Environ. Sci. Technol. 2004, 38, 3351-3359
Environmental Risk Assessment of Paroxetine VIRGINIA L. CUNNINGHAM,* DAVID J. C. CONSTABLE, AND ROBERT E. HANNAH Corporate Environment, Health and Safety, GlaxoSmithKline, 2200 Renaissance Boulevard, Suite 105, King of Prussia, Pennsylvania 19406
Paroxetine hydrochloride hemihydrate (the active ingredient in Paxil) is a pharmaceutical compound used for the treatment of depression, social anxiety disorder, obsessive compulsive disorder, panic disorder, and generalized anxiety disorder. Paroxetine (PA) is extensively metabolized in humans, with about 97% of the parent compound being excreted as metabolites through the urine and feces of patients. Therefore PA and metabolites have the potential to be discharged into wastewater treatment systems after therapeutic use. PA and its major human metabolite (PM) were investigated using studies designed to describe physical/chemical characteristics and determine their fate and effects in the aquatic environment. A significant portion of the PM entering a wastewater treatment plant would be expected to biodegrade given the higher activated sludge solids concentrations present in a typical wastewater treatment plant. The potential for direct photolysis of PM is also possible based on photolysis results for PA itself. These results provide strong support for expecting that PA and PM residuals will not persist in the aquatic environment after discharge from a wastewater treatment facility. This conclusion is also supported by the results of a USGS monitoring study, where no PM was detected in any of the samples at the 260 ng/L reporting limit. The results presented here also demonstrate the importance of understanding the human metabolism of a pharmaceutical so that the appropriate molecule(s) is used for fate and effects studies. In addition to the PA fate studies, PM was investigated using studies designed to determine potential environmental effects and a predicted no effect level (PNEC). The average measured activated sludge respiration inhibition value (EC50) for PM was 82 mg/ L. The measured Microtox EC50 value was 33.0 mg/L, while the Daphnia magna EC50 value was 35.0 mg/L. The PNEC for PM was calculated to be 35.0 µg/L. Fate data were then used in a new watershed-based environmental risk assessment model, PhATE, to predict environmental concentrations (PECs). Comparison of the calculated PECs with the PNEC allows an assessment of potential environmental risk. Within the 1-99% of stream segments in the PhATE model, PEC values ranged from 0.003 to 100 ng/L. The risk assessment PEC/PNEC ratios ranged from ∼3 × 10-8 to ∼3 × 10-3, indicating a wide margin of safety, since a PEC/PNEC ratio < 1 is generally considered to represent a low risk to the environment. In addition, * Corresponding author phone: (610)239-5262; fax: (610)239-5250; e-mail:
[email protected]. 10.1021/es035119x CCC: $27.50 Published on Web 05/14/2004
2004 American Chemical Society
Microtox studies carried out on PM biodegradation byproducts indicated no detectable residual toxicity. Any compounds in the environment as a result of the biodegradation of PM should be innocuous polar byproducts that should not exert any toxic effects.
Introduction Speculation and public concern have increased considerably over the past several years as improved analytical techniques and limits of detection have confirmed the presence of pharmaceuticals in the environment (1-8). However, because the U.S. Food and Drug Administration (FDA) has required environmental assessments (EAs) for new pharmaceuticals since 1977, the FDA has been using these assessments to determine the likelihood that a new pharmaceutical compound may have a significant effect on the environment. These EAs are publicly available from the FDA under the Freedom of Information Act. The environmental fate chemistry and effects information in this paper was taken in part from the Environmental Assessment of the Paxil (paroxetine hydrochloride hemihydrate) New Drug Application #020031. This application was approved by the U.S. Food and Drug Administration (FDA) on December 29, 1992, and a “Finding of no Significant Impact (FONSI)” was issued. Paroxetine hydrochloride hemihydrate, CAS [110429-351], is the generic designation for (-)-trans-4R-(4′-fluorophenyl)-3S-[(3′,4′-methylenedioxyphenoxy)methyl] piperidine hydrochloride (BRL 29060A). PM used in these studies is the designation for (-)-trans-4-[4-(4′-fluorophenyl)-3-piperidinylmethoxy]-2-methoxyphenol hydrochloride (BRL 36610A). The suffix A in the GSK code designates the hydrochloride salt form of the compound. The chemical structures of paroxetine and its metabolite are shown in Figure 1. Mode of Action and Metabolism. PA hydrochloride hemihydrate (the active ingredient in Paxil) is a pharmaceutical compound used for the treatment of depression, social anxiety disorder, obsessive compulsive disorder, panic disorder, and generalized anxiety disorder. Its mode of action is presumed to be linked to the potentiation of serotonergic activity in the central nervous system resulting from inhibition of neuronal reuptake of serotonin (9). PA is completely absorbed after oral dosing of the hydrochloride salt and is extensively metabolized (10). The principal metabolites are readily cleared since they are polar and conjugated products of oxidation and methylation. Because the relative potencies of PA’s major metabolites are at most 1/50 of the parent compound, they are essentially inactive. Metabolism is initiated in the liver by oxidation at the methylenedioxyphenyl carbon atom. Identification of metabolic end products in the urine of mice, Rhesus monkeys, and humans indicated that this is the primary metabolic process for all species. The catechol intermediate resulting from the initial oxidation was too unstable to isolate. It is further metabolized, in part, by methylation at the meta-position (BRL 36610), followed by conjugation of the free phenolic group with glucuronic acid or sulfate to produce the major metabolites in plasma, urine, and bile. Some methylation at the para position and cleavage of the ether linkage also occurs. The proposed metabolic pathway in humans is illustrated in Figure 2. Approximately 64% of a 30 mg oral solution dose of PA was excreted in the urine with 2% as the parent compound and 62% as metabolites over a 10-day post dosing period. About 36% was excreted in the feces, mostly as metabolites VOL. 38, NO. 12, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3351
Experimental Section
FIGURE 1. Chemical structures of paroxetine and PM.
FIGURE 2. Proposed metabolic pathway in humans for PA (BRL 29060) (10). and less than 1% as the parent compound (10). Because these compounds have the potential to be discharged into wastewater treatment systems after therapeutic use, it is important to characterize their environmental chemistry, fate and effects in the aquatic environment. It has been reported (11-13) that pharmaceutical conjugates may be hydrolyzed back to the parent compound in wastewater treatment systems; therefore, PA itself (BRL 29060) and its major metabolite, PM (BRL 36610), were the compounds selected for study. Environmental Fate and Effects. The fate of a compound in the environment is a function of its molecular structure, polarity, solubility in water, volatility, and susceptibility to hydrolysis, photolysis, and biodegradation. Additionally, its water-solids partitioning behavior in wastewater treatment facilities and in subsequent surface waters contributes to its fate in the environment. The information presented here was collected to characterize and understand the potential fate of PA and PM in the aquatic environment. Since most studies reported to date have focused on the fate and presence of pharmaceuticals, there are at present limited published data on environmental effects and on the application of fate and effects data to environmental risk assessments. Since PM has the potential to be discharged into wastewater treatment systems after therapeutic use, it is important to characterize its environmental effects in the aquatic environment. This allows the estimation of a predicted no effects level (PNEC). Fate data were then used in a new watershed based environmental risk assessment model, PhATE (14), to predict environmental concentrations (PECs). Comparison of the calculated PECs with the PNEC allows an assessment of potential environmental risk. 3352
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 12, 2004
Materials. The PA hydrochloride hemihydrate (BRL 29060A) and PM hydrochloride (BRL 36610A) test materials used for the fate and effects studies were prepared by GlaxoSmithKline. Unlabeled test materials and reference standards were 99.9% or greater purity for PA hydrochloride hemihydrate (87.7% as free base). PA hydrochloride hemihydrate is a nonvolatile solid with a melting point range of 120° to 138 °C. Unlabeled PM hydrochloride salt test material was 93.7% pure (84.4% as free base) for the sample used for preliminary studies and 92.1% pure (83.0% free base) for the sample used as an analytical reference standard and for definitive studies. 3,5-Dichlorophenol was obtained from Aldrich Chemical. The specific purity was 97%. Environmental Fate Studies. Analytical Methods - Fate. Analysis for PA and PM was performed using a reverse-phase high-performance liquid chromatography (HPLC) method. A DuPont Zorbax RX-C8 column (4.6 × 250 mm, 5-µm particle size, Mac-Mod Analytical, Chadds Ford, PA) was used for separation. The mobile phase was 65% TFA‚H2O (aqueous trifluoroacetic acid)/35% acetonitrile for PA and 70% TFA‚ H2O (aqueous trifluoroacetic acid)/30% acetonitrile for PM. The separation was conducted under isocratic conditions at a flow rate of 1.2 mL/min with ultraviolet detection at 295 nm. Injection volume was 1 µL. The compounds were quantified by peak area comparisons using multipoint standard curves. The limit of detection was 1 µg/L (ppb), while the limit of quantification was 3 µg/L (ppb). Solubility Studies. The water solubility of PA was determined by the oversaturation and undersaturation shake flask method (15). The amount of PA to be dissolved in each test solution was determined from the results of a preliminary water solubility study. The pH of the control solutions was measured for each group at the end of the study. The solutions were considered to be at equilibrium when the concentration of the over- and undersaturated solutions remained within (5% of each other for three consecutive samplings. Acid Dissociation Constant. The acid dissociation constant of PA was determined according to the procedures given in Albert and Serjeant (16). Potentiometric titration was used despite the marginal solubility of PA at pH 9 because (1) UV/vis spectrophotometry was not applicable and (2) the expected pKa of PA of ∼9.4 was considered to be too high for the conductimetry method. A pKa was calculated from the pH versus titrant volume data using the method of Albert and Serjeant (16) according to
pKa ) pH + log([BH+]/[B])
(1)
where [BH+] is the molar concentration of protonated PA base and [B] is the molar concentration of PA free base. Ultraviolet-Visible Absorption in pH 5, 7, and 9 Aqueous Buffers. The ultraviolet-visible absorption spectra for PA were determined in triplicate at 25 °C in pH 5, 7, and 9 aqueous buffers using a Hewlett-Packard Model 8452A diode array spectrophotometer using OECD Method 101 (17). Absorbance spectra were determined against appropriate blank solutions. Values for lambda max, bandwidth, and molar absorptivity were determined from the spectral data. Octanol/Water Distribution Coefficient. The n-octanol/ water distribution coefficient was determined using the method described by ref 18. The n-octanol used in the study (99.9% pure) was obtained from Aldrich Chemical Co. and was used without further purification. A Hewlett-Packard model 8452A diode array spectrophotometer was used to quantify PA in the test samples. Sorption to Biomass. The sorption of PA to biomass was determined according to ref 19. The biomass used in this study was collected from the GlaxoSmithKline R&D site, King
TABLE 1. Physical/Chemical Properties of PA in Deionized Water and pH 5, 7, and 9 Buffers parameter water solubility (25 ( 0.5 °C) oversaturation undersaturation pKa vapor pressure estimate Henry’s Law constant estimate UV/vis spectrum λmax (nm) bandwidth (nm) (SD) Log Dow Log Kp
5
7
9
D.I. water (pH 6.5)
7881 mg/L 5696 mg/L
1132 mg/L 1131 mg/L
341; 430 mg/L 318; 426 mg/L
6804 mg/L 5050 mg/L
234, 292 33, 30 3732, 3828 1.15, 1.09
234, 292 29, 28 3823, 3817 1.30, 1.35
234, 292 29, 28 3806, 3797 3.29, 3.26
9.6 8.25 × 10-6 Torr
