Speciation and Isotope Dilution Analysis of ... - ACS Publications

Article pubs.acs.org/est

Speciation and Isotope Dilution Analysis of Gadolinium-Based Contrast Agents in Wastewater Lena Telgmann,† Christoph A. Wehe,† Marvin Birka,† Jens Künnemeyer,† Sascha Nowak,‡ Michael Sperling,†,§ and Uwe Karst†,* †

Institute of Inorganic and Analytical Chemistry, University of Münster, Corrensstraße 30, 48149 Münster, Germany Institute of Physical Chemistry, University of Münster, Münster Electrochemical Energy Technology (MEET), Corrensstraße 46, 48149 Münster, Germany § European Virtual Institute for Speciation Analysis (EVISA), Mendelstraße 11, 48149 Münster, Germany ‡

S Supporting Information *

ABSTRACT: The fate of Gadolinium (Gd)-based contrast agents for magnetic resonance imaging (MRI) during sewage treatment was investigated. The total concentration of Gd in influent and effluent 2 and 24 h composite samples was determined by means of isotope dilution analysis. The balancing of Gd input and output of a sewage plant over seven days indicated that approximately 10% of the Gd is removed during treatment. Batch experiments simulating the aeration tank of a sewage treatment plant confirmed the Gd complex removal during activated sludge treatment. For speciation analysis of the Gd complexes in wastewater samples, high performance liquid chromatography (HPLC) was hyphenated to inductively coupled plasma sector field mass spectrometry (ICP-SFMS). Separation of the five predominantly used contrast agents was carried out on a new hydrophilic interaction liquid chromatography stationary phase in less than 15 min. A limit of detection (LOD) of 0.13 μg/L and a limit of quantification of 0.43 μg/L could be achieved for the Gd chelates without having to apply enrichment techniques. Speciation analysis of the 24 h composite samples revealed that 80% of the Gd complexes are present as Gd-BTDO3A in the sampled treatment plant. The day-of-week dependent variation of the complex load followed the variation of the total Gd load, indicating a similar behavior. The analysis of sewage sludge did not prove the presence of anthropogenic Gd. However, in the effluent of the chamber filter press, which was used for sludge dewatering, two of the contrast agents and three other unknown Gd species were observed. This indicates that species transformation took place during anaerobic sludge treatment.

■

1996, Bau et al. observed a Gd anomaly in German rivers.3 These immense accumulations of anthropogenic Gd were also found in similar studies in Australia, France, the U.S. and Japan.4−10 Interestingly, no anomaly of Gd could be determined in rivers in, for example, North Sweden.3 These rivers had only little or no contact to populated areas. The source of the anthropogenic Gd in highly populated regions was quickly discovered to be the increasing application of MRI contrast agents. This suggests that the treatment in sewage plants is not very effective for the removal of the Gd complexes. It is assumed that they pass the sewage treatment plant unhindered because of the high water solubility of the polar or ionic complexes and their high complex stability. However,

INTRODUCTION

Magnetic resonance imaging (MRI) is an indispensible method for clinical diagnostic used for imaging. The examination of organs often requires the signal enhancement by contrast agents. The structures of the five most often employed Gdbased contrast agents gadopentetate (Gd-DTPA), gadodiamide (Gd-DTPA-BMA), gadobenate (Gd-BOPTA), gadoterate (GdDOTA), and gadobutrol (Gd-BT-DO3A) are shown in Figure 1. Today, almost every second MRI examination worldwide is enhanced by application of these agents.1 The complexes are renally excreted from the patient’s body unmetabolized, quantitatively, and fast via the urine. Depending on the number and activity of clinical centers performing MRI examinations within the catchment area of a sewage treatment plant, the input of Gd into the public sewer is therefore very high. Kümmerer et al. calculated the Gd emission by hospitals and radiology practices in Germany to exceed 1100 kg annually.2 In © 2012 American Chemical Society

Received: Revised: Accepted: Published: 11929

May 24, 2012 August 31, 2012 October 12, 2012 October 12, 2012 dx.doi.org/10.1021/es301981z | Environ. Sci. Technol. 2012, 46, 11929−11936

Environmental Science & Technology

Article

Figure 1. The structures of the five most commonly applied Gd-based MRI contrast agents and the trademarks of the agents.

passing through a concentrated microbial biocenosis including highly reactive organic and inorganic matter might lead to a species transformation. Only a few studies examined the behavior of Gd complexes used as MRI contrast agents with regard of their behavior during sewage treatment and their fate after release into the environment. A detailed study about the behavior of Gd during common sewage treatment was carried out by Verplanck et al. in 2010. Gd concentrations were determined with inductively coupled plasma-mass spectrometry (ICP-MS) in a wide range of samples, including sludge and wastewater in several stages of treatment. An enrichment of Gd was found in water samples, but not in sludge samples.11 In 2010, Lawrence et al. investigated the behavior of Gd complexes in an advanced wastewater treatment plant. The particular plant included microfiltration, reverse osmosis and oxidation with UV irradiation and peroxide. Results showed that Gd is efficiently rejected during reverse osmosis. However, Gd could still be detected in the reverse osmosis permeate, but only in concentrations near the detection limit.12 Speciation of Gd-based MRI contrast agents in wastewater was carried out in 2009 by Künnemeyer et al.13 The study showed for the first time that Gd was at least partly present in the effluent as the Gd complexes used in the clinical facilities. A species transformation could not be proven. Nevertheless, the total amount of Gd in the effluent was higher than the amount of complexed Gd, indicating that not all Gd is present as the known complexes. A transformation of the complexes could therefore not be excluded. In this study, we investigate the fate of the Gd-based MRI contrast agents after the emission into the sewage treatment plant. Isotope dilution analysis and complementary speciation with LC hyphenated to ICP sector field (SF) MS are employed to investigate total and complexed Gd concentrations in influent and effluent samples. The week-day dependent variation of Gd input and output of the treatment plant observed over a period of seven days provides information about the removal capacity of the treatment. Additional simulation of the activated sludge tank and the examination of sewage sludge are carried out to examine the removal capacity of biological wastewater treatment. Further investigations of the behavior of Gd compounds during sewage

sludge treatment are accomplished by means of ion chromatography (IC) ICP-MS.

■

MATERIALS AND METHODS Chemicals and Consumables. Nitric acid (65%, Suprapur), gadolinium standard (1000 mg/L) for ICP-MS, thallium standard (1000 mg/L) for ICP-MS and acetonitrile for HPLC were obtained from Merck KGaA (Darmstadt, Germany). REE multielement standard for ICP-MS (10 μg/L) was purchased from High Purity Standards (Charleston, SC). 158Gd enriched Gd2O3 (98.35% 158Gd) was obtained from Isoflex (San Francisco, CA). Formic acid, diethylenetriaminepentaacetic acid (DTPA) ethylenediamine and ammonium formate were obtained from Fluka Chemie (Buchs, Switzerland). Nitrilotriacetic acid (NTA) and ethylenediaminetetraacetic acid (EDTA) were purchased from Sigma-Aldrich (Steinheim, Germany). 2-Hydroxyisobutyric acid was obtained from Alfa Aesar (Karlsruhe, Germany). Contrast agent infusion solutions were provided by the respective producing pharmaceutical companies: Magnevist (Gd-DTPA, 0.5 mol/L) and Gadovist (Gd-BT-DO3A, 1.0 mol/L) by Bayer-Schering Pharma AG (Berlin, Germany), Omniscan (Gd-DTPA-BMA, 0.5 mol/L) by GE Healthcare Buchler (Braunschweig, Germany), Dotarem (Gd-DOTA, 0.5 mol/L) by Guerbet (Sulzbach, Germany) and Multihance (GdBOPTA, 0.5 mol/L) by Nycomed (Konstanz, Germany). All chemicals were used in the highest quality available. Water was purified through an Aquatron Water Stills purification system model A4000D (Barloworld Scientific, Nemours Cedex, France). Sampling. Wastewater samples were taken during one random week in fall in 2010 during dry and rainy weather in Münster (Germany). The city sewage treatment plant of Münster is collecting the wastewater of about 300 000 inhabitants receiving a volume of 50 000 m3 of water per day. About 1000 m3 of waste sludge are withdrawn daily. The sampling of 2 h composite samples for the balancing of Gd input and output of the sewage treatment plant was carried out by a stationary water sampler ASP Station 2000 (Endress + Hauser, Weil am Rhein, Germany). The sampler for the influent samples was located at the inlet of the wastewater treatment plant. The sampler for the effluent samples was 11930

dx.doi.org/10.1021/es301981z | Environ. Sci. Technol. 2012, 46, 11929−11936


Article

Heinsberg, Germany). At the start of the experiment, Magnevist was spiked into the water to reach a final concentration of about 1.6 μg/L. The first sample was drawn 15 min after spiking to ensure complete mixing of the vessel content. The last sample was taken after 24 h. All samples were filtered as described above. Sewage Sludge. The dehydrated sewage sludge was dried directly after sampling for 24 h at 100 °C. It was then pulverized using a mortar to ensure homogeneity and to ease microwave digestion. About 0.05 g of dried sludge were weighed into a PP microwave vessel. Afterwards, 4 mL of nitric acid and 1 mL hydrogen peroxide were added. After a heating ramp at 400 W from room temperature to 125 °C over 5 min, the temperature was kept constant for 1 min. Then, over 5 min at 800 W, the temperature was increased to 150 °C. The temperature was kept constant for 15 min and then heated to 170 °C over 5 min at 1600 W. The temperature was kept constant for 5 min before the samples were cooled down to room temperature. The dissolved sample was diluted with water. In order to perform the standard addition method, internal standards of the REEs were added to aliquots of the samples to reach the following spike concentrations: 0, 25, 50, 150, 350 ng/L. All samples were prepared in PP tubes. Chamber Filter Press. For LC/ICP-SFMS measurements, the effluent of the chamber filter press was filtered like the wastewater samples. For LC/ESI-MS measurements, 500 μL of the filtrated sample were preconcentrated to ca. 50 μL by evaporation at 120 °C for two hours. For stability studies, 10 μL of a 157.3 μg/L DTPA solution were added to 90 μL of the original sample prior to analysis and the sample was incubated for 24 h. Gd complexes of EDTA and NTA. The respective ligand was diluted with water to reach a concentration of ca. 1 mmol/L. Gd acetate was added to reach a concentration of 0.1 mmol/L. The solution was incubated at 37 °C for 24 h. Prior to analysis, the solution was diluted to cGd = 1 μmol/L (LC/ESI-MS) and cGd = 10 nmol/L (IC/ICP-MS), respectively. Instrumentation. Isotope dilution analysis and standard addition measurements were carried out on an iCAP Qc ICPMS (Thermo Fisher Scientific, Bremen, Germany). The ICPMS was coupled to a sample changer SC4-SL (Elemental Scientific, Omaha, NE, USA). Detailed ICP-MS conditions are displayed in Table S-1 in the Supporting Information (SI) for isotope dilution analysis and Table S-2 in the SI for standard addition experiments. All LC separations were carried out using a Shimadzu LC system (Duisburg, Germany). The system consisted of two LC10ADVP pumps, a DGC-14A degasser, a SIL-HTVP autosampler and a CTO-10AVP column oven. Separation of the Gd complexes was carried out on a solid core HILIC stationary phase (Accucore, Thermo Fisher Scientific). Detailed separation conditions are displayed in Table S-3 in the SI. Exemplary LC/ICP-MS measurements of wastewater samples were also carried out on a zwitterionic HILIC stationary phase (zic-HILIC, Merck Chemicals, Darmstadt, Germany). Detailed separation conditions are displayed in Table S-4 in the SI. LC/ESI-ToF-MS measurements to optimize the LC separation were performed on a time-of-flight (ToF) mass spectrometer (micrOTOF, Bruker Daltonics, Bremen, Germany). Detailed ESI-ToF-MS conditions are displayed in Table S-5 in the SI.

located directly before the outlet of the wastewater treatment plant, after the confluent water of all secondary clarifier basins. The stationary water sampler withdrew a volume of 20.8 mL wastewater every five minutes. The sample container is exchanged every two hours. A schematic overview of the sampled sewage treatment plant is displayed in Figure 2.

Figure 2. An overview of the sampled sewage treatment plant.

Mixed liquor of wastewater and activated sludge was collected from the aeration tank for the batch experiment on a weekday in the afternoon. Sewage sludge was taken as a grab sample after dehydration with the chamber filter press. A water sample was taken as a grab sample directly from the effluent of the press. The samples were taken on a weekday in the morning. All sample containers were made of polypropylene (PP) to avoid adsorption effects. Sample Preparation. Mixing of 24 h Composite Samples. The 24 h composite samples were created from 2 h composite samples by using an aliquot of the 2 h samples proportional to the water flow at collection time. This procedure leads to a more accurate display of analyte concentration than a direct sampling of a time-dependent 24 h composite sample. Directly after mixing, samples were frozen at −30 °C. Filtration. All wastewater samples for all studies were defrosted and filtered before analysis. The wastewater was filtered through 0.2 μm cellulose acetate syringe filters, which were preconditioned with 2 mL of the respective sample. Isotope Dilution Analysis. About 300 mg of 158Gd enriched Gd2O3 were dissolved in 100 mL nitric acid (20%). This stock solution was then diluted to a concentration of 10 nmol/L. 3 mL of the filtered wastewater sample were filled into a PP tube. 100 μL of nitric acid and 100 μL of a 50 nmol/L dilution of Tl were added. 100 μL of the 158Gd enriched stock solution were subsequently filled into the tube. The sample was diluted by water to reach a final volume of 8 mL. To determine the precision of the method, the concentration of one sample was determined 10 times. Sample preparation was performed individually and in the same way as the other wastewater samples. The conditions and settings of the ICP-MS measurements were also the same. Speciation with LC/ICP-SFMS. The filtered water samples were filled into PP vials. PP was used as vial material to avoid the adsorption of the polar analytes on the vessel surface. Batch Experiment. 2 L of mixed liquor of wastewater and activated sludge from the aeration tank were filled into a PP vessel. Thorough and constant mixing of the mixture was achieved by stirring with a magnetic stirrer. The aeration of the simulated tank was accomplished by the insertion of an air pump designed for aquariums (sera air 110 pump, sera, 11931



Article

evening and at night are caused by administration of the Gd complexes for examinations during all day. The load of Gd in the effluent sample ranges between 1.04 and 3.19 g per two hours. The high variations of Gd in the samples of the influent are equalized by the mixing of the wastewater in the large tanks, recirculations and the irregular fluxes from the sewage treatment. For the balancing of Gd input and output of a sewage treatment plant and a display of the week-day dependent variations, 2 h composite samples were taken over a period of seven days. The 2 h composite samples were mixed to 24 h composite samples. The Gd concentrations determined by isotope dilution analysis and the respective Gd load are displayed in Figure 4. Details about Gd concentration in each

For LC/ICP-SFMS measurements, the LC was coupled to an ElementXR (Thermo Fisher Scientific). Detailed conditions of the hyphenation and detection can be found in Table S-6 in the SI. LC/ESI-MS measurements of the chamber filter press sample and the Gd complexes of EDTA and NTA were carried out on an Orbitrap ESI-MS (Exactive, Thermo Fisher Scientific). Detailed parameters and conditions are displayed in Table S-7 in the SI. Ion chromatography of the effluent of the chamber filter press was carried out on a cation chromatography system 881 Compact IC pro (Metrohm, Filderstadt, Germany) with a cation exchange chromatography column (Nucleosil 5 SA− 125/4.0, Metrohm). Detailed separation conditions are displayed in Table S-8 in the SI. For IC/ICP-MS measurements, the IC was coupled to an ICP-MS with an octopole reaction system (ORS3) (7700 Series, Agilent, Santa Clara, USA). Detailed conditions of the hyphenation and detection can be found in Table S-9 in the SI.

■

RESULTS AND DISCUSSION Total Gd Concentration in Influent and Effluent Samples. The total concentration of Gd in the influent and effluent of a common sewage treatment plant was determined by isotope dilution analysis. The load of Gd in 2 h composite samples of a Wednesday is shown in Figure 3. Details about Gd

Figure 4. Gd load in 24 h composite samples, determined by isotope dilution analysis.

sample and respective volumetric flow rate of the wastewater are displayed in Table S-11 in the SI. The amount of Gd in the influent samples ranges in between 2 and 58 g per day. The lower load on Saturday and Sunday can be explained by the fact that radiology practices are commonly closed on the weekends. In the hospitals, only emergency MRI examinations are carried out. Since the excretion of the contrast agents is fast, most of the Gd is excreted into the public sewer during the first hours after the examination. Therefore, only low Gd concentrations can be detected in the influent samples on Saturday and even less in the samples on Sunday. The amount of Gd that leaves the sewage treatment plant ranges from 12 to 49 g per day. The day-dependent Gd concentration is equalized in the effluent samples. However, the lower amounts of Gd in the effluent on Sunday and Monday, which result from the lower input on the weekend, represent the high retention time in the treatment plant, which is up to 24 h depending on the volumetric flow rate. These results confirm prior investigations that stated that most of the Gd input entering the wastewater treatment plant is not removed from the aqueous phase during sewage treatment, but enters the environment to a high degree leading to very high load of anthropogenic Gd in surface water. The balancing of input and output of total Gd indicates that about 90% of the Gd input is delivered into the environment via the sewage treatment plant effluent, whereas 10% are removed during sewage treatment. The low removal capacity of biological wastewater treatment for Gd-based MRI contrast agents is consistent with the high values of anthropogenic Gd in surface

Figure 3. Gd load in 2 h composite samples, determined by isotope dilution analysis.

concentrations in samples and respective volumetric flow rate of the wastewater are displayed in Table S-10 in the SI. The total input and output of Gd was calculated with consideration of the water flow during the respective sampling time. Each bar represents the load of total Gd entering the sewage treatment plant (dark gray) and exiting the sewage treatment plant (light gray). The load of Gd in the influent samples ranges between 0.40 and 9.33 g per two hours. This correlates with the application and excretion of the contrast agents by MRI patients. Usually, in the first 1−5 h after administration, the main fraction of the contrast agents is excreted. Note that the retention time of wastewater within the sewer network is about 2−4 h. In the morning between 8 and 10 a.m., the Gd load is high due to the excretion of MRI contrast agents by patients who were administered contrast agents the day before. At midday, Gd is eliminated from the body of patients that underwent MRI in the early morning. High Gd loads in the 11932



Article

creek several kilometers downstream of the effluent of the sewage treatment plant. The results of this experiment, including concentrations and normalized data, are displayed in SI Table S-12. The concentration of GdNat was extrapolated by comparison of normalized Sm and Tb data according to Bau et al.3 GdNat was determined to be 7.8 ng/L, which represents about one hundredth of the Gd concentration in the wastewater samples. Therefore, the fraction of natural Gd is neglected in our considerations and assessments. Speciation of Gd Complexes in Wastewater Influent and Effluent. The separation of the five predominantly used contrast agents was performed on a solid core HILIC stationary phase in less than 15 min. Figure S-1 in the SI shows the chromatogram of the extracted ion traces of all contrast agents detected by ESI-ToF-MS. The signal of Gd-BOPTA shows two peaks. This behavior has already been observed during measurements with a zwitterionic HILIC stationary phase11 and can likely be explained with the asymmetric structure of the ligand. For LC/ICP-SFMS measurements, the LOD was determined by means of a signal-to-noise ratio of three (S/N = 3) and the LOQ by means of a signal-to-noise ratio of 10 (S/N = 10). The LOD was as low as 0.13 μg Gd complex/L and the LOQ as low as 0.43 μg Gd complex/L for all Gd complexes. Linearity was determined by triplicate measurements in a concentration range of 0.13−15.75 μg/L. This method offers an improvement to formerly published methods; the shorter LC separation is important for the analysis of a vast number of samples and the improved detection limit allows the identification of the Gd complexes in very low concentration. The chromatogram of one exemplary influent wastewater sample is shown in Figure 5. In the influent and effluent

waters that was found all over the world. High Gd concentrations were found in wastewater effluent in the past by Bau,3 Möller,10 Lawrence,5,12 and others. The determination of the precision of the isotope dilution analysis was carried out by measurement of one sample that was prepared ten times separately. The relative standard deviation of the measurements was calculated to be 1.9%. Additionally to the analytical error, the sampling procedure is also an error source. The error of the sampling of composite samples with a stationary water sampler is the higher the more unsteady the water flow during the sampling period. According to the manufacturer, this can be in the range of

Speciation and Isotope Dilution Analysis of ... - ACS Publications

Recommend Documents