadministeredsulfamethoxazoledose.

Occurrence 85

fixed ratio of 1:5 with sulfamethoxazole for example, since it also interferes

with bacterial folate synthesis via direct inhibition of dihydrofolate reductase

enzyme. When used together, sulfonamides and trimethoprim produce a

bactericidal effect, as opposed to the bacteriostatic effect yielded by

monotherapy.[19]

Monitoring studies for antimicrobials have been performed, mainly in

wastewaters and surface waters. While ß-lactams, for example penicillins, seem

to be rapidly hydrolyzed after excretion, members of the other three important

antibacterial groups applied in human medicine (fluoroquinolones, macrolides,

sulfonamides) have been detected in the aquatic environment.[6'2023] Generally,

concentrations of up to low pg/L levels occur in WWTP effluents indicating the

importance of WWTPs as point source.[20'23"29] Most concentrations published

are results of the analyses of grab samples, which limits their significance for

assessing occurrence and fate of antimicrobials in wastewater treatment

facilities. Additionally, no data are published yet for the occurrence of

antimicrobials in sewage sludge, with the exception of fluoroquinolones^ 0]

Such data, however, are crucial to enable evaluation and modeling of their

behavior in wastewater treatment.

The aim of the present study was to investigate the occurrence of sulfonamides,

macrolides, and trimethoprim in municipal wastewater treatment plants in

Switzerland (Table 4.1). Wastewater samples from the raw influents to the final

effluents as well as activated and digested sewage sludge samples were

investigated. The results provided the basis for evaluating the behavior within

individual treatment steps of conventional wastewater treatment. Daily profiles

were investigated to assess the short-term variations of antimicrobial loads

entering wastewater treatment. Furthermore, sorption to sewage sludge is

discussed for the various analytes.



Samples were collected from two municipal wastewater treatment facilities in

Switzerland. The wastewater treatment plant Kloten-Opfikon (WWTP-K) is

near the international airport of Zurich, Switzerland. The plant handles 55 000

population equivalents (PE) combining the domestic wastewater of 25 900

86 Chapter 4

inhabitants in the catchment area and the wastewater from the international

airport, including an unknown number of commuters and passengers. During

sampling the average raw inflow was 16 500 ± 5 500 m là. The second

wastewater treatment plant is located at Altenrhein (WWTP-A) in the canton St.

Gall. It receives the combined sewage of 80 000 PE comprising that of 52 000

inhabitants in the catchment area. The average raw inflow amounted to 21 000

m3/d during dry weather. In both WWTPs primary treatment consists of a

screen, an aerated grit-removal tank, and a primary clarifier. In the case of

WWTP-K secondary treatment is performed in two consecutive activated sludge

units and the final effluent is discharged to the receiving surface water after sand

filtration as tertiary treatment. At WWTP-A, secondary treatment is performed

in two parallel operated treatment units: a conventional activated sludge system

and a fixed-bed reactor, receiving -50% of the primary effluent each. Both

systems are designed to provide wastewater treatment by nitrification and

denitrification. The secondary effluents of both units are combined and treated

by a sand filter before being discharged to the receiving river.

In both wastewater treatment plants (WWTP-K and WWTP-A), primary and

secondary sludges are sedimented in the primary clarifier, partially dewatered

and treated in anaerobic, mesophilic digesters with a residence times of 15 - 25

days. Further details on the treatment technologies applied in the two wastewater

treatment plants are given in Chapter 5.


Sampling campaigns were performed on days without substantial rainfall in

March 2002, February 2003 and November 2003 at WWTP-K and in September

2002 and March 2003 at WWTP-A. Flow-proportional composite samples were

collected by means of automated samplers of raw influents after the screen, of

primary effluents after primary clarification, of secondary effluents after

conventional activated sludge treatment and of tertiary effluents after sand

filtration. Weekly campaigns comprised 3 composite samples, one combining

the 24 h integrated samples from Saturday and Sunday and two combining two

or three weekdays. For daily profiles at WWTP-K, flow proportional 8 h

composite samples of three consecutive intervals were collected in February

2003. All samples were transferred into amber glass bottles and filtered (0.45-

pm cellulose nitrate filters, Schleicher & Schuell) as soon as possible (no later

Occurrence 87

than 6 h after collection). For detailed information on sample treatment see

Chapter 2.

Grab samples of sewage sludge were taken at the end of the aeration tanks

(activated sludge) and at the outlets of the anaerobic, mesophilic digesters fed by

both primary and secondary sludges (digested sludge). In the case of the

activated sludge, grab samples were taken once per week (November 2003),

three and four times per week (September 2002 and March 2003, respectively)

or daily (March 2002 and February 2003), filtered (glass fiber filters, GF8,

Whatman) and freeze-dried. Subsequently, grab samples were mixed in equal

amounts to obtain weekly samples, which were further analyzed (refer to

Chapter 3). Digested sludge samples were taken in February, March, and

November between 20 and 25 days after the respective sampling campaigns

(taking the residence time in the digester into consideration) and directly freeze-

dried without flltration. Consequently, the results obtained for activated sludges

are given in pg/kg dry weight, while those for digested sludges, including the

aqueous phase, are given in pg/L. The concentration of solids in the freeze-dried

digested sludge was determined to range between 15 and 18 g/L.

4.2.3 Analytical Methods

Analytical details, including information on all materials and reagents used, are

described for wastewater samples in Chapter 2 and for sludges in Chapter 3.

Briefly, wastewater samples were enriched by solid-phase extraction (Oasis

HLB, Waters, Bergen op Zoom, The Netherlands), followed by reversed-phase

liquid chromatography - tandem mass spectrometry using positive electrospray

ionization. Recoveries from all sample matrices were generally above 80%, and

the combined measurement uncertainty varied between 2 and 18%. Sample-

based quantification limits depended on the analyte and the sample matrix and

ranged between 1 and 220 ng/L.

Freeze-dried sludge samples were extracted using pressurized liquid extraction

with methanol:water (1:1) as extraction solvent. Sludge extracts were diluted

with water to reduce the methanol content below 5%, and subsequently enriched

on Oasis HLB cartridges and analyzed as described for aqueous samples (see

above). Recoveries from activated sludge were above 80%> in all cases and the

overall precision of the method ranged between 3% and 8%. The limits of

quantification varied between 3 and 30 pg/kg.

88 Chapter 4

4.2.4 Estimation ofSorption Constants

Sorption constants were estimated for antimicrobials present in activated sludge

to characterize their respective distribution between sludge and wastewater

phases. Therefore, the amount present as sorbed fraction, that is associated with

the sludge, is related to the dissolved amount by two different approaches. In the

first approach, concentrations measured in the composite samples of the

secondary effluents from the conventional activated sludge systems at WWTP-K

and WWTP-A are used. These aqueous concentrations (n = 3) were related to

the amount present in the activated sludge sample (n =1) of the identical

sampling week (field experiments). In the second approach, grab samples from

the end of the nitrification compartment were taken in October 2003 and

February 2004 of WWTP-K. In the latter case, the amounts present in the filter

cake (sorbed fraction) and in the filtrate (dissolved fraction) were determined by

duplicate analyses. The pH of the filtrate ranged between 7.5 in October 2003

and 7.0 in February 2004. Since the filter cake is not completely dry after

filtration, it also contains some analytes in the water phase. Calculations showed

that the amount of analytes dried onto the sludge during freeze-drying ranged

between 2 - 5%o of the amount sorbed to the sludge. It was therefore neglected in

all cases.

4.2.5 Calculation ofLoads

Loads were calculated to take into account the influence of flow variations on

the measured antimicrobial concentrations (e.g. due to rain). The loads of the

three sampling intervals were summed to yield a weekly load, which was then

used to calculate an average daily load (g/d). For comparison reasons the results

were normalized to 1000 PE in Table 4.3. The dissolved loads were calculated

by multiplying the measured concentration at a specific sampling point and the

respective water flow during the sampled time period. The amount of

antimicrobials sorbed to suspended solids in the raw influent (130-200 mg/L)

and the primary effluent (80-100 mg/L) was estimated using the highest Kd

value observed for activated sludge for the respective compound. The sorbed

portion of antimicrobials in secondary and tertiary effluents was neglected,

based on the small concentrations of suspended solids (5-20 mg/L).

Occurrence 89

The uncertainty of the average daily loads was calculated using error

propagation. Therefore the uncertainty of the analytical method (see Chapter 2)

and the uncertainty of the water flow measurement (5-10%) was used. For

samples below the limit of quantification, a relative uncertainty of 100% was

applied to the concentration measured. A relative uncertainty of 50% was

assigned to the Kd value used for suspended solids in the raw influent and the

primary effluent, to account for the uncertainty connected with the extrapolation

from activated sludge. The resulting uncertainties for the average daily loads

therefore only represent measurement uncertainties and do not show expected

variations between days or sampling campaigns.

Additionally, theoretical loads of antimicrobials were calculated in the raw

influent using available sales data for Switzerland, the maximal amount of the

compound excreted unchanged (Table 4.1) and the number of inhabitants

(n= 25 900) in the catchment area of the investigated treatment facility.


4.3.1 Occurrence in Wastewater Samples

In part of this study the occurrence of sulfonamides (including one human

metabolite), trimethoprim, and macrolides in various compartments of two

municipal wastewater treatment plants was investigated. The median, minimum

and maximum concentrations found in wastewater samples are summarized in

Table 4.2.

While sulfathiazole and sulfamethazine were not detected in the wastewater

samples, all human used antimicrobials investigated were detected, except for

sulfadiazine, which is only prescribed in very low quantities (Table 4.1).

Sulfamethoxazole was the most commonly detected sulfonamide in our samples.

The fraction present as human metabolite, A^-acetylsulfamethoxazole was taken

into account to better assess the occurrence of sulfamethoxazole in wastewater.

Median concentrations of sulfamethoxazole including the measured amount

present as A^-acetylsulfamethoxazole were up to 1 900 ng/L in the raw influent

and 880 ng/L in the tertiary effluent. The frequent abundance of

sulfamethoxazole in our study, as well as described in the literature, is due to the

high consumption of this compound in human medicine. Maximal influent

concentrations of 520 ng/L in the United States [23], 232 ng/L in Austria [25],

Table

4.2

Concentrations

ofsu

lfon

amid

es,

trimethoprim

and

macrolides

in

wastewater

oftwo

municipal

wastewater

treatmentpl

ants

inSwitzerland

concentrationsmeasured(n

g/L)

raw

infl

uent

bpr

imar

yeffluentc

seco

ndar

yeffluentc

tertiary

effluentc

compounda

median

min

max

median

min

max

median

min

max

median

min

max

SPY

90

60

150

130

90

230

70

20

230

90

40

350

SMX

430

230

570

430

90

640

280

130

840

290

211

860

N4AcSMX

1400

850

1600

890

570

1200

40

<LOQ

150

10

<LOQ

180

SMX+N4AcSMX

1700

940

1900

1200

720

1600

380

190

880

400

210

880

TMP

290

210

440

230

80

340

200

80

400

70

20

310

AZI

170

90

380

150

80

320

140

40

380

160

80

400

ERY-H20

70

60

190

80

40

190

80

50

140

70

60

110

CLA

380

330

600

330

160

440

260

150

460

240

110

350

ROX

20

10

40

20

10

50

20

10

30

10

10

30

a

Sulfadiazine,sulfathiazole,

andsulfamethazinewerenotorveryra

rely

detected(d

atanotshown).

bMedian,minimum

andmaximum

valueofninemeasurements,

sinceraw

influentwas

notsampled

inMarch2002

andSeptember

2002.

c

Median,minimumandmaximum

valueof15measurements.

In

the

caseofAZI

(n=

12)no

resultsare

available

forMarch2002

becauseofan

alytical

interferences.ForN4AcSMX

8outof15

tertiary

effluentsamplesand6outof15secondaryeffluentsa

mple

swere

belowthelimitofqu

anti

fica

tion

(—20ng

/L).

Thesewereincludedas0.5xLOQ

inthecalculationofmedianvalues.

Occurrence 91

580 ng/L in Spain L31J, and 9 000 ng/L in Germany[ZI'^

are reported.

Concentrations for sulfamethoxazole available in literature for treated sewage

range from 50 to 4 700 ng/L.[20,23'25'27'29,311 Since none of the previous studies

included the main human metabolite A^-acetylsulfamethoxazole, these earlier

results must be cautiously interpreted. To the best of our knowledge, the

presence of this possibly re-transformable metabolite has been reported only

once. Hilton and Thomas detected up to 2 235 ng/L of A^-acetyl-sulfamethoxazole in a treated wastewater grab sample and up to 239 ng/L in

receiving surface waters.[32] Contrary to our data, the parent compound

sulfamethoxazole was not detected.

We also detected the synergist trimethoprim in all samples at median

concentration of 290 ng/L and 70 ng/L in raw influents and tertiary effluents,

respectively. Trimethoprim concentrations previously reported range from 9

ng/L to 1 500 ng/L for WWTP effluents. [20>24>25'28>32'33] ln human medicine the

prevailing form administered combines sulfamethoxazole and trimethoprim at a

fixed ratio of 5:1, which is reflected in the consumption data. This ratio was still

visible in the raw influent samples we analyzed (5.4:1). However, in the final

effluent (8.6:1) this ratio is altered presumably by the different behavior of

sulfamethoxazole and trimethoprim in wastewater treatment as discussed in

Chapter 5.

Sulfapyridine was also detected in all wastewater samples with concentrations

up to 150 ng/L in the raw influents and up to 350 ng/L in the tertiary effluents.

The higher concentration in the effluents is probably due to the presence of re-

conjugable substances as described below. A similar median concentration of

sulfapyridine (81 ng/L) was detected in grab samples from eight Canadian

wastewater effluents.[29] Sulfapyridine has otherwise not been investigated in

environmental samples, probably due to the fact that it is rarely used as an

antimicrobial agent itself. Its occurrence in wastewater samples results from the

application of sulfasalazine, which is mainly used for the treatment of ulcerative

colitis and rheumatoid arthritis.[34] In sulfasalazine, sulfapyridine is linked to 5-

aminosalicylic acid via an azo bridge, which is cleaved in the colon - thereby

yielding the two components. Of the administered dose only -10% are excreted

via the urine as sulfasalazine itself, while another 10 to 35% are excreted as

sulfapyridine and between 20 and 40% as V-acetylsulfapyridineJ351All four investigated macrolides were detected in wastewater with

clarithromycin being most abundant. Median concentrations measured for

Table

4.3

Daily

loads

ofsu

lfon

amid

es,

trimethoprim

and

macrolides

inprimary

effluentsoftwo

municipalwastewater

treatmentplants

inSwitzerland

averageda

ilyloada(mg/d/1000PE)

WWTP-K

WWTP-A

compoundb

March

February

November

September

March

2002

2003

2003

2002

2003

SPY

16±7C

38±

15

53±4

33±2

34±5

SMX

42±2

158±4

178±5

39±1

188±8

N4AcSMX

307±7

302±7

264±7

255±13

381±19

SMX+N4AcSMXd

305±7

416±9

404±7

258±13

516±20

TRI

89±7

82±7

84±7

33±4

90±9

AZI

e45±2

47±2

59±4

101±8

ERY-H20

49±2

44±2

24±2

15±1

26±1

CLA

149±7

160±7

96±4

59±4

125±8

ROX

16±2

13±2

5±2

9±1

5±1

a

Average

dail

yloadscalculatedfromweeklyloadsasdescribedintheEx

peri

ment

alSection.Thevalueswerenormalizedtopo

pula

tion

equivalents(PE),whichare55000PE

forWWTP-Kand80000PE

forWWTP-A.

bSulfadiazine,sulfathiazole,

andsulfamethazinewerenotorveryra

rely

detected(d

atanotshown).

cDeviationcalculatedfrommeasurementuncertaintiesasdescribedintheEx

peri

ment

alSection.

dSulfamethoxazoleload,in

clud

ingtheamountpresentasA^-acetylsulfamethoxazole.

eNo

resultsavailablebecauseofanalytic

alinterferences.

Occurrence 93

Clarithromycin were 380 ng/L in raw influents and 240 ng/L in final effluents.

For roxithromycin and dehydro-erythromycin median concentrations ranged

between 20 and 70 ng/L in the raw influents and between 10 and 70 ng/L in the

final effluents, respectively. These concentrations correspond well with previous

findings in Switzerland by McArdell et al., who reported between 57 ng/L and

328 ng/L for clarithromycin, up to 31 ng/L of roxithromycin, and up to 199 ng/L

of dehydro-erythromycin in final effluents of three conventional treatment

plants/261 Similar concentrations have also been observed in wastewater

effluents in other countries.[25'28,29'32] Median concentrations of dehydro-

erythromycin - the most predominant macrolide in the investigated wastewater

effluents - were found to be 137 ng/L in Germany, 109 ng/L in the UK, 80 ng/L

in Canada, and 394 ng/L in Austria.

Median concentrations of azithromycin were 170 ng/L in raw influents and 150

ng/L in tertiary effluents of conventional wastewater treatment. Thus,

azithromycin is the second most abundant macrolide after clarithromycin in

Swiss wastewaters.

Table 4.3 shows the average daily loads from the individual sampling campaigns

determined in the primary effluent of both municipal wastewater treatment

plants, WWTP-K and WWTP-A. The individual loads of the investigated

antimicrobials are similar in both wastewater treatment plants due to their

similarity in size, and range between 0.3 and 41 g/d. As a consequence of human

metabolism, the amount of sulfamethoxazole itself in the primary effluent is low

and varied between 14 and 44% of the total amount. Therefore, it is crucial to

include the fraction present as the human metabolite A^-acetylsulfamethoxazole.

Figure 4.1 illustrates the loads measured in the raw influents and tertiary

effluents of WWTP-K for all antimicrobials investigated. Additionally, it shows

that loads in the raw influent estimated from the available consumption data

available follow the same pattern as the measured data - both mirroring the

annual amounts (1999) used in human medicine in Switzerland (Table 4.1). A

very good agreement was obtained between the measured and calculated loads,

within the limits of uncertainty, with disagreements not greater than a factor of

two. The differences can be inferred by i) local variations in consumption as

compared to Swiss average, Ü) unknown loads of antimicrobials from daily

commuters in the catchment area, iii) unaccounted input loads from air traffic

passengers at the international airport of Zurich and iv) the uncertainty of the

available metabolism data (e.g. percentages excreted unchanged).

94 Chapter 4

Figure 4.1 Daily loads ofsulfonamides, trimethoprim and macrolides in raw

influents andfinal effluents ofthe municipal wastewater treatmentplant, Kloten-

Opfikon, Switzerland

30

26

20 -

115«8

O

10

Ml

X

I

X

I

jlT.

X

T

maximum

75% percentile

median

25% percentileminimum

X estimated bad

If raw influent (n=6)

final efluent (n=9)

intfÏMM 1—^~1

fJ__L _L^^

SPY SMX N4ACSMX SMX-t- TMP

N4ACSMX

AZI ERY-H20 CLA ROX

a

Theoretical antimicrobial loads calculated from consumption data as described in the

text.

4.3.2 Daily Variations

The loads determined in the raw influent of three consecutive 8 h time intervals

vary to different extents for the investigated antimicrobials (Table 4.4). These

variations were smoothed out during wastewater treatment, resulting in

approximately equal amounts in the tertiary effluents at all three time intervals

(data not shown). To illustrate the impact of human urine on the wastewater

composition, an average daily profile of ammonium in the WWTP inflow is also

included in Table 4.4. The values given for the water flow in the respective

sampling period represent a typical flow profile for a dry weather influent.

In the case of the sulfonamides and trimethoprim the distribution of the daily

load correlates well with the respective water flows and typical ammonium

Occurrence 95

Table 4.4 Daily variations of antimicrobial loads, water flow, and ammonium

load in untreated wastewater (WWTP Kloten-Opfikon, Switzerland)

daily load fraction of daily load (%)

parameter /

compound(g/d) 0-8 am 8-4 pm 4-12

SPY 1.7 26 39 35

SMX 3.5 28 42 30

N4AcSMX 20 23 44 33

SMX+N4AcSMX 21 24 43 33

TRI 4.7 34 36 30

AZI 3.9 9 61 30

ERY-H20 1.5 15 49 36

CLA 8.1 19 46 35

ROX 0.4 9 27 64

water flow 21939m7d 24 40 36

ammonium ~320kg/d 18 46 36

loads. With a half-life of -10 h in the human body and a typically prescribed

oral administration of twice a day, these findings correspond well with the

theoretically expected distribution pattern.[36]The macrolides showed a more variable daily profile with the lowest loads

occurring between midnight and 8 am (9 - 19% of the daily load). This is most

likely caused by the respective consumption patterns and excretion rates of these

compounds. In the case of azithromycin the observed pattern can be explained

by the usually prescribed oral consumption of once a day - presumably in the

morning, and a half-life in the human body of 10-14 h.[15] Good agreement is

also observed for clarithromycin, which is normally taken twice a day, assumed

to be in the morning and the evening, and which has a half-life in the human

body of only ~5 h. Erythromycin has a half-life in the human body of 1 -2 h and

is administered between 2-4 times a day, which also correlates well with the

observed daily variations. In addition to oral application, ~20%> of the annual

consumption of erythromycin is used in the form of facial lotions against acne.

Therefore, additional amounts of erythromycin are probably washed of directly

during the day. From the usual prescription pattern (twice a day) and the

96 Chapter 4

reported half-life (-10 h) for roxithromycin a correlation of the respective loads

with the water flow and typical ammonium loads would be expected. The poor

agreement observed in this case, may be due to the low consumption of

roxithromycin in Switzerland resulting in very few patients in the catchment

area.

One has to keep in mind that the daily variations observed only result from one

sampling day divided into three sampling periods. Further investigations using

smaller time intervals and covering multiple days would be necessary to confirm

these findings. However, the results clearly illustrate a possible daily variance

for antimicrobials in wastewater samples. It is therefore crucial to use flow

proportional composite samples over at least 24 h, as done in this study, when

assessing their fate and occurrence in wastewater treatment.

4.3.3 Seasonal Differences

Concerning possible seasonal difference in consumption no conclusive results

were obtained for sulfamethoxazole, including the amount present as i\T-

acetylsulfamethoxazole, and trimethoprim (Table 4.3). In the latter two cases,

higher loads were observed in the primary effluent of WWTP-A in March 2003

compared to September 2002. In WWTP-K, however, similar loads were

measured in all three sampling campaigns for sulfamethoxazole, including the

amount present as A^-acetylsulfamethoxazole, and trimethoprim. Also, no

conclusion can be drawn for sulfapyridine, where the presence of chemically

bound sulfapyridine, for example A^-acetylsulfapyridine, is very likely, but was

not assessed in this study.

For the macrolides the determined loads were generally higher in March and

February than in November and September by approximately a factor of two.

Relative monthly sales data available for the group of macrolides in total, show

a clear periodicity with two times lower consumption in summer compared to in

winter.[37] Highest amounts of macrolides are sold between February and April

each year. Even though no sampling campaign was performed directly in

summer, the different loads measured for macrolides in this study correlate well

with the consumption data. This variability is probably caused by the use of

macrolides primarily against respiratory tract infections. McArdell et al. already

related higher loads of macrolides in winter in the effluent of the same WWTP

to the distinct seasonal consumption.[26] Loads determined in influent samples,

Occurrence 97

as presented in this study, however are needed to check this correlation. A

similar annual fluctuation was reported by Strenn et al. for the concentration of

roxithromycin in the inflow of a wastewater treatment plant, however, the

influence ofpossibly varying water inflow was not taken into account.[38]

4.3.4 Occurrence in Sewage Sludge

The antimicrobials investigated in wastewater samples were also measured in

sewage sludge (Table 4.5) Most of the antimicrobials present in aqueous phase

were also found in the solid samples, with a few exceptions. Even though

A^-acetylsulfamethoxazole is present in high concentrations in the raw influent

of the wastewater treatment plant it could not be detected in any of the sewage

sludge samples. Spiking A^-acetylsulfamethoxazole on activated sludge (-100

pg/kg) results in an increase of the sulfamethoxazole concentration while no

A^-acetylsulfamethoxazole could be detected in the sludge (data not shown).

Since no degradation of A^-acetylsulfamethoxazole occurred, when extracted

under the same extraction conditions from silica sand only, a fast transformation

of this compound in the presence of sewage sludge must be assumed.

Sulfamethoxazole and trimethoprim, each of which were present in aqueous

samples, were also detected in activated sludge with average concentrations of

68 ± 20 pg/kg dry weight and 41 ± 15 pg/kg dry weight, respectively. In

digested sludge, however, no significant amounts of these two compounds could

be detected. Therefore, sulfamethoxazole and trimethoprim seem to be unstable

in anaerobic, mesophilic sludge digestion. Sulfapyridine, however, was detected

in activated (28 ± 3 pg/kg dw) and in digested (1.0 ± 0.1 pg/L) sludge. Two

possible explanations could be a significantly higher stability of sulfapyridine

compared to sulfamethoxazole during sludge digestion or the possible presence

of substances re-transformable to sulfapyridine (e.g. sulfasalazine) in sewage

sludge.

Among the macrolides only azithromycin and clarithromycin were detected in

sewage sludge samples. In activated sludge average concentrations of 64 ± 30

pg/kg dry weight azithromycin and 67 ± 28 pg/kg dry weight clarithromycin

were determined. Concentrations in digested sludge were 2.5 ± 1.0 pg/L for

azithromycin and 0.7 ± 0.4 pg/L for clarithromycin. The low concentrations of

roxithromycin present in the water phase most probably led to concentrations in

sewage sludge below the limit of quantification (LOQ = 3 pg/kg dry weight).

98 Chapter 4

Table 4.5 Concentrations of sulfonamides, trimethoprim and macrolides in

sewage sludges

concentration

compoundactivated sludge

(pg/kg dw)a

digested sludge

(ltg/L)b

average ± SD (n = 5) average ± SD (n = 3)

SPY 28 ±3 1.0±0.1

SMX 68 ±20 ndc

TRI 41 ±15 nq(0.1)c

AZI 64 ±30 2.5 ±1.0

CLA 67 ±28 0.7 ± 0.4

a

Separation of solid (~3g/L) and aqueous phase through flltration before freeze-

drying.bNo separation of solid (15-18 g/L) and aqueous phase through filtration before

freeze-drying.c

nd: not detected, nq: below limit of quantification (estimated value)

In the case of dehydro-erythromycin, high dissolved concentrations were

measured while no significant amounts were detected in sewage sludge samples

(LOQ = 6 pg/kg dry weight). Therefore no significant sorption of dehydro-

erythromycin to sewage sludge seems to take place.

4.3.5 Sorption to Sewage Sludge

The measured concentrations ofantimicrobials in activated sludge, during the

performed sampling campaigns and in grab samples (refer to 4.2.4), were also

used to estimate sorption coefficients (K^) for the particular analytes. The values

obtained, ranged between 114 and 460 L/kg for all compounds investigated

(Table 4.6). For the macrolides the results obtained in the field experiments or

grab samples are generally very similar. For sulfonamides and trimethoprim the

resulting K<j values range between 114 and 418 L/kg in activated sludge. From

literature, data on the sorption of antimicrobials to biosolids is available for

Occurrence 99

Table 4.6 Estimated sorption constantsfor sulfonamides, trimethoprim and

macrolides

Kdfield experiments

'

in activated sludge (L/kg)a

'

grab samples

compound average ± SD (n= 9 - 15) October 2003c February 2004c

SPY 295 ± 145 202 418

SMX 256 ±169 114 400

TRI 208 ± 49 157 375

AZI 376 ± 86 460 352

CLA 262 ± 93 300 400

a

Calculated as the ratio of the sorbed and dissolved fraction measured.

Calculated using concentrations from weekly sampling campaigns (dissolved

fraction, n=3; sorbed fraction, n=l).cCalculated using concentrations measured in grab samples from the end of the

nitrification compartment.

tylosin, a macrolide used in veterinary medicine.[39] Organic carbon normalized

sorption coefficients (Koc) were reported for tylosin in soil and manure ranging

between 110 and 7 990 L/kg. The same review gives K0c values in soil between

48 and 323 L/kg for sulfonamides in general and between 101 and 308 for

sulfapyridine in particular. Since activated sludge consists of about 40% organic

matter, reported KoC values might be expected to be larger than IQ values for

activated sludge by at least a factor of two. In general, however, the significance

of such a comparison is limited due to the structural and compositional

differences in the matrices investigated relative to soil samples used in the

previous study.

For the macrolides, the Kd values obtained in this study vary to a lesser extent

than for the sulfonamides. The results for sulfonamides vary up to 70% in the

field experiments and are between two to four times higher in the grab samples

from February 2004 than those from October 2003. This indicates a possible

non-equilibrium state in the samples. The change in solid to solution ratio may

be caused for example by a different degradation of the analytes in the two

phase or by the wastewater treatment processes combined with slow sorption

kinetics. Different sludge characteristics in the samples taken can also influence

the sorption of the investigated antimicrobials. For soil, there is a strong

100 Chapter 4

dependence of the adsorption of sulfonamides on the pH and on the quantity and

composition of the organic matter, namely the concentration of lipids and lignin

dimers, was described.[39] In the case of macrolides, hydrophobic interactions

were described to be mainly responsible for the sorption to soil. Due to the

predominantly negatively charged surface of activated sludge,[40] ionic

interactions are expected to be significant for the dissolved macrolides, being

positively charged through the protonation of the tertiary amino group (pKa >

8.7 [41]). The sorption of macrolides to activated sludge should therefore be

strong and higher than that of the negatively charged or neutral sulfonamides

(pKa = 5.7-8.4 [17]) However, this was not supported by our data (Table 4.6) and

ionic interactions therefore seem to be of minor importance for the sorption of

macrolides to activated sludge. Further experiments would be necessary to

investigate the parameters affecting the observed sorption of antimicrobials to

activated sludge. Additionally it has to be mentioned that sewage is a very

inhomogeneous medium that varies globally on both a spatial and temporal

scale. Therefore the results obtained primarily apply to the investigated systems.

4.3.6 Mass Balances

The loads determined now provide the basis to investigate the behavior of

Figure 4.2 shows the loads obtained in the second sampling campaign (February

2003) for three exemplary compounds. Average daily loads were calculated at

the different sampling points, with differentiation between the dissolved and

sorbed fraction.

A total load (sorbed and dissolved) of 24.4 g/d sulfamethoxazole was

determined in the raw influent including the fraction present as the human

metabolite /V-acetylsulfamethoxazole, which accounted for -75% of the load.

After primary treatment only -62% of the total load could be assigned to

/V-acetylsulfamethoxazole and after secondary treatment this metabolite was

almost completely absent. Thus, there is a strong indication of re-transformation

of A^-acetylsulfamethoxazole to sulfamethoxazole. Further experiments,

however, are needed to fully confirm this assumption. An elimination of

sulfamethoxazole, including the amount present as A^-acetylsulfamethoxazole,occurred mainly in secondary treatment (-55%), while no significant

elimination was observed in primary and tertiary treatment in February 2003. In

the case of clarithromycin and trimethoprim no significant elimination was

Occurrence 101

Figure 4.2 Mass balances for clarithromycin, sulfamethoxazole including the

fraction present as N4-acetylsulfamethoxazole, and trimethoprim in the

conventional wastewater treatment plant Kloten-Opfikon, Switzerland

(February 2003)

A. Sulfamethoxazole including the amount present as 7V4-acetylsulfamethoxazole

daily load

24,5 g/d

B. Trimethoprim

daily load

5.2 g/d

C. Clarithromycin

daily load

9.9 g/d

l%d

]%s

< 0.2%

l%d

3% s

<1%

d: dissolved, s: sorbed/particulate

102 Chapter 4

observed in primary and secondary treatment, but they were partly eliminated

during sand filtration. For clarithromycin a reduction by 15% was found on the

sand filter ofWWTP-K and by 60% for trimethoprim.

4.4 Conclusions

Overall, an incomplete removal of the selected antimicrobials is observed in

secondary wastewater treatment. The maximum concentrations determined in

the final effluents of two municipal wastewater treatment plants amounted to

over 800 ng/L for sulfamethoxazole (Table 4.2). Concentrations of

antimicrobials in the final effluent of wastewater treatment plants have to be

critically assessed concerning their impact on the aquatic environment. Most of

the microbial toxicity data available originates from acute toxicity studies and

falls in the mg/L range, see e.g. references [9,42-45]. In some cases sublethal

effects in algal species have been reported in the high pg/L range.[46] Therefore

antimicrobials are unlikely to cause acute adverse effects in aquatic

microorganisms even if the dilution factor is low, for example in summer.

However, effects due to the chronic low dose exposure or non-target effects and

mixture effects cannot be ruled out. Another aspect concerns the spread and

maintenance of antibacterial resistance. These latter issues require further

investigation.

A detailed investigation of the elimination observed for sulfonamides,

macrolides and trimethoprim in the individual treatment stages of conventional

activated sludge treatment is the focus of our further studies (Chapter 5).

Additionally, conventional activated sludge treatment will be compared to other

secondary treatment technologies, i.e. a fixed-bed reactor and a membrane

bioreactor operated at three different solid retentions times, with respect to the

elimination of the selected antimicrobials.

Acknowledgments




2000-00047, www.eu-poseidon.com) and the EAWAG project on human-use

antibiotics (HUMABRA) within the framework of the National Research

Occurrence 103

Program on antibiotic resistance funded by the Swiss National Science

Foundation (www.nrp49.ch/pages/). We would also like to thank the Swiss

Agency for the Environment, Forestry and Landscape, the Swiss cantons of

Aargau, Basel Land, Bern, Luzern, Schaffhausen, Schwyz, St. Gallen, Thurgau,

Ticino, Zurich and the WWTPs of Kloten-Opfikon and Altenrhein for additional

financial support. We thank the technical staff of the WWTP Kloten-Opfikon

and of the WWTP Altenrhein for their assistance during sampling. For helpful

comments on the manuscript we acknowledge A. Alder, H. Siegrist, M. Suter, T.

Ternes and M. Dodd.


[1 ] Stan, H. J.; Heberer, T. Analusis 1997, 25, M20-M23.


[3] Halling-S0rensen, B.; Nors Nielsen, S.; Lanzky, P. F.; Ingerslev, F.; Holten

Lützenheft, H. C; J0rgensen, S. E. Chemosphere 1998, 36, 357-393.


938.





[8] Tolls, J. Environ. Sei. Technol. 2001, 35, 3397-3406.

[9] Boxall, A. B. A.; Fogg, L. A.; Blackwell, P. A.; Kay, P.; Pemberton, E. J.;

Croxford, A. Reviews in Environmental Contamination and Toxicology

2004,180,1-91.

[10] Jones, O. A. H.; Voulvoulis, N.; Lester, J. N. Environ. Technol. 2001, 22,

1383-1394.

[11] Ayscough, N. J.; Fawell, J.; Franklin, G.; Young, W. "Review of human

pharmaceuticals in the environment," Environment Agency, 2000.


1998.




104 Chapter 4



[16] Vree, T. B.; Hekster, Y. A. Pharmacokinetics ofsulfonamides revisited;


[17] Vree, T. B.; Hekster, Y. A. Clinicalpharmacokinetics ofsulfonamides and


[18] Berger, K.; Petersen, B.; Büning-Pfaue, H. Arch. Lebensmittelhyg. 1986,

37,85-108.

[19] Poe, M. Science 1976,194, 533-535.


1999,225,109-118.


2001,955,199-210.

[22] Aider, A. C; McArdell, C. S.; Golet, E. M.; Ibric, S.; Molnar, E.; Nipales,

N. S.; Giger, W. In Pharmaceuticals andPersonal Care Products in the




[23] Yang, S.; Carlson, K. Water Res. 2003, 37, 4645-4656.

[24] Andreozzi, R.; Marotta, R.; Paxeus, N. Chemosphere 2003, 50, 1319-1330.

[25] Scharf, S.; Gans, O.; Sattelberger, R. "Arzneimittelwirkstoffe im Zu- und

Ablauf von Kläranlagen," Report of the Umweltbundesamt, Wien, ISBN 3-

85457-624-2, 2002.


Technol. 2003, 37, 5479-5486.


[28] Bund/Länderausschuss für Chemikaliensicherheit (BLAC), "Arzneimittel in

der Umwelt - Auswertung der Untersuchungsergebnisse," Hamburg, 2003.

[29] Miao, X.-S.; Bishay, F.; Chen, M.; Metcalfe, C. D. Environ. Sei. Technol.

2004,55,3533-3541.


5455-5462.

[31] Carballa, M.; Omil, F.; Lema, J. M.; Llompart, M.; Garcia-Jares, C;

Rodriguez, I.; Gomez, M.; Ternes, T. A. Water Res. 2004, 38, 2918-2926.

[32] Hilton, M. J.; Thomas, K. V. J. Chromatogr., A 2003, 1015, 129-141.

Occurrence 105

[33] Metcalfe, C. D.; Koenig, B. G.; Bennie, D. T.; Servos, M.; Ternes, T. A.;

Hirsch, R. Environ. Toxicol. Chem. 2003, 22, 2872-2880.

[34] Astbury, C; Dixon, J. S. J. Chromatogr. 1987, 414, 223-227.

[35] Neumann, J. "Untersuchungen zur Bioverfügbarkeit und Pharmakokinetik

von Sulfasalazin und seinen Metaboliten," PhD Thesis, Free University of

Berlin, 1989.

[36] Neuman, M. Antibiotika-Kompendium; Verlag Hans Huber: Bern, 1981.

[37] Truempi, B., Abbot AG, Baar, Switzerland, personal communication.

[38] Strenn, B.; Clara, M.; Gans, O.; Kreuzinger, N. In Water Pollution VII;

Brebbia, C. A., Ed.; WIT Press: Southampton, UK, 2003; Vol. ISBN 1-

85312-976-3, pp 273-282.

[39] Thiele-Bruhn, S. J. Plant Nutr. Soil Sei. 2003,166, 145-167.

[40] Carberry, J. B.; Englande, A. J. Sludge Characteristics andBehavior;

Martinus Nijhoff Publishers: Boston, The Hague, Dordrecht, Lancaster,

1983.

[41] McFarland, J. W.; Berger, C. M.; Froshauer, S. A.; Hayashi, S. F.; Hecker,

S. J.; Jaynes, B. H.; Jefson, M. R.; Kamicker, B. J.; Lipinski, C. A.; Lundy,

K. M.; Reese, C. P.; Vu, C. B. J. Med. Chem. 1997, 40, 1340-1346.

[42] Holten Lützh0ft, H. C; Halling-S0rensen, B.; Jorgensen, S. E. Arch.

Environ. Contam. Toxicol. 1999, 36, 1-6.

[43] Halling-S0rensen, B. Arch. Environ. Contam. Toxicol. 2001, 40,451-460.

[44] Jones, O. A. H.; Voulvoulis, N.; Lester, J. N. Water Res. 2002, 36, 5012-

5022.

[45] Pfluger, P.; Dietrich, D. R. In Pharmaceuticals in the environment;

Kümmerer, K., Ed.; Springer, 2001.

[46] Brain, R. A.; Johnson, D. J.; Richards, S. M.; Sanderson, H.; Sibley, P. K.;

Solomon, K. Environ. Toxicol. Chem. 2004, 23, 371-382.

[47] Schwartz, D. E.; Rieder, J. Chemotherapy 1970, 75, 337-355.

a **i "'*» n; "3

MC lit '*. (î

S 'V»** I ^4 S* tr V« fl

Chapter 5

Behavior in Wastewater Treatment

The elimination of sulfonamides, macrolides and trimethoprim from raw

wastewater was investigated in several municipal wastewater treatment plants.

Primary treatment provided no significant elimination for the investigated

substances. Similar eliminations were observed in the secondary treatment of

two conventional activated sludge (CAS) systems and a fixed-bed reactor

(FBR). Sulfamethoxazole, including the fraction present as N4-

acetylsulfamethoxazole, was eliminated by approximately 60% in comparison to

about 80% in a membrane bioreactor (MBR) independently of the solid

retention time (SRT). The elimination for macrolides and trimethoprim varied

significantly between the different sampling campaigns in the two CAS systems

and in the FBR. In the MBR, these analytes were eliminated up to 50% at SRT

of 16 ± 2 and 33 ± 3 days. Trimethoprim, clarithromycin and dehydro-

erythromycin showed a higher elimination of up to 90%> at a SRT of 60 - 80

days. One of the sand filters investigated led to a significant elimination of most

macrolides (17 - 23%) and trimethoprim (74 ± 14%), while no elimination was

observed in the other sand filter investigated.

Göbel, A., McArdell, CS., Joss, A., Siegrist, H., Giger, W.

Fate ofSulfonamides, Macrolides and Trimethoprim in Different Wastewater

Treatment Technologies

submitted to Environmental Science and Technology

Behavior 109

5.1 Introduction

Pharmaceuticals have been detected in various compartments of the aquatic

environment as a result of the substantial progress achieved by chemical

analysis. This has lead to an increasing interest in the assessment of fate,

environmental risk and potential regulations of these emerging contaminants,

mirrored by the large number of publications and reviews available, e.g. [1-9].

Within the large group of pharmaceuticals, antimicrobials are of special interest

because of their potential impact on the spread and maintenance of antimicrobial

resistance. Following consumption, the unchanged parent compound and

possible human metabolites are discharged to the sewers and, therefore, residual

concentrations mainly enter the aquatic environment after incomplete

elimination during municipal wastewater treatment.

In the last 40 years, wastewater treatment has continuously been amended to

fulfill the increasing requirements on the quality of the final effluents.[10]

However, the efficiency of distinct wastewater treatment processes, e.g.

conventional activated sludge treatment, fixed-bed reactors or membrane

bioreactors, for the elimination ofpharmaceuticals is mostly unknown.

Joss et al. studied the removal of estrogens in municipal wastewater treatment

processes showing a removal of > 90% for estrogens during activated sludge

treatment.1111 Similar efficiencies were observed for a fixed-bed reactor and a

membrane bioreactor operated with a sludge age of 30 days. The fate and

behavior of antimicrobial agents in wastewater treatment and the aquatic

environment has only been addressed in a few studies.[12"15] McArdell et al.

investigated macrolides in WWTP effluents and a receiving watershed and

observed no significant elimination of clarithromycin in the investigated river

stretch.[14] For fluoroquinolones however, Golet et al. reported removal

efficiencies between 48 and 66% in the same river.[12] Tn conventional

wastewater treatment fluoroquinolones are eliminated by 88 to 91%, with

sorption to sewage sludge being the main process responsible.[13] By comparing

in- and out-flowing loads of a German wastewater treatment plant an

elimination of 94 ± 4% was reported by Ternes et al. for sulfamethoxazole and

ranged between 30 and 61% for the investigated macrolides.[,5] No significant

elimination was observed for trimethoprim (18 ± 14%).

The occurrence and sorption behavior of sulfonamide and macrolide

antimicrobials in activated sludge treatment is discussed in Chapter 4 including

110 Chapter 5

mass flow studies. Sorption to activated sewage sludge was found to be ofminor

importance for these compounds, with estimated sorption constants (Kd)

between 114 and 460 L/kg. In general, compounds with a Kd < 500 L/kg are

eliminated by less than 10% through sorption on to activated sludge at an

average specific sludge production of 200 g/m3.[16]The main aim of this study was to compare the elimination of sulfonamides,

macrolides and trimethoprim in different wastewater technologies using

complete mass flow analyses, including sewage sludge measurements. By

elimination the combination of all processes involved, e.g. transformation and

sorption, is addressed. Two conventional activated sludge treatment plants, a

fixed-bed reactor and a membrane bioreactor pilot plant, operated at 3 different

sludge ages, are compared as well as two different sand filters as tertiary

treatment steps. The results obtained are discussed with regard to general

parameters, e.g. temperature, hydraulic retention time and solid retention time.



The wastewater treatment plant of Kloten-Opfikon (WWTP-K) treats 55 000

population equivalents (PE): the combined sewage of 25 900 residents and of an

unknown number of air traffic passengers in the catchment area (Figure 5.1).

The average inflow (dry weather) amounted to 16 500 m3/d. The main

wastewater characteristics are summarized in Table 5.1. Primary treatment

consists of a screen, an aerated grit removal tank, and a primary clarifier.

Approximately 60%> of the primary effluent passes an activated sludge treatment

system operated at a sludge age of 3 days and a hydraulic retention time of 5 h

(V = 2 500 m3) as a pre-treatment. The main conventional activated sludge

treatment (CAS-K) includes denitrification (V = 1 900 m3) and nitrification (V =

3 700 m3) with a solid retention time of 10 - 12 d. The hydraulic retention time

(HRT) including the secondary clarifier is -15 h. Phosphate is removed by

simultaneous precipitation with Fe3+ in secondary treatment. The effluent is

discharged to the receiving water after filtration in a discontinuously operated

two-layer sand filter (SF-K). SF-K consists of 8 compartments with a total

volume of 288 m3, which are filled with a 1.2 m thick layer of schistose material

(diameter 2-3 mm) and 0.4 m of silica sand (diameter 0.7 - 1.2 mm). The filters

Behavior 111

Table 5.1 Average wastewater characteristics of of the municipal wastewater

treatment plants Kloten-Opfikon (WWTP-K) andAltenrhein (WWTP-A)

WWTP-K WWTP-A

parameter raw tertiary raw tertiary

influent effluent influent effluent

PH 5.8 ±0.9 6.8 ±0.2 7.5 ±0.3 7.6 ±0.1

BOD5 [mg/L] 220 ± 67 2.2 ± 0.7 210 ±45 8.7 ±4.6

COD [mg/L] 360 ± 76 16±2 590 ±300 35 ± 10

Ntot [mg/L] 43 ±6 23 ±4 38 ±6 15 ±7.6

Ptot [mg/L] 5.5 ±0.8 0.4 ±0.1 6.7 ±1.4 0.5 ±0.2

are backwashed daily by pumping a combination of air and filtrate through the

filter bed for ~25 min. The secondary sludge of both units is recycled to the

primary clarifier together with the sand filter backwash, except for the first two

sampling campaigns (March 2002 and February 2003), where the excess sludge

of the main unit was directly transferred to sludge treatment.

The membrane bioreactor pilot plant (MBR, 100 PE) is operated in parallel to

CAS-K using primary effluent at a flow rate proportional to raw water influent

(HRT -13 h). The bioreactor consists of a stirred anaerobic compartment (V = 6

or 8 m3), denitrification (V = 4 m3) and nitrification (V = 6 m3). The solid

retention time was increased in-between sampling campaigns from 16 ± 2 over

33 ± 3 to 60 - 80 d (steady state operation for two to three sludge ages prior to

sampling). For the 60 - 80 d sludge age no steady state was achieved, but the

sludge age was steadily increased over 110 d from a solid retention time of

33 ± 3 d to 60 - 80 d. In the final aerobic compartment, the secondary effluent

(permeate) is generated by three different membrane filtration units with a

maximal flow rate of 1.3 m3/h each. For test purposes a microfiltration plate

membrane module (Kubota A50), and two ultrafiltration hollow-fibre modules

(Mitsubishi Aqua-RM and Zenon ZeeWeed 500-C) were run in parallel. The

respective nominal pore sizes were 0.4 pm for microfiltration and 0.1 pm and

0.04 pm, respectively, for ultrafiltration.

The wastewater treatment plant of Altenrhein (WWTP-A), located in the canton

St. Gall near the Austrian border, handles the combined sewage of 80 000 PE,

including 52 000 inhabitants (Figure 5.2). The average inflow (dry weather)

Figure

5.1Flowschemeofthemu

nicipalwastewater

treatmentplantKloten-Opf

ikon

(WWTP-K),showingsamplingpoints

forwastewaterandsl

udge

.A

conventionalactivatedsludge

system(CAS-K,55000PE)andamembrane

bioreactor(MBR,

100PE)

areoperatedinparallel

raw

influent

primary

influent

effluent

CAS-K

-40%

secondary

effluent

tertiary

effluent

grit

removal

tank

compositesample

pnmary

sludge

^t

grabsample

secondary

(exc

ess)

sludge

sand

filter

4flffi5£

o|/k

>o/o

MBR1J

secondaryeffluent

(permeat)

secondary

(exc

ess)

sludge

a

CAS-K

consistsofdenitrificationand

nitrificationcascadewithahydraulicretentiontimeof-15handa

solidretentiontimeof10

-

12

d.Approximately60%

oftheprimary

effluentpass

firs

tlyth

roug

han

additional

activated

slud

gesystem(CAS),

oper

ated

ata

hydr

auli

cretentiontimeof-5handasolidretentiontimeof3

d.

bMBR

consistsofananaerobictankanda

denitrificationand

nitrificationcascadewithahy

drau

licretentiontimeof-13

h.The

solid

retentiontimeswere16±2d,33±3dand60

-80dduring

there

spec

tive

samplingca

mpai

gns.

Figure

5.2Flowschemeofthemu

nicipalwastewater

treatmentplantAltenrhein(WWTP-A),

showingsamplingpointsfor

wastewaterand

slud

ge.A


treatmentplant(CAS-A,

40000PE)andfixed-bedreactor(FBR,

40000PE)

areop

erat

edinparallel

rawinfluent

screen

grit

removal

tank

^P

compositesample

^t

grabsample

primary

effluent

primary

sludge

O0

o

CAS-Aa

secondary

effluent

secondary

(exc

ess)

sludge

öono

öooo

"~T

Q__°_0__Q_^Q___o_-

I

FBRb

filter

backwash

secondary

clarifier

secondary

effluent

tertiary

effluent

sand

filter

a

CAS-A

consistsofdenitrificationand

nitrificationcascadewithahydraulicretentiontimeof-31handasolidretentiontimeof21

25

d.

FBR

consistsofadenitrificationandnitrificationzonewithahydraulicretentiontimeof-1

h.

114 Chapter 5

amounted to 21 000 m3/d. The main wastewater characteristics are summarized

in Table 5.1. Primary treatment consists of a screen, an aerated grit removal

tank, and a primary clarifier. Secondary treatment is performed in two parallel

operated treatment units: a conventional activated sludge (CAS-A) and a fixed-

bed reactor (FBR), receiving -50% of the primary effluent each. Both systems

are designed for nitrification and denitrification. Conventional activated sludge

treatment includes a denitrifying volume (anoxic, mixed) of 2 300 m~ and

nitrifying (aerobic) volume of 6 800 m3. The solid retention time in CAS-A

ranged between 21 d to 25 d, while no value can be given for the FBR. The

hydraulic retention time was -31 h for the CAS-A including the secondary

clarifier, whereas it ranges below 1 h for the FBR. The FBR consists of 8

Biostyr up-flow cells L17] filled with 3.6 mm Styrofoam beads as biofilm support

(V = 1 500 m3) and an anoxic zone (below) and an aerated zone (above the

aeration nozzles). Excess sludge is removed by daily backwash with secondary

effluent. The secondary effluents of both units are combined and treated by a

continuously operated one-layer sand filter (SF-A) before discharge. SF-A

consists of 8 up-flow filter units containing a 1.5 m high silica sand column with

a total volume of-360 m3. The sand is continuously cleaned by pumping it from

bottom to top with an average turnover time of 6 - 8 h. The secondary excess

sludge of the CAS-A and the FBR is recycled to the primary clarifier.


Figures 5.1 and 5.2 give an overview of the chosen sample locations in both

wastewater treatment plants investigated. Sampling campaigns were performed

in March 2002, February 2003 and November 2003 at WWTP-K and in

September 2002 and March 2003 at WWTP-A. At both locations samples were

taken from the raw influent after the screen, from the primary effluent after

primary settling, from the secondary effluents after the respective biological

treatment and after sand filtration. In all cases flow proportional composite

samples were taken. For weekly sampling campaigns, 24 h composite samples

were mixed flow proportional to yield two samples during the week (combining

two and three days, respectively) and one during the weekend (combining

Saturday and Sunday). In February 2003, also 8 h composite samples were

collected for a period of 24 h. Additionally, grab samples were taken in these

three consecutive 8 h intervals from the anaerobic and anoxic compartment of

Behavior 115

the MBR and from the anoxic compartment of CAS-K. For sludge analysis grab

samples were taken from the aerobic compartment of the CAS-K, the CAS-A

and the MBR, respectively from the filter backwash in the case of the FBR.

After freeze-drying, they were combined for each plant and sampling campaign

to yield one weekly sample.

5.2.3 ChemicalAnalysis

Details on the methods applied to aqueous samples and for sludge analysis

including materials and reagents used are described in Chapter 2 and 3. Briefly,

aqueous samples were filtered (0.45-pm cellulose nitrate filters, Schleicher &

Schuell) and subsequently concentrated by solid phase extraction on polymeric

cartridges. Analysis is performed using reversed phase liquid chromatography

coupled to tandem electrospray mass spectrometry in the positive ionization

mode. Sample-based quantification limits depended on the analyte and the

sample matrix and ranged between 1 and 214 ng/L. Sludge samples were filtered

(glass fiber filters, GF8, Whatman) and the solid fraction was freeze-dried. The

investigated compounds were extracted using pressurized liquid extraction with

water:methanol (1:1) as extraction solvent. The extracts were diluted and

analyzed with the same method as aqueous samples. The limits of quantification

for activated sludge varied between 3 and 41 pg/kg.

5.2.4 Calculated Elimination

Average daily loads were used in all cases unless stated otherwise. They were

calculated from the weekly loads being the sum of the loads determined in the

two or three day sampling intervals (Chapter 4). Elimination (weight

percentage) resulted from comparing the loads entering a specific treatment step

to the sum of all loads leaving the treatment step with the aqueous phase.

Elimination therefore combines sorption onto excess sludge and transformation

due to biological or chemical processes.

The sorbed fraction in the excess sludge was calculated from the measured

concentration in the weekly sludge samples and the specific sludge production

of the secondary treatment. It was assessed using the measured chemical oxygen

demand in the primary effluent and general correlations according to literature

data.[18] During the performed sampling campaigns the average excess sludge

116 Chapter 5

production ranged between 0.1 and 0.2 g/L. The specific sludge production was

used, since the daily withdrawn amount of excess sludge is related mainly to

operational criteria. In this manner, only the amount of newly produced sludge

in activated sludge treatment is taken into account for elimination through

sorption. Since the solid retention times is significantly higher than the hydraulic

retention time, the activated sludge in the reactor (-3 g/L) was considered to be

already in equilibrium with the dissolved fraction. This issue is discussed in

detail by Joss et al.[19]

The highest sorption coefficient obtained for secondary sludge was also used to

estimate the respective loads on primary sludge, since sorption coefficients were

not available for primary sludge. Because of the different sludge characteristics

of primary sludge compared to secondary sludge, the results for primary sludge

have to be regarded as rough estimates. This was considered by assigning a

relative uncertainty of 50% to the used values.

Elimination over the sand filter was calculated correspondingly to secondary

treatment, but no distinction between transformation and sorption was made.

However, sorption to solid particles during sand filtrations is assumed to be

negligible for the investigated compounds due to the small additional sludge

production in the filter.

The uncertainty of the calculated elimination was estimated using error

propagation from the relative measurement uncertainty of the respective in and

out flowing load. A correct calculation of the uncertainty including all factors is

not possible due to the inter-relation of the individual parameters and the small

amount of data sets available. The resulting uncertainty stated for a weekly

sampling campaign, do therefore only include the impact of measurement

uncertainties and do not mirror possible daily variations.


Complete mass balances were performed for sulfonamides, macrolides and

trimethoprim to assess their distribution in wastewater treatment and the

importance of individual treatment steps on the overall removal. The results

obtained for the individual weekly sampling campaigns are used to evaluate

possible differences in elimination due to changing treatment conditions, e.g.

solid retention time and temperature. The uncertainty connected to these values

represents only a calculated measurement uncertainty. To assess the behavior

Behavior 117

and fate of the investigated antimicrobials in a specific treatment step over time,

the elimination observed in different sampling campaigns was combined (e.g. in

Figure 5.3 and 5.6).

5.3.7 Primary Treatment

The elimination of the investigated antimicrobials observed in primary treatment

is generally low (Table 5.2). Varying results were obtained for sulfapyridine,

mirrored in a high standard deviation, while negative eliminations were found

for sulfamethoxazole. This is probably caused by the simultaneous presence of

de-conjugable substances, e.g. human metabolites, of these compounds in the

raw influent (Chapter 4). In accordance with the increase in sulfamethoxazole

loads, a slight elimination was observed for its main human metabolite, AT-

acetylsulfamethoxazole, in primary treatment (9 - 21%). In the case of the

investigated macrolides and trimethoprim an elimination of up to 33% was

observed. Due to the high uncertainty connected with the calculation of the

sorbed fractions in primary treatment as well as with the measurement of such

complex samples, this elimination has to be considered as not significant.

5.3.2 Secondary Treatment

The behavior of the investigated antimicrobials was investigated in two

conventional activated sludge systems and a fixed-bed reactor (Table 5.3 and

Figure 5.3). Additionally, weekly sampling campaigns were performed on a

membrane bioreactor operated at three different solid retention times (Figure

5.4). The observed eliminations combine the reduction due to sorption and

transformation processes. However, the share of elimination caused by sorption

to secondary excess sludge generally ranged below 6%> for all compounds

investigated. Compared to the overall uncertainty connected with elimination

studies, sorption to activated sludge seems to be of minor importance for the

investigated sulfonamides, macrolides and trimethoprim. No distinction between

the processes responsible for the observed elimination rates is therefore made in

the following.

Reported negative elimination results from an observed increase of loads from

the inflow to the outflow of the respective treatment step - mainly observed for

sulfapyridine and sulfamethoxazole. In some cases more than twice the

118 Chapter 5

Table 5.2 Elimination ofsulfonamides, macrolides and trimethoprim in primary

wastewater treatment

compounda

acronym CASRN elimination

(%)b

range

sulfapyridine SPY 144-83-2 -29-20

sulfamethoxazole SMX 723-46-6 -21--5

A^-acetylsulfamethoxazole N4AcSMX9-21

SMX±N4AcSMXc0-9

trimethoprim TRI 738-70-5 -13-31

azithromycin AZI 83905-01-5 10-33

erythromycin ERY-H20 114-07-8 -8-4

clarithromycin CLA 81103-11-9 11-14

roxithromycin ROX 80214-83-1 3-9

a

Sulfadiazine (CASRN 68-35-9), sulfathiazole (CASRN 72-14-0) and sulfamethazine

(CASRN 57-68-1) were not at all or very rarely detected.

b

Range of the results obtained in February 2003 and November 2003 at WWTP-K and

March 2203 at WWTP-A. No raw influent was sampled in the other two sampling

campaigns. Negative values result from an observed increase of loads from inflow to

outflow of the respective treatment step.cSulfamethoxazole load, including the amount present as A^-acetylsulfamethoxazole.

inflowing load was detected for these compounds in the respective secondary

effluents of the conventional activated sludge systems and the fixed-bed reactor

(Table 5.3). For both compounds a very inconsistent picture is obtained with

elimination rates ranging between +72 and -104% for sulfapyridine and ±60 and

-138% for sulfamethoxazole. This can be explained by the presence of

substances, e.g. human metabolites, in the inflow, which are subsequently

transformed to sulfapyridine and sulfamethoxazole during biological treatment

(Chapter 4). Neither sulfasalazine, the administered pharmaceuticals containing

sulfapyridine, nor its main human metabolite, A^-acetylsulfapyridine, were

included in this study. The results obtained, however, strongly suggest the

presence of one or both of them in the influent and a possible transformation to

sulfapyridine in biological treatment.

In the case of sulfamethoxazole the fate of the main human metabolite,

Table

5.3Eliminationofsu

lfon

amides

,macrolidesand

trimethoprim

intwo


systemsanda

fixed-bedreactor

CAS-Ka

CAS-A

a

FBR

a

solidretentiontime

(d)

12

12

10

25

21

--

wastewatertemperature(°C)

14

12

16

19

12

19

12

compound

March

2002

February

2003

elimination(%)b

November

September

2003

2002

March

2003

September

2002

March

2003

SPY

-74±66c

-16±45

-107±8

49±5

72±5

52±5

41±9

SMX

-107±8

9±3

-79±7

-138±15

60±3

-61±

10

29±4

N4AcSMX

94±2

87±1

90±1

96±2

85±1

81±1

89±1

SMX+N4AcSMX

50±3

53±1

-1±3

61±3

76±1

60±2

67±2

TRI

3±5

-1±6

14±5

20±

11

-40±20

17±

11

12±11

AZI

d-26±8

-18±7

55±4

22±11

30±6

-13±10

ERY-H20

6±4

-14±4

-22±4

-6±8

-9±8

7±7

-13±8

CLA

9±4

-45±7

-7±5

4±7

20±6

5.6±6

14±6

ROX

18±4

38±3

-18±6

38±5

5±8

35±6

4±8

aCAS-K=conventionalactivatedsl

udge

system

atthemu

nici

palwastewatertreatmentpl

antKl

oten

-Opf

ikon

,Switerland.

CAS-A=conventionalactivatedsl

udge

system

atthemu

nici

palwastewatertreatmentplan

tAl

tenr

hein

,Switzerland

FBR:

fixed-bedreactoratthemunicipa

lwastewatertreatmentplantAltenrhein,Switzerland.

b

Nega

tive

valuesresultfromanobservedincreaseofloadsfrominflowtooutflowoftherespective

treatment

step.

c

Unce

rtai

ntyestimatedfromrelativemeasurementuncertaintyofinandoutfl

owin

gload.

dNo

resultsavailablebecauseofanalytic

alinterferences.

120 Chapter 5

A^-acetylsulfamethoxazole, was also investigated. A high elimination rate of up

to 96% was observed for this compound in the two conventional activated

sludge systems and the fixed-bed reactor (Table 5.3). The possible

transformation of A^-acetylsulfamethoxazole to sulfamethoxazole and a

simultaneous elimination of sulfamethoxazole itself during biological treatment

probably lead to the observed high variability of sulfamethoxazole elimination.

Taking the fraction present as A^-acetylsulfamethoxazole into account, an

average reduction of sulfamethoxazole to relative residual loads of 32 to 49%

was found in the conventional activated sludge systems and the fixed-bed

reactor (Figure 5.3). In the case of CAS-K only the results from the first two

sampling campaigns were taken into account. In the third sampling campaign in

November 2003 no significant elimination was observed for the

sulfamethoxazole load including the fraction present as A^-acetyl-

sulfamethoxazole, in contrast to March 2002 and February 2003 (Table 5.3). No

explanation could be found comparing wastewater temperature or known

operational parameters, e.g. solid retention time.

In the membrane bioreactor an elimination ofA^-acetylsulfamethoxazole of over

95 ± 2% was detected independent of the prevailing solid retention time (Figure

5.4). For sulfapyridine and sulfamethoxazole similar elimination rates were

observed in all three sampling campaigns. The average relative residual loads

measured were 46 ± 5% and 63 ± 1%, respectively. In contrast to the other

treatment technologies investigated, no increase in sulfapyridine and

sulfamethoxazole loads was observed from the influent to the effluent of the

membrane bioreactor. These findings suggest an effective simultaneous

elimination of these compounds during biological treatment in the membrane

bioreactor. Subsequently, a slightly increased average elimination of 78 ± 10%

was observed for sulfamethoxazole, including the fraction present as 1\-

acetylsulfamethoxazole, in the membrane bioreactor (Figure 5.4) compared to

conventional activated sludge treatment and the fixed-bed reactor (Figure 5.3).

For trimethoprim only a slight elimination of up to 20% was observed in

conventional activated sludge treatment and the fixed-bed reactor, not taking the

deviating result of -40 ± 20 obtained in March 2003 into account. Varying

results, including negative elimination, were obtained for the investigated

macrolides, with no obvious pattern between the sampling campaigns (Table

5.3). The presence of de-conjugable metabolites, however, seems unlikely for

macrolides.f20] Since they are mainly excreted with bile and feces, the load

Figure

5.3Relativeresidualloadsofsu

lfon

amid

es,macrolidesandtrimethoprim

intheef

flue

ntof

twoconventionalactivated

sludge

systemsandafi

xed-

bed

reactora

SPY

SMX

N4AcSMX

SMX

+

N4AcSMX

TRI

AZI

ERY-H20

CLA

ROX

aErrorbarsrepresentthe

standard

deviationfrom

threesamplingcampaigns(CAS-K)and

therangeoftwo

values(CAS-A,FBR),

resp

ecti

vely

.

Figure

5.4

Relative

residual

loadsofsu

lfon

amid

es,

macrolidesand

trimethopr

iminamembrane

bioreactoroperated

at

differentsolidretention

timesa

150

:,

,—.

N4AcSMX

SMX

TRI

AZI

ERY-H20

CLA

ROX

+

N4AcSMX

aErrorbarsrepresenttheuncertaintiesestimatedfromrelativemeasurementuncertaintiesofinandoutfl

owin

gload.

bNo

resultsavailablebecauseofanalyt

ical

interferences.

Behavior 123

entering biological treatment may be underestimated taking only the dissolved

fraction and sorption to the suspended solids into account. With this, the amount

enclosed in feces particles, which may be released during biological treatment,

is neglected. It may, together with possible differences in the sludge

composition, lead to the observed variations in the elimination of macrolides.

The average relative residual loads observed for macrolides after the two

conventional activated sludge systems and the fixed-bed reactor, ranged between

78 and 122%, taking the observed inter-campaign variations into account

(Figure 5.3).

As for the sulfonamides, no increase in loads was observed for macrolides and

trimethoprim in the membrane bioreactor (Figure 5.4). Additionally, elimination

observed for these compounds in the membrane bioreactor tends to be higher

compared to conventional activated sludge treatment and the fixed-bed reactor

(Figure 5.3). At solid retention times comparable to those of the other

technologies investigated, i.e. 16 ± 2 and 33 ± 3 d, similar relative residuals

were obtained in the membrane bioreactor. They ranged between 39 and 72%

for trimethoprim and the investigated macrolides, except azithromycin (96%).

Since sorption to excess sludge is of minor importance, operational differences

between the investigated treatment technologies, e.g. concerning the sludge

retention and redox conditions (additional anaerobic compartment), may cause

the observed differences. They may have an effect on certain wastewater and

sludge characteristics as well as on the biodiversity of the microbial flora

present.

5.3.3 Solid Retention Time

The solid retention time describes the mean residence time of the biomass within

the system. It is positively correlated with the concentration of the biomass if the

volume of the system is constant. Increased solid retention times allow the

enrichment of slow growing bacteria and may therefore lead to a broader

microbial flora potentially resulting in extended physiological capabilities.

Additionally, it leads to an increase of the inert fraction of the sludge via the

accumulation of decay products and inorganic material, and a decrease of

observed sludge production per volume of wastewater, caused by higher sludge

decay.

124 Chapter 5

A possible influence on the elimination of sulfonamides, macrolides and

trimethoprim were investigated in a membrane bioreactor operated at three

different sludge ages (Figure 5.4). In the beginning (March 2002) a solid

retention time of 16 ± 2 d, comparable to that of conventional activated sludge

systems, was investigated. For the second sampling campaign (February 2003) it

was increased to 33 ± 3 d, representing a solid retention time usually applied in

membrane filtration systems. Additionally, the impact of a high sludge age (60 -

80 d) on the elimination of the selected compounds was investigated (November

2003).

No dependence of the elimination on the solid retention time could be seen for

the selected sulfonamides including A^-acetylsulfamethoxazole in the membrane

bioreactor. For trimethoprim and most of the investigated macrolide

antimicrobials a similar elimination was observed at a solid retention time of 16

± 2 and 33 ± 3 d, respectively. However, a two to three times higher reduction

was seen for these compounds (except roxithromycin) at a solid retention time

of 60 - 80 d compared to the lower solid retention times. It ranges between 87

and 90% for trimethoprim, dehydro-erythromycin and clarithromycin and

amounts to 25% for azithromycin in this last sampling campaign. In the case of

roxithromycin, the observed elimination increases from 39% at a solid retention

time of 16 ± 2 days to -60% at the two higher solid retention times.

Similar average eliminations were obtained for all compounds investigated in

both conventional activated sludge systems independently of the solid retention

time, which ranged between 21 - 25 d in CAS-A and between 10 - 12 d in CAS-

K (Figure 5.3). Also in literature, no dependence of the elimination on the solid

retention time was reported for sulfamethoxazole and roxithromycin in

laboratory scale experiments,12'1 which confirms the observations made in the

MBR. An elimination of 70 to 90% was reported for sulfamethoxazole at

varying solid retention times between 1 d and 35 d. For roxithromycin no

significant elimination (9%) was observed at a solid retention time of 1 d, while

it ranged between 39 and 66%> at higher solid retention times.

5.3.4 Substrate Dependencies

The partial elimination observed for sulfonamides in the MBR independent of

the solid retention time (Figure 5.4), suggests a correlation of the elimination

with the substrate concentration in the influent for these compounds, i.e. with

Behavior 125

the ratio of a specific substrate to the sulfonamide concentration. This is

supported by the fact that no further elimination of sulfonamides is observed

during sand filtration as described below.

For trimethoprim and the investigated macrolides (except roxithromycin) the

results obtained in the membrane bioreactor indicate that the transformation of

these compounds may be inversely correlated to the sludge loading. By sludge

loading the ratio of substrate input to the amount of biomass present is assessed.

Since the reactor volume is constant, higher solid retention times result in

increased sludge concentrations, and therefore in a reduced sludge loading. The

combination of high solid retention times and reduced sludge loading may cause

an increase in the biodiversity of the active biomass, which may have an

influence on the elimination of compounds undergoing co-metabolism, as

assumed for antimicrobials.

In the case of A^-acetylsulfamethoxazole an almost complete elimination was

observed in all systems investigated and seems to occur under all treatment

conditions investigated. This may suggest a transformation by a widely available

biological reaction or even by abiotic processes.

However, in all cases direct experimental evidence, e.g. from batch experiments,

would be necessary to confirm possible substrate dependencies of the

eliminations observed for the investigated antimicrobials in biological treatment.

5.3.5 Anaerobic Compartment

Another factor that may influence the elimination of sulfonamide and macrolide

antimicrobials observed in the membrane bioreactor may be the additional

anaerobic compartment not present in the investigated conventional activated

sludge treatment plants. Therefore the elimination of the selected compounds in

the different redox compartments was investigated. The concentrations in the

inflow and the outflow of the respective cascade used for biological treatment

were measured in 8 h composite samples. Grab samples were taken from the

anaerobic and anoxic compartment of the membrane bioreactor operated at a

solid retention time of 33 ± 3 d and from the anoxic compartment of CAS-K in

the same 8 h intervals. From this, expected concentrations, assuming no

elimination in the respective compartment, were estimated. The calculated

concentrations, assuming no elimination, were than compared to the

concentrations measured to investigate a possible elimination under the given

Figure5.5Infl

uenc

eofdifférentredoxconditionsonthetr

ansf

orma

tion

ofsulfonamideandmacrolide

antimicrobialsa

300

250

200

150

O) f

100

O

50

anoxicco

mpar

tmen

tCAS-K

Daerobiccr

jrrp

artm

ertCAS-K

anaerobiccorrpartmertM3R

Bano>dccorrpartmert

h/BR

DaerobiccorrpartmentM3R

SPY

SMX

N4ACSMX

SMX

+

N4AcSMX

TRI

AZI

ERY-H20

CLA

ROX

aErrorbarsrepresentthestandarddeviationoftheresultsfromthree8hsa

mpli

ngintervals.

bComparisonofestimatedconcentrations(c

alcu

late

dassumingnoeliminationinthere

spective

compartment)andmeasured

concentrationsingrab

samplesfromtheanaerobic,

anoxicandaerobiccompartmentsofthemembranebioreactor(MBR)andthe

aerobicandanoxiccompartmentsoftheconventionalactivatedsludge

system(CAS-K)

atWWTP

Klot

en-O

pfik

on.

Behavior 127

redox conditions. Figure 5.5 summarizes the results obtained combining all three

8 h intervals. They have to be regarded with caution, however, since only grab

samples were used.

For AAacetylsulfamethoxazole a significantly lower concentrations than

estimated assuming no elimination, were measured in all compartments,

suggesting a transformation of this human metabolite under all redox conditions

given. For sulfamethoxazole itself, the measured concentrations were higher

than those estimated assuming no elimination, in the anaerobic and anoxic

compartments and lower than expected in the aerobic compartments.

Sulfamethoxazole hence seems to be fransformed mainly under aerobic

conditions. The observed accumulation in the other compartments may be

explained by the postulated transformation of /V-acetylsulfamethoxazole to

sulfamethoxazole. A pattern similar to sulfamethoxazole was found for

sulfapyridine, suggesting the presence of retransformable substances in the

influent of biological treatment also in this case.

No significant differences between estimated concentrations assuming no

elimination and measured concentrations were obtained for trimethoprim,

azithromycin and dehydro-erythromycin. For clarithromycin and roxithromycin

a tendency to lower concentrations measured than those estimated assuming no

elimination was observed in the anaerobic and anoxic compartment of the

membrane bioreactor. The combination of anaerobic and anoxic treatment may

therefore have an impact on the elimination of specific compounds, e.g.

macrolide antimicrobials. However, further experiments, e.g. using composite

samples or laboratory scale experiments, would be necessary to confirm the

observed behavior.

5.3.6 Wastewater Temperature

Sampling campaigns were performed at different times of the year resulting in

different wastewater temperatures. In the case of WWTP-A, sampling was

performed in March 2003 and in September 2002, resulting in an increase of the

wastewater temperature from 12 °C to 19 °C. Generally, no correlation between

the water temperature and the respective elimination was observed for the

investigated compounds (Table 5.3). A tendency to higher elimination with

increased water temperature could be seen for azithromycin and roxithromycin.

However, one has to be careful in interpreting this tendency, since variations at

128 Chapter 5

similar temperatures but different sampling times are in the same range.

Additionally, it is unclear, whether temperature dependencies as commonly

observed for biological treatment,1181 also apply to the transformation of

antimicrobials, or micropollutants in general.

5.3.7 Hydraulic Retention Time

Although the hydraulic retention time in the fixed-bed reactor ranged below 1 h,

a similar elimination as in conventional activated sludge treatment with a

hydraulic retention time of up to 31 h was observed (Figure 5.3). The results

indicate that the lower hydraulic retention time in the fixed-bed reactor is

approximately compensated by a higher bioactive sludge concentration per

reactor volume. In the FBR this is caused by the regular discharge of fast

growing heterotrophs and inert material with the backwash of the filter bed. In

general, it seems that for a given influent composition the elimination of

antimicrobials is not directly related to the hydraulic retention time but to a

similar efficiency in nutrient removal. Another possible factor increasing the

elimination of antimicrobials, if assuming a pseudo-first-order-kinetic depending

on the substance concentration as shown for estrogens/111 may be the strict plug

flow regime in the filter bed. It may lead to higher momentary concentrations of

the antimicrobials compared to the concentration in the fully mixed reactors in

conventional activated sludge treatment.

5.3.8 Sand Filtration

Two sand filters used for tertiary treatment - SF-K in WWTP Kloten Opfikon

and SF-A in WWTP Altenrhein - were investigated with respect to their removal

efficiency for sulfonamides, macrolides and trimethoprim (Figure 5.6).

In all five sampling campaigns higher loads of sulfapyridine were detected in the

tertiary effluents compared to the respective influent loads of the sand filters,

resulting in an average increase of 28 ± 17% for sulfapyridine during sand

filtration. It can be inferred that significant amounts of possible sulfapyridine

precursors, e.g. de-conjugable human metabolites, pass through biological

treatment and are transformed to sulfapyridine during sand flltration. Since only

sulfapyridine itself was analyzed in this study no information on a potential

Figure

5.6Relativeresidualloadofsu

lfon

amid

es,

macrolidesandtrimethoprim

inthe

effluent

oftwosandfiltersusedas

tert

iary

treatmenta

SPY

SMX

TRI

AZI

ERY-H20

CLA

ROX

N4AcSMX

a

Errorbarsrepresentthestandarddeviationfromthreesamplingcampaigns

forthesand

filt

eratWWTP-K

(SF-K)andtherangeoftwo

valuesforthesand

filter

atWWTP-A

(SF-A),re

spec

tive

ly.

130 Chapter 5

simultaneous transformation of the parent compound itself on the sand filter is

available.

In the case of sulfamethoxazole the main human metabolite, A^-acetyl-

sulfamethoxazole, was included in the investigation. However, only very small

concentrations of this compound were detected in secondary effluents due to the

high transformation efficiency of biological treatment for V-acetyl-

sulfamethoxazole. The total sulfamethoxazole load, including the amount

present as //-acetylsulfamethoxazole, did not significantly change during sand

flltration (Figure 5.6). Combining all sampling campaigns, no significant change

of the azithromycin load was observed, however high variations were observed

between sampling campaigns, mirrored by the large error bars.

For dehydro-erythromycin, clarithromycin and roxithromycin comparable loads

were detected in the inflow and outflow of one of the sand filters (SF-A), while

a significant reduction was observed for these compounds on the other sand

filter (SF-K). The average elimination amounts to 20 ± 5% for dehydro-

erythromycin, 17 ± 9% for clarithromycin and 23 ± 4% for roxithromycin during

tertiary treatment with SF-K, located at the WWTP Kloten-Opfikon. A poor

elimination of 15 ± 10% was detected for trimethoprim on the sand filter of

WWTP Altenrhein (SF-A). In the case of WWTP-K, sand flltration (SF-K) is

the main step responsible for a reduction of the trimethoprim load with an

average elimination of 74 ± 14%.

Due to the small amount of suspended solids produced in the sand filter, the

effect of sorption can be neglected for the investigated compounds. The results

therefore suggest a highly effective and diverse biofilm present on the sand

particles of SF-K. Additionally, it is noteworthy that an elimination on SF-K is

mainly seen for those compounds also showing an increased elimination in the

membrane bioreactor operated at a high sludge age, i.e. trimethoprim, dehydro-

erythromycin and clarithromycin.

An elimination between 56 - 70% on SF-K was observed for nonylphenolic

compounds by Wettstein.[22] From laboratory scale studies, the observed

elimination for these substances was assigned to the biological activity of the

sand filter material and not to a catalytic or reactive effective of the sand itself.

Even though both sand filters investigated were operated at a hydraulic retention

time of ~ 25 min and have similar hydraulic loads per surface of biofilm,

significant differences were observed in the elimination of certain

antimicrobials, especially trimethoprim. One possible explanation may be the

Behavior 131

different availability of oxygen during the passage of the two filters. The

secondary effluent of WWTP-K is aerated to oxygen saturation (aerated

flocculation reactor) prior to sand filtration, while the secondary effluents at

WWTP-A directly enters tertiary treatment. Further on, the water entering SF-A

shows a higher biological oxygen demand (BOD5) of 10 - 20 mg/L compared to

~3 mg/L in the case of SF-K. This may lead to additional oxygen consumption

during the sand filter passage. The resulting oxygen limitations in SF-A

compared to SF-K may result in the observed differences in the elimination of

certain antimicrobials. Another possible explanation could lay in the design and

operation of the two sand filters investigated. While SF-A is regenerated

continuously, by turning over the sand bed, only a daily backwash of SF-K is

performed. Together with the oxygen availability, this may result in a more

diverse and stable microbial community in SF-K. Additionally, a possible

formation of specific zones may be more predominant in a two-layer filter

disturbed only periodically (SF-K), compared to a continuously regenerated one-

layer filter (SF-A).

5.4 Conclusions

For selected macrolides and trimethoprim elimination seems to correlate with

increasing solid retention time and, hence, with reduced substrate loading. Both

parameters may lead to an increased biodiversity of the active biomass, resulting

in a broader range of degradation pathways available. For sulfonamides a

positive correlation of the observed elimination to the organic substrate

concentration appears to exist. In the case of sand filtration, used as tertiary

treatment, elimination efficiencies were observed for selected macrolides and

trimethoprim and gave the impression to depend on operational parameters of

the sand filter, e.g. oxygen limitations. The correlations observed for the

elimination of the selected antimicrobials in field studies should be verified

under more controlled conditions, i.e. in batch experiments.

Even though the removal in wastewater treatment may be optimized for certain

compounds, e.g. by an increased sludge age or an additional tertiary treatment

step, the overall elimination of sulfonamide and macrolide antimicrobials in

wastewater treatment was shown to be incomplete. Therefore, residual amounts

of these emerging contaminants are continuously discharged to receiving surface

waters. Our ongoing research aims at investigating the ozonation of wastewater

132 Chapter 5

effluents as a possible technique to further reduce the antimicrobial load

entering the aquatic environment. Additionally, no knowledge on the

transformation products formed in wastewater treatment is available. The latter

topic needs to be investigated further, especially considering the environmental

risk assessment for these compounds.

Acknowledgements




2000-00047)[23] and the EAWAG project on human-use antibiotics

(HUMABRA) within the framework of the National Research Program on

antibiotic resistance funded by the Swiss National Science Foundation/241 We

would also like to thank the Swiss Agency for the Environment, Forestry and

Landscape, the Swiss cantons of Aargau, Basel Land, Bern, Luzern,

Schaffhausen, Schwyz, St. Gallen, Thurgau, Ticino, Zurich and the WWTPs of

Kloten-Opfikon and Altenrhein for additional financial support. We thank the

technical staff of the WWTP Kloten Opfikon and the WWTP Altenrhein for

their assistance during sampling. For helpful comments on the manuscript we

acknowledge M.Suter and T.Ternes.


[1 ] Stan, H. J.; Heberer, T. Analusis 1997, 25, M20-M23.


Lützenhßft, H. C; Jorgensen, S. E. Chemosphere 1998, 36, 357-393.

[3] Daughton, C. D.; Ternes, T. A. Environ. Health Perspect. 1999,107, 907-

938.

[4] Kümmerer, K. Pharmaceuticals in the environment: Source, fate, effects

and risks; Springer: Berlin, Heidelberg, New York, 2001.

[5] Erickson, B. E. Environ. Sei. Technol. 2002, 36, 140-144.

[6] Heberer, T. Toxicol. Lett. 2002, 131, 5-17.

[7] Diaz-Cruz, M. S.; Lopez de Aida, M. J.; Barcelo, D. Trends Anal. Chem.

2003,22,340-351.

Behavior 133

8] Giger, W.; Alder, A. C; Golet, E. M.; Kohler, H.-P. E.; McArdell, C. S.;


9] Boxall, A. B. A.; Fogg, L. A.; Blackwell, P. A.; Kay, P.; Pemberton, E. J.;

Croxford, A. Reviews in Environmental Contamination and Toxicology

2004,180,1-91.

10] Ternes, T. A.; Joss, A.; Siegrist, H. Environ. Sei. Technol. 2004, 38, 392A-

399A.

11] Joss, A.; Andersen, H. R.; Ternes, T. A.; Richie, P. R.; Siegrist, H. R.

Environ. Sei. Technol. 2004, 38, 3047-3055.

12] Golet, E. M.; Alder, A. C; Giger, W. Environ. Sei. Technol. 2002, 36,

3645-3651.

13] Golet, E. M.; Xifra, L; Siegrist, H. R.; Alder, A. C; Giger, W. Environ. Sei.

Technol. 2003, 37, 3243-3249.

14] McArdell, C. S.; Molnar, E.; Suter, M. J.-F.; Giger, W. Environ. Sei.

Technol. 2003, 37, 5479-5486.

15] Ternes, T. A., Habilitation Thesis, Mainz, 2000.

16] Ternes, T. A.; Herrmann, N.; Bonerz, M.; Knacker, T.; Siegrist, H.; Joss,

A. Water Res. 2004, 38, 4075-4084.

17] Rogalla, F.; Lamouche, A.; Specht, W.; Kleiber, W. Water Sei. & Technol.

1994, 29, 207-216.

18] ATV-DVWK Fachausschuss, "Arbeitsblatt ATV-DVKW-A 131," ISBN 3-

933707-41-2, 2000.

19] Joss, A.; Keller, E.; Göbel, A.; Alder, A. C; McArdell, C. S.; Siegrist, H.

submitted to Water Research.

20] Bryskier, A. J.; Butzler, J.-P.; Neu, H. C; Tulkens, P. M. Macrolides;


21] Clara, M.; Strenn. B.; Gans, O.; Kreuzinger, N. In Water Resources

Management II; Brebbia, C. A., Ed.; WIT Press: Southampton, UK, 2003;

Vol. ISBN 1-85318-967-4, pp 227-236.

22] Wettstein, F. "Auftreten und Verhalten von Nonylphenoxyessigsäure und

weiteren Nonylphenolverbindungen in der Abwasserreinigung," PhD

Thesis No. 15315, ETH Zurich, 2004.

23] http://www.eu-poseidon.com.

24] http://www.nrp49.ch/pages/.

c

L'a ^u*^

Chapter 6

Ozonation

Sulfonamide and macrolide antimicrobials are generally not completely

eliminated during wastewater treatment. To further reduce the residual amounts

being discharged to receiving waters, the ozonation of wastewater effluents was

investigated. Extensive oxidation of sulfonamide and macrolide antimicrobials

was achieved at an ozone dose of only 2 mg/L, resulting in a reduction of the

respective loads by over 90%. A reduced reactivity towards ozone was observed

for acetylated sulfonamides, e.g. the human metabolite A^-acetyl-

sulfamethoxazole, illustrating the significance of the aniline moiety in

sulfonamides as the main site of ozone attack. The amount of suspended solids

in the wastewater, varying between 3 and 22 mg/L, had no significant impact on

the oxidation efficiencies for all investigated compounds. As expected,

variations in oxidation efficiencies occurred in relation to the pH of the

wastewater, since the rate constants of the compounds as well as the ozone

stability are pH dependent.

partly published in

Huber, M.M., Göbel, A., Joss, A., Hermann, N., Löffler, D., McArdell, CS.,

Ried, A., Siegrist, H., Ternes, T.A., von Gunten, U.

Oxidation of Pharmaceuticals during the Ozonation of Municipal Wastewater

Effluents: A Pilot Study

Environmental Science and Technology, 2005, in press.

Ozonation 137

6.1 Introduction

In recent years, various studies have reported the occurrence of a large number

of pharmaceuticals in the aquatic environment/1"31 Even though the detected

concentration levels are typically in the ng/L to pg/L range, it cannot be

excluded that molecules designed to be biologically active, affect sensitive

aquatic organisms even at such low concentrations. Immediate effects caused by

pharmaceuticals may be subtle and difficult to detect, but nevertheless could

lead to important long-term consequences in aquatic ecosystems/41 After

consumption, pharmaceuticals are excreted from the human body in the

unchanged form or in the form of human metabolites. In developed countries,

wastewater is usually treated in wastewater treatment plants before it is

discharged into receiving surface waters. Municipal wastewater is therefore the

major source of human pharmaceuticals in the aquatic environment. Since it is

highly unrealistic to prohibit or limit the consumption of any pharmaceuticals,

the improvement of wastewater treatment is an effective option to diminish the

release of these compounds into the aquatic environment. Among the various

classes of pharmaceuticals, antimicrobials are of special interest due to the

potential spread and maintenance of antibacterial resistance, especially in human

pathogens. Based on annual consumption data in Switzerland, the most

important antimicrobial classes applied in human medicine are the ß-lactams

(17.5 t/a), sulfonamides (5.7 t/a), macrolides (4.3 t/a) and fluoroquinolones

(3.9 t/a).[5"7] While ß-lactams seem to be hydrolyzed shortly after excretion,

representatives of the other three classes have been detected in wastewater

effluents/8121 The occurrence and fate of fluoroquinolones in the environment

has been intensively studied by Golet et al.,[13'14] while the occurrence and fate of

sulfonamides, macrolides and trimethoprim is the focus of this dissertation.

They are, as shown in the previous chapters, removed to varying extents in

wastewater treatment. Elimination efficiencies proved to be strongly compound

dependent and also varied with the treatment technology investigated. However,

incomplete elimination in wastewater treatment was observed in all cases.

Residual concentrations of these emerging contaminants are therefore constantly

discharged to the aquatic environment. Advanced treatment technologies would

be necessary to achieve a further removal of sulfonamides, macrolides and

trimethoprim from wastewater treatment plant effluents. Ozonation has shown a

high potential for the oxidation of pharmaceuticals in drinking water[,5'16] and

138 Chapter 6

wastewater/ J In wastewater, ozone doses ranging from 5 to 15 mg/L led to a

complete disappearance of most of the pharmaceuticals except iodinated X-ray

contrast media. Also the antimicrobials investigated were eliminated to below

the limit of quantification by an ozone concentration of 5 mg/L, which equals

elimination rates between 76 and 92%.

The aim of the present study was to investigate the removal of selected

sulfonamide and macrolide antimicrobials, including two sulfonamide

metabolites, and trimethoprim, from wastewater effluents by ozonation. To be

able to determine removal of 95 - 99%, the selected antimicrobials were spiked

to the wastewater. By applying comparatively low ozone doses ranging from 0.5

to 5 mg/L, we aimed at finding minimum doses required for the removal of the

selected pharmaceutical classes and to gain a better insight into the ozonation

process. To investigate the influence of particles, three effluents with different

concentrations of suspended solids were treated in the pilot plant. The secondary

wastewater effluents used exhibit a substantially lower dissolved organic carbon

(DOC) concentration than the wastewater investigated by Ternes et al/17] Such

conditions are more representative for Switzerland, due to the higher dilution of

wastewater in Switzerland by e.g. extraneous wasters.


6.2.1 Ozonation Pilot Plant

The pilot plant consists of two ozonation columns operated in series with an

active reactor volume of 140 liter each and a filling level of 4.8 m (0.193 m

nominal inner diameter, 5.2 m total height). A flow scheme of the plant is given

in Figure 6.1. The water enters the plant on the top of column 1, operated in the

downstream mode and leaves the plant at the top of column 2, operated in the

upstream mode. Tracer experiments with a salt spike showed a slightly better

plug flow behavior in the second column as compared to the first. Modeling the

reactor volume as a series of 3 and 4 fully mixed compartments with comparable

total volume, could best simulate the salinity profile at the outflow of column 1

and 2, respectively. With a flow rate of 2 ± 0.1 m3/h, the total hydraulic retention

time amounts to 4.2 ± 0.2 min in each column. The ozone was continuously

supplied by an ozone generator (Ozomatic SWO 200) fed with technical oxygen.

Ozone containing gas was applied at the bottom of column 1 at a flow of

Ozonation 139

200 ±10 L/h, resulting in a counter flow of water and gas bubbles. Ozone

concentrations in the feed and off gas were measured with a UV ozone monitor

(BMT 936 Vent, 0.1-50 g/m3). By adjusting the power input of the ozone

generator, the desired ozone concentrations were obtained. The respective

concentrations yielded transferred ozone doses of 0.5, 1, 2.5, 3.5 and 5 mg/L in

wastewater. Transfer efficiencies were > 98%. No ozone was applied to column

2, serving as a reaction vessel for the dissolved ozone. The residual ozone

concentration was measured at the interface between column 1 and 2 (C (O3-

Cl)) and at the outflow of column 2 (C (03-C2)).

Table 6.1 Set up ofthe ozonation pilot plant used

column 1 column 2

6.2.2 Wastewater Effluents

The pilot plant was operated on site of the municipal wastewater treatment plant

(WWTP-K) in Kloten-Opfikon, Switzerland, located near the international

airport of Zurich. The combined sewage of 55 000 population equivalents (PE)

is treated using a conventional activated sludge system. A pilot-scale membrane

bioreactor (100 PE) is operated in parallel. A detailed description of the plant

140 Chapter 6

and the treatment technologies applied are given in Chapter 5. For the ozonation

experiments the secondary effluent of the conventional activated sludge system

(CAS) and the final effluent of the membrane pilot plant (MBR) were used. The

respective water was continuously pumped into a 300 L tub, mechanically

mixed, and subsequently fed to the pilot plant. To increase the total amount of

suspended solids (TSS) by -15 mg/L (CAS±TSS), sludge from the conventional

activated sludge system was continuously added to the inflow of the tub in one

experiment. The resulting water quality parameters of the 3 different matrices

investigated are summarized in Table 6.1.

Table 6.1 Average water quality parameters of the three wastewater effluents

investigateda

effluent pH T DOC COD TSS alkalinity

[°C] [mg/L] [mg/L] [mg/L] [mM]

CAS 7.0 ±0.1 16±1 7.7 ±0.5 29 ±3 7±2 3.1 ±0.1

CAS±TSS 6.9 ±0.1 15 ±1 7.0 ±0.5 41 ±1 22 ±2 3.2 ±0.2

MBR 7.5 ±0.1 17 ±1 6.6 ±0.2 22 ±2 3±2 5.4 ±0.2

aErrors represent one standard deviation.


An aqueous solution of the investigated antimicrobials (Figure 6.2 and 6.3) was

prepared and continuously added to the respective wastewater effluent prior to

entering the mixing tub. It was taken care that acetone residuals from primary

stock solutions were low enough not to influence the ozonation process. A -500

fold dilution of the spiking solution with wastewater effluent resulted in a final

concentration of ~2 pg/L for all antimicrobials expect for A^-acetyl-

sulfamethazine (~0.5 pg/L). Due to the limited commercial availability of

azithromycin at the time, this compound was not spiked. Also A^-acetyl-

sulfamethoxazole, the human metabolite of sulfamethoxazole, was not spiked to

eliminate a possible influence on the sulfamethoxazole concentration. A^-acetyl-

sulfamethazine, the analogous metabolite of sulfamethazine, a sulfonamide

primarily used in veterinary medicine, was spiked instead as a model substance

for A^-acetylsulfamethoxazole. Sulfamethazine itself, however,

Ozonation 141

Figure 6.2 Chemical structures ofsulfonamide antimicrobials and trimethoprim

andproposed main sites ofozone attack

o,

r~

h\_/o

-s-

II0

R2

trimethoprim

(TRI)

sulfadiazine

(SDZ)

/ sulfathiazole

\ (STZ)

sulfapyridine

(SPY)

sulfamethoxazole

(SMX)

A^-acetyl^ sulfamethoxazole *

(N4AcSMX)

sulfamethazine *

(SMZ)

A^-acetylsulfamethazine

(N4AcSMZ)

Rl

H

H

H

H

COCH

H

COCH3

R2

N=\

-O

—i J

N=\

-o

-O

. HX

CH3

CH,

* Not spiked, but present in wastewater samples.

was not added. The actual concentration of all antimicrobials investigated, was

determined in each matrix and for each ozone dose applied, in the inflow

(C (In)) and the outflow (C (Out)) of the ozonation pilot plant.

6.2.4 Sample Collection and Chemical Analysis

Samples were taken at the inflow of column 1, the interface between the two

columns and at the outflow of column 2 (Figure 6.1). Dissolved ozone

concentrations were determined on the respective sampling days in the later two

samples, using the indigo method/181 The detection limit was 0.05 mg/L. For the

determination of antimicrobials, 100 mL of inflow and 250 mL of outflow were

taken, adjusted to pH 4, and enriched unfiltered using solid phase extraction on

142 Chapter 6

Figure 6.3 Chemical structure of macrolide antimicrobials andproposed main

site ofozone attack

\/N

03

,.nl«*

azithromycin

H°V/^OH \>*

erythromycin

(ERY)

clarithromycin

(CLA)

roxithromycin

(ROX)

H3C,

CH3

Ri

H

CH3

H

R2

0

O

/V^N/

aPH3

dehydro-erythromycin

(ERY-H20)

1 y

V\

o'y^Z-^H

Not spiked, but present in wastewater samples.

Oasis HLB (Waters) cartridges on the sampling day. The dried cartridges were

then frozen and transported to the laboratory, where they were eluted within one

week. Measurement was performed using reversed-phase liquid chromatography

coupled to electrospray positive tandem mass spectrometry. Duplicate analysis

was performed in all cases. Details on the analytical method used, including

materials and reagents, are given in Chapter 2.

For method validation the accuracy of the method and the sample-based limit of

quantification (LOQ) was determined in the investigated sample matrices. The

accuracy was determined by recovery studies in each wastewater matrix for the

inflow and the outflow after a reaction with 0.5 and 5 mg/L of ozone,

respectively. To non-spiked wastewater, used for the recovery studies, 100 and

200 ng of analytes, respectively, were added prior to extraction (n = 4). The

calculated amount of antimicrobials minus the amount already present before

Ozonation 143

spiking (n = 2) was then compared to the spiked amount. Additionally recovery

rates were verified in spiked wastewater treated with the selected ozone doses

(0.5 - 5 mg/L), by adding 250 ng of analytes prior to extraction to the respective

outflow samples (n = 2). The sample-based limits of quantification were defined

as those concentrations in a sample matrix resulting in a signal with signal-to-

noise (S/N) ratios of 10. The concentration corresponding to the defined S/N

was determined by scaling down, using the measured concentration in the

samples and the assigned S/N of the peak - assuming a linear correlation

through zero. Table 6.2 summarizes the obtained results for method validation in

each wastewater matrix. A^-acetylsulfamethoxazole was accidentally not

included in the recovery studies. Since an isotope labeled standard, A^-acetyl-sulfamethoxazole-û?5 was used, a relative recovery of 100 ± 10% was assumed

for the quantification of this compound. No results are given for sulfamethazine,

which was not spiked to the wastewater and also not detected in any of the non-

spiked samples. Therefore no method validation was performed for this

compound.

6.2.5 Calculation ofRelative Residual

To compensate differences in the input concentrations, the outflow

concentrations are reported as relative values. Therefore, relative residual

concentrations after the application of different ozone doses were calculated by

comparing the measured concentration in the inflow to the respective

concentration in the outflow of the ozonation pilot plant. The average of

duplicate analysis is used in all cases. The respective uncertainty of the relative

residuals was estimated by combining the uncertainty of the in- and out-flowing

concentration. Therefore the square root of twice the square of the measurement

uncertainty was used as relative uncertainty of the residual percentage. The

standard deviation of the recovery studies was used as an estimation of the

measurement uncertainty of the investigated antimicrobials in the three matrices

(Table 6.2).

Table

6.2Methodvalidationparametersforsu

lfon

amid

es,macrolidesandtrimet

hoprim

inwastewater

effluents

compound

relativerecovery(%)

average±SD(n=18-22)

CAS

CAS+TSS

MBR

sample-basedLOQ

(ng/

L)

average(r

ange

)(n=

22)

CAS

CAS±TSS

MBR

SDZ

109±4

105±3

107±4

68

(8-

252)

50(9

-186)

69

(10

-

287)

STZ

102±2

100±3

102±5

83(11-261)

104(17-431)

135(18-534)

SPY

117±6

107±9

112±

18

114(56-135)

77(12-383)

77

(9-

507)

SMX

106±9

99±6

104±5

34

(6-

169)

42(5-163)

39

(4-

209)

N4AcSMX

a

(100±10)

(100±10)

(100

±10)

38(10-145)

51(27-134)

49(18-169)

N4AcSMZ

b118±

11

122±8

121±

11

16(4-43)

25

(6-

87)

18(4-71)

TRIC

98±15

59±15

81±17

7(2-21)

16(2-91)

17

(2-

126)

AZI

114±13

165±25

103±13

3(0

.2-

7)3

(0.6

-

12)

3(0.2

-

8)

ERY-H20

103±

14

118±18

93±16

9(0.1-34)

11(0.4-41)

11(1.4-36)

CLA

99±13

114±11

91±15

10(0.1-67)

12

(1-

47)

8(0.1-53)

ROX

128±9

143±20

128±17

7(0.3

-

30)

5(0.3

-

26)

6(0.3

-

32)

a

N4AcSMXwas

acci

dent

ally

notincludedintherecovery

studies.Estimatedvaluesaregi

veninbrackets.

bA^-acetylsulfamethoxazole-wasusedassurrogatestandardforN4AcSMZ,

fortheothercompounds

refertoChapter

2.

c

InthecaseofTRI,only

therecoveriesfromno

n-sp

iked

samp

leswereconsidered(n=

12).

Ozonation 145



After spiking, the concentrations in the inflow to the ozonation pilot plant were

determined at the same time as the concentrations after the respective ozone

treatment. By comparing the measured concentration in the inflow to the spiked

amount, the spiking procedure and the chemical analysis can be examined. In

the case of azithromycin and //-acetyl sulfamethoxazole, which were not spiked,

the measured concentrations are normalized to the average concentration

measured to assess the general variability. Within one wastewater matrix the

concentrations measured in the spiked wastewater showed a very consistent

picture, generally with a variation of less than ± 30%. The concentrations

measured, also agreed well (± 30%.) with the spiked amounts, as shown for the

five different inflows of the ozonation pilot plant from the experiments with

secondary effluent of the conventional activated sludge system (Figure 6.4). In

the case of dehydro-erythromycin and roxithromycin the measured

concentrations were generally higher than those theoretically spiked, reaching

values of up to 150%. Great variations, however, were observed for

trimethoprim within and between the experiments. This is caused by the high

concentration spiked, which was outside of the linear range of the calibration.

Therefore only the results obtained from non-spiked samples during recovery

studies are discussed in this study in the case oftrimethoprim.

Generally, the spiking procedure proved to be successful and led to reproducible

concentrations in the inflow. However, the results also illustrate the importance

of also measuring respective inflowing concentrations.

6.3.2 Oxidation Efficiencies

In Figure 6.5 the relative residual concentrations of the investigated

sulfonamides, macrolides and trimethoprim are given as a function of the ozone

dose applied to the spiked secondary effluent of conventional activated sludge

treatment. Additionally the results for A^-acetylsulfamethoxazole, the main

human metabolite of sulfamethoxazole, and TA^-acetylsulfamethazine, the

analogous metabolite of the veterinary sulfonamide sulfamethazine, are given.

In the case of trimethoprim results from the recovery studies in non-spiked

Figure

6.4Comparisonoftheoreticalandmeasuredantimicrobialconcentration

inspiked

secondary

effl

uent

(CAS)

atfi

ve

differenttimepointsa

SDZ

STZ

SPY

SMX

N4AcSMZ

TRI

AZIb

ERY-H20

CLA

ROX

N4AcSMXb

aErrorbarrepresenttherangeofduplicatean

alys

es.

b

Compared

toanaverageenvironmentalco

ncen

trat

ion,

sincenotsp

iked.

Table

6.3Residualozone

concentrationsmeasured

inwastewater

effl

uent

streatedwithselectedozonedoses

afte

rthefirst

(CI)

andthesecondcolumn(C2)

oftheozonationpi

lotplant

appl

ied

CAS

(Mo)

a

ozonedose

Cl

C2

5mg/L

2.5

0.7

3.5mg/L

1.4

0.3

2mg/L

-

lmg/L

-

0.5mg/L

-

residualozoneconcentration(mg/L)

CAS

CAS±TSS

CI

1.8

0.14

<0.05

<0.05

C2

0.6

<0.05

<0.05

<0.05

CI

C2

MBR

CI

C2

1.6

0.2

1.7

0.34

0.6

<0.05

0.7

<0.05

0.08

<0.05

<0.05

<0.05

<0.05

<0.05

<0.05

<0.05

<0.05

<0.05

<0.05

<0.05

Due

tothehigh

erdilutionofthewastewater,

resultsobtainedonMonday

aregi

vense

parately

fromthoseonotherweekdays.

148 Chapter 6

matrix are shown, and therefore results are only available for ozone doses of 0.5

and 5 mg/L.

Efficient oxidation (> 90 %) was observed for all antimicrobials at ozone doses

above 2 mg/L. These findings correlate well with the measured ozone residuals

(Table 6.3). For ozone doses below 2 mg/L no significant ozone residuals (<

0.05 mg/L) were detected at the outlet of the first column, indicating that ozone

concentrations in the bulk liquid are presumably close to zero. The observed

reaction of antimicrobials with ozone must therefore mainly take place in the

liquid film surrounding the gas bubbles. Under these conditions, the

antimicrobials have to compete with reactive wastewater components for ozone.

Consequently, even sulfonamide and macrolide antimicrobials with very high

reaction rate constants for ozone/15J are only partly oxidized. From the results

the initial ozone demand of the investigated wastewater was estimated to be ~2

mg/L, since dissolved ozone could be measured in the effluent of column 1 at

higher ozone doses.

As expected, similar oxidation patterns were observed for all sulfonamide and

macrolide representatives (Figure 6.5). Within these classes, compounds are

structurally very similar and it can be assumed that ozone attack takes place on

the same functional groups. The reactive functional group in macrolides and

sulfonamides are the tertiary amino group and the aniline moiety, respectively

(Figure 6.2 and 6.3). Since the chemical environment of these reactive moieties

is in most cases quite similar within one class, it can also be assumed that the

rate constants for the reaction with ozone must be very similar. Therefore also

the oxidation efficiency for all compounds of a class should be comparable.

Consequently, a very concise picture was obtained for the four macrolides -

greater variations were observed for the four sulfonamides investigated, which

might be caused by one of the following reasons: On one hand, it cannot be

excluded that in the case of sulfathiazole the thiazole moiety is more reactive to

ozone than the aniline moiety. On the other hand, the speciation of the

sulfonamides at the given pH may influence their reactivity towards ozone. The

pKa of the /»-amino group ranges between 5.7 (sulfamethoxazole) and 8.4

(sulfapyridine) for the investigated sulfonamides. Consequently,

sulfamethoxazole is present in its anionic and sulfapyridine in its neutral form,

whereas the remaining sulfonamides are present as a mixture of both species.

Anionic species can be many times more reactive towards ozone than their

neutral equivalents - the observed variation in the reactivity of the sulfonamides

Ozonation 149

is therefore surprisingly low. A possible explanation is that the higher electron

density on the acidic nitrogen reflected by a higher pKa, extents to the adjacent

moieties, making them significantly more reactive towards ozone. The reactivity

of the neutral form of sulfapyridine seems, therefore, to be as high as that of the

anion of sulfamethoxazole. Consequently the reasonable agreement in the

oxidation pattern may be a coincidence in the case of the sulfonamides. In

general, significant differences in the extent of parent compound oxidation have

to be expected when the compared compounds exhibit different speciation under

the investigated conditions. In the case of the acetylated sulfonamide

metabolites, the ozone reactive moiety is protected by an acetyl group.

Therefore, its reactivity to ozone is considerably reduced and the oxidation

efficiencies much lower compared to the respective parent compounds. At an

ozone dose of 2 mg/L, for example, only -45% of the human metabolite

A^-acetylsulfamethoxazole were removed by oxidation with ozone, while a

reduction by ~98%> was observed for sulfamethoxazole itself. However, only

small amounts of /V-acetylsulfamethoxazole can be detected in secondary

effluents due to the effective transformation of this compound in biological

treatment (see Chapter 4 and 5)

6.3.3 Influence ofthe Wastewater Matrix

Suspended solids

The present study was performed with three different wastewater effluents. The

water quality parameters of the investigated wastewater effluents are

summarized in Table 6.1. One main objective thereby was to investigate the

effect of suspended solids on the oxidation efficiencies of sulfonamide and

macrolide antimicrobials. The effluent of the membrane bioreactor (MBR)

represents a wastewater practically free of suspended solids, due to the small

pore size (< 0.4 pm) of the membranes. With the secondary effluent of the

conventional activated sludge system (CAS) an average effluent quality was

investigated. By fortifying this effluent with activated sludge (CAS±TSS) an

activated-sludge process with sub optimal clarification was simulated. In Figure

6.6 the relative residual concentrations in all three wastewater effluents are

given as a function of the applied ozone dosage for three selected compounds.

Results are shown for the most predominant representative of the sulfonamides,

150 Chapter 6

Figure 6.5 Relative residual concentrations ofsulfonamides (A), macrolides (B),

sulfonamide metabolites and trimethoprim (C) in CAS effluent for ozone doses

rangingfrom 0.5 to 5 mg/L

05mg/L H"lmg/L a2mg/L o35mg/L ö5mg/L

-100

o

I 75

o

II

I 50

D

e 25

ro

p

___(A)

pB

rii-L

i LSDZ STZ SPY SMX

AZI ERY-H20 CLA ROX

N4AcSMX N4AcSMZ

3

In the case of TRI the results from non-spiked samples are shown.

Ozonation 151

sulfamethoxazole, and of the macrolide antimicrobials, clarithromycin. Similar

patterns were observed for the other representatives of the two classes

investigated. Additionally, results from the main human metabolite of

sulfamethoxazole, A^-acetylsulfamethoxazole, are given. Overall the differences

in oxidation efficiencies are relatively small between the different effluents and

no significant trend can be observed. The amount of suspended solids in the

treated effluents is therefore of minor importance for the oxidation efficiencies

of sulfonamide and macrolide antimicrobials with ozone. This demonstrates

that, especially for low doses, ozone is consumed by dissolved components of

the wastewater before it reaches the sludge particles.

pH dependence

For the macrolides, e.g. clarithromycin, slightly better oxidation efficiencies

were observed for the effluent of the membrane bioreactor at low ozone doses

compared to the other two effluent matrices investigated. This can be explained

by the pH dependent reactivity of these compounds with ozone.tl5J In the case of

macrolides the neutral form is the most reactive species. With a pKa of -9 for

the tertiary amine group, the most predominant form present in the investigated

wastewater effluents, however, is the protonated form. Therefore the reactivity

of these compounds is higher at the increased pH of-7.5 in the effluent of the

membrane bioreactor compared to -7.0 in the other two effluents investigated.

In the case of the sulfonamides the dissociated form is the most reactive species,

present at a pH above the respective pKa of the /?-amino group. However, no

strong dependence of the reactivity with ozone on the speciation can be

observed as discussed above. Significantly lower oxidation efficiencies were

even observed for sulfonamides in the effluent of the membrane bioreactor at

ozone doses below 2 mg/L compared to the other two effluents investigated.

This effect can be attributed to a slightly faster decay of ozone caused by the

higher pH in the effluent of the membrane bioreactor. As a result no ozone

residuals were detected in this matrix at the outlet of column 1 after an ozone

application of 2 mg/L ozone (Table 6.3). Therefore less ozone was available for

the reaction with antimicrobials in the effluent of the membrane bioreactor

compared to the other two wastewater effluents investigated. As a result, also

significantly lower oxidation efficiencies were measured for macrolides in

effluent of the membrane bioreactor at an ozone dose of 2 mg/L (Figure 6.6).

152 Chapter 6

Acetylated sulfonamides

Ozone doses above 2 mg/L resulted in a complete oxidation of sulfonamide and

macrolide antimicrobials in all three matrices. In the case of sulfonamide

metabolites, e.g. A^-acetylsulfamethoxazole, a reduced reactivity with ozone is

observed, due to the protection of the reactive group by an acetyl group. In this

case, the direct reaction with ozone is less important at low ozone doses and the

oxidation with OH radicals becomes the predominant reaction. Therefore low

oxidation efficiencies are observed for A^-acetylsulfamethoxazole at ozone

doses between 0.5 and 2 mg/L that only slightly increase with increased ozone

dose applied. However, the oxidation efficiencies significantly increase at ozone

doses of 3.5 and 5 mg/L. In these cases, also significant residuals of dissolved

ozone were measured in the outflow of column 1, making the reaction with

ozone itself more relevant.

Residual ozone concentration

Additionally, the influence of the wastewater matrix was investigated on the

amount of residual ozone present. After an ozone application of 3.5 or 5 mg/L,

similar ozone residuals were measured at the outlet of column 1 in all three

wastewater effluents investigated. An exception are those ozone concentrations

measured on Monday in the secondary effluent of the conventional activated

sludge treatment, which were higher than those measured on Tuesday and in the

other wastewater matrices on weekdays. This difference can be explained by the

stronger dilution of the wastewater on the weekend, still present on Monday

morning. Generally it can be stated that the influence of suspended solids on the

dissolved ozone concentration is low after the first column. On the other hand,

ozone residuals at the outlet of the second column seem to be influenced by the

pH and the amount of suspended solids of the wastewater matrix. Therefore,

lower ozone residuals were measured after an application of 5 mg/L ozone in the

effluent of the membrane bioreactor (pH) and the fortified effluent of the

conventional activated sludge treatment (TSS). The fact that ozone residuals

measured in the effluent of the conventional activated sludge treatment are

higher than those in the particle free effluent of the membrane bioreactor

indicates that the pH difference is more important than the amount of suspended

Ozonation 153

Figure 6.6 Relative residual concentrations ofsulfamethoxazole, clarithromycin

and N4-acetylsulfamethoxazole in three wastewater effluents as afunction ofthe

ozone dose applied

I CAS iCAS+TSS qMBR

0.5 mg/L 03 1 mg/L 03 2 mg/L 03 3.5 mg/L 03 5 mg/L 03



154 Chapter 6

solids in this case. Similar conclusions can be drawn for the simultaneous

disinfection efficiencies by ozone treatment in wastewater effluents, since

disinfection strongly depends on the exposure to ozone itself.1-191

6.3.4 Comparison ofSpiked and Non-SpikedSamples

To be able to determine up to 95 - 99% removal of sulfonamide and macrolide

antimicrobials during ozonation, the selected compounds were spiked to the

wastewater matrices (0.5-2 pg/L). Thereby native concentrations were increased

by a factor of 2 to 420, depending on the antimicrobial compound. Additionally,

the reactivity of compounds not usually present in the investigated wastewater

samples (e.g. SDZ, STZ, N4AcSMZ) could also be investigated. Additionally,

relative recovery studies were performed in untreated wastewater and after the

application of 0.5 and 5 mg/L ozone, respectively, to verify the accuracy of the

analytical method. For the recovery studies non-spiked wastewater was used

and, consequently, only contained antimicrobials already present in the

wastewater. By comparing the concentrations measured in the untreated

wastewaters to those measured in the ozone treated samples, relative residual

amounts can be calculated for the two ozone doses applied in the recovery

studies. All samples of one matrix were taken on the same day within 4 hours

and measured in duplicate analysis. This provides the opportunity to verify that

experiments conducted with spiked antimicrobials yield similar results as

experiments performed at native concentrations. Additionally, the impact of

variations in the wastewater quality in between different days can be

investigated in the case of the non-spiked compounds. In Figure 6.7, the results

for macrolide and sulfonamide antimicrobials obtained at an ozone dose of 0.5

mg/L from the spiked experiments are plotted versus the results from the

recovery studies (non-spiked). The data points close to the solid line with a slope

of one indicate comparable behavior of the compounds. In both experiments, no

residual amounts of the investigated antimicrobials could be detected after an

application of 5 mg/L ozone. The result obtained in the spiked and non-spiked

experiments correlate very well. Significant differences were, however, detected

for the macrolide antimicrobials in the effluent ofthe membrane bioreactor. This

may be explained by a high turbidity observed in the respective wastewater on

the day ofthe recovery experiments. The cause ofthis turbidity in the membrane

effluent could not be identified, but may result in the observed differences

Ozonation 155

concerning the reaction of macrolides with ozone, probably due to sorption

effects.

Figure 6.7 Comparison of relative residual concentrations obtained in spiked

and non-spiked samples at an ozone dose of0.5 mg/L

0 20 40 60 80 100

relative residual obtained in non-spiked samples (%)

6.3.5 Conclusions

Ozonation of wastewater effluents was shown to be a potential tool for the

removal of sulfonamide and macrolide antimicrobials. An ozone dose of 2 mg/L

leads to an almost complete oxidation (> 90%) of the investigated

antimicrobials. The amount of suspended solids present in the wastewater

effluents showed no significant influence on the observed oxidation efficiencies.

The pH of the wastewater effluent, however, influenced the oxidation

efficiencies, due to the pH dependent stability of ozone. Additionally, pH

differences can affect the reaction rate constants of specific compounds with

ozone.

During the ozonation of wastewater only a partial oxidation of pharmaceuticals

is achieved and therefore could yield biologically still active oxidation products.

156 Chapter 6

Ozonation generally increases the number of polar functional groups and

thereby the overall polarity of the substance. However, in the case of

sulfonamide and macrolide antimicrobials the main sites of ozone attack (Figure

6.2 and 6.3) are also crucial for their effectiveness as antimicrobial agents.

Their ozonation products should therefore be less reactive and not further

promote the spread and maintenance of antibacterial resistance. For other

pharmaceuticals, recent studies on 17a-ethinylestradiol[21] and carbamazepine[have shown that partial oxidation was sufficient to significantly reduce

pharmacological activity and toxicity, respectively.

6.4 References cited

[I] Ternes, T. A. Water Res. 1998, 32, 3245-3260.





938.


1998.




[8] Hirsch, R.; Ternes, T. A.; Haberer, K.; Mehlich, A.; Ballwanz, F.; Kratz,

K.-L. J. Chromatogr., A 1998, 815, 213-223.



Technol. 2003, 37, 5479-5486.

[II] Giger, W.; Aider, A. C; Golet, E. M.; Kohler, H.-P. E.; McArdell, C. S.;


[12] Vanderford, B. J.; Pearson, R. A.; Rexing, D. J.; Snyder, S. Anal. Chem.

2003, 75, 6265-6274.


3645-3651.

[14] Golet, E. M.; Xifra, L; Siegrist, H. R.; Aider, A. C; Giger, W. Environ. Sei.

Technol. 2003, 37, 3243-3249.

i

Ozonation157

[15] Huber, M. M.; Canonica, S.; Park, G.-Y.; von Gunten, U. Environ. Sei.

Technol. 2003, 37, 1016-1024.

[16] Ternes, T. A.; Meisenheimer, M.; McDowell, D.; Sacher, F.; Brauch, H.-J.;

Haist-Gulde, B.; Preuss, G.; Wilme, U.; Zulei-Seibert, N. Environ. Sei.

Technol. 2002, 36, 3855-3863.

[17] Ternes, T. A.; Stüber, J.; Herrmann, N.; McDowell, D.; Ried, A.;

Kampmann, M.; Teiser, B. Water Res. 2003, 37, 1976-1982.

[18] Bader, H.; Hoigne, J. Water Res. 1981,17, 185-194.

[19] von Gunten, U. Water Res. 2003, 37, 1469-1487.

[20] Walsh, C. Antibiotics: actions, origins, resistance; ASM Press:

Washington, DC, 2003.

[21] Huber, M. M.; Ternes, T. A.; von Gunten, U. Environ. Sei. Technol. ASAP.

[22] McDowell, D.; Huber, M. M.; Wagner, M.; von Gunten, U.; Ternes, T. A.

Environ. Sei. Technol. in preparation.

! Blar

Chapter 7

Conclusions and Outlook

Seite Lee,, ;

Blank lc ' i

Conclusions and Outlook 161

Wastewater treatment is the main exposure route for human used

pharmaceuticals to the environment. The principle aim of this work was to

increase the knowledge on the occurrence and fate of sulfonamides, macrolides

and trimethoprim in wastewater treatment. Different treatment steps in

conventional wastewater treatment and newer techniques, i.e. a fixed bed reactor

and a membrane bioreactor, were investigated. The main achievements and

some general conclusions are summarized in the following paragraphs:

(1) Specific and reliable analytical methods were developed and validated for

the trace determination of sulfonamides, macrolides and trimethoprim in

wastewater and sewage sludge. These methods are suitable to

comprehensively investigate the occurrence, behavior, and fate of the

selected compounds in wastewater treatment from the raw influent to the

final effluent. With a few adaptations these analytical methods can also be

applied to other environmental matrices, e.g. hospital wastewater effluents,

surface and ground waters.

(2) This study provides an example for the environmental exposure assessment

of human used pharmaceuticals. The importance of composite samples was

illustrated by investigating daily profiles of the selected antimicrobials. The

loads entering wastewater treatment were shown to correlate reasonably well

with the loads estimated from available consumption data and to vary with

season. Additionally, it proved to be of crucial importance to include human

metabolites or other compounds that may be transformed to the active

pharmaceutical ingredient, in fate and behavior studies. This was

exemplified in this study by including A^-acetylsulfamethoxazole, the main

human metabolite of sulfamethoxazole.

(3) Mass flow analyses showed that sorption to sewage sludge plays a minor

role (< 10%) in the elimination of sulfonamides, macrolides and

trimethoprim in wastewater treatment. The predicted sorption coefficients to

activated sludge generally ranged below 500 L/kg for all substances

investigated.

162 Chapter 7

(4) Incomplete removal was observed for all investigated sulfonamide and

macrolide antimicrobials as well as for trimethoprim, in wastewater

treatment. The observed elimination strongly differed for each compound

class and the various treatment steps.

(5) For selected macrolides and trimethoprim elimination seems to be correlated

with increasing solid retention time and, hence, with reduced substrate

loading. Both parameters may lead to an increased biodiversity of the active

biomass, resulting in a broader range of degradation pathways available. For

sulfonamides a positive correlation of the observed elimination to the

organic substrate concentration seems to exist. In the case of sand filtration,

used as tertiary treatment, elimination efficiencies were observed for selected

macrolides and trimethoprim and appeared to depend on operational

parameters ofthe sand filter, e.g. oxygen limitations.

(6) Independent on the amount of suspended solids, sulfonamides, macrolides

and trimethoprim are efficiently oxidized (> 90%) in wastewater effluents by

ozone doses above 2 mg/L. Ozonation, thus, proves to be an efficient

technique to eliminate residual antimicrobials from wastewater, e.g. if

advisable for the receiving ambient water or for infiltration.

Overall, none of the technologies investigated for secondary wastewater

treatment led to an efficient removal (> 90%) of all investigated sulfonamides,

macrolides and trimethoprim. For this purpose, other techniques, e.g. post-

ozonation, were necessary. Residual concentrations of these emerging

contaminants therefore continuously reach ambient waters. Based on the

knowledge available today, however, adverse effects in the environment seem

unlikely.

The behavior and fate observed strongly depend on the compound investigated,

with similar results obtained within the group of sulfonamides and macrolides,

respectively. However, no correlations could be made between the elimination

in wastewater treatment and the molecular structure of the investigated

compounds. Consequently, no extrapolations of the results to other

pharmaceuticals are possible.

Conclusions and Outlook 163

Based on the results from this study the following additional aspects would be of

interest:

(1) In this study only A^-acetylsulfamethoxazole, as main human metabolite of

sulfamethoxazole, was included. By considering additional metabolites, e.g.

A^-acetylsulfapyridine, the occurrence and behavior of the selected

antimicrobials could be assessed to a larger extent. Additionally, to fully

investigate their fate in wastewater treatment, including ozonation, the

identification of transformation products would be necessary. These

transformation products should be included in the environmental risk

assessment ofthe respective compounds.

(2) Well designed laboratory studies are needed to elucidate significant fate

processes such as biotransformation and the cleavage of adducts. In

particular, the results obtained in this study strongly indicate the

transformation of A^-acetylsulfonamides to the respective parent compound

in wastewater treatment. However, laboratory scale studies under controlled

conditions would be necessary to verify this hypothesis.

(3) The dependences of the elimination of antimicrobials on specific parameters,

such as substrate concentration, solid retention time or redox conditions (e.g.

denitrifying and anaerob), should be further investigated.

(4) Based on the obtained data, the importance of municipal wastewater

treatment plants as exposure route for sulfonamide and macrolide

antimicrobials to the aquatic environment was shown. Further studies in

hospital effluents would be needed to assess their importance as point

sources.

(5) Profound studies on the ecotoxicological effects of sulfonamide and

macrolide antimicrobials in the environment, especially caused by the

constant exposure to low doses and complex mixtures, are needed to

evaluate the overall risk associated with the discharged residual

concentrations of these substances. Additionally, the impact of the low

concentrations in wastewater treatment on the spread and maintenance of

antimicrobial resistance needs to be further investigated.

oei .A

Vielen Dank...

..an alle, genannt oder ungenannt, die auf verschiedenste Weise zum Entstehen

dieser Arbeit beigetragen haben!

..an Walter Giger für die Leitung der Doktorarbeit, für die vielen Hinweise aus

seinem jahrelangen Erfahrungsschatz und das Vertrauen in mich und meine

Arbeit.

..an Christa McArdell für die ausgezeichnete Betreuung. Vor allem dafür, dass

sie immer Zeit für meine Fragen und das perfekte Maß zwischen Freiraum und

Unterstützung gefunden hat.

..an Hansruedi Siegrist und Thomas Ternes für die Übernahme des Korreferates,

insbesondere für die Diskussion und Ratschläge beim Publizieren der

Ergebnisse.

..an Elvira Keller für das Teilen von Labor und aller anfallenden Freude und

Frust - fachlich sowie privat. Die vielen Diskussionen, sowie ihre fachliche und

moralische Unterstützung waren eine große Hilfe für mich.

..an Angela Thomsen für ihr Interesse am Thema dieser Arbeit und für die

hervorragende Arbeit im Bereich der Schlammanalytik. Sie ist ein wichtiger

Baustein dieser Arbeit.

..an Adriano Joss für den unermüdlichen Enthusiasmus und die gute

Zusammenarbeit in praktischen sowie theoretischen Dingen. Gut, dass interne

Telefongespräche nichts kosten!

.. an Thomas Ternes für die ausgezeichnete Leitung des EU-Projektes

POSEIDON, das mir optimale Bedingungen fur eine spannende Dissertation

geboten hat. Allen Teilnehmern des Projekts, insbesondere Marc Huber, danke

ich für die gute Zusammenarbeit über fachliche, sprachliche und kulturelle

Grenzen hinweg.

..an alle „TSQ User" für die gute Zusammenarbeit beim Probleme lösen,

Termine verteilen und unerklärliche Phänomene verstehen. Insbesondere danke

ich Marc Suter und René Schönenberger für ihre Hilfe.

..an die Mitarbeiter der Kläranlagen Kloten-Opfikon und Altenrhein für die

Unterstützung bei der Probenahme, sowie bei Fragen und Sonderwünschen.

..an Bert Reinold und Bruno Tona (ETH) für die geduldige statistische Beratung

bei einer so kleinen Datenmenge.

..an alle ehemaligen und jetzigen CHPler für die tolle Gruppenatmosphäre und

die intensive Hilfe beim Einleben an der EAWAG. „Apéro" ist ein Wort, dessen

Bedeutung ich durch Euch gelernt habe.

..an die ganze EAWAG für die vielen offenen Türen und die unglaubliche

Bereitschaft mit Rat und Tat unkompliziert zu helfen. Dies hat die Zeit an der

EAWAG unvergesslich gemacht.

..an alle Mitdoktoranden/innen für den guten Austausch und die vielseitige

gemeinsame Freizeitgestaltung.

..an meine Familie für ihre Liebe und unermüdliche Unterstützung auf meinem

bisherigen Lebensweg. Ohne Euch wäre dies nie möglich gewesen.

..an Bert dafür, dass er all die Gedanken an die Arbeit ertragen hat, die mich

nach Hause begleitet haben.

Curriculum Vitae

Anke Göbel

geboren am 14.03.1975 in Köln, Deutschland

1981 - 1985 Städtische Katholische Grundschule, Köln-Porz, Deutschland

1985 - 1994 Stadtgymnasium Köln-Porz, Deutschland

17.06.1994 Abitur

1994 - 1998 Studium der Lebensmittelchemie an der Rheinischen

Friedrich-Wilhelms-Universität Bonn, Deutschland

03.12.1998 Erste staatliche Prüfung für Lebensmittelchemiker

1999 Forschungspraktikum an der

University of California, Davis, USA

1999 - 2000 Praktische Ausbildung am Chemischen Landes- und Staatlichen

Veterinäruntersuchungsamt Münster, Deutschland

15.11.2000 Zweite staatliche Prüfung für Lebensmittelchemiker

2001 - 2004 Dissertation an der Eidgenössischen Anstalt für

Wasserversorgung, Abwasserreinigung und Gewässerschutz

(EAWAG) und der Eidgenössischen Technischen Hochschule

(ETH), Zürich, Schweiz

antimicrobials in wastewater treatment27735/eth-27735-02.pdfmakrolid- und trimethoprim-frachten...

Documents