Meeting Report: Pharmaceuticals in Water—An Interdisciplinary Approach to a Public Health Challenge

Background The presence of pharmaceuticals in aquatic environments and in drinking water has prompted significant public interest regarding potential adverse ecological effects and risks to human health. Objectives The Environmental Health Summit held in North Carolina, 10–11 November 2008, explored the issues associated with the presence and relative risk of trace levels of pharmaceuticals in water. More than 150 participants from government organizations and institutions, academia, industry, water utilities, and public interest groups participated in discussions aimed at evaluating the current knowledge on this issue and at identifying research gaps and innovative solution-oriented recommendations. Discussion We present different aspects related to the subject that were discussed at the summit, including the source, fate, and transport of pharmaceuticals, their exposure effects and potential risks to human and ecosystems, and the best management practices to address these issues. Recommendations placed emphasis on research needs as well as education, communication, prevention, and intervention programs, and other public health solutions and actions. Conclusions Despite rising concerns about the presence of trace amounts of pharmaceuticals in drinking water, little evidence is currently available that associates these chemicals with adverse human health risks. In order to prioritize which pharmaceutical chemicals could potentially pose the highest risk to consumers and the environment, the summit participants concluded that more studies are needed to generate meaningful and accurate data.

volume 118 | number 7 | July 2010 • Environmental Health Perspectives Research When drugs are prescribed prophylactically or in response to an acute or chronic illness, only a portion of the active ingredient of the drug is metabolized (Ternes 1998). The non metabolized parent compound, as well as the metabolites, enter the natural aquatic environ ment through waste discharges into receiving streams, which may reach downstream rec reational lakes or even the intakes of drink ing water treatment plants. The ubiquitous worldwide use of pharmaceuticals has led to the presence of low but measurable partper trillion levels of these biochemically active compounds in surface and groundwater. In the United States, the most comprehensive monitoring of these chemicals in water was performed between 1999 and 2000 by the U.S. Geological Survey in many of the nation's streams (Kolpin et al. 2002). This study and others identified many products, including analgesics, antiinflammatories, antibiotics, anti epileptics, betablockers, blood lipid regulators, antidepressants, contrast media, oral contra ceptives, and cytostatic and broncho dilator drugs in sewage, surface water, ground water, and drinking water (Costanzo et al. 2005;Daughton and Ternes 1999;Glassmeyer et al. 2008;Heberer 2002;Stackelberg et al. 2007;Ternes 1998;Ye et al. 2007).
Pharmaceutical compounds are designed to have biochemical activity in target organisms at relatively low concentrations. Therefore, at the low partpertrillion levels found in aquatic environments, which are far below human therapeutic doses, and even compared with the much higher levels of rou tinely measured water quality chemicals, there is concern that some of these compounds could have an ecological and human health effect. There is, for example, evidence from rodent and fish studies that suggests some endocrinedisrupting compounds, includ ing those found in prescribed synthetic hor mones, may contribute to tumor formation in humans (Birnbaum and Fenton 2003). The identification of pharmaceuticals in surface waters has opened a Pandora's box in a way that other discoveries, such as the transmission of waterborne infectious diseases, have not. This might be because the subject brings to light that we may be drinking water that has, in part, been recycled through several other human bodies. Several countries have turned down opportunities to implement waste water recycling to supplement potable water needs during drought conditions (Casani et al. 2005;Wilson and Pfaff 2008), but it is hard to conceal the fact that many people in those countries have been consuming wastewater impacted drinking water for many years with out any evidence of negative exposurerelated health outcomes. Nevertheless, the presence of pharmaceuticals in surface waters and their detection, at much lower levels, in some household drinking waters has prompted sig nificant public and mass media interest (Donn et al. 2008). The potential adverse ecological effects and the indirect human exposure to pharmaceutical chemicals warrant a multi disciplinary evaluation to establish a basis for appropriate risk assessment and potential environmental impact.
In the United States, the Food and Drug Administration (FDA) Center for Drug Evaluation and Research requires drug compa nies to submit environmental assessments for new drug applications (U.S. FDA 2009) with predicted wastewater treatment plant (WWTP) effluent introductory concentrations of > 1 ppb based on highend projected sales and worse case, endofpipe effluent discharges. However, the U.S. Environmental Protection Agency (EPA) is responsible for testing and regulating chemicals in drinking water, but pharmaceuti cals are not currently regulated. In recognition of the growing concerns related to pharmaceu ticals, one antibiotic and several hormones used in drug formulations are now on the Drinking Water Contaminant Candidate List (U.S. EPA 2010). Clearly, more information is needed to guide decisions about whether regulations on these chemicals are needed.
In light of these concerns, the Research Triangle Environmental Health Collaborative organized the Environmental Health Summit in the Research Triangle Park of North Carolina (USA) in November 2008 to identify the major issues surrounding public concern with the presence of pharmaceutically active chemicals in water. The collaborative was established to create a forum to promote part nerships and to set research priorities and pub lic policy. This summit provided such a forum for more than 150 participants from North America and beyond, including stakeholders from the federal, state, and local govern ments, academic institutions, foundations, nonprofit organizations, drug manufacturing companies, and water utilities, as well as environmental and public health leaders. Thus, the goal of the summit was to provide a balanced consid eration of this sensitive subject by recognizing its complexity and by involving multiple dis ciplines to assess all the related issues. During the meeting, a series of questions were tabled and a framework for a research agenda was created and built around the major issues that would address the balance of protecting pub lic health through drug use while preventing unprescribed exposure from environmental input. In this article, we present an overview of the summit discussions.

Sources of pollution, fate, and transport.
Scientific data regarding the source, fate, and transport as well as the environmental and human impact of veterinary and human phar maceuticals would be beneficial to all stake holders, including policy makers, to help assure that current use and disposal patterns offer no threat to society. However, such data as well as exposure assessment are limited and uncertain, in some cases generating challenges that can be considered in two major categories: source characterization and exposure assessment.
Source characterization. Pharmaceuticals are widely consumed by humans and also used in many animal feed commodities for treat ment and prevention of disease and to enhance feed efficiency and growth. However, discrimi nation and apportionment between veteri nary animal, agricultural, and human waste, whether from domestic or industrial sources, are not always clear, particularly because data on animalwaste pharmaceuticals are quite scarce. In the animal production industry, technologies have been developed to remove various waste products, including nutrients, heavy metals, ammonia, pathogenic bacteria, odors, and other parameters that are currently regulated by legislative mandates in some U.S. states. However, current wastewater treatment technologies, including those practiced at WWTPs, onsite processes, and even emerging animal waste treatments, do not specifically target pharmaceuticals for removal, so these chemicals are likely to be present at high lev els in effluents from these processes (Cabello 2006;Dolliver and Gupta 2008;Kemper 2008). In fact, WWTPs are frequently identi fied as the main points of discharge of phar maceuticals in surface waters (Radjenovic et al. 2007), but limited studies have been carried out to track the effectiveness of unit processes at these plants. Conducting such studies has become more challenging in part because plant managers are concerned that data show ing the occurrence of pharmaceuticals in their waters could be mis represented by the media (Donn et al. 2008). The use of stable chemical indicators of pollution for both veterinary and human pharmaceuticals was suggested as a relatively easy first step for tracking the fate and transport of pollutants from these sources without the need to identify every chemical present in the mix.
The quantity and quality of existing envi ronmental occurrence data for pharmaceuti cals are insufficient for decision making and may be questionable if the data are generated from nonstandardized analytical methods with insufficient quality assurance and control. For the most part, such data do not account for human or environmental metabolites or par titioning behavior between the aqueous and sediment phases in the natural environment. Furthermore, singlegrab samples, which are the source of much of the data, do not ade quately capture variation throughout the day, and, when used to measure environmental inputs, may not represent the true concentra tions. Efforts should be made to choose more representative sampling methods that also take into account spatial variability (related to land and drug use patterns) and temporal variability and cycles (including longitudinal changes).
The ability to predict the fate and trans port of chemicals in aquatic systems depends on the availability of physicochemical data that describe, among other characteristics, dis tribution coefficients, soil-water partitioning, and biodegradation rates. As part of the drug approval process in the United States, drug manufacturers must generate much of these data. If access to such data were improved, sci entists would be in a better position to control and improve the design of water and WWTP technologies. The U.S. EPA is currently assembling a knowledgebased inventory that includes physical-chemical properties, occurrence data, analytical methods, WWTP influent and effluent measurements, and a bibliographic database for research.
Exposure assessment. Data on pharmaceu tical degradation rates will assist in the design of models to predict their fate and characteriza tion of processes that influence their transport, such as those in and between soil and ground water or downstream of a pointsource dis charge. Additional studies on the human and ecological risk of pharmaceuticals at environ mentally relevant levels, including synergistic effects with other chemicals, would be benefi cial for an accurate exposure assessment.
Pharmaceuticals represent a class of several hundred chemicals that together with their human and environmental metabolites pres ent a major challenge in terms of developing standardized, reliable sampling and analytical methods with adequate quality control. Such methods could be built around a prioritized list that takes into account production and use, their predicted "impact" based on toxic ity and activity, and their degradation and persistence in WWTP effluents; however, a critical data review process should be in place to ensure methodological credibility and to provide a sound basis for source characteriza tion and exposure assessment.
The bioconcentration of pharmaceuticals in indicator species, such as aquatic life and their predators, could provide a better measure of environmental loading and accumulation because they are closer to sources of pollutants.
Pharmaceutical agents and their human metabolites are being discharged into the aquatic environment at a loading that is related to the size and the age of people in the community and the dynamic relationship between WWTP discharge and environmental exposure. This loading is likely to be affected by three major factors: a) the production of a more concentrated effluent from WWTPs due to the increase in population density and the number of elderly people in the community, b) the reduced dilution of sewage as a conse quence of sustained drought conditions that results in a reduced flow of receiving streams and an increased proportion of treated waste water reaching downstream reservoirs, and c) an increase in the number of communities moving to the use of reclaimed water because of water scarcity and climate change.
Sources of drinking water (both surface and ground) are increasingly being affected by upstream WWTP discharges or soil-aquifer infiltration, but it remains to be demonstrated whether the technology used in processing wastewater for reclamation is selective in removing pharmaceutically active chemicals or whether they can be concentrated through sediment transport. Evidence to date suggests that the simplest and most commonly used conventional WWTPs are only partially effec tive in removing the most hydrophobic com ponents and, in so doing, concentrate a wide range of organics, including pharmaceuti cals (Chenxi et al. 2008), in biosolids that are often sold as fertilizer, which when irrigated could release the chemicals into the ground.
Exposure effects and risks to humans and ecosystems. When assessing environmental and health risks from exposure to chemical pol lutants, it is important to clearly distinguish between humans and ecosystems in terms of both exposure and effects. The effects of parts per billion or lower concentrations on ecosystems can range from changes in gene expression to changes in population structure, although little evidence exists for such adverse effects from most pharmaceuticals. However, two extraordinary examples do exist. The best known was the dramatic decrease in vulture populations in India and Pakistan (95% in 3 years), where vultures that fed on carcasses of cattle treated with diclofenac died from renal failure because they were unable to excrete the drug (Oaks et al. 2004). The other case volume 118 | number 7 | July 2010 • Environmental Health Perspectives involved the deliberate dosing of an entire experimental lake with low levels of the active ingredient in birth control pills, ethinyl estra diol. Within the first year, fathead minnows showed evidence of responses at the cellular and tissue levels and declines in the popula tion; by the second year, the fish population had collapsed completely (Kidd et al. 2007). Other studies have shown evidence of endo crine disruption effects in fish populations (e.g., intersex, histological changes in gonads, feminization of male fish preventing them from reproducing) exposed to effluents that contain chemicals such as ethinyl estradiol (Jobling et al. 1998); the predicted noeffect concentra tion for aquatic effects from this compound is < 1 ppt (Caldwell et al. 2008), but levels of this chemical have been found at up to 27.4 ppt in the river waters of Taiwan (Chen et al. 2007) and at up to 178 ppt in WWTP effluents in Australia (Fernandez et al. 2007).
In aquatic ecosystems, the fact that many of these chemicals are designed to resist bio degradation favors their adsorption onto watershed sediments, so despite the low aque ous levels reported, we do not yet have a grasp of the implications of their constant infu sion into rivers and streams. Could this, for example, be generating a hidden environmen tal risk associated with longterm exposure and combinatory effects? The case of anti biotics could be enlightening; for example, a study in North Carolina (USA) has sug gested that a defined environmental effect, the development of antibiotic resistance, may be evolving in the aquatic-sediment interface where multiple antibiotics are accumulating (Stauffenberg and Weinberg 2006).
The primary exposure pathways to humans other than those from prescribed dosing are through drinking water at partpertrillion levels (Stackelberg et al. 2007;Ye et al. 2007), which for typical daily consumption over a lifetime, would provide exposure to individual compounds well below a single therapeutic dose and suggest little threat to human health (Fent et al. 2006), although the effects on pregnant women and their fetuses are still not clear. Exposure through bathing in con taminated recreational waters, the potential for pharmaceuticals to migrate across the skin barrier, and contact with antibioticresistant bacteria are all potential secondary exposure pathways and examples of research needs for assessing the impact of pharmaceuticals in the environment (Ankley et al. 2007).
Approaches to the evaluation of toxicity associated with chronic lowdose exposure to mixtures of pharmaceuticals vary depending on whether we consider human or ecological hazard risk assessment. As an example, surveil lance of waters receiving wastewater effluents is critical for ecological hazard risk assessment, whereas biomonitoring or chemical analysis of drinking water is more important for human hazard risk assessment. However, it is criti cal to ensure that the correct end points are being considered and that they are plausible to measure in lowdose chronic exposures. There is, therefore, a need to consider non traditional end points of toxicity, such as behavioral, developmental, and reproductive responses, including toxicogenomics, relevant "omics" techniques in general, or others such as "no observable transcriptional effect levels." A need also exists to identify organisms or subsets of populations, including specific life stages (e.g., fetal development), that might be more vulnerable to pharmaceuticals. In partic ular, assessment of pharmaceuticals must con sider the mixture of compounds with which exposure occurs. The assumption that addi tive models do not adequately assess effects of pharmaceuticals has not been rigorously tested, so the applicability of existing screen ing methods must be evaluated. Alternate ways to address or evaluate mixtures can be considered, including a screeningthreshold approach where the threshold of toxicologic concern is the denominator (Daughton 2008;Kroes et al. 2005), the measured or predicted exposure concentration is the numerator, and the resulting ratio is evaluated: If < 1, no fur ther actions are warranted. Alternatively, an integrated epidemiological, biomonitoring, and demographic approach could be used. In fact, such an approach was recently applied to the assessment of a different category of drinking water contaminants, namely, mix tures of disinfection byproducts that were formed during the treatment of drinking water (Simmons et al. 2008). That study could serve as a model to frame the evaluation of mixtures of pharmaceuticals.
Risk assessment could also benefit from a prioritization listing of pharmaceuticals to help determine which pose the highest risk based on specific factors such as mode of action, therapeutic dose, and environmen tal exposure. Kostich and Lazorchak (2008) recently published a model to help implement this pharmaceutical prioritization approach, but no widely accepted prioritization list yet exists. It would also be prudent to test whether the existing human health data on pharma ceuticals can be applied to environmental hazard risk assessment even when the envi ronmental doses are much lower than thera peutic doses. Ecosystem adaptation and the changes in sensitivity of organisms (increased or decreased) with continuous exposure are other critical issues to be considered in future health and ecological risk studies.
Best management practices. Despite the various complex variables that combine to cause pollution of the aquatic environment by pharmaceuticals, actions can be taken to reduce their presence in the environment, such as responsible disposal of leftover drugs from the consumer sector. Some activities have already been sponsored by the U.S. federal gov ernment regarding drug disposal (Daughton 2003). As an example, the White House Office of National Drug Control Policy, in collabo ration with the U.S. EPA and FDA, imple mented the nation's first public guidance on consumer drug disposal (Office of National Drug Control Policy 2009), but guideline standardization is still needed.
The most effective best management prac tices (BMPs) would address source reduction, namely, reducing the amount of medicine that goes unused. The Pharmaceutical Research and Manufacturers of America have evaluated unused medicine disposal options and have concluded that toilet flushing of unused medi cine should be avoided whereas household trash disposal and takeback programs are effec tive at removing the unused medicine contri bution to pharmaceuticals in water (SMAR X T DISPOSAL 2009). Another approach for a more environmentally sound handling of unused drugs is to use "social marketing" (NewtonWard 2007). Marketing technologies would be applied to the analysis, planning, exe cution, and evaluation of programs designed to influence the voluntary behavior of target audi ences toward drug disposal.
Before implementation of BMPs can be assured and widely practiced, the following questions need to be addressed: • What are the roles and responsibilities of players within the various interconnected domains, namely, the industrial, medical, social, and environmental sectors? • Which metrics should be used to design and assess BMPs; are they currently being deployed, and how predictive are they? • What achievable BMPs can be established for contaminants already known to cause adverse effects? • What can be done immediately with what we know today, and which BMPs are prac tical and costeffective to implement? Responses to these questions will dif fer depending on the sector considered. Supplemental Material, Table 1 (available online at doi:10.1289/ehp.0901532) suggests the most feasible BMPs for shortterm imple mentation, which are presented in order of priority according to such criteria as speed of implementation, cost containment, and efficacy.
The production, transport, and fate of pharmaceuticals cross the jurisdiction of mul tiple regulatory agencies in the United States, creating a challenge to balance the societal need for safe and effective drugs against pos sible deleterious human and environmental impacts from the presence of their residues in the environment. Better interagency coop eration, data sharing, a focal point for communication, and overall better federal facilitation are essential to seriously address public concerns on these issues, which include the judicious evaluation and effective solv ing of any questions or problems that arise. During summit discussions, a consensus was reached that some organization (which, in the United States, could be the U.S. EPA or another federal agency) needs to oversee the specific issue of pharmaceuticals in drinking water to ensure justification for action or lack thereof. The U.S. Congress might designate leadership to provide guidance and back ground on existing efforts, but stakeholders have to be involved in the evolution of pro cess, and these should include at a minimum with pharmaceutical companies, water utili ties, regulatory agencies, and academia.

Education and communication.
Three key areas were suggested for education and com munication: transparency, research into com munication, and focus of education.
Transparency. The implications of the presence of pharmaceuticals in water should be communicated to the public in a truthful manner without raising alarm or suspicion and causing mistrust.
Research into communication. The tar get audience has to be identified, as well as their perception of the subject before com muniqués are distributed. Potential audiences would include the general public, legislators, farmers, regulators, health and food industry representatives, water treatment plant and WWTP operators, journalists, and scientists.
Focus of education. Those who prescribe and dispense medications, farmers, and the general public should be educated with the objective of creating interventions with the most likely and immediate impact.
Supplemental Material, Table 2 (doi:10. 1289/ehp.0901532) provides initial recom mendations for better communication and education, prioritized in terms of feasibility for shortterm implementation.
One specific goal established during the summit was to come up with a simple and transparent response to public and media con cerns about the safety of the drinking water supply related to pharmaceutical contamina tion. In response to this task, the summit par ticipants issued the following statement: "If your drinking water meets current U.S. standards, your drinking water is considered safe and drinkable. We recognize that trace amounts of pharmaceuticals in combination with other chemicals have been found in water. These sub stances are coming from a variety of sources and are difficult to completely remove. There is lim ited information on how longterm, lowdose exposures affect humans and wildlife. U.S. stan dards may need to be developed for pharma ceutical compounds in drinking water or in aquatic systems as more information becomes available." (Research Triangle Environmental Health Collaborative 2008).
Although not wishing to convey a false sense of security to the public by declaring the water supply riskfree from the effects of pharmaceuticals, it is important to stress that risk assessment is ongoing. Moreover, even if it turns out that pharmaceuticals are not a major concern, there are still other chemi cals (e.g., personal care products, pesticides, industrial contaminants, etc.) that could pose a human health or ecological risk.

Conclusions and Recommendations
Despite rising fears over the presence of phar maceuticals in drinking water, there is currently little evidence associating them with adverse human health risks, but this summit served to highlight the issues of concern that need to be addressed to confirm this perception. Supplemental Material, Table 3 (doi:10.1289/ ehp.0901532) collates the major research needs in this respect. For example, more and better data are needed to prioritize which pharmaceu tical chemicals could potentially pose the high est risk to consumers and to the environment. The identification of those drugs susceptible to formation of degradation and/or transforma tion products that have equal or more toxicity than the parent compounds is a critical point in this respect. Effectiveness of current waste water treatment, onsite processes, and emerg ing animal waste treatment technologies also needs to be evaluated.
In general, ecosystems appear to be more at risk than humans in part because aquatic organisms may have a higher sensitivity to and may be exposed to higher levels of phar maceutical residues than are humans. The key challenge in assessing the risk of pharmaceuti cals in the environment is to evaluate whether unintended longterm, lowdose exposures to mixtures of these chemicals can potentially affect humans and wildlife. In this respect, questions regarding susceptible subpopulations or lifestage sensitivity (e.g., the fetus) must be considered. Existing frameworks, such as those developed for pesticides and disinfection by products, including the use of tracers or sur rogates for whole groups of pollutants, may be appropriate for use as models assessing hazard or risk of pharmaceuticals in the environment.
BMPs have to be implemented covering the industrial, medical and veterinary, social and agricultural, and environmental spheres. Most BMPs suggested in this review focus on educa tion of the health care community and general public on the proper disposal of unused medi cine and how steps can be taken to reduce the amount of medicine that goes unused. Other BMPs listed include takeback programs, evalu ation of treatment technologies, and increased monitoring of pharmaceuticals in water. An evaluation of the effectiveness of BMPs once they have been implemented would be desir able, but there needs to be supervision or over sight for coordination of all activities related to studying the specific issue of pharmaceuticals in drinking water and their sources.
Even though there is limited information about environmental and public health effects derived from the presence of pharmaceuticals in water, there is a need to communicate what is known. The use of new technologies is rec ommended for facilitating communication and collaboration. However, the information must be clear, without providing mixed messages on the subject of potential yet currently unknown risks, remaining truthful without causing alarm. If the idea that water can be considered safe is communicated to the public, it must be emphasized that this is based on the test data of scientific studies, lack of observed adverse effects where these have been evaluated, and current regulations.
Water scarcity, climate change, aging and increasing population density, increasing use of pharmaceutical products, and rising dependence on water reuse may lead to an increase in the presence of pharmaceuticals in groundwater, surface water, and drinking water in the near future that might pose a risk to water safety or an exacerbation of perceived risk. Scientific techniques to understand and predict the potential for longterm effects of pharmaceutical residues in the environment are continuing to be developed to assess these challenges and to help prevent environmental and human health effects.
The summit provided an interdisciplin ary forum that considered a multipronged approach to a potential public health chal lenge. It demonstrated the benefit of collecting voices from those who would normally work in disconnected environments and provides a "clearinghouse" or umbrella through which congressional offices and other interested parties can seek information on the subject discussed. More such interactions among scientists, public advocacy groups, and representatives from rel evant industries, other stakeholders, and policy makers will help to clearly define the challenges faced by special issues and to provide each sec tor with a better understanding on how best to assimilate and disseminate their knowledge. In turn, this will help scientists to better advocate for public policies.