back to CNR | radiation | rat haus | Index | Search | tree

plain HTML | ASCII text formats

Cancer in the Family:
Does Each Case Require More Than One Cause?
The Likelihood of Co-Action.

John W. Gofman, M.D., Ph.D., Prof. Emer. of Molec. &
Cell Biology, U.C. Berkeley; Chairman of CNR.
Egan O'Connor, Executive Director of CNR.

April 1999.  Keywords: Cancer-Rates USA, 1900-1990;
Cancer Causation (Etiology); Synergism.


Part 1  o  How Do Cancer-Rates Now Compare with Past Rates?

        o  In 1993, there were 2,268,553 deaths in the USA, of which 529,904 were cancer deaths --- 23.4% or almost one death out of every four. The ratio of cancer diagnosis to cancer death is approximately 2 to 1. The ratio varies a lot with the kind of Cancer which is diagnosed.

        o  No one can speak with certainty about the overall ratio, of incidence divided by mortality, because incidence rates are analyzed from only a few states plus a few selected metropolitan areas, whereas mortality rates are analyzed from all 50 states. If the overall ratio is 2 to 1, it follows that about 45% of the population (USA) receives a diagnosis of Cancer. The behavior of cancer rates over time is shown in Table 1 and in Figure 1.

        o  Every family with a cancer problem asks "Why? What was the cause?" This publication addresses the likelihood that a case of Cancer results from the interaction of more than one cause.

Table 1: Years 1900 through 1990: All-Cancer Death-Rates, USA, across Time.
Because these are age-specific rates, per 100,000 population of each age-band, different decades can be directly compared without age-adjusting. Prior to 1933, not all states reported. Entries for 1900 are based on the fewest states (only ten). Sources: Vital Statistic Rates in the U.S. 1900-1940, p.250, by F.E. Linder (1947: Government Printing Office), and Health, United States, 1995, p.132 (1996: National Center for Health Statistics).
Age           1900    1910     1920     1930     1940     1950     1960     1970     1980     1990

Under 1 year   3.2     3.1      3.2      3.1      4.4      8.7      7.2      4.7      3.2      2.3
Ages 1-4       2.9     3.3      3.1      4.1      4.8     11.7     10.9      7.5      4.5      3.5
Ages 5-14      1.8     1.5      1.6      2.0      3.0      6.7      6.8      6.0      4.3      3.1
Ages 15-24     3.2     3.5      3.8      4.2      5.4      8.6      8.3      8.3      6.3      4.9
Ages 25-34    14.0    14.1     14.7     16.7     17.3     20.0     19.5     16.5     13.7     12.6
Ages 35-44    52.5    55.6     56.0     58.9     61.1     62.7     59.7     59.5     48.6     43.3
Ages 45-54   139.1   156.7    155.1    159.6    168.8    175.1    177.0    182.5    180.0    158.9
Ages 55-64   260.9   322.4    341.2    355.6    369.6    392.9    396.8    423.0    436.1    449.6
Ages 65-74   421.0   541.7    607.7    677.1    695.2    692.5    713.9    751.2    817.9    872.3
Ages 75-84   544.7   749.9    900.0  1,019.7  1,161.0  1,153.3  1,127.4  1,169.2  1,232.3  1,348.5
Ages 85+     623.2   810.9  1,017.5  1,196.3  1,319.0  1,451.0  1,450.0  1,320.7  1,594.6  1,752.9
Entries in the three rows below are the age-adjusted National Cancer Mortality Rates per 100,000 population. Cancer-deaths at all ages are included in these age-adjusted rates. Because each entry is adjusted to the same "standard population" (1940), the entries in this row are directly comparable with each other --- despite changes over the century in infant mortality, average lifespan, etc. The male-female difference after 1940 results largely from the males' much higher rate of Respiratory-System Cancers.
Both Sexes    79.6    97.0    104.9    113.4    120.3    127.7    129.1    129.8    131.9    135.0
Males         --      --      --       --       115.0    132.8    145.7    155.1    164.5    162.7
Females       --      --      --       --       126.1    123.2    114.9    111.7    108.5    111.3


Part 2  o  A Popular Assumption Is Probably Wrong

        o  There are several well-established causes of Cancer. For instance, the status of ionizing radiation and smoking, as proven causes of Cancer, is not in doubt. By contrast, the status of some other causes is. For example, the status of low-fibre diets as a cause of Colon Cancer in women, and the status of high-fat diets as a cause of female Breast Cancer, are newly in doubt (New England Journal of Med, 21 Jan 99; + Journal of the Amer Med. Assn, 10 Mar 99).

        o  The term "radiation-induced Cancer" is widely in use. So are terms such as "smoking-induced Cancer," "occupationally-induced Cancer," "pollution-induced Cancer," "estrogen-induced Cancer," etc.

        o  Contrary to popular belief, these terms do not necessarily mean that radiation acts alone to cause a case of Cancer, or that smoking acts alone to cause a case of Cancer, or that a workplace carcinogen acts alone to cause a case of Cancer, or that a hormone acts alone to cause a case of Cancer, or that any carcinogenic agent acts alone to cause a case of Cancer. Indeed, the "general wisdom" is that they do not act alone (Part 3).

        o  Then how should terms like "radiation-induced Cancer" be defined? To allow for the likelihood that a case of Cancer requires more than one cause, one should think of "radiation-induced Cancers" as cases which would be absent (prevented) in the absence of radiation exposure (Parts 7 + 8).

        o  Likewise, occupationally-induced cases are cases which would be absent in the absence of exposure to carcinogens in the workplace. Hormonally-induced cases or diet-induced cases are cases which would be absent in the absence of unfavorable levels of specified hormones or unfavorable nutrition, etc.


Part 3  o  Some "General Wisdom" Which Is Probably Right

        o  The concept of necessary co-actors is certainly not new. In the famous 1964 "Surgeon General's Report" on cigarette smoking as a cause of Lung Cancer, the authors wrote (p.31): "It is recognized that often the co-existence of several factors is required for the occurrence of a disease, and that one of the factors may play a dominant role; that is, without it, the other factors (such as genetic susceptibility) seldom lead to the occurrence of the disease."

        o  The concept, that more than one cause is necessary to produce a case of Cancer, is embraced by the widely accepted initiator-promoter model of Cancer. In that model, inherited or acquired carcinogenic mutations require help from a "promoter" --- for example, a hormone or a virus.

        o  The concept of mutually dependent co-actors is also inherent in the widely accepted multi-mutation multi-step models of Cancer, which propose that a single cell must accumulate multiple mutations before it will produce a fatal Cancer. Current textbooks state that Cancer "is typically a multi-step process resulting from an accumulation of as many as 10 genetic changes in a single cell" (from Understanding Genetics; A Molecular Approach, Norman V. Rothwell, 1993, p.471).

        o  The concept, of more than one cause per case of Cancer, arises from various lines of evidence. For example, inheritance of one mutated copy of a "Breast Cancer Gene" certainly does not guarantee that Breast Cancer will develop in every breast cell --- although every breast cell contains the mutation. One or more additional causes are needed to turn even one of those breast-cells into a Cancer. In human Cervical Cancer, "the cause" is claimed to be infection by the Human Papilloma Virus in 90% of the cases, but evidence indicates that the virus alone does not suffice. It needs help from at least one other co-actor. In Stomach Cancer, the bacterium Helicobacter pylori is estimated to be a cause in about 55% of new cases --- but it needs help from at least one other co-actor.


Part 4  o  How NonRadiation Agents Can Make Radiation Worse

        o  There are several ways in which co-actors can modify each other's potency. For example, a nonradiation co-actor can increase the carcinogenic potency of each unit (rad) of radiation exposure, not by increasing the number of radiation-induced genetic injuries per rad (the number is proportional to dose), but by intensifying the carcinogenic consequences of each radiation-induced mutation. For example, the nonradiation co-actor could block a signal for cell-suicide (apoptosis). Thus, more cancer-prone cells (mutated by radiation) are left alive. Or the nonradiation co-actor could stimulate the cancer-prone cell to divide more often than normal, thus producing a tumor. Or both.

        o  Evidence that co-carcinogens can modify each other's potency has been demonstrated --- it is not just speculation.

        o  The report of the fifth Committee on the Biological Effects of Ionizing Radiation (National Academy Press 1990, p.152) concludes: "As discussed in the previous section, the carcinogenic process includes the successive stages of initiation and promotion. The latter phase, promotion, appears to be particularly susceptible to modulation, with cigarette smoking being a conspicuous example of a modulating factor. Susceptibility to the carcinogenic effects of radiation can thus be affected by a number of factors, such as genetic constitution, sex, age at initiation, physiological state, smoking habits, drugs, and various other physical and chemical agents."

        o  The current Committee, BEIR-6, re-affirms the finding (National Academy Press 1999, p.5): "Radiation carcinogenesis, in common with any other form of cancer induction, is likely to be a complex multi-step process that can be influenced by other agents and genetic factors at each step." BEIR-6 is convinced that synergism occurs between radon-exposure and smoking. It follows that "Some lung-cancer cases reflect the joint effect of the two agents and are in principle preventable by removing either agent" (p.33).


Part 5  o  How Radiation Can Make NonRadiation Agents Worse

        o  In addition to considering how nonradiation co-actors modify radiation's potency, we can also consider how radiation can make nonradiation co-actors more dangerous. For example, radiation can mutate a gene which is essential to let a cell repair carcinogenic damage inflicted on that cell by (say) benzopyrene. The result is a higher frequency of benzopyrene-induced Cancers per unit of subsequent exposure to benzopyrene.

        o  But it is not necessary to disable repair-genes, in order for radiation to increase the carcinogenic potency of benzopyrene. For illustrative purposes, we can imagine Cancers whose occurrence requires the co-action of ionizing radiation and benzopyrene:

        o  Suppose (arbitrarily) that radiation exposure is at a level which produces 3 carcinogenic mutations per 1,000 cells, and benzopyrene is at a level which produces 2 carcinogenic mutations per 1,000 cells. The frequency of cells which receive a mutation from both radiation and from benzopyrene is, therefore, their product: Six cells per million. These are the only cells which can produce Cancers from the co-action of radiation and benzopyrene.

        o  Now we modify the illustration. We double exposure to radiation, so that 6 cells per 1,000 have a carcinogenic mutation due to radiation, but we do not change exposure to benzopyrene. Now (6/1,000 times 2/1,000), or 12 cells per million, receive a mutation from both radiation and from benzopyrene. Twice as many cells per million are capable of becoming Cancers, due to radiation-benzopyrene co-action. The increased level of radiation has modified the carcinogenic potency, per unit of exposure to benzopyrene, upward by a factor of two. In a similar fashion, an increased level of benzopyrene exposure could make each unit of radiation exposure more powerful.


Part 6  o  A "Menu" Consisting of Xrays, Diet, Hormones, Pollution, Workplace, Viruses, Smoking, & Inherited Mutations

        o  CNR has already published evidence indicating that about 75 percent of Breast Cancers in the USA are due to medical radiation (Gofman 1995/96). That analysis embraces the concept that a case of Cancer can have more than one cause, and that co-actors modify the per-unit potency of radiation.

        o  If, in general, more than one cause is necessary per case of Cancer, then what is the meaning of an estimate that 75% of Breast Cancer is due to medical radiation?

        o  The meaning is that 75 cases per 100 cases would have been absent in the absence of medical radiation (Part 2, above). It means that 75% of the cases have medical radiation as a necessary co-actor. And, by definition, if you can remove a necessary co-actor, you can prevent the result.

        o  An estimate, that 75 cases per 100 cases are due to medical radiation, does not imply that xrays acted alone. For illustrative purposes, we can suppose that the total Breast Cancer rate per 1,000 women is 100 cases, and that what causes those 100 cases is co-action.

FIRST LIST of potential co-actors (illustrative):

        o  The list, above, does not attempt to include every type of carcinogen. The abbreviation "diet" refers to below-optimal nutrition. "Hormones" refers to an unfavorable level of one or more hormones.

        o  The meaning of the first row, above, is that xrays + diet + hormones + pollutants each make a necessary contribution to each case of Cancer in the first row. In the absence of any one of the necessary co-actors, the 25 cases in the first row could not occur. That is the meaning of necessary. The meaning of all five rows, above, is similar. If a cancer cell is like an automobile, requiring wheels + an engine + fuel in order to operate, no cell can become a cancer cell if any one of the requirements is absent. Moreover, from a cell's point of view, it does not matter which combination of co-actors supplies the required wheels, engine, and fuel.

        o  SECOND LIST: The first list, above, is very different from a list of 100 cases in which 75 cases are due to xrays acting alone, 3 cases are due to a dietary factor acting alone, 3 cases are due to a hormone acting alone, 3 cases are due to a pollutant acting alone, 3 cases are due to an inherited mutation acting alone, 3 cases are due to a workplace carcinogen acting alone, 3 cases are due to a virus acting alone, and 7 cases are due to cigarette smoking acting alone.

        o  The sum of cases in both lists = 100 cases, but every case in the first list is the result of co-action between multiple causes, and every case in the second list is the result of one cause per case.


Part 7  o  The Concept of Fractional Causation

        o  For the reasons presented in Parts 3, 4, and 5, we (and many others) consider it highly unlikely that Cancer requires just one cause per case. The concept of Fractional Causation, explained below, is based on the requirement for more than one cause per case. Therefore, below, we consider only the FIRST list from Part 6.

        o  The sum of the cases in the first list is 100 cases of various etiologies. Because we specified that 100 cases (per 1,000 women) is the total rate, the sum accounts for 100% of the total. Of course, the sum of cases which exist can never exceed unity (100%).

        o  Out of the mixture of cases in the list, suppose we want to explore how many cases would be prevented if we could remove just one cause, while the other causes remain as they were.

        o  In the illustration above, xrays are required co-actors in (25 + 20 + 15 + 15), or 75 cases per 100 of the total rate of Breast Cancer. Because the absence of a required co-actor prevents the result, 75% of the cases would be absent, in the absence of exposure to medical radiation.

        o  For radiation or any other carcinogen, Fractional Causation is the fraction of cancer cases which would be absent (prevented) in the absence of a specified cause.

        o  Next, we put radiation back into the mixture, and we remove just "poor diet." Then we can calculate Fractional Causation by diet. In Part 6, poor diet is a required co-actor in (25 + 15), or 40 cases per 100 of the total rate of Breast Cancer. Because the absence of a required co-actor prevents the result, 40% of the total cases would be absent, in the absence of poor diet.

        o  In this manner, we calculate Fractional Causation, below, for each of the eight causes in Part 6.

        o  Warning: Because Fractional Causation means the fraction or percentage of cases which would be absent (prevented) by the absence of a specified co-actor, addition of the separate Fractional Causations produces nonsense (a total greater than 100%). Such addition is equivalent to counting the same cases of absent Cancer more than once.

        o  Using the first list from Part 6, we calculate the Fractional Causations for all 8 causes listed as potential co-actors:

  • Xrays: Frac. Causation = (25 + 20 + 15 + 15) = 75%
  • Diet: Frac. Causation = (25 + 15) = 40%
  • Hormones: Frac. Causation = (25 + 25) = 50%
  • Pollutants: Frac. Causation = (25 + 20) = 45%
  • Inher. mutations: Frac. Causation = (20 + 25) = 45%
  • Workplace: Frac. Causation = (15) = 15%
  • Viral agents: Frac. Causation = (15) = 15%
  • Smoking: Frac. Causation = (15 + 25) = 40%

        o  It is obvious that when co-action is operating, Fractional Causation of 75% by xrays does not set a maximum Fractional Causation of 25% by any other cause.


Part 8  o  Implications of Co-Action for Preventing Cancer

        o  One of the interesting implications of co-action is this: Reducing exposure to a single carcinogen reduces the power of all of its partners in causing Cancer. If one can identify a single carcinogen which is a necessary co-actor in many common types of Cancer (which means, the carcinogen will have a high Fractional Causation), then one can make real progress in preventing Cancer by reducing exposure to that single cause.

        o  This year, the Committee will publish strong new evidence that ionizing radiation is such a carcinogen.


Figure 1
Figure 1. The two bar-graphs above are reproduced from p.45 of the SEER Cancer Statistics Review, 1973-1994, which is a 479-page publication issued in 1997 by the National Cancer Institute. SEER stands for Surveillance, Epidemiology, and End Results --- a program initiated under the NCI in 1973 to evaluate trends in cancer incidence (USA).

# # # # #

Committee for Nuclear Responsibility, Inc.
Post Office Box 421993, San Francisco, CA 94142, USA.
An educational group since 1971.

This document is also available in plain HTML and ASCII text formats

and can always be found at

back to CNR | radiation | rat haus | Index | Search | tree