In response: Governmental perspective

The answer to the question “Do human pharmaceuticals present a risk to the environment?” is gaining knowledge and closing data gaps.

In Germany alone there are approximately 1200 human pharmaceutical substances on the market 1 (IMS MIDAS®, 2013), which are likely to occur at physiologically relevant concentrations in the environment. The majority of them lack a proper environmental risk assessment. This is especially true for those substances that received European Medicines Agency marketing authorization before 2006—that is, before environmental risk assessment became obligatory in Europe 2. Even for well-known substances such as ibuprofen, diclofenac, and carbamazepine, there are insufficient data for a complete environmental risk assessment, thus not fulfilling the legal requirements according to the European Medicines Agency guideline.

It is necessary to develop a concept that allows a proper judgment of the environmental impact of these substances, individually and in mixtures with other substances, including pesticides, biocides, and industrial chemicals. So far, this concept is far from being available. Such a concept needs to include all available information—in vitro, in silico, and the translation of existing nonclinical toxicological data. The fundamental requirements from this knowledge base are knowledge regarding the translation of in vitro data into quantitative population information, knowledge regarding the calibration of models for a more accurate read-across, and knowledge of how the effects data and the no-effects data from toxicological studies can be used for environmental risk assessment

To date, however, this knowledge has not been available—not even for the most understood pathway, estrogen receptor activation. A search in literature databanks with the search terms “estradiol” and “fish reproduction” reveals more than 8000 publications (ScienceDirect, June 2014). However, not even for this pathway is a quantitative risk assessment possible. Molecular and biochemical end points usually do not provide sufficient information for a quantitative risk assessment for any specific chemical. Even the extensive knowledge of how a potent estrogen receptor agonist such as 17α-ethynylestradiol (EE2) behaves informs only how one would approach the evaluation of other potential estrogenic compounds in a focused, cost-effective manner 3.

An environmental risk assessment for a human pharmaceutical needs to be conducted prior to market application but at an advanced stage of development of a new substance, meaning that quite a lot of therapeutic information is already available. Unfortunately, this information from preclinical studies often is not usable and not transferrable to an environmental risk assessment. This is mainly because toxicological studies need to answer different questions, and the protection goal is different. For environmental studies, the protection goal is the population, rather than the individual in preclinical toxicological studies. This makes the translation of information difficult in the first instance. Often, it is not possible to extrapolate the assessed end points of toxicological studies to population effects. Table 1 compares the different end points and differences in designs in toxicological and ecotoxicological studies. The same applies to the read-across process; although many drug targets and enzyme molecular structures are conserved in humans (mammals) and other vertebrates (e.g. fish) and even in invertebrates 4, the mechanism of action in wildlife for most pharmaceuticals is not yet completely understood 5.

For most pharmaceuticals, insufficient laboratory studies and/or environmental monitoring data are available to answer the question of whether environmental concentrations of pharmaceuticals are anywhere close to those that produce adverse effects in ecotoxicity tests. For a few pharmaceutical substances, however, the answer is “yes.” An assessment according to the European Medicines Agency guideline identifies a potential environmental risk for some hormones, some neurological drugs, and some cytostatics. Furthermore, as global consumption of pharmaceuticals rises, an inevitable consequence is an increased level of contamination of surface waters and groundwaters with these biologically active drugs and, thus, a greater potential for adverse effects in aquatic wildlife 6.

How can we prioritize and identify pharmaceuticals of the greatest environmental risk? This ranking of pharmaceutical active substances has to include effect and exposure. It is necessary to assess the potential effect concentrations but also the amount of a specific substance actually entering the environment. Two extreme examples include ibuprofen, with a steadily increasing consumption in Germany of now nearly 1 000 000 kg/yr, and EE2, consumption of which is slowly but continuously declining to 47 kg/yr as of 2012 1. The effect concentrations of ibuprofen for the aquatic compartment are in the range of micrograms per liter 7, whereas the highly specific receptor-mediated effects of EE2 are in the low nanograms-per-liter range or even less 8, 9. In addition, fate and degradation also have to be accounted for. Because EE2 is relatively stable, it is found in surface waters in the range of the effect concentrations. On the other hand, ibuprofen is easily degradable under aerobic conditions; but because of the high consumption rate, it can be found in surface waters in the range of the effect concentrations as well 10. In addition, metabolites and transformation products should not be ignored because for several products, such as carbamazepine, different environmentally relevant transformation products exist.

To assess an environmental concentration of a specific substance, normally the predicted environmental concentration (PEC) of an individual product will be estimated. This results in shortcomings in both directions. It either overestimates the environmental concentration, as potential degradation is not included, or underestimates the environmental impact, as the PEC is calculated for individual products only, neglecting the fact that nearly all active ingredients are used by more than 1 pharmaceutical product. This is especially true for the high-consumption products such as the above-mentioned ibuprofen, diclofenac, and carbamazepine. Therefore, for highly consumed products, measured environmental concentrations would be the alternative; however, these are lacking for most pharmaceuticals. Although several monitoring programs measure selected pharmaceuticals on the local, national, and international levels, general coordination is lacking. As a consequence, there is no single pharmaceutical measured all over Europe. The European Community recognized this and is now considering, for the first time, 3 pharmaceuticals—diclofenac, EE2, and 17β-estradiol—as candidates for future control via environmental quality standards. Johnson et al. 11 estimated the concentration of these 3 substances in European rivers using a geographically based water model and found that the concentration of EE2 would exceed the environmental quality standard in 12% of European rivers, whereas concentrations of 17β-estradiol and diclofenac would exceed it in 1% and 2% of rivers, respectively.

After identifying the priority substances for testing, the next step has to be testing to a level and quality acceptable for an environmental risk assessment. For substances receiving marketing authorization before 2006, the Umwelt Bundesamt has recommended for several years a monographic system, through which the results of environmentally relevant data are made publically available in a substance-specific monograph. In this way, the industry, both research-based and generic companies, can share the financial burden. In very rare cases, companies have already formed a consortium to gain the data for a shared environmental risk assessment. This testing can be done even without waiting for the final prioritization list. There are several substances and transformation products other than diclofenac, ibuprofen, and carbamazepine that rank highly in prioritization schemes 12. Testing of these compounds should be started immediately, because the environmental relevance of these substances is more than obvious.

In conclusion, the marketing authorization of pharmaceuticals alone, especially for substances authorized before 2006, is not sufficient to evaluate the long-term potential risk for the environment. The basis for environmental risk assessment is reliable data; therefore, additional postmarketing, environmental monitoring is a suitable option for gaining these data.