ML20049J244

Testimony of Id Bross Re Health Risks to Southern La Populations from Addl Radiation in 1-rod Range.Affidavit, Prof Qualifications & Supporting Documentation Encl

ML20049J244
Person / Time
Site:	Waterford
Issue date:	02/24/1982
From:	Bross I JOINT INTERVENORS - WATERFORD
To:
Shared Package
ML20049J211	List: ML20049J210 ML20049J215 ML20049J219 ML20049J220 ML20049J244 ML20049J246
References
NUDOCS 8203120301
Download: ML20049J244 (100)

Text

f '

r m UNITED STATES OF AMEltlCA N UCLEAIL llEGULATOltY COMMISSION m2 p 10 P139 IEFOllE Tile ATOMIC SAFETY & LICENSING BOARD in the Matter of LOUISIAN A POWEll & LIGilT COMPANY Docket No. 50-382 (Waterford Steam Electric Station Unit 3)

SWOltN STATEMENT OF DR. litWIN D. J. BROSS

1. By whom are you employed and what position (s) do you hold?

Answer. I am employed by Roswell Park Memorial Institute for

' Cancer flesearch as Director of Biostatistics.

2. Is this in a specialized health field? If so, what is the description of the type of health field?

Answer. Yes. Cancer Research and Public IIcalth. .

3. What previous positions have you held?

Answer. From 1952-1959, I was IIcad, Statistical Design and Analysis Section at Cornell University Medical College and the Sloan-Kettering Institute in New York City. From 1949-1952, I was ltesearch Associate in the Biostatistics Department of the Johns llopkins

~

University in Baltimore.

8203120301 820308 ,

PDR ADOCK 050003B2 T PDR

~

f .

4. What are your academic qualifications and degrecs?

Answer. I hold a M.A. and Ph.D. in Experimen%1 Statistics from the University of North Carolina, the latter granted in 1949.

5. llave you done post-doctoral work? If so, in what field or fields?

Answer. No formal post-doctoral work.

6. 'IInve you donc any research in the fields of cancer and/or human exposure to radiation?

Answer. Since 1952 I have been heavily involved in cancer research and since about 1967 in research on health effects of low-level radiation.

7. Please describe your research.

Answer. My direct involvement in research on radiogenic cancer occurred when I became Acting IIcad of Epidemiology at RPMI in addition to my job as Director of Biostatistics. During my 7 years as Acting IIcad,I developed a program in Biometric Research on Cancer Epidemiology which developed the biostatistical techn' ology for radiation research, which was subsequently applied to data from the Tri-State survey. More recently, I have analyzed data from the Portsmouth Naval Shipyard Study on health effects among nuclear submarine workers.

8. What publications have your works appeared in?

l Answer. My more than 300 articles have appeared in many different 1

journals. These journals include the most prestigious journals in general science, general medicine, statistics, epidemiology, public health, cancer research, and other disciplines. (See Bibliography).

9. Which of your published works deal with the areas of your research?

Answer. Almost all of them.

l 10. Do you have any as yet unpublished research data compiled?

Anwswer. Yes.

a.

r

11. llave you participated in any scientific colloquia? If so, where, when, under whose sponsorship, and what topics have you dealt with?

Answer. I participated in many scientific colloquia on many different topics on health hazarch of radiation. I have recently given invited papers to the American Statistical Association (1980), and the Yale Symposium on the health effcets of low-level radiation (1981). In 1979, I spoke at the invited symposium in Dusseldorf on metastasis and at the University of licidelberg on radiation hazards.

12. Ilave you ever appeared as an expert witness in state, federal or congressional hearings or courts?

Answer. I appeared at a 1978 NRC hearing held specifically for the purpose of reviewing our radiation-leukeumia findings. I was a principal witness at Congressional hearings on radiation hazards in 1978 (Serial No.95-179) and at Other Congressional hearings such as one on February 25, 1980 on cancer research. I appeared at a state legislative hearing on December 10,1981 (and on several previous occasions) and before the Ontario Provincial Legislature. I also have been involved in the quasi-judicial NRC and the New York State hearings on licensure.

13. Would you please define for purposes of this discussion:

(a) "DN A" Answer. The genetic information stored in a double helix chemical structure.

(b) " Carcinogen" Answer. An agent capable of causing cancer (here, t.uman cancer).

(c) " Doubling dose" Answe . The dosage of a carcinogen that will double Le risk of cancer (rele.tive to baseline levels for a given category of individuals).

4

14. Would you please define synergism and indicate how this phenomena would affect health risks to a population.

Answer. In general, synergism means that the combination of two risk factors produces a more-than-additive effect on the risk. For specificity, the scale used for measuring risk and the characteristics of the population at risk and the diseases under study may have to be spelled out.

15. What do you mean by "the one rad range"? Ilow does the term rad _

relate to the term rem? Is there any special significance or difference between the two terms in discussions of low level radiation?

Answer. At the 1978 NRC hearing mentioned in Question 12, it was stipulated that for the purposes of that hearing, the terms rad _ and rem could be used interchangeably in referring to diagnostic x-rays and low-level nuclear radiation. The 1-rad range is the range up or down by a factor of 10 from 1 rad (100 to 13,000 millirem). This specifies more exactly what is generally called " low-level radiation". Below 100 millirem is commonly called " background radiation". Above this range is " thera-peutic radiation", although usually this would be 50 rads or more.

16. What do you mean by and what is the significance of " indicator diseases"?

. Answer. By " indicator diseases", we mean lesser diseases that tend to precede the occurrence of more sericus diseases such as leukemia and cancer. For children, the indicator diseases are asthma, urticaria, eczema, pneumonia, dysentery, and rheumatic fever. For adults, heart disease and several other diseases can play this role. The persons reporting prior indicator disease have a much higher risk of developing leukemia from low-level exposures than those who report no indicator disease.

17. Under NRC operating license specifications, light w:ter nuclear power plants are allowed to release radioactive effluents in amounts which will result in radiation doses to the public of 25-75 millirems each year. Ilow does this additional annual radiation exposure relate to the background radiation exposure?

Answer. Background radiation is generally taken as 100 millirem per year, although at particular locations, the actual figure may be somewhat higher or lower. The roughly 10-fold increase in leukemia with each decade of life is attributable, at least in part, to cumulative background exposure (which is directly proportional to age). If the excess radiation to the public is 50 millirem per year, this might be taken as roughly equivalent to aging 50 percent faster per year.

18. The NRC staff has concluded, regarding radiation emissions, that

" ...there will be no reasonable radiological impact on members of the public from routine operation of the station."* Ilow does this risk analysis gompare with the results of your research?

Answer. The risk analysis used by the NRC staff fails to use the current figures for health hazards of low-level radiation and does not take cumulative effects or synergistic effects from chemical pollution into account. Since the new risk estimates are 100 times greater than the ones 1

NRC uses, the cumulative effects are much greater than NRC recognizes l and the probable synergistic effects are much more serious, the NRC statement on radiological impact is at least questionable and in all likelihood is wrong.

t

• U. S. NRC, Final Environmental Statement related to the operation of Waterford SES, Unit No. 3, NUREG-0779, paragraph 5.9.1.2, p. 5-36.

l l

i }

19. Do you accept the biostatistical techniques and the risk analysis of the IIElit !!! report?

Answer. The llEllt III report is unacceptable since it completely ignores the quantitative estimates of radiation risks which can now be derived from biostatistical-epidemiological studies of populations actually exposed to low-level radiation. Extrapolations beyond the range of data is unacceptable from a statistical standpoint when there is actual data in the range, as there now is from more than 30 studies (Yale Symposium).

20. Can you describe the mechanism by which radiation and chemicals cause adverse health consequences 1 What is the operation of that mechanism?

Answer. Basically, the mechanism causing cancer and other effects to occur many years after the original chemical or radiation exposure is genetic damage to the DNA of human genetic material. This can be thought of as a " break point" or defect in the complex chemical structure of the double helix. The defect in the DNA represents misinformation which has little or no effect (so far as the whole organism is concerned) as long as it is confined to a single cell. For the whole-body economy to be affected, it is necessary that the misinformation be reproduced by cloning (approximately 32 doubling times are needed). This is the explanation for the long " latent period" between the initial damage and the clinical manifestation of this damage. Eventually the misinformation (which generally involves the manufacture or control of enzymes involved in the host defense system) can result in the deterioration of the host defense system. This, in turn, allows the damaged cells to eventually become metastatic cancer cells.

21. Is there any difference between the mechanisms by which chemicals and radiation cause these adverse health consequences?

Answer. Yes. The radiation damage is random or non-specific whereas chemical mutagens ordinarily attack the structure of the DNA only at very specific points.

22. Does it matter in terms of public health consequences whether chemical mechanisms or radiation mechanisms are in effect?

Answer. Although the mechanisms are different, the adverse health effects are similar. It probably does not matter greatly whether a particular site of damage is produced by a random radiation effect or a systematic chemical effect as long as there is permanent misrepair of the break that puts misinformation into the genetic structure of the DNA.

23. Ilow would the action of this mechanism be manifested in a population?

Answer. The genetic damage would not be immediately obvious because of the redundancy of biological systems; hence, current " target" theories asume that several break points are required to cause initiation of the cancer process (rather than a single break point). Ilowever, the damage cumulates in the sense that the genetic material of the population is degraded. Thus an increased proportion of the population will have multiple defects in their genetic material and their risks of cancer and other diseases are thereby increased. Suppose, for instance, hypothetically it takes 4 defects to produce cancer. If an individual had 3 pre-existing genetic defects, then it would take only 1 additional defect to initiate the cancer process. The. manifestation of the genetic damage of a population, therefore, is likely to be increased morbidity (e.g., indicator diseases) in

1 the population but not necessarily cancer. Eventually, however, cancer rates go up, as the frequency of persons in the " susceptible group" (e.g., 3 defects) increases and the low-level radiological and chemical exposures produces the additional break-point now needed to initate cancer. ,

24. In your view, is the health risk associated with this mechanism cumulative in a population from generation to generation?

Answer. Yes. As the successive generations are exposed to chemical or radiological mutagens, the proportion of the population in the susceptible group or n,cxt-to-susceptible damage categories builds up.

Thus, there is a cumulative effect.

25. Could this health risk be cumulative over the lifetime of an ind.vidual? What support do you have for this view?

Answer. The cumulative effect of background and other environmental exposures is reflected in a steady increase of cancer risks with age that were noted in Question 17. In a mutagenic environment, the risk that a cell in a susceptible individual will sustain the additional break point needed to initiate cancer is proportional to time and in this sense is cumulative.

26. Could you identify any category (ies) of individuals more likely than the rest of the population to demonstrate health effects from a cumulative risk?

Answer. As previously noted, there is a susceptible group (persons who probably had pre-existing genetic damage) that are more likely than the rest of the population to be affected by low-level radiation or other exposures. We cannot identify these persons positively by the genetic technology now available although we can distinguish these persons in a probability sense by their prior medical history.

r

27. What is the qualitative result of cumulative low level radiation exposure? (l.c., what, if any, diseases are associated with such exposures).

Answer. The list of the diseases is a long one and we do not know where it ends. Leukemia and lymphoma are clearly radiogenic. There are u

also a number of technogenic solid tumors, such as lpn.g cancer, bladder cancer and esophagus cancer. The indicator diseases are also radiogenic.

In general, it looks as though most of the diseases which are called " chronic diseases" are likely to be produced or promoted by mutagens in the environment.

28. Qualitatively, how does the health risk from low-level radiation exposure compare to the risk from relatively high level exposure?

Answer. Quantitatively, the answer to this question is given by the dosage response curve. According to recent evidence, the curve is far from linear. The current data suggests that the curve starts to level off at around 10 rem and is relatively flat for doses in, the vicinity of 100 rem,.

and is rciatively-flat for dcacs in the v;c;..ity of 100 scia, actually turning downward at even higher doses bec/ause the cells are sterilized and cannot clone. Qualitatively, this means that the risks for low-level radiation are not so very different from the risks for high-level radiation.

29. Given I,ouisiana's high cancer mortality rate due to chemical carcinogens present in the Mississippi River, such as chloroform, carbon tetrachloride, dimethylsulfoxide, benzene and others, and in the air between I3aton Rouge and New Orleans, i.e. halogenated hydrocarbons, can you state the nature of the risk to the population posed by the introduction of radiation in the one rad range into this

. i-l environment? Assume for this assessment a radiation dose to the i population of 25-75 millirems / year. ,

Answer. In view of the limitations of our current scientific knowledge on the synergistic effects between specific chemicals (such as those named in the question) and low-level ionizing radiation, I don't think it is possible to give any precise quantitative predictions of specific risks in the exposed poulation. It is, however, possible to make a rough qualitative assessment by extrapolating from the experience in the U.S.S.R., where there are conditions similar to those that would exist with the operation of Waterford Three.

30. Why are the U.S.S.R. conditions similar?

Answer. The policy of siting nuclear reactors on chemically contam-Inated rivers is virtually forced by the geography of th'e Soviet Union. For practical purposes, Russia is a landlocked country. The main water resources for chemical or nuclear plants are the long river systems. Since these plants require large amounts of water, the siting policy in the U.S.S.R. has been to string these plants like beads along these long river systems. This results in a build-up in chemical and radiological contam-ination downstream. Hence, many areas in the U.S.S.R. have been experiencing the conditions that would exist on, the lower Mississippi in Louisiana with the operations of Waterford Three.

31. Has this siting policy with a mix of chemical and nuclear plants along the Soviet Russian rivers had any adverse health effects on the population?

Answer. It seems to be having disastrous effects. In all of the technologically advanc:.ed nations (including the U.S.S.R.) there was a

9 declining infant mortality rate for many years, llowever, recently, in the U.S.S.R. these rates have turned around and are now rising rather rapidly.

The rates now about double U.S. rates. This was first reported by CIA statisticians but tvis since been confirmed by Russian statisticians (accord-ing to newspapcr reports).

32. Are there explanations other than contamination of the river waters for the increase in infant mortality rates in the U.S.S.R.? Why single out pollution?

Answer. There are always many post-hoc explanations for statistical facts and both the CIA and the Russian statisticians have given explanations other than pollution. While these explanations may sound plausible, the turn-around of a national rate requires some exposure to hazards on a national scale. Pollution is nationwide because of the siting strategy of the Communist technocrats and the high density of population along the river syste ms. Ilowever, attributing the turn-around to correction of underreporting in a remote province (the Russian explanation) or to vodka-drinking mothers (the CIA explanation) makes little epidemio-

~

logical sense.

33. Are there positive reasons for attributing some or most of the increased infant mortality to chemical-radiological contamination of the Russian ' river system?

Answer. Y es. Drinking water is the key to infant mortality. The climination of bacterial contamination was the key to the reduction in the mortality from infectious disease. To turn the U.S.S.R. rates around, there tvis to be a replacement of the bacterial contamination by technogenic contamination of the drinking water. l I

l l

l I

i

34. What could be predicted for the Waterford Three siting policy on the basis of the experience with Soviet siting policy?

Answer. First, an increase in infant mortality that would reflect the genetic damage from the chemical-radiological contamination. Second, an increase in deaths of children before adulthood due to the genetic damage.

Finally, an increase in the cancer rates for the adult population. These effects could occur from simple cumulative risks, but they would be greater if there are synergistic effcets. The rapid increase in Soviet infant mortality rates suggests that there well may be synergistic effects from the chemical-radiological pollution in the river systems. Clearly, the U.S.S.R. has adopted a dangerous siting policy which the U.S. can avoid because it has more siting options.

35. Is there any actual scientific evidence that would suggest that there may be synergistic effects for deaths at early ages in the children of persons exposed to radiation?

Answer. Yes. There is strong evidence in a recent report in Science on the children of persons who had been exposed to the Japanese A-bomb (Schull, W.J., Masanori, O, Neel J.V.: Genetic Effects of the Atomic Bombs: A Reappraisal. Science, Vol. 213, pp.1220-1227, September 11, 1981) In this case, of course, both parents were exposed to low-level ionizing radiation to the gonads (less than 10 rems) so it is not an example of synergism between chemicals and radiation. The report in Science found no statistically significant differences, but this was due to the use of a faulty statistical analysis. A straightforward analysis of the same data shows the clear evidence of synergism showing Graph I (See Appendix A, attached hereto.)

t

36. Can you explain Graph I?

Answer. Graph I (shown in Appendix A) demonstrates three things.

First, from the ranges of the 95% confidence intervals (shown as vertical 1

brackets), it is clear that the groups designated are distinct in terms of detectable effects in children from radiation exposure of their parents.

J Second, observing the horizontal dotted lines as the range of infant mortality among controls, it is also clear that only one group's percentage i

mortality falls wholely above the control range: the group in which both parents were exposed to 0-9 rems radiation.

The fact that infant mortality in this group is significantly elevated over that shown for exposures to father and to mother independently indicates a synergistic effect among children. Thirdly, it is important that the 95% confidence intervals for this zero-nine rems-to-both-parents group falls wholely within the upper segment of confidence intervals of the grot (>s in which parents were exposed to much higher combined levels of radiation. So this graph demonstrates that synergism results in greater infant mortality in a group exposed to lower doses of radiation than in those exposed to higher doses. Nor can this result be predicted from the groups in which only one parent was exposed to zero-nine rems.

37. Is there auf other evidence of synergism when both parents are exposed to radiation?

Answer. Yes. We had earlier shown that a similar phenomena occum with diagnostic x-rays where there can be exposure of either parent before pregnancy or exposure of mother and fetus during pregnancy. Certain combinations of exposures showed synergism (Bross, I.D.J., Natarajan, N.:

Cumulative Genetic Damage in Children Exposed to Preconception and Intrauterine Radiation. Investigative Radiology ,15 (1): 52-64, 1980).

1

38. Is there any evidence that both reproductive wastage such as infant mortality and cancer in adults can be produced by the same contam-ination?

Answer. Yes. The chemical contamination at Love Canal produced doubled risks of spontaneous abortion and of birth defects (Bross, I.D.J.:

Muddying the Water at Niagara. New Scientist, Vol. 88, No.1231, pp.

728-729, December 11,1980). In the same area, there is also excess cancer (Janerich, Burnett, Feck, .et al: " Cancer Incidence in the Love Canal A rea". Science, Vol. 212, pp.1404-1407, June 19,1981). Since both

, phenomena are due to genetic degradation, it is not surprising that they tend to go together. Ilowever, infant mortality shows up more quickly (9 months) than solid cancers (15 or more years).

39. Can you specify any subgroups within this South Louisiana population which might be special risk?

Answer. As noted in Question 26, there is a susceptible subgroup which is more likely to report indicator diseases than the general population, but it cannot be precisely identified by genetic markers.

40. Would introduction of radiation in the one rad range contained in liquid and gaseous and particulate emissions from the Waterford Three nuclear power plant aggravate this risk? By what mechansm is the risk enhanced?

Answer. There is now evidence from several studies that the doubling dose for myeloid leukemia in men is around 5 rem (See Yale Journal of Biology and Medicine, " Direct Estimates of Low-level Radiation Risks of Lung Cancer at Two NRC-compliant Nuclear Installations: Why are the New Risk Estimates 20 to 200 Times the Old Official Estimates,

Bross and Driscoll). It is likely that the persons affected by this low-level radiation are the susceptibles with pre-existing genetic damage. The emissions from Waterford Three could aggravate the risk. It should be i

noted that while 25-75 millirem may be an average under normal operating conditions, for a variety of reasons, the individual exposures may be substantially higher. Apart from accidentel releases, there are 2 tors in every system that concentrate as well as dissipate particulate radioactives.

In Pennsylvania, this occurred with cows eating grass downstream from the release. An average e,xposure is likely to be misleading because some people may not get any exposure and some may get 10 or 100 times this exposure.

41. Can you make a statement with regard to the health risk from low level radiation in emissions from Waterford Three as it impacts that portion of the population already at risk from pre-existing genetic damage as evidenced by " indicator diseases"?

Answer. For persons with pre-existing genetic damage as evidenced by " indicator diseases", etc., the risks of leukemia may be much more than doubled. In our studies of childhood leukemia (Bross, I.D.J. and Natarajan, N.: Genetic Damage from Diagnostic Radiation. JAMA, Vol. 237, pp.

2399-2401, May 30,1977), the risks of leukemia in the children where indicator diseases are reported were increased by factors of 10 or more.

42. Can you make a statement with regard to the doubling dose which would affect this population with pre-existing genetic damage (due to chemical carcinogens in the Louisiana environment)?

Answer. Since the persons with pre-existing genetic damage cannot be accurately identified, it is not possible to make a quantitative state-

.i ment about risks in this group. Ilowever, a doubling dose such as 5 rem involves averaging of risks over a population including these persons.

Therefc c, the doubling doses of these persons would, if anything, be substantially lower than 5 rem.

43. What are the stages or steps in the oncological proecss?

Answer. We now know with reasonable certainty the general steps and stages in the cancer process, although there are many details (e.g., of the time frame) which still have to be filled in. The first two steps in the process are initiation and promotion. The initiation of the cancer process occurs when the break-point is put into the DNA of human genetic material by a radiological or chemical process. This step is strictly one of physical science--physics and chemistry. Ilowever, nothing occurs clinically unless the second step, promotion, also takes place. It is during this step that the misinformation which is fixed in the genetic material, probably tiy misrepair of the lesion, is reproduced billions of times by cloning by the damaged cell. This is a biological process rather than a physical process.

! During this phase, the cells are under surveillance of the host defense' l

l system and their growth may be slowed or even aborted. While long-term effects on the host defense system are probably genetic, chemicals and i radiation can also produce immediate effects on the system. Both chemical and radiological insult is used, for instance, to knock out the host defense system of animals so that transplanted human cancer cells can be

used in animal studies. Af ter about 32 doubling times, the clone of damaged cells may be large enough to be clinically detected or to cause symptoms. The later steps in the cancer process include growth of the

primary tumor, local dissemination to the lymph nodes, generalized mestastases, and usually the death of the patient.

44. What roles do radiation and chemical agents play in the oncological process?

Answer. As noted in the previous question, radiation and chemical agents can initiate the oncological proecss by causing genetic damage.

They can also have direct effects on the host defense system which may promote cancer.

45. Are there any other mechanisms in which chemical agents and radiation worx together?

Answer. Animal studies (where the terms " initiation" and

" promotion" have a related but more specialized meaning) distinguish

, " complete" carcinogens from other carcinogens. A " complete" carcinogen can both initiate and promote whereas other carcinogens may do one or the other but not both. Radiation is a complete carcinogen and so is tobacco tar. Ilowever, radiation can also act together with a chemical initiator or a chemical promotor.

46. Is the damage from low-level radiation aggravated by excessive' levels of chemical carcinogens?

Answer. As explained in the previous question, chemical carcinogens can work jointly with radiation effects to produce the combination of initiation and promotion that is needed for the clinical manifestation of cancer.

47. Can you cite any incidence of populations which have.been exposed to risk factors (industrial chemical carcinogens und nuclear power plant emissions) similar to those which exist in south Louisiana with the

. r operation of Waterford Three? What has been the result as far as you know?

Answer. Two examples of populations exposed to risk factors, the Soviet river populations and the residents of Love Canal, have already been mentioned and the adverse health effects have been noted. Because the.

Niagara Falls Area has both chemical and radiological dumpsites, the high technogenic cancer rates in this area might possibly reflect some 4

synergistic action, but this is speculative. What is not speculative is that Niagara Falls is in the upper decile of U.S. counties for the technological cancers such as lung, bladder, and esophagus. My testimony of December 10,1981 ("Why the Assurances that the Water is ' Safe' Have No Scientific Validity") to the New York State Assembly Committee on Environmental Conservation dealt in more detail with these risks.

48. Does synergism exist or operate at low levels of exposure?

Answer. Synergism operates at low levels of exposure (and possibly more efficiently at these levels).

49. What happens to a piece of DNA that has been broken? Is the result a lasting one?

Answer. There is a repair process for break-points in DNA.

Ilowever, animal studies suggest that it is not a very accurate one.

Probably it is a misrepair of the break-point that puts permanerit misinformation into the DNA.

50. Do you know of any biostatistical models which date risk from existing genetic damage in a population to additional risk from radiation in the one rad range, with reference to first and second generation exposure in the same population?

2

i

• Answer. An example of this is given in Questions 35,36, and 37.

51. What is your assessment for the health risk to South Louisiana's population of the introduction of additional radiation in the one rad range resulting from plant operations at the Water Three nuclear generating facility?

Answer. As indicated by the previous answers, it is not possible to give a very precise quantitative assessment of the health risks to Southern Louisiana populations from the additional radiation in the 1-rad range produced by plant operations at Waterford Three, flowever, as is also suggested by preceding questions there is sufficient scientific knowledge about the cancer process, genetic damage, radiation risks, chemical hazards, and potential interactions of chemical and radiological hazards to make a qualitative assessment. In other words, there is sufficient scientific knowledge and past experience (primarily in the U.S.S.R.) to indicate that the policy of siting nuclear reactors on the lower Mississippi River could pose a major public health hazard to the population of Southern Louisiana.

In my view, this evidence is more than sufficient to establish a very strong prima facie case that the siting policy would enda er the public health and safety and at this point, I would appeal to the Primacy Principle: With possible technological hazards, the benefits must go to the public and not to the technology. This principle is discussed in my book, SCIENTIFIC STRATEGIES TO SAVE YOUR LIFE (Chapter 3).

In the U. S. (through not in the U.S.S.R.) there are viable alternatives to a policy of siting nuclear plants on a river with a heavy chemical burden already. Since these options exist for us, an application of the Primacy

i l

Principle indicates that it is clearly in the public interest to locate Waterford Three (or its equivalent) elsewhere. Indeed,I would add that the siting policy of putting nuclear plants on U. S. river systems should t,e reconsidered by NRC and this strategy climinated. Unless this is done, the ,

disastrous situation in the U.S.S.R., where the infant mortality rate is double that of the U. S. and is rapidly rising, could be the shape of things to come in the U.S.

O e

k

i STATE OF NEW YORK COUNTY OF b di d AFFIDAVIT BEFORE M E, the undersigned authority, duly authorized to administer oaths, there did appear DR. IRWIN D. J. BROSS, a person of full age and a resident of the county of 6/dl & , state of New York, who being duly sworn on oath, did state that the answers to questions presented herein are all true and correct to the best of his knowledge, information and belief; and that all such answers were prepared at his direction.

t DR. RMil(D. J. BRO VV SWORN TO AND SUBSCRIBED BEFORE ME, Tills # DAY OF //// b uj- ,1982.

> u VLt bl .) O'Y

N OTA R YgPUBLIC FOR Tile COUNfY OF_ 'r1GG ,

STATE OF NEW YORK. s m o u H. DAt:N 3 r York My commission expires at NeuphM.fl.l7j e

My Comm:nisa Impires Marsli 30. It l

l

Confidence Intervols for Infont and Childhood Mortolity by Potents Goncdol Dose 1

l 9.0 -

T 8.0 -

!  ? Ii i 7.0 o

3 -

.. = = ,,--

4>

tl il 5

$ 6.0 -

,1 ___ . ._,,__

o 5.O -

l 4.0 -

, T i i I i t i f I I -

reine, wic* wic* o-o wic* o-o to* o-o to* to*

u.en., wic* o-o wic* to* o-o wic* to

• o-o vo *

% c. m.,a ca, Gonadol Dose

_,,,,_,-w .

-w, g wN h * * * " " ' * #

a t'

e e=

4 t - -

CURRICULUM VITAE IRWIN D.J. BROSS, Ph.D.

Born: November 13, 1921, Halloway, Ohio, Male Married with 3 children EDUCATION:

Dates Degree Field UCLA, California , 1942 B.A. Mathematics North Carolina State, N.C. 1948 M.A. Experimental Statistics University of North Carolina, 1949 Ph.D. Experimental Statistics North Carolina POSITIONS HELD:

Associate in the Department Johns Hopkins Univ. 1971-Present of Epidemiology Acting Chief of Epidemiology Roswell Park Mem. Inst. 1966-1974 State University of 1961-Present Research Professor of Biostatistics N.Y. at Buffalo Director of Biostatistics Roswell Park Mem. Inst. 1959-Present Head, Research Design & Analysis Sloan Kettering Inst. 1952-1959 Assistant Professor for Public Cornell University 1952-1959 11ealth & Preventive Medicine Med. College Research Associate, Department Johns !!opkins Univ. 1949-1952 of Biostatistics .

MEMBERSIIIPS :

American Association for the Advancement of Science American Statistical Association Biometric Society Society for Epidemiologic Research Fellovr, American College of Epidemiology

4 BIBLIOGRAPIIY Dr. Irwin D.J. Bross

'82 fM910 Pi 39

1. Fiducial Intervals for Variance Components. Biometrics,6,(2)[:)

136-144, June 1950. ;h {j -

2. Study of the Influence of Sex of Donor on the Survival of Erythro-blastotic Infants Treated by Exchange Transfusion. (Coauthors:

• Milton S. Sacks, C.L. Spurling, Elsa F. Jahn). Pediatrics, 6,(5):772-777, November 1950.

3. Estimates of the LD50: A Critique. Biometrics, 6,(4):413-423, December 1950. ,

4. Two-Choice Selection. Journal of the American Statistical Association, 45,:530-540, December 1950.

5. Significance Tests for Certain 2x2 Tables. (Coauthor: M.M.

Delancy). The American Journal of Ilygiene, 55_(3):357-362, May 1952.

6. Sequential Medical Plans. Bi o' met ri cs , 8(3):188-205, September 1952. -

7. Introduction to the Theory of Games: Book Review. Journal of the American Statistical Association, 48,(263):655-657, September 1953.

8. Observations on an Outbreak of Infectious liepatitis in Baltimore During 1951. (Coauthors:

• Abraham M. Lilienfeld, Philip E.

Sartwell). The American Journal of Public IIcalth, 43,(9):1085-1096, September 1953.

9. DESIGN FOR DECISION. The Macmillan Company, New York, New York, 276 paEes, October 1953.

10. A Confidence Interval for a Percentage Increase. Biometrics, 10 (2):245-250, June 1954.

11. Is There an Increased Risk 7 Federation Proceedings, 13(3):815-819, September 1954.

.12. Looking Around: Statistical Decision Making, llarvard Business Review, 32,(5):133-147, September-October 1954.

13. Misclassification in 2x2 Tables. Bicmetrics, 10(4):478-486, December 1954.

• = Principal Author if other than Dr. Bross

i

14. Statistical Analysis of Clinical Results from 6-Mercaptopurine.

Annals of the New York Academy of Sciences, 60(2):369-373, December 6,1954.

15. Therapy for Intellectual Obesity or Common Sense in Reducing Figures. The American Journal of Obstetrics and Gynecology, 69, (2):372-377, February 1955.

16. Ilow to !!ake Decisions. The American Society for Quality' Control National Convention Transactions, pp. 457-466, May 1955.

17. Statistics and th'e Clinician. Annals of Internal Medicine, 43(2):442-450, August 1955,

18. Body Weight as a Factor in the Response of Mice to Botulinal Toxins. (Coauthors:

• Carl Lamanna, Wayne I. Jensen). The American Journal of flygiene, 62(1):21-28, July 1955.

19. A Study of Environmental Factors in Cancer of the Larynx.

(Coauthors: Ernst L. Wynder, Emerson Day). Cancer, 9(1):86-110, January-February 1956.

20. Epidemiological Approach to the Etiology of Cancer of the Larynx.

(Coauthors: Ernst L. Wynder, Emerson Day). Journal of the American Medical Association, 160:1384-1391, April 21, 1956.

21. Important Factors in the Evaluation of Medical Statistics.

Modern Medicine, 24(9):244-248, May 1,1956.

22. The Distribution of Sickness Disability in a !!omogeneous Group of

'llcalthy Adult Men'. (Coauthors:

• Lawrence L. Ilinkle, Jr. , Ruth

!!. Pinsky, Norman Plummer) . American Journal of liygiene, 64(2):

220-242, September 1956.

23. Ridit Analysis of Automotive Crash Injuries. (Coauthor: Rivkah Feldman). Division of Automotive Crash Injury Research, CUMC, October 15, 1956.

24. Lung Cancer in Women. A Study of Environmental Factors. (Co-authors: *Ernst L. Wynder, Jerome Cornfield, Walter E. O'Donnell).

New England Journal of Medicine, 255:1111-1121, December 13, 1956.

25. Rapid Analysis of 2x2 Tables. (Coauthor: Ethel L. Kasten).

Journal of the American Statistical . Association, ~52(227):18-28, March 1957.

26. The Significance of Delay in Relation to Prognosis of Patients ,

with Primary Operable Breast Cancer. (Coauthor:

• Guy F. Robbins).

Cancer, 10(2):338-344, March-April 1957.

i

2 7. . Lung Cancer in Women. Abstract in Modern Medicine, April 1, 1957.

28. Actiological Factors in Mouth Cancer, An Approach to Its Prevention.

(Coauthor: *Ernst L. Wynder) . British Medical Journal, i,:1137-1143, May 18, 1957. )

29. Environmental Factors in Cancer of the Upper Alimentary Tract.

A Swedish Study with Special Reference to Plummer-Vinson (Paterson-Kelly) Syndrome. (Coauthors: *Ernst L. Wynder, Sven lluitberg, Folke Jacobsson). Cancer, 10,(3):470-487, May-June 1957.

30. A Method for Evaluation of Laxative liabits in lluman Subjects.

(Coauthors:

• Theodore Greiner,llarry Gold). Journal of Chronic Diseases, 6,(3):244-255, September 1957.

31. Studies in lluman Ecology: Factors Relevant to the Occurrence of Bodily Illness' and Disturbances in Mood, Thought, and Behav.sr in Three llomogeneous Population Groups. (Coauthors:

• Lawrence E.

Ilinkle, Jr. , Norman Plummer, Rhoda Metraux, Peter Richter, John W. Gittinger, William N. Thetford, Adrian M. Ostfeld, Francis D.

Kane, Leo Goldberger, William E. Mitchell, llope Leichter, Ruth Pinsky, David Goebel, llarold G. Wolff). American Journal of Psychiatry, 114(3):212-220, September 1957.

f 32. A Study of the Etiologic Factors in Cancer of the Mouth. (Co-I authors: *Ernst L. Wynder, Rivkah Feldman). Cancer, 10(6):

~~~

1300-1323, November-December 1957.

33. Ilow to Use Ridit Analysis. Biometrics, 14(1):18-38, March 1958.

34. Sequential Clinical Trials. Journal of Chronic Diseases, 8,(3):

349-365, September 1958.

35. Evaluation of the Performance of Indices of Tumor Growth in the Testing of Tumor Growth Inhibitors Against the Nelson Ascites Tumor. ( Coauthor:

• George S. Tarnowski). Annals of the New York Academy of Sciences, 76,(3):586-600, December 5, 1958.

36. Statistical Problems in a Mass Screening Program. (Coauthors:

Paul Armitage, Marvin A. Schneiderman. Discussant: Irwin Bross).

Annals of the New Yo-k Academy of Sciences, 76(3):903-905, December 1958.

. 37. LA DECISION ESTADISTICA. Aguilar, Madrid, Spanish Translation,.

1958.

38. Note on an Application of the Schumann-Bradley Table. The Annals of Mathematical Statistics, 30(2):581-583, June 1959.

i s

39. The Significance of Early Treatment of Breast Cancer. (Coauthors:

• G.F. Robbins, J.W. Berg, C. DePadua, G.P. Sarmiento). Cancer, 12,(4):688-692, July-August 1959.

40. Ridit Analysis of the Effects of Carcinostatic Chemicals on the Growth Indices of the Nelson Mouse Ascites Tumor. (Coauthor:

• George S. Ta rnowski) . Cancer Research, 19,(6):581-590, July 1959.

41. Cancer and Coronary Artery Disease Among Seventh-Day Adventists.

(Coauthors: *Ernst L. Wynder, Frank R. Lemon). Cancer, 12,(5):

1016-1028, September-October 1959.

42. Controlled Clinical Trials for Cancer Chemotherapy. Chemotherapy of Cancer Syllabus, Sloan-Kettering Institute, VII-12-VII-28, October - November 6, 1959,

43. The Antiemetic Efficacy of Cyclizine (Marezine) and Triflupromazine ,

(Vesprin). (Coauthors: *J. Weldon Bellville, William S. Ilowland).

Anesthesiology, 20(6):761-766, November-December 1959.

44. A Method for the Clinical Evaluation of Antiemetic Agents.

(Coauthors: *J. Weldon Bellville, William S. Ilowland).

Anesthesiology, 20,(6):753-760, November-December 1959.

45. Statistical Criticism. Cancer, 13(2):394-400, March-April 1960.

46. Postoperative Nausea and Vomiting IV: Factors Related.to Post-operative Nausea and Vomiting. (Coauthors: *J. Weldon Bellville, William S. Ilowland). Anesthesiology, 21,(2):186-193, March-April 1960.

47. Postoperative Nausea and Vomiting III: Evaluation of the Antiemetic Drugs Fluphenazine (Prolixin) and Promethazine (Phenergan) and Comparison with Triflupromazine ( Vesprin) and Cyclizine (Marezine).

( Coauthors: *J. Weldon Bellville, William S. Ilowland). The Journal of the American Medical Association, 172:1488-1493, April 2, 1960.

48. A Study of the Epidemiology of Cancer of the Breast. (Coauthors:

• Ernst L. Wynder, Takeushi Iliryama). Cancer, 13,(3), May-June 1960.

49. A Rapid Quantitative Method for the Comparison of Diuretic Agents in Bed-Patients with Congestive Failure. (Coauthors:

• llarry Gold, Nathanici T. Kwit, Argyrios J. Golfinos). American-Journal of the Medical Sciences, 239(6)(43/665 - 58/680), June 1960.

50. DESIGN FOR DECISION. Ilokkaido, Japan, Japanese Translation, 1960.

~ . . m

! r. .

. I e

51. Ilow to Cut the liighway Toll in Ifalf in the Next Ten Years.

Public ifcalth Reports, 75,(7):573-581, July 1960.

t

, 52. Postoperative Nausea and Vomiting V: Antiemetic Efficacy of Trimethobenzamide and Perphenazine. (Coauthors: *J. Weldon i

Bellville, William S. Ilowland). Clinical Pharmacology and i Therspeutics, 1,(5):590-596, September-October 1960.

53. Outliers in Patterned Experiments: A Strategic Appraisal.

Technometrics, 3,(1):91-102, February 1961.

, 54. Tables published in - 6 AUSLAGE - DOCUMiNTA GEIGY - WISSENSCilAI'lLICilE TABLLEN, published by J.R. Geigy. Printed in Switzerland, pp.

170.27, 1961. (and END PAPER).

55. A Study of Etiological P~ actors in Cancer of the Esophagus.

(Coauthor: *Ernst L. Wynder) . Cancer, M (2):389-413, March-April 1961. -

56. PREVISION ET DECISIONS RATIONNELLES.

I Dunod, Paris, French Translation, 1961.

57. Food Additives. (Coauthors: *Morton L. Levin, Paul R. Sheche).

Science, 133:947-948, March 24, 1961.

58. Statistical Dogma: A Challenge. The American Statistician, 5(3):14-15, June 1961.

59. Latin Squares to Balance Immediate Residual, and Other Order, i E f fects. ( Coauthor:

• Paul R. Sheche). Biometrics, 17(3): 405-414, September 1961.

60. On a Graphical Sequential Test. Biometrics, 17(4):649-651, December 1961.

61. A New Approach to Differential Toxicity. (Coauthor: George S.

Ta rnowski) . Cancer Research, 22,(1):46-56, January 1962, i

62. Titration of the Effects of a Group of Selected Chemicals on the Growth of Ascites and Solid Forms of the Nelson Mouse Tumor.

(Coauthor:

• George S. Tarnowski) . Cancer Research, 22,(1):136-155, January 1962.

i

63. Symposium on Clinical Drug Evaluation and lluman Pharmacology:

Part V. An Ametric Approach to Bionssay. Clinical Pharmacology and Therapeutics, 3,(3):369-373, May-June 1962.

64. Discussion on the Foundations of Statistical Inference. (Discussants:

Allan Birmbaum and others). Journal of the American Statistical Association, E7(298):269-305, June 1962.

i.

i .

P

65. Tobacco and Health. (Letter to the Editor) . (Coauthors:

i Morton L. Levin, George E. Moore). Science, 137(3533):925, 926, 928, September 14, 1962.

66. Ilow Can Safety Measures Be Scientifically Evaluated? Passenger Car Design and liighway Safety. Consumers Union of U.S. , Inc. ,

pp. 12-24( et seq.), September 1962.

67. Symposium on Clinical Drug Evaluation and lluman Pharmacology:

Part VIII. An Ametric Approach to Bioassay. C.V. Mosby Company, St. Louis, pp. 49-53, 1962.

68. liighway Risks at Extreme Speeds. Public Health Reports, 78,(1):

27-33, January 1963.

69. Some Epidemiological Features of Malignant Solid Tumors of Children in the Buffalo, N.Y. Area. (Coauthors:

• Donald Pinkel, John E. Dowd). Cancer, 16,(1):28-33, January 1963.

70. Ilow Can Safety Measures Be Scientifically Evaluated? Traffic Sa fety, 62(2):16-18, February 1963.

71. Linguistic Analysis of a Statistical Controversy. The American Statistician, 17(1):18-21, Februa y 1053.

12. PREVENTIVE MEDICINE AND PUBLIC llEALTH. Chapter 3 " Statistics in Public ilcalth and Medical Research". (Author: W.G. Smillie; Revised by: E.D. Kilbourne). 3rd ed. The Macmillan Co., New York, pp. 36-73, 1963.

73. Key Problens in Clinical Testing. (Name of Dr. Bross withheld on publication). Interagency Coordination in Drug Research and Regulation. IIcaring before the Subcommittee on Reorganization and International Organizations of the Committee on Government Operations United States Senate, Eighty-Eighth Congress. First Session. Part 4, pp. 1632-1633, March 21, 1963-

74. Linguistic liabits of Scientists. Perspectives in Biology and Medicine, VI(3):322-346, Spring 1963.

i

75. Futility of liigh-Speed Driving. 80 m.p.h. Speed Sharply Increases Accident Risk with Little Saving in Time. Traffic Digest and Review, 11,(6) :16, June 1963.

76. Studies on flyaline Membrane Disease: I. The Fibrinolysin System in Pathogenesis and Therapy. (Authors: *Clara M. Ambrus, et al.) (Appendix A by Dr. Irwin Bross). Pediatrics, 32,(1):24, July 1963.

i 9

77. Comments on High Speeds. Public Health Reports, 78,(9):821-822, September 1963. i

78. Studies on Topical Chemotherapy of Tumor-Contaminated Surgical Wounds. (Coauthors: *Faud A. Mukhtar, F. Stanicy Hoffmeister, Joe DiPaolo, Paul R. Sheche, Hans Wilkens, George E. Moore). The American Journal of Surgery, 106(5):768-776, November 1963.

79. Statistical and Inductive Probabilities. (Book Review of Hugues Leblanc). Journal of the American Statistical Association, 58,(304):1171-1173, December 1963.

80. Comments on Highway Speeds. Public Health Reports, 79(1): 87-90, January 1964.

81. Prisoners of Jargon. American Journal of Public Health, 54(6):

918-927, June 1964.

82. Highway Risks at Extreme Speeds. (Spanish title: Los Peligros De La Velocidad Excesiva En Las Carreteras*) Boletin de la Oficina Sanitaria Panamericana, LVII(1):55-62, June 1964.

83. Is It Really Wrong? The American Statistician, 18(3):21-22, June 1964. .

84. Comparison of Reserpine and 20-Methoxydeserpidine by a New Method for the Assay of Antihypertensives in Ambulant Patients.

(Coauthors: *N.T. Kwit, D. Mehta, H. Gold, A.J. Golfinos, W.N.

Zahm, E.A. Goessel). Journal of New Drugs, 4(4):200-205, July-August 1964.

85. Taking a Covariable into Account. Journal of the American Statistical Association, 59(307):725-736, September 1964.i

86. Blood Groups in Patients with Multiple Cancers. (Coauthors: *Y.

Tsukada, R.H. Moore, J.W. Pickren, E. Cohen). Cancer, 17(10):

1229-1232, October 1964.

87. Diuretic Effect of Polythiazide and Sodium Mera11uride: Comparison in Bedfast Patients with Edema. (Coauthors: *H. Gold, N.T.

Kwit, D. Mehta). Journal of the American Medical Association, 190(7):571-574, November 1964.

88. Incidence of Second Primary Tumors in Children with Cancer and Leukemia: A Seven-Year Survey of 150 Consecutive Autopsied Cases. (Coauthors:

• William Regelson, Juliet Hananian, Goryun Nigogosyan). Cancer, 18,(1):58-72, January 1965.

89. Nill Career Awards. (Letters) Science, 147_(3664):1395, March 1965.

- 90. BESLU11YORMING EN STATISTIEK. Marka-Bocken, Antwerpen, Dutch Translation of DESIGN FOR DECISION,1965. .

91. DESIGN FOR DECISION. A Free Press Paperback, The Macmillan Company, 1965. .

92. JAK PODEJMOVAC DECYZJE. Panstwowe Wydawnictwo Naukowe, Warsaw, Polish Translation of DESIGN FOR DECISION,1965.

93. Book Review of " Stochastic Models in Medicine and Biology".

Proceedings of a symposium conducted by the Mathematics Research Center. Quarterly Review of Biology, 40(3):327-328, September 1965.

94. Synchronous Tumors in Patients with Multiple Primary Cancers.

(Coauthors: *R. Moore, Y. Tsukada, W. Regelson, J.W. Pickren).

Cancer, 18,(11):1423-1430, November 1965.

95. Clinical Factors in Smoking Withdrawal: Praliminary Report.

(Coauthors: *Arlene L. Plakun, Julian Ambrus, Saxon Graham, Morton L. Levin, Charles A. Ross). The American Journal of Public Ilealth, S6,(3):434-441, March 1966.

96. Electronic Computers for Evaluation of Clinical Data. III.

Interpretation of Data. Cancer Chemotherapy Reports, 50,(5): 157-161, March 1966.

97. BESLUTS PLANERING. Bokforlaget Aldus/Bonniers, Stockholm, Swedish Translation of DESIGN FOR DECISION,1966.

l l

98. Development of Lower Genital Carcinomas in Patients with Anal Carcinoma - A More Than Casual Relaticnship. (Coauthors: *A.

l Cabrera, Y. Tsukada, J.W. Pickren, R. Moore) . Cancer, 19,(4):470-480, April 1966.

99. What is an Objective Response? (Coauthors: *A. A. Rimm, J. K.

Ahlstrom) . Proceedings of the American Association for Cancer Research, 7,:59, April 1966.

Spurious Effects From an Extrancous Variable. Journal of Chronic l 100.

Diseases, 19,:637-647, June 1966.

101. Is Toxicity Really Necessary? I. The Question. (Coauthors:

A. A. Rimm, N.ll. Slack, R.K. Ausman, R. Jones, Jr.) . Cancer, 19, (12):1780-1784, Deceniber 1966. .

102. Is Toxicity Really Necessary? II. Source and Analysis of Data.

(Coauthors: A. A. Rimm, N.ll. Slack, R.K. Ausman, R. Jones, Jr.) .

' Cancer, 19,(12):1785-1795, December 1966.

9 103. Is Toxicity Really Necessary? III. Theoretical Aspects. (Co-authors: A. A. Rimm, N.ll. Slack, R.V. Ausman, R. Jones, Jr.) .

Cancer, 19,(12):1796-1804, December 1966.

104. Letter to the Editor on the Bayesian Pendulum. American Statistician, 21,(1):37-38, February 1967, 105. Tar and Nicotine Retrieval From Fifty-Six Brands of Cigarettes.

(Coauthors: *G.E. Moore, R. Shamberger, F.G. Bock). Cancer, 20,(3):323-332, March 1967.

106. Which Breast Cancer Patients llave Faster Growing Tumors? (Co-authors: *R.L. Priore, L.E. B;umenson, N.ll. Slack). Proceedings of the American Association for Cancer Research, 8,:55, March 1967.

107. Chemotherapy and a Two-Disease Model of Breast Cancer. (Co-author's : L.E. Blumenson, N.l!. Slack, R.L. Priore). First Tenovus Symposium on Factors Influencing Prognosis of Breast Cancer, p. 19, 1967 (Abstract).

108. Adjuvant Chemotherapy for Cancer of the Corpus Uteri (Preliminary Report). (Coauthors: *G.C. Lewis, Jr. , S.J. Nadler, N.ll. Slack) .

Obstetrics and Gynecology, 29,(6):797-902, June 1967.

109. Long-Term Drug Dangers. (Letter to the Editor) . Science, 156(3782):1552-1553, June 1967.

110. Pertinency of an Extraneous Variable. Journal of Chronic Diseases, 20,:487-495, 1967.

111. Experiment--or Stagnate. (Letter to the Editor) . New York Times Magazine, p. 4, July 1967.

112. Tar and Nicotine in Cigarettes. (Abstract of " Tar and Nicotine Retrieval from Fifty-Six Brands of Cigarettes"). Modern Medfcine, 35_(17):64, August 1967.

113. What is Objective Tumor Response? (Coauthors: *A.A. Rimm, J.K.

Ahlstrom). Cancer, 20,(8):1326-1334, August 1967.

114. Letter to the Editor on Fiscal vs. Scientific Criteria. Saturday Review, pp. 104-105, January 1968.

115. Chromosomes and Causation of Iluman Cancer and Leukemia. IV.

Vectorial Analysis. (Coauthors: *A.A. Sandberg, N. Takagi, M.L.

j Schmidt). Cancer, 21(1):77-82, January 1968.

}

l

. __ -.,._.___ _ , _ - _ - _ _ _ _ _ _ _ _ _ _ . _ _ . , - . __ _ _ ~. __ _

t

= l l

116. Design of Future Studies of Surgical Adjuvants in Breast Cancer.

(Coauthors: N. Slack, L. Blumenson, R. Priore). Proceedings of the American Association for Cancer Research, 9,:9, }farch 1968, (Abstract).

117. Recognition of Syntactic Structure by Computer. (Letter to the Edi tor) . Nature, 217(5133):1078, March 1968.

118. What is Pain? A Scientific Approach to Questions of Psycho-pharmacological Language. In PSYC110 PHARMACOLOGY: DIMENSIONS AND PERSPECTIVES. Edited by C.R.B. Joyce for Tavistock Publications, J.R. Lippincott Company. (London, England),1968.

119. DESIGN FOR DECISION. Published in Slavak by Slovenske Vydavatel

'stvo Technickej Literatury in Bratislava, Czechoslovakia, 1968.

120. ' Basic Science' Under Military Control. (Letter to the Editor).

New York Times, May S, 1968.

121. Effect,of Filter Cigarettes on the Risk of Lung Cancer. National Cancer Institute Monograph, (28):35-40, 1968.

122. Chronology and Pattern of Iluman Chromosome Replication. IX.

Metasynchronous DNA Replication in Homologs. (Coauthors: *A.A.

Sandberg, N. Takagi, M.L. Schmidt) . Cytogenetics, 7: 298-332, 1968.

123. A Syntactic Formula for English Sentences: Application to Scientific Narrative. Computers and Biomedical Research,1(6):

565-578, June 1968.

124. A Two-Disease Model For Breast Cancer. (Coauthors: L.E. Blumenson, N.H. Slack. R.L. Priore) . In PROGNOSTIC FACTORS IN BREAST CANCER. (Proceedings of First Tenovus Symposium, Cardiff 12th-14th April 1967). Edited by A.P.M. Forrest and P.B. Kunkler, for E. 6 S. Livingstone Ltd., Edinburgh and London, pp. 288-300, 1968.

125. A Controlled Randomized Comparative Study of Early and Late Adrenalectomy in Women with Advanced Breast Cancer. (Coauthors:

• T.L. Dao, T. Nemoto). In PROGNOSTIC FACTORS IN BREAST CANCER.

l (Proceedings of First Tenovus Symposium, Cardiff 12th-14th April

! 1967). Edited by A.P.M. Forrest and P.B. Kunkler, for E. 6 S.

Livingstone Ltd., Edinburgh and London, pp. 177-185, 1968.

126 Risks of Lung Cancer in Smokers Who Switch to Filter Cigarettes.

(Coauthor: R. Gibson). American Journal of Public Health, 58 (8):1396-1403, August 1968.

"" d - -

-m- - - . . _

i 127. Book Review of " Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability". Cancer Research, i 28(9):1916, September 1968.

128. Leukemia in Children Exposed to Multiple Risk Factors. (Coauthors:

• R.W. Gibson, S. Graham, A.M. Lilienfeld, L.M. Schuman, M.L.

Levin, J.E. Dowd) . New England Journal of Medicine, 279(17):906-909, October 1968.

129. Effects of 5-Fluorouracil (NSC-19893) in 389 Patients with Cancer (Eastern Clinical Drug Evaluation Program). (Coauthors: *G.E.

Moore, R. Ausman, S. Nadler, R. Jones, Jr. , N. Slack, A. A. Rimm).

Cancer Chemotherapy Reports (Part 1), 52(6):641-653, October 1968.

130. Effects of 6-Mercaptopur'ine (NSC-755) in 290 Patients with Advanced Cancer (Eastern Clinical Drug Evaluation Program).

(Coauthors: *G.E. Moore, R. Ausman, S. Nadler, R. Jones , Jr. , N.

Sl ack , A. A. Rimm) . Cancer Chemotherapy Reports (Part 1), 52,(6):

655-669, October 1968.

131. Effects of Chlorambucil (NSC-3088) in 374 Patients with Advanced Cancer (Eastern Clinical Drug Evaluation Program). (Coauthors:

• G.E. Moore, R. Ausman, S. Nadler, R. Jones, Jr. , N. Slack, A. A.

Rimm). Cancer Chemotherapy Reports (Part 1), 52_(6):661-666, October 1968.

132. Effects of Meturedepa (NSC-51325) in 233 Patients with Advanced Cancer (Eastern Clinical Drug Evaluation Program). (Coauthors:

G.E. Moore, R. Ausman, S. Nadler, R. Jones, Jr. , N. Slack, A. A.

Rimm) . Cancer Chemotherapy Reports (Part 1), 52,(6):667-673, October 1968.

133. Effects of Mitomycin C (NSC-26980) in 346 Patients with Advanced Cancer (Eastern Clinical Drug Evaluation Program). (Coauthors:

• G.E. Moore, R. Ausman, S. Nedler, R. Jones, Jr., N. Slack, A.A.

Ri mm) . Cancer Chemotherapy Reports, (Part 1), 52,(6):675-684, October 1968.

134. Carcinoma of the Bladder, 5-Fluorouracil and the Critical Role of a Placebo. ' A Cooperative Group Report I'. (Coauthors: *G.R.

I Prout, Jr. , N.H. Slack, R.K. Ausman) . Cancer, 22,(5): 926-931, November 1968.

135. Cigarette Smoking and Periodontal Disease. (Coauthors: *H.

Solomon, R.L. Priore). Journal of the American Dental Association, -

77,(5):1081-1084, November 1968.

136. Letter to the Editor on Peck Order in the Sciences. American Statistician, 22,(5):37-38, December 1968.

i e

w- +.e .--- ,r--+ ,-y ,, , --- J - - - , m-+ - - , -- --

t 1 .

f 137. Statistical Testing of a Deep Mathematical Model for lluman Breast Cancer. (Coauthor: L.E. Blumenson). Journal of Chronic Diseases, 21:493-506, December 1968.

138. Relation Between Aortic Atherosclerosis and the Use of Cigarettes and Alcohol: An Autopsy Study. (Coauthors: *D.L. Sackett, R.W.

Gibson, J.W. Pickren) . New England Journal of Medicine, 279(26):

1413-1420, December 1968.

139. Ridit Analysis. (Anonymous). New England Journal of Medicine, 279(26):1453-1454, December 1968, 140. A Mathematical Analysis of the Growth and Spread of Breast Cancer. (Coauthor: *L.E. Blumenson). Biometrics, 25,(1): 95-109, March 1969. .

141. Applications cf Probability: Science vs. Pseudoscience. Journal of the American Statistical Association, 64(325):51-57, March 1969.

142. Feasibility of an Automated Reporting-And-Inquiry System With All Transactions in the Natural Language of the User. (Coauthors:

R.L. Priore, P.A. Shapiro, D.F. Stermole, B.B. Anderson).

Proceedings of the ,American Association for Cancer Research, 10,:9, March 1969. (Abstract) 143. Information in Natural Languages: A New Approach. (Coauthors:

• P.A. Shapiro, R.L. Priore, B.B. Anderson). Journal of the American Medical Association, 207(11):2080-2084, March 1969.

144. Dunce Caps for Giant Brains. The New York Statistician, 20(4):

1-2, March-April 1969.

145. Ilow Case-For-Case Matching Can Improve Design Efficiency. The American Journal of Epidemiology, 89(4):359-363, April 1969.

146. Role of Placebo, 5-FU in Study of Bladder Cancer. (Coauthors:

• G.R. Prout, N.ll. Slack, R.K. Ausman) . Modern Medicine, 37,(7):

142, April 1969. (Abstract) 147. A Statistical Solution to an Ethical Problem. (Coauthors: M.L.

Schmidt, T.L. Dao, T. Nemoto). Clinical Pharmacology and Therapeutics, 10(3):307-313, May-June 1969.

148. Feasibility of Automated Information Systems in the Users' Natural Language. (Coauthors: R.L. Priore, P.A. Shapiro, D.F.

Stermole, B.B. Anderson). American Scientist, 57(2):193-205, Summer 1969.

149. Book Review of "Information Retrieval Systems: Characteristics, Testing, and Evaluation", by F.W. Lancaster. American Scientist, 57(2):124A-125A, Summer 1969.

150. Location of Breast Cancer and Prognosis. (Coauthors: *B. Fisher, N.H. Slack, R.K. Ausman (and Cooperating Investigators)) .

Surgery, Gynecology and Obstetrics, 4,:705-716, October 1969.

151. Cancer of the Breast: Size of Neoplasm and Prognosis. (Co-authors: *B. Fisher, N.H. Slack (and Cooperating Investigators)).

Cancer, 24(5):1071-1080, November 1969.

152. Letter to the Editor on Requiring Proof of Efficacy and Safety for all Therapy. American Jaurnal of Public Health, 59,(11):

1978-1979, November 1969.

153. Therapeutic Implications From a Mathematical Model Characterizing the Course of Breast Cancer. (Coauthors: *N.H. Slack, L.E.

Blumenson). Cancer, 24(5):960-971, November 1969.

154. What is Objective Tumor Response? (Coauthors: *A. A. Rimm, J. K.

Ahlstrom). In YEAR BOOK OF CANCER, pp. 357-359. Year Book Medical Publishers, Inc., Chicago, 1969.

155. Changes in Fibrin-Stabilizing Factor Levels in Relation to Maternal llemorrhage and Neonatal Disease. (Coauthors: *C.M.

Ambrus, J.L. Ambrus, K.R. Niswander, D.ll. Weintraub, H.B. Lassman) .

Pediatric Research, 4(1):82-88, 1970.

156. DESIGN FOR DECISION. Kodan Sha Ltd., Tokyo, Japan, Japanese Translation, 1970.

157. The Role of Mathematical Models in Clinical Research. The American Statistician, 24,(1):53-56, February 1970.

158. Cats and Childhood Leukemia. (Coauthor: R. Gibson). Proceedings of the American Association for Cancer Research,11:12, March 1970, (Abstract).

159. Cats and Childhood Leukemia. (Coauthor: R. Gibson). Journal of Medicine, 1(3):180-187, 1970.

160. Statistical Criticism. In Tile QUANTITATIVE ANALYSIS OF SOCIAL PROBLEMS, Edward R. Tufte, ed. , pp.97-108. Addison-Wesicy Publishing Co., Massachusetts, 1970.

161. A Model System for Detecting Drug Impairment of Antitumor llost De fenses. (Coauthors: *E. Mihich, R.M. Mihich, C.A. Nichol).

Cancer Research, 10:1376-1383, May 1970.

7 162. Irradiation and 5-Fluorouracil as Adjuvants in the Management of Invasive Bladder Carcinoma. A Cooperative Group Report After Four Years. (Coauthors: *G.R. Prout, Jr. , N.H. Slack) . Journal of Urology, 104,:116-129, July 1970.

163. Evaluation of Survivors of Respiratory Distress Syndrome at 4 Years of Age. (Coauthors: *C.M. Ambrus, D.H. Neintraub, K.R.

Niswander, L. Fischer, J. Fleishman, J.L. Ambrus). American Journal of Diseased Children, 120:296-302, October 1970.

164. Prof Says Computers Have Been Flops in Science. Reporter, 2, (21):4, February 1971. (Abstract).

165. Does Illness in Pets Increase Risk of Human Leukemia? (Coauthors:

R. Berte11, R. Gibson). American Association for Cancer Research, 12;8, March 1971 (Abstra'ct).

166. Prognostic Factors Affecting Adrenalectomy in Patients with Metastatic Cancer of the Breast. (Coauthors: *M.L. Schmidt, T.

Nemoto, T. Dao). Cancer, 27(5):1106-1111, May 1971.

167. Preoperative Irradiation as an Adjuvant in the Surgical Manage-ment of Invasive Bladder Carcinoma. (Coauthors: *G.R. Prout, N.H. Slack). Journal of Urology, 105:223-231, 1971.

168. Analysis of Prognostic Factors in Patients with Primary Breast Carcinoma. ( Coauthor: *N.H. Slack). Journal of Medicine, 2;93-111, 1971.

169. Letter to the Editor. Science, 172(3982):426, April 1971.

(Abstract). .

170. Letter to the Editor. Open Access Archives. The American Statistician, 25(2):36f, April 1971. ( Abstract) .

171. Previous Tonsillectomy end the Incidence of Acute Leukaemia of Childhood. (Coauthors: *A.I. Freeman, N. Lieberman, J. Tidings, O. Glidewell). Lancet, p. 1128, May 1971.

172. Application of a Mathematical Model to a Clinical Study of the Local Spread of Endometrial Cancer. (Coauthors: *L.E. Blumenson, N.H. Slack). Cancer, 28,(3):735-744, September 1971.

173. Another Look at Bode's Law. " Comment'. Journal of the American Statistical Association, 66,(335):562-564, September.1971.

174. Some Epidemiological Clues in Thyroid Cancer: Tonsillectomy, Acne, Allergy, Ethnicity. (Coauthors: K. Shimaoka, J. Tidings).

Archives of Intern. Med. , 128:755-760, November 1971.

m

175. Predictive Design of Experiments Using Deep Mathematical Models.

(Coauthor: L.E. Blumenson). Cancer, 28,(6):1637-1646, December 1971.

176. Letter to the Editor: ' Mission Scientist Rebuts Academicians'.

Biomedical News II, p. 2, December 1971.

177. Critical Levels, Statistical LangJage, and Scientific Inference.

FOUNDATIONS OF STATISTICAL INFERENCE (A SYMPOSIUM). Edited by V.P. Godambe and D.A. Sprott. Holt, Rinehart and Winston of Canada, Limited, pp. 500-519, 1971.

178. Binding of Positively Charged Particles to Glutaraldehyde-Fixed Human Erythrocytes. (Coauthors: *L. Weiss, R. Zeigel, O.S.,

Jung) . Experimental Cell Research, 70,:57-64, 1972.

179. Accessibility of Archives. Biomedical Computing, 3,(2):113-122, 1972.

180. Scientific Strategies in Human Affairs: Use of Deep Mathematical Models. Transactions of the New York Academy of Sciences, Series 11, 34(3):187-199, March 1972.

181. Low Level Radiation and, Leukemia: Identification of Susceptible Children. American Association of Cancer Research,13,:2, March 1972 (Abstract).

182. How Information is Carried in Scientific Sublanguages. (Co-authors: P.A. Shapiro, B.B. Anderson). Science, 176(4041):

1303-1307, June 1972.

183. Is There an Increased Risk of Metastasis? Surgery, Gynecology, and Obstetr cs, 134:1000, June 1972.

184. Leukemia from Low-Level Radiation: Identification of Susceptible Children. (Coauthor: N. Natarajan). New England Journal of Medicine, 287(3):107-110, July 1972.

185. Predicting Survival Following Surgery for Bronchogenic Carcinoma.

(Coauthors: *N.H. Slack, A. Chamberlain). Chest, 62:433-438, October 1972.

186. Testicular Tumor: Histologic and Epidemiologic Assessment.

(Coauthors:

• K.D. Sharma, J.F. Gaeta, R.H. Moore, G.P. Murphy) .

New York State Journal of Medicine, pp. 2421-2425, October 1972.

187. Pets and Adult Leu'kemia. (C6 authors: R. Berte11, R. Gibson).

American Journal of Public Health, 62(11):1520-1531, November 1972.

I 188. Letter to the Editor: 'A Cancer Research Policy'. The Relevant

Scientist, 2

33-34, November 1972.

189. Self-Criticism and Public Confidence. The Sciences. (Point of View), g(10):28, December 1972.

190. Note on Lung Cancer Among Cigar and Pipe Smokers. (Coauthor:

• J. Tidings). Preventive Medicine, 1,(4):543-546, December 1972.

191. Leukemia and Occupations. (Coauthor: *E. Viadana). Preventive Medicine, 1,(4):513-521, December 1972.

192. An Autopsy Study of Metastatic Sites of Breast Cancer. (Coauthors:

• E. Viadana, R. Cotter, J.W. Pickren). Cancer Research, 33,:179-181, January 1973. .

193. Experiences with Bilateral Primary Carcinoma of the Breast.

(Coauthors: *N.H. Slack, T. Nemoto, B. Fisher). Surgery, Gynecology, and Obstetrics, 36:433-440, 1973.

194. Use of a Mathematical Model to Bridge the Clinic-Laboratory Gap:

Local Spread of Endometrial Cancer. (Coauthor: *L. Blumenson).

Journal of Theoretical Biology, 38,:397-411, 1973.

The American 195. Letter to the Editor: 'The Tices Data-Input System'.

Statistician, j27(1):39, j February 1973.

196. Letter to the Editor: ' Tonsillectomy and Hodgkin's Disease'.

(Coauthors: *K. Shimaoka, J. Tidings). New England Journal of Medicine, 288(12):634-635, March 1973.

197. Position Paper on Health Hazards of Small Cigars. United States Congressional Record-Senate, pp. S4360-S4361, March 1973. (Abstract).

198. Letter to the Editor: Dr. Bross Replies. The Sciences, 13,(4):

3, May 1973.

199. Trans-Science and Responsibility. Science, 180(4091):1122-1124, June 1973.

200. Letter to the Editor: The Rush Toward Atomic Energy: Safety First. The New York Times, July 1973.

201. An Autopsy Study of Some Routes of Dissemination of Cancer of the Breast. (Coauthors: *E. Viadana, J.W. Pickren) . British Journal of Cancer, 27,:336-340, 1973.

202. Unobtrusive Biomedical Data-Input Systems. (Coauthors: A.

Gregory, C. Rathod, D. Bross) . Bio-Medical Computing, (4) : 219-228, 1973.

7 203. How Informa' tion is Carried in Scientific Sub-LanguaEes. (Coauthors:

P.A. Shapiro, B.B. Anderson). Polish Translation. Science,t 1(33), 1973.

204. Languages in Cancer Research. ' Perspectives in Cancer Research, Chapter 19, pp. 213-221, 1973.

205. Computer Assisted Discourse Analysis of Jargon. (Coauthor: D.

Stermole). Computer Studies in the Humanities G Verbal Behavior, 4_(2) : 65-76, August 1973.

206. Another Look at Coffee-Drinking and Cancer of the Urinary Bladuer.

(Coauthor: J. Tidings). Preventive Medicine, 2,:445-451, 1973.

207. Preconception Radiation and Leukemia. (Coauthor: *N. Natarajan).

Journal of Medicine, 4_:276-281, 1973.

208. Leukemia from Low-Level Radiation: Identification of Susceptible Children. (Coauthor: N. Natarajan). Year Book of Cancer, 1973.

(Abstract). .

209. Experiences With Bilateral Primary Carcinoma of the Breast.

(Coauthors: *N.H. Slack, T. Nemoto, B. Fisher) . Ycar Book of Cancer, 1973. (Abstract).

210. Preoperative Irradiation and Cystectomy for Bladder Carcinoma, IV: Results in a Sciccted Population. (Coauthors: *G.R. Prout, N.H. Slack). Seventh National Cancer Conference Proceedings, pp.

783-791, 1973.

211. Letter to the Editor: Cars in the Fuel Crisis: Curb Horsepower.

The New York Times, Sec. 1, p. 42, October 2, 1973. -

212. Plasminogen in the Prevention of Hyaline Membrane Disease.

(Coauthors: *C.M. Ambrus, D.H. Weintraub, T.S. Choi, B. Eisenberg, H.P. Staub, N.G. Courey, R.J. Foote, D. Goplerud, R.V. Moesch, M.

Ray, O.S. Jung, I .B. Mink, J.L. Ambrus) . American Journal of Diseased Children, 127:189-194, February 1974.

213. Carcinogenic Activity of Smoke Condensate From Cigarettes with Ammonium Sul famate-Treated Paper. (Coauthors: *F.G. Bock, I. .

Michelson, R.L. Priore). Cancer, 33,(4):108-114, April 1974.

214. Letter to the Editor: Should Journals Publish Extended Abstracts Only? The American Statistician, May 1974. .

215. An Epidemiological Study of the Relationship of Reproductive Experience to Cancer of the Ovary. (Coauthors: *D.J. Joly, A.M. Lilienfeld, E.L. Diamond) . American Journal of Epidemiology, 99:190-209, 1974.

216. Letter to the Editor: ' Adversary Science in Aliquippa'. llealth i Physics, 26,:581-583, June 1974.

217. Efficacy of Adrenalectomy in Treatment of Patients with Metastatic Cancer of the Breast. (Coauthors: *A. Chamberlain, T. Dao, T.

Nemoto, N.H. Slack). Surgery, Gynecology G Obstetrics, 138:891-895, June 1974.

218.

• Strategies of Science in lluman Affairs: liow to Assure the Quality of Scientific Statements. Clinical Pharmacology and Therapeutics,

_15(6):543550, June 1974.

219. Letter to the Editor: Nuclear. The New York Times, July 7, 1974.

220. Relation Between Benign. Prostatic ilyperplasia and Cancer of the Prostate. (Coauthors: *ll.K. Armenian, A.M. Lilienfeld, E.L.

Diamond). The Lancet, pp. 115-117, July 20, 1974.

221. I.etter to the Editor: Speed Lobby. Buffalo Evening News, September 6, 1974.

222. The Role of the Statistician: Scientist or Shoe Clerk. The American Statistician, November 1974.

223. Abstracts and the Referceing Process (Abstract). The American Statistician, 28_(2):75, 1974.

224. Use of the Medical flistory to Predict the Future Occurrence of

• Leukemias in Adults. (Coauthor: *E. Viadana). Preventive Medicine, 3,:165-170, 1974.

225. Direct input of Toxicologic Data to the Computer. (Coauthor:

• A.R. Gregory). Clinical Toxicology, 7(4):395-400, 1974.

226. Risk of Leukemia in Susceptible Children Exposed to Preconception,

In Utero and Postnatal Radiation. (Coauthor

N. Natarajan).

Preventive Medicine, 3:361-369, 1974.

227. Letter to the Editor: Approach to Societal Problems. Science, 186:478, 1974.

228. Point of View: Professional Ethics: Organizations vs. Individuals.

The Sciences, p. 29, November 1974.

, - 229. Adjuvant Progestogen Therapy in the Primary Definitive Treatment of Endometrial Cancer. (Coauthors: *G.C. Lewis , Jr. , N.ll.

Slack, R. Morte11) . Gynecology Oncology, 2:368-376, 1974, 1

v = w -- , - - ,

230. The Relationship of Ilistology to the Spread of Cancer. (Coauthors:

• E. Viadana, J.W. Pickren) . Journal of Surgical Oncology, 7:177-186, 1975.

231. The Metastatic Sp-cad of Myeloma and Leukemias in Men. (Coauthors:

E. Viadana, J.W. Pickren) . Virchows Archiv A. Pathological Anatomy and llistology, 355:91-101, 1975.

~

232. Letter to the Editor: The New York Sunday Times (Financial Section). February 2, 1975. ,

233. Adrenalectomy with Radical Mastectomy in the Treatment of Ifigh Risk Breast Cancer. (Coauthors: *T. Dao, T. Nemoto, A. Chamberlain).

Cancer, 35(2):478-482, February 1975.

234. Letter to the Editor: Uneven Doses. America, p.102, February 15, 1975.

235. SCIENTIFIC STRATEGIES IN llUMAN AFFAIRS: TO TELL 'IllE TRUTil.

Exposition Press, Inc., Jericho, New York, March 21, 1975.

236. Do Generalized Metastasis occur Directly from the Primary?

(Coauthors: E. Viadana, J. Pickren). Journal of Chronic Diseases, 28:149-159, May 1975. .

237. A Cascade Theory of Metastasis and Its Clinical Implications.

(Coauthors: E. Viadana, J. Pickren) . Sixty-Sixth Annual Meeting of the AACR Proceedings, May 7-11, 1975. San Diego, California, Vol. 16. (Abstract). ,

238. " Smoking and IIcalth. I. Modifying the Risk for the Smoker".

Less liarmful Ways of Smoking. Proceedings of the Third World Conference on Smoking and !!calth. (E.L. Wynder, D. Iloffman, G.B.

Gori, Eds.) U.S. Government Printing Office, Washington, D.C. ,

Vol. 1:111-118, 1975.

239. Professional Ethics: Society vs. Individuals. (Opinion) SIAM, 8(3):368, June 1975.

240. Studies on the Prevention of Respiratory Distress Syndrome of Infants Due to liyaline Membrane Disease with Plasminogen.

(Coauthors: *C.M. Ambrus, T.S. Choi, D.ll. Weintraub, B. Eisenberg, H.P. Staub, N.G. Courey, R.J. Foote, D. Goplerud, R.V. Moesch, M.

Ray, O. S. Jung , I . B. Mink , J . S. Amb rus) . Seminars in Thrombosis and IIemostasis, 11(1):42-51, July 1975. .

241. The Influence of Site of Metastasis on Tumor Growth Response to Chemotherapy. (Coauthor: *N.II. Slack). British Journal of Cancer,'32:78, 1975.

o 242. A Computerized Tumor Registry System for Persons Who Do Not Love Computers.i (Coauthors: *I.C. Johnson, L.E. Blumenson). Journal of Surgical Oncology, 7:375-380, 1975.

243. Marital and Reproductive Experience in a Community-Wide Epidemi-ological Study of Breast Cancer. (Coauthors: A.M. Lilienfeld, J. Coombs, A. Chamberlain). The Johns Hopkins !!cdical Journal, 136:157-162, 1975.

244. Epidemiologic Characteristics of Patients with Prostatic Neoplasms.

(Coauthors: *H. K. Armenian, A.M. Lilienfeld, E.L. Diamond) .

American Journal of Epidemiology, 102(1):47-54, 1975.

245. Comments on " Estimating ' Safe' Levels, A Hazardous Undertaking",

by N. Mantel and M.A. Schneiderman. Cancer Research, " Correspondence",

35_:3212-3213, November 1975.

246. Grammatical Compression in Notes and Records: Analysis and Conputation. American Journal of Computational Linguistics, 2_(4), December 1975.

247. The Role of the Statistician: Scientist or Shoe Clerk. Statistica, XXXV(4) : 801-804, October-December, 1975.

248. Letter to the Editor: ' Dumb Editorial' Riles a Health Researcher.

Buffalo Evening News, January 19, 1976.

249. " Communications from Bross". (Commentary). American Journal of Public Health, 66,(2):161, February 1976. ,

250. Epidemiology and Lung Cancer: What Has Been Accomplished?

Mentor, 7,(1) : 13-15, February 1976. -

251. Metastatic Sites that Produce Generalized Cancer: Identification and Kinetics of Generalizing Sites. (Coauthor: L.E. Blumenson) .

IN: FUNDAMENTAL ASPECTS OF METASTASIS. (Leonard Weiss, Editor, North Holland, Amsterdam), Chapter 21, 359-375, 1976.

252. Early Onset of Oral Cancer Among Women Who Smoke and Drink.

(Coauthor: J. Coombs). Proceedings of the American Association for Cancer Research, p.1, March 1976.

253. The Maximal Size of Avascular Tumors. (Coauthor: *L.E. Blumenson).

Proceedings of the American Association for Cancer Research, (Abstract), p.1, March 1976.

254. Hospital Admission Records: A Source for Identifying Occupational Groups at Risk of Cancer, (Coauthors:

• L. Houten, E. Viadana).

Annals of New York Academy of Sciences, 271:384-387, May 28, 1976.

-- - --%- - - . - - - - - - , , - - - = r- -+--a'- - - + + - * - ~~ *~ *6----- w -

Y* -

• w w-e-' ---'

255. Letter to the Editor: P.B. Medawar's Role at SKI. Murray Kempton:

On Lillian lic11 man. The New York Review of Books, XXIII, (10),

June 19, 1976.

256. Right Answers from Wrong Assumptions. (Editorial). Preventive Medicine, 5(2):203-206, June 1976, 257. Medical Concerns: The Mindless Use of Radiation Technology.

Proceedings of the National Energy Forum, the University of Akron, 1:5-11, July 26-30, 1976.

258. Letter to the Editor: On ' Controlling One's Environment'.

American Journal of Public Ilealth, 66,(9):907-908, September 1976.

259. The Metastatic Spread of Kidney and Prostate Cancers in Man.

(Coauthors: *E. ViadanaI, J.W. Pickren). Neoplasma, 23,(3): 323-333, 1976.

260. A Possible Mechanism for Enhancement of Increased Production of Tumor Angiogenic Factor. (Coauthor: *L.E. Blumenson). Growth, 40,:205-209, 1976.

261. Letter to the Editor: Response to " Definition of Rates: Some Remarks on Their Use and Misuse", by, Regina C. Elandt-Johnson.

American Journal of Epidemiology, 104(2):219-220, 1976.

262. Letter to the Editor

Politics of Mammography. Science News, 110(11), September 11, 1976.

263. Screening Random Asymptomatic Women Under 50 by Annual Mammographics:

Does it Make Sense? (Coauthor: L.E. Blumenson). Society for Epidemiological Research. American Journal of Epidemiology,104 (3):316, 1976 (Abstract).

~

264. Letter to the Editor: ' Major llazard' Scen in N-Plant Closing.

Courier Express, Editorial Page, October 14, 1976.

265. Letter to the Editor: Says Deadly Material Stored at NFS Plant.

Buffalo Evening News, Editorial Page 24, October 11, 1976.

266. Screening Random Asymptomatic Women Under 50 by Annual Mammographics:

i Does it Make Sense? (Coauthor: Leslie Blumenson). Journal of Surgical Oncology, 8,:437-445,.1976.

267. Scientific Strategies in lluman Affairs: Explaining Statistics to the Public. The American Statistician, 30(4):171-175, November 1976.

268. Letter to the Editor: Nuclear Waste: The ' Time Bomb'. The New York Times, November 19, 1976.

w: v - + - .w-- , - - . , - .w.- . - , ~ - -- - . c . - _ , , , . - . , , ,m . ,

i

~

4 269. Cancer Experience of Men Exposed to Inhalation of Chemicals or to Combustion Products. (Coauthors: *E. Viadana, L. Ilouten). t Journal of Occupational Medicine, 18,(12):787-792, December 1976.

270. Early Onset of Oral Cancer Among Women Who Smoke and Drink.

(Coauthor: J. Coombs). Journal of Oncology, 33,:136-139, 1976.

271. The Spread of Blood-Borne Metastases in Malignant Lymphomas of Man. (Coauthors: *E. Viadana, J.W. Pickren) . Journal of Oncology, 33_:123-131, 1976.

~

272. Letter to the Editor: Possible llazards of Radiologic Surveillance of liigh Risk Croups. The New England Journal of Medicine, 296(4):

232-233, 1977.

273. Differences in Breast Cancer Between Japan and the Unit . States.

(Coauthors: *T. Nemoto, T. Tominaga, A. Chamberlain, Z. Iwasa,

11. Koyama, M. llama, T. Dao) . Journal of the National Cancer Institute, 58(2):193-197, February 1977.

274. Letter to the Editor: A Coming Battle? Science, 195(4282): 933-934, March 11, 1977.

275. Exposure of Either Parent to Diagnostic Radiation Prior to Conception Produces Serious Genetic Damage in the Child. (Abstract)

(Coauthor: N. Natarajan). Proceedings of the American Association for Cancer hscarch,18:1, March 1977.

276. What is a Valid Scientific Argument: A Linguistic Approach to an Answer. Methods of Information in Medicine, 16(2):7580, April 1977.

277. The Changing llistopathology of Lung Cancer. (Coauthors: *R.G.

Vincent, J.W. Pickren, W.W. Lane, I. Bross, II. Takita, L. liouten,

~

t A. Gutierrez, T. Rzepka) . Cancer, 39,(4):1647-1655, April 1977.

f 278., Prevention of Ilyaline Membrane Disease with Plasminogen: A l Cooperative Study. (Coapthors: *C.M. Ambrus, T.S. Choi, E.

, Cunnanan, B. Eisenberg, H.P. Staub, D.l!. Weintraub, N.G. Courcy, i

R.J. Patterson, II. Jockin, J.W. Pickren, O.S. Jung, J.L. Ambrus).

Journal of the American Medical Association, ,237(17):1837-1841, April 1977. -

279. Genetic Damage from Diagnostic Radiation. (Coauthor: N. Natarajan).

l Journal of the American Medical Association, 237(22):2399-2401, May 30, 1977. ~

l t

280.

Distribution of Aryl liydrocarbon flydroxlase Inducibility in Cultured lluman Lymphocytes. (Coauthors: *B. Paigen, J. Minowada, ll.L. Gurtoo, K. Paigen, N.B. Parker, E. Ward, N.T. Ilayner, F.

Bock, R. Vincent). Cancer Research, 37:1829-1837, June 1977.

281. Five-Year Follow-Up Results of a Collaborative Study of Therapies for Carcinoma of'the Bladder. (Coauthors: *N.H. Slack, G.R.

Prout). Journal of Surgical Oncology, 9,:393-405, 1977.

282. A Retrospective Survey of Cancer in Relation to Occupation.

(Coauthors: *L. Houten, E. Viadana, P. DeCoufle, K. Stanislawczyk).

NIOSH Research Report, 1977. DHEW(NIOSH) Publication No.77-178.

283. Letter to the Editor. Primacy Principle. Environmental Health Perspectives, pp. 329-331, December 1977.

> 284. Occupational Cancer in Men Exposed to Metals. (Coauthors: *L.

Houten, E. Viadana, G. Sonnesso). Chapter 7, IN: ADVANCES IH EXPERIMENTAL MEDICINE AND BIOIDGY, Volume 91 (G.M. Schrauzer, Ed.), Plenum Press, N.Y., 1977, pp.93-102.

285. Letter to the Editor: Comment on Ridit Analysis by Dr. Bross.

American Journal of Epidemiology, 107(3):263-264, March 1978.

286. Cascade Spread of Blood-Borne Metastases in Solid and Non-Solid Cancers of Man. (Abstract) (Coauthors: *E. Viadana, J.W.

i Pickren). Proceedings of the American Association of Cancer Research, 19,:2, March 1978.

287. Intter to the Editor. Dose-Response Relationship in Radiogenic

, Breast Cancer. Journal-of the National Cancer Institute, 60(4):

727-728, April 1978.

288. Letter to the Editor: Nuclear Monologue 7 The Scier.ces, 18(5):

2, May-June 1978.

289. Preliminary Report on Radiation and Heart Disease. (Coauthors:

M. Ball, T. Rzepka, R. Laws). Journal of Medicine, 9,(1) :3-15, 1978.

290. The Metastatic Spread of Cancers of the Digestive System in Man.

(Coauthors: *E. Viadana, J.W. Pickren). Oncology, 35,(3):114-126, 1978. -

291. An Autopsy Study of Metastatic Patterns of Human Leukemias.

-(Coauthors: *E. Viadana, J.W. Pickren). Oncology, 35,(2): 87-96, 1978. ,

292. Cascade Spread of Blood-Borne Metastases in Solid and Non-Solid Cancers of Man. (Coauthors: *E. Viadana, J.W. Pickren). IN: ,

PULMONARY METASTASIS. (L. Weiss, H. Gilbert, Editors), G.K. Hall and Company, Boston, Massachusetts, Chapter 12, pp. 342-167, 1978.

293. Occupational Cancer in Men Exposed to Dust and Other Environmental Hazards. (Coauthors: E. Vladena, L. Houten). Archives of Environmental Health, 33(6):300-307, 1978.

l l - _ - . , . . -_ __ _ -,_ . . . _ . . _ ~ . . _ _ ._.---2.

3 294. Radiation Standards and Public Ilealth. Proceedings of a Second Congressional Seminar on Low-Level Ionizing Radiation. Washington, D.C. , February 10, 1978.

295, llazards to Persons Exposed to Ionizing Radiation (and to their Children) from Dos' ages Currently Permitted by the Nuc1 car Regulatory Commission. Presentation to the Nuclear Regulatory Commission, April 7, ,1978, Docket #43-FR9395.

296. Effect of Radiation on lluman llcalth:  !!calth Effects of Ionizing Radiation. Volume 1. !!carings Before the Subcommittee on llen1th and the Environment of the Committee on Interstate and Foreign Commerce. Ninety-Fifth Congress, Serial No.95-179, February 1978.

297. Reply by Dr. Brons: On* Biases in Judging Statistical Methods.

American Journal of Epidemiology, 109(1):30-32, 1979.

298. Natural Language and a Computer System in Medical Research.

(Coauthors: *I.C. Johnson, S.L. Tsao,.D.P. Shedd). Methods of Information in Medicine, 18,(1):15-17, 1979.

299. A Dosage Response C'urve for the One Rad Range: Adult Risks from Diagnostic Radiation. (Coatithors: M. Ball, S. Falen). American Journal of Public Ilealth, 69(2):130-136, 1979.

300. A Prospective Study of the Relationship Between Benign Breast Diseases and Breast Carcinoma. (Coauthors:

• L.J . Coombs , A.M.

Lilienfeld, W.S. Burnett). Preventive Medicine, 8(1):40-52, 1979.

301. Letter to the Editor: Comments on " Oral Contraceptives and Birth Defects", by Rothman and Louik. New England Journal of Medicine, 300(1):47, 1979.

302. Assessment of Myelotoxic Effects of Chemotherapy from Early Leukopenic Response: Application of a Mathematical Model for Granulopoiesis. (Coauthor: *L.E. Blumenson) . Journal of Surgical Oncology, 11_:171-176, 1979.

303. U.S. Civilian Casualties of Cold War Radiation Weapons and Myths:

Must There Be Another Million? Testimony to the Senate Government Affairs Subcommittee on Energy, Nuclear Proliferation, and Federal Services, March 6, 1979.

304. Letter to the Editor: Comments on "A Study of Diet and Breast -

Cancer", by A.B. Miller et al. American Journal of Epidemiology, 109(5):619, May 1979.

305. Letter to the Editor: " Low-Level Radiation Hazards..." (Comments on " Electromagnetic Radiation Emitted from Video Terminals" by Weiss and Petersen). American Industrial Hygiene Association Journal, 40,:830-831, September 1979.

] 306. Letter to the Editor: " Nuclear Discord". The Sciences, 19(7):

2, September 1979.

307. Letter to the Editor. ("the Forum..." Section): Comments on Video Terminals and Radiation. American Industrial Hygiene Association Journal, iO(12) : A14, December 1979.

308. When Speaking to Washington, Tell the Truth, the Whole Truth, and Nothing But the Truth, and Do So Intelligibly. The American Statistician, 34(1):34-38, 1980.

309. Cumulative Genetic Damage in Children Exposed to Preconception and Intrauterine Radiation. (Coauthor: N. Natarajan).

Investigative Radiology, 15(1):52-64, 1980.

310. Letter to the Editor. Nuclear Wastes. The New York State Journal of Medicine, 80:144, January 1980.

311. Letter to the Editor: Reassessment of Radiation Hazards: Can Health Physicists Keep Up? Health Physics, 38,:431-432, March 1980.

312. The Role of Brain Metastases in Cascade Processes: Implications for Research and Clinical Management. Chapter 4 in BRAIN METASTASES, pp. 66-80, G.K. Hall & Co., Publishpr, (L. Weiss, H. Gilbert, and J. Posner, Editors), 1980.

313. Letter to the Editor. " Mammography", Canadian Medical Association Journal, 123 (2) :94-95, July 19,1980.

314. Letter to the Editor. Re: Health Effects of Particulate Pollution:

Reappraising the Evidence. American Journal of Epidemiology, 112(4): 571-574, October 1980.

315. Bladder Cancor and Artifical Sweeteners: A Methodological Issue.

(Technical Comment) . Science, 210(4468):447, October 1980.

l 316. Letter to the Editor. " Sweet and Sour." The Sciences, pp. 2-3, November 1980.

317. The Biostatistical and Biologi'c al Basis for a Cascade Theory of Human Metastasis. In: METASTATIC TUMOR GROWTH (E. Grundmann, l Editor) , Cancer Campaign, Vol. 4, pp. 207-221, Gustav Fischer Verlag, Stuttgart, New York, 1980.

318. Muddying the Water at Niagara. (" Forum" Section) . New Scientist, gB (1231) : 728-729, December 11, 1980.

319. Letter to the Editor. (Comments on " Low-Dose Radiation and Leukemia," by A. Linos, et al) . New England Journal of Medicine, 303(14):815, 1980.

320. Letter to the Editor. Background Radiation Paradoxes. Health Physics, 40(January) :127-128, 1981. '

321. SCIENTIFIC STRATEGIES TO SAVE YOUR LIFE (Marcel Dekker, Inc., New York, New York) , March 1981.

322. This Week's Citation Classic ("How to Use Ridit Analysis",

Biometrics 14:18-38, 1958). Current Contents 24(24):17, June 15, 1981.

323. Metatechnology: A Technology for the Safe, Effective, and Economical Use of Technology. Metamedicine, 2:145-153, 1981.

324. Letter to the Editor. (Re: " Comparing Radical Mastectomy with Quadrantectomy, Axillary Dissection, and Radiotherapy in Patients with Small Cancers of the Breast", by Veronesi, et al). New England Journal of Medicine, 305(21):1283, 1981.

325. Direct Estimates of Low-level Radiation Risks of Lung Cancer at Two NRC-compliant Nuclear Installations: Why are the New Risk '

Estimates 20 to 200 Times the Old Official Estimates? (Coauthor:

D.L. Driscoll). The Yale Journal of Biology and Medicine, in press.

326. Automatic Encoding of Clinical Narrative. (Coauthors: *N.

Sager, G. Story, P. Bastedo, E. Marsh, and D. Shedd). Computers in Biology and Medicine, in press.

4 W

99

- - - - vv-

- . . - - - - _ - . -. = - - ,

(+

. si

W ( q L

y ,

) l L n woem p /w-op

{

3 I

/ NCE . Ji , 5i ,

1 More than a yhat ago, the wire services carried a reporf.byH U.T statisticians (CIA) hat the infant mortality rates in the Soviet f

' Union were rising. The . . analysts stressed the contrast to other industrialized nations re these rates are declining. More recently the wire services carried conf-irmation by Soviet statisticians that there has been an upturn in the rates.

1 Both U.S. and U.S.S.R. analysts offered explanations for the upturn but none are convincing. The factors sugge,ted are unlikely to have a visible impact on national rates. For instance, correction of underreporting in one region would not reverse a general national decline.

A more plausible explanation could be called the "technogenic hypothesis" (techno = technology, genic = caused by) . According to this hypothesis, I

, the traditional " baby killers" (the infectious diseases) are being replaced by new ones (the technogenic diseases). The latter are produced by chemical and radioactive contamination of the environment, particularly the drinking water.

If this technogenic nypothesis is correct, how can it account 4

for the infant mortality rates going up in the U.S.S.R. while they are going down in other industrialized nations?

Th'e re are unique circumstances in the U.S.S.R. , particularly j

with respect to its water resources, that would produce an upturn in mortality rates in the U.S.S.R. first and upturns in other industrialized nations some years later. Two factors are particularly important in the i

. - - - ~ - _ _ . . _ _ . _ , . _ . _ _ . _ , _ . _ - _ , . _ _ _ _ _ . _ _ _ _ _

coviet Union; one is geographical and the other involves decisions on the deployment of technologies. Geographically the U.S.S.R. is casen-tially a land-locked country whose primary water resources are its long river systems. Whereas t.he U.S. and other West.orn countries (because of the long coastlines) can dump wastes into the ocean or into relatively short river uyut ems that go into the ocean, t.his ocean-dumping option is much less available to the U.S.S.R.

The combination of geography and centralized decision-making has resulted in deployment of heavy indust.ry and nuclear plants in series along the long river systems. This result.s in repeated reutili-zation of the water, both for industrial purpoues (uuch as cooling) and for human p0rposes (such as drinking). As the water flows downstream the contamination steadily builds up and t.herefore a high proportion of the Russian populat ion is exposed to t.his contamination.

Can the technogenic hypothesis be put t.o a rigorous empirical If scientific test? The Soviet data might be said to be one such' test.

the infections diseases had not been replaced by technogenic diseases the infant mortality rates in the U.S.S.R. might have leveled off but r

i they could not have turned upwards. However, two objections can be made l

to this test of the hypothesis. First, it was post-hoc rather than Second, it might be predictive, the latter being much more convincing.

felt that U.S.S.R. statistics are suspect and that U.S. or other Western l

statistics would be more convincing.

l Can the technogenic hypothesis be test.ed on U.S. infant mor-tality rates? tiot yet, but perhaps in the near future. 71.w U.S r ra*.tnr' w , of

( CW

""Q.~

u af .

ylpO &0

)

m

[,i eg9

/ L' cr ,

00LiEIE P U:L A SIMPLE MECilANISM FOR SYPERGISM IN GENETIC DAMAGE T2 1B910 P1:39 FROM LOW-LEVEL RADIATION OR CllEMICAL MUTAGEllS t

From our studies of the Tri-State Survey data on the relation-ships between exposures to diagnostic radiation and childhood leukemia, there is strong evidence of synergistic effects in combinations of

. . z. - --

exposures (1). He had considered reports of diagnostic x-ray ~to the mother, to the father, and to the fetus and mother and combinations of these exposures. The ricks of leukemia in the child when a single type of exposure is reported are significantly higher than wher. no exposures are reported and certain combinations of exposure are significantly higher than the single exposures. The excess risks appear to be more l than additive and therefore would be called " synergistic".

. These findings have now been confirmed by the data on genetic damage in children of parents exposed to low gonadal doses of radiation from the A-bombing of Hiroshima and Nagasaki. The report in Science failed to find any significant radiation effects but a simple reanalysis of this data shows the effects very clearly for genetic damage as measured by deaths at early ages among the children of the bomb victims. The original Science report (2) failed to detect the effects because it made the simplistic (and erroneous) assumption of a linear hypothesis--that the effects must increase linearly as the dose increases. As a simple graph of the actual data clearly shows, the genetic effects diminish when the dose is very high and this same effect shows in other epidemiological

?

'l ~~

data. It is reported by Mancuso (3) for the data on the IIanford reprocessing plant and for uranium workers. In our studies of the Portsmouth Naval Shipyard nuclear submarine workers we have found simiJar effects but the clearest evidence is for data on breast cancer caused by x-ray treatment of post partum mastitis in Rochester, New Yor .

If the risks of genetic damage in children of A-Lomb victims is plotted by estimated gonadal doses and confidence intervals on the risks are included for case of interpretation, then the synergistic effects when both parents have low-dose exposures stand out very c1carly.

Graph I shows the early deaths in the children by the gonadal dose to the pa rents. The baseline group for the comparisons are the children where neither parent was in the city (NIC) at the time of the A-bomb blasts and the confidence intervals for this series are shown by horizontal dotted lines. It can be seen that when either parent had low gonadal doses there is no genetic effect visible in the graph, but when both parents had low doses (less than 10 rem) the early deaths are significantly higher (the 95% confidence intervals do not overlap) . The distribution of exposures is such that in the significant group the total gonadal exposure (mother & father) would usually be well under 10 rem, much of

~

the time less than 5 rem. The total exposure when a parent has a reported dosage over 10 rem is therefore likely to be larger than that l when both parents have low doses. IIence; what is occurring when both

' parents are exposed can be called a " synergistic" effect. This confirms i

our previous report of synergistic effects with diagnostic x-ray exposures.

d

-- 3 --

The next question is therefore: What is the scientific explanation for the synergistic effects when both parents are exposed to low-level radiation? The purpose of this note is to suggest a relatively simple u cnetic mechanism that could produce this synergistic effect.

A Model for Synergism in Genetic Damage to Children of Parents Exposed to Mutagens The theories of genetic damage from . 3 IcVel ionizing radiation have tended to involve dogmatic assumptions .4 :h as the linear 1 ypothesis 1 _ g/

and to ignore the factual evidence of low-]evel radiation hazards in humans actually exposed to such radiation. They have often been put in abstract mathematical language with little or no attempt to relate the abstractions to events in the real world. An example of this would be abstract algebraic formulations of "mul_tihit th_eory" in which it is

'/,

postulated that some number of " hits" are required to cause cancer or other expression of genetic damage. The underly.ing idea is reasonable and, indeed, will be involved in the simple mechanism proposed here.

!!oweve r , it is important to relate the abstract " hits" to the process involved in the causation of human cancer in the real world, i

l l

To some extent a " hit" corresponds to a break-point in the DNA

! of tie genetic material of a human cell. This is the starting point in a long and complex chain of events which may eventually lead to clinically I

1 detectable human cancer. But it is only the starting point. For the l

multihit theory to be realistic with respect to genetic damage, the rest of the process must also be considered. In general terms (though not in l

t i

// ~

all specific details) the overall process has beccme comprehensible in the past decade. In brief, the hit (or the misrepair of the hit) puts misinformation into the genetic code by a lesion in the biochemical structure of the DNA. For this genetic damage to have any effect at the

! Icvel of the whole organism it must be reproduced millions of times in the clone of the damaged cell. For this to happen, the cell must be I

f able to reproduce "better" (in some sense) thall the undamaged cel'Is.

This view of the overall process imposes some stringent conditions for the expression of genetic damage which are not apparent in the abstract hit theory. Whereas in the original theory all that is necessary t to get cancer is a sufficient number of hits, the requirement that the A '

damaged cell be able to reproduce effectively puts an emphasis on the i quality as well as the quantity of the hits. If tha hits happen to put m

a break point in a part of the DNA needed for reproduction or viable

\_ /

growth, the end result will not be cancer because the process will be i

aborted. Idlercas in the original hit theory, more hits will automatically cause more cancer (in rough agreement with the sacrouanct linear hypothesis) ,

when the cloning phase is taken into account this lot be true.

With sufficient hits (e.g., with exposure to high levels of ionizing radiation) the risk of a visable expression of genetic damage (such as L

carly death in the child) can be expected to diminish. 'fr1 w & & as, In othet v.u t if m hits are required to produce deaths in the early years after birth as an expression of genetic damage, an add!tional hit might produce miscarriage or fetal death that would prevent the later expression of genetic damage. The results in Graph 1 at high radiation exposures are in line with this sort of extension of the original hit theory.

- g-Once the stringent conditions for viability are thoughtfully considered, the hit theory in its original form becomes increasing?y implausible. What,' for instance, would be the chancec of having a series of cancer-enhancing hits without having che process aborted because some part of the DNA essential for viability of the cell or whole organism has been hit? From biological consideratio.3, there must be an enormous cunount of information coded into the DNA that would be essential for reproduction arid develormental processes. While there would be redundancy in the code, there could be relatively few areas where the information would not be essential and where the misinformation would not seriously impair viability.

Exposure to low-level radiation would be like shooting at random at a double helix of "lightbulbs" where a few are yellow (cancer-enhancing hits) and the rest are white (redundant) or red (essential for viability) with a preponderance of red. With random shooting at the

$dV4]Wthdrkh Y &M & tegidA Ua f W A YY"Wh lights it becomes implausible that, say,, four yellow in the helix, h 4 2cI/W f lights will be hit %uomrA hitting a red light stops the process.

~

P How can a multihit theory for genetic damage or cancer be made credible if the possibility of hits stopping the process is taken

, into account? A single hit might be rare but not'as impossibly rare as four successive hits. If the human popuJation is considered rather than f a single individual, then there is a way that a multihit theory could be plausible. If t he human population is exposed to mutagens, there would be a frequency distribution of the number of pre-existing cancer-enhancing hits in the gene pool. In other words there might be many with no pre-

-b existing damage, come with one hit, a few with two hits, and a very few with three hits. If four hits were needed for cancer, the individuals with three hits would be in a vulnerable position since for them the one additional hit would be all that was required. Hence, instead of supposing that all of the required hits occur simultaneously (for example during I

exposure of an individual to high-level radiation), the hits would be V

part of a gradual process of genetic degradation of a human population, and with this extension the nultihit theory might be credible.

There is documentation for this genetic degradation version of the multihit theory from analyses of the Tri-State Survey data. The existence of a susceptible sub-group of the population has been doc?mented for both children and adults with myeloid leukemia. As might be expected if the susceptibles were ner.emn-Wi_th pre-cxisting genetic u ' mage in this Cdd.

(i.e., cance- enhancing hits), there would be increased e is that the damage g group t diseases other than cance to DNA would probably af fect the enzyme that was coded for in the damaged area. This, in turn, might be expected to impair the effectiveness of The l

i host-defense systems or other body' systems that used the enzyme.

t imbalance that would affect the operations of the white cell sub-system f

of the host defense system (and hence would eventually lead to leukemia I

l from failure of the feedback machinery) would be likely to affect other I

sub-systems as well. In other words, the misinformation from the break-point in the DMA would produce a range of effects __ra_th_e_r_ than a single effect such as leukemia. Documentation of this expected co-occurrence

. . - }.

of diseases had been possible from the Tri-Stat e Survey Gata. In adults with myeloid leukemia, co--occurr ence of heart disease and other diseaues was noted or children t.bere was a broad spectrum of such co-occurrent diseaues. The included allergies, xxxxxx, xxxxx, xxxxx, etc. (:[X-h) .

With a r.iodified multihit theory, a new mechanism for nynergism in 1 i sku t o children of exposed persons,comes

- . . - . - . . . - - . . - . . - - - . . . - - -~

into play. Whereas in the exposed persons then.uolves, the rnodified theory limits the degradat ion el to one addi t ion hit, in the children there can be an increase of t.wo or A

more. The combinations of damaged DNA are more complex than simple addition of hits although this may serve as a first approxin.ation.

Since each parent is presumably contributing viable cells, a damaged gene on a chromosome is not a fatal flaw and procumably combinations of darnaged chromosomes would be possible without resulting ia a fatal flaw--

at least not in most cases. In some cases, of course, the combination of two or more damaged chtomosomes could result in fatal damage but the t2 y : ' %5 '

a r >r aji * > 1 ~ 'n ". n chances of viability with two or more additional hits in the DNA of the child would be much greater than the corresponding chances in the DNA~

of an exposed individual, j Thus, if four hits produced an expression of genet.ic damage uuch as deaths in childhood, it would be possible that this could be produced in tho children of parents with two hits each. With the modified

r.ultihit model it. would not be rossible fer the parents themselves to reach the four-hit level from ex1osures to mutagens, even at high levels, o

Hence, with this modified model a synergistic ef fect in the children of exposed parents would be expected. The mathematical details are given in t.he next section.

I

j .

C tjC'. d'1

'82 Wa 10 P1 :39 IS THERE A SERIOUS CANCER HAZARD FROM CHEMICAL -

\ -

CONTAMINATION IN THE NIAGARA FALLS AREA?

A Crash Course in Biostatistics for Legislators

• Dr. Irwin D.J. Brass Department of Biostatistics Roswell Park Memorial Institute Buffalo, New York 14263 December 16, 1981 I testify as an individual and for no institute or organi-zation. As a public health scientist and biostatistician-epidemiologist for more than 30 years, I have published more

- than 300 papers and my latest book, SCIENTIFIC STRATEGIES TO SAVE YOUR LIFE (Marcel Dekker Inc. , ' 1981) , deals in detail with studies of the hazards of low-level ionizing radiation.

For more than 20 years I have been Director of Biostatistics at Roswell Park Memorial Institute for Cancer Research in Buffalo, New York (for 7 years as Acting Chief of Epidemiology) and before that was at Cornell University Medical College and Johns Hopkins.

m y ..-* , - , - -

m -

BACKGROUND INFORMATION This position paper has been written i.t the express request of the Assembly Committee on Environmental Conservation, a request made during the NYPIRG hearing of December 10, 1981, on contamination of the Niagara River. Readers interested in the answer to the title question can skip this note (which deals with the circumstances leading up to the request).

Since the order of the witnesses at the hearing (I was number

14) had been changed to accommodate out-of-town witnesses, it was a curious coincidence that my testimony immediately followed a witness from the New York State Health Department.- It flatly contradicted what was said "in the name of" the Health Department. To avoid any appearance that this critique involves personalities, I will hereafter refer to the previous witness as PW.

PW had acknowledged that he was not an expert on biostatistics or epidemiology (although he was testifying on cancer risks at Niagara Falls). Although I was testifying as an individual scientist and not for the Health Department, I have been Director of Biostatistics at Roswell Park Memorial Institute for Cancer Research for 22 years (and Acting Chief of Epidemiology for 7 years) . Hence the legislators had an opportunity to get a professional critique of the previous testimony and took advantage of it.

The testimony of PW was rather startling to the legislators who had been hearing about the chemical contamination of the Niagara River for more than 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />. PW stated (and repeatedly affirmed under cross-examination) that the quality of the Niagara River was " excellent".

When asked to compare the river water with mountain spring water, he insisted both were equally " excellent". When asked about the cancer hazards in Niagara Falls, he insisted there was "no significant" excess cancer. He even gave the legislators a report by the Bureau of Cancer

~

Control (BCC) which purportedly backed up this surprising claim. He gave the impression that the Health Department believed the figures in the Bureau of Cancer Control report (attached to this report) " proved" there was no excess cancer in the Niagara area (1).

Coincidentally, my testimony was titled "Why the Assurances that the Water is ' Safe' Have no Scientific Validity" (2) (it was written November 19, 1981)r My testimony used federal vital statistics (3) to demonstrate that during the period 1950-1969 there had been hundreds of excess deaths from cancer in Niagara County. I also cited a Science paper from the State Health Department that gave more recent figures showing excess incidence of cancer in Niagara Falls (4). As will shor tly be seen, the BCC' report handed to the legislators by PW to " support" his claim of no significant excess cancer is actually in good agreement with the data that I had cited. I,t shows exactly the opposite of what PW testified it shows.

The body of this paper uses a question-and-answer format to present the answer to the title question (and others) as simply and clearly as possible. The business of a professional biostatistician is to put questions to a body of data and to obtain the answers byv, letting the facts speak for themselves. For this purpose he (or she) ,uses standard, objective statistical methods for separating the message of the data (consistent. effects) from the " noise" (random offects) . This lets the message come through more clearly. In what follows, the answers will be given together with an explanation in plain English of the biostatistical reasoning that leads to the results. In my new book, SCIENTIFIC STRATEGIES TO SAVE YOUR LIFE, which was just published by Marcel Dekker, Inc. (1981), I show how statistics is simply " common sense reduced to calculation" (5). The mystery and magic can be e1Lai-nated from these numerical operations and hence my subtitle.

Why go into this much statistical detail? The answer is I am more concerned with the slovenly bureaucratic ploy that was used here than with the claims of PW that are easily refuted. You might wonder:

1 How could PW expect go get away with handing a BCC report to the Committee that shows the opposite of what he is claiming? The answer is that numbers are confusing. Most of the tLme tha this ploy is used, bureau-crats get away with it because there is no one to explain to legislators or to the public what the numbero really mean. Perhaps if this kind of misreport'ing is pointed out (and something done aoout it), there will be less of it in the future.

?

1 l

QUESTIONS-ANL-ANSWERS ABOUT THE CANCER RISK IN THE NIAGARA FALLS AREA Q: Is there a serious cancer hazard from chemical contamination in the Niagara Falls area?

A. Scientifically speaking there can be little doubt that there is a serious hazard in this area. In addition to the limited data in the BCC report, there is extensive data from federal vital statistics as well as other studies of the New York State Health Department. There is also extensive collateral scientific evidence on contaminants and their hazards. The excess cancer risks have continued over more than 30 years, have involved hundreds of persons in the Niagara Falls area, and have been found in numerous independent studies by the New York State Health Department and others. At this point the question of primary concern to any competent and responsible public health scientist is not "Is there a risk?", but "What can be done about the hazard?" However, in view of the legislative mandate, only the first question will be addressed here.

i Q: The previous witness (PW) testifying "in the name of the Health Department" has stated that there was "no significant" excess cancer in the Niagara Falls area and pointed to the BCC data for males in Niagara County (1974-1976) for all sites as

" proof" of this assertion. How would you answer this testimony?

[

A: The answer is obtained by simply looking at the facts in the BCC report with a little common sense (or a little biostatistical l'

competence). PW had specifically pointed out the line for l "All Sites (140-208)". There are three numbers in this line.

~

The first is the " Observed" average annual number of diagnosed cases for the three-year period, 457.00 diagnosed cases. The actual number of cases reported for the three-year period is I therefore 3 x 457 = 1371 cancers. The second number is the

" Expected" average annual number of diagnosed cases on the

l 4_

basis of the age-sex specific rates for New York State (ex-clusive of New York City), 427.01. In the three-year period, 1281.03 cases would be expected. The excess cancers are the difference between the number of cases " Observed" and the number " Expected", 1371 - 1281.03 = 89.97 or about 90 cases.

Q: Are 90 excess cancers significant er not?

A: The word "significant" has two somewhat different senses, a public health sense and a purely statistical sense. By either meaning the 9,0 excess cancers are "significant" and PW has misreported the BCC results to the Committee.

Q: Are 90 excess cancers "significant" in a public health sense?

A: We are talking about trying to prevent at least 90 premature cancer deaths among men living in Niagara County. This is certainly a "significant" target for the State Health Department.

Q: How do you know that 90 is a real excess and doesn't just reflect the chance fluctuations of numbers from fear to year?

A: This question involves "significant" in the statistical sense.

The simplest way to answer this question is to do u.little calculation called the " Chi-Square Test". For this we square the 90 (=8100) and divide by the " Expected" (1281.03) to get 6.32. If the calculated number exceeds 3.84--which it clearly does here--then it is unlikely (less than one chance in twenty) that this is a chance fluctuation. Here the Chi-Square is so large it shows a chance of about one in 100. A statistician

. calls such a result "significant" or " highly significant".

Q: Why then did PW say that this data was not significant?

A: The facts can show that PW misreported the BCC data but they cannot show why he did it. As a bureaucrat, I might speculate on why this was done. I would guess that PW's unit of the Health Department had previously made a policy decision not to get

7_ olved in the Niagara River issue. The NYPIRG report sug-gested that the unit ought to be invol ed. When PW came on strong with " excellent" Niagara River water and "no significant" cancer hazard, he was trying to " justify" the determination of his unit to " sit on its hands".

Q: Is it Health Department policy to misrepresent the hazards or is it the policy of units within the department?

A: In my testimony, I spoke as an individual scientist and I cannot speak for the department. However, in my opinion, Dr. Axelrod is not involved in this. Indeed, PW in his testimony put some distance between his unit and Dr. Axelrod. For instance, he called Dr. Axelrod "too conservative",(meaning too concerned about potential health hazards). Hence, it sounds like a particular unit that is foot-dragging on its official mission.

Q: Coming back to the general question of the hazards in the Niagara Falls area, even if the excess cancer is "significant" how can you know that it is chemical pollution that is causing this cancer?

A: There is a simple way and a difficult way. There is also a middle way using the BCC report itself.

Q: What is the simple way?

A: Anyone who wants to convince himself (or herself) that chemical

ontamination is the culprit has only to get the geographic e

cancer atlas (3) put out by NCI, look at the maps for chemically-i related cancers (such as bladder cancer or esophagus cancer),

l l and than look for the counties where (a) the cancer rates are j in the upper decile (upper 10 per cent of counties), and (b) where the excess cancer deaths are statistically significant.

Since the atlas uses red (or another distinctive color) to l

identify the high risk counties, they are easy to pick out.

It will be immediately evident that many of these counties i (including Niagara) are ones with a high concentration of chemical industries.

Q:

What are the other ways to tie the excess cancer to chemical contamination?

A: The hard way, which will not be considered here, is to go through the scientific literature on chemical carcinogenesis. The middle way is to look at the sites of cancer in the BCC list.

One can then compare cancers known to be caused by chemical and radiological contamination with those that are relatively insensitive to en- svamental factors. The latter sites give us some idea of what the cancer incidence might be in Niagara Falls were it not f,r ochemical contamination from occupational exposures, dumpsites, and air and water pollution.

Q: Is Niagara Falls an area with an inherently high cancer rate?

A: No, quite the opposite. The data suggest that the area would be healthier than average if it weren't for the chemical contami-nation. To see this we can look at cancers of the large intestine without the rectum (cause of death number 153) .

Here we find the observed number of colon cases are less than the expected number (the observed-to-expected ratio is less than ore . 0.926) . This is probably about where the cancer ratios would be for Niagara Falls if the chemical pollution were eliminated.

l Q:

How does this compare with the .atios for environmentally sensitive f

cancers?

i A: They swing in the opposite direction. For example, lung cancer has a ratio of 1.146, kidney 1.27, pancreas 1.22. Bladder is higher than the intestinal cancer, though only 0.979.

l Q: Are these statistically significant?

A: The numbers at most individual sites tend to be too small for this.

However, lung cancer is a large series. For lung cancer the difference between expected and observed for the three-year period is 325 - 283.5 = 41.5. The Chi-Square is 6.07. Hence, lung cancer is statistically significant in its own right and .

accounts for a substantial part of the excess cancer in Niagara County. This lung cancer effect would be even greater if compared to colon cancer.

Q: Why would you want to do that?

A: To get a better idea of the actual excess deaths. Comparisons with statewide figures include contaminated areas in the averages.

Using the 0.926 figure to adjust the expectations for this gives an estimate of 62.5 excess lung cancers instead of 41.5.

Q:

You seem to be picking out the high ratios and calling them chemi-cally related. Is that what you are doing?

A: No. It is the other way around. The cancers I picked out have been extensively studied in the scientific literature and found to be chemically related. This data, in effect, confirm that the literature findings indeed apply to Niagara Falls chemical pollution.

Q:

Aren't cancers caused by cigarette smoking?

A: Of course, Dr. Ernst Wynder and I showed the cigarette-cancer relationships back in the early 1950's. However, with a strong factor like cigarettes it becomes harder, not easier, to shes effects from chemical pollution or other factors. An adjustment for cigarette effects on lung cancer would make the estimate of preventable (exceus) cancers even larger.

Q:

What about women? How can you explain the ratio of 0.931 for all sites for women if there are effects from chemical pollution at Niagara Falls?

A: Most cancers for women tend to be more hormone-related rather than environmentally-related. For colon cancer in women, which is not closely related to hormones or to the environment, it can be seen that the ratio is very low, 0.822. Breast cancer which is hormonally-related is also very low, 0.858.

These two sites comprise more than one-third of all female cancers and this tends to make thc rate for all sites low.

~

Q: Is there any sign of chemically induced cancer in women?

A: Yes. Looking at lung cancer (1.030) and buccal cancer (1.139) we find ratios that are high relative to the very low rates for colon and breast cancer. These low rates in women, by the way, confirm the inference from the male data that Niagara Falls would be a healthier-than-average place if it weren't for the chemical pollution.

Q: Perhaps the pattern in the BCC data fits roughly into the pattern of chemically-induced cancer but there certainly are numbers which do not fit so well. How do you explain them?

A: A biostatistician's business is working with imperfect data, data where the message is obscured by " static" from random fluctu-ations. Statistical methods can bring out the message more clearly, but they do not eliminate the " static"--particularly when the numbers become small. For many cancers on the list there are, on the average per year, less than 15 cases. These will show a lot of variation. Only in the larger series (such as the lung cancers in men) does the message come through clearly. Moreover this is routine data which was processed by the Bureau of Cancer Control for general purposes and not for the specific purpose of the NYPIRG hearing. It is not ideal for scientific purposes.

Q: Why isn't the data " ideal"?

A: There are quite a few things that obscure the relationships of interest. For example, the chemically-induced cancers tend to occur at a somewhat earlier' age than the naturally occurring cancers. When all ages are combined there is no way to see which cancers are occurr,ing " prematurely". For routine purposes, the BCC data is fine, but when we want to take a close look at what is happening at Niagara Falls (say to monitor clean-up progress) , we need finer statistical analysis.

- . _9_

Q: Have you discussed this with the Bureau of Cancer Control? In your original testimony you had suggested working jointly with Dr.

Janerich.

1 A: When I phoned Dr. Janerich, he explained that his new administrative load made it impossible for him to work with me but that perhaps someone else_in his section might do so. My original suggestion was based on the notion that Dr. Janerich and I are known to have rather different views-about cancer hazards.

Hence, if we could agree on a joint statement, it might therefore have more credibility. Doing some computer runs would not be the shme as trying to develop a joint statement.

Q: In summary, does the BCC Report show evidence of serious excess cancer risks in Niagara County?

A: Yes. However, if the BCC Report were the only eviden e, it could be challenged on various grounds. By itself, it might not be conclusive. Taken together with the other data on cancer risks in the Niagara Falls area, it shows a pattern censistent with the scientific hypothesis that the excess risks in the area (that were previously demonstrated) are continuing into the 1970's. It is completely inconsistent with any hypothesis that the excess risks of cancer in the area have been eliminated.

Q: Can you suggest any way that we can avoid misleading statistical testimony on these health issues?

A: Yes. I would suggest that all persons who want to give " expert" testimony (as opposed to "public opinion") be sworn, be required

• to give their credentials, and be subject to perjury charges.

i t REFERENCES i

1. New York State Department of Health, Bureau of Cancer Control Report, July 18, 1980. -

2. I.D.J. Bross: Why the Assurances that the Water is " Safe" Have No 3

Scientific Validity. Tesitmony submitted to the Assembly on Environmental Conservation, December 10, 1981.

i

3. U.S. Department of Health, Education, and Welfare, Vital Statistics:

U.S. Cancer Mortality by County: 1950-1969. DHEW Publication No.

i (NIH)74-645.

I 4. D.T. Janerich, W.S. Burnett, G. Feck, et al: Cancer Incidence in the Love Canal Area. Science 212:1404-1407, June 19, 1981.

1.

5. I.D.J. Bross: SCIENTIFIC STRATEGIES TO SAVE YOUR LIFE (Marcel Dekker Inc. , New York) , 1981.

l i

f i

i i

e I

- ,~ -- - ._ . _ . . - - - , _ , . _ = . . -- _. - - - , - . - . _ - - - _ - - - - _ . . . - _ .

NEW YORK STATE DEPARTMENT OF HEALTH BUREAU OF CAtiCER CONTROL Average Annual Number of Diagnosed' Cases of Malignant Neoplasms **

Major Sites for Males ,

NIAGARA COUNTY - 1974-1976 ,

(SEMI-FINAL) -

Site Observed Expected *,

Ratio 0/E 457.00 42,7.01 - 1.070 All Sites (140-208)

Buccal Cavity & Pharynx (140-149) 19.'67 ..

17.92 . -1.097 11.67 15.00 0.778 Stoinach (151) .

Large Intestine, I Rectum (153) 42.00 45 37 - 0.926

- 27.33 24.50 1.115 Rectum (154) ,

~

~ ~

16.67 13.65 1.226 Pancreas (157) ,

(,

108.33 '94.50 1.146 Lung.& Bronch'us (162). ,,

72.00 60.19 - 1.196 -

Prostate (185) .

~

3.00 3.76- ", 0.796 Testis (186) -

33.33 34.05 , 0.9'79 Bladder (188) , , ,

15.00 11.83 1.167

'. Kidney (189)

?.33 2.11 0.630 Thyroid (193) 17.67 17.47 1.01h .

Lymp'homas(200-202) , ,

15.33 14.23 1.077 Leukemics (E04-208)

• 73.67 72.23 1.020 All Other December, 1979 Processed ;uly 16, 1980 Reported through

• Expected number derived by applying age-sex specific rates for New York State, excluuve of new York City, to estimated county population.

,( ** Includes 19 . cases reported during the 3 years by Death CerLificate Only, some of whom were diagnosed in earlier years and/or the dia9aosis was not l confirmed.

(7/18/80)

-- , g _. _m _

'_ BUREAU OF CAtlCER C0ilTROL Average Ar.nual Number of Diagnosed Cases of fialignant fleoplasmsN fiajar Sites for Females NIAGARA COUNTY - 1974-1976 (SEMI-FINAL)'

Site -

Observed Expected

• Ratio 0/E All Sites (1402208) 404.33 434.12 0.931 ,, ,

~

Buccal Cavity & Pharynx (140-149) 8.67 7.61 1.139 10.33 9.68 1.067 Stom$ch (151) 41.00 49 . 8 O.822 Large Intestine, i Rectum (153) 17.33 19.12 0.906 Rectum (154) .

9.67 12.23 0.790 -

Pancreas (157) -

" . 33.33 32.35 1.030 Lung & B,ronchus (162) .

107.00 124.73 0.858 Breast (174) ,

39.00 32.94 1.184 .

Cervix Uteri - In. Situ (233.1)***

19.00 16.02 1.186 .

- Invas'ive (180) 42.33 36.73 1.152 Rest of Uterus (179,182) 19.33 20.90 0.925 Ovary (183) 7.33 5.21 1.406 Thyroid '(193) t 14.00 14.95 0.936 Lymphomas (200-202) 9.33 10.40 0.897 .

Leukemias (204-208) 65.67 7,4.15 0.886 All Other .

Reported 1.hrough - December,1979 Processed July 16,1980

~

• Expectad number derived by applying age-sex specific rates for New York State, g exclusive of flew York City, to estimated county population.

q l

• • Includes _ 32 cases reported during the 3 years by Death certificate Only, soi g of y eghom were diagnosed in earlier years and/or the diagnosis was not

^** Carcinoma In Situ of Cervix Uteri not included in tabulations of all sites.

(7/18/80)

- . . . . . . . - . . .-. m

7 ,

ilEW YORK STATE DEPARTMEtlT OF HEALTH

- - 8UREAU OF .CAtlCER CONTROL Averac_ e Annual

. Number of Diagnosed Cases of Malignant Neoplasms **

!!ajor Sites for Males .

ERIE COUNTY - 1974-1976' (SEMI-FINAL)

Site Observed Expected

• Ratio 0/E All Sites (140-208) 1988.67 1891.21 1.052.

Buccal Cavity & Pharynx (140149) 2 95.33 79 06 1.206 C 78.33' 66.36 - 1.180 -

Stomach (151)

Large Intestine, I Rectum (153) 190.00 ,

200.88 0.946 114.00 108.32 1.052 Rectum (1S4) .

~

62.00 60.35 1.027 -

Pancreas (157) .

Lung.& Bronchus (162). 462.33 416.92 1.109 ,

{ 284.00 267.37

~

1'.062 .

Prostate (185) '

19.33 17.51 . i.104 Testis (186)

• 159.67 150.73 1.05s Bladder (188) 60.33 52.23 1.155 Kidney (189) 15.67 9.54 '1.642 Thyroi,d (193) 69.33 78.35 0.885' ,'

Lymp'omas h (200-202) 61.33 63.40 0.967 Leukemias (204-208) 317.00 320.02 0 991 All Other l . .

~

Reported through December, 1979 Processed July 16. 1980 l

l

• Expected number derived by applying age-sex specific rates for New York State, j exclusive of New York City., to estimated county population.

/ ** Includes 245 cases reported during the 3 years by Death Certificate Only,

' some of whom were diagnosed in earlier years and/or the diagnosis was not confirmed.

(7/25/80) l -

o BUREAU OF CAtlCER C0ilTROL Averace Annual Number of Diagnosed Cases of Malignant fleoplasmsN Major Sites for Females ERIE COUNTY - 1974-1976 (SENI-FIN $)

Site Observed Expected

• Ratio 0/E 1944.00 2022.-37 . 0 .9 61 All Sites (140 208) ,

30.67 35.39 0.867 Buccal Cavity' & Pharynx (140-149) 50.67 45.51' l.113 .

Stom$ch (151) 204.00 23438 0.870' Large Intestine, s Rectum (153) 92.67 89.70 1.'033 Rectum (154) . .

Pancreas (157), 60.67 57'.43 1.056 .

0.916 -

h Lung & B,ronchus (162) - 137.67 ,

150.25.

563.33 577.78 0.975 Breast (174) ,

116.00 156.13 0.742 . e Cervix Uteri e In.fitu (233.1)***

69.00 74.38 0.928 - -

- Invas'ive (180) ~

158.00 ~169.74 0 .931 - .

Res't of Uterus (179,182) 101.00 96.94 l'.042 Ovary (1,83) 23.33 24.46 0.954 Thyroid '(193) 60.33 70.21 0.859 Lymphomas (200-202) 48.00 48.79 0.984 ,

. Lcukemias (204-208) 344.67 347.13 0.993 All Other .

~

Reported through December, 1979 Processed July I6.1980

• Expected number derived by applying age-sex specific rates for New York State, exclusive of New York City, to estimated county population.

• • Includes 321 ,ca,ses reported during the 3 years by Death Certificate Only, sa of e$hom were diagnosed in earlier years and/or the diagnosis tas not g

• • • Carcinoma in Situ of Cervix Uteri not included in tabulations of all sites.

(7/28/80) -

t e

nog:r

'82 ""7" 10 P 1 P"9 WHY THE ASSURANCES THAT THE WATER IS " SAFE" HAVE NO SCIENTIFIC VALIDITY i

Testimony submitted to:

Assembly Committee on Environmental Conservation Capitol Building, Room 520 Albany, New York 12248 For,a Joint Public Hearing on New York State Public Interest Research Group's (NYPIRG)

Report on Toxics in the Niagara River -

to be held December 10, 1981

. Testimony of:

Irwin D.J. Bross, Ph.D.

Director of Biostatistics Roswell Park Memorial Institute Buffalo, New York 14263 Nov' ember 19, 1981 I testify as an individual and for no institute or organization.

As a public health scientist and biostatistician-epidemiologist for more than 30 years, I have published more than 300 papers and my latest book, SCIENTIFIC STRATEGIES TO SAVE YOUR LIFE (Marcel Dekker, Inc.,1981) , deals in detail with studies of the hazards

of low-level ionizing radiation. For more than 20 years I have been Director of Biostatistics at Roswell Park Memorial Institute for Cancer Research in Buffalo, New York (for 7 years as Acting Chief of Epidemiology) and before that was at Cornell University Med'ical College and Johns Hopkins.

I o m

NYPIRG has done a good job of compiling the available factual

.information on the chemical pollution of the Niagara River. This compi-lation gives a clear picture of what chemicals are present and in what amounts they are released. Since the data comes from public records, the " reservations" or " qualifications" or " doubts" of the chemical industry concerning the NYPIRG report are in the nature of quibbles.

They do not change the broad picture of what is going into the Niagara River, only relatively minor details. There is little real or substantive disagreement.

Hence, the crux of the argument is not "What is going into the river?", but "What will this do to human beings?" What are the health effects (what deaths and disabilities will be caused) in humans downstream from the releases?

NYPIRG does not make a quantitative assessment of the deaths and disabilities but instead relies on a quantitative argument: The toxics in the releases are known to be hazardous and any excess risk from them is unacceptable. However, this allows industry lobbyists to argue that zero-risk levels are unattainable and unrealistic. They then go on to claim (on television, for instance) that the risks are negligible and the water is " safe". , .

My purpose is to consider the question: Do these claims of safety have any scientific validity?

Two distinctions are essential to avoid confusion on this question: s

I (1) The scientific issue of human health and safety should not be confused with the legal issue of compliance with official levels set by EPA.

(2) This is a public health issue and we must not confuse acute (or short-term) health effects with chronic (or long-term) health effects. It is the latter, deaths from cancer for example, which are more serious hazards to human health and safety from the chemical releases.

On the question of compliance, the issue can be settled by physical measurements. However, on questjons of safety, the issue involves study of health effects in human beings--an entirely different matter. For instance, in a recent federal study of nuclear workers at the Portsmouth Naval Shipyard, the installation was in compliance with officially permissible levels. Indeed, on the average the worker had received only one-tenth of the permissible level. Nevertheless, the data shows serious health hazards at the installation, including a doubled risk of lung cancer. Official levels are not necessarily safe.

Compliance is not the same thing as safety (Attachment A makes this point in detail) .

There seems to be a widely-believed myth (often asserted "in the nene of science") that when a hazardous chemical is diluted with enough water the health hazards simply disappear. It is true that the acute ( short-term) effects of some chemicals do disappear in this way.

However, the chronic (long-term) effects that are the crux of the public health problems are not diluted out of existence and may be made worse by the wider dissemination involved in dilution. This is why standards

. that depend on dilution factors for mutagenic materials may provide no guarantee whatever against serious health effecta. ,

A simple example may clarify this point. Suppose we have a dose of a toxic substance that will produce an acute effect such as nausea. If a single individual gets this dose, he will get sick.

However, if the same amount of the material is diluted by a factor of 1000 and given to 1000 persons, no one will get sick. This is because an acute phyciological effect that, involves the gastrointestinal system requires enough material to affect billions of cells. However, the long-term effects occur when a chemical produces genetic damage to the DNA of a single cell--a break-point in the structure of the DNA. One molecule of a mutagen can cause this damage. Fifteen years later the clone of this single cell could cause clinical canaer. Note that if a dose of the mutagenic chemical is given to a single individual it can only harm this one person. However, if it is diluted and given to 1000 persons, it can cause numerous cancers. Thus, long-term health effects can be increased by dilution.

This point can be made in technical terms by estimating health effects (e.g., death from lung cancer) per unit weight of hazardous material (e.g., grams). We can then ask: Are the effects per gram greater at low doses (high dilutions).or at high doses (low dilutions)?

I'or acute effects, the effects may occur at high doses but not at low doses. For cancer and other chronic diseases, the reverse tends to be true. In a recent study (Attachment A), the effect of a mutagen turns out to be over 100 times greater at low doses than at high doses. So it

_4_

is a dangerous myth that the dilution of toxics in the Niagara River solves the health problems. .

Compliance with standards based on concentration of chemicals in drinking water may be better than non-compliance but compliancc does not guarantee that the health and safety of the public has been protected.

Water meeting these standards may cause serious health problems. Even though we may not see these problems right away, they may occur 20 years later, or they may be moved downstream and into Canadian populations, or they may be postponed to future generations. However, the total deaths and disabilities may bo increased if the toxies are diluted and more widely disseminated.

Perhaps it is now clearer why the public assurances of safety thatareissuedbythespokesmenorthefriendsofthechemicElindustry have been largely based on a confusion of cumpliance with safety and on the dangerous myth that the health hazarda of toxics vanish when they are diluted. Does this mean that the safety assurances that the chemical industry has given on the NYPIRG report.(and on many previous occasions when there have been reports of serious health hazards from dumpsites or i

occupational exposures or other chemical exposures) have no valid scien-tific basis and are worthless?

l I'm afraid so. Since the chemical industry has been issuing bland assurances of safety for the past 50 years in Niagara Falls and l

elsewhere, and since there are a number of studies involving health effects in Niagara Falls and Niagara County, there is a direct way to assess the validity of these repeated assurances of safety that have I

i l

i I

L ., . _. . -1

I been given the public and the workers. According to the defenders of the chemical industry, the politicians, public relations staff, scien-tists, and other industry allies, Niagara Falls should be a healthy place to live despite the concentration of the chemical industry in the city. We can see just how healthy Niagara Falls actually is by taking the death rates, compar ing them to other areas in upstate New York or in the nation, and estimating the excess deaths for causes of death that Public Health scientists have found to be related to chemical exposures.

In this way, we can see how many residents of Niagara Falls have probably been killed by chemical exposures during the period that the spokesmen for the chemical industry have been issuing their many assurances of safety.

In a recent State Health Department study of Love Canal, for instance, the census tracts in Niagara Falls had significantly i- her death rates for lung cancer (and other cancers) than the average for upstate New York census tracts. This was true whether the Love Canal tract was included in the figures.or not. One trouble with these survey figures 13 that the numbers are fairly small and give an erroneous impressica of the death toll that the chemical industry has caused by occupational hazards, dumpsites, and other releases of hazardous chemi-cals. To get a clearer picture of this toll, it is worthwhile to take a quick look at national figures. .

The U.S. Department of Health, Education, and Welfare report (DHEW Puulication No.(NIH)74-645 on "U.S. Cancer Mortality by County:

1950-1969) gives some idea of how many persons in Niagara County have

-- _--. g -

~

. . .. been killed by hazards that the chemical industry claims do not exist.

For example, for lung cancer in men there are 922 deaths and a death rate of 45.5 per 100,000 in Niagara. County. By contrast, the nationwide rate was 38.0 and only 10% of the U.S. counties had a rate over 43.3.

This works out to 160 excess deaths from lung concer. Bladder cancer is another cancer which Public Health scientists have found to be related to chemical exposures.. There are 155 mais deaths and a death rate of 8.2 ir Niagara County. The national rate is 6.8 and the 90% decile is 8.3. There are about 26 excess deaths.

It would be possible to go on with this dismal listing adding excess deaths from other causes and in women as well as men, but perhaps enough has been said to make the point: During the years that the chemical industry has been.strenu6usly denying that there are serious health hazards from exposures to toxic chemicals, hundreds of persons in Niagara County have been killed by these exposures. This record of the safety assurances of the chemical industry does not inspire much confi-dence in the latest pronouncements that, despite the heavy burden of toxics going into the Niagara River, the water is " safe".

Finally let me add one personal comment. I have been a Public Health scientist for more than 30 years, I've published more than 300 articles'and three books, and I know most of the Public Health scientists of stature personally or by reputation. None of this stream of safety assurances has been issued or endorsed by any scientist of stature, and I don't believe that any competent scientist could or would make such an assurance. The evidence of hazards, such as the NYPIRG

?  !

report, is too strong and the evidence of safety is nonexistent.

So I particularly resent it when thes'e false assurances of safety are made .

"in the name of science" because this is bad for genuine Public Health -

and other sciences. Nor do I think it helps the chemical companies in the long run when they issue these incompetent and irresponsible claims of safety that actually endanger the health and safety of the citizens of Niagara Falls and other areas.

e e

b o

e.

e e

9 e

6

t' Irwin D.J. Br:ss. Ph.D. s 6irector cl Birst:tistics Rosw:ll Park M:mori:1 Institut]

666 Elm Street Butf ato. N.Y.14263 [r_ g (i ? -

..,,.i...

u..,,......,,...,,.,i,,.i...,..,=,,,.s.....,...,......

a.. . .. . .w. ~ ~ v s,... s...., o.,.,,,

82 IN 10 P 1 :39 September 16, 1981 4

H. Wade Patterson Editor-in-Chief HEALTH PHYSICS Hazards Control Department Lawrence Livermore Laboratory Livermore, California 94550

Dear Sir:

In " Genetic Effects of the Atomic Bombs: A Reappraisal",

vSchull et al (1) examine a series of indicators of genetic effects,,

including the infant and childhood mortality among the children of parents exposed to ionizing radiation from the A-bombs. "In no instance ,

is there a statistically significant effect of parental exposure." In this situation, the standard practice in biostatistical-epidemiological studies of this kind would be to report the negative findings and then stop. However, in the cited paper the authors go on to attempt to estimate the genetic doubling dose for radiation and report the average

( doubling dose as 156 rems. The authors then assert that this shows that 468 rems might be more appropriate than the recently adopted 100 rems of a committee of the International Commission on Radiological Prr,tection.

The purpose of this letter is to make three points: First, l

the method used by the authors to estimate the doubling dose is statisti-cally invalid. Second, when the data in Table 3 (1) is reanalyzed by a valid statistical method, statistically significant genetic effects are

HJ Wada Patterson HEALTil PHYSICS September 16, 1981 Page 2 found in this Japanese data. Third, when these significant effects are used to estimate the doubling dose for genetic damage from ionizing radiation the new estimate is about 1 rem. Hence, far from signaling a relaxation in current exposure standards, the Ja anese y data when properly analyzed suggest that there should be a drastic tightening of these standards.

The invalid doubling dose estimate is based on a regression coefficient for joint parental exposure (0.00085) which is Icss than half as large as its standard error (0.00213). For an estimate to have credence, the usual standard in biostatistical practice is that it be about double its standard error (which corresponds to statistical signifi-cance at the 5% level for a two-tailed test) . For preliminary estimates it might be acceptable to use " borderline" results where the ratio of the estimate to its standard error was between 1.5 and 2.0, but it is unacceptable to use estimates which have a ratio less than unity. For such estimates the " message" about radiation riske is obscured by the

" noise" (i.e. , sampling error) and the result will be a doubling dose estimate that will be a large number but a meaningless one.

For a valid estimate of doubling dose it would first be neces-sary to demonstrate dose-related genetic effects in the Japanese data that would be statistically significant at the 5% level or.would be reasonably close to this. The linear regression technique used in the

! original analysis of the data on mortality in Table 3 (1) failed to find any significant effects for joint parental radiation exposure (i . e . ,

retrospective dosimetry for gonadal dose for each parent) . However, l

i .

+

HI Wada Patterson ,

HEALTH PHYSICS September 16, 1981 Page 3 there are three possible explanations for this negative result. The first is that there are no dose-related effects, the second that the 1 sampling error in this data (particularly in the cells of the contingency table with relatively few deaths) cbscured real genetic effects, and the third that the statistical methcd used (particularly the underlying mathematical model) was inappropriate for this, data. More specifically, J

if the linear regression model simply did not fit the Japanese facts, a statistical analysis that avoided arbitrary assumptions might give significant results for the Japanese data.

As a first step toward a valid estimate of the doubling dose for genetic damage (a more comprehensive analysis would be preferable but would require going back to the basic data rathey than using the information in Table 3), I consolidated the original 5x5 contingency table (Table 3) into a 3x3 table. This avoided having erratic cells with 5 or fewer deaths and reduced the dosage categories to " low dose" (combination of 0 and 1-9 rem categories) and "high dose" (combination of 10-99 and 100+). The numbers were now large enough to get informative simple standard confidence intervals for the 9 cells in the table and to avoid prior assumptions about the dose relationships. By using this simple and straightforward statistical approach on the data of Table 3,

j. I found clear and statistically significant relationships between parental dose and childhood mortality. This took less than an hour (using my personal TRS-80 microprocessor) . The results are shown in Graph I and

~

any reader can easily verify them. Significant differences occur when series have non-overlapping confidence intervals.

' t

a H. Wada Patterson HEALTH PHYSICS September 16, 1981 Page 4 The first comparison that I made was between the subseries where both parents had received low doses (under 10 rem) and the controls where both parents were "not in city" (NIC) and therefore were not exposed to the A-bomb blasts. There were two reasons why I made this conparison first. For one thing, the numbers in the subseries were large and differences would be casier to detect. For another, the findings we had previously reported in " Cumulative Genetic Damage in Children Exposed to Preconception and Intrauterine Radiation" (which had appeared in Investigative Radiology because these exposures are to diagnostic x-rays) suggested that this cell of the contingency table would show the largest genetic effects (2) .

As'can be seen from Graph 1, the children where both parents had baen exposed to low doses during the A-bomb attacks show significantly higher mortality than the controls (and, indeed, higher mortality than the other cells in the contingency table). The mortality is increased by about 20% and this fact can now be used to estimate the doubling dose 1 (which is obviously far less than the hundreds of rems previously reported).

l Some readers may be surprised that at high doses the mortality I- is less than at low doses, but in a series of pape 3s (3,4 have given l

the scientific rationale and the supporting human data for this phenome-non. It might be noted that a linear regression model cannot fit the facts adequateiy when this phenomenon occurs. This is why the previous analysis of the data missed the genetic effects and gave non-significant results thror.ghout. This kind of mistake can be avoided if one follows l

as. ...r s .. ,

1,lEALTil PilYSICS Saptember 16, 1981 Page 5 the prime directive o1 modern science that "A theory must fit the facts" (rather than the anti-scientific rule of trying to impose an erroneous linear hypothes_s upon the facts) .

The doubling dose can be estimated from the formulas (1.0) Mx"MO +PxXx b where M is the percent mortality in a subseries with an average dose of x

X rem, M is the percent mortality in the controls, I is the proportion O

of deaths due to genetic damage in the controls, and D is the doubling dose. M, M , and X can be estimated from the Japanese data. Ilowever, O

j P is not well, estimated for this or other human data. If a guesstimate is made, however, the equation can then be solved numerically for the estimate of the doubling dose or its inverse, D~.

In the original paper, the authors say "we could assume that approximately one in each 200 live born infants died before reaching maturity because of mutation (point or chromosomal) in the preceding generation", but actually use a figure of one in 400 "to err on the conservative side". Therefore, to err in the opposite direction by making the doubling dose as large as possible, I will use the figure one in 50. Again I will use an average dosage of 10 rem which will also err in the same direction (in this case the number could be directly obtained i

from the basic data). When these guasstimates are used in equation .l.0:

(1.1) 7.28 = 6.19 (1 + 0.02 x 10 x D~ )

t l

H. Wade to. -w.swa HEALTH PHYSICS ,,

September 16, 198I',

Page 6 Hence, solving for the doubling dose D gives an estimate that is about one rem.

Although there is obviously a huge disparity between an estimate of doubling dose of 156 rem and 1 rem, there is no need for a "contro-versy" here. The issue can be resolved very simply. The same data is being used. The only difference is in the biostatistical analysis applied to this data. the American Statistical Association has already dealt with a similar disagreement involving data from the Portsmouth Naval Shipyard. I believe that if the. Health Physics Society so requested, the ASA would perform a similar office in this instance.

• Ve,ry incerely yoprp,

/

~

Ifwi

/ / W

. . Bro s, Ph.D.

' Director of Biostatistics IDJB/mak P.S.: Glossles of a professionally redrawn Graph 1 will be available in three weeks.

S S

4 e

e 9

REFERENCES

1. W.J. Schull, O. Masanori, J.V. Neel Genetic Effects of the Atomic Bombs: A Reappraisal. Sc'ience, 213 (4513) :1220-1227, Sept. 11, ,

1981.

2. I.D.J. Brr>ss and N. Natarajan: Cumulative Genetic Damage in Children

~

Exposed to Preconception and Intrauterine Radiation. Investigative Radiology, 15_(1):52-64, l980.

3. I.D.J. Bross: Dose-Response Relationship in Radiogenic Breast u~.

Cancer. (Letter to the Editor) Journal of the National Cancer Institute, 60,(4):727-728, 1978.

I.D.J. Bross. Mammography. (Letter to the Editor) Canadian j

! Medical Association Journal, 123_(2):94-95, 1980.

t 5-i e

l I,

e

- _ _ . _ ,g , . - . . _ ._ _. . . - _ -. - ~ , - - _. . . . , , 1 -

e g... .. ,g.t.. .. g.. .. I g. ..g..

.. .,.a . . ,1. . .....e ..g. 4 ?

4. I . .

.1... s. o.

g.. ..!. .g....g .... r .

g .. ...A e .. -

.l ... ....

g ag , . ..

9.T,~.

.. I4 e. l .

I.. g II. g g..l . ..

.,.. .I..

.  !. [... , .,.

g ,

+

,1,( .

..1_...,.. .

,1..

1 I. .. .g, . ...e.

-.1 .... ,g. .... .. ..., .

. . .. .,,e ..11 . f'. .

. .. l L. .. . _ . ,.

g, *T , .4_

, l . .,

l. , ,f ...4 ,. ...

... .ga.... !..! .,

li I l,. ..... ..-..

la.... s. .,l, 8 ..,.s .!.L e .

.. ~ . , . . , . ...... . ... 4.

.. ! I, e!! ,.a. L ..1., , I.. .aa .... .al.

a

. . . .. . .. . . .. . . . . . . . _.. .$s,- !r ... ..  !!. . . ,1 ,. .  !! !s... ..

. . 6... -

s. g .. , . i - *f ..* . , . , . . .. . 3- .*

• b-. I,il i*.

..,4 4 a d 6 elt - . <l.4 I j .,.) .

I. L.4 . e .f *.a... .6 ,.

,I.a.. .I.. t . , . ....l. 4.4 . .,eI44,

p. . . . .

.s). . - .*

... .. .l.. , ..

.. g .I 4. ...s .6,,1..$ 3. * . ... t

. 0 44,a .4..

. 4 6 . .

ee,& t

. i ... .. .

66.J, .6 4., ,t ..e. .. .J ,.. .$ l 6

.e. 4 -$ .I.*... . . , *....

s ,6 g... .

ass. .. . .,(

.,.i.....4 46 i.-- . 4.... .. .&.+A{.

4 o. . . , . . I

.ff.....4.

w

. . ...a,

..... .s . . .. . ..

.. ..;... .... .. f...

.. ,. .6..I~l 6

- . 6 4, s. I',.~4 .~

56

• 4,. I

, . g6.9I,.I .. U ,. .... .... ..I .s,. s .l . .+. . .

t s

4 4.

.6,+,

5

,.1 48. , $. .*... . 4 .,

. , , !. .e...

. .. ..g

. .. . .. .13 . . . . 151 y1 ,,6 L

-, . .4 ' .E.3. . , 4.,T.

a*.I .

., u. .. 2 .. . . .

..... .4

,f4. 4h,. ...

.... .., .,.. . .. - ...L,.46 . . 6 ...e.

...9.. 4

.6 . , . ,

.6 sa l ., ..,4,6

.4

... .l.

.6,4

.. 9  !

. . 1.6 ,I . . . I. ,..

. . , .s...

..m.

14

.. ]

4.. .,,.

a.... .

.I ... I,..I . ..

.4,.. . I. . ,a...I. .g . . .- ..

.. . . a,...&.. .

,,.4) .6.

,I 14 94 .6... . .

.a,. .4 ,# . . , .. 4. a ,6 . 4 .4 j6 .

.,& q .

6. ,$4J.-a , ...r . , 4 &.t .,

y' * ** I4 ,

. ** *e, ..."***d_, *,... 4..-***.

g. ,. , C. 44L 4 .. ,

E,e ) 5.

a

,a.. ...e .. {. t**' *e ' 4 *1.,

.L114 ' .4 4. ,, .L,a ',44.

• a.* * *

• 4 ..- 6,r 4 '.g ,4,.,

.."I .

9. ..

9 ., I 45 T4 . . . I..4 s

.. ,.. .A .. . . .,.. .4, (1 ,-

4.. . T.4. I[

. p ' 6..,

., 1.,

.I". .4&.

• . . t 4 &.

4H .L4.. . pU T.,. . ....

1

u. 4.,4

- -4 . $.,4.... 4

. A ,4.

A ,

1I .

I. 3w. ,, y .4 9 A ...

. 4 .-.. . .. .

-- 4 ... I . +. . 4 ._. .

._ , .. 4.,34 + . -I.

,4 .s 1,~,.....

. . I. .4. ,, I. . , 4 4. 6

..3,,

...., . . i, I

.. g J. .

g. .4 . A. . 44 4%. .. b. .'..:6., 8.p

.I , I. . . 6, .j *s M..4s.: 3 *..+..*4

• • 4... . . .

(

. .....*I.*.- - -. . $ .. ...,.. &

S 4 t. . * . . * ..Ilp.4 .,.- I.

..* s l1e .4 . I. . . . . .*. .*. ...*

., .1e6 .s . > . . ....

.. . &. .* ,.4 .

4 l..+.,6 .I .. .I

. . I_. ._6... _c.4 3 .T 1,.. . 4 &t ..4 4I.I4 + 4. .44 s .a 4 .

.a.. . , . .....

3

.A..+

A. ._& .A. .. .,_. .

. .4 .I _. * . 1A,:,& a _..5.

.1. T_64.4.,9.I ., ......,..

. , . .t,44, . . .

,H, i . ..

. ..I.I. I* ..

.. .I ,

pl. ,, . . .. .. ,.

...a.. ..

.. .b. .( . L . ,1 *..

,I .

.I 9 L

. ,.64.. .I.L. .

... . ...~. _.. . . . .&. ...A.. ..a....4. ..... La I .- . .4

. i.4 .. ....

.... .... . ,.3. .4 .. . . 4 .e . ..... . .

& ... A .

...q. , - .- . _ .

.. .. .... 4 .

4. . . ...

.Lt , .s. ....

.. . . . d.. .i, ..... . . . ...... .s..

,. al.4

. ....g... ... .a6. ...g..

m.y.., ... .... . _ .. . ..... 4 . ..

..J 4.I. 4 4. .. ....

I.... .... . .. . . ..

..eome ..rs

.n **

p.~_. .. . _ _ _ .

..L.4

~4 , .4 .' *~* ..* ... . . . . w V. . .

.... .... . . . .. i ** 4

.J.- . ..

• ...* {*, .

.e

.. _ . .. .._,. ... . . ... 4 L . . .

.9.., ...."t.".-*

- . ~,. ~ .

• 59. ......,.

e

,,.. . ... .. . .. ... .. ..i ...

.... . .... .~ . . ... ... . . ..

.... .p.. . . . .

.a 6.... ..., . - . . . . . . . .

...- ~ ~ . * = . . ' '. * *.. * * *..e. * ......

. _ ..g... * -.+. .... . ..

. _* ** *.. . A .. * .

a ... . .. .. .

. . . .. .. ..*t-*

• * -+-* *.**

• -* **=

. . = . _. .=. . =

=== .. . .. .. . .

.a

.. ,1

. . ..-i... ... ..

. .. . , . . .c

.. 7.. . .

... ...)....

. I. ,, . i..

. ... .. .. ..... . 4 , Q. . ........

m

..A....

. J ( . 1. .~. ... . ,

. ... . . 4 .. . . . .

.. O L..

4 .. _. .

f,, ._. .

. r . . .O. ......

. . t- . _ . . .... ... .. .... . ...

t.

a..... a. ...

. 1 ...

9~ .

. . , . .y. C. _a.4.

4 .e.. .....

I. 4u. . . . . ....._..

_.... .. . . .. ..17.. ..., 1., 1+., ,. -T-~

Q ... .. .~. .

.. .. m.. . . ..... . . - ....

C

.... ~,7.~.,,.*

.....I

..'_I ..t.. .. . . .

..n... .. . T*. . ....

p _.

. ,. . .. . . ,. t9, ., ... . . .. . . . . . . .

... t'.1. 1.

. ~... . .

. . . .. . . .a . . . . . . . - ...

a .~ . .

,,,C. . - .

. .. ._. .:.. s .... .

..s. .. .y . .:T, t .. .14 o ... . . U~. .+ . ... . . .

.. .#, ,.. .,11 . . .., .

n * .. ..

.- ,. .., ,,3 ..,,,.. 1 .._. v ,$.. .. . ..

..,u..,.4.

• .. . +. , ...

.....,. ~ ~.4._. .

. ... . L.' ., . .

~, - ._.. ._ 4.. ., .y .

IIU .. .._.....I. . ....

.. .. e. ..*1.. . .4.- . . . .. . . .. .. .. . ..

_.r. ._

.as. .. ..

.e ** . s..L..T~ . .  %.,- .. ...

.I ~. . ,

. o. . . , . . . . . . ...... ~.. .. .. . ....

.,. 4 ..

_+.. ,. .. . g. . ...._ . . . ....I' I. . . ..

. .. . .L . . T} ,.,. ..,t..

... . ....... .w% .

v.. . u

.u. .%. . ......

. r1.. m.

~_. .

4 ... . . .a

. . . ....T1,

. . . _ . ., , . .. L.s sa.

-.a. ..

... .. . . . . .. . p...

.,,It .,, .__

, p. . *tt. ._

T.*t . .. .. . **1 4.,.,,. . . , .. 4_ _ . .

.7p.~_..

.L.. h. . .

. . . _ .1 . ...._. , .

1 .. _.

. '*a T*... , _.. s- s .

.. T. . .

... t. v. .e . ,+ .s . m.. T,w'.T

.u. s. _.r..

. s s,

._ . . . .s. . ., . . , . -.a$ 2.,12 _. . --

6O . . . _.. ..

w _ ._.. ....

4 a., _. . . . . ,

. . . . . .. . . . ._ . , ,*.,14 .. .., 2 ,.. .. ... .T*1 ,

(.'I t .4.A., . . . ... . , , .u. f.* ..a 4.

. ..j%. . . .

m

. , . ..I. t'. . ., .. . . .e .. . . .I.. .4.. 4* . . . ~ ....

4..

.. .. . T .. ......q ,.. . . . -

4.I.,....~ .....,.Js.. .....-... 9. 'L

... . ... ...._. .. _.a. . . ,

.. 4 . ..

.&r,

.J..-. . . . . .-... w, ,. .

, .. . . ...+.. . ....... _ -.

. . . .... .. .... . ~ . . . .- -

. _ ........A.

n ... .... s . .. ,. .. - ~

O d .,...

..s

"'T+ ._ ...

,. . ,. ., .. ., ., ,... T...r. .r..- - ..

. . , . < .'T+'.*..

.. ,...1 . . . . . e.

(".

!.... . . . . . . . ,. i..

f

.... - ... s.

.4.

_L .

.s

.... . ... _ ..... .. ..... _ . . .4

.... .l . . . .g* ....q.. . . . ...... ._,.

. $4 .

... 6..

C _. . .. . .

,4 .. _s . . .

., . . . . .. . ,....... .a.* T ... . . . ...

O t,

. - .. . . . . .a

.a.. _.. .. .

.s g..

......)...Q......

..s.. .. ... .. ...

. . ~ . .3~. --. ... 6 .6 O

...p.. ..u ....

2 (4- .. .. I..-

....s .... .. . . . . .

e g

+

f r.-.

p .4,

..C. ._......

,,. ...a.... _ . .

4., . . .

.. _4.. . . . .. .-

_.. s ..

.s d ..

g r_ . ., . . . . . . e. -. .

4... .. r.

6 . . . . .

") _.4..

e,. ..t,

. , ,k. 4...

. ~. ....- .. .. ,ss .

y y 4~.. .. ~ . 4-. .. . . .

,.<.,*13 s ,

.... ^... . .. .

_ . . .. s .. . . . . ...

.9 ..

_=e-.<

s

() r ,.. ., ,

.. g .

.... ..y.,... . , _ , ... ... ...

.. ..t. . L. . , _ .. .

.. g. . . .

...  %. . .- e... ..

4, ,p. .

, 4- ...... . . . . . .. . . 6. .

y .6. .

.y) '. . .... ,_ ..

4..., . _ , 24

.*.. . + -

..y ...+.-:. .

. **. *+

(4 . t*..I )7 . . * . - *d-.. . .+. ~ .

4

. .e.a . . . - - , . .

q .w. ..6 C., ...... . .. **... ..

. g .. .

6.f6 J . .

.l

.a

m. ., L I* . t...r.. .,i*, ! . L. .. . .. . ..t. f-. .. r., .

. ... . 5 ... . . . .

r d

...s. ... . . . . . . , .... .. .

,3

.r.... *. .

g.. g ,. . ..,. ,,.... ..,

....}.3. ;.H. .

C... .r. ..,. ..

re... . G.a.L. . ... ...

,, 9 .. . .

C )m .. .,. . ..

.. ..s- ..

4, ,. . . .. .

s.

s.. _.4.. ..- .

..A. ..

.. l

....6 s a ., s 1 ... ,s .u., +

..T . .

U.... ... . . ... .

O 4, ,. . . ..... .. .. ...-- ..

, A' ,. ..

. . . .e. ..... .. O. . . ... .

+.

s.. , ...

,... .. .,. .. ,. . ..... .,.... .... .... . d.p!...

.4.....

2.-. .,....

.. .- .. .. . .....e C. ... ...-

-O , .

. . .4. . 4 .

_.... - T..6 ..

..... . . ...T . - .

, ,. . oJ ... . -

. .u.... Lat.. .

I *,

g . . .,

. .. .. .. .. .. . . . ..... ...... ., e,- . . . . .. . . ,

, , , .w'r

.* . . . , . . . ..e.

....s ,..

... ~ .-

... .. . ~.. . . -

.r .'. .. . . . . ~ *..

l.'.....

,O O, . . . ... ...~.

. . . - - ~..'-~~.' ...--

. ,-1. ,.* 4 .4--

., . '. .~. .. '... .-'

. q..

l

-i ,

.....<. . .y. .

. . . .. .e.. . . ~ . . . .

.,.1.. . ..

= . . .

....... 6... l.. . .. ..

G. ..

~

-.. Z.

-,.. ..L..L. ...

..... L.. ,.

... -. . ,..w

,..O

... .I.,. .,.. ...

...W ... .. .

...'.=.8 .* . . . ...

. . .g

. W b

g. ,g

..Q.

6...

g a s ... . ..

.. . .~.

.. ......... . . -.. .M. ...

..M w..

j... g .p .. .. ... .. i.$.. -. ..m...

. pi.

.... ...F . .

.ah. i..... ..e.. . . .

..J. 9 . . . . .

..a4.. .. ...* ...

... e ..

...I.

... . . . .. 4.. . .

. ... .a .e..

.*6-*

g

. ..$. . . - ... . g.' * . M..

h.

.. . ...g

..... . .. ... .M. ... . . & S .

.4... .

......i = . . . .

.. ..e .. ..

. ..- .gg h..

g9,%.

...h .M.. . . .e g

..g ...

a.

. .4.*- . . - . . . .

. . .. .~,.. .

.... . . ... . . 4. .

_......- 4.

.._. .Q

.... -a". .

2

... ..,.o..

l .._ .. . . . .

t.. ..

. . . . . . . . .e.....-

..t . .. .. . . . . . .

(..

l .

.a.

. . ...w. . _ . .

l

. . I . ...

.. . . . . ..- . . . .,..[. . .. . . . . .

, . . . . .m. .

,.a . . .

.a .

I.. ... .. ..

_..}.

__. . 4 . .

... j.. .. .. ...

.a_.

.4....

. . _ . . .- 9 ..g

.. .. .e.... g.

. . ./ ...

g.eg

.. ..m.i,

g. .

,. . ;.g. a. Q- -

..-. .l-.

.) . . . , ....

.. . . . ,f . ..g 9, ... . .

n.-.... . . . . . ... .... .. .. . . , .

.w.-

.6 .I e..-

i _.

M m m.

6.

....-hph .

O**'-M -

.hMM.. h. __

- .ig m ..e.

h.

. g g p.

[

i .... < .

.... . . .... . . . .~.

p 4 .

4 . .

. l.... ....

e

... .. ._ . . ., . ... . .., . . e.

. A = . ...

c P

,,., c .

b .$

. . , g

• .a - = -

t t

t

s c. ;- s .s s as s ss. . . s a ns Folder 3T MS 16714 A

• I ( A joy dis t 41 cerfs 22,4 o F),g/

f Direct I"stimates of Low-Leve' Radiation Risk,spf Lung : 1e

Cancer at Two NRC-Compliant Nuclear Installations

Why Are the New Risk Estimates 20 to 200 Times the Old Official Estimates?

o.

IRWIN D.J. BROSS Ph.D.." AND DEBORAH L. DRISCOLL, B.A.*

hitector of Biostafistics; Ros}wellPark MemorialInstitute, Buffalo, New York

. s v [.

l Rettived July 8.1981 I An official report oa the heakh hazards to nuclear submarine workers at the Portsmouth

! Naval Shipyard (PNS), who were esposed to low-level lonizins radiation, was based on a i casualinspection of the data and not on statistical analyses of the dosagerespes*a relation.

', ships. When these analyses are done, serious hasards from fung cancer and olha causes of I death are shown. As a result of the recent studies on nuclear workers, the new risk estimates y have been found to be rauch hisher than the official esilmaies currendy used in aatting NRC e

pervrdssible lesets. De official BElk esiimales are about one luns cancer death per year per minion pere ons per rern [51. De PNS data show II9 jung cancer de,eths per year per mlUen

, persons per rem.

' BRIEF HISTORY OF THE PORTSMOUTH NAVAL SHIPYARD STUDY l An unusual aspect of the Portsmouth aval Shipyard (PNS) study was that all j meetings, memos, and other transactions were, by law, put into a public record. Un-i necessary repetition of organizational, methodological, and other details of the PNS .

. study can therefore be avoided by a brief outline of its history.

. I In February 1978, the Center for Disease Control and a subdivision, the National

• Institute of Occupational Safety and Health (CDC/NIOSH), were given a congres. -

sional mandate to carry out the Portsmouth Naval Shipyard study [1]. CDC/NIOSH j was instructed to confirm or deny, and in a timely manner, a report by Dr. Thomas Najarian of excess leukemia and cancer among the nuclear submarine workers at

. PNS. Dr. Najarian subsequently published his report, with co-author, Dr. Theodore *

. Colton, in The Lancef [2]. The House subcommittee, which had heard Dr. Najarian testify, had also heard testimony from Drs. Thomas Mancuso, Karl Z. Morgan, and Irwin Bross, and named these scientists to an Overught Comraittee for the PNS study, additional scientists being named later.  ;

Presented at a symposium on Effects on Humess of Esposure to I.ow levels of fontaing Radiation.

1. Vale University School or Medicine. May I4.1981.

- Addresa reprint requesta to: Dr. trw.n D.J. Bross. Director i bastatistits. Roswell Park Memonalin-

• siisuta,666 Elm Ssreet. Bufrafo. NY I424s.

s 1

i w

o

I i

'l i l *

's s e s as si.. n.aJih MS 1671 IC Gg h. M . ,

?' Folder 3T joy cards 22,4

+

.8 disk 47

-- . s ,

his advisory committee received a first draft of a " final tcport" on the CDC/-

NIOSH study in September 1980. Several members of the comniittee objacted to Iwo i serious omissions in that report. First, the congressional mandate specificatly in-i volved the dosage response relationship between badge dose expctures and leukemia and cancer deaths.The dosage by-latency tabulations required hko been given only j

for leukemia and a few related causes of death and not for other cancers o causes of death. Second, no statistical analysis of the dosage

• i response relationships had been made.

Despite these and other objections, CDC/NIOSH issued its final report in Decem-ber 1980. Without even notifying the committee, it submitted a much shorter version

,y to The Lances that was published on January 31,1981 [3]. On January 26,1981,(and

1. without knowledge of the publication), Dr. Bross subinitted a critique to CDC/

f NIOSH," Radiogenic Lung Cancer Among Nuclear Workers at Portsmouth N .

l Shipyard" which noted in the public record that:

However this" final report" included only a small part of the Poiismouth Naval Shipyard data that was relevant to the scientific assessment of radia-tion hazards. . . . Thefacts in the dosage by laterey tables that were not in the " final report" flatly contradict the statements made by CDC/NIOSH la the

• final report

• and to the prass. [4}

The primary purpose of this paper is to present the data on lung cancer among the >

j# nuclear shipyard workers and the straightforward statistical analyses which dem-onstrate a strong dosage-response relationship in this data. At an open meeting on .

March 17,1981, significant or highly suggestive s@tical dosage-response analyses for lung cancer, leukemia, and other causes of death were presented to CDC/

I' NIOSH by Dr. Bross and Dr. Colton. However, CDC/NIOSH refused to retract or .

- correct the conclusion of its final report: ,

e Finally, in PNS radiation workers, we found no positive dosage response relationships between ionizing radiation dose and mortality for any cause

, j reported. [3]

The statement was true at the time it was made. Since the data had only been" eye-halled" and not subjected to proper statistical analysis for dosage-response relation-ships nothing had been "found," However, as will be seen in the sequel, strong posi-tive relationships for lung cancer are found by any of several statistical tests.

QUALITATIVE RESULTS OF THE PORTSMOUTH NAVAL SHIPYARD Sn!DY

- " After receiving a complete set of tables from CDC/NIOSH in D-ember 1980, the

• i first table considered by Dr. Bross was the one for lung cancer, As is well known l (and as was noted and referenced in the published CDC/NIOSH seport), lung cancer is radiogenic [3). De effects tend to become detectable after about 15 years.

In his letter to Dr. Bross of March 4. Dr. Robbins, then the NIOSH director, stated

~

that "the report reserves Ndgment . . . because of the small cohort size and short pg-latenegriods necessvy to study organ cancers." Dis statement is consistent w D-or-more-year latent period (but it is not consisten't with a latent period of 10 years

,Yy y)' or less). ,

• / /Ac Table I shows the d'ata on which the first analysis was performed. nis data was

/'*5 8' Il ,

abstracted from the full CDC/NIOSH " dosage x latency" table (81 cases) consider-

?oherf reladec f e ,

/ j'

, lag only those cases with a latent period of over 15 years and by consolidating FAe c, A.,r g, ,'y ,,

g dosage categories under 0.5 rem. Dr.,Karl Z. Morgan discued this consolidation in f

1. ' I detail at the March 17 meeting. As he showed,it is almost meaningfess from a do-

/"

f

• f simetry standpoint to make distinctions between lifetime badge doses un Table I also combines the latency category "over 20 years" (which has very few ca .es

• g.

In this study) with the "15-19 year" category.

f y

• t - - . . . .

g ,,

11 11 51 a...alCl37h ,,,,_

. ' s, p" MS 16714 (GAL 22) .

l F1Mes 37

! i- disk 45 cards 22.4 .

c l .

t

• Tabie 3 also shows tt a corresponding sums of" expectations" obtained by addition f fro:n tt e original CDC/NIOSH tabic. From the observations and expectations the

• Standardized Mortality Ratio (SMR) and the incremental SMR (ASMR) can be 8 calculated as measures of radiation effects. If 4 simplelinear mathematical model

' used, the Incremental SMR would be simply proportional to dosage. This result i derived in Appendix I.

?

Note that if the higher doses (more than I rem) are combined, the expected num.

" i bers are similar to those at the lower doses (less than 0.5 rem). The expectations ar

• 6.56 and 7.40, respectively.Therefore, the observed deaths can be directly com

• l

• l in a very simple way. There were 13 deaths obs '

found even by *eyeba!!ing? This difference is statistically significant (5 percent) b

. r l the Sign Test.The simplest measure-of potential radiation effects at a given dose the observed number of deaths to the number expected, the SMR. Under the null n hypothesis that there is no radiation effect (i.e., the assumption to be tested)

• SMR should be unity. This makes possible another cornparison of the" higher" an ,

l " which lower" dosage categories. This can be done most directly by calcu will have approximately a unit normal distribution (mean zero and variance j

1, unity). The details are shown in Table IB. The difference between the SMRs, .

13/6.56 = 1.982 for the " higher" exposures and 3/7.40 = 0.405 for the" lower" ex-I posures is 1.982 - 0.405 = 1.577. Then dividing by the standard error of the dif-ference, a unit normal test statistic is obtaincJ. The standard assumption that the the observed number of deaths will be equal to its expectation under*gr , tha deaths fo!!ow a Poisson distribution can be made. Thisthe means nut!

. /e

• 8 hypothesis. The variance of the SMRs I is the reciprccal of" the expectation sinc l

O 1 E

/

/p# . 7

. VariancGl=f57 Variance O =lEii =ff *

~

Hence, the variance of the difference is the sum of two variances: '

Variance of difference = h + h = 0.2875

• The standard error is the square root of 0.2875, or 0.536. When the difference is divided by its standard error, the result is 2.94. A value of 1.96 would be needed fo

• ' the 5 percent probability !cvel and 2.57 for the I percentIcvel so the results here are highly significant.No critical assumptions are made in either of the simple method

• this result.Therefore,it is hard to deny the conclusion that the excesslung cancers at higher doses cannot be expf ained away by chance or sampling variation. M Henc

- U represent some kind of real effect.These results do not contradict '

among the shipyard workers were normal or lower than normal compared to the G overall population? Of the 98,223 person-years in the seri:s, only 4,423 person-years involve more than I rem of lifetime exposure and 15 years of latency. Whe CDC/NIOSH diluted this 4.5 percent of relevant person-years with 95.5 percent ir-relevant person-years, nothing was likely to be "found" even if fung cancer was strongly radiogenic. It was an incompete'nt way to examine this data for radiation .

effects.

e

.... . , . s e an.sm -

a

?, A I *,

Folder 3T MS 16716 6.Aly b . a

, joy disk 47 cards 22.l M -

. e

/ff/d .

• . . CONTRAST OF NEW AND OLD RISK ESTIMATES

• If a linear relationship between dose and response is assumed for interpofation-a much weaker assumption than lir. ear extrapolation-it can be shown (see Appen-dia I) that there should be a direct proportionality between the radiation dose and al **

^ l the incremen[SMR.

analysis De slope of th6 data using simple of the iscurve fincar regression isFig.1.

shown in theThe reciprocal response of the doubl i variab e ( r), the incrementa1 SMR in Table 1,is plotted against the mid-dose for the a radiation categories. Th'e regression line through the origin is also shown. The siope of this line is 0.081. The reciprocal of this, the doubling dose,is 12.3 rem. Even here there is some question about linearity but, on this assumption, the doubling dose for q lung cancer is around 12 rem. **

, ,* Although the role of susceptible subgroups will be considered later Fig. I suggssts that some persons are vulnerable to doses as low as I rem and this may be one reason .

for the departure fromiinearity. Note that departures from linearity can only reduce the statistical significance of the linear analysis, but that the slope is nevertheless significant at the 5 percent level.

The official BEIR estirnates are in the vicinity of one lung eencer death per year per million persons per tem [5]. Since the 1980 BEIR repor sed much the same data as the original 1972, this extrapolative estimate has no sanged. The corre-sponding direct estimate can be obtained by considering workers in Table I who

.  ? were exposed to I rem lifetime dose or more. ne procedure is shown in Table I A.

f . There are 13 deaths. Using U.S. vital statistics to adjust for age, CDC/NIOSH

. calculations ind;cate that 6.56 deaths would be e.xpected. Hence, there are 6.44 ex.

cess deaths. The 6.44 excess deaths come from 4,423 persc7ees and an average lifetime dosage of approximately 7.7 rem. Hence, to haie the same units as the BEIR estimate [5], we must civide 6.44 by 4,423, divide again by 7.7, and multiply by 1,000.000. nis gives 189 lung cancer deaths per year per million persons per rem.

This is over 100 times the official estimates and completely changes the picture both

, for protection of workers from radiation-induced lung cancer a.nd for compensa- .

tion.

Might the excess PNS fung cancer be due to lifestyle or hazards of the workplace

, other than radiation? This questiot can be answered by considering the radiation- .

exposed workers with less than 0.5 rem exposure in Table 1. Here 7.40 deaths were expected and li tifestyle or workplace hazard were doubling the risks there should be

• about 15 dea'hs. Actually, there are only three deaths from !ung cancer in this series.

. Apparently, the " healthy worker" bias due to the stringent standards used in select-

. ing and clearing the nuclear workers overrides lifestyle or factors other than radia-l tion exposures. . .

How can the striking differences between the direct estimate from the PNS data ,

and the official estimates of lung cancer risks be explained? The probable scientific reason for this discrepancy is that a small proportion of the population has pre-q cxisting genetic damage that renders it extremely vulnerable to low-level ionizing radiation [6,7]. -

. Note that the PNS workers received much less than the 5 rem per year currently permitted by the Nuclear Regulatory Commission-actually receiving only about 0.5 rem per year [3]. Yet this was enough to greatly increase their risk of lung cancer.

Whereas most of the excess PNS fung cancer deaths are probably from this suscepti-ble subgroup, in the high-dose studies this group would becorne relatively unim-

.portant in the total deaths. However,in setting standards to protect nuclear work- .

ers, it is this subgroup which has to be protected. 3 ,

ne role of the susceptibic subgroups was pointed cut in 1972 to the original BEIR

. commission [Bj. ncse susceptibles were not considered in the new 1980 BElR esti-mate [9). However, their role is crucial for both a scientific and a public health understanding of the health hazards of low-levelionizing radiation. ,

i*T$ "*,

' ,  ! 11 11 31

! F Idcr 37 smelC537s MS 16714 JAL 24 *g j joy disk 47 cards

~~ "

g

. . i ,

~., ,

. Now,let us consider briefly the ines.. . question: Why does the new PNS data t

show excess risk of lung cancer when the older data for the workers at the flanford g

Reprocessing Plant in Washington State is said to show no lung cancer relationaip? ,

g There are two answers to this question, one historical and one current. The his-l torical answer is that when Dr. Thomas Mancuso and his co-workers, Dr. Alice

Stewart and Mr. George Kneale, originally analyzed the Hanford data they did find I excess lung cancer and a doubling dose for lung cancer of about 10 rem, similar to

,l the corresponding estimate for the shipyard workers [10]. Later Hanford analyses made by Dr. Ethel Gilbert and Dr. Sidney Marks at Batte!!c West [11] did not find l the lung cancer relationship.

Dr. Gilbert kindly provided reprints of her original report [11] and a brief update l12] covering the years of Hanford experience after the 1974 cutoff and through v

h 1977. Using these reports, the data table shown in Table 2 was constructed for the **

Hanford workers who died of lung cancer in 1975-1977, it is not possible to .make the PNS data tables and the Hanford tables exactly comparable because, for exam-

.; ple, the Hanford data tables have the lowest dosage category as under 2 rem whereas the PNS uses under 0.5 rem, in the PNS data, effects for leukemia, lung cancer, a nd other causes appear in the range from 0.5 rem to 2 rem. The dose-response relation-ships ,aay be somewhat obscured in the Battelle baseline series because of this 2 rem choice.

As can be seen from Table 2, the Hanford data for 1975-1977 which is the data with comparable latency to the over.15 years' latency used for PNS, tends to con-firm the PNS resuhs although the effects are not quite as c! car. There is an anomalous point in the data for exposures over 15 rem, but there are some technical problems in this dosage category that appear also in the PNS data. For example, ad-ministrative doses can result in spurious high exposures.

An independent direct wimate of lung cancer risks from the Hanford series is .

shown in Table 2A. It is a third of the PNS estimate but many times the off;cial BEIR estimate of one per million. None of these estimates, including BEIR esti. ,

t snates, are very precise. All such numbers should be read as orders of magnitude rather than literally. What can be said with assurance is that the direct estimates of

' risk to nuc! car workers are two logarithmic orders of magnitude greater than the of-ficial risks. When the actuat risks are 100 times greater, the cos't-benefit calculations or permissible levels or environmental impact statements based on the officia' esti-mates cannot protect the health and safety of workers or the public.

DISCUSSION OF RISK ESTIMATES -

Time is now running out for the official inter agency policy that low-level lonizing 8 - radiation is " harmless' and for the risk estimates that support this policy. The reason

- A e for this is that during the " harmless"cra there were many unnecessary and avoidable .

~

cxpm.tres to low-level radiation. In addition to those at Hanford and Portsmouth.

Y

• there were the nuclear weapons tests affecting both servicemeriand civilians and ex-posures to medical X-rays (we Appendix II). Since the 15-year latent period has now

• I clapsed, the leukemia, lung cancer, and other diseases caused by the low-Icvel radia.

tion are now showing up even in the studies carried out by the federal agencies.

l To estimate the actual radiation risks from the data that ar'e coming out of the 3, newer studies is not especia!!y difficult btat does require some competence in the bio-statistical analysis of dose-response data. When the Tri-State Survey Data on ,

. diagnostic X-rays was anatyred in the RPMI Biostatistics Department, it required a l  ! sophisticated analysis using a mathematical model to obtain the estimate of about 5 ,

l 8 rem as the dubling dose for rnyefcid leukemia in men that was reported in 1979 (I3]. However, only simple arithmetic of the kind used for the Portsmouth analysis l was rega ' e4 to confirm this estimate using the Center for Disease Control of ser- i[;

i i

. vicemen v ne Big Smoky tests [14] or the Portsmouth data. e l

l

• - a s . . . .u a ne st 13h

):~ M l F Ider 3T MS 1671.!! [ gat .J a

e joy disk 47 cards 22,4 - p i .

l The 5-rem doubling dose for leukemia, now twice confirmed by Independent j studies reported by federal agencies, is very different from the official estimate in the interagency report [15) of well over 100 rem.This again suggests the actual risks g are more than 20 times the official ones. Indeed, there are now more than 30 stuen (see Appendix !!) where the data show positive relationships in human populations exposed to low level lonizing radiation, results which would be statistically impos-l' sible if the official estimates were correct.

1. Official agencies refuse to use the new data for BEIR and other radiation risk "l estimates because they say ihat the older data (such as the Japanese A-bomb data)is j the "best available" but there is no biostatistical warrant for this claim. Chart A con-

-l traels the general nature of the new data (e.g., at Portsmouth Naval Shipyard and i Haeford) with the older data (e.g., A. bomb data). From a scientific standpoint a -

8 population of healthy workers who have never been exposed to high doses of radia, f tion is much more infor; native than a population of sick persons or survivors of the l A-bomb who may have been exposed to hundreds of rem. Again, continuous and

. concurrent dosimetry for monitoring nuclear workers is far superief to retrospective 1 dosimetry that is based on assumptions which are now it: ser!ous nestion. Finally, good statistical practice says that you never extrapolate far beyond the tange of the data when good data in the right ranse is available.

This is nM a criticism of the original 1972 BEIR report [5]. At that time it made sense to use the high-dose data and linear extrapolation because this was the only j way that risk estimates at low doses could be made. However, with the much better data and direct risk estimates available in 1981, scientific evaluation of radiation risks would replace the obsolete older estimates by the newer ones. That this did not happen in theJatest BEIR report suggests that official estimates are no longer a sci.

entific product but rather a political one.

For radiation technologies, as fur other hazardous techno!cgicf. biostatistical- .

epidemiological studies such as the PNS stindy become bogged down in real or

,. manufactured " controversies" involving methodological Issues.

• interpretations," . ,I

.p 5.ilosophica! questions of causality.

Perhaps the best hope is to invoke the " Primacy Principle's In dealing with potentially hazardous technologies, the benefit of the doubt must go to the public and not to the technology.

In practical terms this means that the critics of a te.:hW y must present a sound primafacie case on the hazard and after this 11.2 burde0 of poof shifu to the pro-ponents to show that the technology is safe.

Consider the 5-rem-per year exposure that the Nuclear Regufatory Commission currently allows for nuclear workers such as theie at PNS. The present report and ,

• *

• other cited studies show that the doubling doses for lung cancer or leukemia are in the dosage range that is currently permitted annually. *Iher e is now much more than ,

a primafacie case that NRC permits doses of radiation that are dangerous-a dose

• that doubles the risk of a fatal disease is a serious public health hazard. By the Primacy Principle i tshould now be up to Ae proponents of such exposures to prove that they are safe. Failure to do this should settle the argument and lead to the reduc-tion of the permissible dose below 0.5 rem per year (as was prop osed at the 1978 con-l "

, gressional hea.wg) [1].

l ,, .

t l l ~

l 1

1..... -

i i t

! 5 I l

! i l

1 i

l i

l l

l I

.. - .gy s n . >

,. ** Ij. *

' ll 1181 1 sle 81Cl37j - -

F:Idcr 3T MS 1671 14 s.41. 26 i

• j joy

. disk 47 cards 22,4 //

~

. /h ,

t

.  ! REFERENCES

1. Effect of Radiation on Human Health: Heahh Effects of Ionizing RaJiation Volume 3. Hearings

. .before the Subec wmittee on Heakh and she Environment of the Committee on Interstate and For.

eign Commerce. Nineiy-Fifit Corigress. Serta! No 95179. February 1978

2. Najarian T, Cohon T: Mortatisy from laukemia and cancer in shipyard workers. Lances I:1018-1020, 1978

3. Rinsky R. Zumealde R. Waa*eiler RJ, et a!: Cancer mortality at a naval nuc! car shipyard. Lancet 1:

f a 231 233. 1981

. 4. Bross IDJ: Radiogenic lung cancer among workers at Portsmouth Naval Shipyard. Memorandum to

. enembers of the Oversight Committee, January 26,1981 .

3. Naisonal Academy of Sciences: Report of the Adwisory Cornmittee on the Biological Effects oflonia. /2 ing Radiation (BEIR ReportlThe Effects on Populations of Esposure to Low Levels of loniairs g Radiation. Washingion. D]Q. National Research Council. L'S Govt Prinsing Omce.1972 g .

th

. 6. Bross IDJ, Natarajaa N: Genetic damage from diagnostic rad.ation. JAMA237.2399-2401, 1977

'7. Bross IDJ Natarajan tv Cumuistive genetic damage in children capos <d to preconception and in- .

trauterine radiation. Investigative Radiology 15(13-32-64, 1990

8. Bross IDJ, Naiarajan N: trukemia from low. level radiation: Identification of susceptible c!didren.

New Engl J Med 237:107-Il0s1972

9. Nailonal Acadeln'y of Sciences: The Effects on Populations of Eaposure to Low Levets oflonizing '3, i Radiation. Committee on the Biological Effects oflonisies Radiation (BEIR). Washinston. Naiion- g.

s! Academy Press,1980 ,' A

10. Mancuso TF. Stewart A. Kneale G: Radiation esposures of Hanford work,ers dying from cancer and other causes. Health Physica 33:369-383,1977

, 11. Gilbert ES, Marka 5:An analysis of the mortality of workers in a musicar facuity. Radiation Re-search 79:122 148,1979 /,

J .

12. Gilbert ES, Marks S:An updated analysia of morta!ity of workersin a nuclear facuity. Radiation Ra-search 33:140-741,1980 /' .,

13. Bross IDJ, Ball M. Falen 5: A dosage response curve for the one red range: Adult risks from diagnostic radiation. Am J Publ Heshh 69:130-136.1979

14. Caldeell CG, Kelley DB, Heath CW Jr: Leukemia among participants in military maneuvers at a nuclear bomb test. A preliminary report. JAMA 244(14):1575-1318, 1980

• 13. Interagency Task Force ca the Heahh Effects ofloniains Radiation: Report of the Work Group on

. Science, June 1979 ',

,f 9

/

e 4

h s

.. .-. . j c .%,.. .. .

VM

~#.

Y; ' " .

'd " I 11 12 31

• l F:Ider 3T MS 167124 Fils clCl37jl ..

h.

g

,g

. l Joy disk 47 cards 22,4

- . APPENDIX I

'i For a given individual, the probability of dying from lung cancer in a given person-I year can be expressed afgebraically as a function of a baseline risk for a white male I

! ofand histheage, R(z),ofand reciprocal the radiation the doubling dose is U,dose then: (X). If the dosage-response curve is linea

P = R(e) (1 + UX) [1]

I-The expectation (E) of the observed number (O)in a given dosage category can be obtained from equation (1) by summing over age specific person-years within the category. If there is no radiation effect (U = 0 and the doubling dose is infinite), '

and if the rates from riational vital statistics are used for R(t), then the evpectation I under this null hypothesis (E.) would be equal to the " expected" numoer in the .-

1 CDC/NIOSH tables. If Uls a positive quantity and if the mid-dose for a dosage

! category is taken as the exposure for all persons in the category, the summation is h taken as the exposure f or all persons in the category, the summation leads to a non-null expectation .(E.) which is simply related to the null expectation (E.) by:

. E.

• E , (1 + UX) ,

[2]

A commonly used statisde in this kind of contingency table is the Standardized -

Mortality Ratio (SMR) which is defined as the ratio of the " observed" number (0) to

{ ghe null expectation (E.), O/E., The SMR will be unity if there are no effects.

. Hence, another measure of effect is the fairemental SMR (ASMR) which is SMR - 1. From equation [2] It follows that the expected value of the increm:ntal SMR is simply proportional to the inverse of the doubling dose (U): .

O EE _ 3 , E(0)& _ 3 , E UL+ W _ 3 , gy (3)

• i Hence, if the incremental relative risks are fitted by a straight line which goes ,

through the origin (0,0) of a graph of Y = ASMR agairat dosage X, the slope of this line wdl give an estimate of U.14 redpocal estimates the doubling dose.

. l

. l l

t !

1 i

{

l [

l i

l l

i I

i

• .. .. ' . t u4 ,...,..

..s ..a..wst/s s a*RINTirm d*. Sj' .

81 Cl37 Biology a Medicine a5 ,3Q

. ,d * ~ # 5 File 81Cl37j ft 11 12 81 f.~

. j Folder 3T MS 1671-26 g e

. joy disk 47 cards 22.4{A

• s

- j

• APPENDIX 11 Biostatistical Studies of Popufgrions Exposed to Low-l.evel lonizing Radiatiott Where Positive Health Effects Appear in fte Data (By Type of Esposure)

Medernt X. Rey Prenatalirradiation and childhood malignancy: a review of British data from Bdhe9 JF. 51 smart AM:

• she Oxford sur=ey. Bris J of Cancer JI:275 237. 8975 ,, ,%

• ' l Bress

• rently 103: Hazards permitted by thetoNuclear persons exposed Regulatory to ionizing Commission. radiation Presentation to (and the UStoNuclea

- ., /"*** 'lshe

. l Commission. Aprit7.1973 Bross loJ. Ball M. Falen 5: A dosage response curveafor the one rad range: adult risks fro ff l radiation. Amer J Pebl tirahn 69130-836.1979 s

• Bross IDJ. Ball M. Riepha T. et at Prehminary report on radiation and heart dissaan. J Med I mg Brosa IDJ. Naiarajan N: Leukemia frora low.- level radission: identification of susceptible child l- Engl J Med 237:107-110.1972 e

Bross IDJ.Natarajan N: Genetic damage from diagnostle radiation. JAMA 237;2J99-2401.1977 3ross IDJ. Natarajaa N: Cumufstive genetic damage la childres esposed to precanceptio trauterine radiation. Investigative Radiology 15:32-64. 1930

. Diamond EL. Schmerter H. Lilienfeld AM: The relationship of intrauterine radiation to subsequent mortality and development of leukemia la children. Amer J Epid 97:233-3i3.1973 j, Gibson R. Craham S. LIhenfeld A. et sh triadiation la the epidemiology of leukemia amoeg

. J Nat Cancer Inst 45.301-311.1972

• Crsham S.1svia MI, Llhenfeld AM, es sh Preconceptica, intra-uterine, and post-metalirradiation a e.

• related to leukemia. Nat Cancer Inst Monograph 19:347-373. 1966

,, g Cens I*W. Atkinson HR: Medu:al radiations and leuksemia: a retrospec6e survey 393.1964 Macht SH. Lawrence PS: National survey of congenita! snalformations resulting frorn exposure to '

442-466, 1955 roe,tgen n radiation. Amer J Roentgenol 73:

MacMahon 8: Prenatal n. ray exposure and childhood cancer. J Nat Cancer last 23:1!71-1 .

Matanoski CM. Seltsa R. Sartwelf/E. et sh The currens mortality rates of radioicsists and o

~

{e 4.' philiclaWpedaGaTdea':E frorn a!! causes and from cancer.

en decades. Aroer Abstract pre. J Epid ICI:G3.t9

• Matanoski CM Settser R. Sartwett TE. et at: Risks of radslegists over sev,14.1979 *

'sented at the meeting of the Society of Epidemiolos! cal Research. Ju*s Reproductive es.

! Meyer MB, Tonascia J: Long term effects of prenatal a ray of humt.a females. Is

, l perience. II. Crowth and development. Ill. Mortality and sr.e t.dit.s Amer J Epid Ilf(3):

1981 .

. 'l e .= e ..e e e

i t

l i

'. e . . .-s l 6ClJtL ,

< [ 'i# MS 167128 [ gal s) 3 ..

! Folder 3T

, joy disk 47 cards 22,d -

?

l Stewart A. Pennybacker W. Barber R: Aduls truksemias and diagnosu a4ay. Brit Med J2.882-890 "

Nuciter Weapons l Abmb clean t,p. Radioactiw Readings 1(3), August 15, 1980

, Cald. ell GG, Keney DB, Heath CW Jr: Leukemia among participants la military ma vers at a nuclear bornb test. A preliminary repo% L'MA 244(14):1575-1575, 1980 Conard RA, et at: Summary of thpJ findings in Marshanese 22 years a' aposure to radioactive fallout. la Rad.ation. Associated Thyroid Carcinoma. Edited by J d Stratton,1977, pp 241-257 New York, Gruns and h

Foege WH: Testimer before the Subcommities on Heskh and the Eavironment,IJS House of Repre.

sectatives, January 25,1978 Johnson KJ: Evaluation of the hazard to residents of areas contaminated whh plutonium. Proceedicas 4th Internat Cong of the Iniernas Radiation Prosertion Assoc,1977, pp 241-246

• Lyon JL, Klauber MK, Gardner JW et at: Childhood 44r4 associated with fattout from nuclear ',-

testics. Nee East J Med 300.397-402,1979 Rallison ML, Dobyns Bad, Keating FR, et a!: Thyroid disease la children. A survey of subjects potendaHy caposed to fallout radiation. Amer J Me,d 56:457-463, 1914 Schull WJ, Unanori 0 Nest JV: Genetic effects of the atomic bombe: A reappraisal.'

Science 283:I220-

• • e* 1127,1981 O

• g*[/ *

• e Solomon N: US marines uked in Nagasaki, The Progressive,1979, pp 2147 #

/L A4 Orewpationel E.uperurae Anderson 1W: Radiation esposure of Hanford wo-ters: a critique sJ the Mancuso, Stewart and Knea!e f report. Health Physks 35]43-750,1978

/O /p Archer VE, Gilliam JD, Wagoner JK: Respiratory disease mortality among uran!um miners. Ann NY Acad Sci 273:280-293,1976 Austin DF:A study of cancer incicence in IJwrence Livermore Laboratory employees. Unpublished ,

report Colton T, Najarian T: Results of second PNS study. Testimoey before the Subco'mmittee on Health and Scienufic Research, June 19,1979

. ,I i .

WICKERSHAM PRINTING-Photo-Comp,

< Biology & Medicine #5 l 81-Cl37 11-12-81 File 81Cl37k Folder 3T MS 1671-29 (dfC3'Q) joy d:sk 47 cards 22,4 // ,

, Evans HJ, Buckton KE, Hamilton GE, es at: Radiatico4aduced chromosome aberrations la nuclear.

dockyard workers. Nature 277.531-534,1979 ,

Gilbert ES, Marks S:An analysis of the WWay of workers la a nuclear facility. Radiation Er.carch 79:122-la8,1979 .

q Gilbert ES, Marks S: Aa updated analysis of enortality of workers la's socitar facility. Radiation Ra vaich 43.740-745,1980

.g Kneale GW, Stewart AM, Mancuso TF: Reanalysis of data relating to the Hanford study of the cancer risks of rad:ation workers. Proceeding. of the International Atomic Energy Agency, Internationa!

Symposium on the Late Biological Effects of lonizing' Radiation. Vienna, March Il-17,1978 Mancuso 1T, Stewart A, Kneale G: Radiation esposures of Hanford workers dying from cancer and other causes. Health Physics n:*69-385,197')

[ Maham S J Occupational mortality la Washington Stsie,1950-1971. HEW Publ No (NIOSH)76-175,

' e Va* I,1976, pp 29-30 a Najarian T Colion T: Mortality fram inkaemia and cancer la shipyard nuclear workers. Lancet I:ICIf. ,

im,197s (p.~2.e,) ' - Najanan T Castleman 5: Islow dose ref ation associated with sayeloma7 New East J Medg l278,1979

~ Rinsky R, Zummalde R Waaweiler PJ, es al: Cancer mortality at a nava! muc! car shipyard. Lancet I:

y p ** 231-235, 1981 I e l

e l

l I

l l

l ._ _ _ _.

.+, ..... .. .

. , . . . 11 13 81 File BICl3711 *'

• b i Foldef 3TT MS 1671 17 ..\l.h g.g 4 g

. joy disk 47 r#rds 224 TABLEI Observed and Espected Deashs from Long Cancer by Radiation Esposure for Workers at the Portsmouth

, Nava! Shipyard. $ssadardued Mortatny Ratios (SMR) and incremental $MRs (ASMR) e also shown.

Radiation Esposure (in Rem)

Ouantay

. 0.001 0. I.f4 3.00- If 0+

a.4,, o.,500-

,, 4.,; 14.,,

Observed Deat hs 3 3 6 6 3 Espected

• Deaths .7.40 1.67 3.69 1.00 1.07 SMR 0.41 1.80 1.63 2.22 2.80 '

ASMR - 0.59 0.80 0.63 f.22 1.40 TABLt 1A Calculation of she Risk of Lun6fancer la BEIR Unks from the Portsmouth Naval Shipyard Data (caposure k 3 rem)

Basic Dere

, Encess deatha frees lung cancer = 6.44

Average lifetime dose (in rem) = 7.70

Perios years = 4,423 l

_ Calculareen l Units are deaths per million per year (!O'/ year)

= 189 deaths per 10*/ year /ree 4,423 y x 7rm t

,e

- WICKERSHAM PRINTING-Photo Comp. 'I 81-Cl37 Biology & Medicine #5 11-13-81 File 81Cl3711 (Il

• Foldef 3TT MS 1671-21 joy cards 22,4 gag 3' TABt.E 2A * .

Calculation of the Risk of Lung Cancer in BEIR Units from ,

q the Haarord Data (caposure a 2 rem) .

Basic Dese Encess deaths from long cancer = 6.10 Average lifetime dose (in rem) = 6.70 .

I Person years = 14,027 e.

o Cniculation Units are deaths per mittion per year (10*/ycar) ,

8 6.8 deaths w 10'

• 14,027 person years a 6.7 rem * ' P I I'

f f

• l , FIG. l. Escess Lung Cancer (ASMR) by Ufetime Radiation Dose (in Rem) for Radiation.

Espose? Workers at Portsmouth Nava! Shipyard (Broken line = data pointsi Sotid Ene = .C,-

fitted regression line). {

l

'. j Il b a rese alt.util

• Folder 37T MS 671 19 GALhg,,pg

. j :y di:k 47 ^

cards 22.4 TABLEIB An Elementary Significance Test for The PNS Lung Cancer Data SMR for 1+ rem i *02 SMR for under 0.3 rem [O,S .

Difference I.a??

Variance of

= h Variance O = h = f /

7 variance of difference

• m /s' "I

+6 = 0.2873 d 8 S.E. of difference s V0.2873 = 0.314 Difference Ig '

S.E. of difference , 0.334 = 2"N'*

Critical Values a.e 1.94 (5%) and 2.35 (!%)

TABLE 2 Observed and Espected Deaths from Lung Cancer by ,

Radiation Esposure for Workers at Hanford Reprocessing

'; P: ant. Standardized Mortality Ratios (SMR) and incrementa!

f SMRs (ASMR) are also shown. (1977 - 197d)

, Radiation Exposure (la Rem)

Quansley 0-2 23 3-13 13+

')bserved Desiha 13 10 6 3 Espected

. Deaths 19.I 3.9 3.6

' 3.d .

SMR O.68 1.69 1.67 0.88

• l ASMR - 0.32 0.69 0.67 - 0.12 .

o IINTING-Photo-Comp.

-Cl37 Biology & Medicine #5 13 81 File 81Cl37ml gl'l alder 3TT MS 167123 GAL ny cards 22,4 CHART A Comparison of the New Data on the Ponsmouth Shipyard Workers with the Data Used la Official Reports (Interagency. BEIR. ICRP. cie.)

Characteristics of the Data PNS Data Official Reports

/ho are the persons under sindy? Nuclear workers unde't normal working Survivors of an A bomb or persons with grave

• conditions disease requiring therapeutic X ray that are the dosages of ionizing radiation? Low level radiation directly pertinent to Dosages la most subjects of mell over 800 rem occupational caposure standards s'has is the quatiry of the dosimetry for persuns Continuous concurrent monitoring of the Retrospective estimaies of caposures mithout under study? esposures with recording of daies. doses. adequate crosschecks or conirol of the h_

etc.. for each individual dosimetry Vhat is the quality of the fdow-up of the - Virtuaffy complete (90%) with full death incomplete and often inadequate follow-up and persons under study? certificale and other information poor qualily of information on individus's .

Nhat was the quatily of the inferrnation used Hosology review enabled use of age sca race- Pick up or biased comparison series (e.g.,in

,for comparisana? cause specific U.S. rates some A-bomb coniparisons. persons esposed

. up to 10 rem are used as conirots)

Whai assuraptions were necessary for estimates Estimates can be made directly without any Esiimates require assumpilon of dubious

. of doubling done or cuher quantitative strong assumptions " linear

• or other hypotheses and are mere?y measures of heshh effects? guesswork

\

\s

.