Sediment Systems for

Pharmacokinetics, Dynamics and Metabolism, Environmental Sciences, .... The radio-purities of the radiolabel test materials for the 3 pharmaceuticals ...
0 downloads 0 Views 215KB Size
Environ. Sci. Technol. 2007, 41, 5803-5811

An Evaluation of the OECD 308 Water/Sediment Systems for Investigating the Biodegradation of Pharmaceuticals JON F. ERICSON* Pharmacokinetics, Dynamics and Metabolism, Environmental Sciences, Pfizer Global Research & Development, MS 8118A-2026, Groton, Connecticut 06340

In recent times, trace levels of pharmaceuticals detected in wastewater effluents and surface waters have raised the level of attention around the ultimate fate and the potential persistence of pharmaceuticals in the environment. We have seen the European Agency for the Evaluation of Medicines (EMEA) recently include more rigorous environmental fate testing in European Union (EU) Environmental Risk Assessment (ERA) guidance to assess the ultimate fate in water/sediment systems. Yet to date, there is little data available that covers the fate of pharmaceuticals in the water/sediment compartment, and little that assess whether current aerobic and anaerobic methods are appropriate for pharmaceuticals. In this study, the biodegradation profiles of 3 Pfizer products were investigated following the latest ERA guidance and its recommendation for OECD 308 water/sediment biodegradation testing. Experiments included 14C-labeled exemestane, azithromycin, and varenicline representing neutral and cationic pharmaceuticals with average Koc values of 3704, 49 400, and 10 483 respectively. Specific HPLC/radioactive monitoring (RAM) methods were used to profile water and sediment samples for biotransformation products. Binding to sediment, as “non-extractables”, was considerable for all three pharmaceuticals, though most notable for the cationic pharmaceuticals varenicline and azithromycin ranging from 52% to 94% at study termination, respectively. In general, for all 3 pharmaceuticals studied, the anaerobic conditions demonstrated less biotransformation and mineralization than the aerobic; though their biotransformation profile (number of metabolites) and amount bound to sediment were similar. Based on these findings and our current understanding of anaerobic biodegradation, we would recommend a tiered approach to the OECD 308 water/ sediment test: with default testing just for aerobic conditions; and then if needed, anaerobic testing only for those compounds potentially amenable to typical anaerobic processes. We suggest that as a simulation test would be better suited in later tier testing under EU ERA guidance. Inherent biodegradation or die-away tests seem better suited to derive biodegradation rate constants for subsequent environmental modeling of water and sediment compartments.

Introduction The fate of pharmaceuticals in the environment has come under more scrutiny in the advent of their detection in the environment in the late 1990s (1-5). Pharmaceuticals are * Phone: 860-441-4864; e-mail: [email protected]. 10.1021/es063043+ CCC: $37.00 Published on Web 07/14/2007

 2007 American Chemical Society

predominately released into the environment through wastewater treatment plants as a result of consumer use, either on wasted sludge or in wastewater effluents (6). The release is continuous, ultimately achieving steady-state levels in the environment at trace (ng-µg/L) concentrations, after the mitigating result of many subsequent fate processes that include sorption, biotransformation, mineralization, and photolysis to name a few (7). Studies within the last 5-10 years have studied the presence of pharmaceuticals beginning at their point of origin, the wastewater treatment plant (810), downstream to surface waters where they may partition into suspended particles and sediments (11-12), and in some cases, for the more mobile anionic pharmaceuticals (13), transport further into ground and potential drinking water sources (14). With this increased level of detection in the environment and potential uncertainty of the persistence of pharmaceuticals, we have seen additional ERA test requirements (15) as part of the EU registration process for human medicines. As a result of this increased focus on environmental fate, we are seeing more results on biodegradation testing of pharmaceuticals (16). It is with these efforts that we are begining to understand the importance of nonextractable residues and biotransformation in the environment and to establish a knowledge base of the “typical” ranges for elimination from the water and sediment compartments in the environment. This study builds upon the current published work by providing environmental fate data on 3 Pfizer products (Table 1). The objective was not only to satisfy our regulatory obligations for these drug candidates, but more importantly to (1) publish the environmental fate results on pharmaceuticals in an area that historically has not been well presented, and (2) provide recommendations as to how better implement the OECD 308 (17) guideline for assessing the aerobic and anaerobic biodegradation of pharmaceuticals in the context of the EU ERA testing guidance. The learning to date would suggest the OECD 308 may not be best suited in Tier A screening tests for the EU ERA, but better suited for subsequent tiers in extended ERA testing when further refinement is needed in the risk assessment, or when more information is needed regarding the disposition and depletion of pharmaceuticals in water and sediment systems. The OECD 308 simulates water and sediment depletion rates, overall disposition of 14C residues, mass balance, and characterization of any biodegradation products. Other methods recently proposed, such as the OECD “3xx” Guideline, July 2006, “Simulation Tests to Assess the Primary and Ultimate Biodegradability of Chemicals Discharged to Wastewaters” may be more appropriate for screening in Tier A fate testing.

Materials and Methods Experimental Setup. The study was performed following the OECD 308 guideline. Each test material was tested on 2 types of water/sediment systems (high organic content/fine texture; low organic content/coarse texture) under aerobic and anaerobic conditions. Each type of water/sediment system was set up in duplicate. Collection and Characterization of Sediments and Waters. One aerobic and one anaerobic sediment sample, each with its associated water, were collected from Turkey Creek (Talbot County, Maryland) and Choptank Rivers (Denton, Maryland) respectively. The Turkey Creek sediment possessed a high organic carbon content (∼7.8%), a fine texture (∼34/34/32 sand/silt/clay), and cation exchange capacity of 17.1-20 meq/100 g while the Choptank River VOL. 41, NO. 16, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5803

TABLE 1. Test Material Specifications and Structures

sediment had a low organic carbon content (0.2-1.5%), a coarse texture (∼96/2/2 sand/silt/clay), and cation exchange capacity of 2.1-6.3 meq/100 g; consistent with types of water/ sediment systems recommended in the study guideline. Aerobic and anaerobic sediments were sampled from the top layer of sediment and sieved prior to use with a 2 m screen following the OECD 308 guideline. Associated waters were taken from the same location at the time of sediment collection. The aerobic sediments and associated waters were stored at approximately 4 °C with free access of air. The anaerobic sediments were sampled to exclude oxygen and stored at approximately 4 °C. All handling of anaerobic sediment after collection and prior to testing was performed under the exclusion of oxygen utilizing a blanket of nitrogen gas. Sediment texture based on particle size (i.e., percentage of sand, silt, and clay), water holding capacity, cation exchange capacity, percentage of organic matter, bulk density, USDA textural class, pH, percent organic carbon, and concentrations of calcium, magnesium, sodium, and potassium were determined by Agvise Laboratories, Inc. (Northwood, ND) prior to initiation of the study. Additionally, the percent moisture and microbial biomass of the four sediments were determined. The microbial biomass was determined using the fumigation-extraction method (18). Characterization of the sediments was performed during field sampling, post handling, start of acclimation, start of the study, and end of test. An example of this for one of the time points is shown in Supporting Information Tables S1 and S2. Preparation of Test Chambers and Acclimation Prior to Study Initiation. A 50 g (dry weight) portion of sediment 5804

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 16, 2007

and 275 mL of water were added to each test chamber so the resulting water/sediment volume ratio was 3:1 and the minimum sediment layer height was approximately 3 cm. The aerobic and anaerobic sediment/water samples were acclimated for 14 days prior to the addition of the test substance to allow sediment conditions such as pH, 02 content, and redox potential to stabilize. Application of the Test Substance. All test chambers containing the sediment/water systems were fortified with the test substance at the start of the test by applying approximately 100 µL of the radiolabeled dosing solution the water layer to achieve a nominal test concentration ranging from 0.04 to 0.5 mg/L. Test concentrations were targeted for the low µg/L range while maintaining sufficient radioactivity for analysis of biotransformation products at 10% parent in the water and sediment phases. The radiopurities of the radiolabel test materials for the 3 pharmaceuticals and the specific activities of each are shown in Table 1. Following the addition of the test material to each test chamber, the water layer was gently mixed, making sure to minimize disturbance of the underlying sediment. Test Apparatus and Operating Conditions. The study requires 3 types of test systems. Two test systems monitor biotransformation and mineralization of the test material under aerobic and anaerobic conditions: one designed to capture CO2 and the other to capture volatiles and methane. A third test system was set up to support characterization of the water/sediment matrices during the study and at study termination. Conical flasks ranging from 500 to 1000 mL served as the test chambers, with the final size for each depending upon the amount of sediment/water required to

meet the 3:1 volume ratio criteria and sampling volumes required for analytical characterization. Each water/sediment system was tested under aerobic and anaerobic conditions at approximately 20 °C. (1) Biotransformation Test Apparatus (Aerobic Biotransformation and Mineralization, Anaerobic Biotransformation; Supporting Information Figure 1). The headspace gases within each of the transformation test chambers were continuously purged with either air (aerobic chambers) or nitrogen gas (anaerobic chambers) at a flow rate of approximately 1 bubble per second (assessed in the KOH traps). The gases exiting each test chamber were passed through a sorbent tube (Anasorb CSC, SKC Inc., Eighty Four, PA), followed by one empty bottle, then two gas washing bottles each containing approximately 100 mL of 1.5 N KOH (CO2 trapping solution), again followed by one empty bottle. The apparatus was designed for measuring biotransformation of test material under aerobic and anaerobic conditions, and subsequent CO2 evolution most notably associated with aerobic biodegradation. (2) Anaerobic Mineralization Test Apparatus (Anaerobic Mineralization; Supporting Information Figure 2). A separate anaerobic mineralization test apparatus was used in parallel to the above to measure total 14C-gases evolved either as CO2 or methane from the anaerobic water/sediment system, as described by Nuck and Federle (19). Briefly, the headspace gases within each of the anaerobic mineralization test chambers were continuously purged with a flow of nitrogen, passed through a gas collection system consisting of two sets of CO2 traps, then combined with a flow of oxygen. Combined gases were channeled into a tube furnace containing a quartz column packed with cupric oxide maintained at approximately 800 °C, combusting methane to CO2. Carbon dioxide was then captured in a final series of traps. While this mineralization test apparatus is very efficient in monitoring volatiles, CO2, and methane, the complexity of the apparatus only makes it suitable for a limited number of samples, such as is needed for the anaerobic system. (3) Water/Sediment Characterization Test Apparatus (Water and Sediment Characterization). The headspace gases within each of the aerobic and anaerobic characterization chambers were continuously purged with air or nitrogen, respectively, and maintained at conditions similar to those described above. The air or nitrogen exiting each characterization chamber was passed through a single gas-washing bottle containing water that was utilized to evaluate the gas flow rate through each chamber. Materials and Apparatus. Solvents obtained from Burdick & Jackson were HPLC-grade acetonitrile (ACN), dichloromethane (DCM), and methanol (MEOH). Reagent and buffers include phosphoric acid (Sigma-Aldrich, 99.999%, 85 wt % solution in water), hydrochloric acid (HCL; J.T.Baker, reagent grade), sodium hydroxide (NAOH) and potassium hydroxide (KOH; Mallinckrodt Baker Inc., analytical reagent grade). IN/US IN_FLOW ES Cocktail was obtained from INUS Systems Inc. Packard Ultima Gold and Ultima Gold XR scintillation cocktails were from Perkin-Elmer Life Sciences. The following apparati were used: Lindberg Mini Mite Tube Furnace, model 55035/31, PFTE tubing, 150 mm flow tubes and high-resolution valves (Cole Parmer). Disposition of 14C-Residues and Mass Balance. Mass balance analysis at each sampling point provided an internal validation for the study by accounting for the disposition of radioactivity found as 14C-volatiles in the sorbent, 14CO2 in KOH traps, and 14C activity in the water and sediment (extractable and non-extractable). Demonstrating that the total radioactivity at each sampling point is greater than 90% provides assurance that all of the activity is accounted for and appropriately characterized.

Determination of Total 14C-Volatiles in Sorbent and 2 in KOH Traps. Sorbent and KOH traps were sampled biweekly according to the schedule outlined in Table 3 of the Supporting Information. The sorbent trap was extracted with acetonitrile for organic volatiles and measured by a Packard TR 2500 liquid scintillation counter (LSC). KOH traps were sampled in triplicate in 1.0 mL aliquots and measured by LSC. Determination of Total 14C-Radioactivty in Water and Sediment Samples. The overlying water of each sacrificed vessel was removed with minimal disturbance to the sediment and the volume was measured. Triplicate 1-mL aliquots of the aqueous layer were assayed by LSC. The remainder of the water was then transferred into a separatory funnel for specific chemical analyses as described below. The sediments were measured for “extractable” and “nonextractable” radioactivity. Typically, the extractable radioactivity was determined by twice extracting with a select solvent in the original test vessels using sonic dismembranation. After each extraction, the vessels were centrifuged prior to decanting the solvent phase into a round-bottom flask. The second extract was combined with the first in the same round-bottom flask. Total radioactivity in the combined extract was determined by LSC analysis of triplicate 1-mL aliquots of the extract. The remaining solvent was processed as described below for the chemical specific analyses. At least three sub-samples of the extracted solids were combusted using a Packard model 307 oxidizer and assayed by LSC to determine the total radioactivity remaining with the solids as “non-extractable” residues. The sediment sample at the last sampling point, day 100, was subjected to extraction by an additional series of solvents ranging from polar to nonpolar to further demonstrate the non-extractable nature of the remaining residue. Chemical-Specific Analysis for Parent Drug and Profiling Biotransformation Products. Determination of 14C-Labeled Parent Drug and Biotransformation Products in Water and Sediment Samples. Water layer was twice extracted with 100 mL of a select solvent: ethyl acetate was used for azithromycin and exemestane, and dichloromethane was used for varenicline. The extracts were combined in a round-bottom flask. Total radioactivity in the extracts was determined by transferring triplicate 1-mL aliquots and assaying by LSC. Comparison of the total radioactivity in the extracts with that in the corresponding overlying water was used to confirm the efficiency of the extraction procedure. The water extracts were rotovapped at ∼40 °C to approximately 2.0 mL volume. This volume was transferred to a 15 mL centrifuge tube with several equal volume solvent rinses of the round-bottom flask where it was finally dried under N2 at 40 °C to a constant residual water volume. Final volume was appropriately adjusted to approximate the mobile phase composition of 95% acetonitrile:5% water. Samples were filtered (0.45 µm) and transferred to auto-sampler vials for final characterization by HPLC/RAM. The sediments remaining in the test systems were extracted twice with 100 mL of select solvents (ethyl acetate for exemestane, acetonitrile for azithromycin, and acidified (0.1 N HCL) acetonitrile for varenicline) by sonicating for 5 min at 50% amplitude. Solvents were transferred to roundbottom flasks and processed for total radioactivity. Samples were rotovapped and evaporated in a fashion to that of the water extracts. Aiquots (100 µL) of the water and sediment extracts were analyzed on a gradient HPLC (Agilent 1100, Hewlett-Packard 1090 or Hewlett-Packard 110) using a YMC-Pack ODS-AM 150 mm × 4.6 mm, 3 µm column operating with a flow rate of 1.0 mL/min at a column temperature of 40 °C. Retention times for azithromycin, exemestane, and varenicline were 7.2, 10.4, and 6.3, respectively and were detected with an

14CO

VOL. 41, NO. 16, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5805

TABLE 2. Data Summary for Depletion of Parent Drug in Aerobic and (Anaerobic) Test Systems watera regression compound exemestane varenicline azithromycin

water/sedimentb at day 100

Ke (days-1)

DT50 (days)

Ke (days-1)

DT50 (days)

0.070 (0.051) 0.032 (0.024) 0.031 (0.041)

10.0 (13.7) 21.9 (28.4) 21.4 (17.1)

0.046

15.1

0.028

24.8

0.030

23.1

% dose at day 100 14C-gasesc

in sediment

parentd water

parentd extractable sediment

19.2 (8.6) 0.30 (0.45) 0.4 (0.5)

34.1 (22.3) 52.3 (72.5) 87.3 (94.1)

0.8e (7.0) 1.1 (4.9) 5.0f (5.2)

0.2 (1.3) 4.9 (1.6) ND (ND)

14C-bound

a Elimination of parent (first-order regression) from water due to biotransformation, sorption, and mineralization. b Elimination of parent (single point estimate at day 100) from the water/sediment system due to biotransformation, sorption, mineralization, and bound residues. c % gases found as 14CO2, 14C-volatiles, and 14C-methane. d % found as parent in water or extractable sediment. e Highly biotransformed; only 2.5% activity in water as parent. f No biotransformation; 100% activity in water as parent.

INUS B-RAM detector. For azithromycin and exemestane, gradient conditions started with 100% A (10% acetonitrile: 90% water (0.1% v/v H3PO4) for 1.0 min; then ramped up to 100% B (95% acetonitrile:5% v/v H3PO4) by 8.0 min and held at those conditions until 12 min. For varenicline, gradient conditions started with 95% A (0.1% H3PO4) and 5% B (acetonitrile) for 2 min; ramped up to 80% B by 7 min, and then ramped up again to 95% B by 7.1 min where conditions were held until 12 min. The gradient was re-equilibrated back to initial gradient conditions for 3 min before starting the next sample. Non-extractable Residues. At study termination, the remaining sediment was extracted by a series of solvents ranging from polar to nonpolar properties. Acidified acetonitrile, and/or a combination of acetonitrile, ethyl acetate, and hexane, were typically used for this purpose. Amounts of residues extracted into the solvents were determined by LSC procedures followed by HPLC/RAM characterization. Estimation of Depletion Rate for Parent. The elimination rate of parent from water (first order regression analysis), and the elimination rate of parent from the test system at day 100 (single point estimate) are provided in Table 2; along with a summary of the % of residues at day 100 found as (1) evolved 14C-CO2, (2) % bound in sediment, (3) % parent in water, and (4) % parent in extractable sediment. The values presented are the averages of values determined from duplicate test systems. The depletion rate accounts for biotransformation, mineralization, and binding as unextractable residues.

Results and Discussion EMEA Phase II Tier A ERA: Use of the OECD 308. The data output from the OECD 308 and the Ready Biodegradation OECD 301 do not adequately support the estimation of environmental concentrations for the Phase II, Tier A ERA (Supporting Information Table S4). The proposed OECD “3xx” Guideline for activated sludge and surface water would be a better test at this point in the risk assessment for providing biotransformation and mineralization rates in the respective matrices and characterization of biodegradation products. The inclusion of an abbreviated OECD 308 study (aerobic, 1 point at day 100) would also be very helpful in this initial tier to establish the amount of unextractable residues in sediment and the amount of unchanged drug in water and sediment at day 100. This would not be as resource intensive as a complete 308 study and yet would provide valuable information regarding the potential need for ecotoxicity testing in sediments and whether more definitive Tier B sediment biodegradation testing would be needed in respect to the potential environmental risks. All three suggested tests could be conducted at this point in ERA testing for a cost less than conducting the one OECD 308. 5806

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 16, 2007

Method. The OECD 308 protocol provides a high level assessment of the disposition and biodegradation of a test material in a water/sediment system. Test endpoints provided depletion of test material from the water and sediment phases, and from the test system itself. The data are helpful in demonstrating its overall biodegradability, and in its application in classification schemes that define whether something is persistent based on its overall depletion, as defined in the European Chemical Bureau Technical Guidance Document (20). The data output from the OECD 308 is in the format of overall depletion, representing the net product of biodegradation and sorption to sediment. A sterile abiotic control for each test system could be used to correct the total depletion for sorption, to provide a biodegradation rate. But it is uncertain to what extent the sorptive properties of sediment would be impacted by sterilization techniques and whether it would still be suitable as a control after sterilization (21, 22). Consequently, the uncorrected depletion data from the study are not always helpful for use in standard environmental models, such as EUSES or GREAT-ER, for refining predicted environmental concentrations that require degradation rates for each compartment and Koc values for the respective matrices. The overall 100 day study period may not be sufficient to fully characterize the depletion of pharmaceutical residues; and in this study it may have been helpful to extend the study period beyond 100 days for those cases where considerable mineralization was continuing at day 100; or for those cases where one may wish to investigate whether additional time would have been helpful in the acclimation process. In doing so, considerations around how best to supplement the system with additional chemical oxygen demand or other food sources to keep the micro-organisms in the system viable would need further investigation. Test System. The aerobic and anaerobic test systems as described provide continuous flows that were designed to ensure that all of the released 14CO2 (or other volatiles) are swept through the system into the appropriate traps. They were easy to maintain and provided good mass balance throughout the study for all of the residues during each of the various sampling intervals. Drawing air through the aerobic system with a vacuum source, as opposed to pushing air through under pressure, also helps to minimize the potential loss of any 14CO2 should a leak occur. The entering air may be humidified via a water trap prior to the test chamber to ensure the system does not dry out during the study period. For the anaerobic system, a cupric oxide furnace was employed to oxidize all volatiles and methane to 14CO2. This approach provided a highly efficient means of trapping the volatile components as 14CO2, rather than implementing

FIGURE 1. Disposition of 14C-total residues in water/sediment systems: % as CO2, sediment extractable, sediment non-extractable, and aqueous; and % depletion of 14C-parent in aqueous phase. traps for other types of volatiles, such as methane and volatile degradants. Mass Balance. The use of radio-labeled test material is of prime importance in running the study. 14C combustion and LSC analyses allows a complete mass balance for each sample during the study, thereby providing an accurate accounting of all of the residues, whether freely dissolved, bound to sediment, liberated as CO2, methane, or as a volatile degradant. Residues remaining on the surface of the test apparatus could also be rinsed and accounted for, if necessary. Residual activity potentially trapped in the soil as 14C-bicarbonate is not uniquely characterized under the described protocol, but is potentially part of what is characterized as “bound”. Procedures are available to liberate this trapped 14C-bicarbonate by acid treatment (23) and would be helpful in further discriminating the fate and disposition of this residue from other “bound” residues. Mass balance (Figure 1) remained >90% through all of the sampling intervals indicating the study was well controlled. Chemical Analysis. Specific chemical analyses by HPLC/ RAM and HPLC/MS/MS are all key in characterizing the profile of biotransformation products present and in un-

derstanding the mechanism of depletion. Such an approach has also been helpful in establishing potential links of environmental biodegradation to human metabolism and to pure culture screening. Study concentrations are targeted to be environmentally relevant when and where possible, but may need to be implemented at higher levels to allow for proper analytical measurement. The concentration selected at the beginning of the study will depend upon the specific activity of the radiolabeled test material, sensitivity of the instrumentation (LC/RAM), amount of water/sediment used in the test system, recoveries of test material from the matrix, extent of anticipated biotransformation, and the potential for inhibition, if applicable. Figure 2 shows representative LC/RAM chromatograms of the 3 pharmaceuticals for standards and one of the overlying water samples. In all cases, the individual metabolites did not exceed 10% of the dose. Analytical method validation was performed for each test material prior to the initiation of the study to demonstrate recovery from the water and sediment phases, as well as at least 90% total recovery from the test system. Recovery from water and sediments is assessed for a series of solvents ranging VOL. 41, NO. 16, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5807

5808

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 16, 2007

FIGURE 2. Sample chomatograms of standards and representative aerobic overlying water samples for exemestane. in polarity (methanol, acetonitrile, ethyl acetate, methylene chloride); with aqueous phase buffered (e0.1 M) as needed to minimize ionization of the test material for optimal recovery. Solvent selection is based on the solvent and pH conditions that provide the best recovery of activity. Stronger treatment with acid or base is reserved for use only during final characterization of “bound” residues where degradation of the matrix is allowed. Sediment Characterization, Study Controls. The water/ sediment characteristics (Tables S2 and S3 in the Supporting Information) remained fairly consistent throughout study. It seems that the sampling interval may be more frequent than what was needed to demonstrate stability and control of the sample conditions. Currently there are no “negative” abiotic controls, nor “positive” controls that represent pharmaceuticals of known “depletion” rates. As a long-term goal, it would be helpful to identify several pharmaceuticals representing a “rapid” depletion, and one representing something more typical. Such controls would be helpful in normalizing seasonal variations of the biological activity of the grab water/sediment samples used in the method. Based on the data from this study, and observations from other studies conducted, exemestane could be suggested as a positive control of a relatively “rapid” depleting pharmaceutical showing extensive biotransformation and moderate mineralization. Negative abiotic controls could be added as well, by sterilizing sediment samples with γ radiation, treating with mercuric chloride, and/or autoclaving (24). Though negative controls could provide a means for assessing abiotic biodegradation, it is uncertain which sterilization treatment(s) and what frequency of treatment(s) would be needed to maintain abiotic conditions for the duration of the study. Disposition of Total 14C Residues: Comparison of Aerobic and Anaerobic Biodegradation. The aerobic and anaerobic depletion profiles for exemestane, varenicline, and azithromycin are provided in Figure 1 showing % total 14C residues found as evolved CO2; extractable and non-extract-

able residues in sediment, dissolved in the aqueous phase; or as 14C-parent in the aqueous phase. Comparisons of the aerobic and anaerobic results show a similar trend for all three pharmaceuticals: similar deposition and biotransformation products with anaerobic results generally showing a slower depletion rate. Given these results and results from other studies that are currently ongoing, we would not perform anaerobic studies as a default for all pharmaceuticals, but rather use a tiered analysis with outcomes from an aerobic study providing input into a decision analysis as to whether there is a need for further anaerobic testing. Such recommendations would be favorable for those compounds that were amenable to dehalogenation (25) or reduction of key functional groups (26). Estimation of Depletion Rate for Parent. Table 2 provides a brief summary of all the testing that was done over the 100-day period, representing many analyses. While rate constants in this definitive test are determined from multiple time points using a regression analysis, one might consider a one-point estimate at day 100 as a way of estimating depletion rates, perhaps as a screening tool. Though one would still require 100 days for the analysis, considerable cost and effort would be saved while generating most of the key information about the disposition and depletion of the test substance. Partitioning and Bound Residues. The compounds with the highest Koc values for soils similarly represent the compounds with the highest amount found as unextractable in sediments. For those cases were little mineralization is observed, the unextractable residues are likely attributed to unchanged drug and/or its biodegradation products irreversibly bound to the sediments rather than CO2 trapped as bicarbonate or radioactivity re-incorporated into the biomass. It is also a likely explanation given the mixed functional groups present with these pharmaceuticals, and the predisposition for ionic exchange to occur with those sediments containing clay and the potential for covalent (and VOL. 41, NO. 16, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5809

other) binding to occur. It is possible under such circumstances to consider these residues as not bioavailable (27) or essentially removed from the test system. Such considerations may be helpful when characterizing whether a pharmaceutical persists or not. Further characterization of the bonds (ionic, charge-transfer, covalent, van-der-Waals) involved and nature of the residues (28, 29) present is possible, but given the complexity of the analyses and expertise needed for the interpretation, it would be difficult to implement such procedures in a routine method. There do not appear to be any methods available for estimating this property of irreversible binding from the physical-chemical properties of the test substances. Exemestane. Of the three pharmaceuticals, exemestane resulted in the most rapid depletion rate (DT50 of 9.9 days) and greatest number of biotransformation products (6 peaks eluting prior to unchanged drug at 10.4 min). Exemestane resulted in 19% mineralization, with less than 3% remaining as parent in both water and sediment. With lowest soil Koc value, it had the greatest amount found in the overlying water. At day 100, CO2 evolution was still occurring. Extending the study period beyond 100 days may have been helpful to assess the full potential of exemestane to mineralize. Given the amount of CO2 liberation and extent of biotransformation, and observed half-life of parent, exemestane is not considered to persist. Varenicline. The extent of biotransformation was not as great as exemestane, resulting in 3 less polar biotransformation peaks eluting around 7.0, 7.5, and 8.0 min, with varenicline around 6.3 min. Though the individual metabolites were less than 10% of the total dose; a preliminary investigation of the predominant less polar peak was initiated. Findings indicated that this peak was consistent with a N-methylated metabolite, consistent with observation made with other secondary amines (unpublished LC-MS/MS data) resulting from sludge and/or water/sediment biodegradation. A significant amount of the residues were bound (50%) at the end of the study, and represent the major mechanism for removal from the test system. No notable mineralization was observed over the 100-day period. Depletion half-life for parent ranged from 22 to 29 days, and therefore, like exemestane, varenicline does not persist. Azithromycin. Of the three pharmaceuticals investigated azithromycin was the only one that did not show significant biodegradation, with elution at 7.2 min. It had the highest Koc value, and the highest amount bound as unextractable residues at day 100, reaching almost 90% of the dose under aerobic conditions. No further residues were released from the sediments at the end of the study when treated with 0.1 N HCL. For the most part, binding after day 14 remained constant and did not decrease at any time thereafter. The study was conducted below the activated sludge EC-10 value of 1.89 mg/L so potential inhibition would not be a concern. Because of azithromycin’s cationic properties, it is likely that the unextractable residues are highly bound to the sediment and not bioavailable. Because of the extent of binding observed, one might consider these residues as essentially removed from the system.

Acknowledgments I thank the contract research organization, Wildlife International Laboratory of Easton Maryland, and Ed Schaefer for performing the azithromycin, exemestane, and varenicline studies.

Supporting Information Available Table S1: Choptank Water/Sediment Characteristics @ Start of Study; Table S2: Turkey Creek Water/Sediment Characteristics @ Start of Study; Table S3: Sampling Schedule for 5810

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 16, 2007

CO2 Evolution, Volatiles, Total Radioactivity and Chemical Specific Analyses for Aerobic and Anaerobic Conditions; Table S4: Physical-chemical, fate and effects studies recommended in Phase II Tier A; Figure S1: Schematic of Aerobic Biotransformation Test Apparatus; Figure S2: Schematic of Anaerobic Mineralization Test Apparatus. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Ternes, T. A. Occurrence of drugs in German sewage treatment plants and rivers. Water Res. 1998, 32, 3245-3260. (2) Halling-Sorensen, B.; Nielsen, S.; Lanzky, P.; Ingerslev, F.; Lutzhoft, H.; Jorgensen S. Occurrence, fate and efffects of pharmaceutical substances in the environment- A Review. Chemosphere. 1998, 36 (2), 357-393. (3) Heberer, T. Occurrence, fate, and removal of pharmaceutical residues in the aquatic environment; a review. Toxicol. Lett. 2002, 131, 5-17. (4) Kolpin, D.; Furlong, E.; Meyer, M.; Thurman, E.; Zaugg, S.; Barber, L.; Buxton, H. Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. streams, 1999-2000: A national reconnaissance. Environ. Sci. Technol. 2002, 36 (6), 1202-1211. (5) Beausse, J. Selected drugs in solid matrices: A review of environmental determination, occurrence, and properties of principal substances. TrAC, Environ. Trends Anal. Chem. 2004, 23 (10-11), 753-761. (6) Versteeg, D.; Alder, A.; Cunningham, V.; Kolpin, D.; MurraySmith, R.; Ternes, T. Environmental Exposure Modeling and Monitoring of Human Pharmaceutical Concentrations in the Environment. In Human Pharmaceuticals - Assessing the Impacts on Aquatic Ecosystems; Williams, R. T., Ed.; SETAC: Brussels, Belgium2003. (7) Kummerer, K.; Ericson, J.; Hannah, R.; Johnson, A.; Sedlak, D.; Weston, J. Environmental Fate and Transport of Human Pharmaceuticals. Human Pharmaceuticals - Assessing the Impacts on Aquatic Ecosystems; Williams, R. T., Ed.; SETAC: Brussels, Belgium2003. (8) Moehle, E.; Kempter, C.; Kern, A.; Metzger, J. W. Examination of the degradation of drugs in municipal sewage plants using liquid chromatography-electrospray mass spectrometry. Acta Hydrochim. Hydrobiol. 1999, 27, 430-436. (9) McArdell, C. S.; Molnar, E.; Suter, M. J.; Giger, W. Occurrence and fate of macrolide antibiotics in wastewater treatment plants and in the Glatt Valley Watershed, Switzerland. Environ. Sci. Technol. 2003, 38, 5479-5486. (10) Furlong, E.; Ferrer, I.; Thurman, M. Identification and quantitation of pharmaceuticals, surfactants and other contaminants by atmospheric pressure tandem mass spectrometry. In Abstracts ACS National Meeting, ACS Div. of Environ. Chem. 2002, 42 (1), 378-384. (11) Weigel, S. Occurrence, distribution and fate of pharmaceuticals and further polar contaminants in the marine environment. Ph.D. dissertation, University of Hamburg, 2002; http://www.meind.de/search.py?20373. (12) Diaz-Cruz, S.; Barcelo, D. Occurrence and analysis of selected pharmaceuticals and metabolites as contaminants present in wastewaters, sludge and sediments. In Handbook of Environmental Chemistry; Springer-Verlag: Berlin, 2004; Vol. 5 (Pt. 1), pp 227-260. (13) Buser, H.; Mueller, M.; Theobald, N. Occurrence of the Pharmaceutical Drug Clofibric Acid and the Herbicide Mecoprop in Various Swiss Lakes and in the North Sea. Environ. Sci. Technol. 1998, 32 (1), 188-192. (14) Bruchet, A.; Hochereau, C.; Picard, C.; Decottignies, V.; Rodrigues, J. M.; Janex-Habibi, M. L. Analysis of drugs and personal care products in French source and drinking waters : the analytical challenge and examples of application. Water Sci. Technol. 2005, 52 (8, 4th World Water Congress: Water and Health, 2004), 53-61. (15) European Medicines Agency, Committee for Medicinal Products for Human Use. Guideline on the Environmental Risk Assessment of Medicinal Products for Human Use; CHMP/SWP/4447/00; London, 01 June 2006. (16) Ternes, T. A. Environmental Fate of Pharmaceuticals in Water/ Sediment Systems. Environ. Sci. Technol. 2005, 39, 52095218. (17) OECD. Aerobic and Anaerobic Transformation in Aquatic Sediment Systems; Guideline for Testing of Chemicals No. 308; OECD: Paris, 2002.

(18) ISO Method 14240-2. Soil Quality - Determination of Soil Microbial Biomass-Fumigation Extraction Method; 1997. (19) Nuck, B.; Federle, T. Batch Test for Assessing the Mineralization of 14C-Radiolabeled Compounds under Realistic Anaerobic Conditions. Environ. Sci. Technol. 1996, 30, 3597-3603. (20) Technical Guidance Document on Risk Assessment Part 11; European Commission Joint Research Centre; EUR 20418 EN/ 2; 2003; p 164. (21) Lotrario, J.; Stuart, B.; Lam, T.; Arands, R.; O’Conner, O.; Kosson, D. Effects of sterilization methods on the physical characteristics of soil: implications for soil sorption isotherm analyses. Bull. Environ. Contam. Toxicol. 1995, 54 (5), 668-75. (22) Stephens, K.; Farenhorst, A.; Fuller, L. Effect of soil sterilization by mercuric chloride on herbicide sorption by soil. J. Environ. Sci. 2002, B37 (6), 561-571. (23) Schwab, E.; Federle, T. Simulation Test for Evaluating Fate in Aerobic Sediments: Application to Surfactants; Poster Presentation at SETAC North America, Baltimore, MD, 2005. (24) Wolf, D.; Dao, T.; Scott, H.; Lavy, T. Influence of sterilization methods on selected soil microbiological, physical and chemical properties. J. Environ. Qual. 1989, 18 (1), 39-44.

(25) Lee, M.; Davis, J. Natural Remediation of Chlorinated Organic Compounds. In Natural Remediation of Environmental Contaminants; Swindoll, M., et al., Eds.; SETAC: Brussels, Belgium, 2000. (26) O’Conner, O.; Young, L. Effects of Six Different Functional Groups and Their Position on the Bacterial Metabolism of Monsubstituted Phenols under Anaerobic Conditions. Environ. Sci. Technol. 1996, 30 (5), 1419-1428. (27) European Crop Protection Association (ECPA). ECPA Position Paper of Soil Non-Extractable Residues; D/00/SuM/5277; Brussels, Belgium, 2000. (28) Northcott, G.; Jones, K. Experimental approaches and analytical techniques for determining organic compound bound residues in soil and sediment. Environ. Pollut. 2000, 108, 19-43. (29) Klein, W.; Still, G.; Kearney, P.; Drescher, N.; Desmoras, J.; Esser, H.; Aharonson, N.; Vonk, J. Non-extractable pesticide residues in soils and plants. Pure Appl. Chem. 1984, 56 (7), 945-956.

Received for review December 21, 2006. Revised manuscript received May 11, 2007. Accepted May 19, 2007. ES063043+

VOL. 41, NO. 16, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5811