Disproof of Any Safe Dose or Dose-Rate of Ionizing Radiation,
with Respect to Induction of Cancer in Humans
This chapter is arranged in eight parts:
Overview -- with a Five-Point Summary, p.1 A Troublesome Trio: Unrepaired, Unrepairable, Misrepaired Injuries, p.2 Evidence on the Capacity and Speed of Repair Systems, p.3 Two Implications from the Repair Studies, p.5 Conclusive Human Evidence below 10 Rads, p.7 A Supplemental Approach via Epidemiological Evidence, p.9 Comparison of Two Explanations for the Observations, p.15 The Bottom Line, p.16
1. Overview -- with a Five-Point Summary
Probably the most important issue in this field is whether or not there exists some low dose, or dose-rate, of low-LET ionizing radiation which produces no cancer at all in exposed populations. In short, is there a safe region below some threshold of danger?
If the idea of a safe dose or dose-rate prevails -- and the idea has some very influential backing (see Chapters 24, 34, 35) -- then both voluntary and involuntary human exposures are bound to increase dramatically above their current levels. If the idea prevails and is false, it could ultimately result in a hundred million or more unnecessary, premature cancers over time, worldwide. Thus, the stakes of the threshold issue are high indeed. And this book does not even open the issue of heritable genetic injuries.
Meaning of "Safe Dose or Dose-Rate" :
Because not all readers of this book will be epidemiologists, we want to be explicit about the meaning of "safe dose or dose-rate." Confusion can arise from two different meanings of safe: (1) free from danger; secure, and (2) having escaped danger or injury; unharmed. For instance, not everyone exposed to battle gets killed. As the battle begins, no one is safe. After the battle, some are dead and some are safe.
With respect to ionizing radiation, the meaning of a safe dose or dose-rate is a dose or dose-rate at which all exposed persons are safe as the exposure occurs, and all are safe afterwards. no fraction will be killed later by radiation-induced cancer. In sharp contrast, "no safe dose or dose-rate" means that no one is safe as the exposure occurs, and afterwards, some fraction of the exposed persons will die from radiation-induced cancer, and the rest will be safe from it.
Our Approach to the Subject :
Elsewhere (Chapters 24 and 34), we have assembled statements from various members of the radiation community to the effect that (A) there is a reasonable chance that safe doses and dose-rates do exist for low-LET radiation, and (B) it is impossible to resolve the threshold issue from the existing evidence.
By contrast, we think human evidence and logic combine to make a case which is already conclusive -- by any reasonable standard of proof -- against the existence of any safe dose or dose-rate of ionizing radiation, with respect to cancer-induction.
The disproof of any safe dose or dose-rate, presented in this book, represents an expansion and closer examination of the case presented in Go81 (pp.404-411) and in Go86 (Section 2 and Technical Appendix 1). Our analysis ignores high-LET radiation, because almost everyone admits that there is no safe dose or dose-rate from alpha particles and other high-LET radiations.
We will summarize the case in a single chapter (this one), and will place the supporting evidence, calculations, and "what if" materials, into auxiliary chapters of their own. Some readers will prefer to read Chapters 19, 20, and 21 before this summary chapter, and other readers will prefer to read this chapter first.
For curious readers, we have also provided an overview (Chapter 35) of the kinds of evidence cited by various authors who have been speculating that perhaps, someday, a net benefit will be discovered for human health, from low-dose exposure to ionizing radiation. Readers will see for themselves that this is sheer speculation, centering largely around possible stimulation of repair and immune responses.
In contrast with such conjecture is the real-world human epidemiological evidence, discussed in Chapter 21, which shows fatal cancer-induction even at minimal doses and dose-rates of ionizing radiation. When excess fatal cancer is observed in humans after such exposures, the excess has occurred despite any possible stimulation of the repair- and immune- responses by low-doses. The net result is injury, not benefit. I wish it were otherwise.
A Five-Point Summary
The argument against any safe dose or dose-rate is summarized below. For the sake of brevity, various statements are necessarily omitted.
o -- 1. The dose from low-LET ionizing radiation is delivered by high-speed electrons, traveling through human cells and creating primary ionization tracks. One track is the least possible disturbance which can occur at the cellular level. "High dose" means many tracks per cell; "low dose" means few tracks per cell; "low dose-rate" means few tracks per cell per unit time. Whenever there is any dose at all, it means some cells and cell-nuclei are being traversed by tracks (see Chapter 19).
o -- 2. Single, primary ionization-tracks, acting independently from each other, are never innocuous with respect to creating carcinogenic injuries in the cells which they traverse. Every track -- without help from any other track -- has a chance of inducing cancer by creating such injuries (see Chapter 19, including "The Yalow Model").
o -- 3. This implies that there can never be any safe dose or dose-rate. However, if every carcinogenic alteration induced by tracks were successfully and invariably "un-done" by repair processes, then there would be an inherently safe dose or dose-rate. The key question is: Does repair of carcinogenic injuries operate flawlessly, when dose is sufficiently low and slow?
o -- 4. Human epidemiological evidence shows that repair fails to prevent radiation-induced cancer, even at doses where the repair-system has to deal with only one or a few tracks at a time, and even at dose-rates which allow ample time for repair before arrival of additional tracks. See Chapter 21, and this one. Such evidence is proof, by any reasonable standard, that there is no dose or dose-rate which is safe . . . unless we can find some wholly additional cancer-prevention mechanism which is perfect whenever the repair-mechanism at low doses fails.
o -- 5. The radiation-induced cancers arising from the unrepaired lesions at low doses do not wear a little flag identifying them as any different from cancers induced by higher doses of radiation, or induced by causes entirely unrelated to radiation. Therefore, threshold proponents cannot argue that the cancers arising from the lowest conceivable doses of radiation will somehow be eliminated by the immune system or any other bodily defenses against cancer. Such an argument would require the elimination of cancer in general by such defenses. Instead, we observe that cancer is a major killer (roughly 15-20% of many populations). So the proposition would lead to a non-credible consequence, and must be rejected. This means that repair is the key, and that Point 4 stands: There is no dose or dose-rate which is safe with respect to human carcinogenesis.
The Heart of the Issue :
Points 1 and 2 above are explained in detail in auxiliary chapters, and are not controversial. We regard Point 5 above as self-evident. Therefore, this chapter will focus on discussing Points 3 and 4. "Repair" is at the heart of the threshold issue.2. A Troublesome Trio :
Unrepaired . . .
Unrepairable . . .
In 1914, Theodore Boveri suggested the hypothesis that imbalance of chromosomal information is a central feature of carcinogenesis (Bov14). Imbalance could include missing genetic information, erroneous genetic information, excess genetic information, and genetic information in the wrong locations. In other words, the idea that carcinogenesis is tied to defective or inappropriate genetic information, in the cell nucleus, is an idea which goes back many decades. With advances in research techniques and tools, the evidence in favor of the idea is rapidly increasing.
Today, in our field, the underlying presumption is that the carcinogenic lesions caused by ionizing radiation are occurring in genetic material -- namely, in the DNA or chromosomes of cell nuclei. UNSCEAR uses this presumption in its 1986 report, after making the following comments (Un86, pp.12-13):
"Cancer initiation is believed to be a uni-cellular process occurring at random in single cells. This is . . . a working hypothesis that has not yet definitely been proved." In speaking of radiation effects on cells, UNSCEAR adds: "These effects involve the cells' genetic material, which is also thought to be the primary target for cancer initiation."
While not everyone accepts these premises, most do -- and I am emphatically one of those who do, for many reasons which do not need discussion in this particular book.
The remarkable scientific work on DNA of the past few decades has established convincingly that biological repair of injury to DNA molecules undoubtedly takes place. There is a large literature on the subject, and the knowledge and detail are quite sophisticated. Some repair of chromosomes also occurs, although here the evidence is not so certain about the details.
No one welcomed the findings about DNA and chromosome repair more than did many segments of the radiation community, for the prospect seemed newly bright that people could be irradiated with no cancer-consequences. It was -- and still is -- widely suggested that repair could certainly take care of low doses of radiation or low dose-rates. But while hope springs eternal, there is the nasty problem of scientific reality: "Does repair always succeed?"
And this is why we call attention to three warnings from the literature.
o -- Brackenbush and Braby (Brack88, p.256) state the following: "Since most cells repair radiation damage with a characteristic time ranging from a few minutes to a few hours, it is evident that irreparable or misrepaired damage must dominate the low-LET radiation effect at low dose-rates."
o -- UNSCEAR (Un86, p.179) comments as follows on repaired, unrepaired, and misrepaired carcinogenic lesions induced by radiation: "The error-free repair of the DNA, which is the most likely target involved, leaves some fraction of the damage unrepaired and the error-prone repair may produce misrepaired sequences in the DNA structure."
o -- Albrecht Kellerer (Kelle87, p.346) describes a type of radiation-induced lesion which would be difficult to repair: "A simple example would be two neighboring single-strand breaks on opposing strands of DNA, which interfere with excision repair." Kellerer's warning is confirmed by Feinendegen and co-workers (Fein88, p.29) who, reporting on irradiated cells, say "not all double-strand breaks are fully repaired."
Important cause for concern is contained in these statements, with their reference to "unrepaired," "misrepaired," and "irreparable" lesions inducible by radiation.
And there is no basis for limiting the concern to DNA, when it belongs to chromosomes too. Although we undoubtedly see chromosomes with apparent repair (for instance, we sometimes see the rejoining of breaks in the strands), it is far from clear that all the pre-injury links have been restored perfectly to their original state. Some apparent repairs may really be incomplete or functionally incorrect, even if residual damage cannot be visualized microscopically. It is appropriate to suggest that some portion of "repair" is really misrepair.
No matter how great the capacity and speed may be of genetic repair mechanisms in a cell nucleus, there are no laws stating that every type of injury is repairable, and that misrepair can never happen. Indeed, imperfect repair is the potential "Achilles Heel" for any concept of a safe dose or dose-rate with respect to radiation-induced cancer. This should have been evident for a long time, as we shall see.3. Evidence on the Capacity and Speed of Repair Systems
In auxiliary Chapter 20, Table 20-M, readers will find tables which estimate the average number of primary ionization tracks which traverse a cell-nucleus at a specified dose. Of course, the derivation of those tables is also shown, step-by-step. The results of the work are used in several parts of this chapter, including this part.
For disproof of any safe dose or dose-rate, it is more important to establish the dose in terms of the average number of tracks per nucleus, than to establish it in terms of rads. The reason is that the lowest conceivable dose or dose-rate with respect to repair is not a millionth or any other tiny fraction of a rad or centi-gray. The lowest conceivable dose or dose-rate is one track per nucleus plus sufficient time to repair it. (For more about dose-rate, see Chapter 20, Part 3.)
Below, in this chapter, we will be citing studies of repair conducted on human cell-cultures following X-irradiation, typically from a 250 kilovolt machine. In such studies, reported doses are often "a few Grays" -- a few hundred rads. In the next chapter, Table 20-"O" tells us that, at a dose of 3 Grays or 300 rads, each cell-nucleus is feeling the disturbance from about (2.3 tracks / rad) x (300 rads), or some 690 tracks on the average. If the dose is an acute one, the tracks are virtually simultaneous, of course.
Our purpose in citing the studies which follow is to establish (A) whether or not human cells are running short of repair-capacity at doses where radiation-induced cancer has been epidemiologically demonstrated, and (B) how much time the cells take to finish all the repair of which they are capable.
Repair-Capacity -- Undiminished at Doses of Several Grays :
Kellerer, who is a leading expert in the field of microdosimetry, has reported:
" . . . there is little or no evidence for an impairment of enzymatic repair processes at doses of a few grays. Studies, for example, by Virsik et al. [Vir82] on chromosome aberrations, have established characteristic repair times that are substantially constant up to 10 Gy, that is, up to the highest doses investigated. Similar observations have been obtained in various cell survival studies. Most of the enzymatic DNA repair processes that are known are of the catalytic type. The enzymes are not used up in the repair process, and under usual conditions it is safe to assume that the concentration of enzymes is sufficient to maintain constant repair efficiency at the concentration of lesions produced by several grays . . . " (Kelle87, p.358-359).
Writing about studies of human cells irradiated with doses up to ten Grays (1,000 rads), Kellerer emphasizes that reduced success of repair at high doses is not the result of insufficient repair-capacity (but rather, the greater frequency of injuries which are very close to each other):
" . . . there is, at present, no experimental evidence for a reduction of the repair capacity or the rate of repair at doses of a few gray which are relevant to cellular radiation effects. Reduced efficiency of repair or enhanced misrepair are apparent at elevated doses of sparsely ionizing radiations and at all doses of densely ionizing radiations, but they can be understood in terms of the greater proximity of sublesions of DNA and the resultant failure of DNA repair. A simple example would be two neighboring single-strand breaks on opposing strands of DNA, which interfere with excision repair. Such interference with repair due to spatial proximity is, in a somewhat loose terminology, included in the general notion of the `interaction' of sublesions" (Kelle87, p.346).
(Our auxiliary Chapter 19, Part 4, points out, however, that such interactions do not require the presence of two separate tracks -- they can also occur between lesions along a single track.)
The key point is that Kellerer, who has looked closely at the evidence on repair, concludes as recently as 1987 that " . . . it is safe to assume that the concentration of enzymes is sufficient to maintain constant repair efficiency at the concentration of lesions produced by several grays . . . " (see above).
Keeping Order in the Genetic Library :
With respect to natural doses of ionizing radiation, such massive repair-capacity would never be needed today. But it exists nonetheless. One can speculate that, during the epoch when DNA was evolving, the natural doses were much higher than now, or that there were viral "vandals" in the "library" of genetic information, and a system with great capacity was necessary to restore order. Whatever the history, what we observe today in the cell-studies is that ionizing radiation tears "books" from their shelves in the genetic library, but massive squads of vigilant librarians very rapidly restore order -- with only an occasional book overlooked or misplaced.
Activation of Repair-Capacity :
No one is implying that repair-enzymes are always present in a cell-nucleus at a constant concentration, whether the menace is one track or 700 tracks. The concentration of the enzymes probably varies with the stimulus. It has been suggested by Goodhead and others that possibly the repair-system needs a "kick" to get started. For instance: " . . . it is also conceivable that the cell would repair relatively more efficiently if there were more damage to stimulate its repair processes" (Good88, p.234-5). If the suggestion is someday confirmed, it would seem to imply difficulties for the safe-dose proponents. At doses or dose-rates too low to provide adequate stimulation, repair of carcinogenic injuries might operate the least efficiently of all, or even be entirely absent. Presently, such matters remain in the realm of speculation.
Repair-Times of 8 Hours and Much Less :
Studies of human cells in vitro, following X-irradiation, indicate that whatever repair is achieved, is complete within 6 hours or less after irradiation, even at doses of 100 and 200 rads (for instance, in Ben82; Nata82; Pres80). Indeed, in the references cited, all the repair occurs within the first two hours after irradiation, and by three hours, the repair curve is flat.
Other sources of evidence are in good agreement on the issue of the time required for a variety of repair functions following radiation injury.
We already cited Brackenbush and Braby (Brack88, p.252) who stated: "Since most cells repair radiation damage with a characteristic time ranging from a few minutes to a few hours, it is evident that irreparable or misrepaired damage must dominate the low-LET radiation effects at low dose-rates."
Upton (Up88, p.606, in his Figure 1) shows that,
a. Repair of sublethal damage is such that the surviving fraction of cells has reached as high a value as it ever will, by between 4 and 6 hours post-irradiation following, doses of 400 rads in vitro for C3H10T1/2 cells.
b. For chromosome aberrations in the same cell system and at 400 rads, repair has reduced the number of chromosome aberrations to as low a value as it would ever achieve, somewhere between 4 and 6 hours after irradiation.
Bender, discussing repair of chromosome breaks, reports repair half-times which are "typically of the order of 1 or 2 hr" (Ben84, p.286).
Bond , in a discussion of single-hit and multi-hit phenomena in irradiated cells, refers to the half-time for repair as "frequently in the range of hours or less" (Bon84, p.393).
Feinendegen and co-workers report, "Whereas the majority of single-strand breaks and base changes are very efficiently and quickly repaired with half-times less than 1 h, the reconstitution of a double-strand break probably lasts much longer, perhaps up to several hours, and not all double-strand breaks are fully repaired" (Fein88, p.29).
Rat Experiments at 1200 Rads :
Burns and Sargent have presented data concerning DNA repair times in rat epidermis (Burns81). These studies involve whole-animal irradiation of rat epidermis with 0.8 MeV electrons. It was found that single-strand breaks in DNA were removed after irradiation with a half-time of 21 minutes in vivo. Concerning the period within 60 minutes post-irradiation with 1200 rads, the authors state:
"By 60 min [their abbreviation] after irradiation, the breaks had returned approximately to the value in unirradiated controls, indicating essentially complete repair."
Reporting on some separate studies of radiation carcinogenesis in rat skin, the authors state that split-dose experiments indicated the half-time for repair of carcinogenically-related events was between 110 and 240 minutes.4. Two Implications from the Repair Studies
Speed -- An Implication :
The cell studies indicate that repair systems finish their work within about 3 to 6 hours, even after acute doses up to 400 rads (about 900 simultaneous tracks per cell-nucleus). The in vivo study with rats indicates half-time for repair is about 2 to 4 hours, which means that essentially all possible repair-work is completed in less than 24 hours, in spite of very high doses indeed.
The dazzling speed of repair has an extremely important implication for settling the threshold issue. It means that certain high-dose evidence can reveal a great deal, as we will explain.
If a radiation dose is received within the time-frame required for repair, and if repair operates flawlessly and leaves no carcinogenic damage, then the net effect of that radiation-dose toward cancer-production is obviously zero, by definition.
Opportunity to Prove a Safe Dose :
So we can describe a scenario in which repair is flawless at a specified low dose, and in this scenario, individuals receive their first exposure to this dose on Monday. Repair is flawless and complete within hours. The individuals have no cancer-risk due to the first exposure. On Tuesday, these individuals receive their second dose of the same size. Since repair is perfect, there is still no risk of radiation-induced cancer from the combined exposures, Monday and Tuesday.
Under these circumstances, such individuals could gradually accumulate a very high dose from dozens or hundreds of low-dose exposure-sessions, and yet the group would show no excess (radiation-induced) cancer when followed-up post-irradiation.
If the same very high dose, received all at once, is known to cause cancer, and if the studies of very high doses accumulated through a series of low doses were large enough and long enough to show a cancer-excess, and if such studies typically failed to show any excess cancer, we would conclude: Division of the very high total dose into low doses and dose-rates permitted repair to produce a safe dose and dose-rate -- even though repair was overwhelmed and flawed when the same number of ionization tracks was received all at once. The weight of such evidence would be persuasive in favor of a threshold.
Opportunity to Disprove a Safe Dose, Too :
High-dose studies of this same type could also provide proof that repair is not flawless at the low doses and dose-rates tested. If the follow-up studies typically revealed excess cancer in spite of the fact that the high total dose had been received in a series of low doses, we would necessarily conclude that repair had not worked perfectly at low doses.
Indeed, our disproof of any safe dose or dose-rate includes four studies which are ostensibly high-dose studies, but really tell "all" about minimal doses and dose-rates. We are not alone in recognizing their implication. Upton refers to " . . . the dose-dependent excess of breast-cancer, which is of similar magnitude per unit dose in a) women exposed to A-bomb radiation, b) women given therapeutic irradiation for postpartum mastitis, c) women subjected to multiple fluoroscopic examinations of the chest during the treatment of pulmonary tuberculosis with artificial pneumothorax, and d) women exposed occupationally to external gamma radiation in the painting of luminous clock and instrument dials" (Up87, p.300-301). In groups (c) and (d), serial exposures to low doses accumulated into high doses.
Upton echoes the NIH Report (Nih85, p.26) when he concludes: "The similarity of the dose-incidence relationships in all four groups of women, in spite of marked differences among the groups in the duration of exposure, implies that the carcinogenic effect of a small dose on the breast is largely irreparable and that the effects of successive doses are additive" (Up87, p.301). Upton appropriately regards the results from successive small doses as " . . . support for the hypothesis that there may be no threshold in the dose-incidence relationship" (Up87, p.300; full context provided in our Chapter 34).
Capacity -- An Implication :
The abilities demonstrated by nuclear DNA repair-mechanisms might, at first, appear very helpful to proponents of a safe dose with respect to cancer-induction. "Just observe! Repair takes only minutes to hours, and repair can handle 400 rads, even 1000 rads, without evidence of inadequate capacity. A system which can take care of 500 rads delivered acutely (some 1,150 simultaneous tracks per nucleus) is surely going to have no difficulty coping with 10-20-50 tracks in the low-dose region!" I know that such a repair-capacity gladdened a few hearts within the nuclear enterprise . . . and mine, too.
One can look with awe, humility, and gratitude at a system of repair with the capacities demonstrated by the DNA repair-system. But an independent analyst, or a realist of any stripe, does not casually dismiss the troublesome trio:
Many physiological systems are amazing and awesome without being perfect: Reproduction, immune defenses, digestion, temperature control, and on and on. Imperfection is the rule rather than the exception. One cannot fault the repair-system in cell-nuclei for leaving a relatively small number of injuries unrepaired, or misrepaired, or for having some inherent inability to repair every conceivable type of injury inflicted at random by the tracks of high-speed electrons.
One could look at all the epidemiological studies (the A-bomb survivors and numerous others) which have demonstrated radiation-induced human cancer at doses between 10 and 400 rems. Those are doses where -- if cell-studies mean anything at all -- we should not anticipate any shortage of repair capacity, and yet excess cancer did occur. If capacity of the repair-system were the issue, we would not even be discussing radiation carcinogenesis today, since it looks as if repair has enough capacity to handle all the carcinogenic damage between 10-400 rems.
Moreover, the human epidemiological evidence on dose versus cancer-response provides no support for the speculation that repair makes each rad less carcinogenic as dose falls. If that were the net result of repair, the shape of dose-response would be concave-upward. But what is seen in the A-Bomb Study and in others (see Chapter 22) is not concavity-upward. The finding is either supra-linearity or linearity -- both of which are inconsistent with the speculation that repair processes make each rad less carcinogenic as dose and dose-rate fall.
Our entire experience with human radiation carcinogenesis should have made it evident that the problem we might be facing is that -- regardless of dose-level -- some fraction of radiation injury to nuclei is unrepaired . . . some fraction is unrepairable . . . and some fraction is misrepaired.
If this is the problem, and if the fraction is about the same over the entire dose-range, we can never expect any safe dose or dose-rate.5. Conclusive Human Evidence below 10 Rads
In the five-point outline in Part 1, the fourth point deserves repeating here: Human evidence shows that repair is failing to prevent radiation-induced cancer, even at doses where cells have to deal with only one or a few tracks at a time, and even at dose-rates which allow ample time for repair before arrival of additional tracks. See Chapter 21. Such evidence is proof that there can be no conceivable dose or dose-rate which is safe . . . unless we invoke a cancer-prevention mechanism wholly additional to repair mechanisms (a reference to the summary's fifth point).
From Chapter 21, we have brought Table 21-A forward, onto the next page. Table 21-A provides the bottom line from nine separate epidemiological studies in which radiation-induced cancer has been observed in the human following dose-rates which delivered twelve tracks or fewer, on the average, to cell-nuclei. In the four high-dose studies where multiple low-dose exposures occurred, there was ample time between exposure-sessions for repair to be completed before the arrival of additional tracks.
Because epidemiology is inherently inexact, it is inevitable that some analysts may challenge the goodness of one study or another, but the case against any safe dose or dose-rate does not rely on a single study. Far from it. We are presenting nine studies, which reinforce each other.
There may be additional human studies, now or in the future, suitable for this type of analysis. It is very probable that, among 20 suitable studies, there will be at least one (5 percent) which will not confirm the finding of no safe dose or dose-rate. In epidemiology -- as in so many other fields -- it is reasonable to have confidence in the weight of the evidence, and to regard it as proof. If I were to have any other attitude toward the evidence here, I would have to question my own objectivity.
It should be noted that the case against any safe dose or dose-rate does not depend upon exactitude in the estimate of tracks per nucleus in Table 21-A. If one accepts the presumption of genetic molecules as the site of radiation injury for carcinogenesis, then there is clearly a vast amount of excess repair-capacity in all nine studies of Table 21-A. Thus, even if the number of tracks per nucleus were higher, the disproof of any safe dose or dose-rate would not be undermined. (In Chapter 20, we show exactly how our track-estimates were derived, and in Chapter 33, we compare them with estimates made by others.)
The nine studies entered in Table 21-A demonstrate that the following doses are not safe, with respect to cancer-induction: 9.0 rads, 7.5 rads, 4.6 rads, 1.6 rads, 1.0 rad, 0.9 rad, 0.5 rad, and 0.1 rad. Claims abound that epidemiological evidence for human cancer-induction is absent at low doses, but such claims are clearly mistaken.
Many proponents of a safe dose and dose-rate will immediately say, about the evidence in Table 21-A: "But the table omits all the medical, occupational, and natural-background studies which show no excess cancer from low-dose exposure at slow dose-rates! You cannot just ignore them!"
And we do not ignore them. Studies of that type are examined in Chapter 21, Part 2.
Such studies -- which are potentially infinite in number -- are simply irrelevant to settling the threshold issue, as explained in Chapter 21, and as usually admitted by their authors, and as admitted by the BEIR-3 Committee. Irrelevant material does not belong in Table 21-A.
The relevant studies are those which are capable of settling the issue. For instance, as we explained in Part 4 of this chapter, the ostensibly high-dose studies included in our disproof of any safe dose or dose-rate, are capable of helping to settle the issue. Those studies had their opportunity to provide powerful evidence in favor of a safe dose and dose-rate, and they failed. Instead, they contribute powerful evidence against any threshold.
=========================================================================== | Col.A Col.B | ( Col.A times Col.B )| | | | |Number Tracks- | | |Assigned Rads per- | Average Number of | |in the per Nucleus | Tracks-per-Nucleus | |Text Exposure at 1 Rad | from Each Exposure | | | | |=========================================================================| |1. Nova Scotia 7.5 1.3378 | 10.0335 10 | | Fluoroscopy | Rounded | |=========================================================================| |2. Israeli Scalp- 7.5 1.3378 | 10.0335 10 | | Irradiation | Rounded | | (Authors' revised est.) 9.0 1.3378 | 12.0402 12 | | | Rounded | |=========================================================================| |3. Massachusetts 4.6 1.3378 | 6.1539 6 | | Fluoroscopy | Rounded | |=========================================================================| |4. Canadian 4.6 1.3378 | 6.1539 6 | | Fluoroscopy | Rounded | | (Excludes Nova Scotia) | | |=========================================================================| |5. Stewart In-Utero 0.5 1.3378 | 0.6689 < One | | Series | | | | 51 % with | | | no track. | |=========================================================================| |6. MacMahon In-Utero 0.9 1.3378 | 1.2040 ~ One | | Series | | |=========================================================================| |7. British Luminizers 0.1 2.9370 | 0.2937 < One | | | | | | 75 % with | | | no track. | |=========================================================================| |8. Harvey Twins 1.0 1.3378 | 1.3378 ~ One | | In-Utero Series | | | | | |=========================================================================| |9. Israeli Breast-Cancer 1.6 1.3378 | 2.140 ~ 2 | | in Scalp-Irradiation Study | | ===========================================================================
In the nine studies tabulated in Table 21-A, the observation of radiation-induced cancer means that repair is failing to become flawless even when it has to cope with average track-frequencies per nucleus of only 12 tracks, only 10, only 6, only 2 tracks, only 1 track, only 0.67 track, and only 0.29 track. If repair had been flawless, it would have successfully un-done every carcinogenic lesion, and so there would have been no excess cancer at all, in any of the nine studies. Yet large excesses were observed (see Chapter 21, Part 1).
The evidence suggests operation of the troublesome trio: The persistence of some unrepaired, unrepairable, or misrepaired carcinogenic injuries which occur at low doses in proportion to tracks, right down to the lowest conceivable dose and dose-rate.
By any reasonable standard of proof, the combined evidence in Table 21-A is conclusive that there is no safe dose or dose-rate of ionizing radiation with respect to cancer-induction. Given the redundancy of repair "equipment" which is implied by the cell studies (and by analogous physiological responses), it would not be reasonable to exclude any of the nine studies on the ground that repair of the carcinogenic injuries might have been flawless, "if only there had been fewer tracks at once."
From the evidence tabulated in Table 21-A, it would be overwhelmingly more reasonable to conclude that there is no safe dose or dose-rate.6. A Supplemental Approach -- Also via Epidemiological Evidence
We expect every sort of effort will be made to deny that the threshold issue is already settled by existing human evidence. (See Chapters 24, 34, 35.) Therefore, we are going to explore an additional piece of evidence and line of reasoning below. We do not regard this Part 6 as essential to disproof of any safe dose or dose-rate, which has been done in Part 5 by any reasonable standard of proof. We regard Part 6 as a supplemental approach. On an issue of such enormous importance to human health, we think every reasonable, alternative approach is worthwhile.
The approach to be demonstrated in Part 6 uses Baverstock's British Luminizer Series (Bav81), a series recognized as important evidence in this field by the radiation community (for instance, Nih85, p.26, and Up87, p.301).
We recognize, of course, that only about 20 % of the study-population has died of any cause, so far, and that further follow-up of the women could either weaken or strengthen the observations of excess breast-cancer mortality. (On the other hand, even zero as a lifespan excess would not alter the interim observation of an early excess -- which is a very serious radiation-effect in itself; see Chapter 12, Part 3.) Further follow-up of the UK Radium Liminizer Survey is planned. See our Chapter 21, Part 1.
Meanwhile, it is reasonable to use the existing data -- and in so doing, we will demonstrate a general method which could be applied to any other suitable data which may develop in the future.
The "Above Average" Rate of Tracks :
A line of resistance which we anticipate, to the disproof already presented, can be stated as a question: "Is it not possible that all the excess cancer in the nine studies above arose only from the cell-nuclei which experienced the `above average' rate of tracks -- and that there might have been no excess cancer if 1, 2, 6, 10, or 12 tracks per nucleus had been the maximum number instead of the average number?"
Let us consider the five studies where the average frequency of tracks per nucleus ranged from less than 1 track, to 2 tracks at once. By definition, some nuclei experience an "above average" frequency -- say 10 tracks at once. In view of the repair-capacity demonstrated in cell-studies, we regard it as self-evident that the repair-system could not be overloaded by the stress of 10, 20, 50 tracks at once. But as we stated at the outset, supplemental ways of settling the threshold issue deserve consideration.
The Essence of Flawless Repair :
The essence of perfect repair of carcinogenic damage from a track is that the carcinogenic damage disappears. Repair completely "un-does" it. With respect to any extra cancer-risk from the track, it is as if the track never traversed the nucleus at all.
It follows that, if repair were routinely and invariably perfect in every nucleus which received 4 tracks or fewer, and if the dose or dose-rate were such that no nucleus ever received more than 4 tracks (with sufficient time for flawless repair before arrival of additional tracks), then all doses and dose-rates which never deliver more than 4 tracks per nucleus would be safe.
If repair could "deliver" a threshold, everyone -- ourselves included -- would like it to occur at as high a dose as possible (to enlarge the range of safe doses). Since dose is proportional to tracks, we will explore flawless repair of damage from as many tracks, per nucleus, as the evidence might conceivably allow. Therefore, when we are referring to this supplemental approach to the threshold issue, we can call it "Max Trax."
Readers who have already studied Chapter 20, Part N, have seen how the Poisson equation can tell us -- when the average frequency of tracks per nucleus is, say, 1.0 -- how many nuclei per million receive no track, 1 track, 2 tracks, 3 tracks, 4 tracks, 5 tracks, etc. Either a nucleus is traversed somewhere by a track, or it is not. In that sense, there are no fractional tracks; fractional tracks are just an artifact from averaging the experience of many nuclei (including, of course, the nuclei which completely escape traversal).
Frequency of Multiple Nuclear Tracks in the Luminizer Study :
In the study of the British Luminizers, the average dose per exposure-session (work-day) was 0.1 rad and the average number of tracks per nucleus was 0.2937 during each exposure-session. In Chapter 20, Part N, we demonstrated how to use the Poisson equation to find out the distribution of tracks when the average frequency per nucleus has been determined. In Chapter 21 (Table 21-C), we used the Poisson equation to show the following:
Poisson Distribution of Tracks, Luminizers :
Mean number of tracks = 0.2937 per cell-nucleus. How many nuclei will get 0,1,2,3,4,5 tracks? 0.745500 = chance of exactly 0 track per nucleus. 0.218953 = chance of exactly 1 track per nucleus. 0.032153 = chance of exactly 2 tracks per nucleus. 0.003147 = chance of exactly 3 tracks per nucleus. 0.000231 = chance of exactly 4 tracks per nucleus. 0.000013 = chance of exactly 5 tracks per nucleus. 0.035545 = chance of 2 or more tracks per nucleus. (Since the probability of more than five tracks per nucleus is so low, it is reasonable to refer to two-to-five tracks as "two or more.")
We anticipate that threshold defenders will say, looking at the Poisson calculations, "Ah ha! In the Luminizer Study, 13 out of every million nuclei received five tracks per nucleus. So the excess cancer in this study does not absolutely rule out a safe dose when no nucleus receives 5 tracks at once! Maybe repair works flawlessly and provides a safe dose when there are up through four tracks per cell-nucleus, but above four tracks per nucleus, it fails to be flawless. The fifth track is the overload where trouble begins for repair. And in the Luminizer Series, there are 13 nuclei per million which have a fifth track per nucleus, and so those nuclei remained capable of causing the radiation-induced cancers observed in the Luminizer Study."
Before we can examine such a speculation, a few additional statements and definitions will be needed.
Exposures Which Increase Breast-Cancer by 50 % :
The Max Trax analysis which follows is based on comparison of the British Luminizer Study with the A-Bomb Study.
Since the conclusions do not depend upon on exactitude of input, readers are urged to accept the following approximation, which we borrow from Upton (Up87, pp.300-301). Upton states that the observed excess of breast-cancer is "of similar magnitude per unit dose" in "women exposed to A-bomb radiation" and in "women exposed occupationally to external gamma radiation in the painting of luminous clock and instrument dials" (Upton specifically includes the Baverstock study of British Luminizers, 1981).
Recently, Baverstock has reported a relative risk for breast-cancer of 1.5 in the luminizers who, at a mean age of 20, received a total dose of 40 breast-rads, at the rate of about 0.1 rad per day (see Chapter 21, Part 1). We will make the approximation that 1.5 is also the relative risk for breast-cancer in the A-Bomb Study for 20-year-old women ATB who received 40 breast-rads. This is the same as an excess relative risk of 0.5 from 40 rads, a doubling dose of 80 rads, and a K-value of 0.0125 per rem (or rad).
We insert the following interruption, however, to avert any possible misunderstanding. We recognize full well that the appropriate K-value in the A-Bomb Study might be lower than 0.0125. Nonetheless, since it does not matter here exactly what risk-coefficient we use for the A-bomb women (as we shall demonstrate shortly), we will just use Upton's generalization that the risk-coefficients are "of similar magnitude."
At 40 Rads, Number of Tracks per Nucleus :
A-Bomb Women: From Table 20-L, we will use the estimate that there were about 5.41 tracks per nucleus, on the average, from a rad of A-bomb radiation. So, (5.41 tracks per nucleus / rad) x (40 rads) = 216.4 tracks per nucleus. Since A-bomb exposure was acute, a nucleus felt all 216.4 tracks at about the same time.
British Luminizers: From Table 21-A, which is reproduced in Part 5 above, we will use the estimate that there are about 2.937 tracks per nucleus, on the average, from a rad of radium-226 gamma-rays. So, (2.937 tracks per nucleus / rad) x (40 rads) = 117.48 total tracks per nucleus. This total was spread over about 400 exposure-sessions (work-days), each delivering about 0.1 rad or 0.2937 track per nucleus, on the average.
Definition of "L" and "Effective Track" :
Meaning of "L": For the explanation which follows, we define the symbol "L" to be whatever number of carcinogenic lesions happens to occur along one primary ionization track, as it traverses a nucleus.
Meaning of "Effective Track": Within a nucleus, this means any primary track which may remain effective in producing cancer because its carcinogenic lesions (L) receive no guarantee of perfect repair. Its carcinogenic lesions may, or may not, be perfectly repaired by attempted repairs. This nuclear track has no repair-warranty.
When five nuclear tracks are present and we say that the carcinogenic lesions from four nuclear tracks are perfectly repaired, we mean that no fewer than (4 x L) or at least (4 x L) carcinogenic lesions are perfectly repaired -- regardless of which track actually produced them.
Does Repair Work Perfectly
on 4 Nuclear Tracks, but Not on 5 ?
Now the foundation has been laid for answering the original question: Is it possible that repair is invariably flawless in every nucleus where it has to deal with a maximum of only four tracks at once?
We can handle the analysis by discussing the exposure of a million breast-cells by each type of radiation: A-bomb and radium-226.
We know from the Poisson calculation above that, when a million breast-cells were exposed in the Luminizer Study, only 13 nuclei per million experienced 5 tracks per exposure-session (per work-day). Since there were 400 exposure-sessions, a total of (13 x 400), or 5,200 nuclei per million experienced 5 tracks per work-day.
We are testing the proposition that all carcinogenic injury from FOUR tracks is flawlessly repaired within the nucleus. Perfect repair makes the carcinogenic lesions from four tracks disappear from the nucleus. Therefore, in these 5,200 nuclei which received five tracks each, the equivalent of only one effective track is left per nucleus. Thus, this scenario has a total of only 5,200 effective tracks per million breast-cells -- or 5,200 nuclear tracks which are potentially effective toward cancer-production.
Since we are considering a million breast-cells at risk, we need to account now for (1,000,000 minus 5,200), or the other 994,800 nuclei. The Poisson table above shows that all of them received four tracks or even fewer. Indeed, 74.55 % of them received no track at all. Since we are testing the proposition here that repair is flawless when it is challenged by only four tracks or fewer, it means that there are no tracks left potentially effective toward cancer-production in these 994,800 nuclei.
It follows that all the excess breast-cancer in the Luminizer Study arose from a total of 5,200 nuclear tracks (per million breast-cells at risk). And it must be emphasized that "a total of 5,200 nuclear tracks" has a totally different meaning from "5,200 tracks per nucleus."
By contrast, the A-bomb women received 216.4 tracks per nucleus, all in one acute dose. Since the proposition is that 4 were flawlessly repaired, this left 212.4 effective tracks in every nucleus. With a million breast-cells at risk, this means that the the excess breast-cancer arose from 212,400,000 effective tracks. So:
o -- A-Bomb Women (Acute Exposure) :
212,400,000 effective tracks per million breast-cells provoked a 50 % increase in breast-cancer.
o -- British Luminizers (Slow Exposure) :
5,200 effective tracks per million breast-cells provoked a 50 % increase in breast-cancer.
Guaranteed perfect repair of 4 tracks "un-did" so many of the tracks, during slow delivery, that the Luminizers had to cope with (212.4 million / 5,200), or 40,846-fold fewer effective tracks than the A-bomb women. And since both groups of women showed an equal cancer-effect, we would have to conclude that each effective ("no warranty") track was 40,846 times more potent (more likely to result in cancer) in the Luminizer Study than in the A-Bomb Study. Now we will adjust for the "built-in" part of this finding. Per rad of dose, A-bomb radiation delivers 5.41 tracks per nucleus, on the average, while radium-226 delivers 2.937. So if an equal cancer-effect comes from an equal dose (40 rads), it must mean that each nuclear track from radium-226 is more potent by a factor of (5.41 / 2.937), or 1.84. So we divide the 40,846-fold disparity by the initial 1.84, and thus reduce it to 22,199.
Conclusion about Flawless Repair at 4 Tracks per Nucleus :
It is not credible that each effective Luminizer track is 22,199-times more likely to lead to cancer than each effective A-bomb track. A proposition which leads to a non-credible conclusion must be false, and so we conclude that repair does not operate flawlessly upon all carcinogenic injuries from four tracks per nucleus -- does not invariably make all carcinogenic damage from at least 4 tracks per nucleus "disappear." In other words, doses and dose-rates which deliver a maximum of only 4 tracks per nucleus are not safe doses or dose-rates. But we can still hope that repair operates perfectly at three tracks per nucleus.
Why the Exact Risk-Value Does Not Matter :
Before testing the next proposition, we will show the same result even if 80 rads (not 40 rads) were needed to provoke a 50 % increase in breast-cancer in the A-Bomb Study. Delivery of 2-fold more energy requires 2-fold more tracks: 432.8 tracks per nucleus instead of 216.4 . (The subtraction of 4 tracks -- whose carcinogenic damage is perfectly repaired -- is negligible.) In the luminizers, cumulative tracks per nucleus would remain 117.48, total. So, the relative potency of each effective track in the luminizers would appear to rise by 2-fold. But the correction-factor by which we divide at the end would also rise by 2-fold. It was (5.41 x 40) / (2.937 x 40), or 1.84. It would become (5.41 x 80) / (2.937 x 40), or 3.684 -- which is 2-fold greater than 1.84. When the factor of 2.0 operates once in each direction, it cancels itself out.
Does Repair Work Perfectly
on 3 Nuclear Tracks, but Not on More ?
One should explore the implications of speculating that repair can work flawlessly up through three tracks per nucleus, but becomes flawed beyond three. Again, we can calculate the relative track-potency under such a proposition.
British Luminizers: Out of a million breast-cells, the Poisson tabulation tells us that 231 nuclei will feel 4 tracks per exposure-session (meaning per work-day). We are testing the proposition that the carcinogenic alterations along three tracks are flawlessly repaired, which leaves one effective track in each of these 231 nuclei per exposure session, times 400 exposure-sessions, or 92,400 effective tracks per million breast-cells at risk.
In addition, there are 13 nuclei per million breast-cells which receive 5 tracks per nucleus in each session. Since damage from three tracks in each nucleus is flawlessly repaired, this leaves 13 nuclei with 2 effective tracks in each, or 26 effective tracks per exposure-session, times 400 exposure-sessions, or 10,400 additional effective tracks per million breast-cells at risk.
Total effective tracks per million breast-cells at risk = 92,400 + 10,400 = 102,800 effective tracks. In all the other nuclei, the proposition is that repair is flawless because it is not strained by an overload and it can repair the lesions along 3 tracks perfectly. Thus, in all the other nuclei, there could be no tracks left potentially effective toward cancer-production.
The A-Bomb Women: These women received 216.4 tracks per nucleus. Since the proposition is that all carcinogenic alterations inflicted by 3 tracks are flawlessly repaired, there are 213.4 effective tracks left in every nucleus. Per million breast-cells at risk, there are 213,400,000 effective tracks left.
o -- A-Bomb Women (Acute Exposure) :
213,400,000 effective tracks per million breast-cells induced a 50 % increase in breast-cancer for the A-bomb women.
o -- British Luminizers (Slow Exposure) :
102,800 effective tracks per million breast-cells induced a 50 % increase in breast-cancer for the British Luminizers.
Conclusion: Each effective track in the Luminizer Study must be 2,076 times more potent in inducing cancer than each effective track in the A-Bomb Study. Then we adjust the ratio by the 1.84 correction factor, and reduce the disparity in track-potency to 1,128-fold.
Conclusion about Flawless Repair at 3 Tracks per Nucleus :
It is not credible that each effective track in the Luminizer Study is 1,128 times more potent, in terms of carcinogenesis, than each effective track in the A-Bomb Study.
A proposition which leads to a non-credible conclusion must be false, and so we rule out the speculation that repair operates flawlessly up through three tracks per nucleus. In other words, doses and dose-rates which deliver a maximum of only 3 tracks per nucleus are not safe doses or dose-rates.
Does Repair Work Perfectly
on 2 Nuclear Tracks, but Not on More?
Now that readers are familiar with the necessary steps in testing the various propositions, we will abbreviate as we test the proposition that repair is flawless up through two tracks per nucleus, but becomes flawed beyond two.
British Luminizers: Out of a million breast-cells, the Poisson tabulation tells us that 3,147 nuclei will feel exactly 3 tracks per exposure-session. We are testing the proposition that the carcinogenic alterations along two tracks are flawlessly repaired, which means:
3147 nuclei / million will have one effective track left. For 400 sessions, we have (400 x 3147 x 1), or 1,258,800 effective tracks.
231 nuclei / million will have 2 effective tracks left. For 400 sessions, we have (400 x 231 x 2), or 184,800 effective tracks.
13 nuclei / million will have 3 effective tracks left. For 400 sessions, we have (400 x 13 x 3), or 15,600 effective tracks.
Total effective tracks per million breast cells in the Luminizer Series : 1,258,800 + 184,800 + 15,600 = 1,459,200 effective tracks.
The A-Bomb Women: These women received 216.4 tracks per nucleus. Since the proposition is that all carcinogenic alterations inflicted by 2 tracks are flawlessly repaired, there are 214.4 effective tracks left in every nucleus. Per million breast-cells at risk, there are 214,400,000 effective tracks left.
o -- A-Bomb Women (Acute Exposure) :
214,400,000 effective tracks per million breast-cells induced a 50 % increase in breast-cancer for the A-bomb women.
o -- British Luminizers (Slow Exposure) :
1,459,200 effective tracks per million breast-cells induced a 50 % increase in breast-cancer for the British Luminizers.
Conclusion: Each effective track in the Luminizer Study must be 147 times more potent in inducing cancer than each effective track in the A-Bomb Study. Then we adjust the ratio by the 1.84 correction factor, and reduce the disparity in track-potency to 78-fold.
Conclusion about Flawless Repair at 2 Tracks per Nucleus :
It is not credible, in our opinion, that each effective track in the Luminizer Study is 78 times more potent, in terms of carcinogenesis, than each effective track in the A-Bomb Study.
A proposition which leads to a non-credible conclusion must be false, and so we reject the speculation that repair operates flawlessly upon two tracks per nucleus. In other words, doses and dose-rates which deliver a maximum of only 2 tracks per nucleus are not safe doses or dose-rates.
Does Repair Work Perfectly
upon 1 Nuclear Track, but Not upon More?
Using the same steps which we used in the previous tests, we will test the proposition that repair of carcinogenic injuries is flawless on one track per nucleus, but becomes flawed beyond one.
British Luminizers: Out of a million breast-cells, the Poisson tabulation tells us that 32,153 nuclei will feel exactly 2 tracks per exposure-session. We are testing the proposition that the carcinogenic alterations along one track are flawlessly repaired, which means:
32,153 nuclei / million will have one effective track left. For 400 sessions, we have (400 x 32,153 x 1), or 12,861,200 effective tracks.
3147 nuclei / million will have 2 effective tracks left. For 400 sessions, we have (400 x 3147 x 2), or 2,517,600 effective tracks.
231 nuclei / million will have 3 effective tracks left. For 400 sessions, we have (400 x 231 x 3), or 277,200 effective tracks.
13 nuclei / million will have 4 effective tracks left. For 400 sessions, we have (400 x 13 x 4), or 20,800 effective tracks.
Total effective tracks per million breast-cells in the Luminizer Series : 12,861,200 + 2,517,600 + 277,200 + 20,800 = 15,676,800
The A-Bomb Women: These women received 216.4 tracks per nucleus. Since the proposition is that all carcinogenic alterations inflicted by 1 track are flawlessly repaired, there are 215.4 effective tracks left in every nucleus. Per million breast-cells at risk, there are 215,400,000 effective tracks left.
o -- A-Bomb Women (Acute Exposure) :
215,400,000 effective tracks per million breast-cells induced a 50 % increase in breast-cancer for the A-bomb women.
o -- British Luminizers (Slow Exposure) :
15,676,800 effective tracks per million breast-cells induced a 50 % increase in breast-cancer for the British Luminizers.
Conclusion: Each effective track in the Luminizer Study must be 13.74 times more potent in inducing cancer than each effective track in the A-Bomb Study. Then we adjust the ratio by the 1.84 correction factor, and reduce the disparity in track-potency to 7.47-fold.
Conclusion about Flawless Repair at 1 Track per Nucleus :
This finding (of a 7.47-fold disparity) is plausible enough. When we "ask" repair to work perfectly on the carcinogenic lesions from only one track per nucleus, the scenario leaves so many effective tracks that no absurd disparity develops in "track-potency." In other words, when we push the analysis to the most extreme possible speculation, it simply becomes inconclusive -- given the epidemiological evidence which is available.
However, this type of analysis is capable of addressing even the the "one track perfectly, but not two" question, if studies come along in which 40 rads are accumulated at a slower rate than 0.1 rad per exposure-session, and if the studies have enough persons and follow-up so that radiation-induced cancer could be epidemiologically detectable. If exposure extends over 10, 20, or 30 years, the radio-sensitivity of the exposed persons will gradually decline DURING the exposure, of course, and this decline will somewhat reduce the final radiation-induced excess cancer, in comparison with exposure to the same total dose received over a shorter period.
If the British Luminizers had happened to accumulate their 40 rads at a dose-rate of 0.02937 track per nucleus on the average -- instead of 0.2937 track -- the ultimate disparity in track-potency would have risen from 7.47-fold to 68.7-fold, if we disregard a gradual decline in radio-sensitivity. (The bottom of Table 20-N provides the Poisson distribution needed by readers who may want to do the other calculations.)
It would not be credible for each effective track in the Luminizer Study to be 69-fold more potent, in terms of carcinogenesis, than each effective track in the A-Bomb Study. So if the luminizers' dose-rate had been ten-fold slower, the Luminizer Study might have addressed even the most extreme scenario, and might have ended possible speculation that repair might handle a maximum of one nuclear track flawlessly, but not two.
Conclusion from the Entire Supplemental Approach :
For convenience, we shall summarize the findings here from the comparisons made above:Perfectly Relative Carcinogenic Repairable Potency Tracks of Effective Tracks per Nucleus (Luminizer / A-Bomb) ------------ ---------------------- 4 yes, but not 5 or more 22,199 3 yes, but not 4 or more 1,128 2 yes, but not 3 or more 78 1 yes, but not 2 or more 7
The supplemental evidence provides strong, additional confirmation of the conclusion reached in Part 5 of this chapter: Repair of carcinogenic alterations does not become invariably perfect in every injured nucleus, even when the maximum strain per nucleus falls to minimum levels. Strain, in this context, means the number of primary ionization tracks per nucleus, per exposure-session.
The distinction between maximum strain and average strain per nucleus is crucial. In the British Luminizer Study, when the average strain was 0.2937 track per nucleus per exposure-session, strain was absent (no track at all) in about 75.55 percent of the nuclei, the maximum strain in 22 percent of the nuclei was one track, and the maximum strain in about 3.5 percent of the nuclei was two to five tracks, per exposure-session (work-day).
The supplemental approach to the threshold issue, demonstrated here in Part 6, shows that speculation about "perfect repair" leads to non-credible consequences, even when the repair-system is tested for perfection in handling a maximum of only 4 tracks, 3 tracks, or 2 tracks per nucleus. In other words, the supplemental approach says that doses and dose-rates which deliver a maximum of only 4, 3, or 2 tracks per nucleus are not safe doses or dose-rates. With respect to a maximum rate of one track per nucleus, the British Luminizer data are incapable of shedding additional light, one way or the other.
We conclude Part 6 with a reminder. Disproof of any safe dose or dose-rate was achieved in Part 5, by any reasonable standard of proof. It did not depend on the British Luminizers. Far from it. Part 6 has been a supplemental approach, presently tied to the British Luminizer Series. However, the general Max Trax method demonstrated above is applicable to any suitable studies which might become available in the future.7. Comparison of Two Explanations for the Observations
Although the Max Trax approach was inconclusive with respect to perfect repair when we tested the available evidence for a maximum strain of one track per nucleus, there are overwhelming reasons to rule out any proposition that doses or dose-rates which inflict a maximum strain of one track per nucleus are safe.
For instance, such a safe-dose proposition would have to be plausibly reconciled with the indication by Max Trax that repair is not perfect when the maximum strain per nucleus is only four, three, or two simultaneous tracks per nucleus. To achieve reconciliation, safe-dose proponents might have to propose that the repair-system runs short of some necessary enzyme or other necessity, as the strain on the system rises from one track up to two, three, or four tracks per nucleus, per work-day.
However, this proposition would lead to a non-credible consequence, namely gross incompatibility with the evidence from cell-studies that -- even in the presence of hundreds of simultaneous tracks -- there is no shortage of repair-capacity for DNA and chromosome injuries.
(On the other hand, if repair is especially poor at low doses because of insufficient stimulation, then safe-dose proponents are in real difficulty, too. Perhaps one track per nucleus stimulates the least repair-capacity per track.)
Rescue for the Safe-Dose Hypothesis ?
Altogether, the findings in this chapter would seem to require that threshold supporters develop a drastically revised defense for their safe-dose hypothesis. Their casual references to "repair" -- as if "repair" would automatically assure some safe dose -- have become at variance with the actual epidemiologic evidence. Some supporters may propose a chain of new speculations, along the following line.
A. They might deny that carcinogenesis is related to injury of DNA or chromosomes at all. They might dismiss the huge and growing body of evidence which suggests that it is. Then they could deny that evidence from cell-studies, about DNA and chromosome repair, is relevant to radiation carcinogenesis.
B. But this would amount to their proposing that cancer induced by ionizing radiation, in the nine studies of Table 21-A and in all the higher-dose studies too, results from some wholly unknown mechanism at some wholly unknown site.
C. Then -- in order to suggest the possibility of a safe dose -- they would need to propose that this wholly unknown process also has its own wholly unknown repair-system.
D. But, because it is clear (from the failure of dose-response to be concave-upward) that this wholly unknown repair-system is not reducing the cancer-risk per rad as dose approaches zero, and because it is clear (from the nine studies) that this wholly unknown repair-system is not perfect even when track-frequency is very low, they would need to propose that this repair-system becomes saturated at very low doses and that saturation accounts for repair's failure in the nine studies.
E. And lastly, in order to rescue the safe-dose hypothesis, they would need to propose that there is some frequency of ionization tracks per cell -- an average frequency even lower than in the nine studies -- at which this wholly unknown repair-system is un-depleted and also flawless.
Although this chain of speculations may seem far-fetched to some readers, points A and B have already been proposed as a response to our disproof of any safe dose or dose-rate. Points C, D, and E would seem to follow from A and B, since threshold proponents have an obligation to take account of the real-world human epidemiological evidence in some way, if not in our way.
Denial of the premise (about the role of DNA and chromosomes in radiation's carcinogenic action) cannot restore plausibility to the safe-dose hypothesis. It can only change the number of tracks per exposure, somewhat, by postulating that the cytoplasm, not the nucleus, is the site of carcinogenic injury.
Even so, if there is any radiation dose at all, primary ionization tracks occur and inflict injury in random fashion. So unless flawless repair suddenly were to occur at track-rates below the very low rates of the nine studies, there would be unrepaired, unrepairable, or misrepaired carcinogenic injuries (located in the cytoplasm), right down to the lowest conceivable dose or dose-rate.
A Competing Explanation of the Evidence :
Scientists worldwide are familiar with the great principle of economy in logic known as "Ockham's Razor," enunciated (in Latin) by the 14th century English philosopher, William of Ockham. The principle warns against fabricating many explanations when one is sufficient: "Entities [explanations] should not be multiplied beyond what is needed."
Instead of fabricating a series of speculations like paragraphs A through E above, and instead of dismissing without any basis a whole body of evidence (which links genetic information with cancer), we think it is far more reasonable for us to suggest a highly plausible -- virtually obvious -- explanation of all the observations which relate to the threshold issue:
Whenever an ionization track traverses a nucleus, there is always a chance that it will cause a carcinogenic lesion and that the lesion will be unrepaired, inherently unrepairable, or misrepaired. In short, there is an inherent failure-rate in the repair-system.
This hypothesis requires no denial of all the evidence linking genetic information with cancer. Moreover, it is consistent with the observation of imperfect operations in other physiological systems. And it would explain all sorts of the specific observations in this field.
For instance, this hypothesis would explain why radiation-induced cancer is found in a host of human studies between 10-400 rems -- a dose-range where cell-studies indicate no shortage of genetic repair-capacity.
For instance, this hypothesis would explain the observation (in several human studies) of radiation-induced chromosome aberrations persisting in persons who received their doses at minimal dose-rates -- from weapons-fallout, elevated background doses, and routine occupational exposures.
For instance, this hypothesis would explain the observation of radiation-induced cancer in the nine studies (Table 21-A) where repair was challenged by so very few tracks per nucleus, on the average.
And this hypothesis would explain the supplemental Max Trax results, which indicate that the repair-process is not routinely and invariably flawless even when it has to cope with a maximum track-rate per nucleus of only four, three, or two tracks.
This hypothesis is, of course, incompatible with any safe dose or dose-rate of ionizing radiation. If there is any dose at all, tracks occur and inflict injuries in a random fashion. So if there is some inherent chance of failure in the repair of carcinogenic damage whenever there are ionization tracks, then this chance will be present right down to the lowest conceivable dose or dose-rate.8. The Bottom Line
No one denies anymore that low-LET ionizing radiation is a human carcinogen. The threshold question is: Does it stop being a human carcinogen when the dose or dose-rate is sufficiently low?
There are a multitude of low-dose human studies which are inherently incapable of helping to settle the threshold issue (discussion in Chapter 21, Part 2). However, we have assembled nine human epidemiological studies which ARE capable.
Together, they amount to proof that repair of radiation-induced carcinogenic lesions, at the cellular level, fails to "deliver" a safe dose or dose-rate of ionizing radiation with respect to human cancer-induction, even when the strain on the repair-system is minimal. Indeed, in five of the nine studies, the strain per cell-nucleus is an average of only one or two simultaneous tracks . . . with an average even below 1.0 in some studies.
The nine studies are supplemented by an analysis which also relies on epidemiological observations. The supplemental analysis indicates that the hypothesis of perfect repair leads to non-credible conclusions, and that there is a failure-rate in the repair-system for carcinogenic lesions -- even when a nucleus has to cope with a maximum of only 4, 3, or 2 primary tracks per nucleus, per exposure-session.
And so the human epidemiological evidence establishes -- by any reasonable standard of proof -- that there is no safe dose or dose-rate . . . unless there exists some wholly separate, post-repair system in the body which also needs consideration.
The Immune System, or Other Post-Repair Defense Mechanisms :
We all know the common refrain that the body has surveillance systems which are constantly rejecting cancers which are constantly being formed in the body, and "without the immune surveillance system, everyone would die of cancer."
So we will consider the proposition that, if the immune or other surveillance system has such prowess in preventing cancer, such a system must be easily able to "take care of" the residual problem left by repair's small failure-rate.
How we wish this attractive proposition were true. But clearly it is not. The reality is that, in countries like the United States, about twenty percent of the population dies of cancer. The percentage of persons with cancers -- not prevented by immune or other defense mechanisms -- is even larger if the non-fatal cancers are included. Obviously, the immune or other surveillance mechanisms are failing to prevent huge numbers of cancers, since they are failing to prevent about one in five persons from being killed by this disease.
Is it possible to reconcile this reality with the hope that, somehow, the same flawed defense mechanisms "take care of" every potential radiation-induced cancer missed by repair, provided the carcinogenic injury occurred along a track received at a very low dose? Reconciliation would require some pure fantasies.
For instance, potential cancers induced by ionizing radiation would have to look different from other potential cancers which are watched by the surveillance systems. Otherwise, the surveillance systems would be unable to select them out for special (perfect) treatment.
In addition, in order for this last line of defense to work perfectly for low-dose radiation when it obviously does not work perfectly at higher doses, the potential cancers induced by tracks at low doses would have to sprout a little flag identifying themselves as low-dose products.
Since such fantasies would strain even the greatest credulity, we must discard this last hope for a safe dose or dose-rate.
The Initial Five Points, Condensed :
Below, we shall condense the five-point summary (from Part 1 of this chapter) even further:
- One primary ionization track is the least possible disturbance which can occur at the cellular level from ionizing radiation. Without a track, there is no dose at all.
- Every primary ionization track has a chance of inducing cancer by inducing carcinogenic injuries; it needs no help from any other track.
- This means that there is no conceivable dose or dose-rate which can be safe, unless (A) the repair-system always successfully un-does every carcinogenic lesion, when the dose or dose-rate is sufficiently low, or (B) every failure of the repair-system, at low doses, is always successfully eliminated by some post-repair defense-system.
- Human epidemiological evidence shows that the repair-system for radiation-induced carcinogenic lesions has a failure-rate even under minimal strain.
- Observation and logic show that post-repair defense-systems (for instance, the immune system) cannot possibly be perfect with respect to providing a safe dose or dose-rate of ionizing radiation.
It follows that there is no safe dose or dose-rate of ionizing radiation, with respect to induction of human cancer. The risk is related to dose, right down to zero dose.
Beyond A Reasonable Doubt :
The existing human evidence shows, beyond a reasonable doubt, that there is no conceivable dose or dose-rate of low-LET ionizing radiation which is safe, with respect to producing fatal cancer in humans.
From 9 rads right down to 0.1 rad, the epidemiological evidence speaks for itself, without reliance on any hypothesis or presumption at all. The evidence includes adults, children, high-energy gamma rays, diagnostic X-rays, acute delivery and very slow delivery. And between zero dose, and the doses tested directly by the nine studies, the calculations which disprove any safe dose or dose-rate rely on only one extremely reasonable assumption -- namely, that radiation-induced cancer originates from events in the nucleus.
In the face of the evidence, I could not possibly suggest in this book that the safe-dose question cannot be answered at all or that it could readily go in either direction.
One might watch out for inconsistent attitudes toward what constitutes proof. Chapter 22 shows that, without any proof of correctness whatsoever and in the face of contrary human evidence of good quality, much of the radiation community has routinely divided observed per-rad cancer-risks by numbers like 2 to 10 when estimating per-rad risk at low doses and dose rates. Now, when it comes to settling the safe-dose issue, it would appear inconsistent if similar segments of the radiation community were to demand unreasonably large amounts of human evidence, or were to become "ultra-careful" about acceptable levels of proof.
Real vs. Imaginary Radiation-Casualties :
By reasonable standards of proof, the safe-dose hypothesis is not merely implausible -- it is disproven.
Disproof of any safe dose or dose-rate invalidates suggestions that, whenever an analyst calculates a number of radiation-induced cancers to be caused by very low-dose exposure, the cancers are just "hypothetical," "speculative," "theoretical," "non-existent," or "imaginary."
It is true, of course, that radiation-induced cancers in a population from very low doses will rarely if ever be detectable epidemiologically, because of the signal-to-noise ratio (see Chapter 21). But it does not follow (from the lack of direct observation) that the cancers are therefore unreal, hypothetical, speculative, theoretical, non-existent, or imaginary. No rational person will deny that one of the most commonplace (and important) functions of science is to let people know what is really happening when direct observation is impossible.
We conclude with a warning: Disproof of any safe dose or dose-rate means that fatal cancers from minimal doses and dose-rates of ionizing radiation are not imaginary. They are really occurring in exposed populations. Proposals, to declare that they need not be considered, have health implications extending far beyond the radiation issue, as pointed out in Chapter 24, Part 10 and Chapter 25, Part 5.