Sampling for Pharmaceuticals and Personal Care Products (PPCPs

Environ. Sci. Technol. 2010, 44, 6024–6035

Sampling for Pharmaceuticals and Personal Care Products (PPCPs) and Illicit Drugs in Wastewater Systems: Are Your Conclusions Valid? A Critical Review C H R I S T O P H O R T , * ,† M I C H A E L G . L A W R E N C E , † ¨ RG RIECKERMANN,‡ AND ADRIANO JOSS‡ JO The University of Queensland, Advanced Water Management Centre (AWMC), QLD 4072, Australia, and Eawag, Swiss Federal Institute of Aquatic Science and Technology, CH-8600 Du ¨ bendorf, Switzerland

Received March 10, 2010. Revised manuscript received June 24, 2010. Accepted July 6, 2010.

The analysis of 87 peer-reviewed journal articles reveals that sampling for pharmaceuticals and personal care products (PPCPs) and illicit drugs in sewers and sewage treatment plant influents is mostly carried out according to existing tradition or standard laboratory protocols. Less than 5% of all studies explicitly consider internationally acknowledged guidelines or methods for the experimental design of monitoring campaigns. In the absence of a proper analysis of the system under investigation, the importance of short-term pollutant variations was typically not addressed. Therefore, due to relatively long sampling intervals, potentially inadequate sampling modes, or insufficient documentation, it remains unclear for the majority of reviewed studies whether observed variations can be attributed to “real” variations or if they simply reflect sampling artifacts. Based on results from previous and current work, the present paper demonstrates that sampling errors can lead to overinterpretation of measured data and ultimately, wrong conclusions. Depending on catchment size, sewer type, sampling setup, substance of interest, and accuracy of analytical method, avoidable sampling artifacts can range from “not significant” to “100% or more” for different compounds even within the same study. However, in most situations sampling errors can be reduced greatly, and sampling biases can be eliminated completely, by choosing an appropriate sampling mode and frequency. This is crucial, because proper sampling will help to maximize the value of measured data for the experimental assessment of the fate of PPCPs as well as for the formulation and validation of mathematical models. The trend from reporting presence or absence of a compound in “clean” water samples toward the quantification of PPCPs in raw wastewater requires not only sophisticated analytical methods but also adapted sampling methods. With increasing accuracy of chemical analyses, inappropriate sampling increasingly represents the major source of inaccuracy. A condensed step-by-step Sampling Guide is proposed as a starting point for future studies.

* Corresponding author e-mail: [email protected]; tel: +61 (0)7 3345 4730; fax: +61 (0)7 3365 4726. † The University of Queensland. ‡ Eawag, Swiss Federal Institute of Aquatic Science and Technology. 6024

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010

Introduction Since pharmaceuticals and personal care products (PPCPs) were first detected in “clean” natural water bodies and treated wastewater in the late 1990s, there have been vast developments in analytical methods and the application of these methods. As a result of numerous studies, there is now a consensus that the primary and constant sources of PPCPs to the aquatic environment are sewage treatment plant (STP) discharges (1), while combined sewer overflows (CSOs) can account for discharges of PPCP loads during short periods (2). To investigate occurrence and fate of PPCPs in the built environment, analytical methods were applied to a variety of different wastewater samples; these comprise, but are not limited to (i) STP influents and effluents to quantify removal efficiencies, (ii) industrial effluents to identify point sources and manage trade waste agreements, (iii) effluents of hospitals to evaluate source control measures, and most recently (iv) raw sewage for epidemiological studies to estimate illicit drug use in communities. “The four basic factors which affect the quality of environmental data are sample collection, sample preservation, analyses, and recording. Improper actions in any one area may result in poor data from which poor judgments are certain.” (U.S. EPA, 1982; 3). This fairly general, commonly acknowledged postulation was used in this review to assess how these factors have been addressed in 87 peer-reviewed journal articles (4-90) reporting results from 267 sites (see Figure 1 for a summary of this analysis). All selected studies investigated the fate of PPCPs or illicit drugs and quantified full-scale pollutant concentrations or fluxes based on samples from sewers. Sewers, including influents to STPs, are a particularly unpleasant or even hazardous working environment and exhibit often unique transport characteristics, highly dynamic flows, and variable pollutant loads. In almost all papers (99%, see Figure 1B) an increasingly sophisticated analytical method is described, referred to, or both. These descriptions usually extend over several paragraphs up to several pages. Preservation and laboratory specific measures to prevent and control contamination are addressed in three-quarters of the studies; ranging from one line to whole paragraphs. In contrast, sampling is usually covered in one or two sentences. Only 11% of the papers provide any justification for the choice of sampling protocol; in general, these statements are very short and more implicit than explicit. The elements that are insufficiently elaborated upon to assess data quality and to validate conclusionssat 10.1021/es100779n

 2010 American Chemical Society

Published on Web 07/26/2010

FIGURE 1. (A) Required elements to properly assess an environmental system (adapted from U.S. EPA, 1982; 3). (B) The percentage indicates the fraction of papers addressing the corresponding element. Eighty-seven journal articles (4-90) investigating the fate of pharmaceuticals and personal care products at 267 different sewer sites (including sewage treatment plant influents) were analyzed. (black: objective evaluation according to the scheme presented in Figure 2; gray: subjective judgment).

FIGURE 2. Categorization of 87 peer-reviewed journal articles (4-90) with regard to the description of sampling, preservation, and analytical methods to obtain environmental data (PPCPs and illicit drugs in sewers). Y ) Yes, N ) No. Brief definitions of sampling modes can be found in Table 1. Footnotes: 1 In most cases where only “composite sample” was specified it is very likely that samples were collected with a time-proportional sampling mode. 2 Samples which were referred to as grab samples but specifying a sampling frequency were counted as time-proportional samples when samples were pooled before chemical analysis. 3 Due to the lack of a detailed description of the sewer system under investigation (6 of 11 papers), it cannot be finally judged if their justification for sampling is sufficient to validate the conclusions; at least the authors of these papers have made an attempt to think about and address sampling issues. 4 These grab samples were analyzed individually and reported either as time series (individual values, not pooled) or at least with a median value, a minimum, and a maximum which indicates the observed range. Grab samples which were analyzed individually but only reported as one value without any indication of the observed range were counted as composite samples. least accessible for the reader in the final published manuscriptsare (i) a sound description of the environment under investigation, i.e., a (sewer) systems analysis, and (ii) a complete sampling protocol which is suitable to judge how (unknown) dynamics in sewers were accounted for.

Sophisticated theories on sampling were elaborated (e.g., 91) and it has been recognized that sampling can be a dominant source of uncertainty in many applications (e.g., 92). In the context of sampling for PPCPs or illicit drugs in sewers, this aspect has not been previously explored. VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6025

Preliminary (55, 57) and more comprehensive recent work (93) now clearly reveal that sampling uncertainty can be significant, and that insufficient emphasis is given to the sample collection. Guidelines and literature focusing on sampling from sewers and PPCPs are sparse and rarely cited; explicit specifications of variations to be expected or the magnitude of sampling uncertainty are, to a large extent, lacking. This review is structured according to the two main aims which are to • describe the two factors (concentration and flow variations) generating the fluctuations of PPCP loads in sewers (and STP influents) and analyze how they were accounted for in the scientific literature • present a first, possible step-by-step approach (subsequently referred to as “Sampling Guide”) to make an efficient and informed choice on the appropriate sampling mode and frequency to minimize sampling errors and maximize data quality of future studies.

Concentration Variations Determine the Sampling Frequency The sampling frequency as referred to in the context of this review is the sampling frequency to obtain one composite sample (typically a 24-h average sample) and not the frequency in a monitoring program (i.e., number and distribution of composite samples or sampling days throughout a year). “The concentration of the various determinands in [a] stream will vary due to random and systematic changes. The best technical solution, to determine the true values, would be to use an on-line automatic instrument providing continuous analyses of the determinand of interest.” (ISO, 1992; 94). Until now there is no online instrumentation to analyze for PPCPs directly in sewers and the following precautionary piece of advice, which can be found in similar phrasings in other norms and guidelines, should have been heeded: “The times and frequencies of sampling in any programme can be properly decided only after detailed preliminary work, in which a high sampling frequency is necessary...” (ISO, 1980; 95). Due to the high analytical costs per sample, 24-h composite samples are most commonly reported (75% of analyzed papers). Therefore, very little is known about short-term variations of PPCPs in sewers. Any aquatic chemist would agree that rainfall events influence the pollutant concentrations in rivers. These are also obvious events in sewers and STPs. Some authors explicitly mention that samples were taken during dry weather conditions, knowing that rain can impact the occurrence and fate of pollutants, particularly in combined sewer systems. Analogously, within sewer systems, toilet flushes and wastewater from other household appliances that contain the majority of the pharmaceutical load (1) should also be considered as “events”. These events can lead to significant short-term variations in combined and separate sewers in the range of minutes. The only statement in water and wastewater guidelines explicitly indicating this is: “Choose sampling intervals on the basis of the expected frequency of changes, which may vary from as little as 5 min to as long as 1 h or more.” (APHA, 1998; 96). To our knowledge, only three studies explicitly assessed and reported such short-term variations of micropollutants in sewers and the possible effects on sampling uncertainty (55, 57, 93): a modeling approach was suggested and experimental studies were performed. The latter were carried out at sampling frequencies of 30 s to 2 min over a total sampling period of 20 min to 4 h. The measured concentration time series clearly reveal high short-term variations: individual values deviated from the corresponding average concentration as much as -70% to +160% (benzotriazole, 6026

9


an anticorrosive in dishwasher detergents, widely applied) and -100% to +2300% (gadolinium, an MRI contrast agent, less frequently used). The observed variations can be attributed to intermittent “wastewater pulses” which compose the nonhomogeneous wastewater stream in sewers. At the source, these pulses last a few seconds (toilet flushes) to minutes (emptying a bathtub or pumping water out of washing machines, etc.). Due to dispersion in gravity-drained sewers (97) the pulses extend over approximately 1-10 minutes or more at the influent of a STP, depending on hydraulic conditions and flow distances. The number of wastewater pulses containing the PPCPs of interest over the sampling period is a decisive variable. In gravity-drained sewers this variable primarily depends on the number of people related to the excretion of a pharmaceutically active compound or the application of a personal care product in the catchment of a STP. If only a few such wastewater pulses are expected over the course of a day, it is apparent that a high sampling frequency is necessary to capture these pulses. Intuitively this does not seem to be the case if numerous pulses can be expected during the sampling period. However, even hundreds of pulses can result in a highly fluctuating pattern and a high sampling frequency is necessary to capture such fluctuations in a representative manner. This has been demonstrated with the modeling approach (a useful planning tool also referred to in the Sampling Guide, point 3) and measurements. The intermittent operation of pumps in pressurized sewers results in different patterns: one pulse of pumped wastewater reflects a fraction of a whole subcatchment rather than an individual toilet flush only. However, depending on the shortest pump cycle to be expected the required sampling frequencies can be similar as in gravitydrained sewers. Therefore, the sampling frequency is a crucial parameter for the collection of wastewater samples from all sewers. Inadequately low sampling frequencies unnecessarily increase random sampling uncertainty or systematic overor underestimations of pollutant loads. Our literature review reveals that a sampling frequency was only reported in every third study (see Figure 2). In less than 10% of all studies was the sampling frequency shorter than one hour; in view of individual pollutant peaks (toilet flushes or pump events) extending over a few minutes. Without any pre-experiments carried out at high frequency (samples analyzed individually) or without a sound systems analysis of the sewer system under investigation the magnitude of the sampling error is difficult, if not impossible, to quantify. The potential sampling error could only be estimated retrospectively with the modeling approach if sufficient information were available.

Flow Variations Determine the Sampling Mode “If the waste water...varies in quantity or quality with time, a continuous-flow discharge record is necessary to obtain a reliable estimate of the load.” (ISO, 1980; 95). Some studies use discharge data, measured online at high frequency, to determine the total volume of wastewater during the sampling period. This integrated volume is sufficient to calculate the environmental pollutant flux from an average concentration in a sample after sampling. However, if flow variations occur within the sampling period, they must be taken into account during sampling to weight individual subsamples and obtain a representative 24-h composite sample () average concentration). The selection of four examples of flow variations in Figure 3 is not exhaustive, and these examples are neither exceptional, nor extreme, but show the possible variety that may be encountered (corresponding catchment characteristics are summarized in the caption). Each STP and sewer system is unique due to different topology and drainage layout, which includes retention tanks and actuators

VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6027

1 Indicates what equipment is required besides sampling bottles, suction hose, and power supply. 2 Is a flow meter required for taking samples (external flow signal)? To calculate environmental loads from sampled (average) concentrations it always needs a flow meter. 3 Ideally a robust peristaltic pump with fine increments to accurately control speed with external flow signal. Linearity of pump speeds and performance (suction and pressure height) need to be checked for conditions that apply in the field. 4 Sampling volume of individual samples needs to be checked for linearity over the whole range of discharge in the sewer. 5 Check variation of individual sample size.

TABLE 1. Visualization and Brief Description of Different Sampling Modes (Adapted from 100)

FIGURE 3. Diurnal flow patterns in the influent of four STPs with different drainage system layouts and designs of inlet works infrastructure. (A) Mainly gravity fed sewers (approximately 100,000 inhabitants, 20,000 m3 d-1, the distinct peaks are caused by the intermittent operation of the fine screen, flow is measured after primary treatment). (B) Sewers with real time control (storage capacity in the sewer system is used to hold peak flows back and release wastewater at slower rate during lower flows, 150,000 population equivalents, 40,000 m3 d-1). (C) Mainly pressurized sewer system (limited storage capacity in a pump sump at the inlet of the STP, whenever this is full, the water is pumped to the STP, 4000 m3 d-1). (D) Mainly gravity fed sewers (30% of the wastewater volume is collected from pressurized sewers, stored in a large tank before the STP, and treated during the night; this STP is operated with two sequenced batch reactors (SBR) while all other STPs are flow through; approximately 45,000 inhabitants; the graph shows a wet weather situation with 20,000 m3 d-1, during dry weather the inflow volume is 10,000 m3 d-1, and the influent pattern is lowered by a constant offset of approximately 120 L s-1). (e.g., pumping stations, flow regulators, etc.). It is important to note how the fluctuations occur at different time scales: these can be over the course of a day (Figure 3A and B) reflecting “smooth” diurnal variations between the night minimum and the maximum dry weather flow with factors of about 2 (Figure 3B) or 10 (Figure 3A); or they can be dominated by pumps with an amplitude between 0 and the maximum pump rate where the frequency of pump cycles dominates the diurnal variation (Figure 3C and D). To experimentally assess the fate of a compound in fullscale streams, at least two sampling locations are necessary. For several reasons, the flow at the downstream location can be larger or smaller than that upstream. Without the proof of two streams being equal (flow measurements or description of system), a direct comparison of concentrations is not meaningful and a quantification of mass fluxes at all locations is required. While under normal conditions a STP influent more or less equals the effluent, the difference in sewers can be huge between upstream (e.g., effluent of a hospital) and downstream locations (e.g., STP influent) or influenced by losses due to CSOs or leaky sewer pipes. Furthermore, the steady consumption of some PPCPs leads to fairly constant daily loads (98) whereas concentrations can vary depending on water consumption, meteorological conditions, and type and conditions of sewers (combined vs separated systems, infiltrating groundwater). Ideally, per capita normalized pollutant loads are reported since they allow for a meaningful comparison of results at one location (variable daily water volumes due to rain events) and among different locations (different catchment sizes). Figure 2 shows that only 27% of all studies took flow variations into account (flow- or volume-proportional sampling mode, see Table 1 for a more detailed description of all sampling modes). In all other cases flow was either not adequately reported or not considered during sampling (i.e., grab samples, “composite samples” without further specifications, or a time-proportional sampling mode). Conceptually it is clear that a time-proportional sampling mode will systematically under- or overestimate pollutant loads when the flow varies, and when flow and concentration are positively or negatively correlated (93, 99). The magnitude 6028

9


of this systematic sampling error is highly dependent on the combination of flow and concentration patterns, and will be different in each catchment (see Sampling Guide at the end of this review for more details).

Discussion Efforts to Improve Chemical Analysis. The trend from analyzing “clean” samples to raw wastewater is a great challenge. This has been recognized, and advances in analytical chemistry (e.g., 101-103) are well documented with quality assurance and control (QA/QC) protocols. Substantial resources were and will be invested to detect more compounds, provide more sensitive methods, and improve trueness and precision in increasingly difficult matrices. Chemical analysis is a key part in the process of gathering environmental data but this review demonstrates that it cannot compensate for deficiencies in any preceding steps (Figure 1). This is emphasized with a comparative study described in 93 which proved that sampling uncertainty can clearly overwhelm analytical uncertainty. Some explanations, the selected approach, and findings of this study are summarized subsequently. Analytical Uncertainty Vs Sampling Uncertainty. Analytical accuracy can be assessed in the laboratory under largely controllable (repeatable) conditions. With independent replicates (subsamples), the precision of the analytical method can be quantified, and with standard addition or by spiking labeled reference compounds, the trueness can be enhanced. In contrast, the daily pollutant loads and the shortterm concentration patterns of PPCPs in a sewer cannot be assumed as constants, which complicates the quantification of “sampling precision”. Furthermore, it seems almost impossible to spike sewers (or hundreds of household appliances such as toilets, showers, etc.) with labeled reference compounds to generate a realistic PPCP pattern with a known daily mass to assess “sampling trueness”. Therefore, a unique experimental setup to provide unambiguous evidence for sampling uncertainty was required for the comparative study in 93: a number of independent sampling devices had to be operated simultaneously in various sampling modes. This approach implied that all

environmental factors, even if variable and unknown, are the same for all sampling devices. Theoretically, if sampling modes and frequencies did not matter, the variation of concentrations among the different samples of the same day should not exceed the variation (precision) due to the chemical analysis. However, for several compounds, the total observed variation was significantly larger. The variation caused by sampling was larger than 20% (up to a maximum of over 100%, single standard deviation) in more than a quarter to a half of all values investigated in two catchments. Furthermore, for substances contained in a small number of wastewater pulses (e.g., X-ray and MRI contrast agents) it was demonstrated that pollutant loads can be systematically (and grossly) underestimated when applying an inadequate sampling frequency. How relevant sampling uncertainty was/ will be in another catchment, for other substances relative to the (in)accuracy of other analytical methods can, unfortunately, not be generalized. Are Your Conclusions Valid? Ninety-five percent of the studies analyzed in our review failed to report sufficient details on sampling in the final publication to properly judge the quality of measured data and to assess to which degree the conclusions are valid (see Justification Figure 2). This is in line with the findings in ref 104: in the context of the compilation of a comprehensive database on the fate of PPCPs in STPs, details on sampling and other relevant information were also collected (if available); with the qualitative conclusion that “...this information is rarely described in a comprehensive manner...”. In an extensive review on entry routes of pharmaceuticals (1) and in two recent reviews on biodegradation of PPCPs (105, 106) sampling is only mentioned once, in a subordinate clause, and not listed as a critical factor or research need. However, in view of the results in 93, sampling uncertainty appears likely to provide at least a partial explanation for the following statement in ref 106: “Generalizing compound behavior in these [wastewater treatment] systems would allow further characterization of the fate and risk associated with PPCPs in the environment, yet this description of trends is impeded by the wide variation in removal efficiencies across therapeutic classes, treatment processes, and even among separate studies for the same individual compounds”. As a conclusion to our work, reviewing more than a decade of research on PPCPs in wastewater systems, this emphasizes the need for all authors to consistently provide more details than is currently the norm in order to judge reproducibility; either by means of a sound systems analysis, with suitable experimental preinvestigations or by applying a precautionary high-frequency, flow-proportional sampling mode to collect a representative 24-h composite sample. The additional effort required to conscientiously plan a sampling campaign, or for performing a sufficiently accurate a posteriori analysis of sampling uncertainty, is in most cases small compared to the time and effort for chemical analyses, evaluation of results, etc. With this little extra work or additional information provided, a large gain in terms of interpretation of results-or in other words “avoiding overinterpretation”-can be obtained. Together with improved chemical analyses, this will enhance the understanding of occurrence and fate of PPCPs and illicit drugs and provide a sound basis for the discussion of removal strategies. As a first step toward standard methods for determining an appropriate sampling methodology we compiled the subsequent step-by-step Sampling Guide. It can serve as a starting point encouraging others to have a closer look at sampling and to consider or reduce related uncertainties in their research. Innovative approaches could overcome, or at best eliminate, some of the difficulties with traditional sampling. Passive samplers, deployed directly in the water stream (e.g., 107), or online sensors providing continuous,

direct measurements (e.g., 108) are alternative technologies, which may be adapted to raw wastewater streams for the detection of specific PPCPs in the future.

Sampling Guide “...samples should be taken at times which will adequately represent the quality and its variations.... This approach contrasts with the choice of sampling frequency based on either subjective considerations or the amount of effort available for sampling and analysis.” (ISO, 1980; 95). In contrast to analytical methods, which were developed under controllable conditions and are described extensively to prove “sound science” with the goal of transferability to other laboratories, the composition of raw wastewater and each sewer site is more or less unique and “uncontrollable” from the experimenter’s point of view. Therefore, this sampling guide consists of a series of aspects which should be considered before sampling and reported in the final publicationseven if only in the Supporting Informationsas a sort of “sampling QA/QC” accompanying results and conclusions as commonly accepted for chemical analysis. Reduce Sampling Uncertainty with a Minimum of Effort. Norms and guidelines need to be of general validity. It is the experimenters’ responsibility to adapt them to their specific cases, accounting for the site-specific environment and dynamics: each system under investigation has its own characteristics and a universal solution cannot be provided. However, experimentally assessing short-term variations of PPCPs in sewers, as suggested with general recommendations in guidelines, can be prohibitively expensive because hundreds or thousands of samples would require analysis to cover a reasonably long period at intervals in the range of minutes for each sampling location. In view of the lack of “PPCP- and sewer-specific” sampling protocols, this Sampling Guide summarizes the current knowledge in the form of a checklist supporting laboratory personnel involved in the planning and execution phase and scientists for evaluating sampling programs. 1. Goal of the Study. Typically, scientific studies analyzing raw wastewater samples can be categorized in one of the three classes: (i) method development, (ii) demonstrating presence or absence of a contaminant, or (iii) quantification of the fate of pollutants. If the goal of the study is to assess the fate of PPCPs, involving a sampling location in the influent of a STP, or the quantification of pollutant loads at a point further upstream in the sewer system, please proceed to point 2. If the goal of the study is to develop a new analytical method for real wastewater that will be applied to determine the fate of certain PPCPs at a later stage any kind of composite sample (or even a grab sample) may be justified, because the sample is exclusively needed to assess matrix effects during chemical analysis and not to quantify full-scale pollutant fluxes. Changes in the matrix (night/day, dry/rain weather), may require several grab samples to cover the expected range. A grab sample may also be sufficient if the occurrence of a compound is to be assessed qualitatively. If the substance is detected, and sample integrity is assured (i.e., uncontaminated), this proves presence. However, if the substance is not detected, absence of the compound remains unproven. Conclusions from a composite sample with discrete samples at, e.g., 1-h intervals would strictly speaking be limited to the 24 individual points in time (pooled to reduce analytical cost) but not representative for the whole 24-h period. The simplest solution to obtain a composite sample and not miss any water packets is to run a peristaltic pump at a constant flow rate. A continuous sampling mode is recommended for screening studies (but a constant flow rate is not sufficient to assess fate, see point 2). VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6029

2. Evaluation of Flow Measurements and Choice of Sampling Mode. Besides reporting total flows per day to calculate loads, flow patterns at high temporal resolution should be considered. In view of Figure 3, it is almost certain that the flow rate in a sewer will vary significantly over time during the course of a day when sampling takes place. Therefore, a flow-proportional sampling mode must be applied. It is the only sampling mode that correctly weights the individual subsamples, which are combined to form a composite sample, and analyzed to obtain an average concentration. The volume-proportional sampling mode (mistakenly often referred to as flow-proportional) takes samples more frequently during higher flows and less frequently during lower flows; but the sampling volume remains constant. Consequently, it cannot provide a true average concentration because only the frequency changes but the individual samples are not weighted properly according to the flow in the sewer (see results in 93). This also holds true for the time-proportional sampling mode, where both frequency and sampling volume are constant. To assess diurnal variations (time-dependent loadings in the system), hourly samplessor whichever temporal resolution is required to answer the research questionsmust be collected and analyzed. If intrahour variations can be demonstrated to be negligible (with high-frequency grab samples) or assumed to be attenuated in a present retention tank (full mixing, no hydraulic shortcuts), hourly grab samples may be adequate. However, most automated sampling devices can be equipped with a rack of 12-24 bottles and in each bottle a (high-frequency) composite sample over one or two hours can be collected and should be preferred to grab sampling. It is not the sampling frequency that determines the analytical effort, but the temporal resolution necessary to answer the relevant research question. A representative 24-h composite sample is often deemed to be appropriate since with this type of sample a longer period can be covered at the same analytical cost, if diurnal variations are less important. If the underlying (true) concentration pattern is unknown, it is impossible to precisely quantify the systematic bias for a certain compound and location. Generally, in the situation of STP influents, a time-proportional mode implies that low flows, with a higher proportion of less polluted (extraneous, infiltrating) water during the night, are over-represented in a composite sample; consequently, influent loads will generally be underestimated. Note. Hydraulic pulses propagate quickly through a STP (as shown in 109) and therefore the effluent pattern of a STP can be similar to its influent. Due to the hydraulic retention time and mixing in the reactors of a STP, the concentrations in the effluent do not vary as quickly as in the influent. These circumstances imply that the frequency may be reduced (see point 3) but flow variations still need to be accounted for also when sampling in the effluent of a STP. Further recommendations on monitoring PPCPs in STP effluents can be found in 110. It is important to note that measuring flows in a sewer (or at the influent of a STP) is not a trivial task and systematic and random errors around 20%, even under optimal conditions, are reported in the literature (111-114). Usually, well maintained flow meters in pressurized pipes (completely filled) measure flow more accurately than devices in open channels (gravity flow, partially filled). Although systematic errors do not severely affect the sampling rate (because the relative fluctuations are still more or less proportionally captured), they must be assessed when calculating wastewater volumes and pollutant mass fluxes. In contrast to (short-term) random uncertainty in flow measurements they do not cancel out, and the results will, therefore, be biased. It is necessary to seek advice from professional experts 6030

9


familiar with such devices and their installations in sewers. Furthermore, it is recommended to request calibration protocols for existing flow monitoring installations. In some STPs calibrations are performed on a routine basis. 3. Determine the Appropriate Sampling Frequency. If the concentration of an analyte in the sewer over the course of a day could be shown to be constant, a grab sample would be sufficient, and equal to the “average concentration in a composite sample”. However, to prove this, a large number of grab samples collected at very short intervals (1-5 min over several hours) are necessary in a pre-experiment, and it is still not guaranteed that the obtained results can be transferred to future sampling campaigns. To be on the safe side without preinvestigations, a flow-proportional continuous sampling mode (“infinite sampling frequency”, as described in 56 and 93) should be applied. The only assumptionscommon to all sampling methods applied in sewerssis that the wastewater is completely mixed over the whole cross section of the sewer. Mathematically and statistically the continuous flow-proportional mode does not result in a “sample”, because it is not discrete (see 93 for more details). Disadvantages of this sampling mode are as follows: The continuous side stream implies either a large sampling volume or a slow sample flow rate. In the latter case it is not suitable to sample representatively for solids. Additionally, a continuous mode may not be suitable for long-term monitoring because the small tube bores may be prone to biofilm growth (the sampling hose is not regularly purged). Particle clogging is possible but unlikely to occur due to slow sample flow rates. Usually, scientific sampling campaigns are carried out over a limited number of days, not years. This contrasts with monitoring campaigns to regularly check STP performance and to check for regulatory compliance. If no continuous flow-proportional sampling mode can be applied, it should still be flow-proportional, but with a high sampling frequency. Without prior knowledge or investigations it cannot be determined what a “high” sampling frequency is. This depends on the catchment and substance under investigation, and mainly on the number of pulses containing the substance of interest, e.g., number of toilet flushes at the sampling location. This can be directly derived from prescription data (if number of scripts is available as, e.g., in Australia) or estimated according to eq 1 from annual national consumption data (Consnat) and a typical daily dose per patient (doseday). According to ref 115 it is assumed that a person goes to the toilet approximately five times per day (nT): Npulses in STP catchment )

Consnat PopSTP 1 · ·n · 365 Popnat doseday T (1)

where Popnat is the total population (relevant for the national consumption data) and PopSTP is the population in the catchment of the STP (relevant for the sampling location). As an example, in a catchment with 10,000 inhabitants, a total of approximately 50,000 toilet flushes per day would be expected. However, only a certain fraction of the population takes a specific drug. In a 10,000-inhabitant catchment in Switzerland only approximately 130 “carbamazepine-related” toilet flushes per day would be expected to be relevant to sample for carbamazepine loads (based on eq 1) and approximately 250 in Australia (based on number of prescriptions). It is obvious that this cannot be generalized for all countries or all substances. The modeling approach presented in ref 93 to determine the required sampling frequency (acceptable sampling uncertainty, respectively) suggests that there is no large difference if thousands to ten thousands of pulses are expected between sampling fre-

quencies of 15 min to 2 h. However, if the number of pulses is smaller than 1000 (or unknown) it is strongly recommended to use a sampling frequency of 15 min or smaller to minimize sampling uncertainty (see point 4 for other options). Some compounds (e.g., gadolinium; ref 93) are contained in a very small number of toilet flushes (or pump events). If such compounds are under investigation, a sampling frequency of 15 min is likely to be insufficient. An individual toilet flush lasts 5-10 s at the source and depending on the state of the house connection and the type and layout of the sewer system may pass the inlet works of a STP within 1-10 minutes. The smallest discrete sample volume that can be taken repeatedly with the sampling device and the sampling frequency define the required volume of the container for the composite sample. If it is larger than the largest vessel fitting in the (refrigerated) autosampler, it may be necessary for the vessel to be emptied more than once per day, and weighed to obtain a proper 24-h composite sample pooled in the laboratory. Important. The sampling frequency should not be reduced just because the storage capacity in the available sampling device is limited. This may lead to an inadequate sampling frequency and cannot be compensated for with a large number of samples, or a sophisticated analytical technique or statistical analysis. In view of the time and money spent on chemical analysis and subsequent interpretation of data, it is counterproductive to attempt to save costs for sampling personnel and equipment! Note. If dealing with a pressurized sewer system, it can be argued that pump sumps act as retention basins and attenuate the highly fluctuating patterns caused by individual toilet flushes. While this is true, close attention needs to be paid to the operation of the pumps (or intermittent discharges of industries, if effluents are related to PPCPs). The pumped water packets pass the influent of a STP in approximately the same period of time as it needs to empty the pump sump. If this happens intermittently with the shortest individual pump cycles lasting, e.g., 10 minutes (but from experience they can also be much shorter), the sampling frequency at the influent of the STP should be shorter than the shortest pump cycles to capture at least parts of these pulses at the influent of the STP. Missing such pumping events would imply that the pollutant loads from entire subcatchments will not be properly accounted for. The operation of pumps is usually well documented and can be obtained from sewer or STP operators. 4. Sampling Location. Access to sewers is limited, both physically and by law (confined space, occupational health and safety regulations). Irrespective, the selection of the sampling site should be an informed decision. It is often convenient for a researcher to accept (inadequate) sampling locations, and to use existing equipment without checking their suitability for purpose, which ultimately can compromise their study. This emphasizes the need for a profound systems analysis, either experimentally prior to the “real” sampling campaign or by applying a sophisticated sampling mode (see points 2 and 3). If samples should be taken in the sewer, there is often not much choice for the best location, as it needs to be set up where power is available (alternatively a generator to operate the sampling device is required) and where a flow meter can be installed to properly measure flow (to control the sampling device during sampling and calculate loads from measured concentrations). In the case of STPs, there is usually more freedom. Although untreated raw wastewater can theoretically be collected only at the influent of a STP, it may, for many studies, also be opportune to take samples after the primary clarifier (as suggested in ref 116) where concentration variations are expected to be attenuated. Depending on the layout and the hydraulic retention time this can be a

justification for longer sampling intervals. Although some particulate matter is removed in the first treatment step, concentrations of dissolved compounds are unlikely to be significantly affected. Filtration of samples before chemical analysis also removes particulate matter which is often not analyzed any further. Internal recirculation flows may enter before the effluent of the primary clarifier and need to be taken into account because they will affect the mass balances of influent loads. 5. Single vs Consecutive Days. It is advisible to perform a sampling campaign over a number of consecutive days, rather than nonconsecutive days. This particularly holds true when different locations are to be compared: Two examples are (i) the comparison of STP influent and effluent (effect of hydraulic or sludge retention time) or (ii) the comparison of hospital effluent (or an industrial site) and STP influent (variable travel times in sewers). If water packets from one day are missed in the influent or effluent, they will be captured the next day and “errors“ are limited to the first and last day of the sampling campaign. All these measures to obtain representative samples enhance the quality of monitoring campaigns: “Monitoring is collecting information on an object through repeated or continuous observation in order to determine possible changes in the object.” (117). Evidently, the fate of an analyte can only be proven if the variation in measured loads in comparative samples is larger than the (in)accuracy of one observation (e.g., one 24-h composite sample). Repeated sampling or sampling over a longer period may be necessary to determine weekly and seasonal variations, e.g., industrial vs domestic input, holidays or seasondependent consumption of pharmaceuticals. 6. Report the Complete Sampling Protocol and Communicate Expected Uncertainty. In the final manuscript (even if only in the Supporting Information), all above points should be adequately addressed. Analytical uncertainty is usually specified in method papers. However, in environmental studies, values are frequently reported without this uncertainty. If a range is indicated, the “(xy” often derives from averaging of multiple observations, and does not allow the reader to distinguish between variability and uncertainty. A complete uncertainty assessment for environmental concentrations would include at least an estimate of sampling uncertainty, analytical uncertainty, and for the calculation of loads, uncertainty of flow measurements. Optimally the accuracy of the final result is estimated by error propagation (e.g., based on a Monte Carlo analysis (2)). 7. Particulate Matter (Suspended Solids). In most studies, samples were filtered before analyses, and solids were not further analyzed. If pollutant concentrations include suspended solids, other requirements related to sample collection (e.g., 118) and sample preparation/analysis (e.g., 119, 120) become relevant. However, comprehensively covering this aspect is beyond the scope of this brief guide. 8. Alternative Technologies. Passive samplers, continuously being exposed to the water, have been suggested to overcome some of the difficulties with active sampling. However, to our knowledge, the current application of passive samplers in raw wastewater is still limited due to several practical reasons such as varying flow velocities, or clogging which influences the uptake rate (e.g., 121). Online sensors providing continuous, direct measurements were successfully applied in raw wastewater streams, but they are not yet selective for individual PPCPs (e.g., 108). 9. Sampling Literature. While the discussion of sampling for aquatic micropollutants is very sparse in the literature, numerous text books and lecture materials are available on more general sampling strategies and prerequisites for environmental data to justify the application of certain statistical analyses. The following list of references is limited to generally available scientific literature, norms, and guideVOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6031

lines (not regarded as exhaustive). The crude categorization may help you to quickly find more, useful information on issues related to sampling: • General sampling/experimental design literature: 91, 92, 99, 117, 122-126 • ISO norms (water quality sampling from different water bodies): 3, 94, 95, 100, 127 • Sampling in surface waters: 128-130 • Assessing reliability of sampling strategies in sewer systems and receiving waters: 131 • Monitoring STP effluents: 110

Acknowledgments We thank the three anonymous reviewers whose time, insight, and critical comments improved the final manuscript. Furthermore, we acknowledge the financial support by the Urban Water Security Research Alliance and the Swiss National Science Foundation (Grant PBEZP2-122958 awarded to C. Ort).

Literature Cited (1) Daughton, C. G.; Ruhoy, I. S. Environmental footprint of pharmaceuticals: the significance of factors beyond direct excretion to sewers. Environ. Toxicol. Chem. 2009, 28 (12), 2495–2521. (2) Joss, A.; Siegrist, H.; Ternes, T. A. Are we about to upgrade wastewater treatment for removing organic micropollutants? Water Sci. Technol. 2008, 57 (2), 251–255. (3) U.S. EPA. Handbook for Sampling and Sample Preservation of Water and Wastewater; 600/4-82-029; Environmental Monitoring and Support Laboratory, 1982. (4) Andersen, H.; Siegrist, H.; Halling-Sorensen, B.; Ternes, T. A. Fate of estrogens in a municipal sewage treatment plant. Environ. Sci. Technol. 2003, 37 (18), 4021–4026. (5) Baronti, C.; Curini, R.; D’Ascenzo, G.; Di Corcia, A.; Gentili, A.; Samperi, R. Monitoring natural and synthetic estrogens at activated sludge sewage treatment plants and in a receiving river water. Environ. Sci. Technol. 2000, 34 (24), 5059–5066. (6) Batt, A. L.; Kim, S.; Aga, D. S. Enhanced biodegradation of iopromide and trimethoprim in nitrifying activated sludge. Environ. Sci. Technol. 2006, 40 (23), 7367–7373. (7) Bendz, D.; Paxeus, N. A.; Ginn, T. R.; Loge, F. J. Occurrence and fate of pharmaceutically active compounds in the environment, a case study: Hoje River in Sweden. J. Hazard. Mater. 2005, 122 (3), 195–204. (8) Bernhard, M.; Muller, J.; Knepper, T. R. Biodegradation of persistent polar pollutants in wastewater: Comparison of an optimized lab-scale membrane bioreactor and activated sludge treatment. Water Res. 2006, 40 (18), 3419–3428. (9) Boillot, C.; Bazin, C.; Tissot-Guerraz, F.; Droguet, J.; Perraud, M.; Cetre, J. C.; Trepo, D.; Perrodin, Y. Daily physicochemical, microbiological and ecotoxicological fluctuations of a hospital effluent according to technical and care activities. Sci. Total Environ. 2008, 403 (1-3), 113–129. (10) Boleda, M. R.; Galceran, M. T.; Ventura, F. Monitoring of opiates, cannabinoids and their metabolites in wastewater, surface water and finished water in Catalonia, Spain. Water Res. 2009, 43 (4), 1126–1136. (11) Bones, J.; Thomas, K. V.; Paull, B. Using environmental analytical data to estimate levels of community consumption of illicit drugs and abused pharmaceuticals. J. Environ. Monit. 2007, 9 (7), 701–707. (12) Buser, H. R.; Poiger, T.; Muller, M. D. Occurrence and environmental behavior of the chiral pharmaceutical drug ibuprofen in surface waters and in wastewater. Environ. Sci. Technol. 1999, 33 (15), 2529–2535. (13) Carballa, M.; Omil, F.; Lema, J. M. Calculation methods to perform mass balances of micropollutants in sewage treatment plants. Application to pharmaceutical and personal care products (PPCPs). Environ. Sci. Technol. 2007, 41 (3), 884–890. (14) Carballa, M.; Omil, F.; Lema, J. M.; Llompart, M.; Garcia, C.; Rodriguez, I.; Gomez, M.; Ternes, T. Behaviour of pharmaceuticals and personal care products in a sewage treatment plant of northwest Spain. Water Sci. Technol. 2005, 52 (8), 29–35. 6032

9


(15) Carballa, M.; Omil, F.; Lema, J. M.; Llompart, M.; Garcia-Jares, C.; Rodriguez, I.; Gomez, M.; Ternes, T. Behavior of pharmaceuticals, cosmetics and hormones in a sewage treatment plant. Water Res. 2004, 38 (12), 2918–2926. (16) Castiglioni, S.; Bagnati, R.; Fanelli, R.; Pomati, F.; Calamari, D.; Zuccato, E. Removal of pharmaceuticals in sewage treatment plants in Italy. Environ. Sci. Technol. 2006, 40 (1), 357–363. (17) Castiglioni, S.; Zuccato, E.; Crisci, E.; Chiabrando, C.; Fanelli, R.; Bagnati, R. Identification and measurement of illicit drugs and their metabolites in urban wastewater by liquid chromatography-tandem mass spectrometry. Anal. Chem. 2006, 78 (24), 8421–8429. (18) Celiz, M. D.; Perez, S.; Barcelo, D.; Aga, D. S. Trace analysis of polar pharmaceuticals in wastewater by LC-MS-MS: comparison of membrane bioreactor and activated sludge systems. J. Chromatogr. Sci. 2009, 47 (1), 19–25. (19) Chiaia, A. C.; Banta-Green, C.; Field, J. Eliminating Solid Phase Extraction with Large-Volume Injection LC/MS/MS: Analysis of Illicit and Legal Drugs and Human Urine Indicators in US Wastewaters. Environ. Sci. Technol. 2008, 42 (23), 8841–8848. (20) Clara, M.; Kreuzinger, N.; Strenn, B.; Gans, O.; Kroiss, H. The solids retention time - a suitable design parameter to evaluate the capacity of wastewater treatment plants to remove micropollutants. Water Res. 2005, 39 (1), 97–106. (21) Clara, M.; Strenn, B.; Ausserleitner, M.; Kreuzinger, N. Comparison of the behaviour of selected micropollutants in a membrane bioreactor and a conventional wastewater treatment plant. Water Sci. Technol. 2004, 50 (5), 29–36. (22) Clara, M.; Strenn, B.; Gans, O.; Martinez, E.; Kreuzinger, N.; Kroiss, H. Removal of selected pharmaceuticals, fragrances and endocrine disrupting compounds in a membrane bioreactor and conventional wastewater treatment plants. Water Res. 2005, 39 (19), 4797–4807. (23) Feldmann, D. F.; Zuehlke, S.; Heberer, T. Occurrence, fate and assessment of polar metamizole (dipyrone) residues in hospital and municipal wastewater. Chemosphere 2008, 71 (9), 1754– 1764. (24) Gautam, A. K.; Kumar, S.; Sabumon, P. C. Preliminary study of physico-chemical treatment options for hospital wastewater. J. Environ. Manage. 2007, 83 (3), 298–306. (25) Gobel, A.; McArdell, C. S.; Joss, A.; Siegrist, H.; Giger, W. Fate of sulfonamides, macrolides, and trimethoprim in different wastewater treatment technologies. Sci. Total Environ. 2007, 372 (2-3), 361–371. (26) Gomez, M. J.; Bueno, M. J. M.; Lacorte, S.; Fernandez-Alba, A. R.; Aguera, A. Pilot survey monitoring pharmaceuticals and related compounds in a sewage treatment plant located on the Mediterranean coast. Chemosphere 2007, 66 (6), 993– 1002. (27) Gupta, P.; Mathur, N.; Bhatnagar, P.; Nagar, P.; Srivastava, S. Genotoxicity evaluation of hospital wastewaters. Ecotoxicol. Environ. Saf. 2009, 72 (7), 1925–1932. (28) Hartmann, A.; Alder, A. C.; Koller, T.; Widmer, R. M. Identification of fluoroquinolone antibiotics as the main source of umuC genotoxicity in native hospital wastewater. Environ. Toxicol. Chem. 1998, 17 (3), 377–382. (29) Heberer, T.; Feldmann, D. Contribution of effluents from hospitals and private households to the total loads of diclofenac and carbamazepine in municipal sewage effluents - modeling versus measurements. J. Hazard. Mater. 2005, 122 (3), 211– 218. (30) Heidler, J.; Halden, R. U. Mass balance assessment of triclosan removal during conventional sewage treatment. Chemosphere 2007, 66 (2), 362–369. (31) Heidler, J.; Sapkota, A.; Halden, R. U. Partitioning, persistence, and accumulation in digested sludge of the topical antiseptic triclocarban during wastewater treatment. Environ. Sci. Technol. 2006, 40 (11), 3634–3639. (32) Jones, O. A. H.; Voulvoulis, N.; Lester, J. N. The occurrence and removal of selected pharmaceutical compounds in a sewage treatment works utilising activated sludge treatment. Environ. Pollut. 2007, 145 (3), 738–744. (33) Joss, A.; Andersen, H.; Ternes, T.; Richle, P. R.; Siegrist, H. Removal of estrogens in municipal wastewater treatment under aerobic and anaerobic conditions: Consequences for plant optimization. Environ. Sci. Technol. 2004, 38 (11), 3047– 3055. (34) Kasprzyk-Hordern, B.; Dinsdale, R. M.; Guwy, A. J. Illicit drugs and pharmaceuticals in the environment - Forensic applications of environmental data. Part 1: Estimation of the usage

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

(43)

(44)

(45)

(46)

(47)

(48)

(49)

(50)

(51)

(52)

(53)

of drugs in local communities. Environ. Pollut. 2009, 157 (6), 1773–1777. Kosonen, J.; Kronberg, L. The occurrence of antihistamines in sewage waters and in recipient rivers. Environ. Sci. Pollut. Res. 2009, 16 (5), 555–564. Kreuzinger, N.; Clara, M.; Strenn, B.; Kroiss, H. Relevance of the sludge retention time (SRT) as design criteria for wastewater treatment plants for the removal of endocrine disruptors and pharmaceuticals from wastewater. Water Sci. Technol. 2004, 50 (5), 149–156. Kummerer, K.; Steger Hartmann, T.; Meyer, M. Biodegradability of the anti-tumour agent ifosfamide and its occurrence in hospital effluents and communal sewage. Water Res. 1997, 31 (11), 2705–2710. Kunnemeyer, J.; Terborg, L.; Meermann, B.; Brauckmann, C.; Scheffer, I. M. A.; Karst, U. Speciation Analysis of Gadolinium Chelates in Hospital Effluents and Wastewater Treatment Plant Sewage by a Novel HILIC/ICP-MS Method. Environ. Sci. Technol. 2009, 43 (8), 2884–2890. Kupper, T.; Plagellat, C.; Braendli, R. C.; de Alencastro, L. F.; Grandjean, D.; Tarradellas, J. Fate and removal of polycyclic musks, UV filters and biocides during wastewater treatment. Water Res. 2006, 40 (14), 2603–2612. Langford, K. H.; Thomas, K. V. Determination of pharmaceutical compounds in hospital effluents and their contribution to wastewater treatment works. Environ. Int. 2009, 35 (5), 766– 770. Leclercq, M.; Mathieu, O.; Gomez, E.; Casellas, C.; Fenet, H.; Hillaire-Buys, D. Presence and Fate of Carbamazepine, Oxcarbazepine, and Seven of Their Metabolites at Wastewater Treatment Plants. Arch. Environ. Contam. Toxicol. 2009, 56 (3), 408–415. Lenz, K.; Mahnik, S. N.; Weissenbacher, N.; Mader, R. M.; Krenn, P.; Hann, S.; Koellensperger, G.; Uhl, M.; Knasmuller, S.; Ferk, F.; Bursch, W.; Fuerhacker, M. Monitoring, removal and risk assessment of cytostatic drugs in hospital wastewater. Water Sci. Technol. 2007, 56 (12), 141–149. Lindqvist, N.; Tuhkanen, T.; Kronberg, L. Occurrence of acidic pharmaceuticals in raw and treated sewages and in receiving waters. Water Res. 2005, 39 (11), 2219–2228. Lishman, L.; Smyth, S. A.; Sarafin, K.; Kleywegt, S.; Toito, J.; Peart, T.; Lee, B.; Servos, M.; Beland, M.; Seto, P. Occurrence and reductions of pharmaceuticals and personal care products and estrogens by municipal wastewater treatment plants in Ontario, Canada. Sci. Total Environ. 2006, 367 (23), 544–558. Loganathan, B.; Phillips, M.; Mowery, H.; Jones-Lepp, T. L. Contamination profiles and mass loadings of macrolide antibiotics and illicit drugs from a small urban wastewater treatment plant. Chemosphere 2009, 75 (1), 70–77. Mari, F.; Politi, L.; Biggeri, A.; Accetta, G.; Trignano, C.; Di Padua, M.; Bertol, E. Cocaine and heroin in waste water plants: A 1-year study in the city of Florence, Italy. Forensic Sci. Int. 2009, 189 (1-3), 88–92. Matamoros, V.; Arias, C.; Brix, H.; Bayona, J. M. Preliminary screening of small-scale domestic wastewater treatment systems for removal of pharmaceutical and personal care products. Water Res. 2009, 43 (1), 55–62. Matamoros, V.; Hijosa, M.; Bayona, J. M. Assessment of the pharmaceutical active compounds removal in wastewater treatment systems at enantiomeric level. Ibuprofen and naproxen. Chemosphere 2009, 75 (2), 200–205. Maurer, M.; Escher, B. I.; Richle, P.; Schaffner, C.; Alder, A. C. Elimination of beta-blockers in sewage treatment plants. Water Res. 2007, 41 (7), 1614–1622. McAvoy, D. C.; Schatowitz, B.; Jacob, M.; Hauk, A.; Eckhoff, W. S. Measurement of triclosan in wastewater treatment systems. Environ. Toxicol. Chem. 2002, 21 (7), 1323–1329. Mullot, J. U.; Karolak, S.; Fontova, A.; Huart, B.; Levi, Y. Development and validation of a sensitive and selective method using GC/MS-MS for quantification of 5-fluorouracil in hospital wastewater. Anal. Bioanal. Chem. 2009, 394 (8), 2203–2212. Musolff, A.; Leschik, S.; Mo¨der, M.; Strauch, G.; Reinstorf, F.; Schirmer, M. Temporal and spatial patterns of micropollutants in urban receiving waters. Environ. Pollut. 2009, 157 (11), 3069– 3077. Nakada, N.; Tanishima, T.; Shinohara, H.; Kiri, K.; Takada, H. Pharmaceutical chemicals and endocrine disrupters in municipal wastewater in Tokyo and their removal during activated sludge treatment. Water Res. 2006, 40 (17), 3297–3303.

(54) Okuda, T.; Yamashita, N.; Tanaka, H.; Matsukawa, H.; Tanabe, K. Development of extraction method of pharmaceuticals and their occurrences found in Japanese wastewater treatment plants. Environ. Int. 2009, 35 (5), 815–820. (55) Ort, C.; Gujer, W. Sampling for representative micropollutant loads in sewer systems. Water Sci. Technol. 2006, 54 (6-7), 169–176. (56) Ort, C.; Lawrence, M. G.; Reungoat, J.; Eaglesham, G.; Carter, S.; Keller, J. Determining the fraction of pharmaceutical residues in wastewater originating from a hospital. Water Res. 2010, 44 (2), 605–615. (57) Ort, C.; Schaffner, C.; Giger, W.; Gujer, W. Modeling stochastic load variations in sewer systems. Water Sci. Technol. 2005, 52 (5), 113–122. (58) Paxeus, N. Removal of selected non-steroidal anti-inflammatory drugs (NSAIDs), gemfibrozil, carbamazepine, beta-blockers, trimethoprim and triclosan in conventional wastewater treatment plants in five EU countries and their discharge to the aquatic environment. Water Sci. Technol. 2004, 50 (5), 253– 260. (59) Peng, X. Z.; Wang, Z. D.; Kuang, W. X.; Tan, J. H.; Li, K. A preliminary study on the occurrence and behavior of sulfonamides, ofloxacin and chloramphenicol antimicrobials in wastewaters of two sewage treatment plants in Guangzhou, China. Sci. Total Environ. 2006, 371 (1-3), 314–322. (60) Quintana, J. B.; Weiss, S.; Reemtsma, T. Pathways and metabolites of microbial degradation of selected acidic pharmaceuticals and their occurrence in municipal wastewater treated by a membrane bioreactor. Water Res. 2005, 39 (12), 2654–2664. (61) Radjenovic, J.; Petrovic, M.; Barcelo, D. Analysis of pharmaceuticals in wastewater and removal using a membrane bioreactor. Anal. Bioanal. Chem. 2007, 387 (4), 1365–1377. (62) Radjenovic, J.; Petrovic, M.; Barcelo, D. Fate and distribution of pharmaceuticals in wastewater and sewage sludge of the conventional activated sludge (CAS) and advanced membrane bioreactor (MBR) treatment. Water Res. 2009, 43 (3), 831–841. (63) Rodriguez, I.; Quintana, J. B.; Carpinteiro, J.; Carro, A. M.; Lorenzo, R. A.; Cela, R. Determination of acidic drugs in sewage water by gas chromatography-mass spectrometry as tert.butyldimethylsilyl derivatives. J. Chromatogr. A 2003, 985 (12), 265–274. (64) Santos, J. L.; Aparicio, I.; Callejon, M.; Alonso, E. Occurrence of pharmaceutically active compounds during 1-year period in wastewaters from four wastewater treatment plants in Seville (Spain). J. Hazard. Mater. 2009, 164 (2-3), 1509–1516. (65) Simonich, S. L.; Federle, T. W.; Eckhoff, W. S.; Rottiers, A.; Webb, S.; Sabaliunas, D.; De Wolf, W. Removal of fragrance materials during US and European wastewater treatment. Environ. Sci. Technol. 2002, 36 (13), 2839–2847. (66) Smook, T. M.; Zho, H.; Zytner, R. G. Removal of ibuprofen from wastewater: comparing biodegradation in conventional, membrane bioreactor, and biological nutrient removal treatment systems. Water Sci. Technol. 2008, 57 (1), 1–8. (67) Spongberg, A. L.; Witter, J. D. Pharmaceutical compounds in the wastewater process stream in Northwest Ohio. Sci. Total Environ. 2008, 397 (1-3), 148–157. (68) StegerHartmann, T.; Kummerer, K.; Schecker, J. Trace analysis of the antineoplastics ifosfamide and cyclophosphamide in sewage water by two-step solid-phase extraction and gas chromatography mass spectrometry. J. Chromatogr. A 1996, 726 (1-2), 179–184. (69) Stumpf, M.; Ternes, T. A.; Wilken, R. D.; Rodrigues, S. V.; Baumann, W. Polar drug residues in sewage and natural waters in the state of Rio de Janeiro, Brazil. Sci. Total Environ. 1999, 225 (1-2), 135–141. (70) Tauxe-Wuersch, A.; De Alencastro, L. F.; Grandjean, D.; Tarradellas, J. Occurrence of several acidic drugs in sewage treatment plants in Switzerland and risk assessment. Water Res. 2005, 39 (9), 1761–1772. (71) Ternes, T. A. Occurrence of drugs in German sewage treatment plants and rivers. Water Res. 1998, 32 (11), 3245–3260. (72) Ternes, T. A.; Bonerz, M.; Herrmann, N.; Teiser, B.; Andersen, H. R. Irrigation of treated wastewater in Braunschweig, Germany: An option to remove pharmaceuticals and musk fragrances. Chemosphere 2007, 66 (5), 894–904. (73) Ternes, T. A.; Hirsch, R. Occurrence and behavior of X-ray contrast media in sewage facilities and the aquatic environment. Environ. Sci. Technol. 2000, 34 (13), 2741–2748. VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6033

(74) Ternes, T. A.; Stumpf, M.; Mueller, J.; Haberer, K.; Wilken, R. D.; Servos, M. Behavior and occurrence of estrogens in municipal sewage treatment plants - I. Investigations in Germany, Canada and Brazil. Sci. Total Environ. 1999, 225 (1-2), 81–90. (75) Thomas, K. V.; Dye, C.; Schlabach, M.; Langford, K. H. Source to sink tracking of selected human pharmaceuticals from two Oslo city hospitals and a wastewater treatment works. J. Environ. Monit. 2007, 9 (12), 1410–1418. (76) Thompson, A.; Griffin, P.; Stuetz, R.; Cartmell, E. The fate and removal of triclosan during wastewater treatment. Water Environ. Res. 2005, 77 (1), 63–67. (77) van Nuijs, A. L. N.; Pecceu, B.; Theunis, L.; Dubois, N.; Charlier, C.; Jorens, P. G.; Bervoets, L.; Blust, R.; Neels, H.; Covaci, A. Spatial and temporal variations in the occurrence of cocaine and benzoylecgonine in waste- and surface water from Belgium and removal during wastewater treatment. Water Res. 2009, 43 (5), 1341–1349. (78) Vasskog, T.; Berger, U.; Samuelsen, P. J.; Kallenborn, R.; Jensen, E. Selective serotonin reuptake inhibitors in sewage influents and effluents from Tromso, Norway. J. Chromatogr. A 2006, 1115 (1-2), 187–195. (79) Vieno, N.; Tuhkanen, T.; Kronberg, L. Elimination of pharmaceuticals in sewage treatment plants in Finland. Water Res. 2007, 41 (5), 1001–1012. (80) Vieno, N. M.; Tuhkanen, T.; Kronberg, L. Seasonal variation in the occurrence of pharmaceuticals in effluents from a sewage treatment plant and in the recipient water. Environ. Sci. Technol. 2005, 39 (21), 8220–8226. (81) Waltman, E. L.; Venables, B. J.; Waller, W. Z. Triclosan in a North Texas wastewater treatment plant and the influent and effluent of an experimental constructed wetland. Environ. Toxicol. Chem. 2006, 25 (2), 367–372. (82) Watkinson, A. J.; Murby, E. J.; Costanzo, S. D. Removal of antibiotics in conventional and advanced wastewater treatment: Implications for environmental discharge and wastewater recycling. Water Res. 2007, 41, 4164–4176. (83) Watkinson, A. J.; Murby, E. J.; Kolpin, D. W.; Costanzo, S. D. The occurrence of antibiotics in an urban watershed: From wastewater to drinking water. Sci. Total Environ. 2009, 407, 2711–2723. (84) Weissbrodt, D.; Kovalova, L.; Ort, C.; Pazhepurackel, V.; Moser, R.; Hollender, J.; Siegrist, H.; McArdell, C. S. Mass flows of X-ray Contrast Media and Cytostatics in Hospital Wastewater. Environ. Sci. Technol. 2009, 43 (13), 4810–4817. (85) Wick, A.; Fink, G.; Joss, A.; Siegrist, H.; Ternes, T. A. Fate of beta blockers and psycho-active drugs in conventional wastewater treatment. Water Res. 2009, 43 (4), 1060–1074. (86) Xu, W. H.; Zhang, G.; Li, X. D.; Zou, S. C.; Li, P.; Hu, Z. H.; Li, J. Occurrence and elimination of antibiotics at four sewage treatment plants in the Pearl River Delta (PRD), South China. Water Res. 2007, 41, 4526–4534. (87) Yu, C. P.; Chu, K. H. Occurrence of pharmaceuticals and personal care products along the West Prong Little Pigeon River in east Tennessee, USA. Chemosphere 2009, 75 (10), 1281– 1286. (88) Yu, J. T.; Bouwer, E. J.; Coelhan, M. Occurrence and biodegradability studies of selected pharmaceuticals and personal care products in sewage effluent. Agric. Water Manage. 2006, 86 (1-2), 72–80. (89) Zorita, S.; Martensson, L.; Mathiasson, L. Occurrence and removal of pharmaceuticals in a municipal sewage treatment system in the south of Sweden. Sci. Total Environ. 2009, 407 (8), 2760–2770. (90) Zuccato, E.; Chiabrando, C.; Castiglioni, S.; Bagnati, R.; Fanelli, R. Estimating community drug abuse by wastewater analysis. Environ. Health Perspect. 2008, 116 (8), 1027–1032. (91) Gy, P. M. Sampling - the Foundation-Block of Analysis. Mikrochim. Acta 1991, 2 (1-6), 457–466. (92) Ramsey, M. H.; Thompson, M. Uncertainty from sampling, in the context of fitness for purpose. Accredit. Qual. Assur. 2007, 12 (10), 503–513. (93) Ort, C.; Lawrence, M. G.; Reungoat, J.; Mueller, J. F. Sampling for PPCPs in wastewater systems: A comparison of different sampling modes and optimization strategies. Environ. Sci. Technol. in press, doi: 10.1021/es100778d. (94) ISO. Water quality - Sampling - Part 10: Guidance on sampling of waste waters; ISO 5667-10; ISO: Gene`ve, Switzerland, 1992. (95) ISO. Water quality - Sampling - Part 1: Guidance on the design of sampling programmes; ISO 5667-1; ISO: Gene`ve, Switzerland, 1980. (96) Clesceri, L. S.; Greenberg, A. E.; Eaton, A. D. Standard Methods for the Examination of Water and Wastewater; American Public 6034

9


(97)

(98)

(99) (100) (101)

(102) (103)

(104)

(105)

(106)

(107)

(108)

(109)

(110)

(111)

(112)

(113) (114)

(115)

(116) (117) (118) (119)

(120)

Health Association: Washington, DC, 1998; Vol. 20, ISBN 0875532357 Rieckermann, J.; Neumann, M.; Ort, C.; Huisman, J. L.; Gujer, W. Dispersion coefficients of sewers from tracer experiments. Water Sci. Technol. 2005, 52 (5), 123–133. Ort, C.; Hollender, J.; Schaerer, M.; Siegrist, H. R. Model-Based Evaluation of Reduction Strategies for Micropollutants from Wastewater Treatment Plants in Complex River Networks. Environ. Sci. Technol. 2009, 43 (9), 3214–3220. Minkkinen, P. Weighting error - is it significant in process analysis? WCBS3, Third World Conference on Sampling and Blending, 22-25 October, 2007; 2007, 1/2007, 59–68. ISO. Water quality - Sampling - Part 2: Guidance on sampling techniques; ISO 5667-2; ISO: Gene`ve, Switzerland, 1991. Morley, M. C.; Snow, D. D.; Kuppig, V. Emerging chemicals and analytical methods. Water Environ. Res. 2005, 77, 561– 658. Richardson, S. D. Water Analysis: Emerging Contaminants and Current Issues. Anal. Chem. 2009, 81 (12), 4645–4677. Wong, C. S.; MacLeod, S. L. JEM spotlight: recent advances in analysis of pharmaceuticals in the aquatic environment. J. Environ. Monit. 2009, 11 (5), 923–936. Mie`ge, C.; Choubert, J. M.; Ribeiro, L.; Euse`be, M.; Coquery, M. Fate of pharmaceuticals and personal care products in wastewater treatment plants - Conception of a database and first results. Environ. Pollut. 2009, 157 (5), 1721–1726. Kagle, J.; Porter, A. W.; Murdoch, R. W.; Rivera-Cancel, G.; Hay, A. G. Biodegradation of Pharmaceutical and Personal Care Products. Adv. Appl. Mircobiol. 2009, 67, 65–108. Onesios, K. M.; Yu, J. T.; Bouwer, E. J. Biodegradation and removal of pharmaceuticals and personal care products in treatment systems: a review. Biodegradation 2009, 20 (4), 441– 466. Macleod, S. L.; McClure, E. L.; Wong, C. S. Laboratory calibration and field deployment of the polar organic chemical integrative sampler for pharmaceuticals and personal care products in wastewater and surface water. Environ. Toxicol. Chem. 2007, 26 (12), 2517–2529. Hoppe, H.; Messmann, S.; Giga, A.; Gruning, H. Options and limits of quantitative and qualitative online-monitoring of industrial discharges into municipal sewage systems. Water Sci. Technol. 2009, 60 (4), 859–867. Ort, C.; Siegrist, H. Assessing wastewater dilution in small rivers with high resolution conductivity probes. Water Sci. Technol. 2009, 59 (8), 1593–1601. Zhang, Z.; Grover, D. P.; Zhou, J. L. Monitoring of Pharmaceutical Residues in Sewage Effluents (Chapter 14). In Handbook of Water Purity and Quality: Monitoring of Pharmaceutical Residues in Sewage Effluents; Ahuja, S., Ed.; IWA: 2009. Bertrand-Krajewski, J. L.; Bardin, J. P.; Mourad, M.; Beranger, Y. Accounting for sensor calibration, data validation, measurement and sampling uncertainties in monitoring urban drainage systems. Water Sci. Technol. 2003, 47 (2), 95–102. Gromaire, M. C.; Chebbo, G. Pollutant concentration measurement uncertainties in sewage. Houille Blanche-Revue Internationale De L Eau 2001, (6-7), 109–114. Port, E. Requirements for the in-house control of WWTPs (in German). WAR TH Darmstadt 1994, 75, 353–361. Thomann, M., Data control of WWTPs with mass balances, experiments and statistical methods (in German). Dissertation, ETH Zu ¨ rich, No 14824. 2002. Rauch, W.; Brockmann, D.; Peters, I.; Larsen, T. A.; Gujer, W. Combining urine separation with waste design: an analysis using a stochastic model for urine production. Water Res. 2003, 37 (3), 681–689. Ternes, T.; Joss, A. Human Pharmaceuticals, Hormones and Fragrances: The challenge of micropollutants in urban water management; IWA Publishing, 2006; ISBN: 1843390930. de Gruijter, J.; Brus, D.; Bierkens, M.; Knotters, M., Sampling for Natural Resource Monitoring; Springer: New York, 2006. Larrarte, F. Suspended solids within sewers: an experimental study. Environ. Fluid Mech. 2008, 8 (3), 249–261. Clark, S. E.; Siu, C. Y. S. Measuring solids concentration in stormwater runoff: Comparison of analytical methods. Environ. Sci. Technol. 2008, 42 (2), 511–516. Deo, R. P.; Halden, R. U. Comment on “The removal of pharmaceuticals, personal care products, endocrine disruptors and illicit drugs during wastewater treatment and its impact on the quality of receiving waters” by Barbara KasprzykHordern, Richard M. Dinsdale, Alan J. Guwy. Water Res. 2010, 44 (8), 2685–2687.

(121) Tan, B. L. L.; Hawker, D. W.; Muller, J. F.; Leusch, F. D. L.; Tremblay, L. A.; Chapman, H. F. Comprehensive study of endocrine disrupting compounds using grab and passive sampling at selected wastewater treatment plants in South East Queensland, Australia. Environ. Int. 2007, 33 (5), 654– 669. (122) Gy, P. M. The analytical and economic importance of correctness in sampling. Anal. Chim. Acta 1986, 190 (1), 13–23. (123) Leardi, R. Experimental design in chemistry: A tutorial. Anal. Chim. Acta 2009, 652 (1-2), 161–172. (124) Minkkinen, P. O.; Esbensen, K. H. Grab vs. composite sampling of particulate materials with significant spatial heterogeneity--A simulation study of “correct sampling errors”. Anal. Chim. Acta 2009, 653 (1), 59–70. (125) Paakkunainen, M.; Reinikainen, S. P.; Minkkinen, P. Estimation of the variance of sampling of process analytical and environmental emissions measurements. Chemometrics Intell. Lab. Syst. 2007, 88 (1), 26–34. (126) Petersen, L.; Esbensen, K. H. Representative process sampling for reliable data analysis - a tutorial. J. Chemometrics 2005, 19 (1112), 625–647.

(127) ISO. Accuracy (trueness and precision) of measurement method and results - Part 1: General principles and definitions; ISO 5725-1; ISO: Gene`ve, Switzerland, 1994. (128) Allan, I. J.; Vrana, B.; Greenwood, R.; Mills, G. A.; Roig, B.; Gonzalez, C. A “toolbox” for biological and chemical monitoring requirements for the European Union’s Water Framework Directive. Talanta 2006, 69 (2), 302–322. (129) Hanke, G.; Wollgast, J.; Loos, R.; Jime´nez, J. C.; Umlauf, G.; Mariani, G.; Mu ¨ ller, A.; Huber, T.; Christoph, E. H.; Locoro, G.; Zaldı´var, J. M.; Boidoglio, G. Comparison of Monitoring Approaches for Selected Priority Pollutants in Surface Water; JRC Scientific and Technical Reports, 2007. (130) U.S. EPA. Load estimation techniques (Chapter 7). In National Management Measures to Control Nonpoint Source Pollution from Agriculture; Washington, DC, 2003. (131) Rossi, L.; Rumley, L.; Ort, C.; Minkkinen, P.; Barry, D. A.; Che`vre, N., Sampling-helper: a web-based tool to assess the reliability of sampling strategies in sewer systems and receiving waters. Water Sci. Technol. accepted.

ES100779N

VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6035

Sampling for Pharmaceuticals and Personal Care Products (PPCPs

Recommend Documents