Environmental Science and Pollution Research

, Volume 19, Issue 3, pp 689–699

Occurrence and removal of lidocaine, tramadol, venlafaxine, and their metabolites in German wastewater treatment plants


    • Department of Environmental Analytical Chemistry, Institute of Atmospheric and Environmental SciencesJ.W. Goethe University Frankfurt am Main
  • Wilhelm Püttmann
    • Department of Environmental Analytical Chemistry, Institute of Atmospheric and Environmental SciencesJ.W. Goethe University Frankfurt am Main
Research Article

DOI: 10.1007/s11356-011-0614-1

Cite this article as:
Rúa-Gómez, P.C. & Püttmann, W. Environ Sci Pollut Res (2012) 19: 689. doi:10.1007/s11356-011-0614-1



Some of the pharmaceuticals that are not extensively investigated in the aquatic environment are the anesthetic lidocaine (LDC), the analgesic tramadol (TRA), and the antidepressant venlafaxine (VEN). LDC metabolizes to 2,6-xylidine (2,6-DMA) and monoethylglycinexylidine (MEGX), TRA to O-desmethyltramadol (ODT), and VEN to O-desmethylvenlafaxine (ODV). Within this study, the distribution and behavior of these compounds in German wastewater treatment plants (WWTPs) were investigated.


Samples of influents and effluents from WWTPs in Hesse, Germany were collected between January and September 2010. Analytes were extracted from wastewater samples by solid-phase extraction and from solid samples by sonication. Extracts were measured using gas chromatography/mass spectrometry.


LDC, TRA, VEN, ODT, and ODV were detected in all analyzed influent and effluent samples. 2,6-DMA could not be identified. MEGX was not detected. TRA and ODV were present in untreated wastewater at the highest concentrations (max, 1,129 (TRA) and 3,302 ng L−1 (ODV)), while the concentrations of LDC and VEN were all significantly lower (mean, 135 (LDC) and 116 ng L−1 (VEN)). All of the analytes were only partially removed in the WWTPs. The mean ratios between the concentrations of the metabolites and their respective parent compounds in influents were 4.7 (ODV/VEN) and 0.7 (ODT/TRA). These values remain approximately constant comparing influents and effluents.


LDC, TRA, VEN, ODT, and ODV are only partially removed from sewage water by WWTPs and thus are continuously discharged in respective recipient rivers. A further transformation of TRA and VEN into the known metabolites during treatment in the WWTPs is not observed.


WastewaterLidocaineTramadolVenlafaxineMetabolitesWastewater treatment plantRemoval rates

1 Introduction

A wide range of pharmaceutical compounds is used for the treatment of various diseases in both humans and domestic animals. These compounds and/or their metabolites are excreted through urine and feces into domestic wastewater and make their way to wastewater treatment plants (WWTPs). Some of them are not totally removed in the WWTPs and remain in the effluents entering surface waters (González Alonso et al. 2010; Zhou et al. 2010) and even groundwater (Einsiedl et al. 2010). The potential for a drug to reach the aquatic environment depends on three factors: the sales amount, the pharmacokinetic behavior (half life, urinary and fecal excretion, metabolism, etc.), and the rate of degradation of the compounds in the sewage system and in WWTPs.

In the meantime, a high amount of drugs and related metabolites have been detected in the aquatic environment and the list is continuously increasing (Choi et al. 2008; Gros et al. 2007; Kasprzyk-Hordern et al. 2008). Some non-extensively investigated drugs are the amino compounds lidocaine, tramadol, and venlafaxine. Lidocaine (2-(diethylamino)-N-(2,6-dimethylphenyl)ethanamide (LDC)), originally marketed by ASTRA AB as Xylocaine®, is widely used as a local anesthetic and antiarrhythmic agent by cardiac disorders. Both humans and animals metabolize LDC to various metabolites, most of them quite unstable at room temperature. One of its major active metabolites is 2-(ethylamino)-N-(2,6-dimethylphenyl)ethanamide (monoethylglycinexylidide (MEGX)), which has demonstrated stability at room temperature (Tam et al. 1990). MEGX metabolizes further into 2,6-dimethylaniline (2,6-xylidine (2,6-DMA)), suspected to be a potential carcinogen (Bangoluri et al. 2005). Tramadol ((1RS, 2RS)-2-(dimethylaminomethyl)-1-(3-methoxyphenyl)cyclohexanol (TRA)), marketed by Grünenthal GmbH as Tramal®, is a centrally acting analgesic. The metabolic fate of tramadol is unusually complex, O-desmethyltramadol (ODT) being its major active metabolite (Chytil et al. 2009). Venlafaxine (1-[2-(dimethylamino)-1-(4-methoxyphenyl)ethyl]cyclohexanol (VEN)), marketed by Pfizer as Effexor®, represents a new form of antidepressants which have the ability to affect the uptake of the neurotransmitters serotonin and norepinephrine. Its major active metabolite is O-desmethylvenlafaxine (ODV), a demethyl form of the parent compound.

According to their physicochemical properties, LDC, TRA, VEN, and the metabolites 2,6-DMA, MEGX, ODT, and ODV are expected to remain mainly in the water phase: the water solubilities of these compounds are relatively high and volatilization from the water phase into air is negligible because of the low Henry coefficients (Table 1). Furthermore, their n-octanol/water partition coefficients indicate a tendency to remain in the water phase instead of accumulation in sewage sludge or in aquatic organisms. However, all of the compounds can be toxic to aquatic organisms after direct ingestion. The reported 96-h LD50 values are 106 mg L−1 (Zebra fish) for LDC (Astra Zeneca GmbH 2001), 6.2 mg L−1 (Gold fish) for TRA (Sigma-Aldrich 2009), and 143.3 mg L−1 (Brachydario rerio) for 2,6-DMA (Sigma-Aldrich 2008). Because of the quantities of the compounds released into the environment, acute toxic effects of LDC, TRA, VEN, and their metabolites on aquatic organisms are unlikely. It would be useful to test the toxicity of the drugs in low doses, as chronic toxicity may develop long-term effects on aquatic species. Unfortunately, ecotoxicological data are available only for <1% of currently used pharmaceuticals (Sanderson et al. 2004).
Table 1

Physicochemical data of the substances and their metabolites analyzed in this study (BLAC 2003; Hanisch et al. 2004; Howell et al. 1993; Hummel et al. 2006; Kasprzyk-Hordern et al. 2008; Keenaghan and Boyes 1972; Lajaunesse et al. 2008; Paar et al. 1997; SRC PhysProp Database 2010)


In recent studies LDC, TRA, and VEN have been detected in effluents from WWTPs and in some rivers and lakes in both Europe and North America. However, there is only a limited data available about the reduction of these compounds and their metabolites in the WWTPs. A report from Eawag (German acronym for the Swiss Federal Institute of Aquatic Science and Technology) indicates the presence of LDC, TRA, and VEN at various locations in Lake Constance, Switzerland, at concentrations of 1, 9, and 2 ng L−1, respectively (Eawag 2009). Kasprzyk-Hordern et al. (2008) detected TRA at concentrations up to 7,000 ng L−1 in rivers of South-Wales, UK, impacted by treated wastewater discharges. Lajaunesse et al. (2008) reported the occurrence of VEN and its demethylated metabolite in municipal wastewater and in recipient waters in Montreal, Canada. More recently, VEN has been identified in various surface waters samples in Colorado, USA, at concentrations ranging between 102 and 690 ng L−1 (Schultz et al. 2010). Moreover, VEN and ODV were shown to be present in influents and effluents from WWTPs in southern Ontario, Canada, with effluent concentrations up to 800 (VEN) and 1,600 ng L−1 (ODV). The removal rates for both compounds in the WWTP were about 40% (Metcalfe et al. 2010).

The aim of the present study was to investigate the distribution and behavior of LDC, TRA, VEN, and the dominating metabolites in three WWTPs in Hesse, Germany. Daily samples from influents and effluents from the WWTPs were collected and analyzed for LDC, 2,6-DMA, MEGX, TRA, ODT, VEN, and ODV using solid-phase extraction (SPE) followed by gas chromatography–mass spectrometry (GC-MS). Moreover, solid samples obtained from the micro-screen rotating drum filter system (MRF) of one WWTP were extracted by sonication, in order to analyze the possible absorption of the analytes in particulate matter.

2 Experimental

2.1 Chemicals

LDC, 2,6-DMA, TRA hydrochloride, VEN hydrochloride, and the internal standard squalane were purchased from Sigma-Aldrich (Steinheim, Germany). MEGX was kindly supplied by AstraZeneca (Wedel, Germany). ODT, ODV, and the internal standard d6-TRA were provided by Toronto Research Chemical Inc. (Ontario, Canada). Individual stock solutions (1 μg μL−1) of target compounds and of the internal standard d6-TRA were prepared in methanol. The stock solution of the internal standard squalane (1 μg μL−1) was prepared in hexane. Working standards of the analytes were prepared from these stock solutions by dilution with methanol. All reported concentrations of target compounds and internal standards are concentrations of the hydrochloride-free analytes.

Acetone was obtained from LS Labor-Service (Griesheim, Germany). All other organic solvents were analytical grade (Carl Roth, Karlsruhe, Germany) and were distilled before use. Ultrapure water was produced from de-ionized water using an Astacus ultrapure water purification system (MembraPure, Bodenheim, Germany).

2.2 Study sites and sample collection

Twenty-four-hour composite samples of influents (raw sewage) and effluents (final treated sewage) were collected from three WWTPs in Hesse, Germany, over the period from 20 Jan 2010 to 09 Apr 2010. Details regarding the population served, treatment technologies, and the sampling period for these WWTPs are summarized in Table 2 where the WWTPs are coded with an identification number. The WWTPs treat wastewater of residential and industrial origin from a population of approximately 250,000 (WWTP1), 200,000 (WWTP2), and 25,800 (WWTP3). The WWTPs investigated consist of a preliminary clarification, a conventional activated sludge system (nitrification, denitrification), phosphate removal and a final clarification. After the final clarification, WWTP1 uses a micro-screen rotating drum filter system (MRF), which removes fine particles that remain in wastewater despite treatment in the secondary clarifier. The material retained in this process (90% liquid and 10% solid) is recirculated to the primary clarifier in WWTP1. In a second occasion, 24-h composite samples of influent and effluent of WWTP1 were collected from 16 Sept 2010 until 23 Sept 2010. During this time, the MRF had an operation time of only 30 min per day. Grab samples of the material recirculated by the MRF were collected from 21 Sept 2010 until 23 Sept 2010.
Table 2

Characteristics and operating conditions of the WWTPs sampled for influent and effluent





Purification steps




Sampling period

20 Jan 2010–02 Mar 2010 and 16 Sept 2010–23 Sept 2010

20 Jan 2010–26 Feb 2010

23 Feb 2010–09 Apr 2010

Population served (PE)

250,000 (urban)

200,000 (urban)

25,800 (rural)

Average flow (m3 day−1)




Wastewater type

68% R and 32% I

62% R and 38% I

94% R and 6% I

Collection system




HRT (h)




SRT (day)




Significant sources

5 hospitals

2 hospitals

1 pharmaceutical lab

12 specialist clinics

1 hospice

1 nursing home

5 nursing homes

6 nursing homes


M mechanical, B biological, N nitrification, D denitrification, P phosphorus removal, MRF micro-screen rotating drum filter system, PE population equivalents, R residential, I industrial/commercial, HRT hydraulic retention time, SRT solid retention time

The wastewater samples were collected using refrigerated autosamplers. Defined volumes of influent and effluent samples were collected every 60 min to obtain a composite sample over a 24-h period. In WWTP2, the volume collected every 60 min was proportional to the flow rates at a collection time on the influent and on the effluent, respectively. In WWTP1 and WWTP3, the sample volume was the same every 60 min. The water samples were collected in 1-L brown glass bottles and stored cooled at 4°C in the dark until processing in laboratory within 7 days after sampling. The bottles were cleaned before use with a laboratory glassware washer G 7835 CD (Miele, Gütersloh, Germany), then flushed with distilled methanol and dried by heating to 110°C for a minimum of 1 h.

In order to study the variation of the concentrations of the analytes in the influent during a 24-h period, composite samples were collected on 18 Mar 2010 every 2 h from one sampling point on the sewerage system of WWTP2 using a refrigerated autosampler. This sample point receives residential wastewater from a population of about 4,000 inhabitants.

2.3 Sample preparation

SPE was carried out using PPL Bond Elute extraction cartridges (1 mL, Varian, Darmstadt, Germany). Extraction volumes were 250 and 500 mL for sewage influent and effluent, respectively. Insoluble components and suspended particulate matter were removed by filtering the samples through a 1-μm borosilicate glass fiber filter (Type A/E, Pall, Dreieich, Germany). The filters were pre-washed with dichloromethane and then heated in an oven for 2 h at 400°C. Before sample extraction, the solid-phase adsorbent was conditioned with 1 mL methanol, 1 mL methanol/acetone (1/1, v/v), and 1 mL distilled water. The samples were introduced to the cartridges by means of PTFE tubes at low pressure (approximately 750 mbar). After sample loading, the SPE cartridges were dried in a nitrogen stream and then eluted three times with 333 μL methanol/acetone (1/1, v/v). The extracts were dried and then dissolved in 400-μL methanol. Subsequently, 2-μg squalane and 2-μg d6-TRA were added to each sample as internal quantification standards.

The samples from the MRF (90% liquid and 10% solid) were first centrifuged at 2,500 rpm for 30 min. The liquid fractions were prepared for analysis using SPE. The solid fraction was freeze-dried using a lyophilization apparatus Christ Alpha I-5 (Martin Christ, Osterode, Germany) for 24 h. Due to the small amount of solid fraction per sample, all fractions were combined and analyzed as one. The solid fraction was extracted three times with 20 mL methanol/dichloromethane (1/1, v/v) using sonication. The sample was immersed in an ultrasound cleaning bath SONOREX SUPER RK 510 H (Bandelin, Berlin, Germany) at a frequency of 35 kHz for 15 min. After the extractions, the sample was centrifuged at 2,500 rpm for 5 min, and then the liquid extracts were combined and evaporated to dryness under vacuum at room temperature. The residue was dissolved in 400-μL methanol; 2-μg squalane and 2-μg d6-TRA were added to each sample as internal quantification standards.

2.4 GC-MS analysis

The samples were analyzed by a Trace GC Ultra gas chromatograph coupled to an ion trap mass spectrometer ITQ 900 (Thermo Scientific, Dreieich, Germany). The gas chromatograph was equipped with a TG-5MS capillary column (30 m length, 0.25 mm inner diameter, 0.25 μm film thickness, Thermo Scientific, Dreieich, Germany). Helium (≥99.999%) was used as the carrier gas at a constant flow of 1.1 mL min−1. The initial column oven temperature (80°C) was increased at 4°C min−1 to a final temperature of 300°C, which was maintained for 30 min. The sample solution (1 μL) was injected using a Tripluss auto sampler in the splitless mode at an injector temperature of 240°C. The mass spectrometer was operated in electron impact mode with 70 eV ionization energy, and full scan spectra were collected in the range from m/z 50 to 650.

Identification of analytes was achieved by mass spectra analysis. The compounds were identified by comparison of their mass spectra with published mass spectra in the library (Wiley Registry/NIST 2008, 8th edn., New Jersey, USA) and by comparing the retention time of each compound with the corresponding reference compounds. Xcalibur Software version 2.0.7 (Thermo Scientific, Dreieich, Germany) was used for data processing. Target compounds were quantified using the internal standard method (Fries 2002), whereas response and correction factors for target compounds were determined at any GC-MS series of measurements. Mass fragments used for identification and quantification of substances are given in Table 3.
Table 3

Main electron impact ionization fragment masses used for identification of the substances and their metabolites


Fragment masses (m/z)


58, 86, 120, 234


77, 91, 106, 120, 121


58, 77, 106,120, 163


58, 77, 135, 263


58, 77, 114, 121, 249


58, 91, 119, 134, 179


58, 77, 91, 120

Italicized fragment masses were used for quantification

2.5 Determination of recoveries

To determine recovery rates over the instrument (GC-MS), 1-L surface water samples (n = 6) were extracted by SPE and the extracts were spiked with the analytes and the internal standards. Extraction procedure by SPE enrichment and measurements by GC-MS were performed as described in Sections 2.3 and 2.4. The recoveries were determined in comparison to a non-enriched standard solution providing values of 86 ± 6% (LDC), 54 ± 5% (2,6-DMA), 64 ± 10% (MEGX), 86 ± 8% (TRA), 95 ± 7% (VEN), 85 ± 10% (ODT), and 95 ± 3% (ODV). The spiked extracts were placed in a fume hood for 24 h at room temperature and then assessed again. Except for 2,6-DMA, no noteworthy differences in the peak areas were found. For 2,6-DMA, the observed peak area in the spiked extracts remeasured was 98 ± 1% lower than the peak area observed in the spiked extracts measured initially. Thus, the metabolite of lidocaine 2,6-DMA appears to be unstable at room temperature and cannot be identified in water samples using this method (because the compound is degraded before analysis).

In order to determine recovery rates over the total method, 1-L surface water samples (n = 6), 1-L treated wastewater samples (n = 6), and 0.2-L untreated wastewater samples (n = 6) were spiked before extraction with the analytes. SPE enrichment and detection by GC-MS were performed as described above. Mean recoveries in surface water were: 70 ± 6% (LDC), 61 ± 3% (MEGX), 73 ± 6% (TRA), 95 ± 8% (VEN), 60 ± 9% (ODT), and 96 ± 3% (ODV); in treated wastewater (n = 6): 65 ± 8% (LDC), 52 ± 8% (MEGX), 62 ± 5% (TRA), 81 ± 3% (VEN), 56 ± 4% (ODT), and 85 ± 3% (ODV); and in untreated wastewater: 56 ± 4% (LDC), 49 ± 9% (MEGX), 61 ± 6% (TRA), 79 ± 6% (VEN), 51 ± 9% (ODT) and 71 ± 11% (ODV).

Recoveries of LDC, TRA, and ODT calculated over the instrument are higher than the recoveries calculated over the total method, indicating that the extraction procedure could be enhanced using another SPE material. For MEGX, the recoveries calculated over the instrument showed no appreciable difference to the recoveries calculated over the total method. Thus, matrix-induced signal suppression appears to be responsible for the loss of MEGX by GC-MS.

Recovery of d6-TRA was calculated over the total method in surface water samples (n = 6) resulting in a value of 55 ± 4%. Because recoveries of the analytes exceeded significantly the recovery of d6-TRA, the use of d6-TRA as surrogate standard would lead to an overestimation. For this reason, this compound was used as internal standard and was added after SPE.

In order to calculate recovery rates of the analytes from solid fractions, 0.5-g freeze-dried activated sludge samples (n = 6) were spiked with the analytes and then stirred to spread over the spiking solution. The spiked samples were placed in a fume hood for 24 h at room temperature. Extraction procedures were performed as described in Section 2.3 for solid fractions. The recoveries were determined in relation to a non-enriched standard solution. Mean recoveries of the analytes in activated sludge were 85 ± 13% (LDC), 60 ± 5% (MEGX), 73 ± 9% (TRA), 83 ± 14% (VEN), 61 ± 7% (ODT), and 76 ± 9% (ODV), at a spiking level of 200 ng g−1.

2.6 Limit of quantification and calibration

Calibration curves for each substance in water samples were prepared by spiking 1-L of ultrapure water (n = 10) with various amounts of stock solutions of target compounds to yield final concentrations between 1 and 2,000 ng L−1. The samples were extracted and analyzed in the same way as field samples. Blank samples were analyzed and the target analytes were not detected. The limit of detection (LOD) of each analyte was calculated in accordance with the German standard method DIN 32645 German Institute for Standardization (1994), with a confidence interval of 99% using the standard deviation of the linear regression curve. The limit of quantification (LOQ) was estimated as three times the LOD and provided 15 (LDC), 48 (MEGX), 26 (TRA), 18 (VEN), 40 (ODT), and 28 ng L−1 (ODV).

For the activated sludge samples, a six-point calibration was used ranging from 1 to 200 ng g−1. LOQs were calculated as three times the LODs in accordance to DIN 32645 (1994) providing values of 26 (LDC), 63 (MEGX), 35 (TRA), 25 (VEN), 61 (ODT), and 39 ng g−1 (ODV).

3 Results and discussion

3.1 Analytes in wastewater

Concentration levels of the analytes found in influent and effluent samples collected between 20 Jan 2010 and 09 Apr 2010 at three WWTPs are summarized in Fig. 1. LDC, TRA, VEN, and the metabolites ODT and ODV were detected in all analyzed samples (n = 236). The metabolites 2,6-DMA and MEGX were not detected in any influent or effluent sample. The experiments carried out for determination of the recovery rates using the standard of 2,6-DMA (metabolite of LDC) showed its instability at room temperature. This suggests that 2,6-DMA might also be unstable in the aqueous solution of the sewerage system after excretion by the human and animal body. The determination of the recovery rate of MEGX has shown that this compound is stable in water. The overall absence of MEGX in sewage water indicates that this compound is possibly released to the sewage water at very low concentrations. The mean concentrations of the detected analytes in influent samples were WWTP1—217 (LDC), 485 (TRA), 331 (ODT), 153 (VEN), and 828 ng L−1 (ODV); WWTP2—96 (LDC), 304 (TRA), 345 (ODT), 99 (VEN), and 506 ng L−1 (ODV); WWTP3—91 (LDC), 239 (TRA), 64 (ODT), 97 (VEN), and 226 ng L−1 (ODV). TRA and the metabolite ODV were present in untreated wastewater at the highest concentrations exceeding 1,000 ng L−1, while the concentrations of LDC and VEN were both significantly lower. These results are consistent with the pharmaceutical sales of LDC and TRA reported in Germany (Table 1). For VEN, although the annual consumption in Germany is unknown, such information may be approximately calculated from data available from other countries. Annual sales of VEN in Canada in 2007 were about 22,000 kg (Metcalfe et al. 2010). Considering that 40% of the world pharmaceutical production is consumed in North America and 32% in Europe (IMS World Review 2008) and with the data of population for North America, Europe, Canada and Germany (DSW German Foundation for World Population 2010), the consumption of VEN in Germany in 2010 was calculated and resulted in 25,000 kg. Despite this value being speculative, it would explain the high concentrations of the metabolite ODV found in untreated wastewater (considering that 29% of an administered dose of VEN is eliminated as ODV).
Fig. 1

a Concentrations (ng L−1) of target analytes in WWTP influents. b Concentrations (ng L−1) of target analytes in WWTP effluents. Sample numbers in parentheses

The measured concentrations of TRA, VEN, and ODV in wastewater from the investigated WWTPs are consistent with other studies conducted in Europe and North America. Lajaunesse et al. (2008) reported the presence of VEN and ODV in untreated wastewater in Canada at similar concentrations. Metcalfe et al. (2010) and Schultz et al. (2010) reported much higher concentrations of VEN and ODV in treated wastewater of WWTPs in North America (>1,000 and >2,500 ng L−1, respectively). The lower concentrations of the compounds in samples from WWTP1, WWTP2 and WWTP3 may be explained by higher consumption of pharmaceuticals in North America, different sewage collection systems or variations in rainfall rates. According to the weather station of the German Meteorological Service at the Frankfurt Airport (DWD German Meteorological Service 2011), irregular mild raining events occurred during the first sample collection (between 20 Jan 2010 and 09 Apr 2010). All of the investigated WWTPs use a sewage collection system through which wastewater and rain water are mixed before treatment. Thus, concentrations of the pharmaceuticals and their metabolites could have been slightly diluted by rainfalls. Hummel et al. (2006) found TRA in both influents and effluents of German WWTPs at concentrations of 1,500 and 610 ng L−1, respectively. However, the data obtained in the present study from WWTP1, WWTP2, and WWTP3 indicate that the demethylated metabolite ODT has also an environmental relevance in wastewater.

Concentrations were generally higher at the WWTP1 influent while WWTP3 had significantly lower concentrations than the other WWTPs. These results may be explained by the potential sources of the respective pharmaceuticals in each WWTP (Table 2). The number of entities associated with the use of pharmaceuticals, e.g., hospitals and laboratories, is greater in the catchment area of the WWTP1 than in the catchment areas of WWTP2 and WWTP3. Thus, the more entities related with the use of pharmaceuticals in the catchment area of a WWTP, the higher the entry of the pharmaceuticals into the sewerage system.

The concentration data and the information about effluent flows have been used to calculate the loads of analytes which are daily discharged into the respective recipient waters. The loads ranged between 0.9 and 8.8 g day−1 for LDC, 1.8 and 18.3 g    day−1 for TRA, 0.6 and 20.9 g    day−1 for ODT, 0.5 and 6.1 g    day−1 for VEN, and 2.2 and 38.7 g    day−1 for ODV. These values confirm WWTPs as a major input source of the investigated analytes and their metabolites into surface waters.

The efficiency of removal of pharmaceuticals in WWTPs depends on many factors such as: type and quantity of wastewater, the wastewater retention time and the treatment technology (Daughton and Ternes 1999; Halling-Sørenson et al. 1998). Figure 2 illustrates the removal rates of the target analytes in the investigated WWTPs, calculated considering the loads in influents and effluents. The data indicate that LDC, TRA, VEN, and the metabolites ODT and ODV are partially removed in the WWTPs. WWTP1 appears to have the highest removal efficiencies (50% (LDC), 50% (TRA), 36% (ODT), 49% (VEN), and 46% (ODV)). Sewage treatment technologies are similar in all investigated WWTPs including preliminary clarification, biological treatment: aerobic (introduction of oxygen) and anoxic (denitrification), chemical phosphate removal and a final sedimentation tank. As mentioned above, WWTP1 had a MFR after the final sedimentation tank. The MRF was operating at the beginning of the sampling from 20 Jan 2010 to 09 Apr 2010 for about 22 h/day. During the second sampling campaign between 16 Sept 2010 and 23 Sept 2010, the MRF was operating only 30 min/day. Higher removal rates have been calculated for samples collected when MRF was operated almost the whole day than from samples when MRF was operated only 30 min/day (37% (LDC), 41% (TRA), 24% (ODT), 48% (VEN), and 29% (ODV)). Analysis of the material recirculated by the MRF showed in the liquid fraction the presence of all the compounds at relevant concentrations (Table 4). Considering that during the two sampling periods there were no significant operational changes in the other cleaning steps carried out in WWTP1, it can be assumed, that the MRF had a positive influence on the removal process of the analytes. The recirculation of material by the MRF to the primary clarifier appears to cause an optimization of the removal process for most of the compounds by the WWTP1. As expected based on their chemical properties (Table 1), the investigated substances were not detected in the solid fraction of wastewater recirculated by the MRF (Table 4).
Fig. 2

Percentage removal of target analytes in WWTPs. Average percentage removal for each WWTP in parentheses

Table 4

Mean concentrations (ng L−1) of the target analytes in the samples of wastewater recirculated by the MRF

MRF samples







Liquid fraction (n = 3)

161 (194)


395 (568)

26 (30)

63 (77)

476 (578)

Solid fractiona







Maximum concentrations are given in parenthesis

aThe 3-day fractions were mixed and analyzed as one

By observing the distribution of compound concentrations in influents (Fig. 1a) and removal rates in each WWTP (Fig. 2), it was quite noticeable that WWTP1 shows higher concentrations in the influent and also higher removal rates. In order to establish a possible relationship between these factors, the concentrations of the analytes in influents were plotted against the calculated removal rates (Fig. 3). It was observed that the higher the concentrations of ODV (metabolite of VEN) in the influent, the higher its removal efficiency in the WWTP. However, this effect was only noticeable when concentrations were relatively high (approximately higher than 1,200 ng L−1). Apparently, increasing concentrations of the compound in the influents (and thus in the aerobic biological step) favors the interaction of the analyte with the activated sludge, and thus the removal efficiency of the metabolite ODV. This effect was not observed for LDC, TRA, VEN, and ODT, which have much lower concentrations.
Fig. 3

Percentage removal of the target analytes related to concentrations (ng L−1) of the analytes in influents in WWTP1

For pharmaceutical compounds such as ibuprofen and naproxen a relationship between the length of time that the wastewater remain in the WWTP (hydraulic retention time (HRT)) and the removal rate of the compound during the process has been reported (Metcalfe et al. 2003; Tauxe-Wuersch et al. 2005). A similar observation is obtained from the present study in case of TRA and VEN (Fig. 4). First-order equations for percentage removalTRA = 17.089e(0.0187 × HRT) (r2 = 0.75) and percentage removalVEN = 20.487e(0.0196 × HRT) (r2 = 0.95) can be used to predict removal rates of TRA and VEN from data on HRTs. The data for LDC and the metabolites ODT and ODV show no trends relating removal rates to HRT.
Fig. 4

Percentage removal of target analytes related to HRT of WWTPs sampled between 20 Jan 2010 and 09 Apr 2010. Trend lines are provided for removal data for TRA and VEN; percentage removalTRA = 17.089e(0.0187 × HRT); percentage removalVEN = 20.487e(0.0195 × HRT)

3.2 Concentration profiles from 24-h sampling

LDC, TRA, VEN, and the metabolite ODV were found in all analyzed samples, while ODT could be quantified only in 83% of the samples and the metabolite MEGX was not detectable in any sample. In the 12 random samples collected on 18 Mar 2010, the investigated compounds were found at the following mean concentrations: 265 (LDC), 752 (TRA), 393 (ODT), 341 (VEN), and 872 ng L−1 (ODV). It is noteworthy that no raining events occurred on the sampling day. Concentrations provided a remarkable variation, but it was not possible to identify any tendency over the 24-h sampling period (Fig. 5).
Fig. 5

Concentrations of the target analytes on the sewerage from WWTP2 during the day profile

3.3 Parent compounds and their metabolites

The calculated ratios between the concentrations of the metabolites and their respective parent compounds TRA and VEN in influents of WWTP1 and WWTP2 (Table 5) are consistent with the data of percentage of dose excreted by humans, which were reported in Table 1. According to Howell et al. (1993), in a medical treatment with VEN, 29% of the applied dose is excreted as the metabolite ODV while only 5% of the dose is excreted as the parent compound. For TRA, 15% of the applied dose is excreted as the metabolite ODT, while the parent compound TRA can be excreted with amounts ranging between 15% and 35% of the applied dose (Kasprzyk-Hordern et al. 2008; Paar et al. 1997). Similar proportions of the parent compounds and the metabolites are observed in the influents of WWTP1 (ODT/TRA = 0.8 and ODV/VEN = 5.8) and WWTP2 (ODT/TRA = 0.9 and ODV/VEN = 5.3). The difference in the ratios found in WWTP3 influent (ODT/TRA = 0.3 and ODV/VEN = 2.9) cannot be explained.
Table 5

Calculated ratios between the concentrations of the metabolites and the parent substances measured in WWTPs influents and effluents


WWTP1 (n = 74)

WWTP2 (n = 66)

WWTP3 (n = 82)








0.8 (0.3–1.4)

1.0 (0.6–1.5)

0.9 (0.6–4.1)

0.9 (0.8–2.9)

0.3 (0.1–0.8)

0.3 (0.1–0.7)


5.8 (1.7–12.3)

6.3 (2.3–10.3)

5.3 (2.0–13.2)

5.9 (1.3–9.3)

2.9 (0.5–6.0)

4.2 (0.5–8.9)

Minimum and maximum ratios are given in parentheses

For each WWTP, the ODT/TRA ratios calculated in influents are similar to the ratios calculated in effluents. The ODV/VEN ratios are slightly higher in the effluents compared to the influents of all three WWTPs. In WWTP1 the ratio increased from 5.8 to 6.3, in WWTP2 from 5.3 to 5.9 and in WWTP3 from 2.9 to 4.2. Given that in all investigated WWTP the removal rates of the metabolites are lower than the removal rates of their respective parent compound, a small increase of the ratios calculated in the effluents in comparison with the ratios calculated in the influents can be expected. Only a very high increase in the ratios calculated in the effluents with respect to the ratios calculated in the influents would imply further degradation of the parent compound in the WWTP, which is not the case for the analyzed pharmaceutical compounds. Thus, ODT and ODV are metabolized only in the liver and are not formed through the cleaning processes carried out at the investigated WWTPs.

4 Conclusions

The results of the study show that LDC, TRA, VEN, and the metabolites ODT and ODV are present in both influents and effluents from three WWTPs in Hesse, Germany. The concentration of analytes in influents differed between the plants, which may be explained by the diverse pharmaceuticals sources in each WWTP. As expected for pharmaceutical compounds, remarkable variations in concentrations in the influent of a WWTP during a 24-h period are found. LDC, TRA, VEN, ODT, and ODV are only partially removed trough cleaning processes at the investigated WWTPs and thus are continuously discharged into the respective recipient waters. The recirculation of treated wastewater to the primary clarifier appears to have a positive effect on the removal process of some analytes. The results indicate that during the treatment of wastewater in the WWTPs both parent compounds and metabolites are degraded. An enrichment of metabolites in relation to the parent compounds TRA and VEN is not observed. 2,6-DMA appears to be unstable in wastewater and could not be identified. MEGX was not detected in any effluent or influent samples. It is probably excreted at very low concentrations and/or degraded rapidly in the sewerage system. Further work is needed to determine the distribution of the investigated compounds and their metabolites in surface water, groundwater, and drinking water.


The first author thanks Deutscher Akademischer Austauschdienst and particularly Katholischer Akademischer Ausländerdienst for financial support. The support of colleagues and staff of the Institute of Atmospheric and Environmental Sciences at the J.W. Goethe University Frankfurt am Main is gratefully acknowledged. Thanks are due to the staff from the three WWTPs for support with sample collection.

Copyright information

© Springer-Verlag 2011