w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Impact of in-sewer transformation on 43 pharmaceuticals in a pressurized sewer under anaerobic conditions a,b, Aleksandra Jelic a, Sara Rodriguez-Mozaz a,*, Damia Barcelo a Oriol Gutierrez a
Catalan Institute for Water Research (ICRA), Scientific and Technologic Park of the University of Girona, Emili Grahit 101, 17003 Girona, Spain b Institute of Environmental Assessment and Water Research (IDAEA-CSIC), Jordi Girona 18-26, 08034 Barcelona, Spain
article info
abstract
Article history:
The occurrence of 43 pharmaceuticals and 2 metabolites of ibuprofen was evaluated at the
Received 5 March 2014
inlet and the outlet of a pressure sewer pipe in order to asses if in-sewer processes affect
Received in revised form
the pharmaceutical concentrations during their pass through the pipe. The target com-
31 July 2014
pounds were detected at concentrations ranging from low ng/L to a few mg/L, which are in
Accepted 20 September 2014
the range commonly found in municipal wastewater of the studied area. The changes in
Available online 7 October 2014
concentrations between two sampling points were negligible for most compounds, i.e. from 10 to 10%. A higher decrease in concentrations (25e60 %) during the pass through
Keywords:
the pipe was observed for diltiazem, citalopram, clarithromycin, bezafibrate and amlodi-
Pharmaceuticals
pine. Negative removal was calculated for sulfamethoxazole (66 ± 15%) and irbesartan
Wastewater
(58 ± 25%), which may be due to the conversion of conjugates back to their parent
Sewage system
compounds in the sewer. The results show that microbial transformation of pharmaceuticals begins in sewer, albeit to different extents for different compounds. Therefore, the in-sewer transformation of pharmaceuticals should be assessed especially when their concentrations are used to estimate and refine the estimation of their per capita consumption in a catchment of interest in the sewage epidemiology approach. © 2014 Elsevier Ltd. All rights reserved.
1.
Introduction
Pharmaceutical compounds (PhACs) have been given much attention in the last decade due to their documented presence in the environment and concern about their potential adverse effects to humans and wildlife. This was largely owing to the
* Corresponding author. Tel.: þ34 972 18 33 80. E-mail address: [email protected] (S. Rodriguez-Mozaz). http://dx.doi.org/10.1016/j.watres.2014.09.033 0043-1354/© 2014 Elsevier Ltd. All rights reserved.
development of advanced analytical techniques and instruments, which enabled the trace-level detection of these compounds in different environmental compartments including surface and groundwater, sediment and soil (Ku¨mmerer, 2009). The issue of “pharmaceuticals in the environment” has only been covered by regulations developed by the Swiss environmental agency for Switzerland (Eggen
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
et al., 2014), but it has been recognized as an emerging problem in the Directive 2010/84/EU and Regulation 1235/2010 on pharmacovigilance that recommend the monitoring and evaluation of the risk of environmental effects of medicinal products. Within the Urban wastewater system (UWS), wastewater treatment plants (WWTP) have been identified as a point of discharge of pharmaceuticals into the environment. Numerous studies have been conducted to assess the magnitude of pharmaceuticals that reach the environment through the wastewater and sludge discharge, and to evaluate the efficiency of typical wastewater treatment plants regarding the removal of these compounds from raw wastewater (Jelic et al., 2011; Joss et al., 2005; Radjenovic et al., 2009). The wastewater quality and its pharmaceuticals content, particularly, have been typically monitored at the influent and/or the effluent of a treatment plant, where these micropollutants were detected at concentrations ranging from low ng/L to a few mg/L (Verlicchi et al., 2012). A wide variation in removal efficiencies has been reported for individual compounds in separate studies as well as across therapeutic classes and treatment processes (Onesios et al., 2009). The actual and major source of PhACs is households, upstream of the catchment, where pharmaceuticals are consumed and eventually released to the sewer systems. Sewers are underground infrastructure composed by drains, pipes and pumping stations that transport wastewater to the treatment facilities (Hvitved-Jacobsen, 2002). Two types of sewer pipes are found depending on the topography: 1) gravity pipes, where wastewater flows naturally from higher to lower elevation in partially-filled sections, and 2) pressure pipes (force mains) e completely-filled tubes e with a pump station at their inlet to mechanically push sewage through flat and sloping terrains. Sewers have been traditionally considered as just a sewage collection and transport system, but in fact they are also active chemical and biological reactors (Huisman et al., 2002; Rudelle et al., 2011; Schilperoort et al., 2012). In gravity sewers, having a relatively long transport time, a significant reduction in the amount of biodegradable substrate and production of biomass may occur under aerobic conditions (Hvitved-Jacobsen et al., 2002). Completely-filled pressured pipes are prone to development of septic conditions which make them preferred grow-locations for anaerobic biofilms that generate detrimental compounds such as sulfide (H2S), volatile organic sulfur compounds (VSOC) and methane (CH4) (Guisasola et al., 2008; Gutierrez et al., 2010; HvitvedJacobsen et al., 2013; Sutherland-Stacey et al., 2008). Micropollutants in sewers are still scarcely reported in the literature. Microbial and chemical processes that occur in sewers have been generally neglected. This issue has been recognized in the field of sewage epidemiology, which basically aims at estimating the actual illicit drug use and trends from the concentrations detected in influent wastewater (Daughton, 2001). It has been emphasized that the main limitation of the approach, which directly influences its accuracy, is the lack of comprehensive data on drug marker stability in sewer systems and during storage (Castiglioni et al., 2012; Khan and Nicell, 2011; van Nuijs et al., 2011). The stability of illicit drugs during sample storage has been assessed in several studies (Baker and Kasprzyk-Hordern, 2011; Castiglioni et al.,
99
2011; van Nuijs et al., 2012). Lai et al. (2011) quantified the overall uncertainty of the sewage epidemiology approach and included five prescription pharmaceuticals (atenolol, carbamazepine, gabapentin, hydrochlorothiazide and venlafaxine) to refine the estimation of the number of people that contributed to wastewater in the monitored catchment. However, possible in-sewer transformation of the pharmaceuticals was not considered. Only Thai et al. (2014) have studied how insewer transformation processes alter the concentrations of organic pollutants, in particular illicit drugs and their metabolites, under different conditions, at laboratory-scale sewer reactors. They found that the biofilms in the lab-scale sewer have significantly enhanced the degradation rate of some compounds (i.e. cocaine, and a metabolite of heroin, 6-acetyl morphine) in comparison to the rates in wastewater only, and they suggested that the fate of illicit drugs will depend on sewer conditions. As mentioned before, pressure pipes are the most biologically-active sections of sewer networks due to higher biofilm presence and biofilm-sewage contact (HvitvedJacobsen, T., 2002). Under anaerobic conditions in completely-filled pipes, septic biofilm grows attached to the sewer walls. Sewer biofilm captures the compounds dissolved in wastewater and change their concentrations in the flowing sewage water (Huisman et al., 2002). In addition, for a similar length of pipe, hydraulic retention time (HRT) in pressure pipes is typically longer than in gravity sections since pressure pipes need to be completely filled and act as a plug flow reactor needing higher volumes to reach similar distances. This can lead to higher biofilm-sewage contact time that can promote the transformation processes. To date, there has been no comprehensive study on whether in-sewer biotransformation of pharmaceuticals occurs and to which extent. In this work, we aimed at assessing the occurrence and the extent of biotransformation of 43 pharmaceuticals and 2 metabolites of ibuprofen in pressure sewer pipes. The study was carried out at a flat 7620 m long pipe transporting domestic wastewater from the municipality of Sant Pere Pescador (Catalonia, Spain). The target compounds belong to commonly prescribed and over-the-counter pharmaceuticals (i.e. nonsteroidal anti-inflammatory agents and analgesics, lipid modifying agents, psycholeptic and antiepileptic drugs etc.) and were selected on the basis of their high consumption and/or frequently reported detection in wastewaters, and the possibility to be analyzed under the same experimental conditions.
2.
Material and methods
2.1.
Chemicals
All pharmaceuticals and their corresponding isotope-labeled standards were of high purity grade (>90%). Detailed information on the chemicals used, the providers of the analytical standards, and the preparation of the mixture solutions is given in the Supporting Information (Table S1).
2.2.
Description of the sewer system
The studied collection system is situated in Sant Pere Pescador, a small catchment in the north-east coastal region of
100
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
Catalonia (Spain) (Fig. 1). The system consists of a pump station (Point 1) that collects the wastewater of 3000 inhabitants during winter, and up to around 35,000 during summer months. Estimated population in the sampling period was 7500 inhabitants. The pump station at Point 1 collects both raw sewage from a close neighborhood (HRT 0.5 h) and from 2 pressure pipes upstream of the catchment with average HRTs of 1.5 and 2 h. The average volume of sewage pumped per event is 23.3 m3, where pump events last around 4 min. The wastewater is pushed through the pressure pipe until l’Escala WWTP (Point 2). The length of the sewer pipe is 7620 m, and its internal diameter is 500 mm. The average wastewater flow was 62.7 m3/h, which resulted in an HRT of 21 ± 2 h (more details in Supplementary Information). The system has no lateral streams and is in good condition, therefore infiltration and exfiltration can be considered negligible. Wastewater characteristics were typical of domestic sewage. Dissolved concentrations of sulfate (SO2 4 ) and sulfide (H2S) were used as indicators of the biological activity of the anaerobic sewer biofilm (Supporting information S3).
2.3.
Sampling
Raw sewage water samples were collected daily as twentyfour hour composite samples at the pumping station e Point 1 (Fig. 1) and at the influent of WWTP L'Escala (Point 2, Fig. 1). The start of sampling at Point 2 (i.e. WWTP influent) was always delayed for 21-hour (i.e. HRT of the pipe) in order to capture the portion of the wastewater stream measured at Point 1. The sampling was carried out over five consecutive days in July 2013. The weather was dry and stable, with the environmental temperatures ranging from 14 to 30 C. Time proportional samples were collected employing 2 portable automatic refrigerated samplers Hach-Lange Bu¨hler BL 2000 with 24 PE containers of 1 L. The autosamplers were programmed to collect a 50 mL sample each 10 min to form a 7.2 ± 0.4 L composite daily. The frequency of pump events and the HRT variation are shown in Figure S1 (Supp. Information).
The sampler tube at Point 1 was installed deep in the pumpstation wet well to ensure the collection wastewater samples, it contained always at least 5.5 m3 of raw wastewater. The samples at Point 2 (outlet) were taken from a tap installed 5 m before the discharge to WWTP and that always contained around 1 m3 wastewater. Collected wastewater, i.e. 4 inlet and 4 outlet samples, were maintained at 4 C to avoid biological degradation. The samples were extracted and processed within 12 h, and analyzed immediately after the preparation.
2.4.
Sample preparation and analysis
Samples for the analysis of PHACs, i.e. 4 inlet samples and 4 outlet wastewater samples, were filtered through a 1-mm glass fiber filter and a 0.45-mm nylon fiber filter (Whatman, U.K.). A Na2EDTA solution (0.1 M) was added to 50 mL aliquots of the filtered wastewater samples (all in triplicate) to a concentration of 0.1% (g solute/g solution). Then, the samples were spiked with a solution of surrogate standards to a concentration of 200 ng/L and preconcentrated by solid phase extraction e Oasis® HLB (Waters, Milford, MA, USA), using a Baker vacuum system (J.T. Baker, Deventer, The Netherlands). Analytes were eluted by 6 mL of pure methanol and the extracts were gently dried under nitrogen and reconstituted in 1 mL of a methanol/water mixture (10/90, v/v) and fortified by a mixture of internal standards (Table S1) to a final concentration of 20 ng/mL. The samples were analyzed by a Waters Acquity Ultra-PerformanceTM liquid chromatography system (Waters, Milford, MA, USA) coupled to a 5500 QTRAP hybrid triple quadrupole-linear ion trap mass spectrometer (Applied Biosystems, Foster City, CA, USA) with a turbo Ion Spray source, according to a previously reported method (Gros et al., 2012). Quantitative analysis was performed by internal standard approach, and two isotope-labeled compounds, i.e. sulfadoxine-d3 for positive ionization mode, and ketoprofen-d3 for negative ionization mode, were added to all the samples before the extraction as control standards. Linear least
Fig. 1 e Location and layout of the studied sewer. Point 1: pump station collecting wastewater from the Sant Pere Pescador municipality; Point 2: wastewater treatment plant.
Table 1 e Method performance characteristics, i.e. method quantification limits (MQL), matrix effect (given as a percentage of signal suppression), average recovery, intraday precision, standard error (SE) of replicate measurements, uncertainty associated with chemical analysis, and the day-to-day variation in concentrations of the target pharmaceuticals. Therapeutic group
Nonsteroidal anti-inflamatory agents/Analgesics (M01)
Antiepileptics (N03) Psycholeptics (N05) Psychoanaleptics (N06) Drugs for peptic ulcer and gastro-oesophageal reflux disease (A02) Beta-blocking agents (C07)
Diuretic (C03) Antidiabetic (A10) Agents acting on the renin-angiotensin system (C09) Antithrombotic agents (B01) Drugs for obstructive airway diseases (R03)
MQL
Matrix effect (mean, n ¼ 6)
Recovery
Intraday precision
SE of replicate measurements
Uncertainty associated with chemical analysis %
ng/L
%
%
%
%
(Mean, n ¼ 6)
(Mean, n ¼ 6)
(Mean, n ¼ 6)
(SE, n ¼ 6)
(n ¼ 9)
Ketoprofen Naproxen Ibuprofen Indomethacin Diclofenac Propyphenazone Oxycodone Codeine Bezafibrate Gemfibrozil Pravastatin Fluvastitatin Atorvastatin Carbamazepine Lorazepam Citalopram Venlafaxine Ranitidine Cimetidine
4 27 37 12 10 4 5 14 20 15 9 9 7 14 10 10 14 12 10
12 50 15 32 32 60 38 28 38 40 22 13 17 80 28 70 50 32 50
111 66 105 52 67 69 104 117 109 110 89 40 39 108 107 115 122 106 123
6 6 6 6 4 3 4 4 1 6 4 3 3 1 3 5 2 6 3
5 3 5 10 5 7 9 6 8 3 3 2 4 2 7 7 4 10 6
Atenolol Sotalol Propanolol Metoprolol Furosemide Torasemide Glibenclamide Amlodipine Losartan Irbesartan Valsartan Clopidogrel Salbutamol
5 7 11 7 14 4 12 10 19 5 5 4 4
13 70 81 70 62 60 78 81 46 10 30 78 63
91 77 82 90 85 122 64 78 45 58 38 62 74
3 1 4 3 3 8 5 4 4 5 7 3 3
4 6 10 5 6 5 4 9 5 10 2 7 4
Day-to-day variation in concentrations % Inlet
Outlet
8 7 8 11 7 8 10 7 8 7 6 4 6 3 8 9 5 12 7
2 10 8 12 8 58 12 14 41 15 7 5 8 13 9 5 13 15 14
5 4 10 19 11 59 8 11 32 6 8 4 13 22 13 9 8 8 29
5 6 11 6 7 10 7 10 7 11 7 8 6
9 13 21 7 7 12 1 15 14 12 17 12 21
11 6 8 4 12 16 3 9 13 4 16 7 22
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
Lipid regulators (C10)
Compound
(continued on next page)
101
102
16 12 10 8 13 8 17 22 10 4 29 7 4
Outlet
Ibuprofen metabolites
Calcium channel blockers (C08) Antihelmintic (P02)
Antibacterials for systemic use (J01)
Erythromycin Azithromycin Clarithromycin Tetracycline Ofloxacin Ciprofloxacin Sulfamethoxazole Trimethoprim Diltiazem Levamisole Thiabendazole 1-hydroxy Ibuprofen 2-hydroxy ibuprofen
21 5 16 10 7 7 8 7 30 5 10 35 37
90 10 80 48 61 72 50 70 90 80 80 10 15
87 80 57 107 82 110 96 73 73 72 48 115 75
7 4 4 5 4 7 2 2 4 4 5 4 6
8 7 9 8 7 8 3 4 7 6 8 4 2
11 8 10 9 9 11 4 5 8 7 10 6 7
34 8 13 9 17 15 23 27 7 12 31 6 4
% Inlet % (n ¼ 9)
% %
(SE, n ¼ 6)
% %
(Mean, n ¼ 6)
ng/L
(Mean, n ¼ 6)
(Mean, n ¼ 6)
Recovery Therapeutic group
Table 1 e (continued )
Compound
MQL
Matrix effect (mean, n ¼ 6)
Intraday precision
SE of replicate measurements
Uncertainty associated with chemical analysis
Day-to-day variation in concentrations
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
squares regression calibration curves were constructed using 9 calibration standard mixtures of the target compounds at concentrations ranging from 0.05 to 200 ng/mL, and for all the compounds it was linear over the range from 0.1 to 100 ng/mL. Method quantification limits (MQLs) were calculated as ten times signal-to-noise, and the values are reported in Table 1. To determine the relative recoveries of the analyzed compounds during the applied method, six wastewater samples were spiked with the pharmaceutical standard mixture to a final concentration of 50 ng/mL. Since the samples may contain target compounds, blanks (non-spiked samples) were also analyzed and their quantified concentrations were subtracted from those of spiked samples. The recoveries and the standard errors (n ¼ 6, confidence interval (CI) 68%) are summarized in Table 1. Instrumental intraday precision of the applied method was calculated from 6 injections of a standard solution at 50 ng/mL over the course of day. It is reported as the standard error of the mean (n ¼ 6, 68% CI) in Table 1. Analytical methods for the monitoring of parameters related to anaerobic activities in sewer systems included the analysis of dissolved sulfur species, volatile fatty acids (VFAs), chemical oxygen demand (COD), total and volatile suspended solids (TSS and VSS, respectively), pH, conductivity, temperature and dissolved oxygen (DO). For the analysis of dissolved sulfur species, 1.5 mL of wastewater was filtered through a 0.22 mm filter and enriched by 0.5 mL preserving solution of sulfide anti-oxidant buffer (SAOB) (Keller-Lehmann et al. 2006). Samples were analyzed by an ion chromatograph (IC) with UV and conductivity detector (Dionex ICS-5000). Online sulfide measurements were carried out by an UVeVIS spectrometer probe s::can spectro::lyser (Messtechnik, GmbH, Austria) according to the method of Sutherland-Stacey et al. (2008). VFAs were measured by gas chromatography coupled to FID detector (Thermofisher Scientific). COD analysis was performed using a standard photometric test kit with commercially available reagent (LCK 114, Hach Lang). Absorbance readings were conducted on DR 2800 Hach Langue spectrometer (method 814). TSS and VSS were analyzed according to the standard method 2510D (APHA 1998). Temperature, pH, conductivity, and DO concentration were measured by Portable probes Unisense Model YSI Digital Pro.
2.5.
Uncertainty estimation
The uncertainty associated with the measured PhACs concentrations was calculated from the individual uncertainties from PHACs chemical analysis and from sampling. The uncertainty of chemical analysis was estimated from the relative recoveries (six spiked samples), triplicate analysis of the samples, intra-day instrumental precision (six injections of standard at 50 ng/mL every 4 h) and other uncertainty factors (i.e. 2%, (Kovalova et al., 2012)). The mixture of chemical standards was prepared just before the analysis, so the error associated with the stability of the solution could be considered negligible. The uncertainty due to sampling was estimated by considering the population served by the sewer and their pharmaceutical consumption pattern, and the applied sampling procedure as proposed in the literature (Castiglioni et al., 2012; Ort et al., 2010a, 2010b; Weissbrodt et al., 2009). The
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
population number and their pharmaceutical consumption pattern figure in the expected Number of wastewater pulses containing the compounds of interest (Np), and it was calculated according to Ort et al. (2010b). The Np may not represent realistic pharmaceuticals consumption pattern for the catchment area, but is useful to get a general idea. The sewer system served approx 7500 population equivalents during the sampling period, calculated from the actual wastewater flow data (i.e. 1505 ± 26 m3/day) by assuming that a person produces 200 L of wastewater per day. The consumption data were obtained from the technical reports published by the Spanish Agency of Medicines and Healthcare Products ~ ola de Medicamentos y Productos Sanitarios). (Agencia Espan The consumption data and the calculated Np, together with the assumed values of sampling uncertainty are provided as Supp. Information (Section S4).
3.
Results and discussion
3.1.
Concentrations of PhACs in the pressure sewer pipe
Concentrations of the target pharmaceuticals, 43 pharmaceuticals and 2 metabolites of ibuprofen, i.e. 1-hydroxyibuprofen (1-OH-Ibuprofen) and 2-hydroxy-ibuprofen (2-OHIbuprofen), in analyzed wastewater samples ranged from low ng/L (e.g. 6e7 ng/L of glibenclamide) to a few mg/L (e.g. 4e6 mg/L of ibuprofen). Fig. 2 summarizes the concentrations of the compounds at the inlet and the outlet of the studied pressure pipe during four consecutive days. The figure shows the daily concentrations with associated uncertainties in the analytical measurement (error bars). Sampling uncertainty was not included in the error bars for a clearer picture of the changes in pharmaceuticals concentrations, and the compounds with sampling uncertainty larger than 20% are underlined. Day-today variations in the measured pharmaceuticals concentrations were between 3 and 16% for 80% of the analyzed compounds, at both sampling points (Table 1). Higher day-to-day variations in concentrations (20e59%) were observed for antimicrobials erythromycin, sulfamethoxazole and trimethoprim, lipid modifying drug bezafibrate, anti-acid drug cimetidin, and antifungal drug thiabendazole. This may be associated with the consumption of these pharmaceuticals in the studied area, which is connected with the results in a way that lower consumption introduces higher sampling uncertainty (Supporting information, Table S4). If these compounds are expected to be excreted in only a few wastewater pulses daily, then their concentrations, detected in the samples taken by the employed sampling protocol, may easily be underestimated (Ort et al., 2010a). Non-steroidal anti-inflammatory drugs ibuprofen and naproxen, and antihypertensive drug valsartan were detected at the highest concentrations at both sampling points. Average concentrations at the inlet and the outlet of the sewer pipe were 5.3 and 4.9 mg/L for ibuprofen, 3.5 and 3.4 mg/L for naproxen, and 1.9 and 1.9 mg/L for valsartan, respectively. These concentrations are in the same range as those previously reported for WWTP influents of the studied area (Ferrando-Climent et al., 2012; Gros et al., 2012). A human metabolite of ibuprofen, 2-hydroxy-ibuprofen, was detected
103
at similar concentrations as its parent compound, i.e. 5.1 and 5.0 mg/L at the inlet and the outlet of the sewer, respectively. Quantified ratio of 2-OH-ibuprofen to ibuprofen (i.e. 1:1 in average) at the inlet was lower than a typical human excretion ratio of 1.7:1 (Lienert el al., 2007; Weigel et al., 2004), and it was the same at the outlet. The proportion varied from study to study (Ferrando-Climent et al., 2012; Weigel et al., 2004), which may be due to stereoselective metabolism of ibuprofen (Rudy et al., 1991) and its administration routes (i.e. oral and topical, whereas the latter introduces active, unmetabolised, compound to sewer (Daughton and Ruhoy, 2009)). Day-to-day variation in the concentrations of the given compounds ranged from 4% (ibuprofen) to 17% (valsartan) for inlet wastewater, and from 4% (naproxen and 2-OH-ibuprofen) to 16% (valsartan). Based on consumption data and the applied sampling protocol, we assumed that the sampling uncertainty for these compounds was less than 5e10 % (CI 68%) on individual days. The lowest concentrations (
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Impact of in-sewer transformation on 43 pharmaceuticals in a pressurized sewer under anaerobic conditions a,b, Aleksandra Jelic a, Sara Rodriguez-Mozaz a,*, Damia Barcelo a Oriol Gutierrez a
Catalan Institute for Water Research (ICRA), Scientific and Technologic Park of the University of Girona, Emili Grahit 101, 17003 Girona, Spain b Institute of Environmental Assessment and Water Research (IDAEA-CSIC), Jordi Girona 18-26, 08034 Barcelona, Spain
article info
abstract
Article history:
The occurrence of 43 pharmaceuticals and 2 metabolites of ibuprofen was evaluated at the
Received 5 March 2014
inlet and the outlet of a pressure sewer pipe in order to asses if in-sewer processes affect
Received in revised form
the pharmaceutical concentrations during their pass through the pipe. The target com-
31 July 2014
pounds were detected at concentrations ranging from low ng/L to a few mg/L, which are in
Accepted 20 September 2014
the range commonly found in municipal wastewater of the studied area. The changes in
Available online 7 October 2014
concentrations between two sampling points were negligible for most compounds, i.e. from 10 to 10%. A higher decrease in concentrations (25e60 %) during the pass through
Keywords:
the pipe was observed for diltiazem, citalopram, clarithromycin, bezafibrate and amlodi-
Pharmaceuticals
pine. Negative removal was calculated for sulfamethoxazole (66 ± 15%) and irbesartan
Wastewater
(58 ± 25%), which may be due to the conversion of conjugates back to their parent
Sewage system
compounds in the sewer. The results show that microbial transformation of pharmaceuticals begins in sewer, albeit to different extents for different compounds. Therefore, the in-sewer transformation of pharmaceuticals should be assessed especially when their concentrations are used to estimate and refine the estimation of their per capita consumption in a catchment of interest in the sewage epidemiology approach. © 2014 Elsevier Ltd. All rights reserved.
1.
Introduction
Pharmaceutical compounds (PhACs) have been given much attention in the last decade due to their documented presence in the environment and concern about their potential adverse effects to humans and wildlife. This was largely owing to the
* Corresponding author. Tel.: þ34 972 18 33 80. E-mail address: [email protected] (S. Rodriguez-Mozaz). http://dx.doi.org/10.1016/j.watres.2014.09.033 0043-1354/© 2014 Elsevier Ltd. All rights reserved.
development of advanced analytical techniques and instruments, which enabled the trace-level detection of these compounds in different environmental compartments including surface and groundwater, sediment and soil (Ku¨mmerer, 2009). The issue of “pharmaceuticals in the environment” has only been covered by regulations developed by the Swiss environmental agency for Switzerland (Eggen
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
et al., 2014), but it has been recognized as an emerging problem in the Directive 2010/84/EU and Regulation 1235/2010 on pharmacovigilance that recommend the monitoring and evaluation of the risk of environmental effects of medicinal products. Within the Urban wastewater system (UWS), wastewater treatment plants (WWTP) have been identified as a point of discharge of pharmaceuticals into the environment. Numerous studies have been conducted to assess the magnitude of pharmaceuticals that reach the environment through the wastewater and sludge discharge, and to evaluate the efficiency of typical wastewater treatment plants regarding the removal of these compounds from raw wastewater (Jelic et al., 2011; Joss et al., 2005; Radjenovic et al., 2009). The wastewater quality and its pharmaceuticals content, particularly, have been typically monitored at the influent and/or the effluent of a treatment plant, where these micropollutants were detected at concentrations ranging from low ng/L to a few mg/L (Verlicchi et al., 2012). A wide variation in removal efficiencies has been reported for individual compounds in separate studies as well as across therapeutic classes and treatment processes (Onesios et al., 2009). The actual and major source of PhACs is households, upstream of the catchment, where pharmaceuticals are consumed and eventually released to the sewer systems. Sewers are underground infrastructure composed by drains, pipes and pumping stations that transport wastewater to the treatment facilities (Hvitved-Jacobsen, 2002). Two types of sewer pipes are found depending on the topography: 1) gravity pipes, where wastewater flows naturally from higher to lower elevation in partially-filled sections, and 2) pressure pipes (force mains) e completely-filled tubes e with a pump station at their inlet to mechanically push sewage through flat and sloping terrains. Sewers have been traditionally considered as just a sewage collection and transport system, but in fact they are also active chemical and biological reactors (Huisman et al., 2002; Rudelle et al., 2011; Schilperoort et al., 2012). In gravity sewers, having a relatively long transport time, a significant reduction in the amount of biodegradable substrate and production of biomass may occur under aerobic conditions (Hvitved-Jacobsen et al., 2002). Completely-filled pressured pipes are prone to development of septic conditions which make them preferred grow-locations for anaerobic biofilms that generate detrimental compounds such as sulfide (H2S), volatile organic sulfur compounds (VSOC) and methane (CH4) (Guisasola et al., 2008; Gutierrez et al., 2010; HvitvedJacobsen et al., 2013; Sutherland-Stacey et al., 2008). Micropollutants in sewers are still scarcely reported in the literature. Microbial and chemical processes that occur in sewers have been generally neglected. This issue has been recognized in the field of sewage epidemiology, which basically aims at estimating the actual illicit drug use and trends from the concentrations detected in influent wastewater (Daughton, 2001). It has been emphasized that the main limitation of the approach, which directly influences its accuracy, is the lack of comprehensive data on drug marker stability in sewer systems and during storage (Castiglioni et al., 2012; Khan and Nicell, 2011; van Nuijs et al., 2011). The stability of illicit drugs during sample storage has been assessed in several studies (Baker and Kasprzyk-Hordern, 2011; Castiglioni et al.,
99
2011; van Nuijs et al., 2012). Lai et al. (2011) quantified the overall uncertainty of the sewage epidemiology approach and included five prescription pharmaceuticals (atenolol, carbamazepine, gabapentin, hydrochlorothiazide and venlafaxine) to refine the estimation of the number of people that contributed to wastewater in the monitored catchment. However, possible in-sewer transformation of the pharmaceuticals was not considered. Only Thai et al. (2014) have studied how insewer transformation processes alter the concentrations of organic pollutants, in particular illicit drugs and their metabolites, under different conditions, at laboratory-scale sewer reactors. They found that the biofilms in the lab-scale sewer have significantly enhanced the degradation rate of some compounds (i.e. cocaine, and a metabolite of heroin, 6-acetyl morphine) in comparison to the rates in wastewater only, and they suggested that the fate of illicit drugs will depend on sewer conditions. As mentioned before, pressure pipes are the most biologically-active sections of sewer networks due to higher biofilm presence and biofilm-sewage contact (HvitvedJacobsen, T., 2002). Under anaerobic conditions in completely-filled pipes, septic biofilm grows attached to the sewer walls. Sewer biofilm captures the compounds dissolved in wastewater and change their concentrations in the flowing sewage water (Huisman et al., 2002). In addition, for a similar length of pipe, hydraulic retention time (HRT) in pressure pipes is typically longer than in gravity sections since pressure pipes need to be completely filled and act as a plug flow reactor needing higher volumes to reach similar distances. This can lead to higher biofilm-sewage contact time that can promote the transformation processes. To date, there has been no comprehensive study on whether in-sewer biotransformation of pharmaceuticals occurs and to which extent. In this work, we aimed at assessing the occurrence and the extent of biotransformation of 43 pharmaceuticals and 2 metabolites of ibuprofen in pressure sewer pipes. The study was carried out at a flat 7620 m long pipe transporting domestic wastewater from the municipality of Sant Pere Pescador (Catalonia, Spain). The target compounds belong to commonly prescribed and over-the-counter pharmaceuticals (i.e. nonsteroidal anti-inflammatory agents and analgesics, lipid modifying agents, psycholeptic and antiepileptic drugs etc.) and were selected on the basis of their high consumption and/or frequently reported detection in wastewaters, and the possibility to be analyzed under the same experimental conditions.
2.
Material and methods
2.1.
Chemicals
All pharmaceuticals and their corresponding isotope-labeled standards were of high purity grade (>90%). Detailed information on the chemicals used, the providers of the analytical standards, and the preparation of the mixture solutions is given in the Supporting Information (Table S1).
2.2.
Description of the sewer system
The studied collection system is situated in Sant Pere Pescador, a small catchment in the north-east coastal region of
100
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
Catalonia (Spain) (Fig. 1). The system consists of a pump station (Point 1) that collects the wastewater of 3000 inhabitants during winter, and up to around 35,000 during summer months. Estimated population in the sampling period was 7500 inhabitants. The pump station at Point 1 collects both raw sewage from a close neighborhood (HRT 0.5 h) and from 2 pressure pipes upstream of the catchment with average HRTs of 1.5 and 2 h. The average volume of sewage pumped per event is 23.3 m3, where pump events last around 4 min. The wastewater is pushed through the pressure pipe until l’Escala WWTP (Point 2). The length of the sewer pipe is 7620 m, and its internal diameter is 500 mm. The average wastewater flow was 62.7 m3/h, which resulted in an HRT of 21 ± 2 h (more details in Supplementary Information). The system has no lateral streams and is in good condition, therefore infiltration and exfiltration can be considered negligible. Wastewater characteristics were typical of domestic sewage. Dissolved concentrations of sulfate (SO2 4 ) and sulfide (H2S) were used as indicators of the biological activity of the anaerobic sewer biofilm (Supporting information S3).
2.3.
Sampling
Raw sewage water samples were collected daily as twentyfour hour composite samples at the pumping station e Point 1 (Fig. 1) and at the influent of WWTP L'Escala (Point 2, Fig. 1). The start of sampling at Point 2 (i.e. WWTP influent) was always delayed for 21-hour (i.e. HRT of the pipe) in order to capture the portion of the wastewater stream measured at Point 1. The sampling was carried out over five consecutive days in July 2013. The weather was dry and stable, with the environmental temperatures ranging from 14 to 30 C. Time proportional samples were collected employing 2 portable automatic refrigerated samplers Hach-Lange Bu¨hler BL 2000 with 24 PE containers of 1 L. The autosamplers were programmed to collect a 50 mL sample each 10 min to form a 7.2 ± 0.4 L composite daily. The frequency of pump events and the HRT variation are shown in Figure S1 (Supp. Information).
The sampler tube at Point 1 was installed deep in the pumpstation wet well to ensure the collection wastewater samples, it contained always at least 5.5 m3 of raw wastewater. The samples at Point 2 (outlet) were taken from a tap installed 5 m before the discharge to WWTP and that always contained around 1 m3 wastewater. Collected wastewater, i.e. 4 inlet and 4 outlet samples, were maintained at 4 C to avoid biological degradation. The samples were extracted and processed within 12 h, and analyzed immediately after the preparation.
2.4.
Sample preparation and analysis
Samples for the analysis of PHACs, i.e. 4 inlet samples and 4 outlet wastewater samples, were filtered through a 1-mm glass fiber filter and a 0.45-mm nylon fiber filter (Whatman, U.K.). A Na2EDTA solution (0.1 M) was added to 50 mL aliquots of the filtered wastewater samples (all in triplicate) to a concentration of 0.1% (g solute/g solution). Then, the samples were spiked with a solution of surrogate standards to a concentration of 200 ng/L and preconcentrated by solid phase extraction e Oasis® HLB (Waters, Milford, MA, USA), using a Baker vacuum system (J.T. Baker, Deventer, The Netherlands). Analytes were eluted by 6 mL of pure methanol and the extracts were gently dried under nitrogen and reconstituted in 1 mL of a methanol/water mixture (10/90, v/v) and fortified by a mixture of internal standards (Table S1) to a final concentration of 20 ng/mL. The samples were analyzed by a Waters Acquity Ultra-PerformanceTM liquid chromatography system (Waters, Milford, MA, USA) coupled to a 5500 QTRAP hybrid triple quadrupole-linear ion trap mass spectrometer (Applied Biosystems, Foster City, CA, USA) with a turbo Ion Spray source, according to a previously reported method (Gros et al., 2012). Quantitative analysis was performed by internal standard approach, and two isotope-labeled compounds, i.e. sulfadoxine-d3 for positive ionization mode, and ketoprofen-d3 for negative ionization mode, were added to all the samples before the extraction as control standards. Linear least
Fig. 1 e Location and layout of the studied sewer. Point 1: pump station collecting wastewater from the Sant Pere Pescador municipality; Point 2: wastewater treatment plant.
Table 1 e Method performance characteristics, i.e. method quantification limits (MQL), matrix effect (given as a percentage of signal suppression), average recovery, intraday precision, standard error (SE) of replicate measurements, uncertainty associated with chemical analysis, and the day-to-day variation in concentrations of the target pharmaceuticals. Therapeutic group
Nonsteroidal anti-inflamatory agents/Analgesics (M01)
Antiepileptics (N03) Psycholeptics (N05) Psychoanaleptics (N06) Drugs for peptic ulcer and gastro-oesophageal reflux disease (A02) Beta-blocking agents (C07)
Diuretic (C03) Antidiabetic (A10) Agents acting on the renin-angiotensin system (C09) Antithrombotic agents (B01) Drugs for obstructive airway diseases (R03)
MQL
Matrix effect (mean, n ¼ 6)
Recovery
Intraday precision
SE of replicate measurements
Uncertainty associated with chemical analysis %
ng/L
%
%
%
%
(Mean, n ¼ 6)
(Mean, n ¼ 6)
(Mean, n ¼ 6)
(SE, n ¼ 6)
(n ¼ 9)
Ketoprofen Naproxen Ibuprofen Indomethacin Diclofenac Propyphenazone Oxycodone Codeine Bezafibrate Gemfibrozil Pravastatin Fluvastitatin Atorvastatin Carbamazepine Lorazepam Citalopram Venlafaxine Ranitidine Cimetidine
4 27 37 12 10 4 5 14 20 15 9 9 7 14 10 10 14 12 10
12 50 15 32 32 60 38 28 38 40 22 13 17 80 28 70 50 32 50
111 66 105 52 67 69 104 117 109 110 89 40 39 108 107 115 122 106 123
6 6 6 6 4 3 4 4 1 6 4 3 3 1 3 5 2 6 3
5 3 5 10 5 7 9 6 8 3 3 2 4 2 7 7 4 10 6
Atenolol Sotalol Propanolol Metoprolol Furosemide Torasemide Glibenclamide Amlodipine Losartan Irbesartan Valsartan Clopidogrel Salbutamol
5 7 11 7 14 4 12 10 19 5 5 4 4
13 70 81 70 62 60 78 81 46 10 30 78 63
91 77 82 90 85 122 64 78 45 58 38 62 74
3 1 4 3 3 8 5 4 4 5 7 3 3
4 6 10 5 6 5 4 9 5 10 2 7 4
Day-to-day variation in concentrations % Inlet
Outlet
8 7 8 11 7 8 10 7 8 7 6 4 6 3 8 9 5 12 7
2 10 8 12 8 58 12 14 41 15 7 5 8 13 9 5 13 15 14
5 4 10 19 11 59 8 11 32 6 8 4 13 22 13 9 8 8 29
5 6 11 6 7 10 7 10 7 11 7 8 6
9 13 21 7 7 12 1 15 14 12 17 12 21
11 6 8 4 12 16 3 9 13 4 16 7 22
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
Lipid regulators (C10)
Compound
(continued on next page)
101
102
16 12 10 8 13 8 17 22 10 4 29 7 4
Outlet
Ibuprofen metabolites
Calcium channel blockers (C08) Antihelmintic (P02)
Antibacterials for systemic use (J01)
Erythromycin Azithromycin Clarithromycin Tetracycline Ofloxacin Ciprofloxacin Sulfamethoxazole Trimethoprim Diltiazem Levamisole Thiabendazole 1-hydroxy Ibuprofen 2-hydroxy ibuprofen
21 5 16 10 7 7 8 7 30 5 10 35 37
90 10 80 48 61 72 50 70 90 80 80 10 15
87 80 57 107 82 110 96 73 73 72 48 115 75
7 4 4 5 4 7 2 2 4 4 5 4 6
8 7 9 8 7 8 3 4 7 6 8 4 2
11 8 10 9 9 11 4 5 8 7 10 6 7
34 8 13 9 17 15 23 27 7 12 31 6 4
% Inlet % (n ¼ 9)
% %
(SE, n ¼ 6)
% %
(Mean, n ¼ 6)
ng/L
(Mean, n ¼ 6)
(Mean, n ¼ 6)
Recovery Therapeutic group
Table 1 e (continued )
Compound
MQL
Matrix effect (mean, n ¼ 6)
Intraday precision
SE of replicate measurements
Uncertainty associated with chemical analysis
Day-to-day variation in concentrations
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
squares regression calibration curves were constructed using 9 calibration standard mixtures of the target compounds at concentrations ranging from 0.05 to 200 ng/mL, and for all the compounds it was linear over the range from 0.1 to 100 ng/mL. Method quantification limits (MQLs) were calculated as ten times signal-to-noise, and the values are reported in Table 1. To determine the relative recoveries of the analyzed compounds during the applied method, six wastewater samples were spiked with the pharmaceutical standard mixture to a final concentration of 50 ng/mL. Since the samples may contain target compounds, blanks (non-spiked samples) were also analyzed and their quantified concentrations were subtracted from those of spiked samples. The recoveries and the standard errors (n ¼ 6, confidence interval (CI) 68%) are summarized in Table 1. Instrumental intraday precision of the applied method was calculated from 6 injections of a standard solution at 50 ng/mL over the course of day. It is reported as the standard error of the mean (n ¼ 6, 68% CI) in Table 1. Analytical methods for the monitoring of parameters related to anaerobic activities in sewer systems included the analysis of dissolved sulfur species, volatile fatty acids (VFAs), chemical oxygen demand (COD), total and volatile suspended solids (TSS and VSS, respectively), pH, conductivity, temperature and dissolved oxygen (DO). For the analysis of dissolved sulfur species, 1.5 mL of wastewater was filtered through a 0.22 mm filter and enriched by 0.5 mL preserving solution of sulfide anti-oxidant buffer (SAOB) (Keller-Lehmann et al. 2006). Samples were analyzed by an ion chromatograph (IC) with UV and conductivity detector (Dionex ICS-5000). Online sulfide measurements were carried out by an UVeVIS spectrometer probe s::can spectro::lyser (Messtechnik, GmbH, Austria) according to the method of Sutherland-Stacey et al. (2008). VFAs were measured by gas chromatography coupled to FID detector (Thermofisher Scientific). COD analysis was performed using a standard photometric test kit with commercially available reagent (LCK 114, Hach Lang). Absorbance readings were conducted on DR 2800 Hach Langue spectrometer (method 814). TSS and VSS were analyzed according to the standard method 2510D (APHA 1998). Temperature, pH, conductivity, and DO concentration were measured by Portable probes Unisense Model YSI Digital Pro.
2.5.
Uncertainty estimation
The uncertainty associated with the measured PhACs concentrations was calculated from the individual uncertainties from PHACs chemical analysis and from sampling. The uncertainty of chemical analysis was estimated from the relative recoveries (six spiked samples), triplicate analysis of the samples, intra-day instrumental precision (six injections of standard at 50 ng/mL every 4 h) and other uncertainty factors (i.e. 2%, (Kovalova et al., 2012)). The mixture of chemical standards was prepared just before the analysis, so the error associated with the stability of the solution could be considered negligible. The uncertainty due to sampling was estimated by considering the population served by the sewer and their pharmaceutical consumption pattern, and the applied sampling procedure as proposed in the literature (Castiglioni et al., 2012; Ort et al., 2010a, 2010b; Weissbrodt et al., 2009). The
w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 9 8 e1 0 8
population number and their pharmaceutical consumption pattern figure in the expected Number of wastewater pulses containing the compounds of interest (Np), and it was calculated according to Ort et al. (2010b). The Np may not represent realistic pharmaceuticals consumption pattern for the catchment area, but is useful to get a general idea. The sewer system served approx 7500 population equivalents during the sampling period, calculated from the actual wastewater flow data (i.e. 1505 ± 26 m3/day) by assuming that a person produces 200 L of wastewater per day. The consumption data were obtained from the technical reports published by the Spanish Agency of Medicines and Healthcare Products ~ ola de Medicamentos y Productos Sanitarios). (Agencia Espan The consumption data and the calculated Np, together with the assumed values of sampling uncertainty are provided as Supp. Information (Section S4).
3.
Results and discussion
3.1.
Concentrations of PhACs in the pressure sewer pipe
Concentrations of the target pharmaceuticals, 43 pharmaceuticals and 2 metabolites of ibuprofen, i.e. 1-hydroxyibuprofen (1-OH-Ibuprofen) and 2-hydroxy-ibuprofen (2-OHIbuprofen), in analyzed wastewater samples ranged from low ng/L (e.g. 6e7 ng/L of glibenclamide) to a few mg/L (e.g. 4e6 mg/L of ibuprofen). Fig. 2 summarizes the concentrations of the compounds at the inlet and the outlet of the studied pressure pipe during four consecutive days. The figure shows the daily concentrations with associated uncertainties in the analytical measurement (error bars). Sampling uncertainty was not included in the error bars for a clearer picture of the changes in pharmaceuticals concentrations, and the compounds with sampling uncertainty larger than 20% are underlined. Day-today variations in the measured pharmaceuticals concentrations were between 3 and 16% for 80% of the analyzed compounds, at both sampling points (Table 1). Higher day-to-day variations in concentrations (20e59%) were observed for antimicrobials erythromycin, sulfamethoxazole and trimethoprim, lipid modifying drug bezafibrate, anti-acid drug cimetidin, and antifungal drug thiabendazole. This may be associated with the consumption of these pharmaceuticals in the studied area, which is connected with the results in a way that lower consumption introduces higher sampling uncertainty (Supporting information, Table S4). If these compounds are expected to be excreted in only a few wastewater pulses daily, then their concentrations, detected in the samples taken by the employed sampling protocol, may easily be underestimated (Ort et al., 2010a). Non-steroidal anti-inflammatory drugs ibuprofen and naproxen, and antihypertensive drug valsartan were detected at the highest concentrations at both sampling points. Average concentrations at the inlet and the outlet of the sewer pipe were 5.3 and 4.9 mg/L for ibuprofen, 3.5 and 3.4 mg/L for naproxen, and 1.9 and 1.9 mg/L for valsartan, respectively. These concentrations are in the same range as those previously reported for WWTP influents of the studied area (Ferrando-Climent et al., 2012; Gros et al., 2012). A human metabolite of ibuprofen, 2-hydroxy-ibuprofen, was detected
103
at similar concentrations as its parent compound, i.e. 5.1 and 5.0 mg/L at the inlet and the outlet of the sewer, respectively. Quantified ratio of 2-OH-ibuprofen to ibuprofen (i.e. 1:1 in average) at the inlet was lower than a typical human excretion ratio of 1.7:1 (Lienert el al., 2007; Weigel et al., 2004), and it was the same at the outlet. The proportion varied from study to study (Ferrando-Climent et al., 2012; Weigel et al., 2004), which may be due to stereoselective metabolism of ibuprofen (Rudy et al., 1991) and its administration routes (i.e. oral and topical, whereas the latter introduces active, unmetabolised, compound to sewer (Daughton and Ruhoy, 2009)). Day-to-day variation in the concentrations of the given compounds ranged from 4% (ibuprofen) to 17% (valsartan) for inlet wastewater, and from 4% (naproxen and 2-OH-ibuprofen) to 16% (valsartan). Based on consumption data and the applied sampling protocol, we assumed that the sampling uncertainty for these compounds was less than 5e10 % (CI 68%) on individual days. The lowest concentrations (
