Kevin J. Bisceglia, PhD, Assistant Professor of Chemistry, Hofstra University
Sewers are mostly invisible, and we like it that way. No one likes to think about what goes on in one, or more accurately, what flows through one. Instead, sewers discretely carry away our waste to be treated, allowing many of us to remain in general ignorance of their very existence. This is a shame, as the proper collection, separation, and treatment of water and wastewater is almost certainly the single most effective thing humanity has done to reduce mortality from infectious disease (Cutler & Miller, 2005). Perhaps even more than modern medicine, it may be the reason you’re alive and healthy enough to read this article today. Don’t fret if you’ve never had warm thoughts about your sewer system, however. Good water sanitation practices have been in place for most of the United States since the 1930s (Cutler & Miller, 2005) – odds are that you probably can’t even remember life before them. You might also be one of more than 30 percent of Americans who drink bottled water exclusively (Huerta- Saenz et al., 2012), and might think that public sanitation is of less concern for you. Finally, and let’s be honest here, sewers remind us of their presence in pretty awful ways – backed-up toilets, contaminated waterways, and that distinctive, awful smell. Given all that, it’s only natural to think, “Sure, sewers have protected my family’s health for the past 90 years, but what can they do for me now?” The answer may be more than you’d expect.
The Challenge of Obtaining Community-Level Health and Lifestyle Data
Most of what we know about the health and lifestyle choices of our communities comes from oral and written population surveys (Daughton, 2011). Additional knowledge about drug abuse and medical conditions is also gleaned from arrest statistics and hospital visits, but essentially, current practices rely on our willingness to share sensitive, potentially incriminating information with perfect strangers who are (often) acting as agents of the government. Self-reporting errors and sampling biases are just two acknowledged concerns with this approach. Studies that are able to independently verify self-reported data consistently find that it underestimates the prevalence of actual behavior, often by more than 50 percent (Magura, 2010). Biases arising from participant selection, nonresponses, and even the format of survey questions have been well-documented (Bowling, 2009).
Even when error and bias can be adequately characterized, survey-based monitoring of community-level public health requires significant investments of time, manpower, and money (SAMHSA, 2013). These barriers make it extremely difficult to collect, interpret, and disseminate results in short time frames (e.g., less than one month), and limit both the quantity and spatial resolution at which data can be collected. In short, real-time, continuous collection of public health data is just not possible with surveys alone.
Another potential approach to monitoring health and lifestyle choices at the community level relies on a discomforting fact: If you live in a city or densely populated suburb (such as Nassau County), your sewer system acts as a repository for countless chemical and biological agents excreted by you and your neighbors. Examples include drugs consumed for medicine or recreation, infectious viruses and bacteria, hormones and other endogenously-produced (bio)chemicals. Many of these agents are used in the medical community as biomarkers of specific activities (e.g., drug abuse), as well as aspects of physiological and emotional health (e.g., certain types of cancer, levels of oxidative stress).
I belong to a growing community of scientists investigating whether sewerbased monitoring of health and lifestyle biomarkers can complement traditional surveys in monitoring community-level public health. Compared to surveys, sewer-based monitoring is likely to be less intrusive, more quantitative, and more anonymous (in a sewer, your biomarkers mix with everyone else’s and most contain no personally identifiable features). Moreover, sewer-based monitoring could be done in near real-time, on a continual basis, and at the neighborhood level. Increasingly, this approach is being called sewage epidemiology.[quote style=”boxed” float=”right”]“ … If you live in a city or densely populated suburb (such as Nassau County), your sewer system acts as a repository for countless chemical and biological agents excreted by you and your neighbors.”[/quote]
Using Sewage to Drug Test a City
Techniques already exist to monitor for chemical biomarkers in municipal sewage in near real-time and with greater spatial resolution than oral and written surveys (Castiglioni et al., 2013). They were developed by chemists and engineers to evaluate municipal wastewater as a source of human-derived micropollutants (e.g., plasticizers, hormones, personal care products, flame retardants, antibacterial agents) in the environment. Widespread, low-level contamination by such chemicals has been identified as one of the principal environmental challenges of our time (Schwarzenbach et al., 2006), and environmental chemists (myself included) have spent countless hours in an ongoing effort to identify and characterize those that are of the most concern. Pharmaceuticals are a major subclass of human-derived micropollutants that, over time, has grown to include drugs of abuse as well (Daughton, 2011). All that was required was a shift in perspective, from viewing such chemicals as environmental contaminants, to viewing them as a source of information about human activity (Figure 1).
The first researchers to make this shift did so with cocaine (Zuccato et al., 2005); their approach has become standard for converting sewage-based measurements of occurrence into estimates of drug abuse. Essentially, a sewage sample is processed to remove unwanted material (you can guess what this material is) and concentrate biomarkers of interest. For cocaine, the most common biomarker is benzoylecgonine (BE), its major metabolite. After processing, each biomarker is separated based on its physicochemical properties using a technique called chromatography. Then, both the abundance and identity of each biomarker are determined by mass spectrometry, which is in some ways akin to taking a chemical “fingerprint” (Figure 2). Finally, measurements of biomarker occurrence are converted into estimates of drug consumption by adjusting for factors such as metabolism (ƒex, if the biomarker is a metabolite), molecular weight (MW), wastewater flow (Qw), and the number of people within the sewer catchment (P):
Since 2005, the sewage epidemiology approach has been employed to estimate community-level consumption of more than 20 drugs of abuse, in more than two dozen cities (Castiglioni et al., 2013). Consumption estimates have generally agreed with results obtained via conventional means (including surveys). Equally important, the approach has been demonstrated to have sufficient spatial and temporal resolution to catch changes in consumption that occur during weekend festivals and local sporting events (Castiglioni et al., 2013). One of the most striking findings to come from sewage epidemiology is that drugs of abuse are just as abundant as prescription and over-the-counter (OTC) drugs. Figure 3A contains representative data for the occurrence of drugs of abuse in municipal sewage from Baltimore, MD (Bisceglia et al., 2010a). Plotted from left to right are: cocaine and its principal metabolite, benzoylecgonine; methamphetamine (aka meth); MDMA (aka ecstasy); morphine, the principal metabolite of heroin; oxycodone (aka Oxycontin); and cotinine, the principal metabolite of nicotine.
For comparison, representative data for the occurrence of prescription and over-the-counter drugs (Bisceglia et al., 2010b) are presented in Figure 3B. From left to right are: acetaminophen (aka, Tylenol, an anti-inflammatory agent); metoprolol (aka Toprol, used to treat hypertension); carbamazepine (aka Tegretol, used to treat seizures and bipolar disorder); brompheniramine (aka Bromfed, Dimetapp, an antihistamine); diazepam (aka Valium, used to treat anxiety); and terbinafine (aka Lamisil, an antifungal agent).
Both figures plot occurrence on a logarithmic scale, as nanograms of drug per liter of sewage (ng/L). For perspective, 1 ng/L is equivalent to one molecule of drug in one trillion “molecules” of sewage, or one postage stamp in an area the size of New York City – all five boroughs. These concentrations may seem miniscule, but modern analytical instrumentation is sensitive, so much so that we were able to measure cocaine and other drugs presented above without preconcentration. Moreover, low concentrations add up to moderate estimates of consumption in large sewer catchments, as is demonstrated for Baltimore, MD, in Figure 4. As with occurrence, sewer-derived estimates of cocaine consumption are roughly one-tenth that of heavily used OTC drugs like acetaminophen, and well within the range of common OTC and prescription drugs.
A caveat: Data in Figures 3 and 4 are means of composite daily samples, and are not necessarily representative of long-term drug use in Baltimore. Nonetheless, our estimate of cocaine consumption for Baltimore agrees with values from detailed sampling campaigns in Europe, which range from 2 to 3 mg/inhab-d (Castiglioni et al., 2013). Unfortunately, there is very little sewage-derived data on drug consumption in the United States. At the time of its publication, ours was only the second sewage epidemiology investigation conducted in the United States. As of 2013, there are still only a handful of U.S. investigations; ours remains the most comprehensive in terms of the number of drugs of abuse and drug metabolites monitored – 23 in total (Bisceglia et al., 2010a).
Minimizing Uncertainty, Moving Toward Routine Analysis
The purpose of sewer systems is simple. Frustratingly, the conditions that exist inside sewers are anything but. It is no surprise, then, that proponents of sewage epidemiology have devoted most of their recent efforts toward identifying and minimizing uncertainties associated with the approach. While the magnitudes of uncertainties are biomarkerspecific, many sources of uncertainty are likely to be universal. They include variations in sewer flow and biomarker loading, analytical uncertainty, biomarker stability (i.e., transformation and/or partitioning that occurs in sewers), estimation of population size and, for metabolites, variability in metabolic excretion.
Castiglioni et al. (2013) have summarized efforts to characterize these sources of uncertainty for drugs of abuse. As long as a biomarker is stable in sewage and not subject to metabolic variability (i.e., not a metabolite), overall error can usually be kept below 10 percent for estimates within the same sewer catchment. This makes routine sewer monitoring a distinct possibility for methamphetamine, cannabinoids, oxycodone and hydrocodone, antiepileptic drugs, certain antibiotics, and a great many other legal and illegal drugs. Comparisons between catchments are also possible, provided that one can suitably estimate real-time population size in each sewer system (for additional details, see Brewer et al., 2012).
Even when biomarker stability is an issue, accurately and routinely estimating usage may still be possible. For example, cocaine rapidly degrades in wastewater to form benzoylecgonine (Figure 5). While benzoylecgonine is stable, in-sewer production from cocaine can cause over-bias consumption estimates by up to 50 percent, depending on conditions (Plósz et al., 2013). Even worse, benzoylecgonine is a metabolite of cocaine. Its occurrence in sewage is also affected by interpersonal variations in cocaine metabolism, which depend on genetics, lifestyle choices, route of administration, and other factors. This variability in excretion can also introduce uncertainties of up to 50 percent (Castiglioni et al., 2013).
Recently, colleagues at Johns Hopkins University and the National Institute of Standards and Technology (NIST) and I have developed an approach for minimizing these uncertainties that takes advantage of cocaine’s proclivity to break down in sewage (Bisceglia et al., 2012). By heating sewage samples under basic pH (thus encouraging a chemical process called hydrolysis), we are able to collapse cocaine and all of its major metabolites to ecgonine, the cocaine molecule’s “backbone” (Figure 5). This backbone appears to be stable in sewage; monitoring cocaine utilization via its presence in posthydrolysis samples reduces uncertainty from both in-sewer transformations and metabolic variability to about 10 percent – in keeping with sources of uncertainty that do not depend on biomarker identity (Bisceglia & Lippa, 2013). Finally, as a bonus, the hydrolysis procedure dramatically streamlines cocaine analysis by eliminating the need to pre-concentrate samples.
Looking Toward the Future: Applications of Sewage Epidemiology in Law Enforcement and Public Health
Feedback on Policies and Practices
Sewage epidemiology has substantial, largely untapped, potential to address the efficacy of a variety of important public health, law enforcement, and legislative strategies. It is currently being used to monitor trends in the use (and potential abuse) of attention deficit disorder (ADD) drugs on a college campus (Burgard et al., 2013). Realtime, quantitative data on drug utilization might prove useful in answering a variety of other important questions.
- Do needle exchange programs encourage drug abuse?
- Will legalizing medical marijuana, as New York state is considering, increase illegal marijuana use as well?
- What impact did New York City’s controversial “stop and frisk” program have on drug abuse?
- Are large drug seizures publicized by law enforcement effective in reducing levels of abuse? If so, does the abuse of another drug increase in compensation?
More generally, sewer-based monitoring could be systematically employed to evaluate the efficacy of drug treatment and control strategies on a continual basis.
Public Health Surveillance
David Satcher, U.S. surgeon general from 1998 to 2002, said “[i]n public health, we can’t do anything without surveillance. That’s where public health begins” (Buehler, 2012). Sewage epidemiology seems like a natural extension for government programs designed to monitor the prevalence of infectious disease and other metrics of public health. Currently, such programs collect information from hospitals, electronic health records, even sales of cold medicines (Heffernan et al., 2004). A primary objective – as yet unobtained – is to monitor the prevalence of infectious disease in real time. This is challenging to do by tracking sales of cold and diarrheal meds, but it might be possible by monitoring for spikes in the occurrence of these and other products (especially antibiotic and antiviral agents) in sewer systems (Singer et al., 2013).
Another approach for monitoring outbreaks is to conduct sewer-based monitoring for infectious agents themselves. Techniques to monitor for the presence of specific classes of viruses and bacteria in sewage are well developed, but generally more resource and time intensive than techniques for monitoring chemical biomarkers. Recent efforts have been made to track the prevalence of adenoviruses (Bibby & Peccia, 2013) and parechoviruses (Lodder et al., 2013), which can cause diarrhea and flu-like symptoms, in European sewage. Unlike many chemical biomarkers, however, the abundance of such viruses may not necessarily correlate with disease prevalence (Bibby & Peccia, 2013).
Sewage epidemiology may also prove useful for monitoring general metrics of lifestyle and wellness at the community level. Trends in hormone levels and other endogenously produced biochemicals (such as isoprostanes) might provide indications of community-level stress (Daughton, 2011). Composite measurements of anti-inflammatory agents or antibiotics might be used to similar effect. Sewerbased monitoring can also be used to monitor for emerging drugs of abuse before they become problematic. Cathinone-derived “bath salts” (Gunderson et al., 2013) and synthetic cannabinoids (Huang, 2012) are just two classes of drugs whose rising popularity was detected too late for some.
Source Tracking Illegal Activity
A final, admittedly more speculative, application of sewer-based monitoring involves source tracking of illegal activity. Using a hydraulic model of the sewer system and knowledge of how the relevant chemical signature(s) behave in sewer environments, it may be possible to identify the location of drug processing and terrorist-related activities by the strategic placement of passive sampling devices.
Passive sampling techniques suitable for municipal sewers have only recently been developed (Birch et al., 2013). The devices require no external power source, and could be innocuously placed at sewer access points (i.e., man holes). They could later be removed, analyzed for the presence of chemical signature, and placed in new locations until the source of activity has been pinpointed. Success in such an effort would depend on the signal-to-noise ratio of the chemicals being monitored, but unlike most epidemiological applications, a simple yes/no for its presence may be sufficient. To my knowledge, source tracking of drug processing or terrorist activity has not been attempted. However, a model for the approach might be the New Jersey Department of Environmental Protection’s efforts to identify sewer-based sources of PCB pollution (Belton et al., 2005).
Sewage epidemiology holds great promise for providing near real-time, quantitative data on human activity at the community level. The idea is young, but sewer-derived estimates of drug abuse consistently agree with survey findings whenever such comparisons can be made (Castiglioni et al., 2013). To provide just one more example of agreement, my colleagues and I used the occurrence of cotinine in Baltimore wastewater to estimate the number of cigarettes consumed per smoker per day in the city. Our sewer-derived result is in good agreement with the most recent survey-based estimate, which, I might add, is from 2002 (Table 1).
It is important to note that there are currently no technological impediments to the routine analysis of chemical biomarkers in sewage. Online, automated procedures already exist (Daughton, 2011), making it a simple matter of cost. The approach is gaining momentum in the European Union and Australia. The EU has even devoted a special branch of its center for drug abuse (EMCDDA) to refining sewage epidemiology techniques. Interest in the United States, in contrast, has been quite limited.
My hope and suspicion is that the relative silence on our side of the Atlantic can be attributed mostly to a lack of awareness about this disgusting yet promising concept. Therefore, I’m ending this article with an invitation to start a dialog among law enforcement and public health officials in and around Hofstra University. Can sewer-based monitoring of health biomarkers help you with a research or professional question you have? What are the shortcomings, and how can they be fixed? You see, we environmental chemists have been developing this approach for a few years now, but the intended users have always been you. So, what do you want out of it?
Belton, T., Stevenson, E., Lippincott, L., et al. (2005). Trackdown of Polychlorinated Biphenyls (PCBs) in a Municipal Sewer System: Pilot Study at the Camden County Municipal Utility Authority (CCMUA). Proceedings of the Water Environment Federation 2005:212-229.
Bibby, K., & Peccia, J. (2013). Prevalence of respiratory adenovirus species B and C in sewage sludge. Environ Sci: Processes Impacts 15:336-338. doi:10.1039/C2EM30831B
Birch, H., Sharma, A. K., Vezzaro, L., et al. (2013). Velocity dependent passive sampling for monitoring of micropollutants in dynamic stormwater discharges. Environ Sci Technol 47:12958-12965. doi: 10.1021/es403129j
Bisceglia, K. J., & Lippa, K. A. (2013). Stability of cocaine and its metabolites in municipal wastewater – The case for using metabolite consolidation to monitor cocaine utilization. Environ Sci Pollut Res Advance online publication. doi 10.1007/s11356-013- 2403-5
Bisceglia, K. J., Roberts, A. L., & Lippa, K. A. (2012). A hydrolysis procedure for the analysis of total cocaine residues in wastewater. Anal Bioanal Chem 402: 1277-1287. doi: 10.1007/s00216-011- 5553-7
Bisceglia, K.J ., Roberts, A. L., Schantz, M. M., & Lippa, K. A. (2010a). Quantification of drugs of abuse in municipal wastewater via SPE and direct injection liquid chromatography mass spectrometry. Anal Bioanal Chem 398: 2701-2712. doi: 10.1007/s00216-010- 4191-9
Bisceglia, K. J., Yu, J. T., Coelhan, M., et al. (2010b). Trace determination of pharmaceuticals and other wastewaterderived micropollutants by solid phase extraction and gas chromatography/mass spectrometry. J Chromatogr A 1217:558- 564. doi: 10.1016/j.chroma.2009.11.062
Bowling, A. (2009). Research methods in health: Investigating health and health services. McGraw Hill/Open University Press, Maidenhead, Berkshire, England; New York, NY
Brewer, A. J., Ort, C., Banta-Green, C. J., et al. (2012). Normalized diurnal and between-day trends in illicit and legal drug loads that account for changes in population. Environ Sci Technol 46: 8305-8314. doi: 10.1021/es202447r
Buehler, J. W., Centers for Disease Control and Prevention. (2012). CDC’s vision for public health surveillance in the 21st century. Introduction. MMWR Surveill Summ 61 Suppl:1-2.
Burgard, D. A., Fuller, R., Becker, B., et al. (2013). Potential trends in Attention Deficit Hyperactivity Disorder (ADHD) drug use on a college campus: Wastewater analysis of amphetamine and ritalinic acid. Sci Tot Environ 450-451: 242-249. doi: 10.1016/j. scitotenv.2013.02.020
Castiglioni, S., Bijlsma, L., Covaci, A., et al. (2013). Evaluation of uncertainties associated with the determination of community drug use through the measurement of sewage drug biomarkers. Environ Sci Technol 7:1452-1460. doi: 10.1021/es302722f
Cutler, D., & Miller, G. (2005). The role of public health improvements in health advances: The twentieth-century United States. Demography 42:1-22. doi: 10.1353/dem.2005.0002
Daughton, C. G. (2011). Illicit drugs: Contaminants in the environment and utility in forensic epidemiology. In Whitacre, D. M. (Ed.), Reviews of Environmental Contamination and Toxicology, Volume 210. Springer, New York, pp. 59-110.
Gunderson, E. W., Kirkpatrick, M. G., Willing, L. M., & Holstege, C. P. (2013). Substituted cathinone products: A new trend in “bath salts” and other designer stimulant drug use. J Addict Med 7: 153-162. doi: 10.1097/ ADM.0b013e31829084b7
Heffernan, R., Mostashari, F., Das, D., et al. (2004). Syndromic surveillance in public health practice, New York City. Emerg Infect Dis 10:858-864. doi: 10.3201/eid1005.030646
Huang, L. (2012, Spring). “Spice” tales: Good chemists vs. evil. Hofstra Horizons, pp. 28-34.
Huerta-Saenz, L., Irigoyen, M., Benavides, J., & Mendoza, M. (2012). Tap or bottled water: Drinking preferences among urban minority children and adolescents. J Community Health 37:54-58. doi: 10.1007/s10900-011-9415-1
Lodder, W. J., Wuite, M., de Roda Husman, A. M., & Rutjes, S. A. (2013). Environmental surveillance of human parechoviruses in sewage in the Netherlands. Appl Environ Microbiol 79: 6423-6428. doi: 10.1128/AEM.01917-13
Magura, S. (2010). Validating self-reports of illegal drug use to evaluate National Drug Control Policy: A reanalysis and critique. J Eval Prog Plan 33:234-237. doi: 10.1016/j.evalprogplan.2009.08.004 Plósz, B. G., Reid, M. J., Borup, M., et al. (2013). Biotransformation kinetics and sorption of cocaine and its metabolites and the factors influencing their estimation in wastewater. Water Research 47:2129-2140. doi: 10.1016/j. watres.2012.12.034
Schwarzenbach, R. P., Escher, B. I., Fenner, K., et al. (2006). The challenge of micropollutants in aquatic systems. Science 313:1072-1077. doi: 10.1126/ science.1127291
Singer, A. C., Järhult, J. D., Grabic, R., et al. (2013). Compliance to oseltamivir among two populations in Oxfordshire, United Kingdom affected by influenza A(H1N1)pdm09, November 2009 – A waste water epidemiology study. PLoS One 8:e60221. doi: 10.1371/journal. pone.0060221
Substance Abuse and Mental Health Services Administration. (2013). Results from the 2012 National Survey on Drug Use and Health: Summary of National Findings, NSDUH Series H-46, HHS Publication No. (SMA) 13-4795. R.
Zuccato, E., Chiabrando, C., Castiglioni, S., et al. (2005). Cocaine in surface waters: A new evidence-based tool to monitor community drug abuse. Environmental Health 4:14. doi: 10.1186/1476-069X-4-14
Kevin Bisceglia is an assistant professor of chemistry at Hofstra University. Prior to joining Hofstra in fall 2013, Dr. Bisceglia founded the Chemistry Department at Bard High School Early College Queens, a satellite campus of Bard College that grants AA degrees to socioeconomically diverse New York City public school students; he also worked in the Chemical Sciences Division of the National Institute of Standards and Technology (NIST ). He holds a BS and ME in environmental engineering from Manhattan College and a PhD in environmental engineering and chemistry from Johns Hopkins University.
In addition to refining techniques for sewage epidemiology, Dr. Bisceglia’s research focuses on understanding how suburban lifestyle and land use practices influence water quality and chemical cycling on Long Island. His long-term goal is to find ways to modify these practices to increase the sustainability of suburban living. At Hofstra, he teaches courses in general chemistry, environmental chemistry, and instrumental analysis. In his spare time, he enjoys hiking and exploring Long Island’s coastal ecosystems with his wife and two young children.