CANCER ON TAP: THE RISKS OF CHLORINATED DRINKING WATER

By Paul Fleckenstein (pflecken@zoo.uvm.edu)

May 18, 2001

I. The History of Water Chlorination: From Harmless to Poisonous

II. Making New Toxins: How Chlorination Creates Organochlorines

III. The Search for Health Effects: The Art of Risk Assessment

IV. Assessing Your Water and Treatment Alternatives

I. The History of Water Chlorination: From Harmless to Poisonous

By the early 1920's, cities across the United States were for the first time providing municipal water customers with safe drinking water. This historic achievement was the result of water plant operators mixing in chlorine, a spectacularly powerful killing agent, which eliminated disease causing bacteria, viruses, and other microbes. Today over 200 million people in the U.S. use chlorinated tap water.

As with many novel chemical innovations, however, what was once thought perfectly harmless has turned out to be poisonous. Over the past 25 years, scientists have discovered that while chlorine is killing microbes, it is also reacting with organic matter already in the water to form toxic chemicals called organochlorines. To date, several hundred known organochlorines have been found in drinking water but many times that number are chemically unidentified.

Federal government monitoring programs show that many U.S. cities, including this writer's hometown, Burlington, Vermont, have elevated levels of organochlorines at the tap--concentrations substantially higher than most public water systems. This is especially disconcerting because numerous of these organochlorines (also known as disinfection by-products) are known carcinogens and mutagens. Epidemiological research has also directly linked chlorinated drinking water to cancer and possibly miscarriages.

The United States Environmental Protection Agency (EPA) recently enacted marginally stronger regulations to control disinfection by-products that are scheduled to take effect in 2002. There will be new standards for trihalomethanes and haloacetic acids, which are two groups of organochlorines that account for one-quarter to one-half the total organochlorine content of tap water. Regulators use these two easily tested for groups as indicators of total organochlorine concentration. How safe the standards are is unclear, but it is clear that the regulations reflect an accommodating view toward organochlorines.

Drinking water chlorination is a difficult regulatory area because killing waterborne pathogens has huge public health benefits to weigh against harms. But as with the thousands of other organochlorine chemicals released into the environment, from dioxins and dry cleaning solvents to pesticides, however, industry officials and regulators operate with an innocent until proven guilty prejudice. Even though virtually every organochlorine tested has been shown to have one or more toxic properties according to the American Public Health Association, regulation procedes to various degrees with the requirement of proven harms before substantial government action to reduce exposure. The problem is that harms from long-term and/or low level exposure to environmental toxins such as disinfection by-products are very difficult or impossible to conclusively find when they exist.

II. Making New Toxins: How Chlorination Creates Organochlorines

Like most drinking water in the U.S., Burlington's water comes from a surface water source, that is, a lake or river. Vermont's Lake Champlain, like many lakes and rivers, is rich in invisible organic matter produced by decaying leaves and algae. During disinfection chlorine randomly attaches to this organic matter to form thousands of new chemicals called organochlorines. The organochlorine typically found at the highest concentration is the carcinogen chloroform, which is a combination of chlorine and methane. Chloroform is one of several compounds known as trihalomethanes that is formed by adding chlorine to methane.

Chloroform illustrates the salient feature of chlorination or reacting chlorine with organic matter: the organochlorine is toxic. "Chlorination virtually always increases toxicity," explains molecular biologist Joe Thornton author of the recently published book Pandora's Poison on organochlorines in the environment. "Consider the trihalomethanes. Methane is not toxic or carcinogenic or mutagenic. The trihalomethanes are."

For most life on the planet, including humans, organochlorines are new to the environment. According to Thornton, some plants and animals, mostly algae and microorganisms, produce organochlorines. These natural organochlorines generally serve as chemical deterrents to predators and parasites. They virtually always occur in minute amounts and are always produced in tightly regulated processes to prevent toxic environmental effects. Only one naturally occurring organochlorine is known to generally circulate in the biosphere. Our current continuous exposure to thousands of organochlorines is radically unnatural. Free chlorine (as in chlorine gas or laundry bleach) and organochlorines have been produced in massively increasing amounts only over the last 100 years. Though water disinfection accounts for only a few percent of total global organochlorine production, the effect on human health is proportionately greater because exposure to chlorinated drinking water is large and continuous. It is piped right into our homes.

III. The Search for Health Effects: The Art of Risk Assessment

It hardly requires more than a few minutes of browsing EPA documents on disinfection by-products to discover a great shortage of information on health effects. This lack of information would be astonishing to anyone assuming standards for disinfection by-products were set at demonstrably safe levels. The EPA disinfection by-product rule acknowledges that risk assessment "relies on inherently difficult and preliminary empirical analysis." Animal studies are very incomplete. Not a single by-product chemical has been assessed for the range of possible effects, including cancer, reproductive toxicity, neurological damage, and immune system disruption. The animal data available for a handful of chemicals has uncertain relevance for humans. Lastly, the EPA notes that single chemical testing on animals is "insufficient" to characterize the risks from exposure to a mix of thousands of organochlorines in chlorinated drinking water. There is evidence that organochlorines can have synergistic effects in which a combination of chemicals is disproportionately more toxic in a mixture.

Most research attention goes to the identified compounds that occur at relatively high concentrations, like trihalomethanes or haloacetic acids, both of which are toxic in animals. Other hazardous organochlorines in drinking water may be unknown or completely unstudied. One recently discovered compound called MX, which is present at only a fraction (one-tenth of one percent) of the concentration of the trihalomethanes, is a potent carcinogen in rodents. Research suggests that MX may be responsible for up to one-half of the mutagenic effects of chlorinated drinking water.

A growing number of studies have linked chlorinated drinking water to cancer and reproductive harms in humans. The most respected cancer study is a compilation of 10 separate epidemiological studies on chlorinated drinking water and cancer known at the Morris study. It found disinfection by-products in chlorinated water to be responsible for 9% of all bladder cancers and 15% of rectal cancers in the U.S. This translates into 10,000 additional deaths per year for just these two organs, a figure the Morris researchers believe to be an under-estimate.

A 1998 California Department of Health Study found that pregnant women with high exposure to chlorinated drinking water nearly doubled their risk of miscarriage, from a rate of 9.5% to 16%. The at-risk group drank water with greater than 75 parts per billion trihalomethanes (which is in the range of Burlington's water). A part per billion is a grain of salt in a swimming pool. Several other studies have found linkages to miscarriages and also to neural tube birth defects. According to the EPA's Stig Regli at the Office of Ground Water and Drinking Water, the results of reproductive studies are less convincing than cancer studies. Results have been inconsistent and more studies are needed. There is currently an attempt underway to replicate the California study.

While the EPA denies that there is "conclusive" causation linking chlorinated drinking water and health effects, the point is not reassuring. Strict conclusiveness in which a particular chemical is proved to cause a certain illness is difficult or impossible to achieve. Organochlorines in the water are always present in a complex mixture, so no single chemical can be singled out as responsible for a disease. While the California study linked trihalomethane levels to increased miscarriages, this wouldn't implicate trihalomethanes specifically since the toxic effects may be from other disinfection by-products, or a combination of trihalomethanes and other chemicals. In addition, long-term health effects may take decades or generations to show up. Health effects could also be subtle (for instance, immune suppression, reduced fertility or neurological damage) that may be impossible to consistently clinically identify. Finally, diseases induced by disinfection by-products may arise from other causes as well, so unless the relative contribution of by-product chemicals is very large, it is impossible to detect with confidence. The effect of all the factors (genetic, diet, exposure to other pollutants) in addition to the one under study, such as chlorinated drinking water, tends to obscure causal relationships.

"Epidemiology is a very insensitive measure," according to David Ozanoff, Chairman of the Department of Environmental Health at the Boston University of Public Health. Ozanoff is a member of the EPA rule-making committee that produced the new disinfection by-product standards. The insensitivity of measures to assess health impacts means that there uncertainty about actual risks says Ozanoff and that there is "lots of reason to be concerned at levels around the current [2002] standard."

Joe Thornton sees the regulatory concern with conclusive evidence as misplaced: "To require conclusive epidemiological evidence before we can judge health to be at risk makes sense only if science can provide such evidence whenever damage is taking place." A lack of certainty may reflect inherent limits of assessing health effects as much as whether or not there are health effects.

Another member of the EPA rule-making committee, Erik Olson, of the Natural Resources Defense Council, acknowledges that there may be problems with the current regulatory strategy: "In thirty years we may slap our heads and say we can't believe we didn't do anything about this!" Olson argues that there is already adequate evidence for still lower standards based on a "pretty strong linkage to cancer and a growing indication that there is a connection to birth defects and miscarriages."

According to Thornton, the history of public health standards for lead provides a classic example of over-confident standard setting. In the 1920s toxicologists determined that 80 micrograms per deciliter of blood was a safe threshold for lead contamination. This standard was based on the levels in men with severe lead poisoning. As late as 1968 leading toxicologists stated that 80 micrograms was a safe level. But beginning in the 1970s, as more sophisticated research began to produce more data showing harms, the threshold of safety began to fall. It declined to 60 in the 1970s, to 25 in the 1980s, and to 10 in the 1990's. There is now research indicating that levels below 10 may impair cognitive development and that there may be no safe level of lead in our bodies at all.

Ozanoff argues that there are reasons to think that there is also no threshold of safety for disinfection by-products. He does think that that reducing levels will reduce harms, however, since there is likely a linear relationship between exposure and disease–that is, there are higher and lower risks caused by higher and lower exposure.

In addition, it is important to recognize that the EPA did not set the standard for disinfection by-products in drinking water based only on their health effects. The standard was set based on a balance of factors including health effects, but also including the economic cost to water utilities of reducing disinfection by-products with advanced treatment technologies (potentially large) and the benefits of killing pathogens with chlorine (undeniably substantial).

Olson believes that the decision on standards was overly weighted in favor of chlorination benefits. "It doesn't have to be a trade off between disinfection by-products and microbiological safety. It is no longer accurate to say that decreasing disinfection by-products increases microbiological health risks." He points to new treatment technologies, such as microfiltration and carbon filters to remove organic compounds before disinfection (thus leaving chlorine with less to react with to form organochlorines) and alternative disinfection technologies to kill pathogens like UV light and ozone. He expects that the costs of advanced treatment technologies will drop substantially and that water systems will become more sophisticated in various other operational strategies to reduce disinfection by-products.

IV. Assessing Your Water and Treatment Alternatives

According to the EPA, most exposure (80%) to disinfection by-products occurs through drinking the water (including its use with coffee, tea, juice concentrates, and soups). But people are also exposed through a variety of other activities since disinfection by-products enter the body through inhalation and skin absorption. Additional exposure occurs through bathing and showering, humidifiers, cooking, washing dishes, dishwashers, and swimming in chlorinated pools. Researchers have found greatly elevated levels of the disinfection by-product chloroform in swimmers' blood after leaving the pool.

An effective kitchen sink activated carbon filter, costing between $200 and $300, can remove organochlorines from drinking and cooking water, but other exposures remain. For this reason, it makes a lot of sense not to create organochlorines at the water treatment plant to begin with.

Disinfection by-products in chlorinated water coming from nutrient-rich surface water sources such as Lake Champlain are often relatively high since the source water contains a large amount of organic matter. Over the past three years, annual average disinfection by-products levels in Burlington, for instance, were as high as 70 parts per billion (ppb) trihalomethanes and 61 ppb haloacetic acids. The EPA's new standards for trihalomethanes and haloacetic acids are 80 ppb and 60 ppb, respectively.]

For comparison EPA figures show that most public water systems in the U.S. have lower levels. Surface water systems have median concentrations of 40 ppb for trihalomethanes and 25 ppb for haloacetic acids. For ground water systems, the median values are 10 ppb for trihalomethanes and 5 ppb for haloacetic acids. People drinking non-chlorinated water from a household well or spring would have no trihalomethanes or haloacetic acids in their water.

To obtain the concentration of organochlorines or disinfection by-products in your city's water, contact your local water utility. Test results for many larger cities are available from the EPA at http://www.epa.gov/enviro/html/icr/icr_query.html.

One of Burlington's water utilities, the Champlain Water District (CWD), provides some examples of steps that can be taken to reduce disinfection by-product levels. CWD has implemented non-chlorine control methods for zebra mussels on intake pipes and reduced the time treated water (continually forming disinfection by-products while waiting to be used) sits in storage tanks before going to the tap. Along with better management of chlorine disinfection at the plant, these steps have steadily reduced by-product levels (both averages and peaks) over the past three years.

One aspect of reducing by-product levels is also reducing seasonal variation. Warm weather levels may be as much as 50% higher. Some possible health risks identified by the EPA such as miscarriages and birth defects may be influenced by seasonal variation occurring during key periods of fetal development during pregnancy.

CWD has also researched switching to ozone for disinfection and is embarking on the development of a 10-year master plan that may include other alternative treatment options such as UV disinfection. CWD treatment practices to date, however, do not include treatment technologies such as microfiltration and carbon filters (for removing organic matter), or disinfectants like UV and ozone that may deliver large disinfection by-product reductions. Twenty large water systems in the U.S. use ozone for disinfection instead of chlorine, but UV and advanced filtration technologies are not commonly used by public water systems.

Many public health officials and water system operators in the U.S. are opposed to no chlorine at all, especially in the distribution system where there is risk of accidental re-contamination. But if small amounts of chlorine were added just to prevent pathogen growth in distribution systems following improved filtration and non-chlorine disinfection, this could contribute to less use of chlorine and a large reduction of organochlorines in our water.


Back to Top