Pharmaceuticals and personal care products: Research needs for the next decade


  • Published on the Web 9/2/2009.


Pharmaceuticals and personal care products (PPCPs) include numerous classes of chemicals with unique physiochemical properties and biological activities. Over the past decade research on the occurrence, fate, effects, risk assessment, and management of PPCPs in the environment has peaked. It is important to appreciate the utility of traditional approaches for examining contaminant hazard and risk while understanding relevant limitations and important research needs to advance environmental risk assessment (ERA) and management efforts for PPCPs. Spurred initially by the critical reviews of Halling-Sorensen et al. [1] (755 citations as of 6 July 2009) and Daughton and Ternes [2] (778 citations as of 6 July 2009), this special issue of Environmental Toxicology and Chemistry includes a timely collection of manuscripts examining the environmental chemistry, toxicology, risk assessment, and management of PPCPs.

Occurrence, fate, and exposure

A vast number of PPCPs have now been detected in surface waters across the world. For human PPCPs, effluent-dominated ecosystems appear to represent worst-case scenarios for waterborne exposure and potential adverse effects [3]. For veterinary medicines, inputs from manure application to soils and the use of aquaculture are probably the most important exposure routes [4]. The nature of exposure to human PPCPs and veterinary medicines are also very different. Human PPCPs are typically released continuously from wastewater treatment plants and are therefore considered to be pseudo-persistent [5] with increased effective exposure duration in effluent-dominated systems [6], whereas veterinary drugs are either transported to surface waters in pulses during rainfall (livestock treatments) or are applied directly to aquatic systems in a single dose. The risk assessments of human use substances and veterinary medicines should therefore be addressed very differently. While most monitoring studies have focused on wastewater treatment systems and surface waters, researchers are beginning to recognize that other environmental compartments are also important. Hence, recent studies have explored the occurrence of PPCPs in biosolids, soils, sediments, biota, drinking water, and even food crops [7]. For example, in this issue Ramirez et al. present findings from a National Pilot Study of PPCPs in Fish Tissue from the United States, which expanded previous observations of fish bioaccumulation in an urban effluent-dominated river [8].

In addition to environmental monitoring studies, a number of studies have investigated the fate and transport of PPCPs. Most transport studies have been on terrestrial systems and have investigated the movement of PPCPs from soils to surface waters in runoff and drain flow [9] and to groundwaters via leaching. The focus has been on veterinary medicines although researchers are now recognizing that these transport processes are also important for PPCPs applied to land associated with biosolids or during irrigation with wastewater [9]. It is clear from these studies that the application matrix (e.g., manure, biosolids) has a significant impact on the transport behavior of many compounds although the underlying reasons for this have yet to be established.

For pharmaceuticals in particular, traditional approaches to characterize partitioning to soils and sediments (i.e., relating sorption to a compound's hydrophobicity and to soil or sediment organic carbon content) are often inadequate. Because many pharmaceuticals are weak acids, weak bases, or zwitterions, sorption processes are due not only to hydrophobic interactions but also are driven by other binding processes such as cation exchange, cation bridging, surface complexation, and hydrogen bonding. In fact, site-specific water, soil and sediment conditions such as pH and cation exchange capacity can have a major influence on PPCP sorption behavior in terrestrial and aquatic environments [10,11] as can the presence of biosolids and manure [12]. A few studies have attempted to model the interaction of pharmaceuticals with different environmental matrices. For example, Aristilde and Sposito [13] used molecular dynamic simulations to model metal complexation interactions of fluoroquinolone antibiotics. Ter Laak et al. [14] developed relationships between soil properties and the sorption behavior of antibiotics. However, far more modeling work of this type will be required before we have the predictive tools needed for risk assessment.

The degradation behavior of PPCPs can be complex, but recent studies are beginning to characterize the degradation rates and pathways of PPCPs in the environment. Persistence in solid matrices such as biosolids, manures, soils, and sediments indicate that for many PPCPs the observed dissipation is not due to degradation but rather to the formation of nonextractable residues [15]. Questions are being raised over the implications of these bound residues. The importance of PPCP metabolites and transformation products is also being increasingly recognized. In this issue Monteiro and Boxall, Williams et al., and Buth et al. specifically examine relevant questions of importance to defining environmental fate and transport processes of PPCPs in aquatic and terrestrial ecosystems.

Effects and risk

In developed countries most human PPCPs lack acute toxicity or fail to elicit mortality responses under typical environmental postconsumer exposure conditions [6]. Exceptions may be observed, however, in developing countries when environmental management and regulatory practices are not as robust. Carlsson et al. report in this issue ambient aquatic toxicity resulting from exposure to effluents from bulk manufacturing facilities in India. Exceptions are also apparent for select veterinary medicines with intended use as insecticides; for example, ivermectin, a parasiticide, is acutely toxic to crustaceans at low or even part per trillion levels [16]. Chronic exposures to PPCPs occur particularly for organisms residing in effluent-dominated systems [3], and potentially adverse ecological responses to select substances may occur at environmentally relevant levels at much lower concentrations than mortality thresholds [6].

Thus, it is critical to consider a priori the mechanisms of action (MOA) for PPCPs in the environment. Leveraging pharmacological safety information, for example, appears useful for supporting such efforts, but for a response to be meaningful as a measure of effect in ecological risk assessment, MOA-related responses should be linked to the population level of biological organization, with endangered or threatened species being an important exception. Ankley et al. [6] presented a framework for performing such examinations. In this issue, Zellinger et al. describe how synthetic estrogens adversely affect fish reproduction at low or even sub-part per trillion concentrations, which could be anticipated based on the MOA of these compounds. Although the approach of Ankley et al. [6] appears promising, it is important to note that a therapeutic MOA may be more predictive for nontarget vertebrates than invertebrates and algae. Brooks et al. [17] highlighted this observation because lower adverse effect thresholds were determined for algae growth following exposure to the antidepressant fluoxetine than endpoints employed in standardized crustacean and fish testing methods. Such an observation was interesting because green algae do not possess the therapeutic target of fluoxetine, though this therapeutic is known to have antimicrobial properties [17].


Occurrence, fate, and exposure

Although most PPCP literature has focused on environmental occurrence, several key areas require additional investigation. Municipalities and other water resource providers increasingly request guidance for monitoring target analytes in potable source waters, fisheries, and effluent discharges. We submit that target analytes should be prioritized for environmental monitoring based on either of two criteria: Does a target analyte such as carbamazepine provide an indicator for other contaminants; or does the compound represent a chemical class either known or suspected to present hazard/risk to aquatic organisms? Rather than selecting target analytes simply because analytical standards (including isotopically labeled compounds) are available, thus potentially expending limited resources, selection of targeted analytes for environmental monitoring should be informed initially by hazard-based, and ultimately by risk-based approaches. This problem is further addressed in the Effects and risk section below. These prioritization approaches should be applied to parent PPCPs and to metabolites formed in animals and humans and transformation products formed in wastewater treatment systems, surface waters, soils, and drinking water treatment plants, because in some cases these compounds may pose a greater risk than the parent compound [18].

Despite the recent work of Carlsson et al. (this issue) much less is known about the contribution of drugs in effluent from production facilities located in the developed world. Similarly, the contribution of unused medications discarded via sewerage to aquatic and terrestrial exposure is less understood, particularly that from health care facilities. Selected input routes of veterinary drugs (runoff from the farmyard) also have been largely ignored. Additional studies are needed to understand climatological influences on temporal and spatial PPCP exposures, particularly for aquatic and terrestrial systems receiving municipal biosolid and veterinary medicine runoff, and in effluent-dominated or dependent water bodies [19], in which PPCP exposures may be maximal under low-flow conditions [3].

Studies are also needed to explain how partitioning among solids (suspended, dissolved, colloidal, sediments, soils), including irreversible binding and hysteresis, and ultimately bioavailability, is influenced by site-specific factors and physiochemical properties of various PPCPs. These studies should aim to determine the impacts of manure and biosolids on environmental behavior. Ultimately, we should aim to develop relationships between PPCP chemical properties, environmental properties and fate characteristics to predict the mobility and persistence of a PPCP in a particular environment.

Ramirez et al. (this issue), describe accumulation of PPCPs in fish and other aquatic organisms residing in effluent-dominated or influenced water bodies, in which the magnitude of exposure to PPCPs detected in edible fish filets is well below human therapeutic levels [8]. The potential uptake of PPCPs into crops is increasingly being recognized, and calculations by Boxall et al. [20] indicate that foodborne exposure may be much more significant than drinking water. Human health risk assessment efforts are needed to fully characterize risks associated with such exposures, particularly over longer-term scenarios than typical pharmacotherapy windows and for sensitive subpopulations. It is important that the Society of Environmental Toxicology and Chemistry Pharmaceuticals Advisory Group has targeted the need for research specifically focusing on refining human health risk assessment of pharmaceuticals in the environment. Our understanding of why a PPCP accumulates in a nontarget organism is also poorly developed, so we should begin to build on initial monitoring studies of fish and crops to understand the underlying processes affecting uptake by organisms. However, future studies should be able to better leverage pre-existing target organism knowledge (e.g., in mammals) on the absorption, metabolism, and elimination of specific pharmaceuticals.

Exposure models are now being used routinely in the risk assessment of PPCPs. Some work has been done to evaluate these models [10,11], which seem to work well for many PPCPs; in other cases, however, they greatly underpredict exposure. We especially need to understand the discrepancies between model outputs and experimental observations and work to improve the models, particularly at a watershed scale, to better characterize risk.

Effects and risk

Because many PPCPs are not acutely toxic to organisms at environmentally realistic concentrations in developed countries, Ankley et al. [6] proposed approaches to determine whether response variables may be employed as measures of effect in ERA and to incorporate PPCP MOAs a priori to guide tier-based examination of the ERA process. Precedence for such approaches includes the evaluation of endocrine-disrupting and -modulating compounds, which include select PPCPs [6]. Evolutionary conservation of therapeutic targets, which varies among nontarget organisms, presents opportunities to consider the potential response of an organism to a PPCP based on known MOA in target species [21]. If a toxicological or pharmacological target is present in a nontarget organism, if the functional interaction with the target is understood, and if target interaction results in physiological responses linked to the population level of biological organization, then potential hazards to nontarget organisms must be considered [6].

Although regulatory programs use standardized approaches to examine characteristics such as biotransformation and ecotoxicity, it is necessary to develop rapid in vitro and in vivo screening methods for prioritizing compounds for higher tier testing. An initial step towards this approach would include developing fate and effects data for representative chemicals with common MOAs or chemicals from PPCP classes. Such ecotoxicological approaches could support selection of target analytes in environmental monitoring and risk management of PPCPs (Supporting Information Figure S1; If higher tier testing is necessary, species sensitivity distributions may be developed for responses linked to a specific chemical MOA. For example, Caldwell et al. [23] employed a species sensitivity distribution of reproduction thresholds to identify a HC5 value of 0.352 ng/L for 17α-ethinylestradiol. Because natural ecosystem experiments, such as a recent groundbreaking study with a whole lake exposed to one concentration of 17α-ethinylestradiol [24], cannot establish lowest observed adverse effect levels and no observed adverse effect levels, inclusion of experimental field testing for PPCPs may also be considered at higher tiers. However, it is critical to incorporate organisms and endpoints, guided by lower tier measures of effect that are linked to a PPCP MOA, for development of appropriate weight-of-evidence approaches in ERAs. For example, Fulton et al. (this issue) performed an assessment of triclosan in experimental streams, which included riffle, run, and pool habitats.

Huggett et al. [21] proposed an initial approach for examining pharmaceuticals for potential aquatic effects, which included comparing predicted plasma concentrations in fish to human therapeutic concentrations of a drug. In human and veterinary target organism assessments, plasma concentration represents the sum of many processes and thus is the measure by which internal exposure is evaluated. More information on tissue-based accumulation, metabolism, and elimination would be helpful in identifying potential pharmaceuticals for prioritized study. Tissue-based residues could be used to develop a critical tissue residue approach for supporting environmental risk assessment and management efforts and would build on approaches based on plasma concentrations [22].

Mixture interactions of PPCPs should not be ignored. In this issue Painer et al. show that environmentally relevant concentrations of antidepressant mixtures affect predator avoidance behavior of juvenile fish. Hence, future studies should not just consider toxicological interactions but should also examine environmental fate-ecotoxicity interactions. For example, Monteiro and Boxall (this issue) report data indicating that the presence of antibiotics in soils at environmentally relevant concentrations significantly decreased the degradation of naproxen.

A recently completed project in the European Union (ERAPharm) examined a model veterinary medicine (ivermectin) and a model human pharmaceutical (fluoxetine). This multidisciplinary study included expertise and numerous representatives from academia, government, and industry. Similar studies have not been performed in other geographic regions, but appear necessary to advance understanding of prospective and retrospective ecological risks posed by PPCPs in the environment. For example, the threshold of toxicological concern concept has been applied to predict toxicity thresholds of estrogen agonists [25] and may support such prioritization efforts. Further engagement among environmental toxicologists, chemists, engineers, risk assessors, managers and communicators is required to support environmental management decisions for PPCPs based on sound scientific, risk-based information.


Fig. S1. A conceptual model for identifying pharmaceuticals and personal care products for environmental monitoring and risk management.

Found at DOI: 10.1897/09-325.S1 (57 KB PDF).