1 Workshop on the occasion of Dr. Mar n Tondel visi ng Osaka /20
On the descrip on about childhood thyroid cancer in UNSCEAR White Paper 2016
Tomoya Yamauchi Graduate School of Mari me Sciences, Kobe Univreisty 21st April 2019 Shin-Osaka Maru-biru Bekkan, 5-3 On the descrip on of childhood thyroid Cancer in UNSCEAR White Paper 2016 2/20 Focusing on two paragraphs of 111 & 112 in UNSEAR 2016 White Paper, we examine by what way they deny the analy cal result of the excess of childhood thyroid cancer in Fukushima Prefecture based on the conven onal epidemiology.
As the report style, these paragraph is completely out of usual scien fic manner.
They have never adequate knowledge about epidemiology, especially on the role of regional classifica on as an opera on variable, an importance of external comparison based on ordinal incidence, and me interval a er the exposure.
The main paper in the present discussion is that published by Tsuda et al. in Epidemiology in 2016. ORIGINAL ARTICLE
5IZSPJE$BODFS%FUFDUJPOCZ6MUSBTPVOE"NPOH 3FTJEFOUT"HFT:FBSTBOE:PVOHFSJO'VLVTIJNB +BQBOUP
Toshihide Tsuda,a Akiko Tokinobu,b Eiji Yamamoto,c and Etsuji Suzukib
the assumption that the rest of examinees were disease free, an incidence Background: After the Great East Japan Earthquake and Tsunami ORIGINAL ARTICLE rate ratio of 12 has already been observed (95% CI = 5.1, 23). in March 2011, radioactive elements were released from the Fuku- Conclusions: An excess of thyroid cancer has been detected by shima Daiichi Nuclear Power Plant. Based on prior knowledge, con- ultrasound among children and adolescents in Fukushima Prefecture cern emerged about whether an increased incidence of thyroid cancer within 4 years of the release, and is unlikely to be explained by a among exposed residents would occur as a result. screening surge. Methods: After the release, Fukushima Prefecture performed ultra- 5IZSPJE$BODFS%FUFDUJPOCZsound6MUSBTPVOE thyroid screening on all residents"NPOH ages ≤18 years. The first (Epidemiology 2016;27: 316–322) round of screening included 298,577 examinees, and a second round 3FTJEFOUT"HFT:FBSTBOE:PVOHFSbegan in April 2014. WeJO analyzed'VLVTIJNB the prefecture results from the first and second round up to December 31, 2014, in comparison with the +BQBOUPOn the descrip on of childhoodJapanese annual incidence and the incidence within thyroid a reference EpidemiologyCancer area t 7PMVNF in UNSCEAR he/VNCFS Fukushima .BZ White Daiichi NPaper 2016uclearThyroid Cancer Power Among Plant Young People released in Fukushima radio3/20- in Fukushima Prefecture. Tactive elements after the Great East Japan Earthquake and Tsunami on March 11, 2011. As the wind shifted direction over Toshihide Tsuda,a Akiko Tokinobu,b Eiji Yamamoto,Results: Thec highestand Etsuji incidence Suzuki rateORIGINAL ratio,b using ARTICLE a latency period of 4 years, was observed in the central middle district of the prefecture com- time, 131I, 134Cs, and 137Cs, in addition to other radionuclides, were pared with the Japanese annual incidence (incidence rate ratio = 50; 95% released to both the northwest and the south of the plant.1 The rela- the assumptionconfidence that interval the rest [CI of] examinees = 25, 90). Twerehe prevalence disease free, of anthyroid incidence cancer was tive amounts of radioactive material released were estimated to be Background: After the Great East Japan Earthquake and Tsunami 131 137 134 rate ratio6055IZSPJE of per 12 millionhas already examinees$BODFS been observed (95%%FUFDUJPO CI (95% = 302, CI =1,082) 5.1,CZ 23). and6MUSBTPVOE the prevalence "NPOH9.1% I, 17.5% Cs, and 38.5% Cs. Compared with Cher- in March 2011, radioactive elements were released from the Fuku- Conclusions:odds ratio A comparedn excess with of thyroid the reference cancer district has beenin Fukushima detected Prefecture by nobyl where one reactor melted down, at Fukushima three reac- shima Daiichi Nuclear Power Plant. Based on prior knowledge, con- 3FTJEFOUT"HFT:FBSTBOE:PVOHFSJO'VLVTIJNB ultrasoundwas 2.6 among (95% children CI = 0.99, and 7.0). adolescents In the second in Fukushima screening round, Prefecture even under tors melted down.2 Radiation released into the atmosphere from cern emerged about whether an increased incidence of thyroid cancer +BQBOUP within 4 years of the release, and is unlikely to be explained by a the Fukushima accident was estimated to be approximately 900 among exposed residents would occur as a result. screeningTsudaEditors’ et al.surge. Note:Toshihide A commentary Tsuda,a onAkiko this Tokinobu, article appearsb Eiji Yamamoto, on page 323.c and Etsuji SuzukibpetabecquerelEpidemiology (131I: 500 petabecquerel, t 7PMVNF 137C/VNCFSs: 10 petabecquerel). .BZ Methods: After the release, Fukushima Prefecture performed ultra- Editors’ Note: Letters to the Editor regarding this article can be found on the 131 FIGURE. .BQ PG 'VLVTIJNBThe radiologic1SFGFDUVSF equivalence to I International Nuclear Event sound thyroid screening on all residents ages ≤18 years. The first (EpidemiologyEPIDEMIOLOGY 2016;27: website 316–322) in the May 2016 issue at http://journals.lww.com/BOETDSFFOJOHBSFBTGPSUIFmSTUSPVOEPG the assumption that the rest of examinees were disease free, an incidence Background: After the Great East Japan Earthquake and Tsunami TDSFFOJOHGSPNScaleUPmTDBM wasZFBST approximately one-sixth of the 5,200 petabecquerel round of screening included 298,577 examinees, and a second round epidem/pages/default.aspx rate ratio of 12 has already been observed (95% CI = 5.1, 23). in March 2011, radioactive elements were released from the Fuku- 3 Submitted 25 January 2015; accepted 10 August 2015.Conclusions: An excess of thyroid cancercalculated has been detected to have by been released by the131 Chernobyl accident. began in April 2014. We analyzed the prefecture results from the first TABLEshima Daiichi 2. N uclear1SFWBMFODF Power Plant. Based1SFWBMFODF on prior knowledge,0EET con- 3BUJPT 103 BOEinhalation*ODJEFODF was the highest3BUF among3BUJPT all estimated *33 radiation dosesJO &BDIwere due%JTUSJDU to substantialVQ I concentrationsUP%FDFNCFS in the south area of From the aDepartment of Human Ecology, Graduateultrasound School among of E childrennvironmental and adolescents in Fukushima Prefecture cern emerged about whether an increased incidence of thyroid cancer to the thyroid, ground shine was Ithen second its highest,health and ingesrisk- assessment,the plant, together with exposurethe World differences Healthbetween radioac Organi- - and second round up to December 31, 2014, in comparison with the and Life Science, Okayama University, Okayama,within Japan;4 years ofbDepartment the release,tion wasand ofthe is lowest.unlikely The to report be explainedindicated that by the a proportion of tive iodine and the total air dose rate. among exposed residents would occur as a result. 9–11 he Fukushima Daiichi Nuclear Power Plantscreening released surge. radio-exposure via ground shine increased as time advanced. InternalIn additionComparison to Japanese sourcesExternal that were cited Comparison by the Japanese annual incidence and the incidence within a reference area Epidemiology, Graduate School of Medicine, Dentistry and Pharmaceu- zation predicted that an excess of thyroid4 cancer12 cases would Methods: After the release, Fukushima Prefecture performed ultra- PrevalenceAside from the screeningof Thyroid in Fukushima Cancer Prefecture that is World Health Organization, Unno et al. reported chrono- c the subject of this study, Watanobe et al.6 conducted a screening logical changes in 131I radioactivity levels in fallout per day soundactive thyroidtical screening elements Sciences, on all after residentsOkayama the ages G University,≤reat18 years. E astThe Okayama, first Japan ( EpidemiologyE Japan;arthquake and 2016;27: Department and 316–322) result6 from radiation-exposed children based ona a prelimi- in Fukushima Prefecture. T in various cities; in 131I radioactivity levels in spinach, cow’s Areasround of and screening Districts included 298,577(1) to examinees,(9) and a second round exerciseCases from 2012 toper 2013 10 including (95% thyroid CI)ultrasonography for POR (95% CI) IRR (95% CI) of Information Science, Faculty of Informatics, Okayama University of 4,5milk, and chicken eggs; and in tap water pollution with 131I Results: The highest incidence rate ratio, using a latency period of 4 Tbegansunami in April on 2014. March We analyzed 11, the 2011. prefecture A resultss the from wind the first shifted direction over1,137 Fukushima residentsnary ages dose 18 years andassessment. younger at the time When the World Health Organization 131 Science,134 Okayama,137 Japan. of the accident. No thyroid cancer was detected in this screening. from March to May of 2011 in various areas of east Japan. years, was observed in the central middle district of the prefecture com- time,and second Iround, Cup s,to Decemberand C31,s, 2014, in addition in comparison to with other the radionuclides, wereIn regions of Japan otherreported than Fukushima, a preliminarythe Japanese Minis- Tdosehey did notestimation consider radioiodine in exposure 2012, through it inhalaestimated- NearestPresented area (1) earlier (2011 aspects fiscal year)of this research at conferenceshe of Fukushimathe International Daiichi N Sociuclear- Power359 Plant (201, released 592) radio- 1.5 (0.63, 4.0) 30 (17, 49) Japanese annual incidence and the incidence within a reference area 1 try of Environment conducted thyroid screening of 4,365 chil- tion. They also measured radioiodine concentrations in breast ety for Environmental Epidemiology in Basel (2013) and Seattle (2014). milk from 119 volunteer lactating women residing within 250 pared with the Japanese annual incidence (incidence rate ratio = 50; 95% releasedMiddlein Fukushima area to Prefecture. both(2012 the fiscal northwest year) and the south of theT plantactive . elements The rela after -dren the andGreat adolescents East402 Japantheages (304, 3–18 E meanarthquake years 522) living population in and three prefectures dose1.7 (0.81, for 4.1) the more-affected33 (25, locations 43) km of the Fukushima nuclear power plant between April 24 Results:T Thehe highestauthors incidence report rate no ratio, conflicts using a latencyof interest. period of 4 Tsunami on March 11, 2011.(Aomori, As Yamanashi, the wind and shifted Nagasaki) direction using ultrasound over in the 2012 7 8 confidence interval [CI] = 25, 90). The prevalence of thyroid cancer was tive N orthamounts middle of district radioactive (2) material released were estimated131 134 to137 befiscal year ; one thyroid237within cancer(123, case 414)Fukushima was detected. We summa Prefecture- and1.0 May (0.40, 31, (excluding 2011. 2.7) Seven of 23 areaswomen who less 20were (10,examinedthan 35) 20 km years, was observed in the central middle district of the prefecture com- time, I, Cs, and Cs, in addition to other radionuclides, were 131 Supplemental digital content is available through direct URL citationsrize previously collected data on thyroid screening including that in April secreted a detectable level of I in their breast milk. 605 per million examinees (95% CI = 302, 1,082) and the prevalence 131 137 134 1 4 9.1% pared Central with theI ,middleJapanese 17.5%in the annual district HT incidenceCMs, L(3) and (incidence PDF38.5% rate versions ratio =C 50; s. of95% C this ompared released article to (www.epidem.com).with both the C northwesther-in Chernobyl and the in south the605 eTfrom ableof (302, the 1 (http://links.lww.com/ plant the 1,082). Tplant,he rela-E DwhichE/A968). were2.6 T (0.99,heimmediately National 7.0) Institute of evacuatedRadiological50 Sciences (25,), estithe90)- less- confidence interval [CI] = 25, 90). The prevalence of thyroid cancer was tive amounts of radioactive materialThree yearsreleased and 10were months estimated after the to accident, be the main mated equivalent doses in mothers and infants from the data odds ratio compared with the reference district in Fukushima Prefecture nobyl whereT hisone content reactor is notmelted peer-reviewed down, at or Fukushimacopy-edited; itthree is the reac sole -respon- affected remainder of of FukushimaUnno et al.,12,13 based on Prefecture, an acute ingestion model. neighboring14 These 605 Koriyama per million examinees City district (95% CI (4) = 302, 1,082) and the prevalence 9.1% 131I, 17.5% 137Cs, andobjective 38.5% of this134C study462s. C was ompared(299, to establish 683) with accurate Cher- and quantita- 2.0 (0.87, 4.9) 39 (25, 57) was 2.6 (95% CI = 0.99, 7.0). In the second screening round, even under sibility 2of the authors. tive estimates from the Fukushima experience and to plan for estimated doses ranged from 119 to 432 mSv among mothers torsodds ratiomelted compared down. with the reference Radiation district in released Fukushima Prefecture into the nobylatmosphere where one reactor from melted down, at JapaneseFukushima three prefectures, reac- theand rest from 330 of to 1,190Japan, mSv in theirneighboring infants for those living countries, 45 to SouthCorrespondence: middle district T (5)oshihide Tsuda, Department of Human Ecology, Gtheraduate future public 486health needs(210, of the957) population. 2.1 (0.7, 6.0) 40 (17, 80) thewas 2.6Fukushima (95% CI = 0.99, 7.0).accident In the second was screening estimated round, even to under be approximatelytors melted down.2 R900adiation released intoand the atmosphere the rest fromof the world.2204 A km map south or ofsouthwest, the includingthree I wakivariously City in the Fuku exposed- Least contaminatedSchool of areaEnvironmental (2013 fiscal and year) Life Science, theOkayama Fukushima University, accident was 3-1-1 estimated 332to be METHODS (236,approximately 454) 900 shima Prefecture,– Ibaragi Prefecture, and Chiba28 Prefecture. (20, 38) 131 137 However, Nagataki et al.15 reported that thyroid radia- Editors’ Note: A commentary on this article appears on page 323. Tsushima-naka, Kita-ku, Okayama, 700-8530, Japan. E-mail:131 tsudatos@ 137 petabecquerel Editors’ Iwaki Note: C Aity commentary district ( onI :(6) this500 article petabecquerel, appears on page 323. Cs: 10petabecquerel petabecquerel). ( I: 500 petabecquerel,Exposure Estimation 451areasCs: (282, 10 petabecquerel). within 682) Fukushima tion1.9 Prefecture doses (0.84, in children 4.8) inis the shown evacuation andin deliberate 38Figure (24, evacu 57). - Editors’ Note: Letters to the Editor regarding this article can be found on the Editors’ Note:md.okayama-u.ac.jp. Letters to the Editor regarding this article131 can be found on the The radiologic equivalence toE xposure131I International information N onuclear 131I from Event the Fukushima ation areas were estimated to be 10 mSv in 95.7% of children TheEPIDEMIOLOGY radiologic website in equivalence the May 2016 issue at to http://journals.lww.com/ I International Nuclear Eventrelease has been uncertain becauseThe of the World 8-day half-life Health of Organization estimated the thyroid PIDEMIOLOGY Southeastern least contaminated district (7) Scale was approximately one-sixth of 236 the 5,200(95, petabecquerel486) (maximum:1 (reference) 35 mSv) among 1,083 by screening20 and(7.9, intake 41) E website in the May 2016 issue at http://journals.lww.com/ epidem/pages/default.aspxCopyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. This131I andis anthe destruction of monitoring sites as a result of the 3 scenario. The timing of evacuations from heavily contami- epidem/pages/default.aspx ScaleSubmitted was25 January approximately 2015; accepted 10 August one-sixth 2015. of the 5,200calculated petabecquerel to have been releasedevent. T oby explain the C differenceshernobylequivalent accident.in the regional doses distributions in 2011 of nated to areas be within 100–200 20 km, and frommillisieverts additional contaminated (mSv) in FromWestern theopen-access aDepartment least of contaminatedHuman article Ecology, distributed Graduate district School under (8) of E thenvironmental terms of the Creative Commons Attri- 305 (146, 561) 1.3 (0.49, 3.6) 25 (12, 47) b In its health3 risk assessment,estimated internal the exposuresWorld Health (through Organi inhalation- and inges- areas mainly northwest of the Fukushima plant, occurred Submitted 25 January 2015; accepted 10 August 2015. calculatedand Lbution-ife Science, toN Okayamaonhave Commercial- been University, released Okayama,No Derivatives Japan; by theDepartment CLicensehernobyl of 4.0 ( CCaccident.BY-NC-ND), where it the131 more affected areas and 10–100 mSv in the3 rest of Fukushima a Northeastern least contaminated district (9) zation predicted that an tion,excess for example, of thyroid of 0 cancerI(0,) and 595)external cases exposures would (for example between0.00 (0.0,March 12 2.6) and mid-June 2011. Many0.00 residents (0.0, were 50) From the Department of Human Ecology, Graduate School of Environmental Epidemiology, Graduate School of Medicine, Dentistry and Pharmaceu- 134 137 1 is permissible to download and share thec work provided it is properly cited.Cs T andhe Cs), Torii et al. suggested that the differences evacuated to areas within Fukushima Prefecture, especially to b tical Sciences,In its Okayama health University, risk Okayama,assessment, Japan; and Departmentthe World result Health from radiation-exposedOrgani- children Prefecture based on a prelimi as - delivered by inhalation, external exposure from and Life Science, Okayama University, Okayama, Japan; Department of ofa Information Science, Faculty of Informatics, Okayama University of 4,5 Thework IRR cannots were basedbe changed on diagnosis in any byway cytology. or used When commercially. basednary doseon histologically assessment. ©Whenconfirmed 2015 Wolters the World Kluwer cases Health,Health that Inc. were Organization All rights operated reserved. on, the IRRs for 4external comparisonswww.epidem.com using a latent | 317 duration of zationScience, predicted Okayama, Japan. that an excess of thyroid cancer cases would ground shine, and ingestion. In the most contaminated areas Epidemiology, Graduate School of Medicine, Dentistry and Pharmaceu- 4 years IwereSSN :28 1044-3983/16/2703-0316 (95% CI = 15, 47) in the nearest area (excludingreported one a benignpreliminary case), dose 30 estimation(95% CI = in 22, 2012, 39) it in estimated the middle area, and 16 (95% CI = 10, 24) in the least contaminated area c Presented earlier aspects of this research at conferences of the International Soci- tical Sciences, Okayama University, Okayama, Japan; and Department resultfor whichety forDO E from nvironmentaltheI: 10.1097/secondary radiation-exposed EpidemiologyE DexaminationE.0000000000000385 in Basel (2013) of cytology and children Seattle (2014). positive based casesthe is onmean incomplete. a population prelimi dose- for the more-affectedjust outside locations 20 km from the plant, the proportion of exposure by of Information Science, Faculty of Informatics, Okayama University of The authors report no conflicts of interest.4,5 within Fukushima Prefecture (excludingCopyright areas © 2015 less Wolters than 20Kluwer km Health, Inc. Unauthorized reproduction of this article is prohibited. Supplemental digital content is available through direct URL citations nary dose assessment. When the World Health Organization 4 Science, Okayama, Japan. in the HTML and PDF versions of this article (www.epidem.com). from the plant, which were immediately evacuated ), the less- Presented earlier aspects of this research at conferences of the International Soci- reportedT316his content a | preliminaryis notwww.epidem.com peer-reviewed or dose copy-edited; estimation it is the sole respon in- 2012,affected it remainderestimated of Fukushima Prefecture, neighboringEpidemiology t 7PMVNF /VNCFS .BZ sibility of the authors. ety for Environmental Epidemiology in Basel (2013) and Seattle (2014). middleCorrespondence: area, Toshihide and Tsuda, 16 Department (95% of Human CI E=cology, 10, G raduate24) inJapanese the least prefectures, contami the rest- of Japan, neighboring countries, the mean population dose for the more-affected locations4 DISCUSSION The authors report no conflicts of interest. School of Environmental and Life Science, Okayama University, 3-1-1 and the rest of the world. A map of the three variously exposed withinnatedTsushima-naka, areaFukushima Kita-ku,for which Okayama, Prefecture 700-8530,the secondary Japan. (excluding E-mail: tsudatos@ examination areas less than of cytology-20 km Although precise measurements of both external and Supplemental digital content is available through direct URL citations areas within Fukushima Prefecture is shown in Figure. md.okayama-u.ac.jp. The4 World Health Organization estimated the thyroid in the HTML and PDF versions of this article (www.epidem.com). frompositiveCopyright the © 2015 casesplant, Wolters Kluweriswhich incomplete. Health, were Inc. A llimmediately rights reserved. This is evacuatedan ), the less- internal radiation exposure in Fukushima were not obtained, Copyright © 2015 Woltersequivalent Kluwer doses Health, in 2011 Inc.to be 100–200Unauthorized millisieverts reproduction (mSv) in of this article is prohibited. This content is not peer-reviewed or copy-edited; it is the sole respon- open-access article distributed under the terms of the Creative Commons Attri- affectedbution-NonT Commercial-he remainder highestNo Derivatives ofIRR License Fukushima 4.0 in (CC BY- theNC -ND), external Prefecture, where it the comparisons more neighboring affected areas and was 10–100 mSvin in theexternal rest of Fukushima comparison, we observed an approximately 30-fold sibility of the authors. Japaneseis permissible to downloadprefectures, and share the thework providedrest ofit is properlyJapan, cited. neighboring The Prefecture countries, as delivered by inhalation, external exposure from Correspondence: Toshihide Tsuda, Department of Human Ecology, Graduate observedwork cannot be changed in the in any centralway or used commercially. middle district of the prefecture, 50 to4 increase in the number of thyroid cancer cases among children ISSN: 1044-3983/16/2703-0316 4 ground shine, and ingestion. In the most contaminated areas School of Environmental and Life Science, Okayama University, 3-1-1 and60DOI :km 10.1097/the westrestEDE.0000000000000385 of from the world. the Fukushima A map of thepower three plant, variouslyjust outside where 20 exposed km residents from the plant, the proportionand adolescents of exposure by using the area/district of residence to provide Tsushima-naka, Kita-ku, Okayama, 700-8530, Japan. E-mail: tsudatos@ areas within Fukushima Prefecture is shown in Figure. were316 | www.epidem.comnot evacuated (IRR based on positive cytologyEpidemiology was 50; t 7PMVNF a /VNCFS surrogate .BZ for exposure information (Table 2). In the early md.okayama-u.ac.jp. The World Health Organization estimated the thyroid Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. This is an 95% CI = 25, 90). Prevalence in the district was 605 cases per reports on excess thyroid cancer from Chernobyl, place and open-access article distributed under the terms of the Creative Commons Attri- equivalent doses in 2011 to be 100–200 millisieverts (mSv) in 24–26 million examineesCopyright © 2015 (95% Wolters CI Kluwer = 302, Health, 1,082), Inc. Unauthorized and thereproduction POR ofcom this article- is prohibited.time were also used as a proxy for exposure information. bution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it thepared more with affected the areas southeastern and 10–100 least mSv contaminated in the rest of Fukushima district was Excesses of thyroid cancer in the central middle district by is permissible to download and share the work provided it is properly cited. The Prefecture as delivered by inhalation, external exposure from work cannot be changed in any way or used commercially. 2.6 (95% CI = 0.99, 7.0). both external and internal comparison were observed, although ground shine, and ingestion.4 In the most contaminated areas ISSN: 1044-3983/16/2703-0316 Although the second round of screening that began in the PORs were relatively lower. The finding that southernmost DOI: 10.1097/EDE.0000000000000385 just outside 20 km from the plant, the proportion of exposure by the 2014 fiscal year was not included in the tables, the num- districts within the middle and the least contaminated areas 316 | www.epidem.com bers of subjectsEpidemiology were as follows: t 7PMVNF total /VNCFSsubjects, 218,397;.BZ actu- had higher IRRs than the northernmost districts was consistent ally screened, 106,068 (49%); among them, for 75,311 (71%), with the flow of 131I being primarily in a southern direction it had already been decided whether the secondary exami- from the Fukushima release. nation was necessary or not; positives in the screening, 611 By considering the prevalence (detected cases per Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. (0.8%); examined in the secondary examination, 377 (62% of 1,000,000), IRRs in Table 2, and years between the accident and the positives); finally diagnosed by the secondary examina- screening—4 to16 years in Chernobyl and less than 4 years in tion, 262 (70%); examined by fine needle aspiration, 22 (8%), Fukushima—we could infer that the incidence of thyroid cancer and detected eight new thyroid cancer cases by cytology up to in Fukushima rose more rapidly than expected based on the cumu- December 31, 2014. All of the eight cancer cases (four males lative attributable thyroid cancer risk over 15 years as estimated and four females with ages at the accident ranging from 6 to by the World Health Organization.5 The radiation burden to the 17 years at the time of the accident in 2011) underwent the thyroid in Fukushima Prefecture might have been considerably first round screening. In three among the eight cases, a ≤5.0- higher than estimated,4 as suggested by other measurements.12 mm nodule and/or a ≤20.0-mm cyst were detected in the first The variability of prevalence in Chernobyl may also result from round. The one cancer case was already operated on (histo- variability in years between the accident and screening. logical type was papillary carcinoma). Mean age of the cases One concern is that the approximately 30-fold increase in 2011 was 12.1 ± 3.4 years in the second round, whereas it observed in the number of thyroid cancer cases in external was 14.8 ± 2.6 years in the first round. Even under the assump- comparison might be the result of a screening effect. This con- tion that the remaining 75,303 (75,311 minus 8 cancer cases) cern is based on the potential presence of silent thyroid cancer are disease free, an excess IRR for external comparison with 3 among children and adolescents in the unscreened regions of years as a latent duration which was maximum time since the Japan. However, the magnitude of the IRRs was too large to first round was observed (12, 95% CI = 5.1, 23). be explained only by this bias. Furthermore, according to the
320 | www.epidem.com © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. On the descrip on of childhood thyroid Cancer in UNSCEAR White Paper 2016 4/20
Secretariat: Infinity Conference Group (Contact: Doreen Albertson) 1035 Sterling Road, Herndon, VA 20170 703-925-0178 (Voice) 703-925-9453 (Fax) [email protected] (E-Mail) Following the publica on of Tsudas’ study, www.iseepi.org (website)
Francine Laden, Sc.D., President the president of Interna onal Society for Verónica Vieira, D.Sc., Secretary-Treasurer Manolis Kogevinas, M.D., Ph.D., President-elect Environmental Epidemiology (ISEE) has
The following correspondence was sent on January 22, 2016, to authorities in Japan including send a le er to the Director of the Ministry Ms Tomoko Kitajima, Director General, Environmental Health Department, Ministry of the Environment Government of Japan of the Environment Government of Japan,
Mr. Hiroyuki Kobayashi, Chief, Department of Health and Welfare, Section for Fukushima Health Management Survey and Fukushima Prefecture. Ms Tamayo Marukawa, Minister of the Environment (ascertained), Ministry of the Environment Government of Japan On behalf of environmental epidemiologist, we at the International Society for Environmental Epidemiology (ISEE), the largest international professional organization for such scientists, are concerned about the recent scientific evidence suggesting an increase in the risk of thyroid cancer among residents of Fukushima that is much higher than previously anticipated. It said “The recently published study The recently published study1 demonstrates a 12‐fold higher risk of developing thyroid cancer among residents of Fukushima compared to the rest of the Japanese population. This is an exceptionally high risk, as pointed out in the demonstrates a 12-fold higher risk of commentary to the published paper. This study builds on previous concerns about the lack of appropriate data and studies to monitor the long‐term impact of the Fukushima nuclear disaster on local residents. Preliminary results of developing thyroid cancer among residents the study were presented at a special Symposium, organized at the ISEE Annual Meeting in Sao Paolo in September 2015. The discussion at the Symposium demonstrated great scientific interest of the Society members in follow up of health consequences of Fukushima accident. of Fukushima compared to the rest of the The study illustrates the need of the on‐going, systematic screening of the population affected by the accident, enabling early detection and treatment of the cases of thyroid cancer. Besides direct benefits to the affected population, such Japanese popula on. This is an prospective study has great value for building up global knowledge about risks of ionizing radiation. We appeal to the government as the stakeholder serving the interest of the people, to develop a series of measures excep onally high risk, as pointed out in to scientifically document and follow up the health of residents of Fukushima and to better understand and estimate the risks from the accident that happened in 2011. We believe that detailed monitoring of population exposure to radiation the commentary to the published paper”. possibly remaining in the environment after the accident remains necessary both for scientific and preventive reasons. Such studies would provide invaluable contribution to the global body of knowledge on health consequences of nuclear accidents and ways for reduction such risks in affected populations. ISEE would be available to assist and support activities where needed by utilizing the expertise of its members. We would be interested to know if, and how, you would envision the involvement of ISEE as an independent international It also said “ISEE would be available to professional organization. We would appreciate hearing back about your perspective regarding our letter and your future plans regarding this assist and support ac vi es where needed important matter.
Sincerely by u lizing the exper se of its members”.
Francine Laden PhD President of ISEE Japanese Government and Fukushima
cc. World Health Organization Prefecture have made no response to this. 1 Tsuda T et al, Thyroid Cancer Detection by Ultrasound Among Residents Ages 18 Years and Younger in Fukushima,Japan: 2011 to 2014. Epidemiology 2015 DOI: 10.1097/EDE.0000000000000385 On the descrip on of childhood thyroid Cancer in UNSCEAR White Paper 2016 5/20 111. One paper [T17] (and a subsequently published response to cri cisms [T16]) claimed to demonstrate that there had been a radia on-induced increase in thyroid cancer incidence: the authors reported a 50-fold (95% CI: 25, 90) excess in Fukushima Prefecture. However, the study design and methods were too suscep ble to bias [J2] to warrant this interpreta on. Tsuda et al. [T17] did not adequately account for the impact of the sensi ve ultrasound screening of the thyroid upon the observed rate of thyroid cancer. Their conclusions were based on a comparison of the rate of thyroid cancer among those people screened by FHMS with the rates found elsewhere in Japan where few children had undergone thyroid screening.
FHMS: Fukushima Health Management Survey [T17], [T16] [J2], [H3], [T6], [A2], [T5] On the descrip on of childhood thyroid Cancer in UNSCEAR White Paper 2016 6/20 111. ・・・・ Studies of other popula ons screened in childhood, par cularly those who underwent ultrasound screening in three unexposed Japanese prefectures [H3], as well as other screening studies of young people in Japan [T6], found baseline rates of thyroid cancer in the absence of radia on exposure that were similar to the FHMS rates. Similarly, the Republic of Korea experienced an apparent large increase in thyroid cancer rates once they ins tuted universal screening [A2]. It is also likely that some of the cancers detected by screening may have existed before the radia on exposure [T5].
[T17], [T16] [J2], [H3], [T6], [A2], [T5] On the descrip on of childhood thyroid Cancer in UNSCEAR White Paper 2016 7/20 112. Wakeford et al. [W2] carried out an analysis of the data in the Tsuda et al. paper by comparing the thyroid cancer prevalence among children studied by FHMS who were residing in locali es with rela vely low, medium, and high exposures as a result of the accident, as defined by Tsuda et al. The analysis by Wakeford et al. did not show any dose–response trend. In fact, the ra o of thyroid cancer prevalence between the locali es with the highest and lowest exposures was only 1.08 (95% CI: 0.60, 1.96) [W2]. Other inconsistencies between Tsuda et al. and the substan al body of data on radia on-induced thyroid cancer in childhood include: (a) the Tsuda et al. paper reported excesses within 1–2 years a er radia on exposure, whereas studies a er the Chernobyl accident and other studies with much larger doses to the thyroid did not show excesses within 3–4 years;
[W2] On the descrip on of childhood thyroid Cancer in UNSCEAR White Paper 2016 8/20 112. ・・・ (b) all the thyroid cancers in the FHMS occurred among those 6–18 years old at radia on exposure, while other studies show the greatest incidence of thyroid cancer induc on was among those with early childhood exposure (before age 5); and (c) the measured doses to the thyroid were much too low to be consistent with the high prevalence they reported [T6, W2]. Because of these weaknesses and inconsistencies, the Commi ee does not consider that the study by Tsuda et al. presents a serious challenge to the
findings of the 2013 report. [W2], [T6] The occurrence of a large number of radia on induced thyroid cancers in Fukushima Prefecture—such as occurred a er the Chernobyl accident—can be discounted, because absorbed doses to the thyroid a er the FDNPS accident were substan ally lower than those a er the Chernobyl accident. UNSCEAR 2013 222 [T17], [T16] [J2], [H3], [T6], [A2], [T5].[W2] References of 111 & 112 in the order of appearance 9/20 [T17] Tsuda, T., A. Tokinobu, E. Yamamoto et al. Thyroid cancer detec on by ultrasound among residents ages 18 years and younger in Fukushima, Japan: 2011 to 2014. Epidemiology 27(3): 316-322 (2016). [T16] Tsuda, T., A. Tokinobu, E. Yamamoto et al. Response to the Commentary by Professor Davis and the Seven Le ers. - a well-known fact should be disseminated to remedy the problems. Epidemiology 27(3): e21-23 (2016). [J2] Jorgensen, T.J. Re: Thyroid cancer among young people in Fukushima. Epidemiology 27(3): e17 (2016). [H3] Hayashida, N., M. Imaizumi, H. Shimura et al. Thyroid ultrasound findings in children from three Japanese prefectures: Aomori, Yamanashi and Nagasaki. PLoS One 8(12): e83220 (2013). Survey of childhood thyroid cancer in Aomori, Yamanashi and Nagasaki prefectures [T6] Takamura, N. Re: Thyroid cancer among young people in Fukushima. Epidemiology 27(3): e18 (2016). [A2] Ahn, H.S., H.J. Kim and H.G. Welch. Korea's thyroid-cancer “epidemic”-- screening and overdiagnosis. N Engl J Med 371(19): 1765-1767 (2014). The New England Journal of Medicine Screening and over-diagnosis in Korea for adults [T5] Takahashi, H., T. Ohira, S. Yasumura et al. Re: Thyroid cancer among young people in Fukushima. Epidemiology 27(3): e21 (2016). [W2] Wakeford, R., A. Auvinen, R.N. Gent et al. Re: Thyroid cancer among young people in Fukushima. Epidemiology 27(3): e20-21 (2016). References of 111 & 112 in the order of appearance in the text 10/20 Letters Epidemiology t 7PMVNF /VNCFS .BZ Epidemiology has opportunity for Epidemiology t 7PMVNF /VNCFS .BZ Letters Disease Control and Prevention” therein. were members of an International Expert elevate cancer rates as much as Tsuda et This source is for policy making, not for Working Group established by the World al.1 claim. open discussions prior to the final scientific research. The author states Health Organization to perform an initial The situation in areas of the for- graph further highlights that grouping a Emerald Group Publishing; 2013. Lost in of association with odds ratios and “However, the National Academy of Sci- assessment of the health consequences mer USSR heavily contaminated follow- number of clearly “different” inequal- translation: rethinking the inequality equiva- relative rates—risk metrics typically ences has reported that childhood cancers of the Fukushima Dai-ichi accident,2 and publica on, using lence criteria for bounded healthle ers variables. to editor. ing the Chernobyl accident in 1986 is ity measurements under a common pp. 3–32. Available at: http://www.emeraldinsight. employed in case–control and cohort have a period of 1 to 10 years. Therefore, we have serious concerns over this inter- of relevance here: in these areas, many title “relative inequality measures” is com/10.1108/S1049-2585(2013)0000021002. studies, respectively. These two alter- based on the best available scientific evi- pretation of Tsuda et al.1 children received high thyroid doses Accessed September 5, 2015. unhelpful for those trying to understand native study designs are much more dence and the following methodology Thyroid disease screening with (much higher than those following the their implicit value judgments. The use- reliable because they are based on indi- presented in this revised White Paper ultrasound can have a dramatic effect Fukushima accident) and there is a clear vidual dose data and, therefore, not fulness of the OR as an inequality mea- on Minimum Latency and Types or Cat- on the detection of thyroid nodules. and large excess of thyroid cancer in this sure depends on how easy it is for policy prone to be influenced by factors that 3F5IZSPJE$BODFS egories of Cancer, the administrator A 15-fold increase in the incidence of group. The thyroids of 13,127 Ukraini- makers to understand its implicit value vary between communities. selected a minimum latency of 1 year for thyroid cancer occurred in South Korea ans, 17 years old or younger at the time judgments and how closely its inequal- "NPOH:PVOH1FPQMF The thyroid cancer rates in the use in the evaluation of cases of child- after the introduction of a national can- ity equivalence criterion matches[J2] public Tsuda study are not consistent with of the accident, were screened between JO'VLVTIJNB 5 perspectives of what changes in health the risks found in earlier cohort and hood cancer for certification in the WTC cer screening program in 1999, with 1998 and 2000. Based on this study, 105 Health Program.…” the incidence rate in regions increasing would leave inequalities unchanged. To the Editor: case–control studies and, therefore, (95% CI: 30, 258) background cases of I would be happy if the present let- in direct proportion to the percentage Given that the inequality equivalence read with interest the article by Tsuda do not seem to be credible based on thyroid cancer would be expected from ter might help readers correctly under- of screened people.3 Consequently, it is criterion of the OR is a mixture of the Iet al.1 Nevertheless, I was very disap- our prior experience with radioactive the first screening in Fukushima prefec- 6 attainment-relative and shortfall-rel- pointed that it failed to identify itself as a iodine. This should be bluntly stated stand the results presented in the last inappropriate to compare the data from ture. The good agreement between this ative measures, it may not add much classic ecologic study and acknowledge because the media do not appreciate column of Table 2. the Fukushima screening program with point estimate and the number of 112 to the picture but instead overcompli- that its findings were, therefore, vulner- that all study designs are not equally Yoshisada Shibata cancer registry data from the rest of cases that has been detected up to the cate things. This is an issue for further able to the ecologic fallacy. Likewise, valid, and the public needlessly pan- Nagasaki University Japan where there is, in general, no such end of March 2015 in Fukushima Prefec- research. the accompanying commentary, by Scott ics over studies of limited scientific Nagasaki, Japan large-scale screening. The proper com- ture1 does not permit the inference that value. Gustav Kjellsson Davis, failed to point out the ecologic [email protected] parison is between different screened an effect of radiation exposure has been 2 I must, therefore, concur with Department of Economics study design. areas within Fukushima Prefecture, demonstrated. A more plausible con- University of Gothenburg The flawed inferential logic, Dr. Davis that “these findings do not REFERENCES since significant radioactive contamina- add anything new regarding radiation- clusion is that the screening program is Centre for Health Economics at the known as ecologic fallacy, threatens all 1. Tsuda T, Tokinobu A, Yamamoto E, Suzuki tion from the accident was confined to University of Gothenburg (CHEGU) induced thyroid cancer.” But I would fur- finding an anticipated increase in thyroid studies that draw risk inferences based E. Thyroid cancer detection by ultrasound a relatively small part of the prefecture. Gothenburg, Sweden ther add that publishing studies that use -FUUFSTamong residents ages 18 years and young- cancer detection across the prefecture. [email protected] on community incidence rates without There is no statistically discern- ecologic study designs without acknowl- er in Fukushima, Japan: 2011 to 2014. Richard Wakeford individual dose data, yet that is but one Epidemiology. 2016;27:316–322. ible difference in thyroid cancer preva- Ulf-G Gerdtham 3 edging the issue of ecologic fallacy is a The University of Manchester of problems with ecologic studies. 2. Rothman KJ. Modern Epidemiology: lence between the low, intermediate, and [W2] Department of Economics Despite the well-known limitations disservice to the people of Fukushima, An Introduction. New York, NY: Oxford Manchester, United Kingdom Health Economics & Management high contamination areas of Fukushima of this study design, a bibliometric who have already suffered greatly and University Press; 2002. Centre for Occupational and Institute of Economic Research do not need the added burden of ground- 3. Tsuda T. On the results of the thyroid screen- Prefecture. The prevalence ratio for the Environmental Health review of ecological studies published ing carried out by Fukushima Prefecture. In: Lund University less worry about their risk of thyroid highest to lowest contamination areas Institute of Population Health Lund, Sweden in major epidemiologic journals found Executive Committee of “Meeting of Nationwide Faculty of Medical and Human Sciences Center for Primary Health Care Research that only 69% of articles clearly speci- cancer—a risk level that most epidemi- Pediatricians,” ed. Nationwide Pediatricians was 1.08 (95% confidence interval [CI]: The University of Manchester Lund University/Region Skåne fied their study’s design by mentioning ologists would consider very small, not- Shall Protect Children from Radiation 0.60, 1.96), and the highest prevalence Exposure from the Scientific Viewpoint. Osaka, Manchester, United Kingdom Lund, Sweden the word “ecological” or “ecologic,” withstanding the Tsuda study. was seen in the area with an intermediate Japan: Executive Committee of “Meeting of [email protected] Dennis Petrie and 49% failed to acknowledge sus- Timothy J. Jorgensen Nationwide Pediatricians”; 2013:27–40 (in level of contamination (prevalence ratio Centre for Health Policy ceptibility to ecologic fallacy as a Health Physics and Radiation Japanese). = 1.21 [95% CI: 0.80, 1.82]). Further- Anssi Auvinen School of Population and major weakness.4 Protection Program 4. Greenland S, Rothman KJ. Measures of oc- more, the measured levels of radioactiv- University of Tampere currence. In: Rothman KJ, Greenland S, Global Health The reluctance of authors to label Department of Radiation Medicine ity in thyroids in Fukushima Prefecture Tampere, Finland Georgetown University Medical Center Lash TL, eds. Modern Epidemiology, 3rd The University of Melbourne their studies as ecologic is understand- ed. Philadelphia, PA: Lippincott Williams & 4 Washington, DC were far lower than would be needed to R. Nick Gent Melbourne, VIC, Australia Wilkins; 2008:47–48. able given the negative reception such [email protected] Public Health England studies often receive. But the Tsuda Chilton, United Kingdom REFERENCES article goes beyond failing to acknowl- REFERENCES To the Editor: Drs. Wakeford and Kesminiene have received Peter Jacob 1. Kjellsson G, Gerdtham UG, Petrie D. Lies, edge that it is ecologic. It actually 1. Tsuda T, Tokinobu A, Yamamoto E, Suzuki research funding from the European Commis- suda et al.1 reported the current find- damned lies, and health inequality measure- hides its design by using “the residen- E. Thyroid cancer detection by ultra- sion. Dr. Wakeford has received a grant from Helmholtz Zentrum München ments: understanding the value judgments. sound among residents ages 18 years and Children with Cancer UK (current) and is a Munich, Germany tial address of the subjects in March Tings of a large-scale thyroid disease Epidemiology. 2015;26:673–680. younger in Fukushima, Japan: 2011 to 2014. screening program in Fukushima Pre- Member of Technical Working Party of UK 2. Popham F. Re: Lies, damned lies, and health 2011…as a surrogate for individual Epidemiology. 2016;27:316–322. Compensation Scheme for Radiation-linked Ausrele Kesminiene inequality measurements: understand- [dose],” and then reports measures 2. Davis S. Screening for thyroid cancer after the fecture, Japan, following the release of Diseases (current). He has served as consultant International Agency for Research on ing the value judgments. Epidemiology. Fukushima disaster: what do we learn from such radionuclides, in particular iodine-131, to US Electric Power Research Institute (past), Cancer 2016;27:e14–e15. an effort? Epidemiology. 2016;27:323–325. from the accident at Fukushima Dai-ichi Horizon Nuclear Power Ltd (past), and the UK Lyon, France 3. Wagstaff A. The bounds of the concentration The author reports no conflicts of interest. 3. Morgenstern H. Ecologic studies in epidemiol- Treasury Solicitor (past). index when the variable of interest is binary, ogy: concepts, principles, and methods. Annu nuclear power station in March 2011. Dominique Laurier with an application to immunization inequal- Copyright © 2016 Wolters Kluwer Health, Inc. All Rev Public Health. 1995;16:61–81. They suggest that these findings indicate Copyright © 2016 Wolters Kluwer Health, Inc. All ity. Health Econ. 2005;14:429–432. rights reserved. 4. Dufault B, Klar N. The quality of modern rights reserved. Institute for Radiological Protection and 4. Kjellsson G, Gerdtham U-G. Research on ISSN: 1044-3983/16/2703-0e17 cross-sectional ecologic studies: a bibliometric an increase in cases of thyroid cancer ISSN: 1044-3983/16/2703-0e20 Nuclear Safety Economic Inequality. Vol 21. Bingley, England: DOI: 10.1097/EDE.0000000000000465 review. Am J Epidemiol. 2011;174:1101–1107. that is attributable to the accident. We DOI: 10.1097/EDE.0000000000000466 Fontenay aux Roses, France
© 2016 Wolters Kluwer Health, Inc. All rights reserved. XXXFQJEFNDPN ] e17 e20 ] XXXFQJEFNDPN © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. References of 111 & 112 in the order of appearance 11/20 [T17] Tsuda, T., A. Tokinobu, E. Yamamoto et al. Thyroid cancer detec on by ultrasound among residents ages 18 years and younger in Fukushima, Japan: 2011 to 2014. Epidemiology 27(3): 316-322 (2016). Main Paper [T16] Tsuda, T., A. Tokinobu, E. Yamamoto et al. Response to the Commentary by Professor Davis and the Seven Le ers. - a well-known fact should be disseminated to remedy the problems. Epidemiology 27(3): e21-23 (2016). Response [J2] Jorgensen, T.J. Re: Thyroid cancer among young people in Fukushima. Epidemiology 27(3): e17 (2016). cri cal Le er [H3] Hayashida, N., M. Imaizumi, H. Shimura et al. Thyroid ultrasound findings in children from three Japanese prefectures: Aomori, Yamanashi and Nagasaki. PLoS One 8(12): e83220 (2013). Survey of childhood thyroid cancer in Aomori, Yamanashi and Nagasaki prefecturesReference of Main Paper [T6] Takamura, N. Re: Thyroid cancer among young people in Fukushima. Epidemiology 27(3): e18 (2016). cri cal Le er [A2] Ahn, H.S., H.J. Kim and H.G. Welch. Korea's thyroid-cancer “epidemic”-- screening and overdiagnosis. N Engl J Med 371(19): 1765-1767 (2014). The New England Journal of Medicine Screening and over-diagnosis in Korea for adults Reference of Le er & Response [T5] Takahashi, H., T. Ohira, S. Yasumura et al. Re: Thyroid cancer among young people in Fukushima. Epidemiology 27(3): e21 (2016). cri cal Le er [W2] Wakeford, R., A. Auvinen, R.N. Gent et al. Re: Thyroid cancer among young people in Fukushima. Epidemiology 27(3): e20-21 (2016). cri cal Le er References of 111 & 112 in the order of publica on 12/20 [H3] PLoS One 8(12): e83220 (2013). Reference of Main Paper Survey of childhood thyroid cancer in Aomori, Yamanashi and Nagasaki prefectures [A2] N Engl J Med 371(19): 1765-1767 (2014). Reference of Le er & Response Screening and over-diagnosis in Korea for adults [T17] Epidemiology 27(3): 316-322 (2016). Main Paper
[J2] Epidemiology 27(3): e17 (2016). cri cal Le er [T6] Epidemiology 27(3): e18 (2016). [W2] Epidemiology 27(3): e20-21 (2016). [T5] Epidemiology 27(3): e21 (2016).
[T16] Epidemiology 27(3): e21-23 (2016). Response
UNSCEAR wrote 2016 White Paper based on the Le ers, completely ignoring the Response by the authors. UNSCEAR completely failed to describe the discussions in Epidemiology as an academic and expert journal. UNSCEAR insulted the academic ac vity in Epidemiology. The role of regional classifica on as an opera on variable 13/20 111. One paper [T17] (and a subsequently published response to cri cisms [T16]) claimed to demonstrate that there had been a radia on-induced increase in thyroid cancer incidence: the authors reported a 50-fold (95% CI: 25, 90) excess in Fukushima Prefecture. However, the study design and methods were too suscep ble to bias [J2] to warrant this interpreta on.
The study design and method was constructed by Fukushima Medical University, FMU, not by the authors of paper [17].
The Le er [J2] insisted the study is an “ecological study” but FMU has started this as a “Cohort study”.
The Le er [J2] claimed that individual dose have never measured, but this can not be a ributable to the authors of paper [17]. The role of regional classifica on as an opera on variable 14/20
Since the paper [17] is u lizing the regional classifica on as an opera on variable to express exposure level, it is adequate analysis from the view point of Epidemiology.
One example of 2x2 table, without miss classifica on. It is assumed that we know real exposure level for individual.
exposed non-exposed Totals
Cases 40 10 50
Noncases 60 90 150
Totals 100 100 200
Odds for exposed : 40/60 = 0.667 Odds for non-exposed : 10/90 = 0.111 Odds ra o : (40/60)/(10/90) = 6.6 We recognize the causal rela onship between the exposure and the cases. The role of regional classifica on as an opera on variable 15/20 The regional classifica on is never match with the real classifica on. Let’s consider the case when 20% of exposed to be treated as non- exposure. (40 -> 32, 60 ->48, 100 -> 80) exposed non-exposed Totals
Cases 32 18 50
Noncases 48 102 150
Totals 80 120 200
Odds for exposed : 32/48 = 0.667 Odds for non-exposed : 18/102 = 0.176 Odds ra o : (32/48)/(18/102) = 3.78
The regional classifica on has a bias to close the odds ra o to one (no effect). When one find an effect using regional classifica on, it is enough. The paper [T17] found the effect, applying the regional classifica on. UNSCEAR do not have enough knowledge on Epidemiology. An importance of external comparison based on ordinal incidence 16/20 111. ・・・・ Studies of other popula ons screened in childhood, par cularly those who underwent ultrasound screening in three unexposed Japanese prefectures [H3], as well as other screening studies of young people in Japan [T6], found baseline rates of thyroid cancer in the absence of radia on exposure that were similar to the FHMS rates. In [H3] one case of childhood thyroid cancer was found from 4,365 persons. The simple ra o par 1,000,000 persons is 229. UNSCEAR insists this ra o is comparable to that in FHMS. But this kind of discussion lacks sta s cal considera on. In Poisson distribu on, lower and upper limits of 95% confidence interval of count 1 are 0.025318 and 5.57163, respec vely. The observa on in [H3] is concordant with the ordinal incidence between 5.8 and 127.6 persons par 1,000,000 persons. Age group in [H3] was between 3 and 18 years old. The lower group was not involved, different from FHMS. It is difficult to make any scien fically meaningful discussions. An importance of external comparison based on ordinal incidence 17/20 111. ・・・・ Studies of other popula ons screened in childhood, par cularly those who underwent ultrasound screening in three unexposed Japanese prefectures [H3], as well as other screening studies of young people in Japan [T6], found baseline rates of thyroid cancer in the absence of radia on exposure that were similar to the FHMS rates. Le er [T6] said three cases of thyroid cancer were found from 2,307 new comers to Okayama University in 2013. Since they were university students, the age group is different from that in FHMS. Response [T16] indicated the fact that no addi onal cases were found between 2012 and 2015 from 36,927 persons, including elder students. UNSCEAR ignored this response when they prepare the White Paper in 2016. An importance of me interval a er the exposure 18/20 112. Wakeford et al. [W2] carried out an analysis of the data in the Tsuda et al. paper by comparing the thyroid cancer prevalence among children studied by FHMS who were residing in locali es with rela vely low, medium, and high exposures as a result of the accident, as defined by Tsuda et al. The analysis by Wakeford et al. did not show any dose–response trend. In fact, the ra o of thyroid cancer prevalence between the locali es with the highest and Epidemiologylowest exposures was only 1.08 (95% CI: 0.60, 1.96) t 7PMVNF /VNCFS .BZ [W2]. Letters
700 To the Editor: o oshihide Tsuda and his colleagues1 600 Le er [W2] paid no Treport that the annual incidence rate 500 of a en on on the me thyroid cancer observed from Sep- tember 2011 to March 2014 among ca. 400 300,000interval a er the children in Fukushima ages 300 18 years or less as of March 2011 is approximately 30-fold higher than the 200 meanexposure, nevertheless annual incidence rate in Japan
100 reported by the National Cancer Center
Thyroid cancer prevalence per Mi of they spoken about the Japan. However, the formula they use 0 in Table 2 for deriving the incidence rate low middle high ratio,dose-response in which the prevalence is approxirela on.- FIGURE. 1SFWBMFODFPGUIZSPJEDBODFSJOBSFBTPG'VLVTIJNB1SFGFDUVSFXJUIEJGGFSFOU mately equal to the product of “incidence SBEJBUJPODPOUBNJOBUJPO rate” and “latent duration of disease,” has not been validated to date, although Alfred Korblein. Re: Thyroid cancer among young people in Fukushima. they citeEpidemiology 27(3): e18-e19 (2016). 2nd edition of Rothman2 prob- To the Editor: 2007.2 In Korea, however, it is 18.3 ably for authorizing their formula. n the recent article by Tsuda and col- among men and 87.4 among women in Rothman2 presents two formulas Ileagues, the authors performed two 2010.3 Kweon et al.4 described how the involving incidence rate: (1) Risk equals types of comparisons, internal and exter- high age-standardized incidence rate is incident rate × time, and (2) prevalence, nal. In the internal comparison, no dif- mainly explained by enhanced detec- if small, is approximately equal to inci- ference among nine areas in Fukushima tion (screening effect) and the changes dence rate × mean duration of the dis- was observed. In contrast, in the exter- in medical practice patterns rather than ease. They coined the above-mentioned nal comparison, extremely high incident by specific factors. Even higher age- formula by replacing risk in (1) with risk ratios—between 20 and 50 (except standardized incidence could be prevalence and by replacing the mean in one area)—compared with the rates observed if the participation rate is duration of disease in (2) with latent from national cancer registries in Japan extremely high, as in the study by Tsuda duration of the disease, which cor- were reported. et al. responds to time in (1). Since Tsuda For valid external comparison, For valid causal inference, more et al. also use formula (2) in the previ- comparability should be discussed. In detailed data on age and sex distribution ous presentation,3 I wonder why they this case, the system of case finding are needed. abandon it in the present article.1 One in Fukushima and other area is not the Sadao Suzuki possible reason may be that they noticed same, and therefore a direct comparison Department of Public Health the following statement in Saunders could be misleading. Nagoya City University Graduate School of and Rothman4 regarding the formula in In the study, the participa- Medical Sciences (2): “the formula presenting the three tion rate was as high as 81%, and this Nagoya, Japan quantities does not apply to age-specific rate affects the incidence rate (i.e., a [email protected] prevalence.” The prevalence of thyroid screening effect). For example, age- cancer is indeed age specific. REFERENCES standardized incidence rate of thyroid Incorrect citation is also seen in 1. Tsuda T, Tokinobu A, Yamamoto E, Suzuki cancer per 100,000 in Japan was 2.2 E. Thyroid cancer detection by ultrasound the reference (numbered 31), which they among men and 7.9 among women in among residents ages 18 years and young- refer to for defending their assertion er in Fukushima, Japan: 2011 to 2014. against the criticism that 4 years are too Epidemiology. 2016;27:316–322. 2. Katanoda K, Matsuda T, Matsuda A, et al. short for radiation exposure to induce An updated report of the trends in cancer in- thyroid cancer. The author of the cited The author reports no conflicts of interest. cidence and mortality in Japan. Jpn J Clin reference (numbered 31) is “World Trade Oncol. 2013;43:492–507. Copyright © 2016 Wolters Kluwer Health, Inc. 3. Jung KW, Won YJ, Kong HJ, Oh CM, Seo HG, Center (WTC) Health Program Adminis- All rights reserved. This is an open-access article Lee JS. Cancer statistics in Korea: incidence, trator” and we do not read “Centers for distributed under the terms of the Creative Com- mortality, survival and prevalence in 2010. mons Attribution-Non Commercial-No Derivatives Cancer Res Treat. 2013;45:1–14. License 4.0 (CCBY-NC-ND), where it is permis- 4. Kweon SS, Shin MH, Chung IJ, Kim YJ, The author reports no conflicts of interest. sible to download and share the work provided it is Choi JS. Thyroid cancer is the most com- properly cited. The work cannot be changed in any mon cancer in women, based on the data Copyright © 2016 Wolters Kluwer Health, Inc. All way or used commercially. from population-based cancer registries, rights reserved. ISSN: 1044-3983/16/2703-0e19 South Korea. Jpn J Clin Oncol. 2013;43: ISSN: 1044-3983/16/2703-0e19 DOI: 10.1097/EDE.0000000000000462 1039–1046. DOI: 10.1097/EDE.0000000000000461
© 2016 Wolters Kluwer Health, Inc. All rights reserved. XXXFQJEFNDPN ] e19
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Letters Epidemiology t 7PMVNF /VNCFS .BZ An importance of me interval a er the exposure Response [T16] 19/20
The first point is about minimum TABLE. 3FTVMUTPG*OUFSOBM$PNQBSJTPOGSPNUIF'JSTU3PVOEPG4DSFFOJOHJO&BDI latency. Our inference on minimum "SFBPS%JTUSJDUVQUP+VOF 8IFOUIF&GGFDUPGUIF-FOHUIPG5JNF&MBQTFE latency was not solely based on the Cen- #FUXFFOUIF"DDJEFOUBOE5JNJOHPG4DSFFOJOH8BT$POTJEFSFE ters for Disease Control-World Trade Cen- 20 1st Cancer ter report but also on the Chernobyl data Examinees Cases POR (95% CI)a showing the excess thyroid cancer cases within 3 years after the 1986 Chernobyl Nearest area (1) (2011 fiscal year) 41,810 15 4.6 (2.2, 11) 1 year accident.10–13 The excess was mainly in Middle area (2012 fiscal year) 139,338 56 2.6 (1.2, 6.0) 2 years teenagers within 1980s, in contradiction North middle district (2) 50,618 12 1.5 (0.65, 3.9) to the statement, “many patients were Central middle district (3) 18,194 11 3.9 (1.6, 10) younger than 10 years old.”4 Koriyama City district (4) 54,063 25 3.0 (1.4, 7.2) Second, we discuss the issue of South middle district (5) 16,463 8 3.1 (1.2, 8.4) screening effect. In contrast to the second Least contaminated area (2013 fiscal year) 119,328 42 - 3 years round, sensitivity analysis was performed Iwaki City district (6) 49,429 24 2.1 (0.92, 5.2) by employing the latent duration in the Southeastern least contaminated district (7) 29,820 7 1 (reference) first round between the date when thy- Western least contaminated district (8) 33,720 11 1.4 (0.54, 3.8) roid cancer became detectable by screen- Northeastern least contaminated district (9) 6,359 0 0 (0, 2.5) ing (i.e., more than 5.1 mm in diameter) aWe obtained these results by correcting the observed cancer cases. With regard to central middle district (3) and Koriyama city district (4), we show the means of corrected PORs and 95% CIs because we cannot use numbers with a and the date when it could be diagnosed decimal point in EpiInfo 7. clinically. Latent duration assigned in our CI indicates confidence interval; POR, prevalence odds ratio. article was 4 years,9 which may be long for the particular hypothesis tested. One can Prevalence of the lengthO ofdds time Ra os are highest in Nearest area and lower in elapsed between the contaminated areas in Ukraine” using a assign any number of years, even 20, in accident and timing of screening. We 14 the sensitivity analysis and still observethe least contaminated districts. Higher POR is found in Iwaki the linear excess relative risk model. Given mentioned as the second limitation of our remarkable excesses. The effect of preclin- the substantially larger sample size, how- City, where radio iodine plume had passed away. Iwaki city is not 9 ical tumors can be quantitatively estimated article, that this potential confounding ever, direct estimation from ultrasound using an assigned proportion of preclinicalso contaminated by radio cesium. leads to an underestimation. As a sensi- screening data among 47,203 examinees tumors among the detected cancer cases tivity analysis, we assigned the time from in the unexposed or relatively low con- (eAppendix; http://links.lww.com/EDE/ the accident to screening in each area to taminated areas in Ukraine would be more B28). According to the pathological details adjust for the elapsed time: 1 year for appropriate, where no cancer cases were of the 96 thyroid cancer cases operated at the nearest, 2 years for the middle, and detected (95% confidence interval: 0–78 Fukushima Medical University, released 3 years for the least contaminated areas per million examinees),16–19 as shown on August 31, 2015,21 only eight cases (Table). The adjusted prevalence odds in eTable 1 of our article.9 Furthermore, (8%) were free of lymph node metastasis, ratios indicated a dose–response relation- although disregarded by some of the extrathyroidal extension, and/or distant ship. In general, individual-level data can letters,2–5 comparability, for example by only be derived for ingestion in field set- metastasis. Although suggested by some age and diagnostic criteria, should be 2–5,7,8 tings, and we employed areas and districts letters, a screening effect due to pre- considered when using the findings from as a surrogate for individual radiation clinical or slow-growing tumors can be South Korea.15 Screening in South Korea dose measurements.9 As we mentioned in excluded as a plausible interpretation. was conducted among adults with differ- Third, we discuss the issues of our article,9 this potential nondifferential ent diagnostic criteria from Fukushima, dose–response relationship and individ- exposure misclassification would lead to where one quarter of surgical patients had ual radiation doses. As for the relatively underestimation in our findings. tumors less than 5.0 mm in diameter,15 low risk estimates in the nearest area to Finally, we respond to concerns whereas no cancers in this size range were the accident, we should consider the effect about information regarding the unexposed 4 population. When comparing the Ukraine detected in Fukushima. Takamura pre- experience14,22 with that in Fukushima, sented another example of inappropriate Wakeford et al.5 cited Jacob et al.22 for the comparison with the all-school screening The authors report no conflicts of interest. program started at Okayama University, Supplemental digital content is available expected number of thyroid cancer cases through direct URL citations in the HTML in the first screening (i.e., first 4 years) in Japan in 2012. Although the Okayama and PDF versions of this article (www. Fukushima. This was estimated from data study did detect three thyroid cancer cases epidem.com). recorded 12–14 years after the Chernobyl by palpation among 2,307 freshmen (ages Copyright © 2016 Wolters Kluwer Health, Inc. All accident among 13,127 residents aged 18 18 or older) in 2012, no other cases were rights reserved. ISSN: 1044-3983/16/2703-0e22 or younger at the accident, through extrap- detected among the total of 36,927 stu- DOI: 10.1097/EDE.0000000000000468 olation to the unexposed from “heavily dents enrolled between 2012 and 2015. e22 ] XXXFQJEFNDPN © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. 20 Conclusion /20 Descrip on about childhood thyroid cancer in UNSCEAR 2016 white paper （111&112）was examined. The descrip on by UNSCEAR has failed to correctly reconstructed the scien fic discussions which were made in the academic and expert journal of “Epidemiology”. UNSCEAR have never enough knowledge about epidemiology, especially on the role of regional classifica on as an opera on variable, an importance of external comparison based on ordinal incidence, and me interval a er the exposure for dose-resopnse rela on. We can not accept the descrip on in 111 & 112 of UNSCEAR2016 white paper, from the view points of sciences. For Japanese reading persons 本発表に関連する論考が『科学9月号』（岩波書店）に掲載されています。 以下のサイトから無料でダウンロードできます。 h p://www.lib.kobe-u.ac.jp/handle_kernel/90005240 神戸大学学術成果リポジトリ