Is cancer risk of radiation workers larger than expected?

Transcription

1 See Editorial, p Hemholtz Zentrum Münhen, Institute of Radiation Protetion, Neuherberg, Germany; 2 Federal Offie for Radiation Protetion, Department of Radiation Protetion and Health, Obershleißheim, Germany; 3 Johannes Gutenberg University Mainz, Institute of Medial Biostatistis, Epidemiology and Informatis, Mainz, Germany Correspondene to: P Jaob, Hemholtz Zentrum Münhen, Institute of Radiation Protetion, D Neuherberg, Germany; Jaob@helmholtz-muenhen.de Aepted 13 May 2009 Published Online First 30 June 2009 This paper is freely available online under the BMJ Journals unloked sheme, see oem.bmj.om/info/unloked.dtl Is aner risk of radiation workers larger than expeted? P Jaob, 1 WRühm, 1 L Walsh, 2 M Blettner, 3 G Hammer, 3 H Zeeb 3 ABSTRACT Oupational exposures to ionising radiation mainly our at low-dose rates and may aumulate effetive doses of up to several hundred milligray. The objetive of the present study is to evaluate the evidene of aner risks from suh low-dose-rate, moderate-dose (LDRMD) exposures. Our literature searh for primary epidemiologial studies on aner inidene and mortality risks from LDRMD exposures inluded publiations from 2002 to 2007, and an update of the UK National Registry for Radiation Workers study. For eah (LDRMD) study we alulated the risk for the same types of aner among the atomi bomb survivors with the same gender proportion and mathed quantities for dose, mean age attained and mean age at exposure. A ombined estimator of the ratio of the exess relative risk per dose from the LDRMD study to the orresponding value for the atomi bomb survivors was 1.21 (90% CI 0.51 to 1.90). The present analysis does not onfirm that the aner risk per dose for LDRMD exposures is lower than for the atomi bomb survivors. This result hallenges the aner risk values urrently assumed for oupational exposures. Oupational and medial diagnosti exposures to ionising radiation are mainly due to Roentgen rays and gamma rays, whih belong to so-alled lowlinear energy transfer (LET) radiation. The exposures may aumulate over a lifetime to doses of the order of 100 mgy. For example, in the 15- ountries ollaborative study on radiation workers in the nulear industry, about 10% of the study members reeived external doses exeeding 50 mgy, while only 0.1% reeived doses exeeding 500 mgy. 1 Exposures with doses in the range of mgy are onsidered here to be moderate in omparison with the high-dose groups of the atomi bomb survivors from Hiroshima and Nagasaki. Within an hour, whih is the timesale for ellular repair proesses, doses from oupational and medial diagnosti exposures do not generally exeed the order of 10 mgy. Thus, these exposures our at low-dose rate. It follows that estimates of health risks, in partiular of aner risks, related low-dose-rate, moderate-dose (LDRMD) exposures are of entral importane for pratial radiation protetion. Current estimates of aner risks from LDRMD exposures are mainly based on risk oeffiients derived from the Japanese atomi bomb survivors, that is, from persons with aute, high-dose exposures, whih are then ombined with a dose and dose-rate effetiveness fator (DDREF). 2 3 Values for DDREF have mainly been dedued from What this paper adds Review Oupational exposures to ionising radiation our normally at low-dose rate and may sum up to moderate doses in the order of 100 mgy. Limits of oupational exposures are based on the assumption that aner risk fators are lower than for the atomi bomb survivors by a fator of two. Twelve reent epidemiologial studies on aner after low-dose-rate, moderate-dose exposures were inluded in this analysis of aner risks related to suh exposures. The studies provide evidene that aner risk fators for oupational exposures are not lower than for atomi bomb survivors. The new evidene for aner risks should be taken into aount in optimisation proedures for the use of radionulides and ionising radiation at the work plae and in mediine. experiments with laboratory animals and from radiobiologial measurements. Speifially, the International Commission on Radiologial Protetion (ICRP) derived estimates of the exess aner risk after low-dose exposures and after exposures with higher doses but low-dose rates by reduing the orresponding risk value for the atomi bomb survivors by a DDREF of The BEIR VII Committee of the US National Researh Counil used a DDREF of During the past few years, a number of epidemiologial studies have been published, whih provide major information on aner risk after LDRMD exposures. The statistial power of eah of these studies is not strong beause of the relatively low risks of the doses involved. Therefore, the present study fouses on studies of larger groups of aners. More speifially, studies of all aner, all aner exluding leukaemia, all solid aner and all solid aner exluding bone aner have been inluded. In the present paper, values of the exess relative risk (ERR) per dose in LDRMD studies of aner risks from exposures to low-let radiation are ompared with those alulated for the atomi bomb survivors for the same grouping of aner types, gender distribution, average age at exposure, average age attained and dose quantity. A ombined estimator of the resulting risk ratios is alulated. Based on this risk estimator, aner lifetime risks are assessed. Oup Environ Med 2009;66: doi: /oem

2 In some of the LDRMD studies, ERR-per-dose distributions inlude a value of zero, whih would orrespond to an infinite value of the DDREF. In order to avoid resulting instabilities of the alulations, the inverse DDREF value, Q, that is, the ratio of the ERR-per-dose value in the LDRMD study to that for the atomi bomb survivors, is alulated here. METHODS Literature searh A systemati literature searh for primary epidemiologial studies was onduted in the PubMed database in January 2008, overing the period January 2002 to Deember The searh terms radiation and aner were ombined with alternatives of the terms oupation, work, personnel, and environmental or emergeny. A number of exlusion terms were speified to limit the findings to ionising radiation effets in the oupational, environmental or emergeny setting. An initial seletion of 714 papers was identified. The PubMed searh was augmented by a manual searh for referenes, by whih a paper on Chernobyl emergeny and lean-up workers 4 and a paper on Oak Ridge National Laboratory (ORNL) workers 5 were identified. Further, stimulated by a suggestion of a reviewer, a reent study on the UK National Registry for Radiation Workers 6 was inluded in the analysis, beause of its outstanding importane. Results without inlusion of this study are also reported below. The initial seletion was then restrited to ohort and aseontrol studies and epidemiologial reviews, whih left 123 papers. Further eliminations were made of studies on exposures to alpha radiation (beause most of the oupational exposures are due to external radiation), foused on hildren or individual aner sites, or without dosimetry. Further, nine publiations were not inluded in our analysis mainly beause relative risk estimates and their standard deviations ould not be derived, 7 11 beause there were many ohort members with high exposures, beause no data of the Life Span Study (LSS) were available for the orresponding group of aners among the atomi bomb survivors 14 or the required information on age at exposure and age at risk were not obtained. 15 If a study ontained results for different aner outomes, then the outome losest to solid aner was hosen. Espeially, inlusion of leukaemia was avoided as far as possible beause of differenes in height of risk and in shape of dose response, if ompared with solid aner. Conerning the 15-ountries ollaborative study of aner risk among radiation workers in the nulear industry, 16 the present analysis inludes only results, whih are not based on the Canadian data, beause problems with the appliation of the Canadian data within the 15-ountries study have been reported (Norman Gentner, personal ommuniation, 2008). ERR per dose for atomi bomb survivors The publily available atomi bomb survivor datasets for aner mortality from 1950 to 2000 (DS02an.dat) and aner inidene from 1958 to 1998 (lssini07.sv) from the Radiation Effets Researh Foundation ( were used to alulate ERR-per-dose values for aute exposures. Only survivors with doses below 4 Gy of shielded kerma were used in the risk analysis. The atomi bomb survivor data for the aner ategories used in an LDRMD study, i were fitted with a model inluding an expliit ERR-per-dose parameter, b lssi, a male fration, f i, an age at exposure, e i and an age-attained, a i : l(d i, s, e, a) =l 0 (s, e, a) [1+ b lssi d i r i (s, e, a)] (1) with r i (s, e, a) =h i (s) exp[a i (e 2 e i ) + v i ln (a/a i )] (2) and h i (s) = 1+ h is f i,ifs = female (3) h i (s) = 12 h is (12 f i ), if s = male Here l is the total mortality/inidene rate, l 0 the baseline rate, d i the dose (see below), s gender, e the age at exposure, a the age at risk and a i, v i, and h is are parameters. For e i, the average age at start of follow-up in the LDRMD study was hosen as a surrogate for average age at exposure. The modelling of age-at-exposure and age-attained dependenes in equation (2) is the way the age parameters are treated in reent A-bomb papers, for example, by Preston et al. 17 We based the risk alulations for the atomi bomb survivors on the dose to that organ as it was used in the orresponding LDRMD study, if the study was based on an organ dose the skin dose, if the LDRMD study was based on film badge or TLD readings. Neutron doses were weighed by a fator of 10. The Poisson regressions were performed with the programme AMFIT of the software pakage EPICURE (HiroSoft International Corp., Seattle, Washington, USA). Ratio of ERR-per-dose values The ratio of the ERR-per-dose value, b ldrmdi, in an LDRMD study i and the orresponding value for the atomi bomb survivors was alulated as: q i = b ldrmdi /b lssi. (4) Normal distributions were assumed for b ldrmdi and b lssi with average values orresponding to the best estimates given in the publiations (for b ldrmdi ) or obtained in the Poisson regression (for b lssi ). Standard deviations of the single estimates were estimated by dividing the width of their respetive onfidene interval by twie the appropriate quantile of the normal distribution. Perentiles and the variane V i of q i were alulated from 1000 samples from eah distribution generated with the Monte Carlo software pakage Crystal Ball (Deisioneering, Denver, Colorado, USA). Combined estimator of the risk ratio A ombined estimator of the ratio of the ERR-per-dose values for LDRMD and aute exposures was obtained by the inverse variane method for alulating a weighted average of the ratios for the single LDRMD studies. where n is the number of LDRMD studies onsidered. The ratio Q was alulated separately for studies of aner mortality and for studies of aner inidene. Some of the LDRMD mortality studies had part of the data in ommon. In order to avoid a double ounting of suh mortality data, two analyses inluding only independent studies were performed. In the first analysis, LDRMD studies with the larger number of aner mortality ases were used. In the seond analysis, instead of these, LDRMD studies with the smaller number of ases were inluded. Out of the three analyses (two for aner mortality and one for inidene) the ombined risk estimator with the narrowest unertainty range (the ratio of the upper and the lower boundary of the 90% onfidene interval) was defined to 790 Oup Environ Med 2009;66: doi: /oem

3 Table 1 List of aner mortality studies, whih were inluded in the analysis No Referene, ountry Population Follow-up, aner ases Type of exposure Caner outomes be the main analysis. Sensitivity analyses were performed by exluding single studies from the main analysis. Study heterogeneity was assessed by alulating Cohran s Q statisti and the orresponding p value. Lifetime risks The BEIR VII ommittee performed a probabilisti alulation of the lifetime-solid-aner mortality and inidene risks per dose for low-dose-rate exposures to external radiation aording to: lr BEIRVII = lr lss /DDREF BEIRVII (6) where lr lss is the lifetime risk per dose for aute, high-dose exposures as derived for most aner sites from the inidene data of the atomi bomb survivors from Hiroshima and Nagasaki, transferred to the Amerian population. 3 DDREF BEIRVII has a mode of 1.5 and a 95% CI of 1.1 to 2.3. Lifetime-solid-aner mortality and inidene risks per dose for LDRMD exposures have been alulated here as: lr ldrmd = lr BEIRVII DDREF BEIRVII Q (7) In the alulation, lr BEIRVII and DDREF BEIRVII were assumed to be negatively orrelated with a orrelation oeffiient of In order to hek the impat of this subjetive hoie, limiting alulations were also performed for values of the orrelation oeffiients of 0 and 21. The ICRP has defined the detriment-adjusted nominal risk oeffiient as a weighted sum of lifetime risks per dose for fatal and non-fatal aner, severe heritable effets, and length of life lost. The oeffiient is alulated by: ERR per dose, b ldrmd (Gy) 21, best estimate and 90% CI 1 Boie 2006, USA 18 Workers at Roketdyne 21999, 3066 External and internal All aner exluding 0.0 (21.9 to 2.4)* leukaemia 2 Cardis 2007, 14 Radiation workers in Variable, 6119 External All aner exluding 0.6 (20.1 to 1.4) ountries{ 16 nulear industry leukaemia 3 Ivanov 2001, Russia{ 21 Chernobyl lean-up , 515 External Neoplasms ICD (1.3 to 2.9)* workers 4 Ivanov 2006, Russia 4 Chernobyl lean-up , 651 External Solid aner 1.5 (0.2 to 2.9)* workers 5 Krestinina 2005, Russia 24 Teha River residents 21999, 1842 External and internal Solid aner exept bone 0.9 (0.2 to 1.7)* aner 6 Muirhead 2009, UK 6 Radiation workers 22001, 6959 External Malignant neoplasms 0.3 (0.02 to 0.6) exluding leukaemia 7 Stayner 2007, USA 5 ORNL workers 21984, 225 External All aner exluding 4.8 (0.4 to 13.3)1 leukaemia 8 Telle-Lamberton 2007, Frenh nulear workers , 721 External All aner exluding 1.5 (20.5 to 4.0) Frane" 19 leukaemia 9 Wing 2005, USA 20 Hanford workers 21994, 2265 External and internal All aner 0.3 (20.3 to 1.0) ERR, exess relative risk; ORNL, Oak Ridge National Laboratory. *95% onfidene interval. {Canadian data exluded from 15-ountries study. {Summarised by Ivanov et al 25 in 2007, and therefore inluded in our analysis. 1After orretion for dose unertainties. The result without this orretion is 5.4 (0.5 to 12.6) Gy 21. "Results for all aner exluding leukaemia were supplied by personal ommuniation with Telle-Lamberton (2008). Table 2 List of aner inidene studies, whih were inluded in the analysis No Referene, ountry Population Follow-up, aner ases Type of exposure Caner outomes d ICRP = d lss /DDREF ICRP (8) where d lss is the detriment-adjusted nominal risk oeffiient for aner after aute, high-dose exposures as derived mainly from the inidene data of the atomi bomb survivors. 2 DDREF ICRP has the value of 2. Taking aount of the aner risk per dose in LDRMD epidemiologial studies, a detriment-adjusted nominal risk oeffiient for aner was assessed here aording to: d ldrmd = d ICRP DDREF ICRP Q (9) RESULTS Studies of low-dose-rate, moderate-dose exposures All 12 studies seleted for the analysis were ohort studies. The nine mortality studies (table 1) and three inidene studies (table 2) inluded seven studies on radiation workers, three studies on emergeny and lean-up workers after the Chernobyl aident and two studies on the residents of villages loated along the banks of the Teha River Although a number of Chernobyl liquidators have obtained high-dose-rate exposures, the studies are inluded here, beause the vast majority had only low-dose-rate exposures. None of the 12 studies inlude a onsiderable number of ohort members with umulative exposures exeeding a few hundred milligray. The best estimates of the ERR were positive in all studies (in one study it was 0.0). In seven of the 12 studies the exess aner risk was signifiantly related to the radiation exposure. ERR per dose, b ldrmd (Gy) 21, best estimate and 90% CI 10 Ivanov 2004, Russia 22 Chernobyl lean-up , 1370 External Solid aner 0.3 (20.4 to 1.2)* workers 11 Krestinina 2007, Russia 23 Teha River residents , 1836 External and internal Solid aner exept bone 1.0 (0.3 to 1.9)* aner 12 Muirhead 2009, UK 6 Radiation workers 22001, External Malignant neoplasms exluding leukaemia 0.3 (0.04 to 0.5) *95% onfidene interval. Oup Environ Med 2009;66: doi: /oem

4 Table 3 Datasets, parameters and risk per dose for the atomi bomb survivors orresponding to the low-dose-rate, moderate-dose studies of aner mortality in table 1, and the risk ratios, q i No Population Male fration Average age at start of follow-up Average age at end of follow-up ERR per dose for atomi bomb survivors The ERR-per-dose estimates for the atomi bomb survivors mathed by ategories of aner mortality, sex ratios, average ages at exposure and average ages at risk of the LDRMD studies vary by more than a fator of 2.5 (tables 3 and 4). The highest estimate orresponds to the onditions in the aner inidene study of the Teha River residents: a value of 0.59 (95% CI 0.49 to 0.69) Gy 21 is obtained for relatively young average age at first exposure (25 years) and a large fration of females (0.57). Also, the risk estimation is related to the dose in a relatively wellshielded organ (stomah). The lowest estimate orresponds to a mortality study of Chernobyl liquidators: a value of 0.23 (95% CI 0.11 to 0.34) Gy 21 is obtained for all solid aner and a high male fration (100% males). Further, the risk is related to the relatively high dose in skin. Comparison of ERR-per-dose values for different types of exposure Generally, the unertainties of the ERR estimates in the LDRMD studies are muh larger than the orresponding estimates for atomi bomb survivors (figs 1 and 2). In six of the 12 LDRMD studies, the best estimate of the ERR per dose is larger than that for the atomi bomb survivors by more than a fator of 1.5, in five studies it is omparable, and only in one study it is smaller by more than a fator of 1.5. The risk ratio, q, is signifiantly larger than 1.0 for the two mortality studies of Chernobyl lean-up workers In the remaining 10 LDRMD studies, the aner-risk-per-dose values are ompatible with those from the study of the atomi bomb survivors. Combined estimator of the risk ratio No statistial heterogeneity was deteted between the estimated ratios, q i, inluded in eah of the three analyses (table 5). Dose quantity ERR per dose (Gy 21 ) in LSS, b lss, best estimate and 90% CI Risk ratio, q, best estimate and 90% CI 1 Workers at Roketdyne Skin dose 0.26 (0.16 to 0.35)* 0.00 (27.25 to 7.33) 2 Radiation workers in nulear Colon dose 0.49 (0.30 to 0.67) 1.19 (20.34 to 3.12) industry 3 Chernobyl lean-up workers{ Skin dose 0.47 (0.29 to 0.65)* 4.49 (2.79 to 7.17) 4 Chernobyl lean-up workers Skin dose 0.23 (0.11 to 0.34)* 6.66 (1.67 to 14.7) 5 Teha River residents{ Stomah dose 0.54 (0.42 to 0.65)* 1.71 (0.52 to 3.04) 6 UK radiation workers Skin dose 0.30 (0.21 to 0.39) 0.91 (0.01 to 2.01) 7 ORNL workers Skin dose 0.25 (0.16 to 0.33) 19.6 (26.38 to 51.3) 8 Frenh nulear workers Skin dose 0.33 (0.23 to 0.43) 4.59 (22.34 to 12.6) 9 Hanford workers Skin dose 0.41 (0.33 to 0.49) 0.73 (20.87 to 2.35) ORNL, Oak Ridge National Laboratory. *95% onfidene interval. {Calulations performed for all aner, beause out of 515 neoplasms (ICD ) there were only three non-aner ases (ICD ). {Calulations performed for all solid aner, beause mortality data with DS02 were not available for bone aner. It should be noted, however, that the power of the test is not strong in view of the small number of studies inluded. The unertainty range of the ombined estimator for the larger mortality studies and for the inidene studies had the same width. The analysis of the larger mortality studies was hosen as the main analysis beause it inludes more studies. The main analysis inludes seven aner mortality studies, five of nulear workers, one of Chernobyl emergeny and lean-up workers 21 and one of Teha River residents. 24 A risk ratio, Q, of 1.21 (90% CI 0.51 to 1.90) is obtained. The best estimate for the smaller mortality studies is larger; the differene is, however, not signifiant (p = 0.16). The ombined estimator for the inidene studies is relatively lose to the result of the main analysis. Leaving out one of the studies hanged the best estimate of Q in the main analysis at most by 26%. The lowest risk ratio with a value of 0.96 (90% CI 0.12 to 1.80) was obtained when the study of the Teha River residents was exluded. The highest risk ratio with a value of 1.44 (90% CI 0.48 to 2.41) was obtained when the study of the UK radiation workers was exluded. Lifetime risks Based on assessments of BEIR VII for lifetime aner risks after aute exposures and on the results of the present analysis (equation 7), a number of about 14 (90% CI 6 to 31) or 24 (90% CI 9 to 49) exess solid aner ases among 1000 males or females, respetively, is obtained for LDRMD gamma-ray exposures with a dose of 100 mgy. It is further estimated that there would be about seven (90% CI 3 to 15) or 11 (90% CI 4 to 23) exess fatalities from solid aner among males or females, respetively. If lr BEIRVII and DDREF BEIRVII were assumed to be not or ompletely anti-orrelated, then the best estimates of the Table 4 Datasets, parameters and risk per dose for the atomi bomb survivors orresponding to the low-dose-rate, moderate-dose studies of aner inidene in table 2, and the risk ratios, q i No Population Male fration Average age at start of follow-up Average age at end of follow-up Dose quantity ERR per dose (Gy 21 ) for aute exposure, b lss, best estimate and 90% CI Risk ratio, q, best estimate and 90% CI 10 Chernobyl lean-up workers Skin dose 0.33 (0.21 to 0.46)* 0.99 (21.10 to 3.25) 11 Teha River residents Stomah dose 0.59 (0.49 to 0.69)* 1.70 (0.54 to 2.92) 12 UK radiation workers Skin dose 0.37 (0.29 to 0.46) 0.71 (0.09 to 1.42) *95% onfidene interval. 792 Oup Environ Med 2009;66: doi: /oem

5 Figure 1 Exess relative risk per dose for aner mortality in nine studies of low-dose-rate, moderate-dose exposures (red symbols), as ompared with aute, high-dose exposures (atomi bomb survivors of Hiroshima and Nagasaki) (blue symbols). The error bars indiate 95% CIs for the studies of workers at Roketdyne, the Chernobyl emergeny workers and the Teha River residents, and 90% CIs for all other studies. lifetime risks are essentially the same and the onfidene intervals are inreased or dereased by about 30%, respetively. The radiation protetion system of the ICRP is based on the effetive dose. For whole body exposures with low-let radiation, the effetive dose in the unit Sievert (Sv) is numerially equal to the absorbed dose in the unit Gray (Gy) as it was used by BEIR VII. Based on the assessment of the ICRP for the detriment-adjusted nominal aner risk oeffiient for aute exposures and on the result of the present analysis (equation 9), an estimate of the detriment-adjusted nominal risk oeffiient for workers of about 10 (90% CI 4 to 16) Sv 21 is Figure 2 Exess relative risk per dose for aner inidene in three studies of low-dose-rate, moderate-dose exposures (red symbols), as ompared with aute, high-dose exposures (atomi bomb survivors of Hiroshima and Nagasaki) (blue symbols). The error bars indiate 95% CIs for the Chernobyl emergeny workers and the Teha River residents, and 90% CIs for the UK National Registry for Radiation Workers. obtained for LDRMD exposures. Representing essentially a sum of exess aner fatalities and of weighted exess non-fatal aner ases, this value is slightly larger than the sex-averaged result for the mortality risk as desribed above. DISCUSSION Strengths and limitations of the present study It is the strength of the analysis to have extrated the following ommon information from a number of reent epidemiologial studies of aner after LDRMD exposures: Oup Environ Med 2009;66: doi: /oem

6 Table 5 Ratios of the exess relative risk per dose in low-dose-rate, moderate-dose studies and for the atomi bomb survivors as alulated in three analyses (main analysis in bold) Endpoint Criterion to selet independent studies Numbers of studies inluded* There is evidene for an exess aner risk after LDRMD exposures to ionising radiation. There is no indiation that the exess aner risk per dose for LDRMD exposures is smaller than for the atomi bomb survivors. These results still hold if single studies are exluded from the analysis. Most of the studies inluded in the present analysis have methodologial limitations espeially onerning dosimetry. It is impossible to predit how improvements of dosimetry would or will hange the results of the single LDRMD studies. A Monte Carlo simulation study inorporating unertainty in the dose parameters estimated for study of ORNL workers found very little impat of these unertainties on ERR-per-dose estimates. 5 Further, if future hanges of the results of several LDRMD studies do not go in the same diretion (inreasing or dereasing the risk), then impliations for the general results of the present analyses are expeted to be low, beause the risk ratios in the three different analyses presented in table 5 are quite onsistent; Risk ratio, Q, best estimate and 90% CI Mortality Larger number of aner 1, 4, 5, 6, 7, 8, (0.51 to 1.90) 0.79 ases Mortality Smaller number of aner 1, 2, 3, (1.16 to 3.01) 0.21 ases Inidene 10, 11, (0.41 to 1.54) 0.49 *Compare tables 1 and 2. Figure 3 Ratio Q of exess relative risk-per-dose values for aner after low-dose-rate, moderate-dose exposures and after aute, high-dose exposures as reommended by the International Commission on Radiologial Protetion (ICRP), 2 used by BEIR VII (95% CI), 3 and derived in the present analysis from epidemiologial studies (epi-risk, 90% CI). p Value for heterogeneity the risk ratio of the main analysis is not strongly affeted by a single study. Another severe limitation of the LDRMD studies is the nonavailability of data on risk fators other then radiation, espeially of smoking data. Suh risk fators may onfound the results. Sine, however, neither the LDRMD studies nor the analyses of the atomi bomb survivors take suh risk fators into aount, the risk ratios derived in the present paper may be less affeted by the missing information than the risk estimates themselves. A main limitation of the present analysis is the inlusion of results for different exposed groups and different groups of aner types. Indeed, the relative risks among the atomi bomb survivors mathing the onditions of the LDRMD studies vary by more than a fator of 2.5. There is no obvious way to avoid this limitation beause the available single studies and even the large 15-ountries pooled analysis do not have enough statistial power to allow onlusions as drawn in the present paper. However, the alulation of risk ratios for omparable onditions (groups of aner types, male fration, age at exposure, age attained, dose quantity used in the risk analysis, mortality or inidene data) in the present paper and the determination of a ombined estimator for these ratios alleviate the problem with heterogeneous study onditions and endpoints. Another limitation is the fat that published risk estimates were used instead of individual data from the inluded studies. Aess to individual data from some of the exluded studies is possible via the Comprehensive Epidemiologi Data Resoure ( However, for the urrent analyses suh extensive data aquisition and analysis ould not be undertaken. Finally, in the omparison of risks from protrated and aute exposures, the definition of age at exposure is problemati. In the present analysis, the average age at the start of follow-up has been used in the omparison. An older effetive age at exposure would be more orret, but ould not be estimated in this study. Using an older effetive age at exposure would result in lower ERR-per-dose estimates for aute exposures and thus in even higher q i values than obtained in the present analysis. In summary, the value of the present study is a general estimation of impliations of published studies rather than a quantitative risk evaluation. Comparison with low-dose-rate, high-dose exposures Two papers have been published in the past few years on large ohort studies of solid aner risk due to low-dose-rate, but high-dose exposures. One study inluded workers at the Mayak Prodution Assoiation in the Southern Urals, Russia, whih produed plutonium for the atomi weapons of the former Soviet Union. 13 These workers were exposed to external radiation and to plutonium whih exposed mainly lungs, liver and bone. A first analysis of the aner mortality with regard to other 794 Oup Environ Med 2009;66: doi: /oem

7 organs yielded an estimate of the ERR per external dose whih was onsiderably lower than that for the atomi bomb survivors. It may, however, be noted that leukaemia risks per dose were quite omparable. The seond study inluded residents of northern Kazakhstan who were exposed to the fallout and also to external radiation from atomi bomb explosions performed at the nulear Semipalatinsk test site. 12 The best estimate of the exess relative aner mortality risk per dose was onsiderably higher than that for the atomi bomb survivors. In summary, these high-dose studies do not provide ontraditory evidene for the present evaluation of LDRMD exposure studies. Comparison with BEIR VII and ICRP reommendations Aording to BEIR VII, aner risk after LDRMD exposure is expeted to be by a fator of 1.5, aording to the ICRP by a fator of 2, smaller than among atomi bomb survivors. However, the best estimates of the aner risk in 11 of the 12 LDRMD studies are larger than both expetations (tables 3 and 4). Due to low statistial power most single studies are onsistent with the BEIR VII and ICRP reommendations: the 90% onfidene ranges of 10 of the 12 risk ratios, q i, inlude the value of 0.67, orresponding to the inverse DDREF value used by BEIR VII; eight inlude the ICRP value of 0.5. Aording to the main analysis in the present paper, the ombined estimator of the risk ratio, Q, is ompatible with the DDREF used in BEIR VII, although the BEIR VII risk estimates are in the lower range (fig 3). The risk value reommended by the ICRP is smaller than the present result for LDRMD exposures. This result is borderline signifiant on the 90% onfidene level. The ICRP and BEIR VII base their DDREFs mainly on radiobiologial results inluding animal data, whih, in their majority, suggest a harateristially low risk for low-dose-rate exposures. It remains an open question as to why this harateristi is apparently not refleted in the human epidemiologial data. Impliations The reent epidemiologial studies analysed here provide some evidene that aner risks assoiated with LDRMD exposures to ionising radiation may be greater than those published by BEIR VII and the ICRP. The ICRP rationale for radiation protetion is based on three onepts: justifiation, dose limitation, and optimisation. The results of the new epidemiologial studies highlight the need for justifiation of the use of radionulides and ionising radiation in mediine, industry and researh. Derivation of dose limits for radiation protetion is a omplex proess inluding, for example, omparisons of oupational exposures with exposures to radiation from natural soures, or of radiation risks with other oupational health and mortality risks. Compared with earlier reommendations, the ICRP deided in 1991 to onsiderably redue the reommended limit on effetive dose for oupational exposures to 20 msv per year, averaged over 5 years (100 msv in 5 years). 26 Estimates of aner risks related to exposures with umulated doses of 100 msv have been given in the Results setion. The ICRP has defined optimisation as the soure-related proess to keep the likelihood of inurring exposures, the number of people exposed, and the magnitude of individual doses as low as reasonably ahievable, taking eonomi and soietal fators into aount. 2 The new epidemiologial results may influene optimisation proedures for future use of radionulides and ionising radiation. Probability-of-ausation alulations play an important role in the adjudiation of laims of ompensations for aner diseases after oupational radiation exposures. The omputer ode IREP made available by the US National Institute for Oupational Safety and Health ( is widely used for these alulations. The IREP inludes a DDREF, whih lowers the probability of ausation for low-doserate exposures. 27 Use of suh a fator in these alulations is questioned by the new epidemiologial studies. Indeed, in the UK ompensation sheme it is not assumed that low-dose exposures result in a lower risk per dose than aute, high-dose exposures. 28 Aknowledgements: The authors would like to thank Vitor Ivanov, Colin Muirhead and Maylis Telle-Lamberton for making available unpublished details of their studies, and Norman Gentner for disussions on the Canadian radiation workers study and its implementation in the 15-ountries study. Competing interests: None. Provenane and peer review: Not ommissioned; externally peer reviewed. REFERENCES 1. Cardis E, Vrijheid M, Blettner M, et al. Risk of aner after low doses of ionising radiation: retrospetive ohort study in 15 ountries. BMJ 2005;331: International Commission on Radiologial Protetion. The 2007 Reommendations of the International Commission on Radiologial Protetion. ICRP Publiation 103. Annals of the ICRP 37, Nos. 2 4, Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation, National Researh Counil. Health risks from exposures to low levels of ionizing radiation. BEIR VII phase 2. Washington, DC: National Aademies Press, Ivanov VK, Tumanov KA, Kashheev VV, et al. Mortality in Chernobyl lean-up workers: analysis of dose-effet relationship. Radiation and Risk 2006;15: (In Russian.) 5. Stayner L, Vrijheid M, Cardis E, et al. A Monte Carlo maximum likelihood method for estimating unertainty arising from shared errors in exposures in epidemiologial studies of nulear workers. Radiat Res 2007;168: Muirhead CR, O Hagan JA, Haylok RGE, et al. Mortality and aner inidene following oupational radiation exposure: third analysis of the National Registry for Radiation Workers. Brit J Caner 2009;100: Atkinson WD, Law DV, Bromley KJ, et al. Mortality of employees of the United Kingdom Atomi Energy Authority, Oup Environ Med 2004;61: Habib RR, Abdallah AM, Law M, et al. Caner inidene among Australian nulear industry workers. J Oup Health 2006;48: Hwang SL, Guo HR, Hsieh WA, et al. Caner risks in a population with prolonged low dose-rate -radiation exposure in radioontaminated buildings, Int J Radiat Biol 2006;82: Sigurdson AJ, Doody M, Rao RS, et al. Caner inidene in the US radiologi tehnologists health study, Caner 2003;97: Silver SR, Daniels RD, Taulbee TD, et al. Differenes in mortality by radiation monitoring status in an expanded ohort of Portsmouth Naval Shipyard workers. J Oup Environ Med 2004;46: Bauer S, Gusev BI, Pivina LM, et al. Radiation exposure due to loal fallout from Soviet atmospheri nulear weapons testing in Kazakhstan: Solid aner mortality in the Semipalatinsk historial ohort, Radiat Res 2005;164: Shilnikova NS, Preston DL, Ron E, et al. Caner mortality risk among workers at the Mayak nulear omplex. Radiat Res 2003;159: Auvinen A, Pukkala E, Hyvönen H, et al. Caner inidene among Finnish nulear reator workers. J Oup Environ Med 2002;44: Jartti P, Pukkala E, Uitti J, et al. Caner inidene among physiians oupationally exposed to ionizing radiation in Finland. Sand J Work Environ Health 2006;32: Cardis E, Vrijheid M, Blettner M, et al. The 15-ountry ollaborative study of aner risk among radiation workers in the nulear industry: Estimates of radiation-related aner risks. Radiat Res 2007;167: Preston DL, Ron E, Tokuoka S, et al. Solid aner inidene in atomi bomb survivors: Radiat Res 2007;168: Boie JD, Cohen SS, Mumma MT, et al. Mortality among radiation workers at Roketdyne (Atomis International), Radiat Res 2006;166: Telle-Lamberton M, Samson E, Caer S, et al. External radiation exposure and mortality in a ohort of Frenh nulear workers. Oup Environ Med 2007;64: Wing S, Rihardson DB. Age at exposure to ionising radiation and aner mortality among Hanford workers: follow up through Oup Environ Med 2005;62: Ivanov VK, Gorski AI, Maksioutov MA, et al. Mortality among the Chernobyl emergeny workers: estimation of radiation risks (preliminary analysis). Health Phys 2001;81: Oup Environ Med 2009;66: doi: /oem

8 22. Ivanov VK, Gorski AI, Tsyb AF, et al. Solid aner inidene among the Chernobyl emergeny workers residing in Russia: estimation of radiation risks. Radiat Environ Biophys 2004;43: Krestinina LY, Davis F, Ostroumova EV, et al. Solid aner inidene and low-dose-rate radiation in the Teha River ohort: Int J Epidemiol 2007;36: Krestinina LY, Preston DL, Ostroumova EV, et al. Protrated radiation exposure and aner mortality in the Teha River Cohort. Radiat Res 2005;164: Ivanov VK. Late aner and nonaner risks among Chernobyl emergeny workers of Russia. Health Phys 2007;93: International Commission on Radiologial Protetion Reommendations of the International Commission on Radiologial Protetion. ICRP Publiation 60. Annals of the ICRP 21, Nos. 1 3, Land C, Gilbert E, Smith JM. Report of the NCI-CDC Working Group to revise the 1985 NIH radioepidemiologial tables. US Department of Health and Human Servies, National Institutes of Health, Bethesda, Wakeford R, Antell BA, Leigh WJ. A review of probability of ausation and its use in a ompensation sheme for nulear industry workers in the United Kingdom. Health Phys 1998;74: Zankl M. Personal dose equivalent for photons and its variation with dosimeter position. Health Phys 1999;76: International Commission on Radiologial Protetion. Conversion oeffiients for use in radiologial protetion against external radiation. ICRP Publiation 74. Annals of the ICRP 26, Nos. 3/4, APPENDIX 1: RELATION OF WHOLE BODY DOSE AND SKIN DOSE The term whole body dose, as used in a number of epidemiologial studies of workers exposed to ionising radiation, relates to the dosimeter dose worn in front of the trunk of the worker. Values for this dose quantity are not available for the atomi bomb survivors. The main exposure of the atomi bomb survivors is due to Roentgen rays or gamma rays in the energy range of 100 kev to a few MeV. In this Appendix an organ is identified, for whih dose values are available and whih may serve asa surrogate for the whole body dose among atomi bomb survivors. Zankl published onversion oeffiients for the whole body dose, or more speifially for the personal dose equivalent, H p (10), per air kerma free in air, K a (in Sv Gy 21 ) for a typial dosimeter position, for monoenergeti photons inident in various irradiation geometries. 29 The ICRP published onversion oeffiients of 15 organs in an anthropomorphi phantom per kerma free in air for monoenergeti photons inident in various irradiation geometries. 30 We alulated the ratios of these two sets of onversion fators for two irradiation geometries: parallel from the front (anterior-posterior) and parallel from all horizontal diretions (rotational invariant). For the photon energies and irradiation geometries of interest, the onversion oeffiients for skin were found to be similar to the onversion oeffiients for whole body dose: for both irradiation geometries and the whole energy range the oeffiients agree within 10%. Oup Environ Med: first published as /oem on 30 June Downloaded from on 21 April 2018 by guest. Proteted by opyright. 796 Oup Environ Med 2009;66: doi: /oem