ASBESTOS IN DRINKING WATER

Asbestos is a remarkable material. It is a mineral composed of elements much like those of sand, but it comes in the form of long, thin, submicroscopic fibers. Materials made from asbestos are light, fireproof, flexible, and tough. During the decades of the 50’s to 70’s, when it was referred to as “the magic mineral” (Roggli and Coin, 2004), it was used in building materials, roofing tiles, pipe insulation, brake liners, fireproof clothing and mitts, and sometimes appeared in the kitchen as a heatproof stovepad. It was also used in asbestos/concrete pipe for water transport. In this case, it was valued not for its fireproofing ability, but because the fibers helped bind the concrete. At 25% asbestos, 75% concrete (OSHRC Docket No. 01-0261), asbestos made the pipes lighter and easier to install. This pipe was intended for such purposes as irrigation, but because it was plentiful and cheap, asbestos/concrete or AC pipe, also known by its trade name, Transite, was used to transport drinking water in municipal water systems across the United States and Canada. There are an estimated 200,000 miles or more of asbestos/concrete pipes in US water systems today (Olson, 1974, cited in Millette et al., 1980).

In the 1970’s, another property of asbestos became too obvious to ignore. Exposure to asbestos can cause death.

Workers exposed to asbestos in their jobs became ill and died of particular diseases. Over time, it was discovered that exposure to airborne asbestos fibers could lead, decades later, to asbestosis, lung cancer, and mesothelioma, a cancer of the lining of the body cavities. Mesothelioma is a hallmark of asbestos exposure; it is a distinctive and, even now, extremely rare form of cancer. About 15-20 cases occur per million men each year in North America, and far fewer women develop the disease. Almost every person who has ever developed mesothelioma has had occupational exposure to asbestos at least 15 years prior, and sometimes as long as 70 years before. The peak in incidence occurs 30 to 40 years after exposure (Sporn and Roggli, 2004).

.

Other kinds of cancer also showed up decades after asbestos exposure. By the early 80’s, five studies had shown that cancers of the stomach and other parts of the gastrointestinal (GI) system occurred at higher rates in asbestos workers. The risk ranged from somewhat higher than non-exposed control workers (1.32 in London factory workers) to almost 3 times the usual risk in New York and New Jersey insulation workers (Safe Drinking Water Committee, 1983). It became clear that at least some GI cancers are likely related to asbestos exposure (Rolston and Oury, 2004).

What was not clear was whether inhaled asbestos fibers had migrated through the body from the lungs to cause these cancers, or whether fibers from the lungs, coughed up and then swallowed, might have entered the stomach and other digestive organs more directly, from ingestion.

This question was important then, and is still important now, because, as noted above, hundreds of thousands of miles of asbestos/concrete pipe are used to transport the drinking water for many Americans. This pipe is known to shed asbestos fibers (Safe Drinking Water Committee, 1983). Even at relatively low exposures, if swallowed asbestos can cause cancer, exposing millions of people to asbestos in their drinking water could be considered dangerous (CCERP, 1987).

This report reviews the scientific literature relating to the possible carcinogenicity (cancer-causing ability) of asbestos taken by mouth, as in drinking water. Because reviews of this literature in the 1980s had a substantial impact on public policy regarding asbestos in drinking water, I summarize that literature first. I then review literature to 2005 and discuss the impact of the more recent findings on our current knowledge concerning the carcinogenicity of asbestos in drinking water.



STUDIES ON INGESTED ASBESTOS AND CANCER THROUGH THE ‘80s



A number of scientific studies done in the 1970’s and 1980’s addressed the question of whether swallowed asbestos could cause cancer. Since it is unethical to deliberately expose people to suspected cancer-causing agents unless it is considered beneficial to them, the studies used animal models and “natural experiments” in which humans had been exposed to asbestos in their drinking water. The many approaches to this question in animal models included direct injection or surgical implantation of asbestos into the body followed by examination for tumors, testing of whether asbestos given by mouth could penetrate body tissues, and administration of asbestos in food or water followed by examination for tumors.



DIRECT INJECTION OF ASBESTOS: EFFECTS

Several studies established that direct injection or surgical implantation of asbestos into the body cavities of rats and hamsters produced tumors. Forty-one rats injected with asbestos into the lung cavity, for example, developed 5 mesotheliomas in the abdomen in addition to 20 mesotheliomas of the lining of the lung (pleural) cavity (LaFuma et al., 1980). Injection of asbestos into the peritoneal (abdominal) cavity of 20 European hamsters produced 15 abdominal tumors, while 3 were seen in the control group. Rabbits were also shown to develop mesothelioma 5 and 6 years after peritoneal injection of asbestos fibers (Pott et al., 1980). Monkeys developed asbestosis and pretumorous lesions from asbestos injection into the trachea and pleura (Pylev and Shabad, 1973; cited in Pott, 1980).



PENETRATION OF ASBESTOS FIBERS INTO THE BODY FROM THE GUT

“During the 1970s, potential health risks associated with exposure to asbestos in drinking water became a national concern. This concern developed primarily because of reports of some drinking water fiber concentrations in the range of 10,000,000-10,000,000,000 fibers/liter (Millette et al., 1979) and published observations of increased risk of gastrointestinal and peritoneal cancer among asbestos workers (Seilikoff, 1979).” This statement introduced a review paper (1983) by Philip Cook, an employee of the US Environmental Protection Agency (EPA). He continued: “One of the key questions that arose from debate over whether ingestion of mineral fibers could result in increased gastrointestinal cancer risk was whether fibers can penetrate the gastrointestinal mucosa and thus have some chance of residing in tissue. It is likely that such movement of a large number of fibers is a necessary precursor for carcinogenesis following ingestion of asbestos.”

Cook reviewed 19 reports in which tissues from animal and human subjects exposed to asbestos taken by mouth had been examined. These reports, dating from 1965 to 1982, varied a great deal. The subjects were rats, baboons, fish, and people, and several different types of asbestos were given in water, food, or by stomach tube for periods ranging from a single dose to lifetime exposure. Colon, blood, brain, lung, urine, kidney, lymph fluid, muscle, spleen, mesentery, jejunum, and esophagus were examined, mostly using transmission or scanning electron microscopy on concentrated preparations.

The results were mixed. Some studies found no indication of asbestos fibers in the tissue examined, while others clearly did. Cook observed that those studies that did report the presence of fibers were those best able to detect them, and were also least likely to have been affected by contamination. In his conclusion, Cook stated, in the cautious language of scientists, “It is difficult to conclude on the basis of the studies … that asbestos fibers do not cross the intestinal barrier.” He also estimated, based on the little evidence available, that in humans, about one fiber of every thousand in drinking water passes into the urine (Cook and Olson, 1979).



ASBESTOS GIVEN BY MOUTH: EFFECTS

Cook’s review was published in 1983, in Volume 53 of Environmental Health Perspectives. The same volume also contained a review covering studies of the cancer-causing potential of orally administered asbestos. Lyman Condie, also with the EPA, reviewed eleven studies published between 1967 and 1981. All but one used rats or hamsters. The single nonhuman primate study lasted only 5 years, a latency period Condie considered too short to allow cancer to occur if baboons reacted similarly to humans (Webster, 1974). Of the remaining 10 reports, Condie noted, “Many of the studies suffered from an insufficient number of experimental animals and from an inadequate exposure time to asbestos. Another major drawback of many of the studies was that they were not lifetime studies.”

Of the studies of 50 or more animals exposed to asbestos for 5 months or longer and observed at for least 20 months, two used rats (Gibel et al., 1976; Donham et al., 1980) and one used hamsters (Smith et al., 1980). Gibel’s group fed rats 20 milligrams daily of asbestos in the form of filters mixed with their food. The 42 rats examined developed 12 malignant tumors against only 2 in controls, a significant difference. Tumors in the experimental group occurred in the lung, kidney, nodes, and liver, and the authors noted in their summary “The localization of the tumors varied and so conclusions about organ specificity can not be drawn, when asbestos is [applied] orally.” Condie’s review was rather dismissive, saying “The authors stated that no conclusions could be made from their test results regarding the pathogenesis of the tumors caused by the oral intake of asbestos material.” This issue of tumor localization with respect to the issue of how ingested asbestos might cause cancer is discussed later.

A study of 180 hamsters given 1300 million fibers/liter water throughout their lifetime produced one peritoneal mesothelioma and two carcinomas of the forestomach as well as a lung carcinoma in the group, but the authors regarded their results as inconclusive since another 180 hamsters given 10 times that dose did not develop tumors (Smith et al., 1980). Since mesothelioma is a rare cancer generally associated with asbestos exposure, however, this does give grounds for suspecting carcinogenesis from asbestos exposure.

In the study done by Donham’s group, the bowel and rectum of 189 rats fed 10% asbestos in their diet and followed to 32 months were examined microscopically (other organs were not examined). Although the number of malignant colon tumors was not significantly different in the asbestos and control groups, when other lesions of the colon and rectum were considered, including ulcers and inflammation, the rats fed asbestos had significantly more lesions. One of the asbestos-fed rats developed mesothelioma. In light of their results, which included evidence for gut penetration by asbestos fibers and altered cell regulatory mechanisms. these authors stated “…we feel there is suggestive evidence that ingested asbestos may have some role in colon carcinogenesis….”

One of the observations made by Donham and his co-workers has great relevance to the suitability of rats as an animal model, though it was not mentioned in the review by Condie and associates. The Donham group chose the F344 strain of laboratory rats for its longevity; this strain lives an average of 30 months, and they carried their study out to 32 months. “Significantly,” they observed, “the bowel tumors were found only in very old animals (25-30 months). Investigators should consider this fact when designing future studies. Delaying until at least two years before killing any animals may enhance the detection of neoplastic or preneoplastic bowel lesions related to asbestos.”



CONCLUSIONS DRAWN FROM ANIMAL EXPERIMENTS TO 1983

From the experiments published through 1983, it was clear that direct application of asbestos into the body would cause cancer in animals. It also seemed likely that asbestos fibers penetrate the wall of the gastrointestinal tract and migrate into tissues, given that the best studies showed positive results, though some negative studies had also been published. Logically, it might be expected that if asbestos can pass into the body from the gut, and if the fibers cause cancer when in the body, animal models given asbestos by mouth should develop cancer at a greater rate than untreated controls. Some studies did show this, while others gave negative results.

Although Condie, who reviewed the oral carcinogenesis studies, pointed out serious problems with many of these studies he also noted, “One can question the suitability of the animal models employed in evaluating the human response to oral exposure to asbestos, since sufficient time may be lacking between exposure and the development of malignancies during the animal’s lifetime.” Since asbestos exposure by inhalation had induced cancer in rats, however, he concluded “one would expect to be able to produce a neoplastic [cancerous] response within the lifetime of conventional laboratory animals with massive doses of ingested asbestos such as those employed in some of the studies mentioned in this paper.” Condie left the clear implication that if ingested asbestos was dangerous to humans, this should have been demonstrable in the animal studies. Some alternative deductions are considered in the second part of this report.



HUMAN STUDIES ON THE EFFECTS OF ORAL ASBESTOS

An ethical way to study human reactions to a substance and route of exposure suspected to cause cancer is to study people who have been inadvertently exposed. In the case of water-borne asbestos, epidemiological studies were done comparing cancer rates in communities with substantial amounts of asbestos in the water supply to other communities believed to have less exposure. As of 1987, cancer death rates had been examined in seven areas of the United States and Canada. The areas examined were Duluth, Minnesota, Connecticut, Quebec, the California Bay Area including San Francisco, Utah, Puget Sound, especially Everett, Washington, and Escambia County, Florida. These epidemiological studies were reviewed by a Working Group for the Department of Health and Human Services (DHHS) Committee to Coordinate Environmental and Related Programs (CCERP), Subcommittee on Risk Assessment (Richard Lemen, Chair) (CCERP 1987).

The working group was formed “to assist policy makers in the DHHS determine if adequate information was available for a definitive risk assessment on [potential risk of cancer associated with ingestion of asbestos] … and evaluate if the weight of evidence was sufficient to prioritize this issue for new policy recommendations.”

The kind of epidemiological studies they reviewed, where an attempt is made to correlate exposure to a suspected carcinogen with cancer rates in a geographic area, have numerous problems. People move in and out of the area, carrying their exposure histories with them. Their individual exposures may be increased by occupational exposure or decreased, for example, if they use bottled water. The exposure of the population in general may not be well characterized. Factors such as occupation, race, ethnicity, socioeconomic status, and personal habits such as smoking may not be controlled. The size of the population studied may be too small to show significant differences in cancer rates unless there are extraordinary numbers of excess cancers. For cancer studies such as this one, latency is particularly important; the time from exposure to the suspected carcinogen to the time cancers are counted must exceed the time it takes for cancers to develop.



Each of the studies considered by the working group was flawed, some more than others. In the first two Duluth studies, the maximum exposure time was 15 to 20 years, possibly too short for a disease with an estimated latency of 30 to 40 years to be clearly demonstrable (Masson et al., 1974; Levy et al., 1976).

In the Connecticut studies, the latency was somewhat longer, but the exposure to asbestos may have been low.

The two Quebec studies did have a maximum duration of exposure of more than 50 years, and one study demonstrated more-than-expected (excess) numbers of cancers. These were cancers of the stomach and lung in men, and pancreas in women (Wigle, 1977). However, under the assumption that both sexes should be equally susceptible to cancer at all three sites, the author concluded that a link between cancer and asbestos in drinking water was not supported.

The population studied in the Bay Area had a maximum exposure time of more than 40 years. Significant relations were found between asbestos content in drinking water and lung cancer in men, gall bladder, pancreatic, esophagus, pleura, and kidney cancer in women, and peritoneal and stomach cancer in both sexes (Kanarek et al., 1980). A follow-up study controlling for several potentially confounding factors supported the original findings, reporting a relation between asbestos dose and cancers of the colon and prostate in men, peritoneum and pleura in women, and total cancers and cancer of the esophagus, stomach, and pancreas for both men and women (Conforti et al., 1981). A third San Francisco study suggested that population density might be an important confounding factor. Since 1974, the California Department of Health Services has recommended filters to reduce exposure in any water supply system using asbestos-laden water sources.

The Utah study suffered from an insufficient latent period, and the asbestos concentration in the drinking water was not known.

In Puget Sound, the duration of asbestos exposure was more than 40 years. One of the three studies found non-significant increases in prostate and brain cancer and leukemia in men, and cancers of the small intestine and thyroid in men and women exposed to water-borne asbestos. A case-control study of people who had been diagnosed with cancers of the GI, respiratory, and urinary systems, compared to people living in the same area who had not been diagnosed was done in Everett (Polissar et al., 1983). A significant increase in the odds of developing stomach and pharyngeal cancer in men, but not women, was found, and the authors concluded that their evidence for a link between cancer and asbestos in water was not convincing. Asbestos exposure was individually estimated based on a questionnaire asking about water consumption “5 years ago.” Considering the typical latency of asbestos-related cancer, these exposure estimates may have been inadequate.

The Escambia County, Florida study found no association between asbestos dose and cancer deaths, but the authors noted that a large number of cancer deaths would have been required to reach statistical significance.



“The collective conclusions of available reviews and our panel,” the DHHS committee (CCERP, 1987) wrote, “indicate that the information is inadequate for a credible risk assessment of cancer risks associated with the ingestion of asbestos based on information developed from studies of asbestos ingestion…. Epidemiologic studies have not definitively established the existence of an association between asbestos ingestion and gastrointestinal cancer or other cancers.” They concluded that if there is such an association, “it is probably weak compared to background cancer rates” and that “epidemiologic research methodology is limited in its ability to detect small increases in risk. If additional epidemiological research is undertaken, then it should be recognized that studies will require a substantial amount of time for completion, will be very expensive, and will probably provide only limited information about small increases in risk, possibly providing an upper estimate of risk.”

“Given the available data,” they concluded, “we do not believe that from a qualitative cancer risk assessment perspective the cancer risks associated with the ingestion of asbestos are among the most pressing environmental health hazards in the United States. Nonetheless, this should not be taken to mean that the potential hazard associated with ingested asbestos is an unimportant issue which does not warrant further research. Even if the increased rate of cancer is less than 10% of the background rate and cannot be demonstrated by available research tools, the ingestion of water, food, or drugs laden with asbestos by millions of people over their lifetimes could result in a substantial number of cancers.”

They ended their report saying, “Several of the members of this working group believe it is prudent, preventive public health policy to recommend eliminating possible sources of ingestion exposure to asbestos whenever and to whatever extent possible. This should not be interpreted as a recommendation of the Department of Health and Human Services or its member agencies. Some of the approaches which could be pursued include the following: eliminating asbestos cement pipe in water supply systems, eliminating the use of asbestos filters in the processing of beverages, foods, and medications, and reducing the levels of asbestos fibers in drinking water supplies. It should be noted that on January 29, 1986, EPA proposed prohibiting the manufacture of asbestos cement pipe; therefore, future abatement efforts of this potential exposure source may only involve replacement of existing pipe.”







STUDIES ON INGESTED ASBESTOS AND CANCER FROM THE 1980s - 2005



As matters stood at the time of several reviews written in the 1980s, the cancer-causing potential of ingested asbestos was anything but clear. Although deposition of large quantities of asbestos directly into the bodies of experimental animals clearly caused cancer, the evidence that asbestos fibers could penetrate the gut wall of experimental animals, or that asbestos in their food or water could cause cancer was equivocal. Epidemiological studies of people whose drinking water contained asbestos did not consistently demonstrate a carcinogenic effect, but they did not establish that drinking asbestos-containing water was safe, either. Further study was advised, although some reports noted that additional epidemiological studies of the type conducted thus far would be unlikely to add to then-current understanding (CCERP, 1987).



PENETRATION OF ASBESTOS FIBERS INTO THE BODY FROM THE GUT



One research group, whose negative results with asbestos fiber penetration had been cited in Cook’s review, redid their analysis (Kaczenski and Hallenbeck, 1984). Using larger amounts of tissue and a more sensitive assay for asbestos, they found that asbestos fibers could be clearly demonstrated in the tissues (stomach, heart, spleen, pancreas, blood) of a baboon which had been given asbestos by mouth, and that those fibers must have penetrated the gut wall to get there. With this reversal of one of the previously negative findings, the preponderance of the evidence appeared to show that asbestos can travel through the gut wall and into the body from exposure via the gastrointestinal tract.

A recent study (Haque et al, 2001) has also demonstrated that pups born of pregnant rats fed asbestos through a feeding tube show asbestos fibers in their bodies.



ASBESTOS GIVEN BY MOUTH: EFFECTS



In light of the known ability of asbestos to cause cancer when placed in the body, and its ability to penetrate the gut, the preponderance of negative results in animals given oral asbestos seemed difficult to explain. Donham and his coworkers (1980) had noted that in their experience, precancerous changes appeared primarily in the oldest of the rats they studied. They recommended waiting at least 24 months to look for tumors caused by asbestos ingestion early in life.

Four subsequent studies followed that advice. Two examined the effect of ingesting amosite asbestos, one in hamsters (National Toxicology Program, 1983), the other in rats (National Toxicology Program, 1990), both with essentially negative results.

A study of asbestos given in palm oil to rats (Truhaut and Chouroulinkov, 1989) showed no toxic effects and no polyps of the large intestine after 24 months. The authors noted that a similar study published in 1985 (National Toxicology Program, 1985) had shown a significant increase in polyps, and suggested that the palm oil used as a vehicle for the asbestos may have been protective. In this connection, they noted that Bolton et al. (1982), using a diet to which margarine was added, also failed to observe such polyps.

The study of the effect of chrysotile asbestos in rats (National Toxicology Program, 1985) found that ingestion of chrysotile fibers of intermediate length was associated with polyps of the large intestine, small intestine, and stomach of the males. The authors interpreted their findings as “some evidence of carcinogenicity.”



RECENT REPORTS ON INGESTED ASBESTOS IN HUMANS

In the summary report of the DHHS working group (CCERP, 1987), the epidemiological studies reported through 1987 were regarded as inconclusive. Most epidemiological studies are very blunt experimental instruments, and it is difficult to use them to establish either safety or danger from a particular agent. The advice given by the DHHS CCERP committee to consider carefully the information to be gained from large, ecologic, epidemiological studies before embarking on such expensive undertakings seems to have been followed. Only two such studies appear to have been published since 1987 (Howe et al., 1989; Browne et al., 2005).

A second summary report (Cantor, 1997) again reviewed studies done to that point and concluded, “Overall, the epidemiologic database is not adequate to evaluate the cancer risk posed by asbestos in drinking water. There are suggestions in the cited studies of elevated risks for gastric, kidney, and pancreas cancers, which warrant further investigation.”

The study published in 2005 was of cancer incidence in Woodstock, New York between 1980 and 1998 (Browne et al., 2005). Asbestos contamination in the town’s water, as much as 304.5 million fibers per liter from asbestos/concrete pipes installed in the water system in the 1950s, was discovered in 1985. The response of the New York State Department of Health was to establish “the Woodstock Asbestos Exposure Registry [WAER] to monitor rates of cancer among individuals who lived on the water supply between 1960 and 1985” (Browne et al., 2005). Cancer rates in 2936 registrants were observed from 1980-1998, with a follow-up time of 5 to 20+ years. For various reasons, including moves from the area, the data of 1852 people was actually analyzed. Their cancer rates were compared to that of New York State excluding New York City. Only pancreatic cancer showed a statistically significant increase in Woodstock, with 2.19 times the incidence of the rest of the state (excluding NYC), based on 9 observed cases. This increase was greater in men (3.08 times the expected incidence) than women (1.39 times the usual risk). Esophageal cancer in men had a relative risk of 2.91, but with such small numbers, the only statistically significant difference was in pancreatic cancer, and the authors concluded, “The general pattern of results did not demonstrate a likely link between exposure to asbestos in drinking water and cancer occurrence among participants in the WAER.”

This study, though somewhat better than most others in that only residents who were actually on the water system were included, suffered from small numbers, making statistical significance difficult to establish, and a minimally adequate latency period. The authors also assumed that the population of New York State excluding New York City (which had a level of asbestos below detectable limits; Millette et al., 1980) is relatively free of oral asbestos exposure, although several other cities in New York had asbestos levels ranging from 500,000 to nearly 1,000,000 fibers /liter (Millette et al, 1980).



Some kinds of epidemiological studies can produce more definitive results, however. A population with known heavy exposure to the substance in question can be followed in time, and their cancer incidence can be compared to that of a similar, not so exposed population. After World War II in Norway, lighthouse keepers became the unwitting guinea pigs for eventual studies on the effects of ingested asbestos.

In the 1950s, many of the 133 lighthouses along the Norwegian coast had their roof tiles replaced with asbestos-cement plates composed of 15% asbestos. Two decades later, in the late 70s, the roof tiles were found to be heavily worn. Water collected from the roofs and into the cisterns, which provided nearly all of the household needs for the lighthouse keepers and their families, contained more than 500,000,000 fibers per liter in the seven lighthouses tested (Andersen et al., 1993). This was ten thousand times more than the exposure of any other known population in Norway (50,000 fibers/liter or fewer), although some US municipalities had water supplies with a similar asbestos burden (Millette et al., 1980).

In the first of two studies, the 690 lighthouse keepers who began work between 1920 and 1966 were followed, and the cancers they experienced from 1960 to 1991 were compared to the cancers expected in a similar population of rural Norwegians (Andersen et al., 1993). More than expected numbers of stomach and colon cancer were observed, but this excess was not statistically significant in the group as a whole. Those who had 20 years or more of exposure, however, did have significantly more stomach cancer, 2.4 times as much. The authors cautiously noted that diet, a potential source of excess stomach cancer, was not controlled, the population size was small, and the time when eroding roof tiles began to release larger numbers of asbestos fibers into the drinking water could only be estimated.



In 2004, a medical case report was published of a 76-year-old woman from the Netherlands who had multiple tumors on the peritoneum, confirmed to be malignant mesothelioma. She had no occupational exposure to asbestos, but upon questioning, she told her doctors that for many years, she used to prepare vegetables for cooking in rainwater collected from a roof made of asbestos (van Kesteren et al., 2004A).

Summarizing the literature on asbestos effects in animal and epidemiologic studies to that point, the authors of this case report stated, “The overwhelming evidence that has been found in studies on the association of inhaling asbestos and pleural and peritoneal mesotheliomas has not been observed in those concerning asbestos ingestion. However, the observation that asbestos fibers can penetrate the gastro-intestinal tract and subsequently transmigrate to other tissues (CCERP, 1987) in combination with the recent finding of asbestos fibers in omentum and mesentery in subjects with peritoneal mesothelioma (Dodson et al, 2000) [raises] concern about the malignant potential of ingested asbestos fibers. We think it is therefore important to pay attention to the history of the ingestion of asbestos fibers of this woman with malignant peritoneal mesothelioma. Whether the ingestion of asbestos fibers caused her disease is speculative. The concentration of the asbestos fibers in the water this patient used and subsequently the fiber contamination in her food is unknown as well as the size of the fibers, which is an important carcinogenic determinant (CCERP, 1987). However, we consider the asbestos contamination as very likely regarding the report of a high fiber contamination (500 million fibers/l) found in cisterns supplied with runoff water from asbestos shingle roofs (CCERP, 1987).”

A few months later, a letter in response to this report appeared in the same journal. It was titled: “Asbestos in drinking water does not cause mesothelioma” (Lange and Hoskins, 2004). Based on their interpretation of the literature, they said “…it appears that that [sic] ingestion of asbestos in water, even at high levels, does not result in mesothelioma and other cancers. Reports suggesting a relationship of asbestos in water and cancer, particularly mesothelioma, should be considered spurious.”

“We believe it was important to pay attention to the history of the ingestion of asbestos fibers in our patient,” replied the authors of the case report (van Kesteren et al., 2004B), “because:



1. Malignant peritoneal mesothelioma is clearly related to asbestos exposure (DHHS committee [CCERP], 1987; Antman et al., 2001).

2. Asbestos fibers can penetrate the gastrointestinal tract and subsequently transmigrate to other tissues (DHHS Committee [CCERP], 1987).

3. Asbestos fibers are found in omentum and mesentary in subjects with peritoneal mesothelioma (Dodson et al., 2000).”



They concluded, saying, “We think that one should not ignore the potential risk factors of a disease whose pathogenesis is far from revealed. Therefore, we feel that it is important to question a patient with malignant peritoneal mesothelioma for any kind of asbestos exposure in the past, including asbestos ingestion.”



2005 saw the publication of an update on the Norwegian lighthouse keepers (Kjaerheim et al., 2005). This time, those first employed between 1917 and 1967, and born after 1884 and alive in 1960, were followed up for cancer incidence from 1960 to 2002. These 726 men were subdivided according to exposure; 107 of them had worked at one of the 21 lighthouses that received their water from asbestos-concrete roofs, while 479 had been mostly unexposed and the exposure of 140 was unknown. Because lighthouse reconstruction (and installation of asbestos roofing) took place between 1945 and 1950, 1948 was chosen as the beginning of exposure.

The definitely exposed group had 2.5 times as much stomach cancer as expected, and more cancer of the kidney and small intestine than expected as well. Small numbers (e.g. 6 stomach cancers in the exposed group) prevented the cancer incidence in the subdivided groups from reaching statistical significance, but taken as a whole, the lighthouse keepers showed a significant elevation of stomach cancer (1.6 times that expected) and gastrointestinal cancer generally (1.4 times that expected). Those exposed 20 years or more had elevated rates of stomach, colon, and gastrointestinal cancer. The stomach cancer rates after 20 years of exposure, at 1.7 times the number expected, showed definite statistical significance. (The expected percentage of men who develop gastrointestinal cancer at some time in their lives in rural Norway is 5.9% (calculated from Kjaerheim et al., 2005), while the observed percent was 8.5%, giving an excess incidence of 2.6%.)

The possible role of diet in explaining the excess stomach cancers was considered and found unlikely. The authors concluded, “results support the hypothesis of an association between ingested asbestos and gastrointestinal cancer risk, and stomach cancer specifically.”

Their final statement, “… the limited excess risk found in this heavily exposed group does not indicate that asbestos in drinking water is a major health problem in the general population” is doubtless intended to be reassuring. Perhaps it is, for Norwegians.



About 10% of 406 US cities tested had drinking water with asbestos fiber concentrations above 10 million fibers per liter (Millette et al., 1980), sometimes with over 600 million fibers per liter (Cook et al., 1976, cited in Millette et al., 1980), comparable to the exposure reported for the lighthouse keepers (maximum 500 million fibers/ liter). One tenth of the US population is about 29,780,000 people. If they experience the 2.6% excess gastrointestinal cancers observed in men with similar heavy exposure in their lifetime, nearly 800,000 Americans develop gastrointestinal cancer from asbestos in their drinking water. Presumably, lower doses result in a lower cancer incidence, but roughly 36% of Americans, 100,000,000 of them, drink water containing 500,000 to 10,000,000 asbestos fibers per liter (Millette et al., 1980). Even a low excess cancer incidence would affect many, many people.



Heavy asbestos contamination of household water apparently also leads to increased airborne asbestos (Webber et al., 1988). This is another factor (inhaled asbestos is a well known carcinogen) that has not been seriously considered in the debate surrounding the question of whether asbestos in drinking water can cause cancer.



CONCLUSIONS

As information about the effects of ingested asbestos accumulates, it is becoming clear that although the latency period is long and the effect size is possibly small relative to that of inhaled asbestos, asbestos taken by mouth does cause cancer in humans.

In terms of responsible action to ensure the safety of water users, this means that the 200,000+ miles of asbestos/concrete water pipe in the United States (Millette et al., 1980) will need to be replaced with a safe substitute as quickly as possible. Communities whose water naturally contains asbestos will need to find improved ways to reduce the potential danger. Users of water catchment systems must be made aware of the danger of

using asbestos-containing components in the system, and asbestos/concrete pipes should not be sold, here or abroad, for the purpose of transporting drinking water.


It is interesting to look back and consider what would have had to have been different to avoid the present situation, in which hundreds of millions of people are being exposed to a carcinogen in the water they use for drinking and cooking. A requirement for proof of safety before a new product is put into general use, similar to the present system for testing pharmaceuticals, might have prevented this situation. (See, for example, http://www.asmalldoseof.org/ for a presentation of the precautionary principle.)

There were early indications that asbestos in the body caused cancer, and that fibers could penetrate the gut. The exclusive reliance on rodent animal models of carcinogenesis was evidently unwise, though there were definite warning signs in the finding of precancerous changes in rats with a lifetime exposure to oral asbestos and the occasional findings of mesothelioma, always in experimental rodents, never in controls (Smith et al., 1980; Donham et al., 1980; Hilding et al., 1981). It may be that rats simply do not live long enough to develop full-blown cancer from their exposure, or other species differences may explain why humans get gastrointestinal cancer from oral asbestos when rodents do not do so to a significant extent.

Considering the thalidomide disaster, this will not be the first instance in which limited use of animal models for safety assessment has led to misplaced confidence in product safety. Thalidomide, the drug given to pregnant women in Europe for morning sickness in the late 50s and early 60s, was tested in rats and found to have no adverse effects. An application to sell the drug in the US was submitted, but later withdrawn, when babies born to women taking the drug were stillborn or had severe birth defects- missing or malformed arms and legs. Decades later, it was learned that all nonhuman primate species tested except the distantly related prosimian Galago crassicaudatus are susceptible to birth defects from thalidomide (Hendrickx, 1973; Hendrickx and Newman, 1973; Hendrickx and Sawyer, 1978). Primate models are generally much closer to humans in their responses to carcinogens and toxicants, but the only report on nonhuman primates used to test oral asbestos examined baboons after 5 years or less (Webster, 1974), a latency period likely too short for asbestos-caused cancer to develop in primates.

The history of population studies on the effects of asbestos in drinking water is a virtual textbook demonstration of drawbacks to attempting to use ecologic epidemiological methods to establish the safety or danger of a suspected carcinogen.

The danger of inhaled asbestos was established relatively early because people with a known heavy exposure could be followed individually, though the latency to develop cancer from asbestos inhalation is typically 30-40 years.

In contrast, until the Norwegian lighthouse keepers were studied, inhabitants of cities with asbestos in the water supply were the only populations available for study. Unlike the lighthouse keepers, whose employment records at individual lighthouses, and hence exposure, could be individually determined, the city dwellers, for the most part, were not followed individually; they drank water with asbestos content that varied with rainfall, acidity, and their location in relation to the water system (unless they used bottled water), and, being mobile Americans, were often no longer living in the city under study after 20 to 40 years.

There must have been a great impetus to know if oral asbestos caused cancer at the time its presence in water supplies was discovered; several studies were conducted before sufficient time had passed to expect cancers to have developed (CCERP, 1987).

Even with all their drawbacks, some of the population studies did show statistically significant increases in cancer rates in association with asbestos in drinking water.