Report on pharmaceuticals and personal care products in illinois drinking water

Report on Pharmaceuticals and Personal Care Products
in Illinois Drinking Water
Bureau of Water, Illinois EPA
June 2008
Introduction

While the presence of pharmaceuticals and personal care products (PPCPs) in raw (untreated)
and finished (potable) drinking water has become an issue of concern recently, the original
reports of pharmaceutical chemicals’ presence in water go back three decades. Garrison et al.
(1976) and Hignite and Azaznoff (1977) both reported the presence of clofibric acid, a
breakdown product of several blood lipid regulators, in wastewater, and Hignite and Azaznoff
also found salicylic acid, an aspirin breakdown product, in their study. As analytical techniques
became increasingly sensitive and detection limits approached and sometimes surpassed the low
nanograms per liter (ng/L) or parts-per-trillion (ppt) level, many more PPCPs have been reported
in waste water, ambient water, and drinking water. In one recent survey of 139 U.S. streams,
Kolpin et al. (2002) found PPCPs in 80% of the streams, while in another report Heberer (2002)
reviewed research on pharmaceuticals in water and listed 80 drugs and breakdown products that
had been detected.
The issue of PPCPs in drinking water was brought to the forefront earlier this year when the
Associated Press released a three-part series of reports that found PPCPs in the drinking water of
24 U.S. metropolitan areas serving approximately 41 million residents. Acting on these reports,
Governor Blagojevich requested that the Illinois Environmental Protection Agency (Agency)
monitor water samples for the presence of PPCPs, and that the Agency and the Illinois
Department of Public Health (IDPH) assess the effects on public health of any chemicals that
might be found.
Purpose

Illinois EPA Bureau of Water (Division of Public Water Supplies) staff collected samples of raw
and finished drinking water that were analyzed for the presence of pharmaceuticals, in order to
evaluate whether detectable amounts are present in sufficient concentration to cause adverse
human health effects.
Methodology

Sample Selection – Chicago and four other communities were selected for sampling. Chicago
was chosen because of the large population served, considering the city itself and the numerous
neighboring communities that purchase water from Chicago. Four communities (Elgin, Aurora,
Rock Island and East St. Louis) were chosen because they use surface water (Fox River and
Mississippi River) as a drinking water source and are located downstream near a wastewater
treatment plant discharge. Since the major route for pharmaceuticals’ entry to surface water is
primarily through discharge of treated municipal wastewater, the selected water supplies are
more likely than others to show detectable levels of these substances.

Sample Collection – Samples were collected starting Monday, March 24 and continued through
Thursday, March 27, 2008. The samples were collected following standard procedures by
Agency staff, using bottles provided by the laboratory. Samples were express shipped to the
South Bend, IN office of Underwriters Laboratories on the day of collection. Once the
laboratory received the samples, results of analyses were to be available within 21 to 28 days.
For the initial set of analyses, untreated and potable water samples were collected from Chicago,
Elgin, Aurora, Rock Island, and Illinois American Water Company – East St. Louis Division.
Chemical Analyses – Underwriters Laboratories was selected to perform the analyses of the
water samples, using their certified methods L220 and L221 for Pharmaceutically Active
Compounds. These methods are capable of detecting 56 compounds that are found in many
types of PPCPs, such as pain relievers, antibiotics, anticonvulsants, antidepressants, replacement
hormones, an insect repellant, and chemicals related to coffee and tobacco. Chemicals reported
by these methods, their detection limits, and a brief description of the chemicals are listed in
Table 1.

Screening Levels – Upon receipt of the analyses after final quality assurance from the laboratory,
the results were provided to Agency and IDPH toxicologists for review and interpretation of
whether there are possible adverse human health effects that may be associated with
consumption of the potable water. Since there are no established standards or guidelines for the
chemicals analyzed for this project, it was necessary to develop Screening Levels for these
chemicals. In consultation with IDPH toxicologists and other health professionals, the Agency
chose to develop the Screening Levels for the PPCPs using a conservative risk assessment
approach. This approach drew heavily on the procedures used in the recently finalized
Australian Guidelines for Water Recycling (2008) to develop Drinking Water Guidelines
(DWGs) to be applied to recycled wastewaters in Australia.
The Australian procedures rely on two large sources of toxicological data as the starting point for
deriving the DWGs for pharmaceuticals. The first source is the Acceptable Daily Intakes (ADIs)
developed for human exposures to pharmaceuticals with agricultural and veterinary applications.
The ADIs have been developed by the European Medicines Association Committee for
Veterinary Medical Products, the Joint FAO/WHO Expert Committee on Food Additives, or the
Australian Therapeutic Goods Administration, and are used unaltered in the development of the
DWGs. The Agency and IDPH toxicologists have also chosen to use the unaltered ADIs in
deriving the Screening Levels for this project.
The second source is the Lowest Daily Therapeutic Doses (LDTDs), in milligrams per day
(mg/d), developed for human pharmaceuticals. The LDTD represents a balance between the
beneficial effect of the drug and its known or potential adverse side effects. While human drugs
receive extensive safety evaluations before release, much of the testing data remain confidential
and thus unavailable for use in deriving drinking water criteria. In developing the DWGs,
therefore, the Australians assumed that the LDTD represents the lowest observable effect level
(LOEL) for side effects, and then applied safety factors appropriate to the drug to extrapolate
from the LDTD to a dose that would be without effect even for sensitive subgroups of the
population. For most drugs the safety factor is 1,000, and an additional safety factor of 10 is applied to highly cytotoxic (ex., chemotherapy) or hormonal (ex., birth control) drugs. The Agency and IDPH toxicologists also chose to use this approach, but decided that for developing our ADIs a safety factor of 10,000 is appropriate initially, rather than using a safety factor of 1,000 and additional factors added for specific types of drugs. Thus, the LDTD was divided by a series of four safety factors, each a value of 10, that took into account extrapolation from a LOEL to a no observable effect level (NOEL), intrahuman variability (adults vs. children), short-term vs. long-term effects, and therapeutic use vs. no therapeutic need, to arrive at the ADIs to be used in developing the Screening Levels. Since the LDTDs are expressed in mg/d, it was also necessary to convert this into a dose based on body weight, in milligrams per kilogram of body weight per day (mg/kg/d). We chose to use the average body weight for a young child of 10 kg, as discussed below, in making this conversion. As an example of the development of an ADI for this project, the LDTD for carbamazepine is 200 mg/d, which was divided by the safety factor of 10,000 to obtain a safe level of 0.02 mg/d. This was then divided by the assumed 10 kg body weight to derive the ADI for this project of 0.002 mg/kg/day. Since the units used for the analytical results in this report are nanograms per liter (ng/L), all other units in this report will be converted to nanograms; thus for carbamazepine the ADI of 0.002 mg/kg/d is equivalent to 2,000 nanograms per kilogram per day (ng/kg/d). There also were four chemicals detected that are not human or animal drugs and thus do not have ADIs or LDTDs: caffeine, nicotine, paraxanthine, and DEET. The Agency and IDPH toxicologists determined that there are no appropriate toxicological data available at this time to allow development of an ADI for the first three chemicals. Regarding DEET, the California Environmental Protection Agency has developed a Risk Characterization Document for this chemical, which identified a two-year study with rats that found a NOEL of 100 mg/kg/d for reduced body weight and food consumption and increased cholesterol (Goldenthal, 1995). The Agency and IDPH toxicologists used this study as the basis for developing an ADI, by dividing this NOEL by three safety factors of 10, or a total safety factor of 1,000, to account for extrapolation from animals to humans, for intrahuman variability, and for protection against seizures that have been reported in a small number of children who used large amounts of DEET. Thus, the ADI for this project is 0.1 mg/kg/d, or 100,000 ng/kg/d. It should be noted that California EPA also calculated Annual Average Daily Dosages (AADDs) in various age groups from dermal exposures based on the results of a survey of DEET use, and the ADI falls within the reported AADD range of 37,000-130,000 ng/kg/d. The final step in the process of deriving the Screening Levels was to determine the maximum concentrations of the PPCPs in drinking water that would not result in people consuming amounts of the PPCPs in excess of the ADIs. This was done by using the procedures used by many regulatory agencies to derive drinking water criteria: Criterion (ng/L) = [(ADI x BW)/IR] x RSC, where IR = drinking water ingestion rate (L/d) RSC = relative source contribution (% of daily intake attributable to The Australians used standard risk assessment assumptions for lifetime exposures for the BW
and IR inputs to the equation, assuming an adult body weight (BW) of 70 kg and an adult water
ingestion rate (IR) of 2 liters per day (L/d), but decided that the default RSC of 20% of the daily
exposure derives from drinking water was unreasonable. Instead, they reasoned that the daily
exposure from sources other than water will be zero unless the drug has been prescribed for the
person, so the RSC should be 100%. The Agency and IDPH toxicologists agreed with the RSC
selection, but decided that the BW and IR terms should reflect a young child’s exposure rather
than an adult’s. Therefore, values of 10 kg for BW and 1 L/d for IR were chosen. These
changes resulted in Screening Levels that are 3.5 times more restrictive than the Australian
DWGs for most PPCPs. The Agency and IDPH toxicologists believe that this conservative
approach is very protective of public health. The Screening Levels derived from these
procedures are listed in Table 2.
Results and Discussion
In order to evaluate the PPCP concentrations detected in the samples from the five public water
supplies, the Agency compared the reported concentrations to the Screening Levels to calculate a
Hazard Index (HI) for each chemical. The HI is a ratio of the actual exposure to the acceptable
exposure, and if the HI does not exceed 1.0 the exposure is at an acceptable level.
Concentrations detected in the raw and finished water samples, the Screening Levels, and the
corresponding HIs for the finished water samples are listed in Table 2.
As can be seen from this Table, all HIs are much lower than the critical value of 1.0, ranging
from 0.003-<0.00000001. This indicates that the concentrations of the PPCPs in the samples do
not pose a public health hazard at this time. The largest HI of 0.003, for cotinine (a breakdown
product of nicotine) in the Elgin sample, suggests that there is a margin of safety of at least 333
(1.0/0.003), and likely considerably higher because of the conservative nature of the Screening
Levels, for exposure to this chemical in the drinking water.
There are some interesting features that are apparent from the results. The Chicago sample of
raw water suggests that Lake Michigan is a relatively clean source of drinking water, with less
total numbers of PPCPs detected (4 chemicals) in comparison with the supplies drawing from
river sources (9-14 chemicals). This result may be representative of lakes in general, since
results reported to the Agency for raw water from Lake Springfield, analyzed using the same two
analytical methods as in this project, also are lower (7 chemicals) than the range for the river
samples (chemicals and levels not presented). The Lake Michigan sample also had generally
lower concentrations of the PPCPs that were detected than the corresponding results from the
river sources; concentrations of cotinine, nicotine, and gemfibrozil were higher in the river
samples, while the levels of monensin were comparable.
The results from the untreated water samples from the rivers suggest that agricultural sources
may be important contributors to the load of pharmaceuticals in the source water of these
supplies. Several drugs that are primarily or exclusively used in agricultural or veterinary
treatments (lincomycin, monensin, sulfadimethoxine, and sulfamethazine) were detected in the
river samples, although the HIs were very low. These results suggest a potential control point if
these chemicals become a concern in the future.
The results for the untreated versus finished samples from all facilities except Aurora indicate
that routine water treatments are capable of reducing or eliminating the levels of some of the
PPCPs found in the raw water while other chemicals are only minimally reduced. (The Aurora results are not comparable to the results from the other facilities since the finished water at the time the sample was collected was a blend of approximately equal amounts of water from the river and the facility’s well field). The results listed in Table 2 show that diltiazem, lincomycin, sulfadimethoxine, sulfamethoxazole, and trimethoprim are mostly or fully removed from the raw water by the facilities’ treatments, while the results for caffeine, fluoxetine, paraxanthine, and sulfamethazine are inconclusive because of insufficient or conflicting data. On the other hand, the results show that carbamazepine, cotinine, DEET, gemfibrozil, monensin, naproxen, and nicotine are minimally removed by treatment. These last results are not surprising, as most of these chemicals have been reported to persist in drinking water following treatment in studies of the effectiveness of treatment processes in removing PPCPs (Stackelberg et al., 2007; Westerhoff et al., 2005). While the concentrations detected and HIs calculated for this project were very low, it is likely premature to suggest that the issue of PPCPs in drinking water is resolved at this time, as some uncertainties remain. Obviously, the database developed in this project is small, leaving considerable uncertainty about the potential range of chemicals and concentrations that may be present in untreated and potable drinking waters across the state. The timing of the sample collection (late-March), when the rivers involved in this project were at high flow levels, likely contributed to an underestimate of the levels of the PPCPs that might be present in the water, due to dilution. Indeed, a study by Loraine and Pettigrove (2006) reports a significant difference in the concentrations of some PPCPs between low-flow and high-flow stream conditions, with some chemicals measured at low-flow conditions approaching levels found in wastewater discharges. Another potentially significant uncertainty for this project is that the analytical methods used in this project are not capable of detecting some chemicals/chemical families that have been identified as potential problems because of high use, high levels found in some studies, and/or high toxicity reported in studies of PPCPs in water. Examples include: • codeine – high use (maximum detected in raw water = 1,000 ng/L, Kolpin et al., 2002) • diazepam (Valium) – high use, high toxicity (Screening Level = 500 ng/L) • the anti-acid drug ranitidine (Zantac) – very high use • the beta-blockers bisoprolol and propanolol – high toxicity (bisoprolol Screening Level = 125 ng/L), some high levels found (bisoprolol maximum concentration in raw water = 2,900 ng/L, Daughton and Ternes, 1999) • the chemotherapy drugs cyclophosphamide and isophosphamide – high toxicity, and • the estrogenic hormones 17 beta-estradiol and 17 alpha-ethinyl estradiol – very high estrogenic activity (Screening Levels = 500 and 30 ng/L, respectively).
If funding were to become available, it would be informative to follow up this project with
additional samples to expand the coverage of drinking water sources across space and time, and
to include other PPCPs if appropriate analytical procedures can be identified.

Conclusions
This project has identified 16 PPCPs in the untreated or potable water of five public water
supplies in Illinois. The results for the potable water samples were compared against
conservative Screening Levels developed by Agency and IDPH toxicologists, and were found to not present a public health hazard at this time. These comparisons suggest that even the chemical with the highest Hazard Index has a margin of safety of at least 333, and likely much larger. However, there are also considerable uncertainties that suggest that further sampling is appropriate if funding can be made available. REFERENCES
Australian Guidelines for Water Recycling: Augmentation of Drinking Water Supplies. May 2008. A publication of the Environmental Protection and Heritage Council, the National Health and Medical Research Council, and the National Resource Management Ministerial Council. California Environmental Protection Agency, Department of Pesticide Regulation. September 2000. N,N-Diethyl-meta-Toluamide (DEET) Risk Characterization Document. Document # RCD 00-01. Daughton CG and Ternes TA. 1999. Pharmaceuticals and personal care products in the environment: agents of subtle change? Environ. Health Perspect. 107 (Suppl. 6): 907-937. Garrison AW, Pope JD, and Allen FR. 1976. Analysis of organic compounds in domestic wastewaters. In: Identification and Analysis of Organic Pollutants in Water (Keith CH, ed.). Ann Arbor, MI: Ann Arbor Science Publishers. pp 517-556. Goldenthal EI. (International Research and Development Corp.). 1995. Evaluation of DEET in a two-year dietary toxicity and oncogenicity study in rats. DEET Joint Venture/Chemical Specialties Manufacturers Assoc. DPR Vol. 50191-176, Rec. No. 133986. Heberer T. 2002. Occurrence, fate, and removal of pharmaceutical residues in the aquatic environment: a review of recent research data. Toxicol. Lett. 131: 5-17. Hignite C and Azaznoff DL. 1977. Drugs and drug metabolites as environmental contaminants: chlorophenoxyisobutyrate and salicylic acid in sewage water effluent. Life Sci. 20: 337-342. Kolpin DW, Furlong ET, Meyer MT, et al. 2002. Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. streams, 1999-2000: a national reconnaissance. Environ. Sci. Technol. 36: 1202-1211. Loraine GA and Pettigrove ME. 2006. Seasonal variations in concentrations of pharmaceuticals and personal care products in drinking water and reclaimed wastewater in southern California. Environ. Sci. Technol. 40: 687-695. Stackelberg PE, Gibs J, Furlong ET, et al. 2007. Efficiency of conventional drinking-water-treatment processes in removal of pharmaceuticals and other organic compounds. Sci. Total Environ. 377: 255-272. Westerhoff P, Yoon Y, Snyder S, et al. 2005. Fate of endocrine-disruptor, pharmaceutical, and personal care product chemicals during simulated drinking water treatment processes. Environ. Sci. Technol. 39: 6649-6663. TABLE 1. CHEMICALS REPORTED BY UNDERWRITERS LABORATORIES METHODS L220 AND L221 CHEMICAL DETECTION Method L220
TABLE 1, continued.

CHEMICAL
Method L221
Active metabolite of several lipid regulators Doxycycline 50 Antibiotic Gemfibrozil 0.5 Oxytetracycline 500 Antibiotic Penicillin G Sulfadiazine 50 Antibiotic Sulfadimethoxine 5.0 Sulfamethizole 5.0 Antibiotic Sulfamethoxazole 2.0 TABLE 2. CHEMICALS DETECTED IN RAW AND FINISHED DRINKING WATER, SCREENING LEVELS, AND HAZARD INDICES CHEMICAL DETECTED Aurora (NOTE: Finished water approximately 50:50 surface & well water
Caffeine 50
E St Louis
Rock Island
(1) The screening level pertains to the sum of all sulfa drugs.

Source: http://www.epa.state.il.us/water/pharmaceuticals-in-drinking-water.pdf