ASCII text format
|
|
My two major points can be summarized at the outset (details and references below), with respect to the valuable New York Times' June 29, 2008 article, "Weighing the Costs of a CT Scan's Look inside the Heart," by Alex Berenson and Reed Abelson.
First, Coronary Artery Disease: The article overlooked evidence uncovered in 1999 (Part 1, below) that x-ray-induced mutations in the coronary arteries may well have had a major causal role in initiating and/or accelerating atherosclerosis (Part 4, below). This is called "Hypothesis-2" in the 1999 study.
Second, Cancer: A credible warning (also based on the 1999 research described in Parts 3 and 4), is that accumulated lifetime exposures to medical x-rays --- which include CT and fluoroscopic exams --- may constitute a causal co-actor in more than 50% of current cancer mortality in the USA. This is called "Hypothesis-1" in the 1999 study.
The "over 50%" estimate is at least as scientifically credible as the popular estimate that medical x-rays account for only 1% of American cancer (Doll and Peto 1981 p.1256, Table 20). I defend this assertion below (Part 3). Both estimates are "ballpark" estimates, meaning that the absence of appropriate data necessitates the incorporation of several important approximations and assumptions.
Part 1. Where Does One Find the 1999 Research?
In 1999, the late John W. Gofman, M.D., Ph.D., authored a monograph entitled "Radiation from Medical Procedures in the Pathogenesis of Cancer and Ischemic Heart Disease: Dose-Response Studies with Physicians per 100,000 Population." Please see Reference List at the end. The monograph is available from amazon.com. Its Executive Summary is available, free, online at www.ratical.org/radiation/CNR/RMP/chp1F.html (see standalone file version at www.ratical.org/radiation/CNR/RMP/execsumm.html), as are the main critiques from peer review at www.ratical.org/radiation/CNR/RMP/6critiques.html.
Dr. Gofman was an honored pioneer in both lipoprotein research (for example, Gofman 1949, Gofman 1950, Gofman 1955, Von Euler 1972, Brown 2007, Havel 2007, Krauss 2007) and in radiation research (for example, Archer 1982, Mendelsohn 1995, Radford 1995, Upton 1999).
In view of his track-record, a prudent regard for public health demands very serious consideration of his final monograph --- especially in view of the expanding use of both CT and fluoroscopy in medicine.
Part 2. Why Quantification of Cancer-Risk from CT X-rays Is So Very Uncertain
The chance of benefit was well explored in the New York Times' article. However, the very uncertain size of the cancer-risk deserves much greater discussion, because the magnitude could be far above the common presumption of negligible.
If large, why is it not recognized, after more than a century of using x-rays in medicine? We can consider three of the reasons.
The principal reason is that the lifetime accumulated x-ray dose per organ is unknown and unknowable even for one individual, unless evidence exists of never having received an x-ray and of not having received pre-birth irradiation during pre-delivery x-ray measurements of the mother's birth canal --- one baby in 13 until 1970 in the USA (Gofman 1995/1996, p.89). Lifetime accumulated dose is what matters because most of the x-ray-induced mutations survive cell-division and accumulate.
The failure for 100 years to measure and record x-ray doses undermines the effort to quantify health risk from both medical and non-medical exposures to ionizing radiation, e.g., from occupational exposures, from nuclear pollution, from the Hiroshima and Nagasaki atomic bombings. A key part of every dose-response study is the "matching" of dose-groups for equal accumulated exposure to additional causes of the ailment under investigation (cancer, for instance). Without reasonably accurate matching, the resulting risk estimates can be extremely unreliable --- very much too low or very much too high, compared with the truth. Matching is a giant pitfall in epidemiology. And its failure in radiation epidemiology is particularly worrisome because the most common source of manmade human exposure ever since 1904 is the x-ray, which is the one for which almost no data exist on accumulated doses.
Because unmeasured accumulated x-ray dose is "the 800 pound gorilla" who inflicts mayhem upon risk-estimates from ionizing radiation, even the best and least biased investigator will produce unreliable estimates of risk per dose-unit. Dr. Gofman scoffed at confidence-limits on such estimates because he recognized all such estimates (including his own) were inherently just ballpark values.
The serious "gorilla" problem in the A-bomb Survivor Study --- where about 85% of participants received estimated radiation doses from the bombs that were not higher than estimated doses from common medical procedures --- was acknowledged as "contamination" by some of its guardians. Kato et al wrote that, by 1982, "The doses from diagnostic x-ray examinations had already become significantly great contaminates of the radiation doses from the atomic bombs" (Kato et al, 1998, p.51).
Nonetheless, in the past and even today, most analysts use the A-bomb Survivor Study as the basis for estimating risk per dose-unit from medical x-ray procedures.
Secondly, estimated values for cancer-risk from medical x-rays are "born seriously underestimated" whenever risk values from the A-bomb Survivor Study are used without multiplying those values by factors between two and four. Why? Experimental radiobiology (not epidemiology) has shown that medical x-rays are two to four times more mutagenic per dose-unit than the less densely ionizing high-energy gamma rays from the atomic bombs (among others: Bond 1978 p.433, Gofman 1990 Chapter 13 Part 4, BEIR 1990 p.218, Straume 1995 p.956, Kato 1998 p.71).
Whether or not David J. Brenner makes that upward adjustment for x-rays is unclear to me from the New York Times article. Dr. Gofman explicitly used it in his analysis of radiation-induced breast cancer (Gofman 1995/1996 p.278), to adjust the risk-values from the 1982 follow-up of A-bomb Survivors. The results of his own education in preparing the 1995/1996 monograph alerted him that accumulated x-ray exposures before (and after) the 1945 atomic bombings, might have seriously compromised the A-bomb Survivor Study.
A third reason, for uncertainty in efforts to quantify cancer-risk from medical x-rays, is co-action. Advancing insights undermine earlier assumptions that a single cause suffices to produce a case of cancer. The prevailing, tentative model, is that each clinical case of cancer results from at least a two-step process: Initiation and promotion --- co-action by multiple causal factors. By definition, the absence of one necessary factor would prevent the cancer from occurring clinically (among others: Doll and Peto 1981 pp.1220-1221, Gofman 1999 pp.3-4). The model, of multiple causes per case, results in the expectation that cancer risk per unit dose of one factor (e.g. x-rays) will vary from person to person and from population to population, according to the presence and absence of necessary co-actors.
Part 3: Attempt to Make a Reality-Check on the "Just 1%" Estimate
One result of Dr. Gofman's monograph on breast cancer (Gofman 1995/1996) was the estimate that about 75% of the 1993 incidence of breast cancer in the USA would have been absent in the absence of accumulated exposures of breasts to ionizing radiation, mostly from medical x-rays.
He could not ignore the next natural question: How much of the entire cancer problem in the USA would have been absent, in the absence of accumulated exposures to medical x-rays. He felt obligated to make a reality-check on the ballpark estimate of 1% (Doll and Peto 1981 p.1256). The story of his own failure to see the problem earlier is related with charm in Gofman 1999, Chapter 2, Part 9, "A Very Slow Arrival at Conceiving and Testing Hypothesis-1" (www.ratical.org/radiation/CNR/RMP/chp2F.html) .
For the reasons described in Part 2, above, estimated cancer risks from medical x-rays per unit dose from partial body irradiation have a very wide range. Doll and Peto had necessarily made choices and filled in the blanks --- on average lifetime accumulated organ-doses, and risk per unit dose, and on attributable risk --- with copious assumptions and approximations.
Given what preparation of his 1995/1996 monograph had taught Dr. Gofman about x-ray practice in the USA up to 1960 (including exuberant use of high-dose fluoroscopy during routine check-ups), he recognized that key inputs to the Doll-Peto 1% estimate had very flimsy foundations. Indeed, Drs. Doll and Peto themselves acknowledged some. For instance, they wrote (p.1220): "Evidence for making quantitative estimates of attributable risk [when more than one cause of a cancer is known] is for the most part lacking. So we have made estimates on the basis of extrapolations of uncertain reliability, clinical impressions, and contemporary hypotheses."
A reality-check on their 1% estimate should be based on a very different method, Dr. Gofman decided. A method might produce more credible output if its input would include no one's guesstimates of average accumulated lifetime x-ray doses to each organ and include no one's estimates of cancer risk per dose-unit.
And that is why his input to the 1999 monograph consisted of using data on all nine Census Divisions to find out if a strong positive correlation existed between physicians per 100,000 population in a Census Division and age-adjusted cancer mortality per 100,000 population in the same Census Division. Details of the input, including his study of the stability of relative physician density from one decade to the next, are provided in Chapters 3 and 4 of the monograph; those chapters are available at www.ratical.org/radiation/CNR/RMP/.
His underlying common-sense assumption was that the more physicians per Census Division, the more x-ray exams would be given --- an assumption supported by data elsewhere (Mettler 1987 p.134). And his key approximation was that one could achieve a ballpark estimate, of what the age-adjusted cancer mortality rate would have been in the absence of any physicians (absence of any x-ray exams), by extending the observed dose-response to its intercept with mortality rate at "zero docs per Division." A picture speaks louder than words. Six such graphs can be seen at www.ratical.org/radiation/CNR/RMP/chp1F.html#fig1a. Abbreviations used there: PhysPop = Physicians per 100,000 Population, and MortRate = Age-Adjusted Mortality Rate.
Gofman considered his ballpark method to be at least as credible as the Doll-Peto ballpark method. And he explicitly called attention to his method's important advantage, in a field challenged by conflicts of interest: The Gofman method simply ruled out bias creeping into its input, for there was no estimating average accumulated organ doses and no choosing whose estimated values to use with respect to cancer risk per dose-unit. His input was utterly neutral regarding the issue at hand.
He recognized, and warned, that using the same method beyond 1940 or 1950 would be problematic, due to massive migration from one Census Division to another during World War Two, due to introduction of toxic chemotherapy agents after that war, and due to the soaring impact from smoking upon age-adjusted cancer mortality rates after mid-century. Although he attempted a "smoking correction" beyond 1940, he was dissatisfied with that and called it "far from elegant!"
Part 4: Results for Cancer, IHD, and NonCancer NonIHD Mortality
Gofman 1999 first analyzed the age-adjusted cancer mortality data for all cancers combined, males and females separately. He also analyzed all the major cancers separately, by gender. He found that the higher the physician density (the surrogate for average accumulated x-ray dose), the higher were age-adjusted mortality-rates from cancer.
The results from these 12 distinct cancer studies astonished Dr. Gofman. They produced a credible basis for the hypothesis that medical x-rays were very probably a necessary causal co-actor in well over half of the age-adjusted cancer mortality-rates in 1940.
But would the new method produce approximately the same positive dose-responses, no matter which diseases he tested? To find out, he repeated all the analyses. He studied the combined non-cancer causes of mortality, separately for males and females. Then he studied each major non-cancer cause separately, by gender --- 26 non-cancer studies in all.
The correlations between non-cancer mortality rates and physician density differed unmistakably from the cancer studies. The individual non-cancer correlations were either negative (the more physicians, the lower the death rate), or they were approximately flat (no significant correlation of any type). Visit the graphs again.
But one very important exception revealed itself: The correlation between physician density and ischemic heart disease (also called coronary heart disease, coronary artery disease) was both positive and spectacularly strong --- just like the undeniably linear correlation observed for cancer. The data seemed to scream, "Pay attention! Cancer and ischemic heart disease share a common cause not shared by the other major causes of death in the USA." Visit the graphs again! The IHD data come from 1950 because before that, IHD (CHD, CAD) was not reported as a distinct type of "heart disease."
By themselves, positive dose-responses can suggest a causal relationship in biology, although they can never prove causation, as Dr. Gofman emphasized repeatedly throughout the text. But very strong correlations, such as the ones discovered by his 1999 monograph, demand an explanation. In the final chapters of his monograph (Chapters 67, 68, 69), he examined whether the presumption of a causal relationship would conflict with other existing evidence and whether a non-radiation agent could provide a more reasonable explanation for the correlations. His conclusion: No and No.
Therefore, for both cancer and ischemic heart disease, the data required presenting the hypothesis that very probably more than half of the age-adjusted mortality rates at mid-century would have been absent, in the absence of accumulated exposures to medical x-rays.
Part 5: How Could X-Rays Cause Fatal Cases of Ischemic Heart Disease?
It was well known decades ago that very high doses of x-rays for treating Hodgkins Disease and Breast Cancer exposed the heart to very high doses that damaged the heart. But no one thought that much lower x-ray doses from routine procedures (e.g., x-rays of the upper spine, esophagus, lungs, breasts, etc.) --- that also irradiated the heart --- could cause atherosclerosis and fatal ischemic heart disease. So Dr. Gofman was astounded by the similarity between the dose-responses he uncovered for age-adjusted mortality from cancer and from IHD.
How could this be? To explain this similarity, Dr. Gofman proposed a mechanism similar to radiation carcinogenesis.
His hypothesis is that some of the non-inherited mutations, acquired in the coronary arteries from x-ray exposures, are atherogenic when they render a cell dysfunctional at processing plasma lipoproteins AND endow the same cell with a proliferative advantage. (A single aberrant cell --- among a person's trillions of cells --- is definitely not a cancer-menace either, unless it has acquired the capability to proliferate abnormally.)
Regarding induction of atherosclerosis, Dr. Gofman proposed that the clone of aberrant cells with a proliferative advantage gradually replaces normal cells at a localized patch of arterial wall. Because this patch retains lipoproteins that would otherwise exit normally from the arterial wall (something that massive quantities of plasma lipoproteins routinely do), the patch becomes a site of chronic inflammation, resulting in construction of an atherosclerotic plaque. If the plaque's fibrous cap becomes too fragile to contain the highly thrombogenic lipid-core below it (high blood pressure can help make it too fragile), the cap ruptures and a blood clot --- sometimes fatal --- results.
Note that this proposed model is one of co-action between acquired mutations and certain lipids, as cancer is also the result of co-action between two or more causes. By definition, no case can occur if one of its necessary co-actors is absent. Dr. Gofman was well aware that x-rays are not the only source of acquired mutations, and he certainly did not propose that x-ray-induced mutations became necessary co-actors all cases of IHD and cancer in the USA.
However, he did propose, as a result of the very strong and linear dose-responses he uncovered and the associated intercepts, that x-rays became the dominant cause of both carcinogenic and atherogenic mutations after the introduction of x-rays into medicine in 1900.
Part 6: Bottom Line of This Communication and Back to CT Heart Scans
And what about today? We do not know how average accumulated exposure per capita, to medical x-rays during the past 70 years, compares with the 40 years preceding 1940. We share the assumption that average annual x-ray exposure per capita must have been increasing since the early 1970s, when CT scans were introduced and when the use of fluoroscopy to guide placement of catheters and instruments began increasing.
X-rays remain a uniquely potent mutagen because they can induce uniquely complex and double-stranded DNA damage which the cell cannot always repair correctly, if at all. There is no risk-free level with respect to mutagenesis (see "Three Remarkably Similar Reports on the Safe-Dose Fallacy" at www.ratical.org/radiation/CNR/PBC/chp45.html).
Therefore, serious effort is ethically imperative to acquire good images with the least possible x-ray dose and to avoid unnecessary exams. Methods reducing dose to half or less have been demonstrated (www.ratical.org/radiation/CNR/XHP/WhoSays.html) but will not be applied while the assumption prevails that current doses confer a negligible risk.
As recently as a decade ago, students at some medical schools were still being taught that just a trivial risk comes with x-ray procedures. Many patients did not even realize that mammograms, CT, and fluoroscopy were x-ray procedures, or that they received fluoroscopy while they were under anesthesia for certain surgeries. Alex Berenson and Reed Abelson might explore what today's medical students are taught about the risk from the x-rays they order. And they might check into current training in x-ray offices on how to reduce dose-levels. In 1999, one x-ray equipment salesman told me, "I hate to tell you this, but radiologists almost never ask me about the doses delivered by our equipment. That seems to be the last thing that interests them."
It is currently impossible, in my opinion, to offer credible estimates of either an individual's likely benefit or the likely risks from repetitive CT heart scans --- or from repetitive screenings with x-rays for any disease. But on the risk side, it would be very foolish to brush off the warnings inherent from the mid-century observations in Gofman 1999. They challenge the 1% "negligible risk" assumption in a powerful way. The cautions expressed by several physicians, in the fine article by Alex Berenson and Reed Abelson, are perhaps far too mild.
# # # # #
Addendum: Two Contrasting Comments on the 1999 Monograph
In 2006, the newest report on the Biological Effects of Ionizing Radiation --- known as BEIR-7 --- mentioned receiving the 1999 monograph from Dr. Gofman (BEIR 2006 p.329). Then that publication stated that it did not share his interpretation of the discovered correlations, because he used "data on populations rather than data on individuals" and because "it is not possible to verify the quantitative nature" of the relationship between physician-density and x-ray dose.
We considered exactly those inputs to be the monograph's special merits, for reasons summarized in Part 3 above!
In contrast with BEIR-7's lack of interest, Arthur C. Upton, M.D., chairman of the earlier BEIR-1990 Committee and Director of the National Cancer Institute 1977-1979, wrote to Dr. Gofman about the 1999 monograph, "I find your observations intriguing, and your interpretation of them to be thoughtful and constructively hypothesis-generating. I hope that your book stimulates the productive follow-up research that your findings clearly call for . . . Best wishes for continuing productivity in the new millennium."
Dr. Gofman was unable to carry on his work, due to his wife's failing health and then his own. He died in August 2007. I, who helped with his research for 20 years and edited his last three monographs, feel qualified to have written this communication. In his will, Dr. Gofman explicitly authorized me and/or his son to describe, summarize, and explain his radiation research when he no longer could.
Ms. Egan O'Connor, San Francisco, July 2008.
REFERENCES:
Archer, Victor E., 1982 (March 19). Reviewing John Gofman's 1981 book, Radiation and Human Health, in the Journal of the American Medical Assn (at p.1637), Dr. Archer wrote, "This remarkable and important book enables any intelligent person with a high school education to understand the complexities involved in assessing the risks to man from low levels of ionizing radiation. Gofman not only demonstrates his mastery of this complex subject but carefully explains the basic concepts of epidemiology, genetics, birth defects, carcinogenesis, radiobiology, physics, chemistry and even mathematics, which are necessary to an understanding of the subject."
BEIR 1990. Committee on the Biological Effects of Ionizing Radiation. Health Effects of Exposure to Low Levels of Ionizing Radiation [also known as the BEIR-5 Report]. Washington DC: National Academy Press. ISBN 0309039959. 421 pages.
BEIR 2006. Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation, Board on Radiation Effects Research, National Research Council of the National Academies. Health Risks from Exposure to Low Levels of Ionizing Radiation: BEIR-7 Phase 2. BEIR stands for Biological Effects of Ionizing Radiation, a sporadic series of reports beginning in 1972. This document comes with a sticker on its verso page stating: "This BEIR VIII study was supported by funds from the U.S. Dept. of Defense, U.S. Dept. of Energy, U.S. Environmental Protection Agency, the Nuclear Regulatory Commission, and the U.S. Dept. of Homeland Security." Washington DC: National Academy Press. ISBN 030909156X. 406 pages.
Bond, Victor P., 1978, with Charles B. Meinhold and Harald H. Rossi. "Low Dose RBE and Q for X-ray Compared to Gamma Radiations," Health Physics 34: 433-438.
Brown, W. Virgil 2007 (May), "From the Editor in Chief," Journal of Clinical Lipidology, 1 No.2: 97-99. Referring to Gofman 1955 (below), Dr. Brown writes: This document is "both a scientific tour de force and a historically important presentation of concepts that underpin our field. I was most impressed on rereading the document that he not only discovered relationships previously unknown but defined important questions that remain unanswered even today."
Doll and Peto 1981 (June). Richard Doll and Richard Peto. "The Causes of Cancer: Quantitative Estimates of Avoidable Risks of Cancer in the United States Today," Journal of the National Cancer Institute 66, No.6: 1192-1265 plus 6 appendices.
Gofman, John W., 1949 with Frank T. Lindgren and Harold A. Elliott. "Ultracentrifugal Studies of Lipoproteins of Human Serum," Journal of Biological Chemistry, 179: 973-979.
Gofman, John W., 1950, with Frank T. Lindgren and Harold A. Elliott, et al. "The Role of Lipids and Lipoproteins in Atherosclerosis," Science, 111 No. 2877: 166-171 & 186.
Gofman, John W., 1955, with Oliver DeLalla and Frank Glazier, et al. "The Serum Lipoprotein Transport System in Health, Metabolic Disorders, Atherosclerosis and Coronary Heart Disease," Plasma 2, No.4: 413-484. This classic paper was re-typeset and fully re-published in May 2007 by the Journal of Clinical Lipidology, 1 No.2: 104-141.
Gofman, John W., 1990. Radiation-Induced Cancer from Low-Dose Exposure: An Independent Analysis. San Francisco: CNR Books. ISBN 0932682898. 455 pages.
Gofman, John W., 1995/1996. Preventing Breast Cancer: The Story of a Major, Proven, Preventable Cause of This Disease. San Francisco: CNR Books. The 1995 edition sold out. The 1996 edition is identical except at the end, it adds a section entitled, "Response to Critiques of the First Edition." 1996 edition's ISBN 0932682960. 421 pages.
Gofman, John W., 1999. Radiation from Medical Procedures in the Pathogenesis of Cancer and Ischemic Heart Disease: Dose-Response Studies with Physicians per 100,000 Population. ISBN hardcover = 0932682979, softcover = 0932682987. San Francisco: CNR Books. 699 pages.
Havel, Richard J., 2007. "Introduction: John Gofman and the Early Years at the Donner Laboratory," Journal of Clinical Lipidology, 1, No.2: 100-103. Referring to Gofman 1955 (above), Dr. Havel writes: "It extensively summarizes most of this laboratory's pioneering clinical observations, which energized a new field of biomedical research." In conclusion, Dr. Havel refers to Dr. Gofman's "seminal discoveries on plasma lipoproteins."
Kato, Kazuo, 1998, with Walter J. Russell and Kazunori Kodama. "Medical Radiation Exposures of Atomic Bomb Survivors," Chapter 3 at page 51 in Effects of Ionizing Radiation: Atomic Bomb Survivors and Their Children (1945-1995). Editors: Leif E. Peterson and Seymour Abrahamson. Washingtgon DC: The National Academy Press, Washington DC. ISBN 0309064023. 370 pages.
Krauss, Ronald M, 2007. In the University of California at Berkeley Press Release, Sept. 4, 2007, about John W. Gofman's death, Dr. Krauss is quoted referring to Gofman 1955 (above) as a seminal paper that "basically laid out the whole map for the field of lipids and heart disease for decades to come, which was absolutely remarkable. That's where John Gofman showed the genius of his capabilities as a physician and a physicist. That paper was almost a bible for people in the field" (www.berkeley.edu/news/media/releases/2007/09/04_GofmanObit.shtml).
Mendelsohn 1995. Mortimer Mendelsohn, M.D., Ph.D., was then Associate Director of the Radiation Effects Research Foundation (the joint Japanese-U.S. governmental guardian of the Atomic Bomb Survivor Study). He was asked his opinion of Dr. Gofman as a scientist, during the British Television Channel 3 investigative report, "The X-ray Effect" (broadcast August 3, 1995). Dr. Mendelsohn replied, "Gofman is a superb analyst and has always been at the cutting edge of medical science, particularly when it comes to protecting people."
Mettler, Fred A. Jr. et al, 1987. "Analytical Modeling of Worldwide Medical Radiation Use," Health Physics 52: 133-141.
Radford 1995. Edward P. Radford, M.D. was then professor of epidemiology, and former Chairman of the 1980 Committee on the Biological Effects of Ionizing Radiation (BEIR-3). He was asked his opinion of Dr. Gofman as a scientist, during the British Television Channel 3 investigative report, "The X-ray Effect" (broadcast August 3, 1995). Dr. Radford replied, "Dr. Gofman is owed a debt of gratitude by the scientific community because he was one of the first people to raise the issue of cancer risks from radiation exposure."
Straume, Tore, 1995. "High-Energy Gamma Rays in Hiroshima and Nagasaki: Implications for Risk and Weighting Factor," Health Physics 69, No.6: 954-956.
Upton 1995. Please see "Addendum" above.
Von Euler, Ulf S., 1972. "The 1972 Stouffer Prize is awarded to Dr. John W. Gofman for pioneering work on the isolation, characterization and measurement of plasma lipoproteins, and on their relationship to arteriosclerosis. His methods and concepts have profoundly stimulated and influenced further research on the cause, treatment, and prevention of arteriosclerosis," was Professor von Euler's citation with the award. Professor von Euler was chairman of the Stouffer Prize Committee and former chairman of the Nobel Prize Committee for Physiology and Medicine.
- End -
This document is also available in ASCII text format
and can be found at http://www.ratical.org/radiation/CNR/RMP/heartCT.html