We must preface this review with the same comments that we have made about our review of all other prospective members of BEIR VII. We in no way question Dr. Krewski’s scientific credentials in his area of expertise, his integrity as a regulator or a scientist, nor his right to hold the opinions he does about the risks associated with exposures to chemicals or radiation. The issues about which the National Research Council (NRC) has requested public comment go to whether the addition of Dr. Krewski to the BEIR VII panel solves the balance problems regarding the panel, and whether there are any conflict-of-interest issues regarding Dr. Krewski.

We also must add, as we have previously, that we have been hampered in this evaluation by the NRC’s refusal to release even a full curriculum vita for Dr. Krewski, let alone his conflict-of-interest "disclosure" forms. NRC has merely posted on its web site a four-sentence description of Dr. Krewski. It indicates he is a co-author of more than 300 papers. However, by refusing to release the c.v., and permitting less than three weeks for review and comment, we have been able to identify a small fraction of his work. Surely there is no defensible reason for declining to release the c.v. so that an adequate review could be possible and meaningful comment result. Similarly, in the absence of the conflict-of-interest disclosure form, public comment on those (secret) conflicts is impossible.

We therefore apologize to Dr. Krewski if this evaluation provides an inadequate picture of the whole corpus of his work, but given the refusal to release the c.v. and conflict forms, coupled with the short review period, the public is severely handicapped in its ability to do a full analysis.

Dr. Krewski has spent much of his professional career as a regulator. The four-sentence description of him on the NRC website states that he is currently on the faculty of the University of Ottawa, and that "[p]reviously he served as director of Risk Management at Health Canada." However, the brief biography given for him on the website for BELLE, the hormesis organization on whose Advisory Board he serves, states that he is "currently Acting Director of the Bureau of Chemical Hazards of the Environmental Health Directorate of Health Canada." It would help if the NRC staff were to clarify the apparent contradiction -- is Dr. Krewski currently serving in a regulatory position for Health Canada and, if not, up until how recently did he have such regulatory duties? Perhaps the explanation is that he previously served as director of Risk Management for Health Canada, and now serves as Acting Director of the Bureau of Chemical Hazards. In any case, the NRC brief bio creates an incorrect impression of someone who may have long ago served in a regulatory capacity but for many years has been an academic researcher separate from regulatory pressures. Dr. Krewski has spent much of his career as a government regulator, and based on his bio with BELLE, appears still to have such regulatory duties.

Dr. Krewski’s background as a regulator raises some potentially troubling conflicts-of-interest questions. A regulator is under pressures that are very different from that of a pure scientist. Powerful economic and political interests very frequently are taken into account when setting regulatory standards, rather than purely scientific and public health considerations. If particular carcinogens are proposed to be regulated strictly in order to protect public health, businesses and agencies that release such carcinogens engage in extensive lobbying to block such efforts. At every step of risk estimation, such lobbying occurs about the assumptions to be employed. A regulator who sticks purely to scientific and public health considerations will frequently not last long nor rise high in a regulatory system. Someone whose career has been primarily that of a successful regulator comes to issues of risk differently than does a pure scientist, making it very difficult for the former to separate scientific issues from other factors (such as economic impacts on powerful interests).

Additionally, there are other potentially troubling conflicts in such a role. Dr. Krewski in his capacity as a regulator has been responsible for a number of very controversial regulatory decisions. He has authorized public risks that many felt were excessive. Can he realistically be expected, in a committee addressing in part the degree of risk from current and past permitted levels of radiation exposure, to reach a conclusion that the standards he had helped set and enforce were too lax and exposed people to excessive risks?

There is something troubling about permitting a current or recent regulator to serve on an important scientific committee whose risk estimates will inevitably have critical regulatory implications. BEIR VII is to establish risk estimates that will implicitly either affirm past standards as sufficiently protective or insufficiently so, or, most likely, given the imbalanced composition of the panel, concluding that risk estimates should be lowered, permitting further relaxation of standards. A regulator has too much at stake in such a conclusion to be involved in making it.

The presence of a regulator on a scientific panel relied upon by regulators raises serious questions of conflicts-of-interest. Health Canada has a great deal riding on the outcome of BEIR VII. Should Health Canada be represented on it? If so, why shouldn’t the U.S. Department of Energy, Nuclear Regulatory Commission, and the DOD nuclear agencies have their bureaucrats serving on the panel too? We have already pointed out numerous conflicts of interests with scientists named to the panel being too close to the agencies that will be affected by BEIR VII’s conclusions. The Research Council has implicitly conceded we were correct regarding at least five of the people it had originally named to the panel (and obviously we continue to believe numerous others still on the panel have similar disqualifying conflicts). If those removed had disqualifying conflicts because of ties to these agencies, how then can one place on the panel someone whose ties to an agency dependent upon the BEIR VII outcome are far more direct -- years and years of serving as a high official, either continuing at present or only recently ended?

We should make clear that a regulatory role in the distant past followed by decades of academic positions is a different matter. But Dr. Krewski’s role as a regulatory bureaucrat extends to the present period. He was deeply involved, as we will discuss below, in numerous decisions that could either be justified by or called into serious question by the results of BEIR VII. His institution, Health Canada, will be deeply affected by the outcome of BEIR VII. There is something intellectually improper about the presence of a regulator on a scientific panel relied upon by regulators.

The current composition of the BEIR VII panel is grossly imbalanced. Designed to review the Linear No-Threshold model and the range of viewpoints in the scientific community, only one of the three basic positions to be examined is represented on the panel. Intriguingly, that position (that the strict LNT model overstates risks) happens to coincide with the interests of the financial sponsors of this and other studies performed by the Research Council’s Board on Radiation Effects Research (BRER).

We have some sympathy for the Research Council staff involved, in that they are on soft money, dependent year-by-year for their funding by successfully obtaining subsequent contracts from DOE and other nuclear agencies. However, that creates the obvious conflict that were a panel to be set up, with members well balanced and without significant ties to DOE, and reach conclusions that made DOE unhappy, subsequent contracts (and continued financial support for the livelihood of individual BRER staff-members) would be placed at risk. This is no idle threat, as that is precisely what happened to the NRC radioactive waste board in the 1960s. As documented in Philip Boffey’s book on the Academy, when a Research Council committee issued reports critical of Atomic Energy Commission’s (AEC) waste practices (prescient, it turns out, given what has in recent years been disclosed about contamination at the AEC/DOE complex), AEC threatened to cut off all funding to the Research Council’s waste board. The threat resulted in the Academy caving in to the threat, reconstituting the board with people friendly to AEC, and a marked change in the reports issued therefrom--reports far less critical.

Nonetheless, the Federal Advisory Committee Act (FACA) requires that Research Council committees be balanced and free of conflicts-of-interest. We have provided detailed critiques of the BEIR VII panel composition as initially proposed and as subsequently modified. Those criticisms at their heart indicate that all scientists whose work either suggests risks are strictly linear with dose or greater than presumed by LNT have been excluded from the committee. Of the three basic viewpoints in the scientific community, the panel is dominated by people holding one viewpoint -- that low-dose radiation risks are less than linear, with many of the panel having done work cited by proponents of further relaxing radiation regulations. Those whose work suggests current risk estimates understate risks, and by implication radiation regulations need to be strengthened, have been consistently and, we believe, consciously excluded. This violates FACA’s core.

In the latest round of changes, only one new proposed member has been added, Dr. Krewski. The fundamental question is whether his views at least would provide a single token representative of the excluded views, or whether he reinforces the existing imbalance. Our review of Dr. Krewski’s past actions and publications regrettably leads us to the latter conclusion. We will begin our discussion with one of his most recent actions and publications.

This report was issued just last year by a Joint Working Group of Health Canada and AECB. Dr. Krewski was co-chair of the Working Group. Its purpose was to examine whether radiation protection regulations should be strengthened so as to conform to permissible risk levels established for all other carcinogens. Radiation is currently regulated, in Canada and the U.S., at risk levels orders of magnitude more lax than permitted for chemical carcinogens (1 in 100 lifetime cancer risk permitted for members of the public from exposure to radiation at permissible levels compared to the 1 in 1,000,000 risk permitted for chemical carcinogens.)

The context of the Working Group and its report is as follows: In 1994, the Ontario Advisory Committee on Environmental Standards recommended tritium in drinking water be limited to 100 Bq/L, to approach the maximum permissible risk level allowed for chemical carcinogens. The Ontario Minister of Environment and Energy, however, issued a document establishing tritium limits considerably higher, consistent with general radiation risk limits. The different approaches led the Minister "to request guidance from Health Canada regarding the apparent differences between acceptable risk levels used in regulating radionuclides and chemicals." Dr. Krewski, as indicated above, co-chaired the effort, in which he concluded that RADIATION PROTECTION REGULATIONS SHOULD NOT BE STRENGTHENED. While conceding that radiation was regulated at risk levels far greater than permitted for any other carcinogen, he blocked bringing them into conformance with the stricter risk levels permitted for chemical carcinogens.

The content of Krewski’s report, however, is also disturbing in what it reveals about his views on ionizing radiation. He and his colleagues make a number of statements in their joint report that clearly come down firmly on the side that low-dose radiation is either much less dangerous than presumed by LNT or not dangerous at all. Here are a few of the statements:

"Some attempts have been made to obtain direct estimates of risk from groups that have received low-dose exposures, such as people exposed to radon in homes and nuclear industry workers. Within the limits of sensitivity of these studies, increased cancer risks have not been identified." (p. 6, emphasis added). This is an extraordinarily false statement, and frankly should be disqualifying for Dr. Krewski in terms of possible service on BEIR VII. Numerous studies of nuclear workers have found excess cancers associated with radiation exposures at facilities such as Fernald, Oak Ridge, Hanford, Rocketdyne, and others. Indeed, Dr. Krewski’s own 1999 paper on the Canadian Occupational Radiation Registry found significant increased cancers, and he concedes there his findings are within the range of what effects have been found elsewhere in these dose ranges.

"[E]vidence is growing that there may be an effective threshold below which there are no adverse effects of low doses of radiation..." (p. 6, emphasis added)

Page 16: they support the divergence from LNT by use of a DDREF; even so, they argue this is "conservative," i.e., overstates true risks. They say that although it is "conservatively" assumed for regulatory purposes that any increase in radiation exposure results in a proportional increase in cancer risk and that of genetic disorders, there is "evidence to the contrary."

p. 17: They challenge LNT, saying, "Some evidence from both human and animal studies suggest that in certain cases, notably for the induction of bone cancer by radium-226, a practical threshold dose exists below which the chance of producing a bone cancer within the normal lifespan is virtually zero" (emphasis added)

They even affirm hormesis: "There is also some evidence of a reduction of cancer rates in humans on exposure to very low doses of radioactivity, resulting from the stimulation of repair mechanisms." (emphasis added) They do note that the data as of 1996 on this matter were not sufficient "at present" to be taken into account in setting regulations, but pose a troubling indication of Dr. Krewski’s views on a matter the non-regulatory BEIR committee is to address when it issues its report in 2002 or so.

Referring to the ICRP risk estimates -- a reduction from LNT by a DDREF of 2 -- Krewski says that they "may overestimate risk at low-doses."

He goes on to say that "the linear no-threshold model cannot be used to predict the outcome of actual exposures to an individual or a population," that it is solely a regulatory tool.

He notes that under LNT, "potential societal harm would be determined by the total population dose," but goes on to apparently affirm the controversial recommendation that in calculating total population dose, doses less than 10 microsieverts per year "should be categorized separately from and not added to those of higher individual doses." He then makes the extraordinary statement that "[t]his recommendation reflects the lower level of concern about the health risks of such doses which are considered negligible even if the linear no-threshold hypothesis is assumed to be correct." Just a few sentences earlier he had conceded that if LNT is correct, "If a very large number of individuals were exposed to low doses of radiation from various sources, below any limit set for individuals, the total dose to the population could be appreciable." Yet shortly thereafter he applauds not counting the low doses, saying there is a lower level of concern about the health risks of such doses and they are negligible. Obviously, he doesn’t believe in LNT, but believes low doses, even if spread over many people, have essentially zero effect.

Krewski and colleagues go on to cite the studies which they feel are "the most important" regarding low-dose radiation, citing only the IARC pooled studies which have been heavily criticized for drowning positive findings at individual sites by improperly mixing apples-and-oranges from several sites with non-comparable exposure pathways (e.g., plutonium at one site, iodine at another) and non-comparable exposure records. None of the studies which show positive effects are cited, showing his bias on the matter, it would appear. Additionally, he characterizes the Cardis et al study as having "failed to establish a clear cancer risk at low levels of dose," which goes even beyond how Cardis and her colleagues have characterized their results. We note that Dr. Krewski is a colleague of Dr. Cardis’, representing Canada at IARC, and having co-authored papers with her; so once again, evidence that Dr. Krewski, rather than balancing the bias of Dr. Cardis and others on the panel, comes from the same side of the debate.

p. 23: Dr. Krewski comments favorably on recommendations to treat individual exposures of 0.01 mSv per year to members of the public as de minimus levels, saying, "[e]xposures at this level would be regarded as safe."

p. 25: Dr. Krewski here refers to guidelines, which it would appear from his radon work he was involved in setting, that remedial measures only be taken where the level of radon in a home is found to exceed the amount sufficient to produce a dose of 14 mSv per year, "equivalent to a risk of fatal cancer of about 1 in 15." This is an extraordinarily non-protective standard. Elsewhere Dr. Krewski has written that it is in general not cost-effective, in his view and that of his agency, to remediate radon exposures.

p. 39: Dr. Krewski, in reference to LNT models, says there is "[s]ome evidence for practical threshold effects for specific radionuclides."

p. 42: Dr. Krewski’s report concludes that existing risk management strategies for ionizing radiation provides "a high degree of health protection, based on the absence of observable health effects using epidemiological methodology." Current official risk estimates for those standards is a 1 in 286 lifetime risk of fatal cancer for members of the public exposed to permissible levels of radiation, and 1 in 8 for workers. It is extraordinary that Dr. Krewski would describe this as a "high degree of health protection." Furthermore, it is simply false to state that there is an "absence" of observable health effects in epidemiological studies. Numerous worker studies, including his own 1999 study, showed significant excess cancers at doses far below what is permitted under current regulations.

It is conceivable that Dr. Krewski will claim that others wrote the sections of the report quoted above. However, he was co-chair of the Working Group and signed off on the whole report. If these statements are incorrect or he otherwise disagreed with them, he had a professional obligation to not approve the report or attach his name to it. One must presume it represents his views.

In this testimony regarding high level nuclear waste, Dr. Krewski and his colleagues assert that the LNT model is "conservative," i.e., it overstates true risks. (On questioning, his colleague Dr. Bartlett was questioned about this assertion.) Why, it was asked, if there is uncertainty in the low dose region, might not linear extrapolation indeed be non-conservative, might understate risks, with the shape of the curve being supra-linear. In their written testimony, they asserted that nuclear power plants and nuclear waste do not pose a significant health risk, and that the public’s perception of the risks is greatly in error.

When pressed to present an estimate of collective dose from the proposed high level waste repository, Dr. Krewski and his colleagues at Health Canada made the following statements:

Health Physics Society Newsletter

March 1996

Dr. Krewski has thus taken the position that "there is no scientific evidence to support the conclusion that an individual dose as small as one microSievert could carry any risk at

all. He has implied that it is unreasonable to consider collective dose, the fundamental premise the flows from the LNT model. And he has applauded the controversial Health Physics Society position that collective dose is "highly speculative" and "should not be quantified for the purpose of estimating population health risks." This is a key part of controversial HPS position advanced by Dr. Mossman that made his initial appointment to the BEIR VII committee an indication of such imbalance, given the lack of counter on the committee. Dr. Krewski’s adherence to these views indicate that imbalance is in no way remedied.

This testimony goes on to conclude, "There is no firm scientific evidence to support

the occurrence of any kind of health effects at individual radiation doses that are a tiny fraction of normal background exposure." (emphasis added) Dr. Krewski’s repeated attacks on the LNT principle as a means of estimating real risks places him squarely on the side of that debate that is already overwhelmingly represented on the BEIR committee and fails to provide any of the balance so desperately needed.

In these papers, Dr. Krewski and his colleagues propose abandoning LNT as a means of risk assessment and regulation and replacing it with a procedure for calculating "benchmark doses" and uncertainty/safety factors. Essentially they propose estimating the dose necessary from animal tests to increase tumor incidence by 10%, then setting a regulatory limit at something below that, by assuming an uncertainty factor. The particular factor they suggest would result in about a ten-fold relaxation of current regulations for chemical carcinogens in a number of settings. It is unclear what Dr. Krewski and his colleagues would propose as limits for radioactive carcinogens using their method. But they propose not using LNT to estimate risks at low doses, and getting away from any presumption of risk at low doses, saying, "quantitative statements about low levels of risk generally are difficult to justify."

We recognize that Dr. Krewski earlier in his career has made statements more supportive of LNT. But his recent work seems to move in the other direction.

Dr. Krewski now serves on the Board of BELLE, the primary hormesis organization. He has made statements in the past supportive of hormesis, some of which have been quoted above. See also, Health Physics, Vol. 57, Supplement 1, 1989, p. 324-5 (when asked "I would like to reaffirm the need to incorporate the possibility of hormesis into quantitative models," he responded, "I agree." In response to another question, he said, "It is certainly true that most dose-response models used to describe toxic phenomena, including carcinogenesis, do not provide for the possibility of hormesis. However, to the extent that hormesis acts so as to reduce risk at low doses, current methods of risk estimation should tend to be conservative in the presence of such effects."

Industry, Dr. Krewski, and His General Approach to Carcinogens

Dr. Krewski was a member of a Research Council panel on synthetic carcinogens in foods paid for in part by the food industry (e.g., Nabisco Foods Group), which obviously has a major financial interest in getting a study that concludes their products are safe. The study was highly controversial and frankly somewhat outrageous in its conclusions. The press release was headlined, "Excess Calories Pose More of a Cancer Threat Than Natural or Synthetic Carcinogens in Foods." It began by asserting, "Cancer-causing chemicals that occur naturally in foods are far more numerous in the human diet than synthetic carcinogens, yet both types are consumed at levels so low that they currently appear to pose little threat to human health." Industry could not have been more pleased. Indeed, the pesticide industry, through its lobbying group the American Crop Protection Association, issued a press release trumpeting the findings, claiming that pesticide residues in foods pose essentially no risk. His participation in that controversial Research Council study, funded by and favorable to industry, raises serious questions about how he would fare in a controversial radiation study funded by and heavily tilted in composition towards the interests of the nuclear agencies.

We are aware of two studies of ionizing radiation performed by Dr. Krewski. One focused on indoor radon in one community in Canada. Another, far larger, examined the effects of occupational exposures to external radiation through the National Dose Registry of Canada. The response by Dr. Krewski to his own findings in each case is instructive.

The small radon study produced inconclusive results. ("Case-Control Study of Residential Radon and Lung Cancer in Winnipeg, Manitoba, Canada", Am. J Epidemiol., Vol. 140, No. 4, 310-322. He states it was designed to provide an 80% probability of seeing a 50% increase in lung cancer risks. At best what can be said from his failure to find a statistically significant association is, even assuming that the significant methodological problems didn’t impair the study, the radon exposures didn’t result in a 50% increase in lung cancer. But a 50% increase amongst the general public is a large effect, and radon could be increasing lung cancer rates by 20, 30, or 40% and his study couldn’t see that. Nonetheless, the paper has a good deal of language in it suggesting that indoor radon may not be harmful, a matter which goes far beyond these limited data.

On the other hand, the massive occupational study of all workers in Canada with radiation dose data found a striking excess in cancers with exposure to low dose radiation. "First Analysis of Mortality and Occupational Radiation Exposure based on the National Dose Registry of Canada," Am. J Epidemiol, Vol. 148, No. 6, 1998, 564-574. The study found an effect size an order of magnitude higher than that found from the A-bomb survivors, but consistent with the higher risk levels found by Steve Wing at Oak Ridge, Alice Stewart at Hanford, Morgenstern at Rocketdyne, and Ritz and Fernald (although Krewski only cites to Wing). Yet these results, quite extraordinary in their importance, based as they are on such a large population and with far better exposure data than the radon study, is downplayed to a level that is remarkable. Rather than pointing out how it lends support to the series of studies suggesting current risk estimates may be low by an order of magnitude, he merely says the findings are in the range of all studies, from the very low (e.g., Cardis) up. Note also that this study was based solely on external (i.e., SMR) comparisons, rather than internal comparisons of outcomes of higher exposed versus lower exposed workers. Dr. Krewski and his colleagues demonstrate that the data prove there was a healthy worker effect at work, but fail to draw the conclusion that the effect size they found would therefore still underestimate the true effect of low dose radiation on this population, because the healthy worker effect masked some of it.

If the Research Council were taking seriously its responsibilities under FACA and attempting to correct the marked lack of balance in the BEIR VII committee as currently composed, it would have to add something on the order of a dozen scientists who believe current risk estimates understate true risks to balance off the large number of people already on the panel who believe the contrary. No such additions have been made. Instead, Dr. Krewski appears to be someone who also believes LNT overstates risks, who has sympathy for hormesis arguments, who questions whether there is any risk whatsoever at so-called low doses, who opposes calculation of collective doses. The imbalance remains identical to what it was at the beginning of this process. The currently proposed BEIR VII committee violates FACA just as thoroughly as it did at the beginning. Everyone who has a position is on one side of the debate, with the other side of the scientific community completely excluded. The excluded wing of the scientific community, coincidentally, happens to be the side whose views of radiation risks are less in concert with the economic interests of the study’s funding agencies such as DOE -- which can save huge amounts of money if cleanup standards are relaxed -- and more in accord with the interests of protecting public health. The Research Council continues to violate law and the fundamental principles of science by its continued insistence on a stacked panel on an issue of immense importance to the health and safety of the public.