Effluent from bulk drug production is toxic to aquatic vertebrates


  • Gunnar Carlsson,

    1. Department of Biomedical Sciences and Veterinary Public Health, Division of Pathology, Pharmacology and Toxicology, Swedish University of Agricultural Sciences, P.O. Box 7028, SE-750 07 Uppsala, Sweden
    Search for more papers by this author
  • Stefan Örn,

    1. Department of Biomedical Sciences and Veterinary Public Health, Division of Pathology, Pharmacology and Toxicology, Swedish University of Agricultural Sciences, P.O. Box 7028, SE-750 07 Uppsala, Sweden
    Search for more papers by this author
  • D. G. Joakim Larsson

    Corresponding author
    1. Department of Neuroscience and Physiology, Sahlgrenska Academy, University of Gothenburg, Box 434, SE-405 30 Gothenburg, Sweden
    • Department of Neuroscience and Physiology, Sahlgrenska Academy, University of Gothenburg, Box 434, SE-405 30 Gothenburg, Sweden
    Search for more papers by this author

  • Published on the Web 2/2/2009.


Very high levels of a range of pharmaceuticals have been reported recently in the effluent from a wastewater treatment plant near Hyderabad, India. The plant serves approximately 90 manufacturers of bulk drugs that primarily are exported to the world market. Fluoroquinolone antibiotics were found at levels that are highly toxic to various microorganisms. Even though milligram-perliter levels of drugs targeting human proteins also have been found, it is difficult to conclude whether these levels are sufficiently high to adversely affect fish or amphibians due to the lack of relevant chronic toxicity data for most human pharmaceuticals. To assess potential effects on aquatic vertebrates, tadpoles of Xenopus tropicalis were exposed to three dilutions of effluent (0.2, 0.6, and 2%) over 14 d, starting at developmental stage 51. Additionally, newly fertilized zebrafish (Danio rerio) were exposed to diluted effluent in 96-well plates for up to 144 h postfertilization (hpf). The tadpoles' body lengths, weights, and developmental stages were recorded, whereas a larger number of sublethal and lethal endpoints were studied in the zebrafish. A 40% reduced growth of the exposed tadpoles was demonstrated at the lowest tested effluent concentration (0.2%), indicating potent constituents in the effluent that can adversely affect aquatic vertebrates. The median lethal concentration (LC50) for zebrafish at 144 hpf was between 2.7 and 8.1% in different experiments. Reduced spontaneous movements, pigmentation, and heart rate were recorded within 48 hpf at 8 and 16% effluent concentrations. Treated effluent from a plant that serves as an important link in the global supply chain for bulk drugs is thus shown to cause adverse effects to aquatic vertebrates even at very high dilutions.


Pharmaceuticals are biologically potent chemicals interacting with specific drug targets at low concentrations. Many of these drug targets are conserved across various organisms [1]. Consequently, even the low environmental levels arising from the usage of certain drugs may result in adverse effects on wildlife [2–4].

For human drugs, excretion via urine and feces has been considered the major route to the environment [5]. Levels found in sewage effluents often span from low nanogram-per-liter to low microgram-per-liter ranges [6–8]. However, it has now been recognized that wastewater from pharmaceutical production sites may be in certain cases sources for much higher environmental concentrations [9–11]. India is the largest supplier of bulk drugs to the world market [10], and one of the largest production centers within India is near Hyderabad. In 2007, we reported on the content of pharmaceuticals in treated effluent from the largest common treatment plant for pharmaceutical industries in this region, which receives wastewater from approximately 90 bulk drug factories [9]. The levels of 11 of the 61 investigated drugs exceeded 100 μL, with ciprofloxacin reaching 31 mg/L. Six of the top 11 pharmaceuticals were fluoroquinolones (i.e., broad spectrum antibiotics targeting bacterial DNA gyrase/topoisomerase). The effluent therefore is expected to exhibit a very high chronic toxicity to bacteria. These high levels of antibiotics also raise concerns that bacteria within the treatment plant and in the recipient may acquire a high degree of antibiotic resistance and that their resistance genes eventually can spread to human pathogens [9,10,12].

Several drugs with a range of different vertebrate drug targets also were detected at high levels in the effluent (e.g., losartan, cetirizine, metoprolol, citalopram, and ranitidine). Because there is a dearth of comprehensive effect data on aquatic vertebrates for the great majority of human pharmaceuticals, it is difficult to conclude that these levels, although much higher than those found in normal sewage effluents [6–8], are sufficiently high to adversely affect fish or amphibians.

In our previous paper [9], we did not investigate the toxicity of the effluent to vertebrates (i.e., organisms that have conserved targets for many of the drugs). In the present study, we wanted to investigate whether or not the effluent from Patancheru Enviro Tech (PETL) could be expected to adversely affect fish and amphibians downstream from the treatment plant.

Zebrafish (Danio rerio) and Xenopus frogs have been used widely as laboratory model species to represent their respective animal classes in toxicology testing. Their maintenance is relatively easy in laboratories, spawning is possible all year round, and eggs can be collected in large quantities.

Embryo toxicity was tested in zebrafish based on the method developed by Schulte and Nagel [13], which includes exposure in microwell plates followed by regular embryo observations at different times. The embryo method has been used in various applications including toxicity tests of single chemicals, sediment, and sediment extracts and evaluations of complex mixtures like sewage effluent [14,15]. The method is under consideration within the International Organization for Standardization and Organization for Economic Cooperation and Development as an alternative to acute tests using adult fish [16].

Toxicity tests on amphibian tadpoles offer unique opportunities to assess effects on apical characteristics during the sensitive metamorphosis process, including disruptions of the thyroid hormone system. Xenopus laevis has been the primary amphibian species used for such developmental toxicity studies [17,18], but several studies with its relative Xenopus tropicalis also have shown promising results [19–21]. Because both species share many characteristics, the latter species provides some advantages for use as laboratory species such as higher fecundity, smaller size, higher development rate, and shorter generation time [19–20].


Test samples

Descriptions of the PETL treatment plant, water sampling procedures, and chemical characteristics of the effluent can be found in Larsson et al. [9]. Samples taken simultaneously with those from our previous study were thawed, and parts of them were sterile-filtered through a 0.22-μm filter (Corning). Unfiltered, treated effluent from PETL was tested using zebrafish embryos, whereas filtered effluent was tested using both zebrafish embryos and Xenopus tadpoles.

Tadpole test

The water used for the frogs throughout the study (frog water) was a mixture of 25% carbon-filtered tap water and 75% deionized water. The water was kept at 26°C and continuously aerated before use. We obtained X. tropicalis tadpoles that were 2 d postfertilization (dpf) as described in Pettersson et al. [22]. Briefly, one male and one female were injected with human chorionic gonadotropin (dissolved in 0.9% NaCl solution) in the dorsal lymph sac. Breeding took place within a few hours, and the eggs were left to develop after parental frogs were removed from the breeding aquarium. Prior to the onset of the experiment, the tadpoles were housed in 10-L aquaria with frog water at a density of 50 individuals per aquarium; 75% of the water was replaced at 7 and 10 dpf. Tadpoles were fed at a rate of approximately 8% of their body weight per day with a suspension of Sera Micron (Sera) in frog water.

The tadpole test was based on Carlsson et al. [21]. At 13 dpf, tadpoles in stage 51 (staging according to [23]) were selected and placed individually in plastic jars containing 400 ml of exposure medium. Both the distribution of the individuals and the jar positions were random. The exposure medium treatments consisted of filtered effluent that was diluted with frog water to 0.2, 0.6, and 2%. Further, a control group with only frog water and a propylthiouracil (PTU; 75 mg/L) group were included. Propylthiouracil was added as a positive control group to inhibit the thyroxin synthesis. Specific inhibition of the developmental stage and reduced hind limb length growth but no effects on the overall body growth were expected in the PTU-treated tadpoles [20]. There were eight replicates for each effluent concentration, 12 for the frog water control and six for the PTU group. During the experiment, the exposure medium was renewed three times per week (Monday, Wednesday, and Friday), and the tadpoles were fed Sera Micron suspension daily, corresponding to approximately 8% of the control animals' body weights. All animals were given the same amount of food. Mortalities and presence of malformations were recorded daily. Tadpoles were euthanized in tricaine methane-sulfonate (MS 222; Apoteket AB) after 14 d of exposure, and body weight and developmental stage were recorded. Each individual also was photographed with a digital camera, and body length and hind limb length were measured with an accuracy of approximately ±0.2 mm on the photographs using Adobe Photoshop®.

Zebrafish embryo test

Unfiltered and filtered waters were tested in parallel using zebrafish embryos. Standardized water was prepared according to a method by the International Organization for Standardization for rearing and breeding the adult zebrafish and for diluting the water composition samples for embryo exposure [24]. Adult male and female zebrafish were mixed in groups of 15 to 20 individuals and placed in stainless steel cages within three aquaria. Parental fish and exposed embryos were kept at a photoperiod of 12:12 h of light/darkness. The onset of light in the morning initiated breeding, and eggs were collected 30 min later. Eggs collected from the different aquaria were mixed and rinsed of debris. To start the exposure as soon as possible, the eggs were introduced immediately into 100-ml beakers containing 50 ml of the appropriate effluent dilution. Unfertilized eggs then were removed with the criterion that embryos should have reached at least the four-cell stage to be included in the test. Fertilized eggs were transferred individually to wells of 96-well plates including 250 μl of diluted effluent or clean water. The effluent treatment concentrations were 1, 2, 4, 8, and 16% effluent. Each treatment consisted of 16 replicates, and there were 32 replicates for the controls with standardized water only. The well plates were covered with parafilm and placed in a temperature-controlled room at 26°C. Observations of lethal and sublethal endpoints (Table 1) were made with a stereomicroscope at 24, 48, and 144 hpf. A time-lapse camera photographed the well plates at 1-hour intervals between 48 and 144 hpf. The photographs were examined visually to determine the hatching time for each individual. Hatched fish that were still alive at the end of the exposure period were fixed in phosphate-buffered formalin and photographed at a later time. The same photographic setup (distance and lens focal length) was used for all groups, so a direct comparison of body lengths could be made (Adobe Photoshop). Pigmentation measurements at 48 hpf were scored according to an ordinal scale (1 = full pigmentation; 2 = decreased pigmentation on body and normal in eyes; 3 = decreased pigmentation on body and in eyes; 4 = no pigmentation). To verify the results, two additional tests were performed on unfiltered effluent water, and one additional test was performed on filtered effluent water, where 144 hpf mortality was the only endpoint.


Continuous data were analyzed using one-way analysis of variance (ANOVA), with Dunnett's post-hoc test comparing exposed groups to control groups. Prior to the ANOVA, Levene's tests were used to check the homogeneity of variance, and square transformation of the data was performed where necessary to meet the ANOVA requirements of normal distribution and equal variance. Ordinal data were analyzed using the Kruskal-Wallis test followed by the Mann-Whitney U test with Bonferroni adjustment of p values. The LC50 values were calculated from categorical data by linear regression from probit-transformed proportions. Significance level was set to <0.05, and all data were analyzed using Minitab 15 (Minitab).

Table Table 1.. Description of observed endpoints and effect concentrations (median effective concentration, EC50; median lethal concentration, LC50; no-observed-effect concentration, NOEC) at different observation times in the zebrafish embryo test for unfiltered and filtered effluent from Patancheru Enviro Tech (PETL). Lethal endpoint includes coagulation and lack of heartbeats. Sublethal categorical endpoints include not full eye development, not full tail extension, lack of circulation, presence of edema, and no hatching at 144 hours postfertilization (hpf)
   EC50/LC50/NOEC (%)
EndpointsTime (hpf)Type of response variableUnfilteredFiltered
Lethal or sublethal24CategoricalEC50 =16EC50 = 17
Lethal48CategoricalLC50 = 8.0LC50 = 11
Lethal or sublethal48CategoricalEC50 = 8.2EC50 = 11
Lethal144CategoricalLC50 = 4.4LC50 = 8.1
Lethal or sublethal144CategoricalEC50 = 3.6EC50 = 7.8
Lethal (test 2)144CategoricalLC50 = 3.1LC50 = 2.7
Lethal (test 3)144CategoricalLC50 = 5.3
Pigmentation (score 1–4)48OrdinalNOEC = 4NOEC = 8
Movements (n/min)24ContinuousNOEC = 4NOEC = 8
Heart rate (beats/min)48ContinuousNOEC = 8
Hatching time (hpf)48–144Continuous
Body length (mm)144Continuous


Tadpole test

Body weight was reduced significantly in all groups of tadpoles exposed to the effluent (Fig. 1a). In the 0.2, 0.6, and 2% dilution groups, body weights were reduced to 60, 59, and 43%, respectively, compared with the body weights of tadpoles in the control group (p < 0.001; Dunnett's test). In addition, body length was reduced significantly in all treatment groups (to between 74 and 86% of that of the controls) in a dose-dependent manner (p < 0.001; Dunnett's test), as seen in Figure lb. The hind limb length was also shorter for all groups of tadpoles exposed to effluent, as seen in Figure 1c. In the two lowest concentration groups, the hind limb lengths were reduced to 75% of the control measurement (p < 0.01; Dunnett's test), whereas in the highest concentration group the length was 53% of that of the controls (p < 0.001; Dunnett's test). Tadpoles exposed to a 2% dilution of effluent water were significantly less developed based on the developmental stages reached after 14 d of exposure compared with those of the control (Fig. 1d) (p < 0.001; Mann-Whitney U test with Bonferroni correction). The stages ranged between 56 and 58 in the control group, whereas the tadpoles in the group exposed to 2% effluent water ranged between 54 and 56. No significant effects on the developmental stage were observed in the two lower concentration groups.

Body weight and body length in the PTU group were not significantly different from those of the control group (Fig. 1a and b). However, hind limb length was significantly reduced to less than half compared with that of the control group (p < 0.001; Dunnett's test) (Fig. 1c); further, a significant effect on development was observed (p < 0.01; Mann-Whitney U test with Bonferroni correction). The stages ranged between 54 and 55 in the PTU group (Fig. 1d).

Mortality in the experiment was low. Only one individual from the 0.2% treatment and one from the 0.6% treatment died, both after 2 d of exposure, and no mortality was recorded in the control group. No malformations were recorded in any individual.

Zebrafish embryo test

The results from the exposure of zebrafish embryos to unfiltered and filtered effluent water are summarized in Table 1. Mortality at 144 hpf in the control groups in the various tests ranged between 0 and 9.4%. The 144 hpf LC50 concentrations ranged between 3.1 to 5.3% for the unfiltered effluent water (average 4.3%) and between 2.7 to 8.1% (average 5.4%) for the filtered water.

Exposure to unfiltered water resulted in a reduced number of movements at 24 hpf in both the 8 and 16% treatments. Control zebrafish performed 3.8 movements per minute, whereas fish in the 8 and 16% dilution groups moved 1.8 and 0.4 times per minute, respectively (p < 0.001; Dunnett's test). Pigmentation also was reduced in these treatment groups at 48 hpf. All controls but one were classified with Score 1 (full pigmentation). In the 8% group, 13 of 16 individuals were classified as Score 2 (decreased pigmentation on body and normal in eyes) (p < 0.001; Mann-Whitney U test with Bonferroni correction), and individuals in the 16% treatment were all Score 4 (no pigmentation). Effects such as edemas and lack of circulation also were observed at 48 hpf in these groups. Further, the heart rate was reduced to 72% of the rate in the control group at 48 hpf in the 16% treatment (p < 0.001; Dunnett's test).

The number of movements at 24 hpf in embryos exposed to 16% filtered effluent water were reduced significantly (1.31 movements per minute; p < 0.001; Dunnett's test) compared with those of the control (3.8 movements per minute). Further, pigmentation at 48 hpf was reduced in this treatment. Half of the individuals in the 16% group were classified as Score 2 (p < 0.01; Mann-Whitney U test with Bonferroni correction). Additional sublethal responses from the filtered water at 48 hpf included edema and lack of hatching up to day 6.

In the tests of both filtered and unfiltered effluent waters, no effects were recorded on body lengths of hatched alive individuals (3.8 ± 0.2 mm) at the termination of the test or on hatching time (60.3 ± 8.2 hpf) in any of the treatment groups (p = 0.16, p = 0.11, respectively; one-way ANOVA).


Exposure to the PETL effluent resulted in markedly reduced growth of tadpoles at a dilution as low as 0.2%. Compared with the present zebrafish embryo test and previously conducted short-term tests exposing Vibrio fischeri, Lactuca sativa, and Daphnia magna to the same effluent [9], adverse effects on tadpoles were recorded at a considerably lower concentration. It should be emphasized that all tested concentrations resulted in reduced tadpole body growth. We therefore cannot exclude the possibility that even lower concentrations of effluent might result in some degree of vertebrate toxicity. Taking into account the large amount of effluent released per day (1,500 m3) from PETL and a toxicity at a dilution of 1:500, it would be sufficient to contaminate a volume of 750,000 m3 of water per day to a degree that strongly affects the development of aquatic vertebrates. The flow in the Isakkavagu, Nakkavagu, and Manjira rivers are highly variable over the year due to monsoon and drought, so a dilution factor is difficult to estimate. In a separate study [25], we estimated the initial dilution in Isakkavagu in March 2008 to be approximately 1:5 and an additional twofold dilution when joining the Nakkavagu; however, this is just a snapshot of the dilution situation. The environmental concentrations of contaminants downstream from the treatment plant should be investigated further.

Figure Fig. 1..

Body weight (a), body length (b), hind limb length (c), and developmental stage (d) of Xenopus tropicalis tadpoles after 14 d of exposure to treated effluent from Patancheru Enviro Tech (PETL) [9] at 0.2, 0.6, and 2% dilutions including control (C) and 75 mg/L propylthiouracil (PTU). Values are presented as mean ± standard deviation (a, b, c), and as a box plot where boxes show 25th to 75th percentiles and bars show 5th and 95th percentiles. Asterisks denote significance levels of p < 0.01 (**) and p < 0.001 (***) analyzed using Dunnett's test (a, b, c) and Mann-Whitney U test with Bonferroni correction of the p values (d).

In previous work [9,10], we stressed the risks of releasing high levels of fluoroquinolones from PETL on the local microbial communities and the promotion of resistant bacteria; the present study shows that direct effects on aquatic vertebrates also are to be expected. Investigations on the fauna in the river system therefore are warranted, as well as parallel measures to reduce the toxicity of the effluent. Manjira feeds the Godavari River, India's second largest river. Because this water is utilized heavily for irrigation purposes, the potential effects on terrestrial life are also of concern.

A major challenge for the future is to identify toxic components in the effluent to determine appropriate measures to reduce the environmental impact. In addition to the pharmaceuticals reported earlier [9], we also have let a contract laboratory (Analycen AB) perform a directed search for more than 200 different organic solvents, metals, pesticides, and other persistent pollutants in this effluent. The great majority of the chemicals were below detection limits, and to our best knowledge no individual nonpharmaceutical chemical was found at levels that easily could explain the toxicity observed here (Supporting Information; http://dx.doi.org/10.1897/08-524.S1). In absolute concentrations, the analyzed pharmaceuticals dominated over all other analyzed substances. In contrast with most other chemicals (excluding pesticides and preservatives), pharmaceuticals are intentionally biologically potent compounds. Consequently, it is reasonable to hypothesize that some of the pharmaceuticals in this effluent could cause significant toxicity. Gunnarsson et al. [1] reported a strong conservation of drug targets between humans and aquatic vertebrates, including zebrafish and X. tropicalis. Thus, it is likely that drugs with human targets could affect specific physiological processes in fish and amphibians at concentrations much lower than those required for most chemicals to induce other forms of nonspecific toxicity, such as narcosis. The effects on growth of X. tropicalis at 0.2% effluent support the hypothesis that there are sufficiently high levels of human drugs in this effluent to adversely affect aquatic vertebrates in the recipient.

In water containing a complex mixture of known and unknown chemicals, interaction effects also should be under consideration. A large number of pharmaceuticals with different modes of action are present in high concentrations in the investigated effluent [9]. Most toxicity studies on pharmaceuticals used for environmental risk assessments are performed using single-substance exposure, which may lead to underestimating the actual environmental impact. Exposure for such a complex and potent effluent as that investigated herein is likely to result in mixture toxicity. Nevertheless, as a starting point, we will discuss briefly the potential of some of the known drugs in this effluent to cause the observed effects. Of course, we cannot rule out other undiscovered but highly potent compounds in the effluent that could be driving or contributing to the toxicity.

The fluoroquinolones constituted the most abundant group of analyzed pharmaceuticals in the effluent [9]. Indeed, the largest peak in the total ion chromatogram of the effluent came from ciprofloxacin (up to 31 mg/L), suggesting that ciprofloxacin was one of the most common low-molecular-weight compounds in the effluent (N. Paxéus, GRYAAB, Gothenburg, Sweden, personal communication). Fluoroquinolones, including ciprofloxacin and ofloxacin, have low biodegradability [26]. They are useful as antibiotics because they have much higher affinity for bacterial DNA gyrase/topoisomerase than to vertebrate topoisomerase. Because of this discrepancy in affinity, the toxicity of most fluoroquinolones seems to be comparably low for aquatic vertebrates. Richards and Cole [27] recorded no effects for up to 100 mg/L of two fluoroquinolones, ciprofloxacin and levofloxacin, in a frog embryo teratogenicity assay Xenopus (FETAX). Halling-Sørensen et al. [28] did not record any adverse effects in zebrafish exposed up to 100 mg/L of ciprofloxacin for 72 h. Robinson et al. [29] investigated the effects of seven fluoroquinolone antibiotics using fathead minnow (Pimephales promelas) early life-stage toxicity tests. Four of the tested fluoroquinolones, ciprofloxacin, lomefloxacin, ofloxacin, and enrofloxacin, were detected as major constituents of the PETL effluent [9]. However, none of these four drugs showed adverse effects at the tested concentration of 10 mg/L [29]. The available data do not indicate that fluoroquionolones are causing the reduced growth of the tadpoles.

Another perhaps more plausible hypothesis is that the tadpoles are affected by drugs in the effluent that have vertebrate primary targets [1]. Metoprolol is a β-adrenoreceptor antagonist present at almost 1 mg/L in undiluted effluent from PETL [9]. Exposure to metoprolol in 3- to 4-d-old Japanese medaka (Oryzias latipes) resulted in a 48-h LC50 value of 24.3 mg/L [30], which is well above the concentrations measured in the effluent. However, in the same paper, growth was reduced already at 500 μL. Body length was not reduced in zebrafish in the present study. In contrary, frog tadpoles showed a clear body growth reduction when exposed to a dilution of 0.2% PETL effluent. This dilution corresponds to approximately 1.6 μL metoprolol. Triebskorn et al. [31] recorded ultrastructural changes in liver in rainbow trout (Oncorhynchus mykiss) exposed to 1 μL metoprolol for 28 d. This is just above half of the estimated concentration of metoprolol in the tadpoles exposed to 0.2% effluent. Lower dilutions of PETL effluent that resulted in late mortality of zebrafish embryos also resulted in reduced heart rate, edemas, and absent circulation at 48 hpf. These dilutions (8–16%) correspond to approximately 64 to 128 μg metoprolol/L. Larsson et al. [32] showed that short-term exposure of rainbow trout to 70.9 μL propranolol, another more lipophilic β-blocker, did not affect heart rate in adult rainbow trout, but an intravenous injection of 2 mg/kg propranolol hydrochloride did. This suggests a functional conservation in fish of β-adrenoreceptors and their role in controlling heart rate, but the bioconcentration of propranolol may not be sufficient to reach potent concentrations at the receptors. The zebrafish embryos also appear to regulate their heart rate through adrenoreceptors, because propranolol affected heart rate at 2 mg/L in 56-h-old embryos exposed for 8 h [33]. Taken together, the estimated exposure levels of metoprolol appear to be sufficient to have some effect on aquatic vertebrates even at the lowest dilution tested, and we therefore cannot exclude the possibility that metoprolol has contributed to some of the effects observed both in the tadpoles and in the zebrafish.

Comparing the results from the two species, the tadpoles showed at least a tenfold higher sensitivity. Body length, an endpoint measured in both species, was reduced in the frogs at only 0.2%, whereas this measure was not affected at any of the tested concentrations in the zebrafish. One main difference between the tests is that the zebrafish embryos were exposed prior to external feeding, so they were still dependent on the yolk as a nutrient source, whereas tadpoles were exposed during an active feeding period. Thus, the large reduction in body growth parameters seen in tadpoles might be caused by reduced feed assimilation or altered foraging behavior. A strong and early effect in the current zebrafish test was the reduction of spontaneous movements. This was recorded at 24 hpf, and the individuals in these treatments died later in the test. Movement reduction and behavior alterations could be the result of a general toxic effect but also could be a consequence of exposure to specific, neurologically active drugs. The selective serotonin reuptake inhibitor (SSRI) fluoxetine resulted in reduced growth and feeding rate in juvenile fathead minnow, with effects seen at concentrations as low as a few micrograms per liter [34]. In the present study, fluoxetine was not measured as a major contaminant. However, another SSRI, citalopram, was present at 770 to 840 μL [9] and also has been reported to bind to serotonin transporter sites in fathead minnow brain [35], indicating that fish are target organisms for SSRIs. The dose-response relationships for citalopram on behavioral endpoints in aquatic vertebrates is, however, not yet sufficiently established to estimate its potential contribution to the effects observed here.

The angiotensin II receptor antagonist losartan was found at the second highest concentrations among the pharmaceuticals detected in the PETL effluent [9]. In fish, losartan often acts like partial agonists or inhibitors at high concentrations. The agonistic effects of losartan binding in trout are reported as uncharacteristic of the mammalian receptors and indicate differences between the two vertebrate receptors [36]. To our best knowledge, relevant controlled dose-response experiments with losartan are lacking.

In the present study, tadpoles were exposed to sterile-filtered effluent to minimize the risk of effects due to bacterial infections. This also means that particle-bound chemicals might have been removed from the filtered waters. The zebrafish embryo test showed that the toxicity of both unfiltered and filtered water was in the same range, demonstrating the relevance of using filtered water for the tadpole exposure. If anything, the sterile filtration could have led to an underestimation of the toxicity compared with that of actual field conditions. The method used on X. tropicalis tadpoles is optimized to reveal thyroid disruption but also can be applied to identify other forms of lethal and sublethal toxicities such as malformations and reduced growth caused by other mechanisms than thyroid disruption. The developmental stages correspond to the overall developmental progress in Xenopus. This development is interrupted after exposure to certain thyroid hormone inhibitors [37]; the growth of the hind limb is thyroid-hormone-dependent as well [38]. Endpoints, such as body weight and body length, are important to include to distinguish between decreased developmental rates caused by thyroid disruption and other modes of toxicity. Tadpoles exposed to all concentrations of PETL effluent had reduced hind limb length, and in the highest concentration group the developmental stage was reduced as well. However, body weight and body length were reduced in a similar manner, in contrast to the results for tadpoles exposed to the thyroid synthesis inhibitor PTU. This indicates that the decrease in the rate of the metamorphosis process caused by the effluent is probably not due to a disruption of the thyroid system but is mediated more likely through other modes of toxicity. The main endpoints included in the zebrafish embryo test are lethalities such as coagulation and lack of heartbeats; however, sublethal endpoints that could give indications of mechanisms of toxic response have been included frequently as well [33]. The test is terminated normally after 48 hpf, but further extension of the test can be made to include posthatching exposure and measurements, such as hatching time, presence of spinal deformations, pericardial area, and body or tail length [33,39]. In the present study, the inclusion of early, sublethal endpoints did not actually increase the sensitivity of the test compared with later observations of mortality.

We have investigated the toxicity of the effluent from a major treatment plant that receives process water from a large number of suppliers of bulk drugs to the international market [40]. There is a strong effect of the effluent on the growth of tadpoles at concentrations as low as 0.2%, implying that widespread effects on the local aquatic vertebrate fauna are plausible. The present study calls for further efforts to mitigate the pollution due to pharmaceutical production in the Hyderabad area.


This supplementary information contains analytical data reported by a contract laboratory (Analycen AB) on the same effluent investigated in the present study.

Found at DOI: 10.1897/08–524.S1 (47 KB PDF).


Cecilia Berg (Department of Physiology and Developmental Biology, Uppsala University) kindly provided X. tropicalis tadpoles. We thank the Swedish Foundation for Strategic Environmental Research (MISTRA), the Swedish Research Council (VR-Medicine), and the Swedish Research Council for Environment, Agricultural Sciences, and Spatial Planning (FORMAS) for financial support.