Mad Science: The Nuclear Power Experiment

by Joseph Mangano

  The National Institutes of Health is actively involved in studying the adverse effects of ionizing radiation, and we concur with your view that the risks at low levels need further clarification… The descriptive studies of leukemia clusters around the Pilgrim power plant in Massachusetts, and several plants in the United Kingdom, have led us to initiate a large-scale evaluation of cancer deaths occurring among persons living near the over 100 reactors operating in the United States. We are correlating county mortality data from the 1950s through early 1980s with reactor operations to determine whether the previous reports might be chance occurrences based on small numbers, or whether there might be valid reasons for concern. This evaluation should be completed within about one years.

  The NIH study, assigned to the National Cancer Institute, was not completed “within one year” as Wyngaarden predicted, but nearly three years – giving greater support to the belief that the NIH had done nothing until Kennedy’s letter arrived. Even before the study had begun, Wyngaarden expressed doubt whether results of such a study could be helpful, because of the “very small health effects” involved – a curious opinion given that he had committed the NIH to a major undertaking, and a violation of scientific objectivity prior to an examination of data in a relatively unknown area:

  It is important to stress that useful information about very small health effects, like those associated with very low levels of radiation, is extremely difficult and expensive to obtain. An indirect approach, such as studying populations with higher-level exposures and extrapolating the results to lower levels, tends to be more productive.

  Another statement made by Wyngaarden that revealed bias against finding any radiation–cancer link was the following:

  The most serious health impact of the Three Mile Island (TMI) accident that can be identified with certainty is mental stress to those living near the plant, particularly pregnant women and families with teenagers and young children.

  At the time, nine years after the Three Mile meltdown, not a single journal article examining trends in cancer rates near the reactor had been published.

  In July 1990, the federal report was finally finished, and in September of that year, the National Cancer Institute unveiled results of its study, contained in three volumes about seven inches thick. The eight-member committee that directed the study made its conclusion clear, on the very first page of the summary: “The Committee concludes that the survey has produced no evidence that an excess occurrence of cancer has resulted from living near nuclear facilities.”

  The NCI study had looked at cancer death rates in 107 counties near sixty-two US nuclear plants (ten weapons and fifty-two power plants) during the years 1950 to 1984, covering 880,028 persons who died of cancer. The basic approach taken used “Standard Mortality Ratios” of local vs. US rates; for example, if a local rate was 2% below the US, the SMR would be 0.98. The study examined SMRs in the periods before and after a nuclear plant began operations, both in the 107 “study” counties and in 321 “control” counties that were demographically similar but geographically distant from any nuclear plant.

  Naturally, cheers went up from pro-nuclear advocates, including Michael T. Rossler of the Edison Electric Institute, who said “The study is helpful. It shows there is not a problem to individuals, that there is not a danger from radiation.” Others were not convinced. Kennedy himself commented that the study “is not a clean bill of health for the nuclear industry.” Study conclusions were used for decades after by the nuclear industry as “proof” that reactors operated safely, and did not pose any health problems for local residents.

  While it was important that the federal government finally acknowledged the issue of cancer near nuclear plants and produced some research, the study had a number of limitations. A brief summary of these follows:

  Mortality Only. The study examined cancer deaths near all sixty-two nuclear plants included, but considered cancer cases near only four of these, because only the states of Connecticut and Iowa had established cancer registries for most of the period 1950–1984.

  One-Third of Reactors Excluded. Only nuclear plants that began operations by 1981 were included. This means that of the sixty-five plants (with 104 reactors) now operating, data for twenty-three plants (with thirty-five reactors) started after 1981 were excluded from the study.

  Only Data to 1984. The latest data (1984) were already six years old when the report was released; the fact that no other national study has been done since makes the data over a quarter century old.

  Ignores Wind Direction. A “study” county was selected purely based on proximity to reactors, ignoring other factors such as whether it was downwind from the nuclear plant.

  Whole Counties Only Used. The use of entire counties can be criticized because 1) no sub-county data was used, and 2) only the one or two closest counties were used. (This is a difficult issue to resolve, but some maintained the NCI should have expanded the study to include more counties and/or sub-county data.)

  Cancer Only. The study only considered cancer as a potential health effect of living near a nuclear plant, and excluded others such as underweight births, infant deaths, and birth defects.

  Control Counties Close to Nukes. Some of the NCI “control” counties selected as unexposed areas are actually close to reactors, in some cases adjoining “study” counties, and thus were probably also exposed to releases from reactors.

  The NCI study actually raised some red flags – that is, data suggested that living near a nuclear plant was in fact linked with elevated cancer risk. One good example is Three Mile Island, the site of the worst meltdown in US history (Santa Susana was not included in the NCI study). Officials chose to examine cancer mortality data for three Pennsylvania counties closest to TMI; Dauphin (where the plant is located), Lancaster, and York. York County is also the site of the Peach Bottom nuclear plant.

  The NCI study provides data on the local vs. national cancer death rate for the periods 1970–1974 (before the plant opened) with 1975–1984 (after startup) for the counties closest to TMI. Of the fourteen types of cancer studied, the local-vs.-national Standard Mortality Ratio (SMR) near Three Mile Island showed:

  – ten types of cancer increased after startup

  – two types of cancer decreased after startup

  – two types of cancer were unchanged after startup

  The SMR for the most radio-sensitive cancers all rose, including thyroid cancer (+58%), bone/joint cancer (+13%), and leukemia (+10%). Moreover, the SMR for children dying of cancer by age nineteen rose 10% (below):

  Source: Jablon S. et al. Cancer in Populations Living Near Nuclear Facilities. Washington DC: U.S. Government Printing Office, 1990.

  Another indication that the NCI study showed living near nuclear plants may have harmed Americans involves incidence data. Cancer incidence (cases) is much more helpful than cancer mortality (deaths) in understanding the effects of a pollutant such as radiation. Advances in medical science have enabled physicians to keep many cancer victims alive; for some types of cancer, survival rates are very high, including thyroid and child cancer, which are probably most sensitive to radiation. While all US deaths have been recorded since the 1930s, there is no national cancer registry of diagnosed cases. The only states which had established cancer registries in the years covered by the NCI study were Connecticut and Iowa, which operated four plants (a small fraction of the sixty-two plants with mortality data in the report).

  Even though four plants is not a large number, some interesting results were obtained. Thyroid, breast, and child cancer incidence increased after startup near each of the four plants (below). It appears that lack of incidence data was a major limitation of the NCI study.

  Numerous questions were raised after the 1990 NCI study was unveiled. One of them was that there was no plan by federal regulators to follow the study up and improve its methodology – clearly, it was only done because of pressure applied by Senator Kennedy. Nuclear supporters cited the researc
h as evidence that reactors posed no health risk – even decades later. In July 2010, the Nuclear Energy Institute issued a “White Paper” on radiation safety – a dubious undertaking, since the NEI is the lobbying group for the large corporations who operate nuclear reactors. The White Paper’s section on studies of radiation studies to the public first cited the NCI study – even though the data was between twenty-six and sixty years old.

  For the next two decades, federal regulators took no actions to conduct other studies of cancer near nuclear plants.

  In the mid-1980s, Dr. Jay Gould became involved in the issue of hazards of nuclear reactors. The 1979 partial meltdown at Three Mile Island and the 1986 complete meltdown at Chernobyl sparked his interest. Gould was an economic statistician who served as an expert witness in numerous legal actions. One of the last cases he worked on before retiring involved the Westinghouse Corporation, which along with General Electric manufactured over 90% of the world’s nuclear reactors. Gould read about Westinghouse, and became intrigued by the issue.

  At about the same time, Gould learned about Ernest Sternglass, who had just retired from the University of Pittsburgh and moved to New York City, where Gould lived. The two met one night, and hatched the idea of forming a group of professional researchers independent of government and industry to conduct the much-needed studies of cancer risk from nuclear reactors. They decided to name the group the Radiation and Public Health Project (RPHP).

  Over the ensuing quarter century, RPHP produced twenty-seven medical journal articles, which had to undergo strict peer review of and acceptance by experts before publication. RPHP members recruited by Gould and Sternglass also made extensive efforts to share research findings with the public, producing seven books on the topic of health risks of relatively low dose radiation exposure from reactors. The group also held twenty-three press conferences after the year 2000, leading to considerable exposure from media.

  Some of the principal findings from RPHP are as follows:

  Baby Tooth Study. Beginning in 1998, RPHP collected nearly 5,000 baby teeth from US children, and sent them to a specialty lab that measured levels of strontium-90 in the teeth. The study was Gould’s brainchild, inspired by a similar effort in the 1960s by Washington University in St. Louis examining in-body atom bomb fallout. The RPHP tooth study involved citizens’ groups and celebrities like actor Alec Baldwin and model Christie Brinkley to collect teeth, and results were published in five medical journal articles. The major findings were that children in counties closest to nuclear plants had 30–50% higher levels of Sr-90 than those living far from reactors, and average Sr-90 levels rose 50% from the late 1980s to the late 1990s. Most importantly, trends in Sr-90 near nuclear plants in New York and New Jersey were followed by similar trends in local childhood cancer. The tooth study remains the only study measuring radiation levels in bodies of Americans living near nuclear power plants.

  Notes: Rates adjusted to 1970 US standard population. Counties included are Middlesex CT (Haddam Neck), New London CT (Millstone), Benton/Linn IA (Duane Arnold), Harrison IA (Ft. Calhoun). SIR represents the county rates vs. the state rate.

  Closed Reactor Study. Group members found that when US nuclear plants closed, the health of infants and young children living near and downwind improved immediately and dramatically. Specifically, the first two years after shutdown near all eight plants studied found large declines in infant deaths, birth defects, and cancer diagnosed in children under age five.

  Child Cancer Study. The rising child cancer rates near four nuclear plants in the NCI study inspired more detailed research by RPHP. It examined cancer incidence in children diagnosed by age nine in the period 1988–1997 near fourteen nuclear plants in the eastern United States. Near all fourteen plants, the local cancer rate exceeded the US standard, with a total excess of 12.4%.

  Thyroid Cancer Study. The US incidence of thyroid cancer, which is highly sensitive to radiation exposure, is increasing more rapidly than any other cancer. RPHP researchers examined thyroid cancer rates in the 500 most populated US counties and found that of the eighteen with the highest rates, eleven were located in a small area of eastern Pennsylvania, central New Jersey, and southern New York. This area has sixteen nuclear power reactors (thirteen still operating), the greatest density in the US.

  Coverage of RPHP presented an obstacle to the “selling” of nuclear power, and industry leaders were furious. The tooth study caused some of the strongest vitriol from industry, and for good reason: here was hard evidence of poisonous chemicals produced by reactors in the bodies of American children. The Nuclear Energy Institute, made up of corporate giants like Entergy, Exelon, and Duke Power, embodied the fears of these research findings and how they might sway public opinion and public policy. As part of a twelve-page diatribe against the tooth study placed on its web site, the NEI declared:

  For several decades, a small group of activists has tried to instill fear in the public that a substance called strontium-90 is evidence that low levels of radiation released from nuclear power plants causes cancer and health problems in nearby residents. Since the claims first surfaced some thirty years ago, they continuously have been dismissed by mainstream scientists as scare tactics and ‘junk’ science contributing nothing to finding the real causes of cancer. They are instead manipulations of the public by these groups without any basis in science. These studies are known as the ‘tooth fairy project.’

  In a blunt example of the industry’s hammerlock on government regulators, the NRC posted an eight-page critique on the tooth study on its website at about the same time. The NRC took a position similar to the NEI, using many of the same references and the same wording. The NRC, which employs many physicists and engineers, but no health researchers, made the declaration that low doses are harmless – an opinion that defied the growing trend of studies on X-ray radiation to pregnant women, Americans exposed to bomb test fallout, and exposed workers at nuclear weapons plants:

  If ingested in quantities that produce very large radio-logical dose rates (about a thousand times higher than doses we all receive from natural radiation), Sr-90 is known to increase the risk of bone cancer and leukemia in animals, and is presumed to do so in people. Below these doses, there is no evidence of excess cancer.

  While most scientists and health officials were quiet on the issue, some succumbed to the power of the NEI and NRC, and issued public critiques of the tooth study. But some professionals stood up for the study’s integrity. Donald Louria, a physician who was Professor of Preventive Medicine at the New Jersey Medical School and who was familiar with RPHP tooth study activities in New Jersey, was one of these who asserted the dangers of strontium-90 in the body:

  Studies such as those… measuring body burdens of that dangerous radioactive isotope, strontium, and relating those body burdens to cancer risk are very useful and important.

  The verbal fight between RPHP supporters and detractors continued for years. But the group endured, chiefly because it was not beholden to government for support, as Dr. Thomas Mancuso and Dr. John Gofman had been in the 1970s. There were instances, however, when RPHP sought public funds, and ran into trouble. In 2000, the New Jersey legislature allotted $75,000 to support the baby tooth study in the state. But Governor Christie Whitman used a line-item veto to kill the funds, stating that the tooth project should have been paid for by private sources. The following year, EPA administrator Whitman came under tremendous public fire after she stood at Ground Zero one week after the 9/11 attacks and declared the air was safe to breathe. And after leaving the Bush administration, Whitman co-chaired the Safe Energy Council, an industry-funded effort to encourage the expansion of nuclear power in the US.

  The only other studies of cancer near US nuclear plants, other than the ones that RPHP were churning out, focused on Three Mile Island. The reaction of the research community following the March 1979 partial meltdown was disturbing. By late 1990, nearly twelve years after the accident, not a single medical journal ar
ticle addressed actual cancer patterns in persons living near the stricken Pennsylvania plant. (One article shortly after the meltdown, by former NCI director Dr. Arthur Upton, estimated that TMI would cause an additional 0.7 cases of cancer – essentially zero.) However, there were thirty-one articles that addressed psychological and stress concerns among local residents, in publications like Journal of Trauma and Stress, Psychosomatic Medicine, and Health Psychology. Andrew Baum, a psychologist then working for the Defense Information Systems Agency, was the author or co-author of eight of these articles. This represented a clear effort to relegate any harm from the Three Mile Island meltdown to “just stress.”

  The first journal article with actual cancer statistics only was published because resources had become available from a settlement of a legal action that established the Three Mile Island Public Health Fund. Unfortunately, the Fund and the researchers it chose received their money from the nuclear industry. A group of researchers from Columbia University were chosen by the Fund, and the first journal article utilizing official counts of cancer near Three Mile Island finally went to press in September 1990. The team documented that the number of newly-diagnosed cancer cases among the 160,000 residents within ten miles of the stricken reactor soared 64% from 1975–79 to 1981–85, from 1,722 to 2,831 cases. Numbers jumped for leukemia, lymphoma, lung cancer, and childhood cancer. But the group showed a bias in the article, stating that they believed relatively low dose exposures would not show any risk: “prior expectation based on estimated releases and conventional radiobiology – that no excess cancer would be found – was confirmed.” The following June, their next study ended with the comment that local cancer increases “might reflect the impact of accident stress on cancer progression.”

  The Columbia studies remained the only ones to assess actual cancer rates near Three Mile Island until 1997, when a University of North Carolina team headed by Steven Wing published a journal article. The UNC group had been asked to do this study by 2,000 local residents, many suffering from cancer and other diseases, who had filed a class action suit against Three Mile Island. Using the Columbia team’s data, Wing found that, indeed, there was a correlation between radiation exposure and cancer risk near Three Mile Island. Wing pointed out that the Columbia researchers may have been biased in believing that doses from the accident could not cause an increase in cancer – a point that was not lost on the Columbia team. A fight in the journals broke out. Columbia team member Mervyn Susser fired off a letter to the journal Environmental Health Perspectives, which had published Wing’s study. Suggesting that the Wing article was “advocacy parading as science,” Susser went further: “Wing’s position amounts to a charge that we are either incapable of understanding our data… or that we obfuscate or lie … We are not sure we can say the same for Wing et al …. As we see it, they have done no more than muddy the waters.”