Inconsistencies and Open Questions Regarding Low-Dose Health Effects of Ionizing Radiation · 2008....

Inconsistencies and Open Questions Regarding Low-Dose Health Effects ofIonizing Radiation

Inconsistencies and Open Questions Regarding Low-DoseHealth Effects of Ionizing Radiation

Wolfgang Kohnlein, Munster, Rudi H. Nussbaum, Portland

AbstractThe effects on human health of exposuresto ionizing radiation at low doses have longbeen the subject of dispute. In this paperwe focus on "open questions" regarding thehealth effects of low dose exposures thatrequire further investigations. Seeminglycontradictory findings of radiation healtheffects have been reported for the same ex-posed population or inconsistent estimatesof radiation risks were found when diffe-rent populations and exposure conditionswere compared.Such discrepancies may be indicative: (1)of differences in sensitivities among theapplied methods of epidemiological analy-sis or (2) of significant discrepancies inhealth consequences following comparabletotal exposures of different populations un-der varying conditions.We focus first on inconsistencies and con-tradictions in presentations of the "state ofknowledge" by different authoritative ex-perts. Subsequently, we review studies thatfound positive associations between expo-sure and risks in dose ranges where tradi-tional notions generalized primarily fromhigh dose studies of A bomb survivors orexposed animals would have predictednegligible effects. One persistent notion inmany reviews of low dose effects is the hy-pothesis of reduced biological effectivenessof fractionated low dose exposures, com-pared to that of the same acute dose. Thisassumption is not supported by data onhuman populations.From studies of populations that live incontaminated areas, more and more eviden-ce is accumulating on unusual rates of va-rious diseases, other than radiation induced

63

malignancies, health effects that aresuspected to be associated with relativelylow levels of internal exposures originatingfrom radioactive fallout.Such effects include congenital defects, neonatal mortality, stillbirths and possibly ge-netically transmitted disease. A range ofopen questions challenges physicians andradiation experts to test imaginative hypo-theses about induction of disease by radia-tion with novel research strategies.

I INTRODUCTION1.1 Low dose radiation health effects: de-fining the "state of knowledge"The "state of knowledge" of health effectsfrom low dose exposures to ionizing radia-tion has recently been reviewed in exten-sive reports by three prestigeous nationaland international commissions of scientificand medical experts with partially over-lapping membership, known by theiracronyms UNSCEAR [89], BEIR V [4] andICRP [39]. Publication of' these reportswas followed by a number of summaries inscientific journals, authored by recognizedradiation experts, that purport to present a"scientific consensus" of low dose effectsin a more accessible format for health pro-fessionals. A critical comparison betweenvarious presentations of "accepted views",however, reveals inconsistencies, in bothcategories, that of "established facts" andthat of "unsettled questions" [28].

1.2 Inconsistencies and open questionsIn 1990 the BEIR V Committee (composedof 17 experts on radiation epidemiology,bio effects, and risk estimation) issued a400+ pages report [4] which serves as a

---------------------------_!!!'!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!,~.!"!"_!!,!! ....~....,


widely quoted and prestigeous review oflow dose radiation health effects. In thebody of this report, the Committeeacknowledges some critical areas of uncer-tainty and controversy, particularly withregard to estimates of radiogenic risk per-taining to anthropogenic increases in lowdose exposures above unavoidable naturalbackground levels, both occupational andenvironmental. Obviously, such estimatesare of the greatest importance to guidelinesfor the protection of public health. Yet,within the BEIR V report, we find incon-sistencies between the Committee's con-clusions, as stated on different pages (seesec. 1.2.1 below). Moreover, few of theseobviously unresolved questions found theirway into the most widely quoted ExecutiveSummary. Subsequent authoritative over-views in scientific journals have not onlyglossed over some of these inconsistenciesin the BEIR V report, but they also presentdifferent views of what constitute "wellestablished" and "unproven" aspects of lowdose health effects. We will highlight someof these inconsistencies by quoting or para-phrasing statements from the BEIR V re-port and comparing them with assertions onthe same topics from three subsequentjournal reviews, all citing BEIR V as amajor source. Editorial comments, re-flecting on the citations, have been placedin square brackets. In our discussions,"low doses" means the dose range well be-low 50 cGy.We will select five controversial issues inthe debate about protracted low dose expo-sures, to illustrate our point.

12.1 BEIR V [4]A. Shape of a dose effect curve for can-cer inductionIn several places of its report, the BEIR VCommittee concurs with the large team ofscientists at the Radiation Effects ResearchFoundation in Hiroshima, Japan, which hascollected and analyzed the Life Span Study

64

(LSS) of A bomb survivors for decades:after a one time (acute) exposure, a linear,non threshold relation between excess mor-tality from cancers, except leukemia, anddose gives an excellent fit to the 1950 1985LSS data, if restricted to doses below 200cGy. However, BEIR V "recognizes that itsrisk estimates become more uncertain whenapplied to very low doses" and the Com-mittee concedes rather obliquely that"departures from a linear model at low do-ses, however, could either increase ordecrease the risk per unit dose" (p.6).

B. Dose rate effectiveness factor (DREF)at low doses (see sec. 11.1)In its report, the BEIR V Committee states:"For low LET radiation [low linear energytransfer, such as from beta and gamma ra-diation] , accumulation of the same [total]dose over weeks or months, however, isexpected to reduce the lifetime risk appre-ciably, possibly by a factor 2 or more"(p.6) Such a downward correction for line-arly extrapolated risk values is calledDREF (Dose Rate Effectiveness Factor).On the next page (p.7), however, we read:"While experiments with laboratory ani-mals indicate that the carcinogenic effec-tiveness per Gy of low LET radiation is ge-nerally reduced at low doses and low doserates, epidemiological data on the carcino-genic effects of low LET radiation are re-stricted largely to the effects of exposuresat high dose rates. Continued research isneeded, therefore, to quantify the extent towhich carcinogenic effectiveness of lowLET radiation may be reduced by fractio-nation or protraction of exposure".For decades, findings from animal experi-ments at high and very high doses have gi-ven support to the speculation that thehu-man dose effect relation, for cancer in-duction is strongly concave if low dose ex-posures are accumulated over extendedtime periods (dose fractionation). Such arelation implies a practically zero effect

Inconsistencies and Open Questions Regarding Low-Dose Health Effects ofIonizing Radiation -

threshold at doses of the order of naturalbackground irradiation and a significantlysmaller risk per unit dose at lower than athigher doses.Fifteen pages later, the Committee states:"There are scant human data that allow anestimate of the dose rate effectivenessfactor (DREF)" (p.22).Then, in a subsequent section the reportpicks up the same topic:"Since the risk models were derived pri-marily from data on acute exposures ..., theapplication of these models to continuouslow dose rate exposures requires conside-ration of the dose rate effectiveness factor(DREF) .... For the leukemia data, a linearextrapolation indicates that the lifetimerisks per unit bone marrow dose may behalf as large for continuous low dose rate asfor instantaneous high dose rate. For mostother cancers in the LSS, the quadratic con-tribution is nearly zero, and the estimatedDREFs are near unity. Nevertheless, thecommittee judged that some account shouldbe taken of dose rate effects and in ChapterI suggests a range of DREFs that may beapplicable" (p.l71 4).

C. Biological effectiveness of X rays ver-sus gamma raysReferring to work by a previous authorita-tive radiation commission, the InternationalCommission on Radiation Units andMeasurement [38] (ICRU), BEIR V states:"Most human exposures to low LET ioni-zing radiation are to X rays, while the Abomb survivors received low LET radiationin the form of high energy gamma rays.These are reported to be only half as ef-fective as ortho voltage X rays. While thatis not the conclusion of this Committee,which did not consider this question in de-tail, it could be argued that since the riskestimates that are presented in this reportare derived chiefly (or exclusively) fromthe Japanese experience they should bedoubled as they may be applied to medical,

65

industrial, or other X ray exposures"(p.218).The physical basis for such a possible ef-fect is the roughly four fold higher ioniza-tion density in tissue by medical X raysthan that by high energy gamma rays [42].

D. Role of free radicals in tumorigenesisby ionizing radiation"To the extent that the effects of radiationare mediated by free radicals, which canalso mediate the effects of promotingagents, sequential exposures to radiationmay serve to promote tumorigenesisthrough mechanisms similar to those ofchemical promoting agents" (p.139)The report gives, however, no further con-sideration to the question, whether radio-genic free radical production, in particular,at low doses and low dose rates could linkprotracted low level exposures to variousdiseases or immune depression, known tobe promoted by these highly reactive che-mical species [30].

E. Radiation hormesisOn p. 383 the report states:"Although 'beneficial' effects of radiationhave been alleged on the basis of reducedmortality in high background areas in theUnited States, analyses that include an ad-justment for altitude indicate no,beneficial' effects .... This apparently,beneficial' effect of radiation may, in fact,be an example of confounding ...."

1.2.2 "State of knowledge" summaries af-ter BEIR V .The first of the three summaries discussedbelow was published in a journal for publichealth professionals by members of theBEIR V Committee [91]. Hence its state-ments conform largely with the BEIR V re-port, except for some significant omissions.The other two summaries [33, 52] showdeviations, as well as omissions, comparedto the BEIR V report. They have been di-


Hendee [33]A: Shape of dose effect curve

The non threshold dose in- The linear model furnishes Induction of mutations Incidence hypothesis, first the most conservative (i.e. human cells is a no thresh-supported by the association highest) risk estimates for old linear function of dose,between childhood leukemia exposures to low doses of independent of dose rate.and pre natal diagnostic x irradiation, even though evi- The dose response for in-radiation at doses compa- dence establishing the linear duction of breast cancer israble to natural background, model as the correct rela- linear without threshold.has been extended to other tionship is still relatively in- While there are several epi-malignancies, as well as to conclusive. demiological studies thatgenetically significant muta- have purported to showtions. Data on teratogenic carcinogenic or leukemo-effects (e.g. small brain size genic effects of irradiationor severe mental retarda- in the dose range below 10tion) are also compatible cGy, there are no theoreticalwith a nonthreshold linear reasons, nor are there sup-dose effect curve. porting animal data, or low

dose A bomb survivor datain the range 1 - 9 cGy sug-gesting that there should bea convex upward dose rela-tion, that would be requiredto observe a rapidly risingcancer incidence at very lowdoses, close to natural back-ground.

Upton et al. [91]

B: Doserate effectivenessIn the absence of adequate Suggests, [somewhat obli-human data on the carcino- quely] that a DREF of 2.25genenicity of protracted low (from a 1980 BEIR report)LET irradiation, the BEIR V should be applied to theCommittee was unable to BEIR V risks. [No specificspecify the extent to which justification is given, othertheir projections may over- than that it would reduceestimate the risks of a dose risks closer to earlier esti-of radiation that is accumu- mates.]lated over long periods oftime.

66

Little [52]

The dose rate effect for in-duction of specific genemutations In human cellsmay be significantly lessthan that observed in rodentcells. Nevertheless, whenthe experimental data areconsidered along with limi-ted epiidemiologic data, aDREF of 2has been recom-mended for chronic expo-sures. However, little or nodecrease in risk was obser-ved for induction of breastcancer, when the dose wasreceived In a protractedmanner, as opposed to asingle brief exposure.

Inconsistencies and Open Questions Regarding Low-Do'se Health Effects ofIonizing Radiation

c: X-rays versus Gamma-ra,s , ,[Not mentioned] [X ray exposures of most [Not mentioned]

medical workers far belowprotection guidelines arediscussed, but no mention ofa possibly higher biologicaleffectiveness of X rays,compared to gamma rays onwhich the guidelines are ba-sed.]

D: Free radicals "

[Not mentioned] [Not mentioned] Ionization results in the pro-duction of free radicals thatare extremely reactive andmay lead to permanent da-mage of affected molecules.

E: Radiation hormesisAlthough several studies [Not mentioned] A lack of correlation be-have found that the rates of tween cancer incidence andcancer and other diseases background radiation wasvary inversely with natural observed ill different stu-background radiation levels, dies. Low dose epidemio-which some investigators logic studies in populationshave interpreted as evidence of limited size must be care-of beneficial ("hormetic") fully controlled, and are of-effects of low level irradia- ten prone to bias by con-tion, the relationship does founding factors.not persist after the effectsof altitude and other con-founding variables havebeen adequately controlled.

rected to physicians and radiologists in ge-neral. Quotes or paraphrases from thesereviews have been keyed to the topics A toE, as defined above, for convenient com-parIson.The usefulness of reviewing "unansweredquestions after BEIR V" for the purpose ofidentifying new directions for investiga-tions, was recently recognized by other re-searchers in the field [34].The present contribution is predicated onthe premise that a special focus on unrefu-ted positive associations of very low doseexposures with health effects that are in-

67

consistent with long held notions, will sug-gest unorthodox hypotheses. Testing thesewill require investigations in yet insuf-ficiently explored areas that are likely toreveal a greater than expected complexityof interactions between low dose radiationexposures, other environmental toxics anddisease.Because of their dominance in shaping pre-valent notions about the effects of radia-tion, we briefly review the findings fromthe A bomb survivor study, with particularemphasis on low dose effects. In subse-quent sections we summarize a selection of

Inconsistencies and Open" Questions Regarding Low-Dose Health Effects ofIonizing Radiation

studies that are pertinent to our above sta-ted premise.

II THE FOLLOW UP STUDY OF ABOMB SURVIVORS (Acute Exposures)ILl Evolution of Official Low Dose Ra-diation Risk EstimatesOfficially adopted radiation risk estimatesabout health effects of radiation at low do-ses have been based primarily on extrapo-lations from the continuing follow up studyof about 90,000 inhabitants of Hiroshimaand Nagasaki who had survived the firstfive years after the physical and social de-vastation caused by the atomic bombs.Until the mid 1970s cancer mortalitiesamong survivors with exposures below 100cGy had not shown statistically significantexcesses above Japanese national averages,in contrast to findings at higher exposures.Growing demands for occupational and ge-neral radiation protection standards leadnational and international radiation regula-tory commissions to resort to models fordownward extrapolation to reasonablelevels of occupational exposure from thewell established high dose observations. Byimplicitly postulating the existence of auniversally valid dose effect relation, theICRP [37] UNSCEAR [89] and BEIR III[3] reports in the late seventies, all conclu-ded either explicitly or implicitly that line-ar, no threshold extrapolation from highdose A bomb survivor mortalities would infact overestimate low dose radiogenic risks.For fractionated low dose exposures "doserate effectiveness factors" (DREF's) of atleast a factor 2, were recommended.However, microdosimetric analyses haveshown, that at decreasing doses, the con-cept of dose rate looses its meaning entirelybecause of the discrete nature of the radia-tion " cell interaction: the smallest possibleeffect must be caused by a single cell tra-versal [2, 26].More recently, official evaluations of can-cer risk from ionizing radiation have un-

68

dergone significant upward revisions com-pared to those published about a decadeearlier [4, 39, 89].For the non leukemia A bomb data, RERFanalysts found that a DREF value muchabove one for acute low dose exposures isnot consistent with the updated data [60,61, 92]. Yet, disregarding the new eviden-ce, the conclusions by UNSCEAR [89],BEIR V [4], and ICRP [39] retained theirprevious recommendations to reduce esti-mates of radiogenic risks, based on a lineardose effect model, for protracted low doseexposures by DREF corrections of at least afactor of two (see above).

11.2 A bomb Survivor Study as UniversalStandardThe interpretations of A bomb survivors'cancer mortality or incidence statistics byscientists at the Radiation Effects ResearchFoundation (RERF) in Hiroshima and otherofficial commissions, have become the au-thoritative standard to which all findingsfrom epidemiological studies on other ex-posed populations, such as nuclear workers,have been compared. In particular, studiesthat found substantially higher radiogenicrisks at low doses and low dose rates thanthose officially adopted [96] have beenlabeled "renegade" by some recognizedradiation experts and have been imputed tobe in error by others [67, 83, 97]. Ratherthan questioning the comparability of in-congruent studies, some epidemiologistsinvoke bias of unknown origin in the occu-pational data in order to set aside their ownfindings, if they differ from those derivedfrom LSS statistics [24]. Almost no atten-tion has been given to evidence in theRERF data that these discrepancies mightreflect unrecognized intrinsic incommen-surabilities in health profiles and age distri-butions, between the LSS cohort and aworker population quite apart from thevastly different characteristics of irradiation[77, 78]. Adopting the LSS findings as a


universal standard also implies the untestedhypothesis that a single dose effect rela-tionship can describe all conditions of ex-posure [96].

11.3 Direct Evaluation of IncrementalExcess Cancer Risk from MortalitiesAmong the Lowest Dose SubcohortsLinear extrapolation models used by BEIRV and RERF to predict low dose risk valuescan be checked by a straightforward analy-sis of mortality data, limited to the lowestdose sub cohorts. The methods used in allofficial analyses of A bombrmortality datahave weighted the resulting risk values to-ward those observed in the medium to highdose range [62]. Recently, two groups ofresearchers published independent analysesthat were restricted to cancer mortalitiesamong the A bomb survivors who had beenexposed to less than 50 or 100 cGy [26, 50,56]. These low dose sub cohorts includeabout 80% of the entire LSS cohort. Usingthe 1950 1985 follow up data [71], andcombining new DS86 sub cohorts fromboth cities, these authors have shown sta"tistically significant (p < 0.01) excess mor-talities (for cancers except leukemia) forthe combined ,,6-19 cGy" sub cohort (meancolon dose 10.9 cGy) compared to thecombined ,,0-5 cGy" sub cohort (mean co-lon dose 0.7 cGy) (Fig. 1). The ,,0 - 5 cGy"dose group was chosen for comparison,rather than RERFs "zero" dose group,since the combined sub cohort includessurvivors, nominally unexposed to the ra-diation flash from the explosions, as well asan unknown fraction who at that distancefrom the epicenter were affected by falloutexposures [59]. This additional dose is notreflected in DS86 estimates of individualdoses. Other uncertainties have arisenrecently in regard to the contributions ofneutrons to individual doses of survivors,especially affecting the low dose sub co-horts who were located at large distancesfrom the explosions [63, 80, 81]. For the

69

lowest dose DS86 sub cohorts, we can thusexpect that upward corrections in mean do-ses will have to be made, with the greatestcorrection to the lowest mean doses,decreasing rapidly with increasing DS86mean dose.A graphical display of cancer mortality ver"sus mean dose elucidates more directly therelevant dose response association than theusual display of relative risk versus dose.Weighted linear regression analysis overthe dose ranges listed in Table I anddisplayed in Fig. 1, yields a higher slopefor mortality versus dose (or incrementalrisk per unit dose) for the dose range ,,0-19cGy" than for the dose range ,,6-99 cGy".While statistically only weakly significant,the 1950-1985 survivor mortality data forthe low dose range, suggest that the incre-mental excess cancer risk per cGy forsingle exposures may be greater below 20cGy, than in the medium dose range 20 -100 cGy, for which our estimate of excesslifetime risk (9± 1) per 104 p-cGy (Figure 1and Table I) is consistent with the value ofabout 12 per 104 P cGy published by RERFanalysts [71] or the value of about 7 per 104p-cGy from BEIR V [4]. To check ourconjecture and possible bias from using ag-gregate mortalities, one of RERFs chiefstatisticians applied a more extensive mo-del for fitting excess relative risk that in-cludes stratification for city, sex, age at ex-posure and follow up period. For the mor-tality data below 100 cGy, he found im-provement in the fit for excess relative riskproportional to the square root of dose(convex curve) compared to a linear dosedependence [Donald A. Pierce, privatecommunication 1991]. Unfortunately, up-dated mortality data for 1950-1990, haveyet to be published by RERF. Non uniformupward corrections to sub cohort mean do-ses due to unaccounted for fallout or neu-tron doses might well augment the convexshape of the dose effect relation. In thiscontext, it is noteworthy, that RERF ana-


lysts, studying the issue of a hypothesizedthreshold and the shape of the dose respon-se curve for leukemia (acute lymphocyticleukemia or ALL and chronic myeloid leu-kemia or CML) among the LSS cohort atvery low doses, found a better fit of thedata to a non threshold convex dose effectrelation (logarithmic with dose) than to alinear one with a hypothesized 5 cGythreshold [12].

11.4 Summary of low dose effects fromthe A bomb survivor studyFindings from the A bomb survivor fo]]owup studies (DS86, 1950-1985 fo]]ow up)which contradict the validity of applying aDREF to low dose exposures:* (1) both the A bomb survivor cancermortality (1950 1985) and incidence data(1950-1987) fail to suggest the existence ofa threshold for cancer induction down tovery low doses [17, 72, 92].* (2) doses less than 5 cGy and pro-bably as low as 1.6 cGy have been associa-ted with excess cases of leukemia (ALLand CML) among A bomb survivors [12,85]. Carter [12], found a better fit of thedata to a non threshold convex dose effectrelation (logarithmic with dose) than to alinear one with a hypothesized 5 cGythreshold (p = 0.056).* (3) doses in the range from less thanone to a few cGy have been associatedwith brain damage in pre nata]]y exposedchildren of A bomb survivors [70].* (4) mortality for solid cancers in the,,6-19" cGy dose group (mean colon dose10.9 cGy) is significantly higher (p < 0.01)than it is in the ,,0-5" cGy dose group(mean colon dose 0.7 cGy), and there is asuggestion for a convex dose relation.(section II.3)

III EFFECTS FROM OCCUP ATIO-NAL EXPOSURES (Protracted Exposu-res)

111.1 Critical evaluation of governmentsponsored nuclear worker studiesSo far, practically a]] epidemiological stu-dies of nuclear worker populations in theindustrialized world have been funded andoverseen directly or indirectly by govern-ment agencies that have promoted militaryand civilian nuclear technologies. Histori-ca]]y, production interests in nuclear instal-lations have competed directly with con-cerns for the protection of workers or pub-lic health.The impact of this situation on the qualityof radiation epidemiological research hasbeen amply demonstrated by a critical re-view of 124 U.S. and British governmentstudies undertaken by a task force of twelveindependent physicians and epidemiolo-gists assembled and sponsored by Phy-sicians for Social Responsibility. Their eye-opening report concludes that:(1) "The Department of Energy's (DOE)(and its predecessor agencies') epidemio-logy program is seriously flawed ...(2) There appear to be major inac-curacies, and serious questions as to con-sistency and reliability in the measurementsof the radiation exposures.(3) The nearly exclusive focus onmortality studies .., eliminates from con-sideration virtua]]y an cancers which maybe related to radiation exposure but whichwi11not or have not yet caused death, andthus severely limits our knowledge of thehealth consequences of low level ionizingradiation exposure ....(4) ... the problems and flaws evidentin many investigations are precisely thosewhich tend to produce false negative re-sults." [20].A large number of the mortality studies un-der review found no statistically significantassociation between cancer induction and

70


low dose radiation exposures. Most of themextended over limited follow up periods,too short to observe long latencies. Also,when workers' mortalities are being compa-red to national rates, the findings are biasedtoward lower risk for all causes of deathamong radiation workers (healthy workereffect).Nevertheless, in a few of the reviewed stu-dies and in some that have been publishedmore recently, significant increases in 'spe-cific types of· cancer were found, for ex-ample prostatic cancer [5, 35], multiplemyeloma, lymphatic and hemapoetic neo-plasms, and bladder cancer [74], leukemia[95], multiple myeloma [21, 23, 24], andlung cancer [15, 66]. These positive fin-dings have either been dismissed as due tounknown causes or chance by the authorsor they have been ignored in revisions ofradiation protection standards [55].However, there is no reasonable justifica-tion for ignoring findings of positive asso-ciations of radiogenic risk with exposure onthe basis of their smaller number orbecause of disagreeing with inconclusive ornegative findings, unless specific substan-tial errors in the analysis can be shown.Mutually inconsistent epidemiological fin-dings are likely indicators of essential diffe-rences in sensitivity to detecting small doserelated excess mortalities at low exposureswhich depend critically on the choice ofcase and control populations, on the depen-dability of dose records over long periodsof time, as well as on adequate statisticalcontrols for a variety of selection effectsassociated with mortality rates [73].In evaluating the significance of a particu-lar health study, the uncertainties and am-biguities in epidemiological methods mustbe considered (see table III). For example,a recent published international study usinglarge-scale pooling of cancer mortalitiesfrom UK, U.S., and Canadian nuclear in-stallations by Cardis et a!. [ll] based on amethodology similar to that used before by

71

Gilbert eta!. [23, 24] finds a negative assi-ciation of dose with cancer mortality(except for leukemia). While presented as"the most precise direct radiogenic riskestimates" on the formal basis of its statisti-cal power, the critical reader will realizethat these data originate from widely diver-se work environments using non-uniformtechniques and methods for dose monito-ring and recording. Moreover, incompletecontrol for heterogeneous confounding va-riables across different worker populations,including the effect of age on susceptibility,can reduce significantly the sensitivity for atest of low-dose health effects [20, 79]. TheCardis et al. study is a prime example, illu-strating that statistically defined "highpower" per se does not protect an epide-miological study from an inconclusive orflawed result.

III. 2 Worker studies showing low-doseradiation effects.In contrast, two major U.S. studies didestablish statistically significant excesscancer mortalities at mean exposures farbelow allowable yearly exposures, bothamong Hanford (1944 -1986) [46] and OakRidge workers (1943 -1984) [32, 64, 96].Comparable results were found in a Britishstudy [6]. The risk values obtained fromthese studies are more than an order of ma-gnitude larger than the official values (seeTable II), flatly contradicting the claims ofinternational radiation commissions thatradiogenic risks perunit dose are lower forlow-dose exposure spread over longperiods of time (low dose rates) than equi-valent acute exposures. No wonder theabove findings were met by rejection andheated debates [64, 67, 83, 97].Meanwhile, the U.S. Department of Energy(DOE) in a new promotional publicationseems to have taken account of the abovefindings in its statement on radiation andhuman health. The DOE states: "In general,the risk of adverse health effects are higher


when exposure is spread over a long periodthan when the same dose is received at onetime" [90].

III. 3 Do mutually inconsistent epidemio-logical study results neutralize eachother?There is no reasonable justification forignoring "aberrant" findings unless specificsubstantial errors in the analysis can beshown. Mutually inconsistent epidemiolo-gical findings can often be explained by theinvestigators' choices of different criteriafor data selection, or by using divergentmethods of statistical controls for con-founding variables. Specific methodologi-cal decisions are likely to determine astudy's statistical sensitivity as to whetheror not the existence of a dose-related excesscancer mortality at low exposures can beestablished. Such choices include allowan-ces for individual variations in susceptibili-ty (e.g., due to age at exposure) and cancerlatencies, controlling for selection effectswithin different groups of the workingforceand other socio-economic confounders af-fecting baseline mortality rates [20, 47].For low-dose exposures, an equally impor-tant source of systematic bias, likely to re-duce a study's sensitivity, are ambiguitiesin recorded occupational doses at or justbelow detection limits of radiation monitorsover decades of employment and improve-ments in monitor technology [79,98].For discussions of other relevant occupa-tional radiations studies, including thosedealing with airline flight and medical x-ray personal, we can refer to pervious re-views [57, 58]. For these groups, elevatedcancer risks and chromsome aberrationshave been linked conclusively to low-doseradiadion exposure. Much debate continuesabout postulated genetic effects of paternalexposures, initiated by the findings of leu-kemia and lymphoma clusters amongyoung people near the Sellafield nuclearplant in West Cumbria, Great Britain.

72

Subsequent mutually inconsistent findingsfrom epidemiological studies aroundnuclear installations, or contrasting clinicalreports among populations affected by fall-out, highlight one of the most crucial openquestions regarding long term health con-sequences of continuing radioactive con-tamination of the biosphere. The authorsrecognize the serious problems in estima-ting internal doses, yet without consideringthe biologically more damaging exposuresfrom internally lodged radiQisotopes, com-pared to those from external sources, theissue cannot be resolved. Research in thisarea will be decisive in advancing ourknowledge.

III. 4 Higher risks per unit dose for me-dical x-rays, compared to risk estimatesfrom A-Bomb gamma raysThe biological effects of nuclear radiationin tissue depend in a complicated manneron the density of ionizations and chemicalbond breaking capacities of primary radia-tion and secondary electrons along theirpaths. These processes are determined bythe nature of the primary radiation and theybecome more concentrated at lower andlower energies. Alpha particles and neu-trons produce much more highly concentra-ted damage in tissue than high energyelectrons or photons. A thorough non-technical discussion of various biologicalinteractions of ionizing radiation with li-ving tissue can be found in [26]:(chapter19).A 1986 report by a joint task force fromtwo official international radiation com-missions presented radiobiological eviden-ce that at the same (relative low) dose, 250kVp medical x-rays are about twice asbiologically effective as high-energygamma rays [38]. A more recent publica-tion on the biological effectivenes of A-bomb neutrons also includes informationabout relative biological effectiveness(RBE) of x-rays versus gamma-rays. Using


the frequency of induced chromosomeaberrations in human blood lymphocytes invitro as the indicator, and comparing 250kVp x-rays with Co-60 gamma rays atvarying doses, the x-rays were about 2,7times as effective as Co-60 gamma [16].A-bomb gamma rays with considerablyhigher mean energies in the 3-6 Me V rangeare still less biologically effective than thelower energy Co-60 emission as recentlydemonstrated by Straume in a review sur-veying the relevant literatur [81] and shownin figure 2. This means that the radiologicalrisks per dose for exposures to 250 kVp x-rays and even softer x-rays in the case ofmammography (less than 30 kVp) at lowdoses are between 4 to 5 times higher thanA-bomb gamma rays. It is surprising thatthis warning has been omitted from thesummaries of known effects from low-doseexposures to soft x-rays in influential medi-cal publications.Most of the man made radiation exposureof general populations in industrializedcountries result from application of medicalx-rays [4, 39]. Thus, a medical exposurerisk value four to five times greater thanthat assumed by radiation protection com-missions and used as guidelines by radio-logists, call for revisions in standard patientrisk versus benefit analyses for radiologicalprocedures.

IV CONCLUSIONS AND DISCUSSIONA number of findings reviewed in the pre-vious sections are at variance with thesummaries of the "state of knowledge"(sec. I), which have been primarily basedon official interpretations of the A bombsurvivor follow up study (sec. II). Neitherthe fetal hypersensitivity to radiation [8, 9,25,31,43,44,48,49,99], nor an increasein susceptibility for cancer induction for anaging population [14, 22, 96] are part of theaccepted notions on radiation effects at lowdoses. Nor does this body of assumptionslink low dose exposures resulting from ra-

73

dioactive fallout (either from nuclear tes-ting or from reactor accidents) to any of theobserved congenital effects like infant mor-tality [53, 68, 69, 94] rare childhood can-cers [29] and low birth weight [27]. Whenlevels of fallout contamination over largeareas of the globe became known, local au-thorities everywhere, referring to the pro-nouncements by official national and inter-national radiation regulatory commissions,reassured the populations under their juris-diction that their levels of exposure wouldbe much too low to cause any adversehealth effects. In the light of the new evi-dence, sadly, these statements have nowlost their credibility.Also, on the basis of the foregoing summa-ries of studies, we draw the following con-clusions regarding the five issues selectedin sec. 1.2.1. as having been controversial:

A. Dose effect Relation at Very Low Do-sesWhile the A bomb survivor mortality data19501985 yield a non threshold linear doseeffect relation for cancers (other than leu-kemia) down to about 20 cGy with a sug-gestion of an increased excess relative riskin the lowest dose range, the most recentlypublished cancer incidence statistics 19501987 [17] show a statistically strong nonthreshold linear acute dose effect relationfor all solid tumors down to the 1 10 cSvorgan dose range with an excess relativerisk about 40 % larger than that derivedfrom the mortality data. Some of the epi-demiological studies of protracted occupa-tional exposures with life time accumulateddoses under 50 cSv and mean doses of theorder of natural background find excessrisks per unit dose for cancers substantiallyin excess of those predicted by linear extra-polation from the LSS mortality or the in-cidence data. This apparent discrepancy ininitial slope of the dose effect curve couldbe due to bias from selection effects [77,78], uncertainties in dose assignments in

Inconsistencies and Open Questions Regarding Low,Dose Health Effects ofIonizing Radiation

the LSS cohort, or the accumulated occu,pational doses [45, 83]. However, we liketo emphasize that the hypothesis of a uni,versal dose effect relation, which would re,quire consistency of risk over such widelydifferent population characteristics andconditions of radiation exposures, remainsunproven.

B. Presumed Reduced Biological Effecti-veness of Ionizing Radiation (DREF)The occupational exposure studies re,viewed in [57, 58], the pre natal X ray andexternal background exposure studies [31,49], as well as the studies related to air-borne radioactive emissions [53,69,94] areall inconsistent with the hypothesis of re-duced biological effectiveness of ionizingradiation at protracted irradiation (1.2.1B).

C. Enhanced Biological Effectiveness ofMedical X rays, Relative to High EnergyGamma RaysThis extremely important question in termsof its implications for public health hasonly been touched upon in the BEIR V r~-port by referring to a 1986 review by theInternational Commission on RadiationUnits and Measurement [38], but without indepth discussion. BEIR V [4] suggests,however, that the radiation risk estimates asderived from the acute gamma ray exposu-res of the Japanese survivors which formthe basis for all radiation protection guide-lines may underestimate these risks by afactor of two for medical, industrial orother low energy x ray exposures. In thethree reviews of the current state of know-ledge of radiation effects, cited in sec. 1.2.2,especially directed toward physicians, thistopic is not even listed among the openquestions, implying that the generally ac-cepted risk values (derived from the Abomb studies) are applicable to all medicalexposures as well. Yet, there are welldocumented findings [86, 87] of twice aslarge a mutational effect in Tradescantia for

74

250 kVp x rays compared to Cs 137 gammarays and factors between 2,7 and j for theinduction of chromosomal damage arefound when comparing soft x-rays with A-bomb gamma-rays. [16, 82]. There is aphysical basis for expecting such a diffe-rence in biological effectiveness [26]. Thesignificance of these radio biological fin-dings for human exposures is an unsettledquestion with broad ramifications for ra-diation protection.

D. Free Radicals, Low Dose Exposuresand HealthExcept for mentioning the possible creationof free radicals by ionizing radiation in theBEIR V report (sec.1.2.1 D) and by one ofthe reviews cited (sec. 1.2.2 D), the possi-bility that this interaction could provide astrongly non linear alternative biologicalmechanism [76] to the well known directmutational interactions of radiation withhuman cell nuclei in the induction of disea-se in particular, at very low doses has notbecome part of the discussions of low doseradiation effects, in spite of a burgeoningliterature linking free radicals to a widespectrum of diseases, as well as suggestingpossible treatments [30].

E. The Radiation Hormesis HypothesisAll of the low dose studies of radiation ef-fects in human populations reviewed aboveare inconsistent with hypothesized longterm cancer reducing effects of such expo-sures in excess of unavoidable naturalbackground of human populations (horme~sis) (sec. IV.A.2). One can only speculateabout the continued "popularity" of thisconjecture among some groups of radiationexperts.

Suggestions for New ResearchBy comparing statements about the abovelisted five aspects in different authoritativepresentations of "known" health effects oflow dose exposures, and by focusing on in-


I·

consistencies or selective omISSIOns, wehave identified unsettled questions in themainstream "state of knowledge". Howe-ver, the identification of unsettled questionscan be extended by reviewing findingsfrom a number of unrefuted studies on po-pulations other than the LSS cohort of Abomb survivors, that are inconsistent withtraditional notions and, therefore, havebeen rejected, ignored or glossed over inpurportedly comprehensive reviews of thefield. These inconsistencies raise a rangeof additional questions about the limitationsof currently accepted concepts.Finally, in the aftermath of the widespreadfallout from the explosion of the Chernobylreactor in the former Soviet Union, thereare suspected associations of disease withradiation exposures that have barely beenreported in the scientific literature. An ad-ditional relevant summary of observedhealth effects as a consequence of theChernobyl nuclear explosion, presented atan International Workshop on the Impact ofthe Environment on Reproductive Health(30 September - 4 October 1991), Copen-hagen, Denmark) can be found in [51].While an international team of radiationexperts invited by the Soviet governmentand financed by the IAEA confirmed an in-creased rate of a variety of health problems,but dismissed any possible association withradiation exposure [36, 65]. In the meantime ten years after the accident almost1000 thyroid cancers in children exposed in1986 have been confirmed in the heavycontaminated areas as reported recently atan International Congress in Berlin [40].Many severe health problems other thancancer are seen in the cohorte of the liqui-dators and in the normal population [58].Very recently an increase in germlinemutation at human minisatellite loci hasbeen reported and found to have a positivecorrelation with levels of radioactive con-tamination [18]. High levels of geneticchanges in rodents living near the destroyed

75

nuclear reactor have been observed. Thebase pair substitution rates for mitochon-drial cytochrome b gene are hundreds oftimes greater than those typically found inmitochondria of vertebrates [1]. These fin-dings are not in accord with the state ofknowledge as documented in authoritativereports [33, 52, 91] (8,9,10). In the UnitedStates, only a handful of government fun-ded health studies have been initiatedamong populations ("downwinders") thathave been at risk for internal exposures byvarious pathways as a result of radioactivereleases into the environment from wea-pons production and testing facilities, insome instances possibly in synergism withchemical exposures. These populations atrisk include large groups of civilians andtens of thousands of military personnel,who had been stationed at nuclear sites orwho were involved with nuclear bombtests. Some official epidemiological studieson these populations were admittedly"defensive" in nature [75] (0), respondingto pressures by affected populations for ma-terial compensation. On the other hand, anincreasing number of well researched inve-stigative reports and small scale health sur-veys, organized by members of the affectedpopulations themselves [7, 10, 13, 19, 41,93] document the existence of clusters ofcancers and similar patterns of other serioushealth problems among down winders nearvarious nuclear sites. An .increasing body ofverifiable observations, not matched byreasonable alternative explanations byscientific bodies, presents a challenge topublic health agencies to commence largescale unified health surveys and to radiationexperts to extend their research strategiesinto insufficiently investigated interactionsof radiation with human health. There is anurgent need for the formulation of novelguiding questions that need to be translatedinto testable hypotheses.


Table I1950 - 1985 Radiogenic cancer risk and projected lifetime excess risks per104 person-cGy.a

Subcohort Dose range Dose groups 1950-1985 . Estimated life-Dosimetry [cGy] used in Excess risk per time riskc per

analvsisb 104 u-cGv 104 u-cGv

All cancers exept leukemia

Colon dose 0-49 0,-5,-9,-19,-49 5.0±1.5 18.1±4.90-19 0-5,-19 9.1+1.4 33+56-99 6-19,-49,-99 2.8±0.3 9.3±1.1

a Table I adopted from refs. [50,56].b dose ranges in adjacent cSv intervals: -5=1-5; "9=6-9; -19=10-19, etc., except for the dosegroups a and -5 combined, indicated by 0-5.c a detailed discussion of this estimation is given in refs. 28, 29. The errors shown are standarderrors.

Table IIIChoises and variables to be considered that affect the sensitivity ofepidemiological studies to find health affects os low-dose radiationex osure in the resence of confoundin factors• data selection (exclusions)• heterogeneities in health profiles (selection effects)• recognition of significant controlling variables• stratification of variables• variations of susceptibility with age at exposure• variations in latency periods• socio-economic factors affecting base-line mortality or morbidity• ambiguities in assigning exposure levels• distin uishin between external and internal ex osure

76


Table IISelected radiogenic cancer risk estimates for exposures at low doses,acute or protracted

Exposure conditions Excess cancer riskper 104 p-cGyt

Reference Dose Dose Rate Applicable Applicable Observed LifetimeRange Population Follow-up Risk (estim.)cGy

Nussbaum 1 - 11 Acute Bomb - 90.000 A- 1950-1985 9.1 33-Kohnlein Gamma bomb[56] survivorssame 11 - 69 Acute Bomb same 1950-1985 2.8 9.3

GammaGofman 0-5 Acute Bomb same 1950-1985 5 30[26] Gammasame 0-5 Acute Bomb recalculated 1950-1985 - 26

Gamma for U.S.population

Gilman, 440 for exposures[46] ground) Hanford after age 58

(WA)Wing et al. 1.7 mean low rate - 8.000 1943-1984 Working Life Risk[96] < 5 for nuclear deaths - 110 average for all

68 % workers workers, all agesOak Ridge(TN)

Beral et al. 0.8 mean low rate - 23.000 1951-1982 Working Life Risk[5] British -19 Y - 165 average for all

nuclear mean workers, all agesworkers follow-up

77

ref. [88]ref. [4]ref. [39]


t e.g. for a lifetime excess cancer risk of 30 per 104 person-cGy: exposing 15,000 people to anaverage accumulated dose of 10 cGy (100 mGy) will on average lead to [(30 cancers)/l04 p-cGy](1.5xI04xlO p-cGy) = 450 extra radiogenic cancer deaths over the lifetime of these 15.000people.

# this means that exposed children have a l2-fold risk for developing breast cancer as adults thanunexposed controls.

Table II continued

Range of estimated lifetime risk values for protracted low-dose exposures of normalo ulations b international radiation commissions$

UNSCEAR (1988) 2 - 5 fatal cancers per 104 p-cGyBEIR V (1990) 4 fatal cancers per 104 p-cGyICRP (1990) 5 fatal cancers er 104 -cG

$ including recommended Dose Rate Effectiveness Factore of two, not supported by human studies[26,58]

Fig. 2Biological effectiveness of low-LET radiationThe data are the low-dose linear slopes of the linear-quadratic dose response curves forchromosome dicentrics induced in vitro in human lymphocytes exposed to the radiationindicated and evaluated at the first division [82].

10-'

250-kVp

>.ro Tritiuml-(!) betasl-ID 15-MeVc... electronsID() 10-2l-IDc...(/)()'C H & N gammas.•....cID.~0

10-3

10-3 10-2 10-' 10° 10' 102

Mean energy (MeV)

78


Fig. 11950 - 1985 LSS mortality from all cancers except leukemiaCumulative mortalitiy per 104 survivors for the lowest six DS 86 colon dose sub cohorts 0,1-5, 6-9, 10-19, 20-49, 50-99 cGy [triangles] and for the two combined 0-5 (mean dose0.7) and 6-19 (mean dose 10.9) cGy sub cohorts [squares] versus mean colon dose (cGy).Standard error bars are shown. The increase in mortalities between the 0-5 and 6-19 cGysub cohorts is statistically significant (p < 0.01). The solid line is an error weighted linearfit to the five [triangle] data points below 40 cGy mean dose (line 1, Table I). The twodashed l~nes are weighted linear fits to (1) the two data points for the combined 0-5 and 6-19 cGy dose groups [two squares] (line 2, Table I) and (2) the three data points for thedosegroups 6-19, 20-49, and 50-99 cGy with mean doses above 10 cGy [one square andtwo triangles] (line 3, Table I), respectively. The slopes of the three lines correspond to thethree values of excess risk per 104 person cGy listed in Table I. Data from ref. 71.

Cancer mortality except leukemia from theRERF 1950 - 1985 follow up statistics

CJ)"- 10000>.~ ---------::J ---CJ) 900 .-- 1------'¢ ---0 ---~"-Q.) 800Q.:>.:::cot 7000 ,~

600o 10 20 30 40 50 60 70

Colon Dose (cGy)References1. Baker RJ, Van Den Bussche RA, WrightAJ, Wiggins LE, Hamilton MJ, Reat EP, SmithMH, Lomakin MD, and Chesser RH. Highlevels of genetic changes in rodents ofChernobyl. Nature 380: 707-708 (1996).2. Barendsen GW. Do fast neutrons at lowdose rate enhance cell transformation in vitro?A basic problem in microdosimetry and inter-pretation. lnt J Radiat BioI 47:731 744 (1985).

79

3. BEIR III. National Research Council.Biological effects of ionizing radiation. The ef.fects on populations of exposure to low levelsof ionizing radiation. Washington, DC: NationalAcademy Press, 1980.4. BEIR V. National Research Council.Health effects of exposure to low levels of io-nizing radiation. Washington, DC: NationalAcademy Press 1990.


5. Beral V, Inskip H, Fraser P, Booth M,Coleman D, Rose G. Mortality of employees ofthe United Kingdom atomic atomic energy au-thority, 1946 1979. Brit Med J 291: 440447(1985).6. Beral V, Fraser P, Carpenter L, et al. Mor-tality of employees of the atomic weaponsestablishment 1951-1982. Brit. Med. J 1988;297: 757-770.7. Bertell R. No immediate danger. Toronto,Canada: The Women's Educational Press, 1985.8. Bithell JF, Stiller CA. A new calculationof the radiogenic risk of obstetric X raying. StatMedicine 7: 857864 (1988).9. Bross IDJ, Natarajan N. Cumulative ge-netic damage in children exposed to preconcep-tion and intrauterine radiation. Investig Radiol15(1):5267 (1980).10. Brown P. When the public knows better:Popular epidemiology challenges the system.Environment 35(8):1641 (1993).11. Cardis E, Gilbert ES, Carpender L, et al.Direct estimates of cancer mortality due to lowdoses of ionising radiation; an internationalstudy. Lancet 1994; 344: 1039-1402.12. Carter RL. Low Dose leukemogenic ef-fects of A bomb irradiation. RERF Update 4(1):9 10 (1992) and RERF technical report (TR 991). Hiroshima : Radiation Effects ResearchFoundation, .1993.13. Caufield C. Multiple exposures. Chicago:The University of Chicago Press, 1990.14. Checkoway H, Mathew RM, Shy CM,Watson JE, Tankersley WG. Radiation, workexperience, and cause specific mortality amongworkers at an energy research laboratory. Brit JInd Med 42: 525 533 (1985).15. Checkoway H, Pearce N, Crawford BraunDJ, Cragle DL. Radiation doses and cause spe-cific mortality among workers at a nuclear ma-terial fabrication plant. Am J Epidemiol127:255266 (1988).16. Dobson RL, Straume T, Carrano AV,Minkler JL, Deaven LL, Littlefield LG; AwaAA. Biological effectiveness of neutrons fromHiroshima bomb replica: Results of a collabo-rative cytogenetic study. Radiat Res 1991; 128:143-149.17. Dohy H, Ikeda T, Kamada N,KuramatoR,Kusumi S, Mabuchi K,Matsui T, Nonaka H,Ochikubo S, Preston DL, Ron E, Soda M, Su-

80

gimoto S,Terasaki M, Thompson DE, TokunagaM, Tomonaga M. Cancer incidence in atomicbomb survivors. Rad Res 137(2 supp):SI SI12(1994).18. Dubrova YE, Nesterov VN, KrouchinskyNG, Ostapenko VA, Neumann R, Neil DL,Jeffreys AJ. Human minisatellite mutation rateafter the Chernobyl accident. Nature 380: 683-686 (1996).19. Gallagher C. American ground zero.Cambridge: MIT Press, 1993.20. Geiger HJ, Rush D, Michaels D, BakerDB, Cobb J, Fischer E, Goldstein A, Kahn HS,Kirsch JL, Landrigan, PJ, Mauss E, McLeanDE. Dead reckoning: A critical review of thedepartment of energy's epidemiological re-search. Washington DC: Physicians for SocialResponsibility 1992.21. Gilbert ES, Petersen GR, Buchanan JA.Mortality of workers at the Hanford site: 19451981. Health Phys 56: 11 25 (1989).22. Gilbert ES, Fry SA, Wiggs LD, Voelz GL,Cragle DL, Peterson GR. Methods for analyzingcombined data from studies of workers exposed .to low doses of radiation. Am J Epidemiol 131:917 927 (1990).23. Gilbert ES, Omohundro E, Buchanan JA,Holter NA. Mortality of workers at the Hanfordsite: 1945 1986. Health Phys 64:577 590(1993).24. Gilbert ES, Cragle DL, Wiggs LD. Upda-ted analyses of combined mortality data forworkers at the Hanford site, Oak Ridge NationalLaboratory, and Rocky Flats weapons plant.RadiatRes 136(3): 408 421(1993).25. Gilman EA, Kneale GW, Knox EG, Ste-wart AM. Pregnancy X rays and childhood can-cers: effects of exposure age and radiation dose.J Radiol Prot (GB) 8:3 8 (1988).26. Gofman JW. Radiation induced cancerfrom low dose exposure: An independent ana-lysis. San Francisco: Committee for NuclearResponsibility, Inc. (POB 421993, San Fran-cisco CA, 94142),1990.27. Gould JM, Sternglass EJ. Nuclear fallout,low birthweightt, and immune deficiency. IntJHealth Serv 24: (1994), in press.28. Greenberg M. The evolution of attitudesto the human hazards of ionizing radiation andits investigators. Am J Ind Med 20: 717 721(1991).


29. Haaf G, Kaatsch G, Kreis J, Michaelis J,Berthold F. Fall Kontroll Studie zum Anstiegder Neuroblastom Inzidenz fUr im Jahr 1988geborene Kinder. Europaische Perspektiven dermedizinischen Informatik, Biometrie und Epi-demiologie (Michaelis J, ed). Munchen: MMVVerlag, 1993; 28 32.30. Halliwell B, Gutteridge JMC. Free radi-cals in biology and medicine. New York:OxfordUniversity Press, 1989.31. Hatch M, Susser M. Background gammaradiation and childhood cancer within ten milesof a US nuclear plant. Int J Epid 19: 546 552(1990).32. Hendee WR, Howe GR, Brown RA,Greenspan BS, Marshall BJ, Maienschein FC,Peelle RW, Wing S, et al. Editorials and corre-spondence. JAMA 1991; 265: 1437-1439; Jama1991; 266: 653-654; JAMA 1991; 267: 929-230.33. Hendee R. Estimation of radiation risks:BEIR V and its significance for medicine.JAMA 268 (5):6204 (1992).34. Hoel, DG, Galas DJ, Abrahamson S.Commentary: Unanswered questions after BEIRV. Radiat Res 136: 137 145 (1993).35. Inskip H, Beral V, Fraser P, Booth M,Coleman D, Brown A. Further assessment ofthe effects of occupational radiation exposure inthe United Kingdom Atomic Energy Authoritymortality study. Brit J Ind Med 44:149 160(1987).36. International Commission on RadiologicalProtection. Publication 60: Recommendationsof the ICRP. Oxford: Pergamon, 1991.37. International Physicians for the Preven-tion of Nuclear War. Radioactive heaven andearth. New York: The Apex Press, 1991.38. International Kongress: Tschernobyl - 10Jahre danach: Eine aktuelle Bilanz der Folgen.Book of Abstracts. April 1996, Berlin.39. International Commission on RadiologicalProtection. Publication 26: Recommendation ofthe ICRP, Oxford:Pergamon, 1977.40. International Commission on RadiationUnits and Measurements. The quality factor inradiation protection. ICRU Report 40. Washing-ton, DC: Joint task force of the ICRP and theICRU, 1986.

81

41. International Atomic Energy Agency. Theinternational Chernobyl project: An overview.Vienna 1991.42. Kerr GD. Quality factors. Health Phys 55:241 249 (1988).43. Kneale GW, Stewart AM. Pre natal Xrays and cancer: Further tests of OSCC data.Health Phys 51:369 376 (1986).44. Kneale GW, Stewart AM. Pre natal Xrays and cancer: Further tests of OSCC data.Health Phys :200 (1987).45. Kneale GW, Sorahan T, Stewart AM.Evidence of biased recording of radiation dosesof Hanford workers. Am J Ind Med 20:799 803(1991).46. Kneale GW,Stewart AM. Reanalysis ofHanford data: 1944 1986 deaths. Am J Ind Med23:371 389 (1993).47. Kneale GW, Stewart AM. Factors af-fecting recognition of cancer risks of nuclearworkers. Occup Environ Med 19.95; 52: 515-523.48. Knox EG, Stewart AM, Kneale GW, Gil-man EA. Prenatal irradiation and childhoodcancer. J Soc Radiol Prot (GB) 1987; 7: 3 15.49. Knox EG, Stewart AM, Gilman EA,Kneale GW. Background radiation and child-hood cancer. J Soc Radiol Prot(GB) 8: 9 18(1988) .50. Kohnlein W, Nussbaum RH. Reassess-ment of radiogenic cancer risk and mutagenesisat low doses of ionizing radiation. Adv Muta-gen Res 3:53 80(1991).51. Kulakov VI, Sokur TN, Volobuev AI,Tzibulskaya IS. Malisheva VA, Zikin BI, EzovaLC, Belyaeva VA, Bonartzev PD, SperanskayaNV, Tchesnokova JM, Marveeva NK, KaliznukES, Miusrova LB, Orlova NS Female repro-ductive funktion in areas affected by radiationafter the Chernobyl power station accident. En-viron Health Perspect. Suppl 101 (2): 117-123(1993).52. Little JB . Biologic effects of low levelradiation exposure. In Radiology: Diagnosis,imaging, intervention, Vol. 6 (Taveras JM, Fer-rucci JT, eds.) Philadelphia: JB Lippincott,1993; Chapter 13: 1 15.53. Luning G, Schmidt M, Scheer J, Ziggel H.Early infant mortality in West Germany beforeand after Chernobyl. Lancet 8671: 1081


1083(1989); and Lancet 8682, Ian 20.: 161 162(1990).54. Modan B, Alfandry E, Cherit A, Katz L.Increased risk of breast cancer after lowdose ir-radiation. Lancet Nr. 8639 V01.1: 629-631(1989).55. Morgan KZ. Changes in international ra-diation protection standards. Am 1 Ind Med 25:301 307 (1994).56. Nussbaum RH, Kohnlein W, Belsey RE.Die neueste Krebsstatistik der HiroshimaNagasaki Uberlebenden: Erhohtes Strahlenrisi-ko bei Dosen unterhalb 50 cGy (rad), Konse-quenzen fUr den Strahlenschutz. Med Klin86:99 108(1991).57. Nussbaum RH, Kohnlein W. Inconsisten-cies and open questions regarding low-dosehealth effects of ionizing radiation. Environ-mental Health Perspectives 1994; 102: 656-667.58. Nussbaum RH, Kohnlein W. Health con-sequences of Exposures to ionizing radiationfrom external and internal sources: Challengesto radiation protection standards and biomedicalresearch. Medicine and Global Survival 1995;2: 198-213.59. Okajima S, Fujita S, Harley JH. Radiationdoses from residual radioactivity. In: Atomicbomb radiation dosimetry reassessment inHiroshima and Nagasaki, vol 1 (Roesch WC,ed). Hiroshima: Radiation Effects ResearchFoundation, 1987:205 226.60. Pierce DA, Vaeth M. Cancer risk estima-tes from the A bomb survivors: Extrapolation tolow doses, use of relative risk models and otheruncertainties. RERF Commentary and Review(CR 2 89), Hiroshima: Radiation Effects Re-search Foundation, 1989.61. Pierce DA, Vaeth M. The shape of thecancer mortality dose response curve for the Abomb survivors. Radiat Res 126: 3642 (1991).62. Preston DL, Pierce DA . The effect ofchanges in dosimetry on cancer mortality riskestimates in the atomic bomb survivors. RadiatRes 114:43766 (1988).63. Preston DL, Pierce 0, Vaeth M. Neutronsand radiation risk: A Commentary. RERF Up-date 4(4):5 (1992).64. Prichard HW, Gilbert ES, Wing S, et al.Editorial and correspondence. Health Phys1992; 62: 260-264.

82

65. RERF. IAES's International Chernobylproject outcome corroborates earlier findings.RERFUpdate 1991; 3: 1-2.66. Rinsky RA, Melius 1M, Hornung RW.Case control study of lung cancer in civilianem-ployees at the Portsmouth Navel Shipyard,Kittery, Maine. Am 1 Epidemiol 127:55 64(1988).67. Rojas Burke 1. Oak Ridge cancer findingshotly debated. 1 Nuclear Med 32: IIN 26N(1991).68. Scheer 1. Gould and Goldman: DeadlyDeceit A defense. Health Phys 61: 279 280(1991).69. Scheer 1. Early infant mortality in WestGermany Before and After Chernobyl. Brit MedJ 304: 843 (1992).70. Schull WI, Nishitani H, Hasuo K, Ko-bayashi T, Goto I, Otake M . Brain abnormali-ties among the mentally retarded prenatally ex-posed atomic bomb survivors. RERF TechnicalReport (TR 13 91). Hiroshima: Radiation Ef-fects Research Foundation, 199 I and RERFUpdate 3(4): 34(1991).71. Shimizu Y, Kato H, Schull WI. Cancermortality in the years 1950 1985 based on therecently revised doses (OS 86). Life Span StudyReport 1I, Part 2 (RERF TR 5 88). Hiroshima:Radiation Effects Research Foundation, 1988and Radiat Res 121: 120141 (1990).72. Shimizu Y, Kato H, Schull WJ, MabuchiK. Dose response analysis of atomic bomb sur-vivors exposed to low level radiation. RERFUpdate 4(3):23 (1992).73. Shleien B, Ruttenber AI, Sage M. Epi-demiologic Studies of cancer in populationsnear nuclear facilities. Health Phys 61:699 713(1991).74. Smith Rl. Study of atomic veterans fuelscontroversy. Science 221: 733 734 (1983).75. Smith PG, Douglas AI. Mortality of wor-kers at the Sellafield plant of British NuclearFuels. Sr Med 1293:845854 (1986).76. Sternglass EJ, Gould JM. Breast cancer:Evidence for a relation to fission products in thediet. Int 1 Health Serv 23: 783 804 (1993).77. Stewart AM, Kneale GW. A bomb radia-tion and evidence of late effects other than can-cer. Health Phys 58:729 735(1990).


78. Stewart AM, Kneale GW. A bomb survi-vors: Further evidence of late effects of earlydeaths. Health Phys 64:467 472 (1993).79. Stewart AM, Kneale GW. The Hanforddata: Issues of age at exposure and dose recor-ding. PSR Quarterly 1993; 3: 101-111.80. Straume T. Neutron discrepancies in thedosimetry system 1986 have implications forradiation risk estimates. RERF Update 4(4): 3 4( 1992).81. Straume T, Egbert SD, Woolson WA,Finkel RC, Kubik PW, Gove HE, Sharma P,Hoshi M. Neutron discrepancies in the DS86Hiroshima dosimetry system. Health Phys63:421426(1992);82. Straume T. High-energy gamma rays inHiroshima and Nagasaki: Implications for riskand Wr. Health Phys 1995; 69: 954-956.83. Strom DJ. A critique of "Mortality amongworkers at Oak Ridge National Laboratory".Nucl News (July): 6774 (1991).84. Tankersley W, West C, Reagan JL,Watson JE. Assessment ofradiation exposure ator below the minimum detectable level in epi-demiologic studies. App Occup Environ Hygie-ne:(1994), in press.85. Tomonaga M, Matsuo T, Carter RL. Abomb irradiation and leukemia types: An up-date. RERFUpdate 3(4):5 6 (1991).86. Underbrink AG, Keller AM, Mills RE,Sparrow AH. Comparison of x ray and gammaray dose response curves for pink somaticmutations in Tradescantia clone 02. Radiat En-viron Biophys 13:295 303 (1976).87. Underbrink AG, Edwards FM, LowerWR, Yanders AF, Ranney TK. Absence of de-tectable fractionation effects in Tradescantiasomatic mutations for an x ray dose of 5 rad.Rad Res 101:170 176 (1985).88. United Nations Scientific Committee onthe Effects of Atomic Radiation. Sources andeffects of ionizing radiation. New York: UnitedNations, ] 977.89. United Nations Scientific Committee onthe Effects of Atomic Radiation. Sources, ef-fects and risks of ionizing radiation. New York:United Nations, ] 988.90. U.S. Department of Energy. Closing thecircle on the splitting of the atom. Washington,DC: Office of the Environmental Management,USDOE. January 1995.

83

91. Upton AC, Shore RE, Harley NH. Thehealth effects of low level ionizing radiation.Annu Rev Publ Health ]3: ]27 150 (1992).92. Vaeth M, Preston D, Mabuchi K. Extrapo-lating life span study cancer risk estimates tolow dose radiation exposures. RERF Update4(3):5 6 (1992).93. Wasserman H, Solomon N. Killing ourown. New York: Dell Publishing Co. 1982.94. Whyte RK. First day neonatal mortalitysince 1935: Re examination of the Cross hypo-thesis. Brit Med J 304: 343 346 (1992).95. Wilkinson GS, Dreyer NA. Leukemiaamong nuclear workers with protracted expo-sure to low dose ionizing radiation. Epidemio-logy 2:305 309 (1991).96. Wing S, Shy CM, Wood JL, Wolf S,Cragle DL, Frome EL. Correspondence. JAMA266: 652654 (1991); JAMA 267:929930(1991)and Health Phys 62:260 264(1992).97. Wing S, Shy CM, Wood JL, Wolf S,Cragle DL, Frome EL. Mortality among wor-kers at Oak Ridge National Laboratory: Eviden-ce of radiation effects in follow up through1984. JAMA 265:13971402 (1991).98. Wing S, West CM, Wood JL, et al. Re-cording of .extemal radiation exposure at OakRidge national laboratory: Implications for epi-demiological studies. J Expos Anal EnvironEpidem 1994; 4: 83-93.99. Yoshimoto Y, Kato H, Schull WJ. Risk ofcancer among children exposed in utero to Abomb radiation,1950 84. The Lancet ii:665 669(1988).

page1titlesInconsistencies and Open Questions Regarding Low-Dose Health Effects of Inconsistencies and Open Questions Regarding Low-Dose Wolfgang Kohnlein, Munster, Rudi H. Nussbaum, Portland

page2titlesInconsistencies and Open Questions Regarding Low-Dose Health Effects of

page3titlesInconsistencies and Open Questions Regarding Low-Dose Health Effects of Ionizing Radiation -


page5titlesInconsistencies and Open Questions Regarding Low-Do'se Health Effects of

tablestable1

page6titlesInconsistencies and Open" Questions Regarding Low-Dose Health Effects of






page12titlesInconsistencies and Open Questions Regarding Low,Dose Health Effects of

page13titlesInconsistencies and Open Questions Regarding Low-Dose Health Effects of I·


tablestable1

page15titlesInconsistencies and Open Questions Regarding Low-Dose Health Effects of Table II Selected radiogenic cancer risk estimates for exposures at low doses,

tablestable1

page16titlesInconsistencies and Open Questions Regarding Low-Dose Health Effects of Table II continued Range of estimated lifetime risk values for protracted low-dose exposures of normal o ulations b international radiation commissions$ Fig. 2 Biological effectiveness of low-LET radiation

imagesimage1image2

tablestable1

page17titlesInconsistencies and Open Questions Regarding Low-Dose Health Effects of Fig. 1 1950 - 1985 LSS mortality from all cancers except leukemia Cancer mortality except leukemia from the o Colon Dose (cGy) References

imagesimage1

tablestable1

page18titlesInconsistencies and Open Questions Regarding Low-Dose Health Effects of 80




Date post:	16-Feb-2021
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

Inconsistencies and Open Questions Regarding Low-Dose Health Effects of Ionizing Radiation · 2008....

Documents