Final Report: The Environmental Occurrence, Fate, and Ecotoxicity of Selective Serotonin Reuptake Inhibitors (SSRIs) in Aquatic Environments

EPA Grant Number: R829006
Title: The Environmental Occurrence, Fate, and Ecotoxicity of Selective Serotonin Reuptake Inhibitors (SSRIs) in Aquatic Environments
Investigators: Black, Marsha C. , Armbrust, Kevin L.
Institution: University of Georgia
EPA Project Officer: Page, Angela
Project Period: September 1, 2001 through August 31, 2004 (Extended to April 30, 2007)
Project Amount: $522,892
RFA: Drinking Water (2000) RFA Text |  Recipients Lists
Research Category: Drinking Water , Water Quality , Water


The two primary goals of our investigations were to determine if selective serotonin reuptake inhibitors (SSRIs) could accumulate and persist in the aquatic environment as parent drugs and/or metabolites and if so, at what concentrations would they be toxic to model aquatic organisms.  To accomplish these goals we conducted multiple laboratory experiments to determine the chemical characteristics, degradation potential and fate of these chemicals under simulated environmental conditions.  These experiments were patterned after established methods for determining the environmental fate of pesticides in the environment.  To determine the potential for adverse effects of these chemicals on aquatic organisms we conducted experiments to characterize their acute and chronic toxicities to three taxa, including an aquatic invertebrate, fish and aquatic frog.  Finally to determine the potential for exposure to aquatic organisms to these chemicals and their potential for eliciting adverse effects under realistic conditions, we collected and analyzed sediment and water samples from a local watershed receiving permitted municipal effluents, and analyzed them for SSRI parent compounds and metabolites.


For each SSRI, laboratory environmental fate measurements will include the following: aqueous solubility; octanol-water partition coefficient (Kow); acid dissociation constant (pKa); soil adsorption coefficient (Koc); hydrolysis and aqueous direct photolysis rates at pH 5,7and 9; indirect photolysis rates including the hydroxyl radical rate constant; irradiated and dark aerobic-aquatic metabolism; and ready biodegradability. Major degradation products will be identified wherever possible. The occurrence of each SSRI and major metabolites/degradation products will be measured over one year in raw and treated wastewater, and receiving water at sub-ppb levels. Potential loads to waste treatment facilities will be assessed through surveys of numbers of prescriptions filled by local pharmacies. Acute toxicity and chronic exposure tests for survival and reproduction will be conducted with Ceriodaphnia following EPA protocols on each SSRI as well as on raw sewage water and treated effluent. Toxicity identification evaluation (TIE) methods will be used to determine if SSRIs are responsible for observed toxicity in any sample. Additionally, the mosquito fish, Gambusia, will be chronically exposed to SSRIs in outdoor microcosms to assess their impacts on the fish's spawning, potential and real fecundity, gonadal somatic index and brood clutch size.

Summary/Accomplishments (Outputs/Outcomes):

I. Environmental fate of SSRIs in the Aquatic Environment
            Physicochemical properties measured in the five SSRIs (citalopram, fluoxetine, fluvoxamine, paroxetine and sertraline) can be used to predict their environmental persistence and fate.  Compound hydrophobicity is a key parameter predicting environmental persistence of organic chemicals and can be measured as water solubility or octanol:water partition coefficient (Kow).  As the hydrochloride salt form, all five SSRIs had relatively high water solubilities and relatively low octanol-water partition coefficients (Kow, 1.12-1.39) (Table 1).  Results for these two properties are in agreement and predict that these compounds will have low persistence in the aquatic environment.  Water soluble compounds typically do not sorb to sediments and are more likely to be degraded by microbial processes, reducing their persistence in the environment. However, it should be noted that these measurements were made on the salt form (hydrochloride) of each SSRI, which is expected to be significantly more water soluble that the free base form.
Sorption coefficients measured for the five SSRIs with soils and sediments  indicate a different environmental fate scenario for each compound.  Very high sorption coefficients (presented as log Koc or organic carbon normalized sorption coefficient) were measured for all SSRIs, with a range of log Koc values from 3.82 to 5.63 for fluvoxamine (lowest degree of sorption) and citalopram (highest degree of sorption) (Table 1).  These values indicate that all of the SSRIs have a very high affinity for sorbing to soils and sediments, which in the absence of other degradation processes would increase the half-lives of each of these chemicals in aquatic environments through sequestration in sediments, but greatly reduce their concentration in overlying water.
Other physicochemical measurements of the SSRIs further separate the SSRIs into two types: those that will likely persist in the aquatic environment and those that will likely be degraded into more polar degradation products.  Photolysis is a potentially important route of degradation for several SSRIs.  Photolysis rates and half lives were measured in buffered laboratory water (± synthetic humic acid) and in river water to simulate a more environmentally-relevant condition.  Photolysis half lives for the SSRIs in lake water (Table 1) indicate that fluoxetine was far more resistant to photolysis than the other SSRIs, with a half-life (t½) of 122 days.  Fluoxetine was very resistant to hydrolysis and photolysis in buffered laboratory water, with degradation less than 10 percent. Citalopram and sertraline exhibited intermediate photolytic stability in lake water under simulated sunlight.  Photolysis rates for sertraline and citalopram were very slow pH 5 and 7 (buffered laboratory water), yet complete photodegradation of sertraline occurred at pH 9 by 10 days and similar conditions greatly increased photodegradation of citalopram. Degradation of citalopram was faster in synthetic humic water than in pH 9 buffer or lake waters.  Paroxetine and fluvoxamine were rapidly degraded (half-lives were less than one day) in river water.  The photolysis of paroxetine in buffered laboratory water was accelerated by increasing pH, with two degradation products detected.  Fluvoxamine was moderately degraded by simulated sunlight, and degradation kinetics were biphasic, with slow initial degradation followed by more rapid degradation after 7 days.  Photolysis rates of citalopram and fluvoxamine were greatly increased in synthetic humic waters, indicating photosensitized degradation.
Table 1. Physicochemical properties of SSRIs
Log KOWa
Log KOCb
Photolysis t½c (d)
0.57; 29
aMeasured on salt form (HCl) of each SSRI.
bAverage calculated from experiments with five different soils and sediments.
cAverage calculated from experiments with two different lake water samples (except for fluvoxamine, where both values are presented).
            Experiments to determine water/sediment degradation for SSRIs were conducted according to US EPA methods and in general were in agreement with previous measurements of sorption capacity and photolysis half lives.  In sediment water systems fluoxetine appears to be stable to light in water. Sediments also are an important sink for fluoxetine, as it rapidly adsorbed to pond and creek sediments and sorbed concentrations decreased minimally within 30 days. Nor-fluoxetine, a major degradation product/metabolite, was not detected in any sample. Sertraline can be degraded while in the aqueous phase, but appears to be stable once it is absorbed to sediments. After 30 days of treatment, more than 80 percent of sertraline remained in the sediments, indicating that the sediments are an important sink for sertraline.  Dissipation rates of fluvoxamine from the water phase varied according to light conditions and types of water.  Results indicate that fluvoxamine is somewhat susceptible to light, and the properties of natural water affect its degradation in water. Fluvoxamine residues in sediments were highest in all treatments between days 1 and 3, indicating rapid adsorption.  Concentrations remained constant in sediments, indicating that fluvoxamine likely is stable to microorganisms in sediments. One photoproduct, a photoisomer of fluvoxamine, was detected in water and sediment under light conditions, but decreased to negligible amounts by day 14. Approximately 46% of paroxetine remained in water and sediment under light conditions at day 1 compared to 66-75% remaining in the dark, indicating that paroxetine is degraded by sunlight.  By day 30 of application, very little (0.1-0.2%) of paroxetine remained in water and only 6.0-32.7% remained in sediment, suggesting that paroxetine partitioned to sediment, where it was further degraded in sediment with time.  Dissipation rates of citalopram from the water phase under light conditions were identical to those in the dark, indicating that citalopram is stable to light in water. Although 9-13% of citalopram applied still remained in the creek water phase at day 30, only 0.2 percent remained in the pond water phase at the same day, showing higher adsorption capacity for pond sediments. Of the originally applied citalopram, 45-52% and 52-63% were detected in creek or pond sediments at days 1 and 30, respectively, indicating that citalopram rapidly adsorbed to both sediments and was stable in the sediments.
Ready biodegradability by wastewater treatment was tested with an inoculum from an activated sludge collected from a wastewater treatment plant in Columbus, Mississippi. No degradation was observed in the test period of 28 days for any SSRI.  Thus, it can be concluded that none of these drugs may be classified as readily biodegradable in wastewater treatment plants. No significant hydrolytic degradation was detected for any drug in any aqueous solution.  
Finally, monthly environmental samples (sediment and water) were collected from the vicinity of the wastewater treatment plant (WWTP) in Columbus, Mississippi from April 2005 through March 2006.   The five SSRIs and their two degradation products (norfluoxetine and norsertraline) were analyzed in water samples (upstream, downstream of the WWTP;  influent, effluent from the WWTP) and downstream sediment samples.  In the water samples, the order of detection frequency of SSRIs were: sertraline (88%) > citalopram (50%) > fluoxetine (48%) > norfluoxetine (15%) > norsertraline (10%) > paroxetine = fluvoxamine (2%).  Quantities of each SSRI detected in water samples (ng/L) were as follows: citalopram, 5.7 – 63.7; fluoxetine, 5.6 – 48.0; fluvoxamine, 3.7; paroxetine, 1.6; sertraline, 1.1 – 34.9; norfluoxetine, 3.2 – 31.7; norsertraline, 12.5 – 61.9. Citalopram, fluoxetine, and sertraline were the most frequently detected SSRIs in water samples.  Influent samples had the highest detection frequency, with 51% of the seven target compounds detected, followed by 36% detected in effluent, 20% detected in downstream waters, and18% detected in upstream water. Citalopram was detected in all sediment samples that were collected, ranging from 0.40-2.88 ng/g.  Sertraline was also detected in many sediment samples, ranging from 0.09 to 0.67 ng/g, approximately 4-fold lower than citalopram concentrations.  Fluoxetine was detected at 1.26 ng/g in only one sediment sample.  Results show that the compounds most persistent in laboratory investigations (citalopram, fluoxetine, and sertraline) were most frequently measured in water and sediment environmental samples, and represent a potential threat to aquatic organisms.
II. Toxicity of SSRIs to Aquatic Organisms
Acute and chronic toxicity experiments were conducted on all five SSRIs with the aquatic microcrustacean, Ceriodaphnia dubia. Mortality increased with increasing concentration of each SSRI and the 48-h LC50 values ranged from 0.14 mg/L for sertraline to 3.18 mg/L for citalopram.  The LC50 values for other SSRIs were: fluoxetine, 0.47 mg/L; paroxetine, 0.59 mg/L; and fluvoxamine, 1.26 mg/L.  In chronic tests, the SSRI with the lowest concentration to affect the mean number of neonates produced (LOEC) was sertraline (0.045 mg/L) while the highest LOEC was measured upon exposure to citalopram (4 mg/L).  The time to first brood was not significantly affected after exposure to any of the SSRIs tested; however, the number of broods was reduced after exposure to fluvoxamine, fluoxetine, and paroxetine but not by the other SSRIs at the concentrations tested.  Based on these results and literature values for measured amounts of SSRIs in aquatic environments, acute and chronic toxic effects are not expected in C. dubia at environmentally relevant concentrations. 
A major concern for the ecotoxicology of pharmaceuticals is that exposure to low concentrations of biologically active compounds may exert negative effects on aquatic organisms after multiple generations of exposure.  This issue was investigated in C. dubia, with fluoxetine used as a model compound and reproduction evaluated over four generations.  While results indicated more lethargic swimming in C. dubia exposed to fluoxetine, no differences in reproductive output or timing were detected among generations.
Because organisms are most likely exposed to pharmaceuticals in surface waters receiving effluents from wastewater treatment plants, another important issue is that of toxicity of mixtures of pharmaceuticals. All five SSRIs have been detected in surface waters and all are presumed to act by the same mode of action (as in human medicine), warranting examination of multiple compound exposures of these compounds. The toxicity of binary and quaternary mixtures of SSRIs to C. dubia were investigated and results confirmed that toxicity was additive and most easily predicted with a concentration addition model. These results emphasize the importance of evaluating joint toxicity of mixture components for biologically active compounds such as the SSRIs and other pharmaceuticals.
Experiments were conducted in to determine the acute and chronic toxicity of fluoxetine to the western mosquito fish, Gambusia affinis.   In 7-d acute toxicity tests with neonate (age 24-36 h) G. affinis, the LC50 for fluoxetine was 614 Fg/L.  A 91 d chronic test was conducted with neonate (age 24-48 h) G. affinis based on toxicity data obtained from the acute test, and fish were exposed to 6, 0.6, and 0.06 Fg/L fluoxetine.  Chronic exposure to fluoxetine did not significantly affect survival, sex ratio, or morphology of the male gonopodium.  A second chronic test was conducted in larger tanks (100 L) beginning with older juvenile fish (age 5-6 weeks) and continued until fish were age 140-145 d.  At the high exposure concentration (60 Fg/L fluoxetine), expression of mature sexual characteristics (gonopodium, male fish; black spot on abdomen, female fish) was significantly delayed in western mosquito fish exposed to 60 ppb fluoxetine.
Additional experiments were conducted in the laboratory to investigate the effects of fluoxetine on development of western mosquito fish exposed during specific stages of development.  Fish were exposed to 0, 6, or 60 ppb fluoxetine in three experiments.  In experiment 1, fish were exposed from neonates (age 24-48 h) to an age of 11 d, and then exposure to fluoxetine ended and fish were allowed to develop to an age of 96 d.  In experiment 2, fish were exposed from neonates (age 24-48 h) to age 43 d, and then exposure to fluoxetine was ended and fish were allowed to develop to an age of 96 d.  In experiment 3, fish were exposed for 164 d from an age of 38-42 d to age 202-206 d at which time the experiment was ended.  At pre-selected times during each experiment fish were randomly sampled from treatments for evaluation of morphological features and were preserved for histological examination.  Fish tissues have been sectioned and scored and await secondary scoring/confirmation by a fish pathologist.
Fluoxetine has been reported to inhibit the mammalian thyroid axis.  Because events leading to amphibian metamorphosis are cued by circulating thyroid hormone levels, the African clawed frog, Xenopus laevis has been used as a model organism to assess chemically-induced impacts on thyroid hormone production, through monitoring the time to metamorphosis in exposed individuals. We conducted three separate experiments to investigate the effects of the SSRIs fluoxetine and sertraline on X. laevis growth, development and metamorphosis.   In preliminary experiments, tadpoles exposed to high concentrations of fluoxetine (lowest observable effects concentration, LOEC = 29.5 µg/L) in reconstituted exposure water (FETAX solution) had a 30% increase in time to metamorphosis.  Smaller metamorphs (wet weight basis) were observed at all fluoxetine exposure concentrations (LOEC = 0.59 µg/L), compared with controls.   Subsequent experiments with a lower range of fluoxetine concentrations confirmed that exposure to 1 and 10 µg/L fluoxetine was associated with reduced mass (dry weight) of metamorphs.  In the same experiment all tadpoles exposed to sertraline (LOEC = 0.1 µg/L) had significantly reduced mass at metamorphosis.  Concentrations of both sertraline and fluoxetine exerting an effect on metamorph mass were 3- to 10-fold higher (sertraline) or 15- to 30-fold higher (fluoxetine) than concentrations of the two SSRIs measured in the vicinity of the Columbus, MS WWTP (this study). 
In conclusion, all SSRIs are acutely toxic to C. dubia and western mosquito fish, but only at concentrations far greater than measured environmental concentrations. However evidence of additive toxic effects of SSRIs on C. dubia emphasizes the importance of   quantifying total SSRI concentrations when assessing environmental exposure and subsequent effects.  Transient developmental delays were observed in western mosquito fish, again at concentrations over 100-fold greater than measured environmental concentrations of fluoxetine.  Histopathology results will ultimately reveal whether fluoxetine affected gonad maturation, in addition to external morphology.  Of the organisms tested, X. laevis tadpoles were the most sensitive to SSRI exposure, with significant reductions in mass at metamorphosis in individuals exposed to fluoxetine and sertraline.  Effects concentrations for both sertraline and fluoxetine fall within a 10- to 100-fold safety factor of their measured environmental concentrations, and raise concerns about exposures of these compounds to amphibians in the wild.  Based on our chemical fate studies, these are among the most persistent SSRIs, further emphasizing their potential for causing adverse environmental effects.

Expected Results:

The ultimate benefit of this research will be to provide extensive environmental fate, ecotoxicological, and occurrence information on a class of chemicals for which little data is published in the literature but that have the potential to be present in water bodies, including drinking water sources receiving treated wastewater effluent. This information should allow environmental risk assessments to be conducted for these compounds and could be a model by which pharmaceutical manufacturers could begin to assess their products as they are commercialized. This work may also indicate that these compounds will degrade quickly in the environment and be of no cause for concern or it may indicate that most pharmaceuticals should be tested for environmental and ecological safety with the same rigor as pesticides.

Journal Articles on this Report : 7 Displayed | Download in RIS Format

Other project views: All 44 publications 11 publications in selected types All 9 journal articles
Type Citation Project Document Sources
Journal Article Henry TB, Kwon J-W, Armbrust KL, Black MC. Acute and chronic toxicity of five selective serotonin reuptake inhibitors to Ceriodaphnia dubia. Environmental Toxicology and Chemistry 2004;23(9):2229-2233. R829006 (2003)
R829006 (2004)
R829006 (Final)
  • Abstract from PubMed
  • Journal Article Henry TB, Black MC. Mixture and single-substance acute toxicity of selective serotonin reuptake inhibitors in Ceriodaphnia dubia. Environmental Toxicology and Chemistry 2007;26(8):1751-1755. R829006 (2006)
    R829006 (Final)
  • Abstract from PubMed
  • Journal Article Henry TB, Black MC. Acute and chronic toxicity of fluoxetine (selective serotonin reuptake inhibitor) in western mosquitofish. Archives of Environmental Contamination and Toxicology 2008;54(2):325-330. R829006 (Final)
  • Abstract from PubMed
  • Journal Article Kwon J-W, Armbrust KL. Hydrolysis and photolysis of paroxetine, a selective serotonin reuptake inhibitor, in aqueous solutions. Environmental Toxicology and Chemistry 2004;23(6):1394-1399. R829006 (2003)
    R829006 (2004)
    R829006 (Final)
  • Abstract from PubMed
  • Journal Article Kwon J-W, Armbrust KL. Photo-isomerization of fluvoxamine in aqueous solutions. Journal of Pharmaceutical and Biomedical Analysis 2005;37(4):643-648. R829006 (2004)
    R829006 (2005)
    R829006 (Final)
  • Abstract from PubMed
  • Journal Article Kwon J-W, Armbrust KL. Degradation of citalopram by simulated sunlight. Environmental Toxicology and Chemistry 2005;24(7):1618-1623. R829006 (2005)
    R829006 (Final)
  • Abstract from PubMed
  • Journal Article Kwon J-W, Armbrust KL. Laboratory persistence and fate of fluoxetine in aquatic environments. Environmental Toxicology and Chemistry 2006;25(10):2561-2568. R829006 (2006)
    R829006 (Final)
  • Abstract from PubMed
  • Supplemental Keywords:

    pharmaceuticals, aquatic persistence, reproductive toxicity., RFA, Scientific Discipline, Waste, Water, Ecosystem Protection/Environmental Exposure & Risk, Ecology, Aquatic Ecosystems & Estuarine Research, Environmental Chemistry, Chemistry, Aquatic Ecosystem, Fate & Transport, Environmental Microbiology, Biochemistry, Ecology and Ecosystems, Drinking Water, ceriodaphnia dubia, fate and transport, pharmaceuticals, aquatic organisms, other - risk assessment, selective serotonin reuptake inhibitors (SSRIs), impact of pharmaceuticals, aquatic persistence, aquatic ecosystems, gambusia holbrooki

    Relevant Websites:


    Progress and Final Reports:

    Original Abstract
  • 2002 Progress Report
  • 2003 Progress Report
  • 2004 Progress Report
  • 2005 Progress Report
  • 2006 Progress Report