Medicines have an important role in the treatment and prevention of disease in both humans and animals. But it is because of the very nature of medicines that they may also have unintended effects on animals and microorganisms in the environment. Although the side effects on human and animal health are usually investigated in thorough safety and toxicology studies, the potential environmental impacts of the manufacture and use of medicines are less well understood and have only recently become a topic of research interest. Some of the effects of various compounds—most notably anthelmintics from veterinary medicine and antibacterial therapeutics—are already known (Daughton & Ternes, 1999; Boxall et al, 2003a, 2004a; Floate et al, 2005), but there are many other substances that can affect organisms in the environment. This is further complicated by the fact that some pharmaceuticals can cast effects on bacteria and animals well below the concentrations that are usually used in safety and efficacy tests. In addition, breakdown products and the combination of different biologically active compounds may have unanticipated effects on the environment. Although it may be safe to assume that these substances do not substantially harm humans, we have only recently begun to research whether and how they affect a wide range of organisms in the environment and what this means for environmental health.
…we have only begun to research whether and how they affect a wide range of organisms in the environment and what this means for environmental health
The scope of this potential problem is not to be underestimated. More than 10 million women in the USA alone use oral contraceptives, which eventually find their way into the environment. A wide range of human medicines, including antibiotics, statins or cytotoxins used in cancer treatment, are produced and used, some in the range of thousands of tons per year. It is hard to obtain information on the amount of human medicines used, but recent data from Canada indicates that high‐use drugs include acetominophen, acetylsalicylic acid, ibuprofen, naproxen and carbamazepine (Metcalfe et al, 2004). Large amounts of veterinary medicines, such as antibacterials, antifungals and parasiticides from aquaculture and agriculture, may also contribute to the stress on the environment, particularly as they often find their way directly into soils and surface waters unlike human medicines, which usually go through a water treatment plant first. The use of antibacterials in aquaculture in the USA alone is estimated to be between 92,500 and 196,400 kg per year (Benbrook, 2002), while estimates for the total use of antibacterials in US agriculture range between 8.5 and 11.2 million kg annually (Nawaz et al, 2001; Mellon et al, 2001).
…recent monitoring studies have detected low levels of a wide range of pharmaceuticals … in soils, surface waters and groundwaters
These human and veterinary therapeutics are released to the environment by various routes (Fig 1). Residues released during the manufacturing process may ultimately enter surface waters. After administration, human medicines are absorbed, metabolized and then excreted to the sewer system. They usually go through a treatment works before they find their way into receiving waters or land by the application of sewage sludge. Antibacterials for the treatment of fish or shrimp in aquaculture are directly released to surface waters. Veterinary medicines used to treat pasture animals are excreted to soils or surface waters. In intensive livestock treatments, these medicines are likely to enter the environment indirectly through the application of slurry and manure as fertilizers. Other minor routes of entry include emissions to air and through the disposal of unused medicines and containers.
Although pharmaceuticals have been released into the environment for decades, researchers have only recently begun to quantify their levels in the environment. Using information from different countries and on various usage patterns, several prioritization exercises have identified those pharmaceuticals that are most likely to be released into the environment (Box 1). For example, data from the UK on annual usage of veterinary drugs was combined with information on administration routes, metabolism and ecotoxicity to identify medicines that should be monitored in a national reconnaissance programme (Boxall et al, 2003b). Hilton et al (2003) performed a similar exercise for human medicines using information on annual usage and therapeutic dose along with predictive models. Although these studies are generally based on country‐specific information, they still provide an indication of those substances that should be investigated at the international level. New analytical techniques, such as liquid chromatography coupled with tandem mass spectrometry (LC‐MS‐MS), have allowed us to develop a better understanding of how medicines behave in the environment and to determine concentrations in wastewater treatment plants, soils, surface waters and groundwaters.
Pharmaceuticals that have been identified as a priority for further study
Aminophylline, Beclametasone, theophylline, Paracetamol, Norethisterone, codeine, furosemide, Atenolol, Bendroflumethiazide, chlorphenamine, lofepramine, Dextropropoxyphene, Procyclidene, Tramadol, Clotrimazole, Thiridazine, Mebeverine, Terbinafine, tamoxifen, Trimethoprim, Sulfamethoxazole, Fenofibrate, diclofenac (Hilton et al, 2003)
Amitraz, Amoxicillin, Amprolium, Baquiloprim, Cephalexin, Chlortetracycline, Clavulanic acid, Clindamycin, clopidol, Cypermethrin, Cyromazine, Decoquinate, Deltamethrin, Diazinon, Diclazuril, Dihydrostreptomycin, Dimethicone, Emamectin benzoate, Enrofloxacin, Fenbendazole, Flavomycin, Flavophospholipol, florfenicol, Flumethrin, Ivermectin, Lasalocid Na, levamisole, Lido/lignocaine, lincomycin, Maduramicin, moensin, Morantel, Neomycin, Nicarbazin, Nitroxynil, Oxolinic acid, Oxytetracycline, Phosmet, Piperonyl butoxide, Poloxalene, Procaine benzylpencillin, Procaine penicillin, Robenidine HCL, Salinomycin Na, Sarafloxicin, Sulphadiazine, Tetracycline, Tiamulin, tilmicosin, Toltrazuril, Triclabendazole, Trimethoprim, Tylosin (Boxall et al, 2003b)
Once released into the environment, pharmaceuticals will be transported and distributed to air, water, soil or sediment. A range of factors, such as the physico‐chemical properties of the compound and the characteristics of the receiving environment, will affect their distribution. The degree to which a pharmaceutical is transported between the different environmental media primarily depends on the sorption behaviour of the substance in soils, sediment‐water systems and treatment plants, which varies widely across pharmaceuticals. Reported sorption coefficients for several veterinary medicines in soils range from less than 1 litre per kilogram to more than 6,000 litres per kilogram (Boxall et al, 2004b). Additionally, there is a large variability in the sorption behaviour within an individual matrix—for example, partition coefficients for enrofloxacin in a range of soil types vary by up to a factor of 30 (Boxall et al, 2004b). Moreover, unlike other organic substances, such as pesticides and industrial chemicals, the sorption behaviour of many pharmaceuticals cannot be simply derived from the substance's hydrophobicity or the organic carbon content of the solid material (Tolls, 2001).
Pharmaceutical substances may also be degraded by biological organisms in treatment systems, water bodies and soils as well as abiotic reactions. Generally, these processes reduce the potency of medicines; however, some breakdown products have similar toxicity to their parent compounds (Halling‐Sürensen et al, 2002). Furthermore, degradation varies significantly depending on chemistry, biology and climatic conditions. For example, the half‐life of the antiparasitic ivermectin under winter conditions is six times greater than in the summer and the compound degrades faster in sandy soils than in sandy loam soils (Halley et al, 1993). The natural estrogens 17β‐estradiol and estrone degrade in the aerobic and anoxic tanks of activated sludge systems, whereas 17α‐ethinylestradiol only degrades under aerobic conditions (Ternes et al, 2004). All this adds to the complexity of the problem and calls for individual solutions for individual pharmaceuticals and applications.
Not surprisingly, recent monitoring studies have detected low levels of a wide range of pharmaceuticals, including hormones, steroids, antibiotics and parasiticides, in soils, surface waters and groundwaters (Hirsch et al, 1999; Kolpin et al, 2002; Table 1). The reported concentrations are generally low—usually less then 1 μg l−1 in surface waters—but what is more worrisome is that many therapeutic substances have been found across a wide variety of hydrological, climatic and land‐use settings, and many of the substances have been detected throughout the year. These findings have raised questions about how this mixture of veterinary and human medicines abundant in soils and surface waters has an impact on beneficial organisms in the environment and on human health.
Comparison of these data with therapeutic dose information, drinking water limits and health advisories indicates that the concentrations of therapeutic compounds in surface waters are well below levels that would be of concern to human health (Webb, 2001; Kolpin et al, 2002). It therefore seems that indirect exposure to pharmaceuticals through the water supply is unlikely to pose a risk to humans. However, risks through other routes of exposure, such as uptake from soils into crops and biomagnification through the food chain have yet to be quantified and cannot be ruled out completely.
The impacts on environmental health are more difficult to assess. Since 1980, the US Food and Drug Administration (FDA) requires environmental risk assessments of human and veterinary medicines on the effects on aquatic and terrestrial organisms before they allow a product to the market (Breton & Boxall, 2003), and the EU introduced similar requirements in 1997. These environmental impact studies investigate the potential negative effects on fish, daphnids, algae, bacteria, earthworms, plants and dung invertebrates. Much of the data are publicly accessible—many of the environmental assessments are published on the FDA's web site—and provide a reasonable body of data for further study (Boxall et al, 2004a). However, there are valid questions about the real‐world value of these studies. Risk assessments usually use standard ecotoxicity tests, which are often short‐lived and focus predominantly on mortality as the endpoint. Moreover, aquatic tests tend to focus on the water compartment and do not take into account pharmaceuticals residing in sediments. In general, the effects observed in these studies occur at much higher concentrations than those that are measured in the environment. What is less known are the more subtle effects that therapeutically active substances can have on organisms in the environment, such as growth, fertility or behaviour.
It is clear that current standard ecotoxicity tests are probably inappropriate for assessing the impacts of many pharmaceuticals
Pharmaceutical compounds are designed either to be highly active and interact with receptors in humans and animals or to be toxic for many infectious organisms, including bacteria, fungi and parasites. But this does not mean that they affect only these living forms. Many lower animals have receptor systems similar to humans and animals used in agriculture. Furthermore, many groups of organisms that affect human and animal health, which are targeted by pharmaceuticals, have a crucial role in the functioning of ecosystems. It is therefore possible that pharmaceuticals may cause subtle effects on aquatic and terrestrial organisms that are not detected in standard studies. And as human medicines are almost continuously released to the environment, wildlife organisms are exposed for much longer durations than those used in standard tests. Researchers have therefore begun to look into some of the more subtle effects caused by long‐term, low‐level exposure to pharmaceuticals. A wide range of subtle impacts has been reported so far (Table 2), including effects on oocytes and testicular maturation, impacts on insect physiology and behaviour, effects on dung decomposition, inhibition or stimulation of growth in aquatic plant and algae species, and the development of antibacterial resistance in soil microbes. Steroids from contraceptives are strongly suspected to affect the fertility and development of fish, reptiles and aquatic invertebrates. Equally, antibiotics from human and veterinary use have an effect on soil microbes and algae. Macrocyclic lactones can affect invertebrate larvae in dung at fairly low concentrations; earthworms appear sensitive to the parasiticides used in veterinary medicine and plants may be sensitive to many antibiotics. In addition, macrocyclic lactones have also been shown to elicit many sub‐lethal responses in dung invertebrates, such as reduced feeding, disruption of water balances, reduction of growth rate, inhibition of pupation and the disruption of mating. As dung from livestock contains diverse fauna and provides a fruitful foraging habitat for other species, macrocyclic lactones may therefore indirectly affect other species by depleting the quality and quantity of their food source. The impacts of sediment‐associated pharmaceuticals have also been considered. Using life‐cycle studies, Nentwig et al (2004) showed that carbamazepine affects the emergence of chironomids, an aquatic midge. While many of these observations have been seen at environmentally realistic concentrations, the significance in terms of environmental health has yet to be established—in fact, this will be one of the challenges in the coming years.
Furthermore, pharmaceutical substances are not the only contaminants in environmental systems. Aquatic and terrestrial organisms are exposed to a mixture of medicines and other substances, including pesticides, biocides and general industrial chemicals. A recent US monitoring study (Kolpin et al, 2002) detected the antibacterial agent lincomycin in combination with up to 27 additional chemicals. The study looked only for selected compounds, so many other synthetic substances may also have been present. Interactive effects, such as additivity of substances with similar modes of action and synergism, are therefore possible. As current environmental risk assessments focus on single substances, it is possible that these assessments are underestimating the impacts.
When we begin to consider these interactions, it is important that we do not just focus on toxicological endpoints. It is also possible that the environmental behaviour of a substance could change in the presence of other substances. Antibacterials, for example, have been shown to affect soil microbes, which have an important role in breaking down pesticides. For example, studies indicate that veterinary antibacterials may affect sulphate reduction in soil and inhibit the decomposition of dung (Westergaard et al, 2001). If a veterinary antibacterial were to be applied in slurry to an agricultural field before the application of a pesticide, it is quite possible that the environmental impact of the pesticide could be radically changed.
As very little is known about the impacts of pharmaceuticals on ecological health and the interactions of different compounds, some workers are taking a precautionary approach and are developing methods to reduce the releases of these substances to the environment. Various approaches have been advocated, including the control of pharmaceuticals at the source, the segregation of sources, the treatment of waste products to remove pharmaceutical compounds, the introduction of husbandry practices and the improvement of disposal systems for out‐of‐date medicines and waste containers (Table 3; Ternes et al, 2002, 2004; Daughton, 2003a,b). Source controls include labelling, controlled disposal and urine separation. Segregating sources of pharmaceuticals, such as hospital wastewater, which is likely to be heavily contaminated with pharmaceuticals and antibiotic‐resistance bacteria, should make it possible to focus treatment resources on the most contaminated waters.
Pharmaceuticals can be removed when treated through physical processes, such as sorption or volatilization, biological degradation or chemical reactions, for instance, through treatment with ozone. The suitability of different options are likely to be highly specific for each substance. For example, the antibiotic ciprofloxacin is removed by strong sorption onto suspended solids of sewage sludge whereas diclofenac and 17α‐ethinylestradiol undergo significant biodegradation in aged activated sludge. It is therefore likely that a range of measures will be required to reduce emissions. Many of the treatment methods, whilst removing the pharmaceuticals, may also produce transformation products that are more persistent and mobile than the parent compounds, some of which may also have similar or enhanced toxicity. Little work has been performed to assess the environmental impacts of these transformation products on the environment (Boxall et al, 2004c).
It is clear that during the past few years a wealth of data has become available on the levels of pharmaceuticals in the environment and on their effects on aquatic and terrestrial organisms. There are, however, still many questions that need to be addressed before we can eventually determine whether residues in the environment are a threat to human and environmental health. First, what are the risks of substances that have yet to be studied? Due to resource limitations, only a small proportion of pharmaceuticals in use today have so far been investigated, and there is a great need to understand how other substances affect the environment. Second, how can we better assess ecotoxicity? Current standard ecotoxicity tests are probably inappropriate for assessing the impacts of many pharmaceuticals. The use of more subtle endpoints, such as changed behaviour, physiology and biochemistry, seem to show some merit. Further work should be performed to identify these subtle effects. It is likely that many of the technologies now used by molecular biologists—for instance, proteomics and genomics techniques or large‐scale DNA or protein arrays—could greatly help with this task.
Third, what do these ecotoxicity data mean in the real world? Although many subtle effects have been shown after exposure to pharmaceuticals at environmentally realistic concentrations, we need to establish what these data mean in terms of ecological functioning. Fourth, what are the risks of mixtures? Pharmaceuticals are unlikely to appear in the environment on their own so the current ‘single‐substance’ approach to risk assessment could be underestimating environmental impacts. This also includes possible indirect effects. Little work has been done to determine the uptake of pharmaceuticals into organisms and through the food chain. Such studies are crucial to determine the potential indirect effects of environmental exposure on ecological and human health. A related question is whether we should worry about transformation products. Most work so far has focused on the parent compounds; however, we know that transformation products are produced in the environment and in treatment processes. It is important that we begin to understand the potential impacts of these substances.
Future work must therefore focus on understanding the biotic and abiotic processes underlying the release, environmental fate and effects of pharmaceuticals
Finally, does environmental exposure result in more antibacterial resistance? A wide range of antibacterials has been observed in waters and soils and many of these persist for some time. It is possible that such exposure will result in the formation of resistant microbes, which could pose a serious threat to human and animal health.
It will be impossible to design and carry out studies to answer each of these questions for every single substance that is in use today. Future work must therefore focus on understanding the biotic and abiotic processes underlying the release, environmental fate and effects of pharmaceuticals. Such an understanding should ultimately allow the development of new modelling approaches. For instance, Huggett et al (2004) have proposed a comparative plasma concentration model that bridges mammalian and fish species, which could provide useful information on the probable impacts of pharmaceuticals on fish. Other modelling approaches, such as quantitative structure‐activity relationships, could allow us to estimate the environmental impacts of pharmaceuticals from their chemical structure. Read‐across approaches, where data from closely related compounds are used to identify the impacts of an untested compound, may also help to improve environmental assessment studies. These improved tools should allow us to understand better the impacts of pharmaceuticals on the environment. In the meantime, we should strive to refine the ways in which we use, handle and treat medicines to minimize their releases to the environment.
The author is grateful to the European Commission, the UK Environment Agency, DEFRA and English Nature, who have funded the work that has made this paper possible.
- Copyright © 2004 European Molecular Biology Organization