Water, Air, & Soil Pollution

, Volume 218, Issue 1, pp 413–422

Assessment of Pharmaceuticals Fate in a Model Environment

Authors

    • Departament de QuímicaUniversitat Autònoma de Barcelona
  • Marc Ribera
    • Departament de QuímicaUniversitat Autònoma de Barcelona
  • José Peral
    • Departament de QuímicaUniversitat Autònoma de Barcelona
Article

DOI: 10.1007/s11270-010-0655-y

Cite this article as:
Domènech, X., Ribera, M. & Peral, J. Water Air Soil Pollut (2011) 218: 413. doi:10.1007/s11270-010-0655-y

Abstract

A multiphase model based on the Mackay-type level II fugacity model has been used to predict the behaviour and final environmental concentrations of some of the more consumed pharmaceuticals in Spain. The model takes into account the mean rate of consumption of pharmaceuticals, the percentage of pharmaceutical metabolised, the formation of the corresponding glucuronide, which is assumed to be hydrolysed back to the parent molecule, the partial degradation of each pharmaceutical in a conventional sewage treatment plant, and the fate of these substances in a regional model environmental system. Predicted environmental concentrations in air, water, soil, sediments and suspended matter, and the corresponding residence time for each pharmaceutical have been obtained by application of the model. The predicted concentrations of pharmaceuticals in the water phase are of the same order than the measured experimentally, showing that the simple model used to predict the environmental concentrations is suitable for modelling the environmental fate of high water soluble and low volatile organic compounds such as pharmaceuticals products.

Keywords

Predicted environmental concentrationMultiphase modelPharmaceuticals

1 Introduction

The increasing consumption of pharmaceuticals by humans, which are partially non-metabolised and excreted unchanged into raw sewage, must lead to some concern due to their occurrence in the environment (Khetan and Collins 2007). Actually, pharmaceuticals are considered emerging pollutants because they enter the environment in appreciable amounts, they are moderately persistent and can exhibit deleterious effects in organisms (Khetan and Collins 2007). It is therefore necessary to know the fate of these chemicals once they are released to the environment. The assessment of fate and exposure of a chemical in the environment with the final goal of predicting the concentration of the substance in different environmental compartments is a basic element of any Environmental Risk Assessment (ERA; Maltby 2006; EMEA 2006). Usually, emerging pollutants occur in the environment at very low concentrations and in a wide variety of matrices, making analytical determination very difficult. For this reason, the so-called multimedia fate models have been developed for the purpose of predicting the distribution and concentrations of chemicals once released to the environment (Mackay 1991). The introduction of the fugacity concept in the development of multiphase models (Mackay 1979), has allowed to build tools for ERA (EC 2004; DTSC 1993). Currently, ERA tools, such as the European Uniform Systems for the Evaluation of Substances (EU 2004) or CalTox (DTSC 1993) rely on Mackay-type fugacity models to predict the behaviour and fate of pollutants in the environment. Fugacity models can be built with increasing levels of complexity, from level I, corresponding to equilibrium distribution without transformation, to level IV, including advection and transformation without achieving equilibrium or steady state between phases. Increasing complexity in the models implies handling a larger amount of variables, thus increasing input data requirements. It is up to the user to decide the minimum level of complexity required to fulfil his needs.

The aim of this work is to predict the concentrations of the more consumed pharmaceuticals in the different compartments of an environmental system by means of the use of a level II fugacity model and to compare the obtained results with some available measured concentrations of pharmaceuticals in water in order to validate the proposed multiphase model. The following pharmaceuticals have been analysed: Amlodipine, atorvastatin, bezafibrate, captopril, carbamazepine, celecoxib, ciprofloxacin, citalopram, clarithromycin, clopidogrel, diazepam, diclofenac, diltiazem, erythromycin, fluoxetine, gabapentin, ibuprofen, iopromide, ketoprofen, metoprolol, naproxen, nimodipine, omeprazole, oxazepam, pantoprazole, paracetamol, paroxetine, piroxicam, pravastatin, ranitidine, roxithromycin, sertraline, simvastatin, sotalol, trimethoprim, valsartan, venlafaxine. These pharmaceuticals are the most consumed in Spain, with rates of consumption per inhabitant higher than 1 mg of active substance per year.

2 Model Description

Figure 1 schematically shows the fate of the pharmaceutical once it is consumed. The pharmaceutical is ingested and partly metabolised, giving rise in some cases to a glucuronide as a metabolic product. Due to the high solubility of both the pharmaceutical and its glucuronide they are incorporated into wastewater, part of which is treated in a sewage treatment plant (STP) where the glucuronide is hydrolysed back to the parent pharmaceutical molecule (Ternes 1998). Consequently, the sum of concentrations of the non-metabolised pharmaceutical and the corresponding formed glucuronide has been considered as the concentration of pharmaceutical at the entry of the STP (Ci,in). In this model, an average 80% of the population has been considered to be connected to an STP (EC 2003), the remaining 20% generating wastewaters that are not treated. The effluent from the STP containing the non-biodegraded pharmaceutical, as well as the untreated 20% is directly discharged to the environment.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-010-0655-y/MediaObjects/11270_2010_655_Fig1_HTML.gif
Fig. 1

Scheme of the model system. Pharmaceutical (Ph) is consumed and after being partially metabolised, is incorporated to wastewater, which in part is treated in a sewage treatment plant (STP). Both treated and untreated wastewater fluxes are finally released to the environment. The pharmaceutical contained in both fluxes is partitioned and degraded in the different compartments

Wastewater is assumed to be treated in a plant that uses a conventional activated sludge treatment, consisting of a primary settling tank, followed by an aeration tank and a secondary settling tank. The extent of pharmaceutical removal depends on both physicochemical and biological processes, such as adsorption to activated sludge and biodegradation; volatilisation can be neglected due to the very low value of Henry’s Law constant of all the pharmaceuticals considered in the present work (see Table 1). Assuming that the steady state is achieved in the plant, the concentration at the effluent (Ceff) of each tank is related to the concentration at the inffluent (Cinf), according to Eq. 1 (Salvito et al. 2002):
$$ {C_{\rm{eff}}} = \frac{{{C_{{\inf }}}}}{{1 + {K_d}\cdot {\hbox{SS}}\cdot \frac{\text{HRT}}{{\hbox{SRT}}} + {k_b}\cdot {\hbox{HRT}}}} $$
(1)
Where Kd is the partition coefficient between water and suspended solids (L kg−1), SS is the concentration of suspended solids (kg L−1), kb is the kinetic biodegradation constant (h−1) and Hydraulic retention time (HRT) and sludge retention time (SRT) are the hydraulic retention time and the solids retention time, respectively (h). All these parameters are different for each tank in the STP.
Table 1

Physicochemical parameters of the different pharmaceuticals at 25°C

Pharmaceutical

Kh/atm m3 mol−1

Koc/L Kg−1

kb/h−1

pKa1

pKa2

kOH/h−1

Amlodipine

2.9E-17

198.3

0.083

9.5a

0.97

Atorvastatin

2.4E-23

398

0.14

4.46b

1.2

Bezafibrate

2.1E-15

204.8

0.090

3.75c

0.22

Captopril

3.9E-13

1.98

0.22

3.7a

9.8a

0.48

Carbamazepine

1.1E-10

168.6

0.076

15.44a

0.64

Celecoxib

7.8E-13

2,537

0.029

9.43c

0.078

Ciprofloxacin

5.1E-19

0.99

0.046

6.09b

8.74b

1.7

Citalopram

2.7E-11

1,698

0.026

9.63c

0.50

Clarithromycin

1.7E-29

23.5

0.015

8.99b

2.2

Clopidogrel

2.2E-09

807.4

0.037

5.3b

0.66

Diazepam

3.6E-09

274

0.073

3.4a

0.054

Diclofenac

4.7E-12

404.7

0.053

4.8a

0.89

Diltiazem

8.6E-17

197.8

0.072

7.7a

0.97

Erythromycin

5.4E-29

25.5

0.016

8.88b

2.2

Fluoxetine

8.9E-08

1,505

0.049

10.2a

0.20

Gabapentin

1.8E-10

0.335

0.14

3.68b

10.7b

0.22

Ibuprofen

1.5E-07

224.7

0.13

5.2a

0.064

Iopromide

1.0E-28

0.0213

0.069

6.51a

11.42a

0.35

Ketoprofen

2.1E-11

119.4

0.12

4.6a

0.033

Metoprolol

1.4E-13

29.9

0.094

9.68b

0.79

Naproxen

3.4E-10

93.5

0.15

4.2a

0.63

Nimodipine

2.0E-15

368

0.083

−4.8c

15.45b

0.68

Omeprazole

3.0E-19

871.3

0.058

5.2a

0.52

Oxazepam

5.5E-10

53.4

0.090

1.55b

10.9a

0.75

Pantoprazole

5.8E-20

860.3

0.076

3.92b

8.19b

0.55

Paracetamol

6.4E-13

20.9

0.14

9.38a

0.096

Paroxetine

1.8E-12

1,224

0.076

11.2a

0.88

Piroxicam

2.9E-19

298.6

0.10

6.3b

0.18

Pravastatin

2.0E-15

9.59

0.27

4.23c

1.3

Ranitidine

3.4E-15

13.8

0.044

8.2a

2.0

Roxithromycin

5.0E-31

7.21

0.012

7.66c

1.9

Sertraline

5,1E-08

6,421

0.034

8.74c

0.53

Simvastatin

2.8E-10

935.7

0.10

−4.78a

16.23a

1.2

Sotalol

2.7E-14

4.23

0.097

9.22c

0.55

Trimethoprim

2.4E-14

78.6

0.060

7.2a

1.1

Valsartan

3.1E-18

158.6

0.20

3.81c

5,51c

0.23

Venlafaxine

2.0E-11

208.2

0.032

9.2a

0.69

aYoshida and Topliss 2000.

bToxnet (Accessed 27 April 2010)

cSPARC

As mentioned above, the pharmaceutical reaches the environment by means of treated and untreated wastewaters. The environment considered has been modelled as a standard regional environment made up by five compartments: air, water, soil, sediments, and suspended matter (Fig. 1; EC 2003; see later for system dimensions and other environmental parameters). The only emission of pharmaceutical to the environmental system considered here is the corresponding to the organic dissolved in water, neglecting the emission of pharmaceutical adsorbed in suspended solids in both treated and non-treated wastewater, due to the low concentration of suspended matter in both effluents (Radjenovic et al. 2009).

The environmental system has been modelled taking into account the level II fugacity model (Mackay 1979). Substance-specific physicochemical properties are used to predict the behaviour and distribution between the different compartments, as well as the output flows from the system by means of advection and degradation. In this model, equilibrium between the different phases is assumed and consequently no transfer between compartments is considered.

In steady state, the input flow to the environmental system, I (g h−1), is equal to the sum of output flows, namely degradation (assuming first-order kinetics) and advection in each compartment i (Eq. 2):
$$ I = \sum\nolimits_i {{V_i}{k_i}{C_i}} + \sum\nolimits_i {{G_i}{C_i}} $$
(2)
Where Vi, Ci, ki, and Gi, are the volume (m3), the pharmaceutical concentration (g m−3), the kinetic degradation constant of the pharmaceutical (h−1) and the advective flow of the fluid (air or water) in compartment I (m3 h−1), respectively. If the fugacity capacity of the pharmaceutical in compartment i (Zi, g m−3 atm−1) and the fugacity (f, atm) are introduced, the following equation is obtained:
$$ I = \sum\limits_i {{V_i}{k_i}{Z_i}f} + \sum\limits_i {{G_i}{Z_i}f = f\left( {\sum\limits_i {{V_i}{K_i}{Z_i}} + \sum\limits_i {{G_i}{Z_i}} } \right)} $$
(3)
From Eq. 3, if the input flow of the pharmaceutical is known, the fugacity can be determined, and from the latter, the predicted environmental concentration of the pharmaceutical in compartment i (Ci) can be calculated (Eq. 4):
$$ {C_i} = f \cdot {Z_i} $$
(4)
The fugacity capacity of the pharmaceutical in each compartment of the environmental system is determined from the Z values of each constituting phase of the compartment. Thus, in the water compartment: Zwater = fw · Zw + fsm · Zsm, where fw and fsm are the volume fractions of water and suspended matter in the compartment, respectively, whereas Zsm and Zw are the fugacity capacities of pharmaceutical in suspended matter and water, respectively. For the soil compartment: Zsoil = fa · Za + fw ·Zw + fs · Zs, where fa, and fs are the volume fractions of air and solids, whereas Za and Zs are the fugacity capacities of pharmaceutical in air and soil solids, respectively. For the sediment compartment, which is constituted by solids and porewater: Zsed = fw · Zw + fs · Zs. Finally, for the air compartment, which it is assumed to be only made up by air: Zair = Za. Estimation of Z for solid phases (soil, sediments and suspended matter) is based on the assumption that neutral pharmaceutical molecules adsorb to the organic matter present in these solid phases, according to the expression: Kd = Koc · foc, where Koc is the organic carbon partition coefficient and foc is the fraction of organic carbon in the environmental compartment.
The physicochemical parameters of the different pharmaceuticals considered that are required for the calculations are summarised in Table 1. Due to lack of experimental values of the physicochemical parameters for the pharmaceuticals under study, the values of KH, Koc, kb, and hydroxyl radical reaction rate have been obtained from the recently released version 4.0 of the EPI Suite programme (US EPA 2009). KH values have been estimated from molecular fragments (HENRYWIN programme), whereas Koc have been obtained from KOW values using PCKOWIN programme. For ionic pharmaceuticals, Kh and Koc must be corrected, considering that only the neutral form is actually able to volatilise or to adsorb to organic matter in solids. Thus, if a pH = 7 is considered in both the STP and the environmental system, the correction factor (CF) for acids and bases is calculated according Eqs. 5 and 6, respectively (Sherrer and Howard 1977):
$$ {\hbox{C}}{{\hbox{F}}_{\rm{acid}}} = \frac{1}{{\left( {1 + {{10}^{{7 - {\rm{pKa}}}}}} \right)}} $$
(5)
$$ {\hbox{C}}{{\hbox{F}}_{\rm{base}}} = \frac{1}{{\left( {1 + {{10}^{{{\rm{pKa}} - 7}}}} \right)}} $$
(6)
Some pharmaceuticals have two pKas (see Table 1); in these cases, the ratio of molecules that undergo no dissociation is determined by the pKa value that is more distant to the working pH.

Biodegradation kinetic constants of pharmaceuticals have been obtained from the half-lives estimated by the BIOWIN programme (primary biodegradation; US EPA 2009).

Table 2 displays typical values for hydraulic and solids retention time and the concentration of suspended solids in the STP (EC 2003; Salvito et al. 2002). These values have been used to calculate the concentrations of pharmaceuticals in each tank of the STP, according to Eq. 1.
Table 2

Values of the hydraulic retention time (HRT), sludge retention time (SRT), and suspended solids (SS) of the different tanks of the STP

Tank

HRT (h)

SRT (h)

SS (kg L−1)

Primary settling tank

1

1

0.00045

Aeration tank

7

221

0.0025

Secondary settling tank

1

1

0.0025

Concerning the environmental system, the regional model described in the Technical Guidance Document on Risk Assessment has been used (EC 2003). This model consists of a densely populated area (20 million habitants) of 40,000 km2, with the following compartments: air, water, soil, sediments, and suspended matter in water. In Table 3, the parameters describing these compartments are shown (EC 2003), along with advection residence times in air and water (τi) used to determine the corresponding advective flows in air and water (Gi = Vii).
Table 3

Environmental and compartment-specific parameters

Compartment

fa

fw

fs

V (m3)

fOC

τ (d)

Air

4 × 1013

0.7

Water

3.6 × 109

40

Soil

0.2

0.2

0.6

1.9 × 109

0.02

Sediments

0.8

0.2

3.6 × 107

0.05

Suspended matter

0.9

0.1

2.2 × 105

0.1

fa, fw, fs, are the air, water, and solid fractions of the different compartments; V is the compartment volume, fOC the organic carbon fraction and τ the advective residence time

3 Results and Discussion

Predicted environmental concentrations are estimated from the rates of pharmaceuticals consumption estimated in Spain from the average annual consumption in the period 1999–2006 (KNAPPE 2009; Carballa et al. 2008). The predicted concentrations of pharmaceuticals in wastewater have been estimated from the percentage of pharmaceutical not metabolised and excreted unchanged plus the percentage of glucuronide formed that is hydrolyzed back to the parent pharmaceutical molecule (information in Online Resource 1) and assuming an average production of 200 L wastewater d−1 inh−1 (EC 2003). For those pharmaceuticals for which data on percentages of both not metabolised and glucuronide formed are not known, a 100% of pharmaceutical not metabolised has been considered.

Equation 1 is applied to the three tanks of the STP, allowing us to obtain the steady-state concentrations of dissolved pharmaceuticals in each tank. The concentration of the non-degraded pharmaceutical in the last tank (Ci,out), corresponds to the concentration in the treated effluent leaving the plant. Values for Kd are obtained with the expression Kd = Koc·foc, with foc being 0.4 (Salvito et al. 2002). Since the concentration of microorganisms in each STP tank is higher than in any environmental compartment, the kinetic biodegradation constant is also expected to be higher and must be corrected accordingly. Taking into account the suspended matter concentrations in the tanks (Table 2), the kb value used to predict biodegradation in aeration tank and secondary settling tank is obtained by multiplying by 10 the value of kb reported in Table 1 (Clark et al. 1995), while the kb value in the primary settling tank is obtained by multiplying by 4 the value of kb of Table 1 (US EPA 2000). Finally, the values of Kd and kb for the different pharmaceuticals are used together with HRT, SRT, and suspended matter values in Eq. 1 to obtain the steady-state concentrations of dissolved pollutants in the tanks (Online Resource 2). In Table 4, the predicted concentrations of the pharmaceuticals in the wastewater at the entry of the treatment plant (ci,in) and leaving the plant (ci,out) are shown. The majority of pharmaceuticals are removed with ratios higher than 50%, but few of them, such as the antibiotics clarithromycine, erythromycine and roxithromicine show very low percentages of removal (10–15%; Online Resource 2). In fact, the latter pharmaceuticals are those with an estimated lower biodegradation kinetic constant (Table 1). The main removal route is biodegradation, except for celecoxib, clopidogrel, diazepam, nimipodine, omeprazole, pantoprazole, and simvastatin, for which the adsorption on the suspended matter is the main removal process (Online Resource 2). The latter pharmaceuticals are present in the wastewater in their non-ionised forms and have KOC partition constants higher than 300 L Kg−1 (Table 1).
Table 4

Annual per capita consumption (APC), estimated concentrations of pharmaceutical at the inflow (Ci,in) and at the outflow (Ci,out) of the STP, and predicted (PEClocal) and measured environmental concentrations (MEClocal) on a local scale of the different pharmaceuticals

Pharmaceutical

APC/g y-1

Ci,in/μg L−1

Ci,out μg L−1

PEClocalw/ng L−1

MEClocala/μg L−1

MEClocalb/ng L−1

MEClocalc/ngL−1

Amlodipine

0.013

0.13

0.062

6.1

   

Atorvastatin

0.051

0.70

0.25

22.4

   

Bezafibrate

0.10

0.862

0.41

40.7

0.03/15.06(1.02)

4–37

Captopril

0.024

0.329

0.080

8.0

   

Carbamazepine

0.50

4.66

1.70

56.9

0.08/3.09(1,07)

35/1160

11–90

Celecoxib

0.048

0.50

0.015

0.010

   

Ciprofloxacin

0.091

0.93

0.60

60.2

   

Citalopram

0.04

0.088

0.067

6.1

 

3/120

 

Clarithromycin

0.132

0.81

0.70

69.4

   

Clopidogrel

0.091

1.25

0.20

1.3

   

Diazepam

0.023

0.032

0.0096

0.079

   

Diclofenac

0.81

3.88

2.37

236

0.08/18.74(2.2)

7–50

Diltiazem

0.136

1.36

0.68

38.9

   

Erythromycin

0.203

2.78

2.36

233

0.01/0.57(0.11)

21–71

Fluoxetine

0.105

3.02

0.19

17.5

<20 ng/L

8/44

 

Gabapentin

0.046

0.63

0.22

21.9

   

Ibuprofen

6.9

14.2

5.39

533

0.16/9.89(1.37)

65–289

Iopromide

0.5

6.34

3.46

340

   

Ketoprofen

0.005

0.069

0.027

2.7

0.16/2.71(0.79)

<30–144

Metoprolol

0.058

0.80

0.37

36.5

0.01/0.16(0.05)

  

Naproxen

1.07

11.7

3.90

389

0.02/2.06(0.53)

44–247

Nimodipine

0.033

0.45

0.11

1.6

   

Omeprazole

0.063

0.60

0.082

2.6

   

Oxazepam

0.0008

0.011

0.0046

0.25

 

6/129

 

Pantoprazole

0.018

0.22

0.028

0.34

   

Paracetamol

3.595

29.8

9.98

935

0.06/2.42(0.42)

 

Paroxetine

0.04

0.59

0.29

28.4

<8 ng/L

  

Piroxicam

0.011

0.15

0.059

5.9

   

Pravastatin

0.024

0.33

0.066

6.6

<0.47 ng/L

  

Ranitidine

0,484

3.75

2.46

152

0,01/0.57(0.11)

21–142

Roxithromycin

0.010

0.092

0.080

2.2

   

Sertraline

0.067

0.90

0.048

8.7

   

Simvastatin

0.023

0.23

0.024

0.14

   

Sotalol

0.018

0.25

0.011

11.1

   

Trimethoprim

0.0925

0.70

0.38

26.5

0.02/0.47(0.14)

10–69

Valsartan

0.127

2.02

0.53

51.5

   

Venlafaxine

0.027

0.12

0.083

8.2

 

22/387

 

aMin/max (mean) concentration values of pharmaceuticals downstream STP plants in seven different locations of Llobregat river (Muñoz et al. 2010).

bMinimum and maximum concentration values of pharmaceuticals downstream STP plants in different rivers of Madrid metropolitan area (González et al. 2010).

cMinimum and maximum concentration values of pharmaceuticals downstream STP plants in different locations in the Ebro river basin (Gros et al. 2007)

From the concentration of the pharmaceutical in the effluent of the wastewater plant, and once the treated effluent is spilled into the receiving water, the predicted local concentration (PEClocal,w) of pharmaceutical can be obtained according to the following equation (EC 2003):
$$ {\hbox{PE}}{{\hbox{C}}_{{{\rm{local}},w}}} = \frac{{{C_{{i,{\rm{out}}}}}}}{{\left( {1 + {K_d}.{\hbox{SS}}} \right){\hbox{DILUT}}}} $$
(7)
In this expression, SS is the suspended solid in the receiving water and DILUT is the dilution factor. Values of 15 · 10−6 Kg L−1 and 10 have been used for SS and DILUT (EC 2003). The local concentration corresponds to the concentration of the chemical achieved after complete mixing of the effluent in the receiving waters. Due to the short time passed between effluent discharge and complete mixing, degradation, volatilisation and sedimentation have been neglected as removal processes (EC 2003). The obtained values of local PEC in water for the different pharmaceuticals are summarised in Table 4, together with concentrations of different pharmaceuticals measured downstream of STPs in different Spanish rivers. As it can be seen, the predicted local concentrations in water lie in the ng L−1 level, except for diclofenac, erithromycyne, ibuprofen, iopromide, naproxen, paracetamol, and ranitidine which achieve concentrations higher than 100 ng L−1. In fact, these pharmaceuticals are among the most consumed pharmaceuticals by the Spanish population (Table 4), giving rise to high concentrations in wastewater. As it is showed in Table 4, the predicted local concentrations obtained by means of the application of the model lie in the same range of the measured concentrations in Spanish river waters.
As it has been previously mentioned, 80% of the generated wastewater is assumed to be subject to treatment. This treated effluent, along with the remaining untreated 20% wastewater is discharged to the environment. This means that 40 L wastewater d−1 inh−1 from 200 L wastewater d−1 inh−1, containing pharmaceuticals with a concentration of ci,in are discharged directly without treatment and the remaining 160 L of wastewater d−1 inh−1, with pharmaceuticals at a concentration of ci,out.are discharged once have been treated at the STP. The total input (I) of pharmaceutical to the regional environment system is obtained according to the following equation:
$$ I = \left( {160{C_{{i,{\rm{in}}}}} + 40{C_{{i,{\rm{out}}}}}} \right)\frac{{20\cdot {{10}^6}}}{{24}} $$
(8)
In this equation, the total population (20 millions inhabitants) and 24 h/day factor have been introduced to obtain the total input in g h−1. It has been considered that the sludge generated during the wastewater treatment has been placed in a disposal facility, and the pharmaceuticals contained in the sludge have been assumed not to be liberated to the environment. The obtained values of I (see Table 5), range from 1 g h−1 for oxazepam to 2,357 g h−1 for paracetamol.
Fugacities in the environmental system for the different pharmaceuticals considered are calculated with Eq. 3, using I emission values estimated by means of Eq. 8, and using substance-specific data in Table 1 and the environmental system parameters in Table 3. From the resulting fugacities and the Zi values for the corresponding pharmaceuticals, the predicted environmental concentrations on a regional scale (PECreg) in the different compartments of the environmental system are finally obtained with Eq. 4 (Table 5). Zi values for the pharmaceuticals in each compartment are determined using substance-specific data in Table 1. The predicted concentrations of pharmaceuticals in air are extremely low (well below pg m−3) and they are not summarised in Table 5 because they are not relevant in the discussion. These very low levels of pharmaceuticals in air are consequence of their low Henry’s Law constant values (see Table 1).
Table 5

Total input in the environmental system (I), residence time (τ) and predicted (PECreg) and measured concentrations (MECreg) on a regional scale for the different pharmaceuticals

Pharmaceutical

I/g h−1

τ/h

PECreg/ng L−1

MECreg (water)/ng L−1

water

soil

sediments

suspended matter

Ref. 1

Ref. 2

Ref. 3

Ref. 4

Ref. 5

Amlodipine

12.4

13.5

0.036

0.018

0.038

0.042

     

Atorvastatin

56.8

10.5

0.067

0.18

0.19

0.20

0.80 (1.4)

<0.25

   

Bezafibrate

83.0

11.7

0.240

0.050

0.19

0.22

     

Captopril

21.6

4.8

0.026

0.0052

0.021

0.023

     

Carbamazepine

383

25.6

0.12

4.8

4.0

4.0

4.1 (51)

5.8

<0.8–43.2

0.84

1.37

Celecoxib

18.7

69.2

0.0022

0.66

0.55

0.55

     

Ciprofloxacin

111

22.7

0.63

0.13

0.50

0.57

     

Citalopram

11.8

52.5

0.079

0.17

0.19

0.20

   

0.02

0.04

Clarithromycin

120

67.9

2.0

0.49

1.7

1.9

     

Clopidogrel

67.8

54.4

0.00064

1.9

1.6

1.6

     

Diazepam

2.34

27.7

0.00015

0.033

0.027

0.027

0.43 (0.47)

<0.25

<0,4

  

Diclofenac

445

20.0

2.1

0.61

1.9

2.1

 

<0.25

<0.9–2.5

  

Diltiazem

135

25.4

0.10

1.6

1.4

1.4

1.1 (1.2)

    

Erythromycin

407

67.3

6.2

2.6

6.1

6.7

     

Fluoxetine

35.5

27.4

0.13

0.26

0.30

0.31

0.80 (3.0)

<0.25

   

Gabapentin

50.2

7.3

0.092

0.018

0.073

0.083

     

Ibuprofen

1190

8.8

2.4

0.94

2.3

2.5

  

<0.1–0.6

  

Iopromide

672

16.2

2.4

1.2

2.5

2.7

     

Ketoprofen

5.84

8.7

0.012

0.0031

0.010

0.012

  

<0.3–3.0

  

Metoprolol

75.0

11.1

0.21

0.043

0.017

0.19

   

ND

ND

Naproxen

910

7.0

1.6

0.34

1.3

1.4

0.90 (32)

0.54

<0.1–0.2

ND

ND

Nimodipine

30.0

24.1

0.0030

0.36

0.30

0.30

     

Omeprazole

31.1

33.7

0.012

0.51

0.43

0.43

     

Oxazepam

0,98

20.4

0.00055

0.0092

0.0080

0.0081

     

Pantoprazole

11.0

26.4

0.0010

0.15

0.12

0.12

     

Paracetamol

2324

9.1

3.2

4.9

6.1

6.4

  

<1–210.1

ND

ND

Paroxetine

56.3

13.9

0.19

0.043

0.16

0.18

     

Piroxicam

12.9

10.5

0.032

0.010

0.029

0.032

     

Pravastatin

19.8

4.0

0.019

0.0041

0.016

0.018

     

Ranitidine

453

40.4

0.65

8.1

7.2

7.2

     

Roxithromycin

13.8

161

0.021

1.1

0.92

0.92

     

Sertraline

94.4

57.5

0.029

2.7

2.3

2.3

     

Simvastatin

10.9

19.8

0.00033

0.11

0.091

0.091

 

<1.0

   

Sotalol

23.1

10.8

0.062

0.013

0.050

0.056

     

Trimethoprim

73.4

28.3

0.10

0.87

0.79

0.80

0.80 (11)

1.2

 

ND

0.26

Valsartan

137

5.7

0.16

0.10

0.19

0.20

     

Venlafaxine

14.9

32.8

0.11

0.041

0.11

0.12

   

0.19

0.43

Ref. 1 Pharmaceuticals in source water before being treated in drinking water treatment plants in 19 different locations of USA. Median values are given and maximum values detected in parenthesis Benotti et al. 2009

Ref. 2 Pharmaceuticals in Boulder Basin Lake Mead, NV (USA). The water of Boulder Basin is the primary source of drinking water for the Las Vegas metropolitan area Vanderford and Snyde 2006

Ref. 3 Pharmaceuticals in source drinking water from Hérault watershed (south coast of France) Togola and Budzinski 2008

Ref. 4 and 5 Pharmaceuticals in Lake Ontario: Newcastle Harbour (Ref. 4) and in an open site (Ref. 5; Li et al. 2010)

It can be seen from values summarised in Table 5 that the predicted concentrations are bellow ng L−1 level in all compartments, except for some pharmaceuticals, particularly, for the more consumed ones, such paracetamol, ibuprofen, naproxen, iopromide, and diclofenac. For these pharmaceuticals, the PECreg values obtained are some units of ng L−1 in water as well as in the solid phase compartments. Clarithromicyn achieves a concentration in water of 2 ng L−1 although its rate of consumption is not specially high (Table 4). This relatively elevated concentration in water and also in the solid phase compartments can be ascribed to its low biodegradation constant (Table 1), being responsible of an elevated residence time, 67.9 h. Clarithromicyne is the most persistent pharmaceutical after celecoxib (Table 5). On the other hand, for both carbamazepine and sertraline, significatively higher concentrations in solid phase compartments than in water are predicted, probably as a consequence of their relatively elevated KCO partition values (Table 1). Also, it is interesting to note the case of erythromycine and ranitidine, for which high concentrations are estimated in all compartments in spite of the rate of consumption of both pharmaceuticals, that are relatively low (Table 4). However, ranitidine is excreted 50% unchanged (Online Resource 1), giving rise to a high concentration in wastewater (Ci,in, in Table 4). In the case of erythromycine due to the lack of information of the metabolism of this pharmaceutical, we have considered that 100% of the compound is no metabolised and, correspondingly, the concentration in wastewater is also high (Table 4). For both pharmaceuticals, the biodegradation rate constant is low, giving rise to a low percentage of removal at the STP (Online Resource 2), high input rates in the environment, high residence times and high predicted concentrations in the different compartments (Table 5).

Finally in Table 5, the measured concentrations of some pharmaceuticals experimentally determined by other authors are summarised. The measured values correspond to concentrations in samples taken far away from the influence of urban areas, being more appropriated the comparison with the background regional PEC values. As can be seen, the majority of the measured concentration values lie in the sub ng L−1 level, similarly to the ones obtained for the predicted concentrations in water, thus validating the presented model.

4 Conclusions

The results obtained in this study concerning to the estimation of the environmental fate of pharmaceuticals by using the simple Mackay-type level II fugacity model show that this multiphase model based on the achievement of the steady state for the pharmaceutical, i.e., the input to the environment is equilibrated by its removal by means of biodegradation and advection, is reliable enough for predicting concentrations in different compartments of a regional environmental system. The relatively good concordance between predicted and measured concentrations of pharmaceuticals in water highlights the good quality of the physicochemical parameters of these ionisable organic compounds estimated with the US EPA EPI Suite programme (v 4.00).

From the application of the fugacity model to pharmaceuticals, it can be concluded that their concentrations in the different environmental compartments achieve values in the range of hundredths to some units of ng L−1, except in air where negligible concentrations (well below pg m−3 range) are estimated. Higher values are obtained for predicted local concentrations in water, downstream of emission sources such STPs, after mixing is completed. In these cases, the pharmaceutical concentrations lie well above the ng L−1 range, particularly for the most consumed pharmaceuticals, such as paracetamol, naproxen, ibuprofen, iopromide, diclofenac, erythromycine and ranitidine, that reach concentrations higher than 100 ng L−1, high enough to produce harmful effects to the environment (EMEA 2006).

Acknowledgments

The authors want to thank the financial support received from the Ministerio de Ciencia e Innovación (Spanish Government) through the research project CTQ2008-00178

Supplementary material

11270_2010_655_MOESM1_ESM.doc (69 kb)
ESM Online Resource 1(DOC 69 kb)
11270_2010_655_MOESM2_ESM.doc (72 kb)
ESM Online Resource 2(DOC 71.5 kb)

Copyright information

© Springer Science+Business Media B.V. 2010