Assessing the Radiation Risk to Workers Exposed to Thorium

4 4 7 Assessing the radiation risk to workers exposed to thorium

• C. CHALLETON • A. BIAU (IRSN) • S. FOUCHARD (IRSN) • J.-R. JOURDAIN (IRSN) DE VATHAIRE (IRSN) • P. BOISSON (IRSN) • C. COSSONNET (IRSN) • A. RANNOU (IRSN) • M. BOURGUIGNON (OPRI) Protection of human life

n the context of actual or suspected situations of Handling monazite was brought to a halt in the I exposure to ionizing radiation, requests for health last phase of the company as the lighters were risk assessments have been made of the public assembled from purchased components. So in experts in the field, OPRI and/or IPSN, that were principle two different risk periods relating to merged to form IRSN in 2002. Audits were carried thorium can be distinguished for the site employ- out jointly by OPRI and IPSN at Pargny-sur-Saulx in ees and the local population: response to the demand made by the Prefect of the • in an early period (1934-1970), the potential Champagne-Ardennes Region. In 1997 drums of exposure and contamination risk was severe radioactive waste were found on the site of Orflam- because the thorium ore was being handled; Plast, a company now in liquidation, revealing the • from the 1970s onwards, the risk of external past exposure of its former workforce that had been exposure was greater than that of internal con- unknown and whose severity had to be assessed. tamination on the Orflam-Plast site, although the risk presented by the presence of thorium in the dike leaching waters is an unknown quantity. It was thus essential that a radiological risk assessment be made. IRSN carried out a series of surveys from 1997 un- The records of the former Orflam-Plast manufac- til 2001 for the purpose: turing plant at Pargny-sur-Saulx, revealed that • geographical study of mortality2; from 1934 to the beginning of the 1970s, signif- • control case studies3; icant amounts of imported monazite ore, rich in • radiotoxicological analyses. natural radioactive elements, particularly tho- The epidemiological studies were carried out in rium-232, with a half-life of 14 billion years had conjunction with INSERM’s U521 unit (geo- been handled. At that time, Orflam-Plast chemi- graphical study of mortality) and the health

References cally extracted the ore, cerium, a hard metal used watchdog Institute InVS (control case survey). 1 - “Société Orflam- for making lighter flints. The residual waste of IRSN’s in-house mass spectrometry thorium Plast à Pargny-sur- Saulx: expertise this manufacturing process was in solid form measuring facilities were used to carry out urine des conséquences made up of mineral sand with high thorium con- analyses for radiotoxicity. sanitaires par l’OPRI”, internal tent. The several thousand metric tons of mon- OPRI report, 1997. azite handled in this plant thus led to an initial Geographical study of mortality 2 - F. de Vathaire, C. Challeton de approximation of an almost equivalent amount of Vathaire, J. Ropers, waste. Apparently no precautions were taken An audit carried out by IRSN in 1997, demon- A. Mollié, “Cancer Mortality in the when spreading this ore residue whose radioele- strated that the former employees and neigh- Commune of Pargny-sur-Saulx in ment content was ignored or unknown in the boring population were likely to have been France”, J. Radiol. environment. It was found scattered on the very exposed to ionizing radiation at levels consider- Prot, 18, 23-27, 1998. site of the company and in particular, in the ably higher than naturally occurring radioactivity 3 - C. Challeton de premises where the ore was handled in bulk1. levels, the dose rate occasionally peaking at Vathaire, P. Quenel, A. Zeghnoun, About 7500 metric tons of this residue was used 10 µSv/h, that is over 100 times the natural J.-H. Leblanc, “Étude to build an embankment several hundred meters ambient level1. These estimates, based on the cas-témoins de mortalité par cancer along the River Saulx, both upstream and down- measurements made in the defunct manufactur- du poumon et de la vessie, Pagny-sur- stream of where it crossed the plant site; the ing environment, did not enable any conclusions Saulx, Marne, 1986- residue was thus “put to use” protecting the to be drawn on earlier potential exposure. 1996”, internal OPRI report, June 2000. neighboring meadows from winter floods. Moreover it appeared that the majority of

SCIENTIFIC AND TECHNICAL REPORT 2002 159

was associated with having worked on the Orflam-Plast site in order to go a step forward. Occupational exposure to This epidemiological study of control cases of cancer mortality at Pargny-sur-Saulx was con- ionizing radiation was cited as ( ducted in conjunction with the Environmental potentially responsible for the Health Department and the Health Watchdog surplus of cancers observed. Institute (InVS) and with the agreement of the French civil liberties organization, the CNIL. The individual data on death consolidated for the period 1968-1996 was supplied by INSERM’s SC8, for people living within a 10 km radius of Pargny-sur-Saulx at the time of death. For this control case mortality study carried out on the general population, the immediate, principal and Orflam-Plast employees likely to have been combined causes of each death were obtained. exposed the most were recruited and lived in the The cases were defined as being individuals districts surrounding Pargny. deceased through lung or bladder cancer in the This led to a study into cancer mortality in con- study area and for the period under considera- junction with INSERM’s U521 unit from data tion. from the national mortality database, managed The control cases were defined as being individ- by INSERM’s joint service 8 (SC8) at Le Vésinet. uals who had died from some other cause in the This study was carried out across all the districts area under study for the study period. They were within a 16 km radius of Pargny, for the period selected from a file of causes of death excluding 1968-1994. It revealed an overshoot of 40% over respiratory ailments and genito-urinary appara- the norm of deaths through lung cancer and tus ailments. Secondly these control cases were bladder cancer in men aged over 40 living in the paired up to the cases by sex, age (in 5-year seg- catchment area at the time of death. Moreover, ments) and the period of occurrence of death. this mortality surplus increased significantly as Three such periods of occurrence of death were the study approached the district of Pargny-sur- Saulx (11 additional cases of lung cancer and 6 additional cases of bladder cancer). Measuring ionizing radiation A precise mortality analysis did not reveal a sur- on the Orflam-Plast plant site. plus of deaths through cancer in the neighboring districts of the Saulx downstream of Pargny. This observation enabled us to rule out the risk of a major health implication that might have resulted from thorium contamination carried by leaching waters. Consequently, the significantly higher than average national mortality rate through lung and bladder cancer registered in male residents of Pargny prompted the search for a one-off possibly occupational origin. This is why occupational exposure to ionizing radiation • was cited as potentially responsible for the surplus of cancers observed, and in particular lung cancers, because of the presence of thorium in the dust generated while handling the ore.

Control case study

It was necessary to test the hypothesis according to which death through lung or bladder cancer within a radius of 10 km of Pargny-sur-Saulx

160 INSTITUT DE RADIOPROTECTION ET DE SÛRETÉ NUCLÉAIRE 4

Orflam-Plast site (March 1999). Protection of human life

defined, in principle, on the basis of the mortal- izing access to data relating to the death of peo- ity survey findings for the general population: ple registered in the RNIPP in the context of period No. 1, 1968-1977; period No. 2, 1978- research into health”. Eleven of those 46 indi- 1987; period No. 3, 1988-1994. A fourth period viduals were subsequently identified. (1995-1996) was added to draw on the maximum The statistical analysis of the data consisted in possible number of cases. Three paired control comparing the frequency of membership of the cases were associated with each of the other category of former Orflam-Plast employee in the cases by random selection. real and control cases by making a paired analy- The characterization of exposure to thorium was sis (Epi-info software, version 6.04). carried out as follows. Research was made to find out for all the cases and control cases THE FINDINGS selected, whether they had or had not worked at The mortality data for the study between 1968 the Orflam-Plast plant, however long the period and 1996 are as follows: 3501 deaths of which of employment. This research was based on the 879 cancers, broken down into 154 cases of lung register of former Orflam-Plast employees, cancer (142 men and 12 women) and 39 cases held by the Champagne-Ardennes Regional of bladder cancers (27 men and 12 women). The Employment, Work and Professional Training file of former employees of the manufacturing Office (DRTEFP) in the occupational medical site numbered 692 individuals. The proportion of inspection department. The file was compiled female employees was 74%. Of the 641 employ- from occupational medical records (for employees for whom information was available, 380 ees who worked on the site from 1980 onwards), (59%) were hired before 1970. Of the 599 the work inspection register (register of employ- employees for whom information was available, ees hired and leavers) and the register of 434 (72%) were aged under 20 when they were employees covered by redundancy schemes. hired to work on the site. For each real and control case, a search was • As regards cases of lung cancer, only 3 of the made on the “date of birth” identifier cross-ref- 150 cases to which 3 control cases could be at- erencing the file of medical causes of death tached had worked at Orflam-Plast, whereas 7 against the file of former employees. As this of the 450 corresponding control cases were common identifier could be insufficient to con- former workers on the site. The paired odds firm whether the same individual was pin- ratio (OR) (statistical parameter measuring the pointed, it was necessary to check for excess risk) was 1.29 (confidence interval at consistency between names and dates of death 95% calculated according to the exact Fischer in the official registers. Of the 692 individuals method: 0.27–4.89). listed in the former Orflam-Plast employees’ file, As the confidence interval is either side of 1 for the birth date details of 46 were unknown. which the risk is zero, there is thus no significant INSERM’s unit 472 was requested to supply the statistical association between the fact of death life statistics of these people and find their dates through lung cancer and having worked at of birth, in application of decree 98-37 “author- Orflam-Plast.

SCIENTIFIC AND TECHNICAL REPORT 2002 161

contamination by this radioelement after several

Thorium elimination years have elapsed. Given the fact that excretion from the human body of thorium through the feces is greater than is very slow because through urine, analysis of the former is more appropriate. Unfortunately it is harder to gain this element is retained ( acceptance for this analysis, which is why it was for a long time in certain decided to proceed with urine analyses. organs, in particular Furthermore conventional radiotoxicity examinations (chemical purification and alpha spectrom- the lungs and liver. etry) are very onerous and time-consuming. Only one mass spectrometry method (ICP-MS, Inductively Coupled Plasma-mass Spectrometry) could enable mass radiotoxicity examinations of urine to be carried out in a short time. • There was no case of bladder cancer among A cohort of former Orflam-Plast workers was the former employees and only one in the con- brought together by the DRTEFP occupational trol group of 117 individuals. As only 1 of the medical and labor force inspection service of 39 groups did not conform the OR could not be Champagne-Ardennes. estimated. A site analysis of the Orflam-Plast plant enabled Thus the control case study does not enable us a questionnaire to be drawn up for completion by to explain the higher mortality rate through lung the former workers so that the contour and activ- or bladder cancer observed in men in the geo- ity of the various workstations, in particular dur- graphical mortality study. Furthermore in view of ing the first ore-processing era, could be the fact that the expected number of lung can- accurately defined. Additionally each former cers in the essentially female worker cohort, is employee could list the various jobs held during low – around 3 - , it is unrealistic to suggest a the course of employment at the company. cohort study limited to the factory workers: The cohort, which was reconstructed as fully as effectively it would carry very little weight and possible by the DRTEFP, cannot enable exhaus- would only detect very high excess mortality, a tive radiotoxicity studies to be performed for two factor of 3 or 4, which on the basis of the state reasons. Firstly, a certain number of individuals of the art, is very unlikely. could not be traced because they had moved out of the region. Furthermore, the health monitor- Radiotoxicity studies of the urine ing had been offered to other people who were of former workers free to take part in the study as volunteers. The responses to the questionnaire effectively led The high lung and bladder cancer figures to one workstation being singled out as being revealed during the geographical mortality study particularly exposed to monazite ore dust. Ten prompted the immediate implementation of tho- former employees out of the respondents had rium contamination studies without waiting for worked for a long time at this workstation. the results of the control case study. IRSN carried out a radiotoxicity analysis of the References Thorium elimination from the human body is • urine of these 10 people as a priority applying the 1 - International effectively very slow because this element is following strategy. Commission on Radiological Protection, retained for a long time by certain organs, par- If the results observed in these 10 former employ- “Age-dependent Doses to Members of the Public ticularly the lungs – in the case of inhalation -, ees were negative, the probability of observing a from Intake of the bones and the liver. The biokinetic models set positive result in another, according to the survey Radionuclides : Part 2 Ingestion Dose up by the International Commission for less exposed employee in principle, would be Coefficients”, ICRP Radiological Protection1,2 indicate that while most considered as zero. If the opposite were to apply Publication 67. Annals of the ICRP, 23 (3/4), 1993. of the incorporated thorium activity is eliminated the study should be pursued. 2 - International in the first ten days following acute inhalation, The urine of these 10 employees was collected Commission on Radiological Protection, excretion through urine and feces continues for for 24 hours in two-liter bottles with sulfamic acid "Individual monitoring for internal exposure of many years at very low levels. as a preserving agent. The radiotoxicity analysis workers", ICRP Only radiotoxicity analyses of feces or urine of the urine comprised the following radiochem- Publication 78, Annals of the ICRP, 27 (3/4), 1997. searching for thorium are likely to corroborate ical and count stages:

162 INSTITUT DE RADIOPROTECTION ET DE SÛRETÉ NUCLÉAIRE

The additionofatracer,20mBq • • • The firstworkcarriedoutinvolvedthestudyofa tion stages,makesquantitativeanalysesof initial urinepriortotheradiochemicalprepara- 0 B.l µBq. 300 and and • • presence ofvaryingquantitiesnatural plicated matterasithastotakeintoaccountthe three analyses.Interpretingtheresultsisacom- takes twofullweeksofworktoconductthe This methodologyisverytime-consuming;it asfrcnetainlvl faot1mBq.l days forconcentrationlevelsofabout1 radioelements (severalminutesinsteadofseveral ably reducesthemeasuringtimeforlong-lived alpha-spectrometry, theuseofICP-MSconsider- of sensitivityandresponsespeed.Comparedwith improve radiotoxicityanalysisprotocolsinterms mass spectrometrytechniquetodaylikelyto wider studybemadeusinganICP-MS,theonly These preliminaryfindingsthusdictatedthata mal readingsfornaturalthoriumelimination. 2 oftheformeremployeeshavehigherthannor- spectrometry thorium areestablishedbymakinganalpha- the isotopichalf-lifeandidentificationof rium iscoprecipitatedbylanthanumfluoride; the twoeluatesareputtogetherandtho- hydrochloric acid; nitric medium,theneluatedbyconcentrated resin isspecificallyfixedonTEVAina the thoriumfractionnotboundtoanionic then eluatedbyconcentratedhydrochloricacid; the thorium following concentratednitricacidtreatment, and magnesiumphosphates; an ammoniummediumbycalciumphosphates actinide coprecipitation–andthusthoriumin

228 quantitative analyses of 232 Th possible. Th intheurine(approximately -1 Addition ofatracer,20mBq on average).Theresultsshowthat 4+ count oftheprecipitate. Th istrappedonananionicresin, ( 229 Th, tothe 228 232 -1 Th Th ). • 232 The acid andcoldtemperature. Sample preservationwasprovidedbysulfamic for 24hoursinthetwo-literbottlesprovided. were willingtotakepartandcollectedtheirurine Meuse. Only79workersoutofthe110contacted sur-Saulx –theMarne,Haute-Marneand living inthethreedepartmentsclosetoPargny- June 2000fromformerOrflam-Plastemployees hours wasthencarriedoutbyIRSNattheendof simple dilutionofurine an ICP-MSuraniumanalysisprotocolfollowing The resultsledtotheproposalandvalidationof simple procedureformeasuringuraniuminurine. (mBq/24 hr)in expressed inmass(µg/24hr)andhalf-life get overthematrixeffects.Theresultswere standards werepreparedfromurineblanksto medium forICP-MSmeasurements.The urine inultrapurenitricacidat2%,afavorable ing inafactorof20aliquotpartsml is 5.5ng.l natural uraniumforagivenapparatussensitivity this protocol,thedetectionlimit,forexamplein 232 0.48–2.7 ng/24hr,or2–11.1µBq/24hr. limits obtainedonthe79samplesvaryfrom each measuringcampaign.Thusthedetection depends ontheapparatusadjustmentswith adults notexposedtothoriumsurchargedwith short. Theprotocolwasperfectedontheurineof 232 asthehalf-livesofitsotherisotopesaretoo urine measuring uraniumtothoriuminthe Laboratory) transposedtheprotocolvalidatedfor Internal DoseAssessmentandModeling Orflam-Plast employees,theLEMDI(IRSN In thecontextofhealthmonitoringformer Th and Th. Asamplingcampaignofurineover24 232 3 . ICP-MScanonlymeasurethoriumisotope- Th measurementsweretakenafterdilut- -1 or 0.14mBq.l 229 228 232 Th, made CETFCADTCNCLREPORT TECHNICAL AND SCIENTIFIC Th possible. Th. SensitivityinICP-MS 1,2 by afactorof20.Using -1 . “Determination of CossonnetandJ.Ritt, C. 3 -N.Baglan, 167-181, 2001. Radioprotection Monitoring”, Measurements inWorker Uranium Urinary ICP-MS Protocolsfor “Implementation of J. RittandP.Berard, Cossonnet,F.Trompier, C. 1 -N.Baglan, References Purposes”, Individual Monitoring in UrinebyICP-MSfor 81 (1),76-81,2001. urinesparICP-MS”, les l’uraniumdans de de mesuredirecte Cossonnet,“Protocole C. 2 -J.Rittand 77 (4),455-461,1999. Health Phys. Health Phys.

, 36(2), 232 2002

Th , , 163 4 Protection of human life Individual excretion of natural thorium through the urine

varies widely with the physiological characteristics of

the individuals and the contamination levels of food ingested. (

The results for the former Orflam-Plast employ- through the urine varies widely with the physio- ees show that significant values, i.e. situated logical characteristics of the individuals and the above background level, are only observed in contamination levels of food ingested. 53% of the cases. It should be noted that the 2 Consequently the variations observed in the former employees whose results were higher results may be due to a difference in nutritional than normal when measured by conventional ingestion of thorium, which can be considerable radiotoxicity measurements did not take part in within the population, particularly through min- this voluntary campaign. eral water. The results of the former Orflam-Plast employees It can be concluded that the elimination of thorium were compared with those of 539 workers not in the urine by the 79 former Orflam-Plast employ- exposed to thorium in the Hurtgen1 study. The ees examined seems normal. To confirm this maximum value of the series of measurements result, out of some fifteen urine samples repre- 232 Reference made on the Orflam-Plast site (0.10 mBq/24 hr) senting those with the highest Th readings, the 1 - C. Hurtgen, “Natural was of the same magnitude as the minimum 228Th was measured by spectrometry – after chem- Radioactivity in Bioassay by Alpha Spectrometry value observed in Belgium by Hurtgen in an ical purification. These results enable the existence Measurements”, unexposed population (0.15 mBq/24 hr). We of significant contamination of these employees by J. Radioanal. Nucl. Chem., 248, 477-482, 2001. note that individual excretion of natural thorium • thorium to be ruled out.

164 INSTITUT DE RADIOPROTECTION ET DE SÛRETÉ NUCLÉAIRE 4

Conclusion into the excess mortality revealed by the geographical study, there are two lines of investiga- To sum up, the various studies carried out by tion open: IRSN did not enable: • the first would certainly be to study whether the • any statistically significant association to be mortality excess can be explained by the socio- revealed between having worked in the Orflam- professional characteristics of the local popula- Plast plant and having died from lung or blad- tion; der cancer within a 10 km radius of Pargny-sur- • it would also be useful to study the relation Protection of human life Saulx; between death through lung and bladder can- • any significant elimination of 232Th through the cer within a 10 km radius of Pargny-sur-Saulx urine of former Orflam-Plast employees to be and having worked in another industry in the demonstrated. district (for example, the brickworks). But that Thus the results of these studies are inconclusive • is quite another story. with regard to the high mortality rate through This research could be envisaged if the local lung or bladder cancer observed in men included authorities were to formally request them. in the geographical mortality study. Moreover it enables us to rule out the responsibility for con- Acknowledgements tamination by thorium as the cause of the high rate of cancer observed. IRSN thanks Doctor J.-H. Leblanc (DRTEFP) and L. As regards any continuation of these studies, and Caffé (DDASS) for their aid during this study. more particularly the search for an explanation (

The results of these studies are inconclusive with regard to the high mortality rate through lung or bladder cancer observed in men included in the geographical mortality study.

SCIENTIFIC AND TECHNICAL REPORT 2002 165