Biodegradability of 14C-Labeled Antibiotics in a Modified Laboratory

Antibiotics in a Modified Laboratory. Scale Sewage Treatment Plant at. Environmentally Relevant. Concentrations. THOMAS JUNKER,* , †. RADKA ALEXY, â...
0 downloads 0 Views 280KB Size
Environ. Sci. Technol. 2006, 40, 318-324

Biodegradability of 14C-Labeled Antibiotics in a Modified Laboratory Scale Sewage Treatment Plant at Environmentally Relevant Concentrations T H O M A S J U N K E R , * ,† R A D K A A L E X Y , ‡ THOMAS KNACKER,† AND KLAUS KU ¨ MMERER‡ ECT Oekotoxikologie GmbH, Boettgerstrasse 2-14, D-65439 Floersheim, Germany, and Institute of Environmental Medicine and Hospital Epidemiology, Hugstetter Strasse 55, D-79102 Freiburg, Germany

For all new pharmaceuticals, an environmental risk assessment (ERA) has to be performed according to guidelines developed by the European Medicines Evaluation Agency. An important factor of this procedure is the assessment of the predicted environmental concentration in the aquatic environment, which is significantly influenced by the biodegradability of pharmaceuticals in sewage treatment plants. Established standardized methods for determining biodegradation under laboratory conditions apply to substance concentrations, which are much higher than those expected in reality. Against this background, the laboratory scale sewage treatment plant (LSSTP), as described by the Organisation for Economic Cooperation and Development (OECD) Guideline No. 303A, was modified to construct a lossless system, which allows laboratory testing at realistic concentrations. To verify the experimental setup, the antibiotics benzylpenicillin, ceftriaxone, and trimethoprim were tested at low concentrations (µg/L) using 14C-labeled compounds. The results show that approximately 25% of benzylpenicillin was mineralized, whereas ceftriaxone and trimethoprim were not mineralized at all. Due to the high total recoveries of added radioactivity (g95%) and the fact that the findings comply with available literature data, the lossless operation of the test system could be proved. Consequently, the modified LSSTP is a suitable tool to determine more realistic biodegradation data required for the exposure assessment within the scope of an ERA for pharmaceuticals.

Introduction Antibiotics rank among the most important pharmaceuticals with regard to the amounts used in human and veterinary medicine. After intake, most of them are only partially metabolized and are excreted in the urine and feces. Accordingly, considerable amounts of antibiotics reach the sewage treatment plants (STPs) (1). For example, in Germany antibiotics from hospitals and households are emitted into * Corresponding author phone: +49-6145-9564-60; fax: +49-61459564-99; e-mail: [email protected]. † ECT Oekotoxikologie GmbH. ‡ Institute of Environmental Medicine and Hospital Epidemiology. 318

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 1, 2006

municipal sewage at a total amount of about 277 t/a (calculated for the year 1998), resulting in a predicted environmental concentration (PEC) of approximately 70 µg/L in municipal STP influents (2). Low eliminable pharmaceuticals may pass through the sewage treatment system and end up in the environment, mainly in the aquatic compartment. Thus, drugs of different classes and some of their metabolites have been detected in municipal STP effluents, rivers, and streams (3-7). Antibiotics have been found in the effluent of pharmaceutical companies, hospitals, and households (8-10), in municipal wastewater (8, 11, 12), in surface water (8, 12-16), and even in groundwater (6, 8, 17). Whereas concentrations of antibiotics between 10 and several hundred ng/L have been detected in groundwater, the concentrations in surface water were in the low µg/L range and antibiotics in municipal sewage and STPs have been detected in concentrations of a few µg/L. Municipal STPs were affirmed to be the most important source for the entry of pharmaceuticals into the aquatic environment (11, 18, 19). Because of these and other alarming findings concerning the occurrence of pharmaceuticals in the environment, environmental risk assessments (ERAs) for new drugs are requested by the European Union (EU) (20, 21). More specific guidance on the ERA procedures for veterinary pharmaceuticals has been developed by the International Cooperation of Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products, which ensures a harmonized approach for Japan, the U.S.A., and the EU (22, 23). For human pharmaceuticals, a revised draft guideline on the ERA has been released in 2005 (24). For both veterinary and human pharmaceuticals, a stepwise procedure is proposed for the ERA. In phase II of these procedures, the biodegradability of the substances, in particular the elimination of human pharmaceuticals via STPs, is of high importance. Ready biodegradability tests have shown that most antibiotics are not readily biodegradable (25-28). In these laboratory tests, nonlabeled compounds were used to study their fate. Indeed, neither the test conditions of standardized ready biodegradability tests nor those of the simulation test, according to OECD 303A (29), allow differentiation between the total mineralization of a compound and abiotic elimination by other mechanisms such as sorption (30). However, as emphasized above, detailed knowledge of the fate of the test compounds and their metabolites formed during (bio)degradation is crucial for a sound ERA; therefore, radioactive (14C-labeled) compounds were used at realistic concentrations in the low µg/L range to simulate the fate of selected antibiotics in a laboratory scale sewage treatment plant (LSSTP). For this purpose a routinely used LSSTP applied in studies, according to OECD Guideline 303A, was modified to construct a lossless system (Figure 1). Thus, a balance of input and output by measuring radioactivity in all compartments of the test system should be possible allowing for the determination of exposure concentrations which are much more realistic than those obtained by existing standardized test methods.

Experimental Section STP Simulation. The performance of the simulation tests was conducted as close as possible to the recommended procedures of test guideline OECD 303A (31). Consequently, based on these recommendations and on the results of preliminary tests, the chosen test period was six weeks. After an acclimation period of three weeks (phase I) the test system was operating efficiently (constant dissolved organic carbon 10.1021/es051321j CCC: $33.50

 2006 American Chemical Society Published on Web 11/17/2005

FIGURE 1. Modified laboratory scale sewage treatment plant (LSSTP) for the testing of the biodegradability of radiolabeled antibiotics. Left, experimental setup; right, schematic presentation: 1 ) nutrient solution inflow, 2 ) tap water inflow, 3 ) aeration vessel, 4 ) settling vessel, 5 ) charcoal absorber, 6 ) collection vessel, F1 ) cold trap (aeration vessel), F2 ) cold trap (settling vessel), F3 ) triethyleneglycol (TEG), F4 ) NaOH 1, F5 ) NaOH 2, F6 ) NaOH 3, P1 ) nutrient solution pump, P2 ) tap water pump, P3 ) sludge recycle pump, P4 ) aeration pump, P5 ) washing bottle pump, R1 ) rotameter (exhaust air aeration vessel), R2 ) rotameter (exhaust air settling vessel), R3 ) rotameter (aeration), R4 ) rotameter (washing bottles). (DOC) elimination >80%) and the radiolabeled antibiotic was added to the test plant continuously over a period of another three weeks (phase II), reaching a steady state of >80% DOC elimination after the addition of the test substance. Figure 1 shows the LSSTP which consisted of an aeration vessel [3] containing the activated sludge and a settling vessel (separator) [4]. The activated sludge was obtained from a municipal sewage treatment plant (Niedernhausen, Germany, 18 500 inhabitant equivalents), which received municipal wastewater without effluents from hospitals. Synthetic sewage, consisting of a nutrient solution, which contained the test substance (1000 mL/d) [1] and tap water (15 800 mL/d) [2], was pumped [P1, P2] into the aeration vessel, resulting in an influent volume of 16 800 mL/d. The dry matter content (concentration of suspended solids) in the aeration vessel was maintained at 2.5 ( 0.5 g/L. The sludge was recycled from the settling vessel to the aeration vessel intermittently at regular intervals by a flexible-tube pump [P3]. Both vessels were closed with sealed lids. The activated sludge was kept in suspension by conducting air into the aeration vessel (∼250 mL/min) through a frit at the bottom of the vessel. Due to the limited capacity of the washing bottles, the air was recirculated by pumping air of the headspace of the aeration and settling vessel back to the aeration vessel to facilitate suspension of the activated sludge [P4]. The respective air volumes were controlled by valves and measured by rotameters [R1, R2, R3]. To prevent the accumulation of condensation water in the valves and in the rotameters, two cold traps [F1, F2] were placed into the air streams. The exhaust air from the settling vessel was replaced by ambient air through the overflow pipe due to negative pressure, whereby the oxygen supply was ensured.

To trap 14CO2 and other volatile radioactive components, the air of the headspace of the aeration vessel was pumped [P5] through four washing bottles filled with triethyleneglycol [F3] and 1.5 N NaOH [F4, F5, F6]. The air volume was controlled by a flexible-tube clip and measured by rotameter R4. The uncontaminated air was then emitted into the ambient air. All hoses used in the system, except for those used in the pumps, were made of polytetrafluoroethylene to prevent adsorption of the test substance. The hoses used in the pumps needed to be flexible and were made of silicone. The effluent (about 16 800 mL/d, corresponding to a mean hydraulic retention time of 6 h) was cleared from the radioactive components with a charcoal absorber and was controlled by liquid scintillation counting before disposal. The experimental setup allowed for the assessment of the degree of complete mineralization to 14CO2 as well as for the distribution of either the parent compound or the transformation products or both between the compartments of the test system. For this purpose, samples of the influent, the activated sludge, the effluent, and the exhaust air were taken and analyzed using a liquid scintillation counter (TriCarb 2500 TR, Packard, Dreieich, Germany). Based on the results, a 14C-balance was set up five times during the radioactive test period (phase II). The recovery was calculated by comparison of the overall radioactivity added to the LSSTP to the overall radioactivity measured in the specimens. Total elimination was also monitored by measuring the DOC. Test Compounds. On the basis of consumption, toxicity, evidence in the environment, and other parameters, 18 antibiotics were selected from 243 compounds (31). The antibiotics benzylpenicillin (penicillin G), ceftriaxone, and trimethoprim were available as 14C-labeled compounds. 14CBenzylpenicillin was obtained from Amersham Pharmacia VOL. 40, NO. 1, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

319

TABLE 1. Test Compounds and Their Concentrations test compound group

14C-benzylpenicillin

14C-ceftriaxone

14C-trimethoprim

penicillins

cephalosporins

molecular formula structure

C16H17N2O4SK (free acid)

C18H18N8O7S3

other antibiotics (diaminopyrimidine) C14H18N4O3

solvent radiochemical purity specific radioactivity conc in the influent activity in the influent total conc in the influent (including nonlabeled test substance)

ethyl alcohol 96.8% 5.74 MBq/mg 1.45 µg/L (5.2% of total conc) ∼8300 Bq/L 28.00 µg/La

methyl alcohol (90%) >95.0% 1.89 MBq/mg 4.40 µg/L (31.4% of total conc) 8300 Bq/L 14.00 µg/La

ethyl alcohol (80%) 97.0% 1.48 MBq/mg 5.60 µg/L (18.7% of total conc) ∼8300 Bq/L 30.00 µg/La

a

Corresponding to 20-fold PEC in order to simulate peak loads (2).

Biotech Europe GmbH, Freiburg, Germany; and 14C-ceftriaxone and 14C-trimethoprim were provided by HoffmannLa Roche AG, Basel, Switzerland. Aliquots of the respective alcoholic stock solution were applied to the sterilized (autoclaved at 121 °C for 30 min) nutrient solution through a septum using a syringe. To simulate a peak load and to cover worst-case scenarios in municipal sewage treatment plants, the PECs of the antibiotics estimated by Ku ¨ mmerer und Henninger (2) were multiplied by a factor of 20. To achieve these concentrations in the influent of the LSSTP, additional unlabeled test substance was added to the influent, resulting in total concentrations of 28, 14, and 30 µg/L for benzylpenicillin, ceftriaxone, and trimethoprim, respectively (Table 1). DOC, Sludge Parameters, Dry Matter, Nitrogen Parameters, and Physicochemical Parameters. The DOC elimination was determined according to the test guideline (29). Analyses were performed with a carbon analyzer (TOCOR-2, Maihak, Germany) in order to monitor DOC elimination during the course of the test. For this purpose, samples of the influent and effluent were taken three times a week. Prior to measurement, samples were acidified and diluted, if necessary. DOC samples were stored in a refrigerator until analysis. Sludge index and dry matter content were also determined several times a week according to German standard methods for the examination of water, wastewater, and sludge (32, 33). In addition, the sludge retention time (SRT) was determined. According to Stier and Fischer (34), a SRT of >10 days is required to ensure nitrification. To control nitrification, the nitrogen parameters ammonium, nitrite, nitrate, and total nitrogen were determined regularly using Spectroquant testkits (Merck, Darmstadt, Germany) and a photometer. The oxygen concentration, temperature, and pH were monitored at least once daily. Compound Specific Analysis. Samples of the influent, activated sludge, and the effluent were taken for compound specific analyses of trimethoprim, performed at the Institute of Environmental Medicine and Hospital Epidemiology of the University Hospital Freiburg, Germany, to determine the amount of parent compound and potential metabolites in the compartments of the LSSTP. The samples were stored deep frozen until analysis and were cleaned up and preconcentrated by solid-phase extraction. Analysis was performed by HPLC (Shimadzu, Duisburg, Germany; Autosampler SIL-10AD vp, column oven CTO-10AC, pump LC-10AT, UV-Vis detector SPD-10A vp). 320

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 1, 2006

Solid-Phase Extraction. The samples were adjusted to pH 5 and ethylenediaminetetraacetic acid (EDTA)was added in a concentration of 7 mmol/L. One hundred milliliters of the sample was concentrated on a Chromabond-C18 LV (500 mg) (Macherey & Nagel, Du ¨ ren, Germany). Trimethoprim was eluted with 2 × 600 µL of methanol. The samples were dried with a gentle N2 stream and, afterward, diluted in 1 mL of distilled water. HPLC. We injected 150 µL of the aqueous sample into the HPLC. A gradient with 0.085% phosphoric acid and acetonitrile was used. It rose from 2% acetonitrile to 65% over 30 min. The last 10 min were performed with 2% acetonitrile (isocratic) for re-equilibration of the column. The analytical column was a CC 250/4 Nucleodur 100-5 C18 (Macherey & Nagel), the guard column was a CC 8/4 Nucleosil 100-5 C18 (Macherey & Nagel). Quantification was done by calibration with an external trimethoprim standard (cotrim forte ratiopharm ampule) following the same sample pretreatment (peak area of the UV absorption at 210 nm). The limit of quantification was 10 µg/L and the mean recovery was about 90%. The experimental set-up is summarized in Table 2.

Results and Discussion Operating Parameters. During the simulation tests the temperature in the aeration vessel was maintained within the recommended range. Due to a very high biomass concentration, the oxygen concentration in the aeration vessel fell below the limit of 2 mg/L (single values down to 0.7 mg/L) temporarily. The pH ranged between 7.1 and 7.7 throughout all test periods. The mean dry matter content in phase II was 2.7 ( 0.1, 2.6 ( 0.3, and 2.6 ( 0.2 g/L for benzylpenicillin, ceftriaxone, and trimethoprim, respectively. The sludge index was 90 ( 8, 84 ( 11, and 118 ( 26 mL/g, respectively. No bulking sludge was observed within the radioactive test phases. The mean sludge loading was 0.10 ( 0.01 mg/(mg*d) for benzylpenicillin and ceftriaxone and 0.11 ( 0.01 mg/(mg*d) for trimethoprim. The SRT was >10 days throughout all test periods. During all radioactive test periods, increased concentrations of ammonium were measured and the limit of 2 mg/L); sludge volume index ( 10 d); dry matter content (2.5 ( 0.5 g/L); DOC elimination (>80%); ammonium (