Introduction

Water pollution today represents a major challenge both at the economic and social level. The quality of water must be preserved both for human protection and to safeguard the environment from compounds capable of exerting an effect at low levels of concentration. Among others, pharmaceuticals are of concern because they are consumed in high quantities; many are refractory to traditional wastewater treatment and thus become widely distributed in river waters (Banjac et al. 2015; Zhang et al. 2013). Their presence in the environment has been attributed to the discharge of hospital effluents (Gómez-Canela et al. 2014, Langford and Thomas 2009; Santos et al. 2013; Verlicchi et al. 2010), domestic water (Rabiet et al. 2006), and effluents from wastewater treatment plants (WWTP) (Gómez-Canela et al. 2012; Santos et al. 2013; Verlicchi et al. 2012). This study explores senior residences and sociosanitary centers as a source of pharmaceuticals to environmental waters.

In Europe and over the world, the phenomenon of aging and over-aging has led to societies where 15–20% of the population is over 65 years. Part of this population reside in homes for the elderly or in sociosanitary centers if the level of impairment is high. The homes for elderly or health care institutions are infrastructures that emerge in response to biopsychosocial needs and have become popular in most European countries. On the other hand, sociosanitary centers provide integral attention to people that require special care because of the low health status, either physic, psychic, or intellectual. These establishments have a configuration of typically 50–150 beds and provide lodging, meal services, and health and social assistance. With an estimated consumption of 5–10 pills/patient, much higher than the healthy population, the total consumption of pharmaceuticals is of hundreds of milligrams. These compounds are excreted through urine or feces and are released to the main urban sewerage system without any type of treatment. Then, waters are transported to the WWTP and a fraction is eliminated during primary and secondary treatment, but depending on the configuration of the WWTP, pharmaceuticals can also be discharged to receiving waters. Due to the aging effect and the increased population established in homes for the elderly or health care institutions, senior residences. Thus, senior residences these establishments can represent a point source pollution of pharmaceuticals to the environment.

The European Medicines Agency (EMA) is an agency of the European Union (EU), responsible for the scientific evaluation, supervision, and safety monitoring of medicines developed by pharmaceutical companies for use in the EU (European Medicines Agency (EMA) 2006). Among other activities, they monitor the safety of medicines across their life cycle. In 2006, EMA proposed the calculation of predicted environmental concentrations (PEC) to estimate the presence of pharmaceuticals in environmental waters and recommended to evaluate their risk when PEC values in surface water were equal or above the threshold value of 0.01 μg/L. This model takes into account the consumption of a specific drug, the excretion rates, and the dilution factor in a particular region and permits to prioritize specific drugs with potential to cause toxic effects at specific water concentrations (Fick et al. 2010). The efficiency and applicability of the approach to determine the theoretical presence of pharmaceuticals in surface waters and wastewaters and to prioritize compounds for further monitoring has been demonstrated by the increasing number of research papers that use this methodology, as in Italy (Riva et al. 2015), Germany (Kümmerer and Al-Ahmad 2010), NW England (Booker et al. 2014), France (Besse et al. 2008), Catalonia (Franquet-Griell et al. 2015), the Netherlands (Oosterhuis et al. 2013), Spain (Ortiz de García et al. 2013) and Poland (Oldenkamp et al. 2013). van Nuijs et al. 2015). According to prescription data (van Nuijs et al. 2015), PEC values and toxicological information, it is then possible to determine the potential risk of pharmaceuticals in the environment (van Leeuwen and Vermeire 2007).

Our hypothesis is that senior residences represent an important source of pharmaceuticals to the environment. In this study, we propose an innovative scheme for assessing the loads and discharge of pharmaceuticals from four senior residences and one sociosanitary center from France, Portugal, and Spain. We have followed EMA guidelines for PEC calculation and risk evaluation based on consumption data, excretion, dilution, and toxicity. We finally provide a list of pharmaceuticals consumed in high quantities in senior establishments and discuss their environmental impact.

Methodology

Health Institutions studied

Health Institutions can be classified in different categories, each with its own specialization, particularity, and functioning, as they host people with different types of illnesses. In this study, we have selected five establishments, one in France, two in Spain, and two in Portugal. For comparability purposes among the three countries, the establishments selected had a high number of beds (>50) and were located in urban areas. Selected establishments were (i) residences oriented to housing or day care for independent individuals that do not require help or assistance and also general impairment (for individuals with general loss of activities of daily function) and (ii) a sociosanitary center which has a high level of impairment. The detailed description of each residence (number of beds, type of facility, and the annual water consumption) is indicated in Table 1.

Table 1 Number of pharmaceuticals administered in five establishments for the elderly according to consumption data in 2015 and 2016
Full size table

In each establishment, we interviewed the manager and the head health professional to obtain information on the type of patients, sickness, and level of impairment. All this information was key to defining the typology of the residence and treatments performed. Then, each establishment provided complete data on the consumption of pharmaceuticals, as the number of pills, injections or other presentations of a specific drug, and their concentration. This information was compiled to identify the main pharmaceutical families administered and to calculate the total amount consumed (milligrams per day) in each establishment (Table 2). To compare the consumption rates in the five establishments, consumption data was normalized per patient so that data is given also in milligrams per day per patient. Data correspond to consumptions in 2015, except for F1 and S1 which correspond to 2016.

Table 2 Pharmaceuticals administered in each residence ordered by consumption data (>1000 mg/day, > 400 mg/day in France given the overall low consumptions), PECres, PECgrid, and PEC river
Full size table

Estimation of the predicted environmental concentrations

PEC calculations were adapted from EMA guidelines to determine the predicted concentrations in effluents from senior residences (PECres), the predicted concentration in the sewerage system (PECgrid), and the predicted concentrations in river (PECriv). The former permitted to determine pharmaceuticals released from health-care establishments according to high consumption and high excretion. On the other hand, PECgrid considered the dilution of the effluent from the health-care facilities to the urban sewerage system. Finally, PECriv considered the elimination in WWTP and the final dilution in receiving waters. PEC values are always given in micrograms per liter.

When calculating PECres, one of the main particularities that might affect the discharge of pharmaceuticals is the people wearing diapers. This implies that an inferior amount of pharmaceuticals than the one that was actually consumed will be discharged. This factor is included in the PECres formula:

$$ \mathrm{PECres}=\frac{{\mathrm{DRUG}}_{\mathrm{consumption}}\times {F}_{\mathrm{Excretion}}\times \left(1-{F}_{\mathrm{Pads}}\right)}{{\mathrm{WATER}}_{\mathrm{consumption}}}\times {10}^6 $$
(1)

where

  • DRUGconsumption (g/day) is the quantity of each pharmaceutical delivered in each senior residence.

  • F Excretion is the excreted fraction of the unchanged drug, considering both urine and feces. Excretion data was obtained from Theriaque database (Amiel and Husson, 1994). Selected values ranged from negligible to 100%, depending on the compound.

  • F Pads is the percentage of patients using incontinence pads. In this study, we used the value of 50% as it represents the mean percentage of patients using incontinence pads in the establishments studied.

  • WATERconsumption (L/day) is the water consumed in each residence per patient per year (Table 1).

  • 10 6 is a conversion factor so that PECres are expressed in micrograms per liter.

To evaluate the amount of pharmaceuticals discharged to the river waters, PECriv (in μg/L) were estimated using the formulas:

$$ \mathrm{PECgrid}=\frac{\mathrm{PECres}}{DF_{\mathrm{GRID}}} $$
(2)
$$ \mathrm{PECriv}=\mathrm{PECgrid}\frac{\left(1-{F}_{\mathrm{WWTP}}\right)}{DF_{\mathrm{RIVER}}} $$
(3)

where

  • DFGrid is an expected dilution factor from senior residences to the general sewage grid. To calculate this factor, we determined the percentage of water from senior establishment to the general volume of water in the sewerage system. In Barcelona and Lisbon, the dilution factor is more than 2500, and in Nimes of 1300 times, based on the volume of sewage water in each sewerage system collected in each treatment plant (27,000 m3/day in Nimes, 450,000 m3/day in Barcelona, 350,000 m3/day in Lisbon). We used an agreement value of 100 times for all three sites considering that the total sewage water is divided in 5-10 collectors in each city and the flow can vary substantially according to discharges, rains, etc..

  • F WWTP is the removal fraction in WWTP. Removal data was obtained from EPI Suite by Environmental Protection Agency (EPA) (EPA 2013). Here, when different data were obtained from the bibliography, the lowest value was applied. In the cases that no information was available, a default value of 0 was used indicating that a given compound is not eliminated.

  • DFRIVER is the dilution factor from WWTP effluents to receiving water and was considered 75.73 for France, 25.92 for Spain, and 61.23 for Portugal, as suggested by Keller et al. (2014). This differential dilution factor is used to better estimate PEC values according to the differences in river flows and dynamics among countries.

For compounds with PECriv >0.01 μg/L, as proposed by EMA, the toxicity data was compiled according to ECOTOX (EPA), DrugBank, or toxicological data sheets of Sigma-Aldrich, ScienceLab, and Santa Cruz Biotechnology.

Prioritization and risk evaluation

It is not feasible to monitor all possible pharmaceuticals present in the environment, and it is necessary to prioritize those that can represent the greatest threat (Donnachie et al. 2016). The consumption data permitted to prioritize compounds with the highest potential impact in river waters from France, Spain, and Portugal. The workflow used is based on

  1. 1.

    Listing of pharmaceuticals according to the consumed data in each senior residence.

  2. 2.

    Calculation of the predicted concentration in the effluents of the senior residence (PECres) for all compounds and all residences.

  3. 3.

    Calculation of PEC in river waters and ranking of compounds with PECriv higher than the 0.01-μg/L threshold level proposed by EMA.

  4. 4.

    Toxicity evaluation using Daphnia magna or other species EC50 or LC50 values, depending on available data.

  5. 5.

    Selection of toxic compounds for which risk assessment is needed.

All this information has been compiled in a database that allows to identify those substances discharged from health-care facilities that may produce an environmental effect.

Results

Consumption of pharmaceuticals

Figure 1 shows the consumption of pharmaceuticals in the five senior residences located in France, Spain, and Portugal. The net total amount of pharmaceuticals ranged between 54 and 1801 g/day, with the residences in Spain being the ones with the highest consumption. According to the size of each residence, this corresponds to an average consumption per day per person ranged from 677 mg in F1 to 13,856 and 9755 mg in S1 and S2, respectively, although the levels were quite similar in P1 and P2 (between 1279 and 1714 mg/person/day). Considering these quantities and taking into account that senior residences have become a living preference in many countries, the amounts of pharmaceuticals discharged to the sewerage system can become a real problem. For instance, there are >8000 public senior residences in France, 5339 in Spain, and 4787 in Portugal, besides the number of sociosanitary centers, which suggests that the estimated total consumption of pharmaceuticals from senior residences should not be disregarded in terms of contribution of pharmaceutical load to the sewerage system and indirectly, to the environment. A rough estimation average discharge considers a consumption of 100 g/day in a median residence of 100 patients, which would mean than on a country base, from 478 to 800 kg of pharmaceuticals are discharged daily from health institutions in south-west Europe. These figures are similar to those reported in Germany (Herrmann et al. 2015). Thus, the incurred impact on WWTP or rivers can be high. In comparison, the potential discharge of pharmaceuticals from the normal healthy population dwelling in an urban area is much lower as consumption is not in such high amounts and discharge is not concentrated in time or space.

Fig. 1
figure 1

Total daily consumption of pharmaceuticals in each residence, indicating the number of residents (N) and the families consumed in each establishment (pie diagrams)

Full size image

The number of pharmaceuticals consumed in each senior residence ranged between 116 and 164 (Table 1). Main pharmaceuticals consumed (>1000 mg/day, 400 mg/day in F1) in each of the five studied residences are indicated in Table 2. Observed differences in main consumed drugs in French, Spanish, and Portuguese residences evidenced the different and specific treatments that patients can receive. A total of 397 common pharmaceuticals were consumed in the five senior residences studied, which belong to 90 therapeutic classes. Table SI1 shows all pharmaceuticals consumed in the five senior residences studied, indicating their Anatomical Therapeutic Chemical (ATC) classification code (WHO 2017). On the other hand, the main therapeutic classes consumed are indicated in Fig. 2 and include antidiabetic agents, analgesics, antibiotics, and antiepileptics as the main treatments for the elderly. Figures 1 and 2 also show the high variability of each therapeutic group consumed in the three countries, reflecting specific treatments according to the different health problems in each establishment or either specific medication according to impairment intrinsic of each patient or typology of residence.

Fig. 2
figure 2

Families of pharmaceuticals (in percentage) most widely consumed in health institutions from southwest Europe (France, Spain, and Portugal). N indicates the number of pharmaceuticals dispensed for each family. Macrogol (laxative) is not represented as its consumption ranges from 13 to 580 g/day which would represent 86% of the total pharmaceuticals consumed

Full size image

Flow of pharmaceuticals from senior residences to river waters

Of the total number of pharmaceuticals consumed, we determined the PEC values in the effluents of senior residences. These calculations took into account that approximately half of the people living in the residences wear incontinency pads. At a glance, this appears to minimize the problem of pharmaceuticals discharged into the wastewaters. However, it is important to note that this is adding up to another problem. If diapers are not properly disposed as biohazard waste in the senior residences, a similar amount of residues is polluting other places, e.g., landfills.

Among the total 397 pharmaceuticals administered in all residences, Table 2 indicates those with amounts administered in quantities higher than 1000 mg/day (400 mg/day in France). PECres varied from negligible to hundreds of micrograms per liter (Table 2). Given the large number of pharmaceuticals administered, it is obvious that the ones consumed at the highest concentration and showing high excretion rates will have higher chances to reach surface waters. We observed that compounds consumed in amounts >1000 mg can be discharged to the sewerage system, and even they are diluted and biodegraded in the WWTP, and they can reach receiving surface waters at the nanogram per liter level. This process of dilution and elimination of pharmaceuticals in WWTP would presumably lead to a concentration in river waters close to the EMA value of 0.01 μg/L. Only 20 had PECriv >0.01 μg/L, with 1 being for F1, 9 for S1, 6 for S2, 1 for P1, and 3 for P2. Table 2 indicates these compounds.

PECriv are also indicated in Table 2 and ranged between 0.001 and 6.15 μg/L, with the highest levels found in Spain due to the lower dilution factor. Comparing PECres and PECriv, the concentrations estimated in river waters represent on average 0.01% or less of the initially discharged by the effluents of the health-care establishments. This decrease in concentration is basically due to 100× dilution in the sewerage system and dilution in river. The WWTP degradability for most of the compounds was very low, and thus, a high proportion of pharmaceuticals will be potentially discharged by the WWTP effluents to receiving waters. Figure 3 shows, using a double axis, this difference. For the studied compounds, the percentage of pharmaceuticals detected in river is low in comparison to the effluents of the residences, indicating that dilution is the main process for the removal of pharmaceuticals once discharged to the sewerage system.

Fig. 3
figure 3

PECres (gray, left axis) and PECriv (light gray, right axis) in each country (F1 = France, S1 and S2, Spain and P1 and P2, Portugal). Only compounds with PECriv >0.01 µg/L according to EMA threshold are represented. This represents a decrease in percentage of more than 99%

Full size image

In France, only one compound, metformin, had PECres of 134 μg/L and PECriv of 0.017 μg/L, which represents 0.013% of the concentration initially discharged given a high excretion rate. Paracetamol had high PECres but was highly eliminated in WWTPs and thus had low PECriv. Dabigatran and amoxicillin were highly biodegradable in WWTP, and the rest of the compounds are expected to be found in the effluents of the WWTP, but once diluted to the river their concentration is expected to be at the low nanogram per liter level.

In both Spanish residences, 15 compounds exceeded the EMA threshold value (0.01 μg/L). PECriv ranged from 0.01 to 0.086 μg/L, except for macrogol which had PECriv from 0.459 to 6.015 μg/L. Macrogol is the international nonproprietary name for polyethylene glycol used primarily as laxative or also as excipient in many pharmaceutical products. It is consumed in high amounts in Spain (57–581 g/day in the two Spanish residences or from 0.5 to 5 g/inhab/day) and is rapidly excreted and poorly biodegraded in WWTP. WWTP were only partially efficient in eliminating ibuprofen (29%) and megestrol (30%), but despite the high PECres, they would not be expected to be found in river waters if originating only from health-care facilities. However these compounds are widely used by the overall population and frequently detected in river waters (van Nuijs et al. 2015). Overall, in Spanish residences, compounds with the highest PECriv were macrogol, paracetamol, metformin, acetylcysteine, amoxicillin gabapentin, cyanocobalamin, levetiracetam, levofloxacin, and megestrol. The low dilution factor is mainly responsible for the high PECriv of these compounds.

Finally, in Portugal one compounds in P1 and 3 in P2 were consumed in amounts which produced PECriv higher than the EMA 0.01 μg/L threshold value (Table 2), indicating that despite the high dilution factor compared to Spain (61.23 vs 25.92), the pharmaceuticals consumed in senior residences might contribute to river water contamination. In Portugal, the compounds with the highest PECriv were macrogol, metformin, and piracetam.

When estimating PECriv for compounds with consumption of 1000 mg/day, we observed that dilution in the sewerage system, elimination in the WWTP, and dilution in river waters may not be enough to eliminate all pharmaceuticals. We have then identified that some compounds present in effluents would be presumably detected in river waters at levels of 0.01–6.15 μg/L. The compounds that should be considered as suspect compounds as they could be present in river waters at concentrations >0.01 μg/L are indicated in Table 2 for each establishment.

Considering the three countries, the most consumed drugs and for which PECriv is higher than 0.01 μg/L proposed by EMA are listed in Table 2. Altogether, 12 compounds of the 397 commonly administered in health institutions had PECriv >0.01 μg/L. Of these, only metformin was common in all countries. The rest of the compounds are specific of a given country or even health institution, indicating that there is a wide variability on the pharmaceuticals administered to patients, even though most belong to the same family. Many of these compounds have been previously identified as most commonly detected in the environment. For instance, metformin has been previously identified as one of the main pharmaceuticals in wastewaters in the Netherlands (Oosterhuis et al. 2013). Similarly, Van Nuijs et al. detected metformin, valsartan, and tramadol in sewage water with good correlation with prescribed values (van Nuijs et al. 2015). In Iraq, paracetamol, amoxicillin, and metformin had an annual consumption exceeding 1000 t per year and were expected to produce a risk (Al-Khazrajy and Boxall 2016).

Prioritization of pharmaceuticals for further treatment and risk assessment

Table 3 gives the physico-chemical characteristics of the prioritized pharmaceuticals according to PECriv >0.01 μg/L EMA threshold and includes also common phramaceuticals in the 3 countries with levels >0.001 μg/L. Most of them have high solubility and low logP, indicating that preferentially they will remain in water. Even though pharmaceuticals can be degraded in water (Carlsson et al. 2006), their continuous discharge, even at low concentrations, makes these drugs recalcitrant and environmentally hazardous compounds. Because of the lack of a legislation that controls the levels of drug residues in discharges and in surface waters, it is important to prioritize actions that minimize the impact of these pollutants on the environment. Thus, the theoretical evaluation of presence and risk can provide a new and simple to use tool to predict their presence in the environment. These tools can be extrapolated to other areas with similar problems (e.g., kindergartens, hospitals).

Table 3 Physico-chemical properties of prioritized pharmaceuticals according to PEC data with values higher than 0.001 µg/L and reported aquatic toxicity (Aquatox μg/L)
Full size table

For the 12 compounds likely to be consumed in amounts that can reach river waters, we determined the aquatic toxicity using different organisms according to available data from the open bibliography (Table 3). The toxicity in general is low. However, it has been pointed out that there is scarce information about long-term effects of pharmaceuticals to aquatic organisms, in particular with respect to biological targets (Fent et al. 2006). In the environment, acute toxic effects of pharmaceuticals are unlikely but chronic ecotoxicity data is needed for a correct evaluation of risk (Carlsson et al. 2006).

Main compounds to be released from health-care institutions were laxatives such as macrogol and lactitol; analgesic and antipyretic drugs such as paracetamol and piracetam; antibiotics such as amoxicillin and levofloxacin, gabapentin, and levetiracetam for the treatment of epilepsy and neuropathic pain; and pharmaceuticals for the treatment of diabetes (metformin and insulin). Prioritization according to PECs helps in the implementation of focused monitoring programs that consider only the most abundant and toxic compounds, in an attempt to control and mitigate the release of pharmaceuticals consumed in health care institutions.

Risk management

The new “Urban Water Agenda 2030,” addressed at the Leeuwarden Conference (02.2016), incorporates concerns about wastewater treatment by focusing on emerging contaminants to contribute to the achievement of the good chemical status of water bodies. The main objective is to ensure the quality of water for urban use.

In this study, we have identified health institutions as a point source pollution of pharmaceuticals to the environment, with similar conclusions as those obtained in Germany (Herrmann et al. 2015). The number of homes for elderly people is currently high and is expected to increase in the future. This is worrying because these establishments are a considerable source of emerging pollutants and, hitherto, there are no guidelines or information about the risk management of effluents, which are typically classified as domestic. Nonetheless, the World Health Organization (Chartier 2014) alerts that although a large part of the wastewater from health-care facilities can be considered domestic (because they pose the same risks as domestic wastewater), depending on the service and tasks of the facility, these wastewaters might contribute to the contamination of rivers. This is clearly the case for the sociosanitary centers, where the consumption of pharmaceuticals is high.

Consequently, as for classical hospitals, risk assessments have to be envisaged for elderly residences. Several possibilities can be imagined. One of them could be the application of the recent French guideline for good management of waste produced by health-care establishments (edited by the French Ministry of Health) that propose initiatives to organize the management of liquid waste (http://social-sante.gouv.fr/IMG/pdf/pour_une_bonne_gestion_des_dechets_produits_par_les_etablissements_de_sante_et_medico-sociaux.pdf).

Onsite treatment could be another effective strategy to manage the release of pharmaceuticals to the environment. A future avenue for this area would be to conduct a cost-benefit analysis and sociological studies to know the viability of this strategy.

Daily practices could be also a good but challenging procedure to mitigate the release of pharmaceuticals, and this can be done specifically for incontinency pads.

Overall risk management is a complex issue because it involves many and different types of stakeholders, such as environmental and health authorities, the pharmaceutical sector, water and waste industries, health practitioners, researchers, and elderly home managers and clients, as well as the general public. Environmental, social, and economic objectives have to be considered for identifying the problem and determining the risk clearly and early in the process, so that mitigation actions can be achieved. To facilitate this process, risk assessment and risk management should be integrated activities and should share a common requirement which is effective risk communication (Naidu et al. 2016).

Knowledge on the risks associated with pharmaceutical residues in wastewater is very weak, sometimes controversial, and in part unknown (García-Santiago et al. 2016; Touraud et al. 2011). A perception study on elderly people’s residences is currently in progress, and the first results show a real misunderstanding of the problem (data not shown). Related knowledge, attitudes, and social representations have yet to be established from a social science point of view (Pidgeon et al. 2011). This elaboration process starts in the inquiries about the topic (Lichtenstein and Slovic 2006) and the cultural and social dispositions of people (Kahan 2009). For this reason, the risk and benefits of pharmaceuticals, namely, openness through frequent dialogs, decisions based on the best available science, transparency, timeliness, and responsiveness, should be taken into consideration (Bouder 2011). Furthermore, campaigns to increase risk awareness should be initiated before any alarm episode (Barnett and Breakwell 2003) or crisis (Gaspar et al. 2015).

Conclusions

We have identified senior residences as a source of pharmaceuticals to surface waters at concentrations higher than 0.01 μg/L, which is the EMA threshold for risk assessment. Depending on the size of the elderly people’s home, and taking into account the circumstances and medical treatments usually received, wastewaters contain pharmaceuticals in their effluents at concentrations of hundreds of micrograms per liter. Because these effluents are discharged to sewerage systems and WWTP are mostly inefficient to eliminate pharmaceuticals, residues are discharged to river waters, thus contributing to water pollution. This effect, amplified by the large number of residences in the south-west Europe, indicates the importance of controlling the discharges of pharmaceuticals form senior residences to minimize the impact on aquatic ecosystems. A protocol scheme and risk management actions foreseen should be used to implement focused monitoring and remediation technologies that consider the most toxic compounds to ensure effectiveness in the control and evaluation of the impact of pharmaceuticals released from health-care facilities.