中国云锡矿工的氡致肺癌危险系数与砷的复合作用

中华放射医学与防护杂志 1998年第4期第18卷 特邀专著

作者：孙世荃

Risk coefficient of radon-induced lung cancer

　　Yunnan tin mine of China

Su Shiquan

　　Objective To identify the suitability of excess relative risk (ERR) coefficient of radon-induced miner lung cancer proposed by NIH Publication (1994) which was based on the data of Yunnan tin miners. Methods Using the collected materials of epideminology,underground ventilation,radon progeny monitoring and environmental arsenic pollution,the problems in the process of estimation of cumulative exposure of radon progeny and adjustment of combined effect of arsenic exposure for Yunan tin miners by NIH Publication were analyzed.Results Because of overestimation of the time of underground working and actual exposure to radon progeny for previous Yunnan tin miners by NIH,the resultant ERR coefficient of radon induced lung cancer was underestimated.The relative risks of different radon-arsenic exposure groups were used by NIH to adjust the combined effect of arsenic without considering the difference of environmental pollution in miner and control groups;that was the another reason for the decrease of ERR coefficient.Conclusion The ERR coefficient of radon-induced lung cancer was probably underestimated for Chinese miners by NIH Publication,therefore it is unsuitable to be used in the risk projection and risk management for radon-induced lung cancer in China.It may be premature to use the Yunnan tin miner data as the value for Chinese miners in the combined analysis of world miner studies.

　　 Key words Lung cancer Radon progeny Risk assessment

　　1 Introduction

　　The UNSCEAR 1994 Report (Annex A) presents the cumulative exposure of radon progeny (WLM) and excess relative risk coefficient (ERR/WLM) of radon-induced lung cancer in 11 miner cohorts with an average of 0.49% per WLM.It was about half of the values in ICRP Publication 32,50,65 and BEIR IV.Among the data from 11 cohorts,the lowest one was from China,0.16% per WLM.These data were referred from the US NIH Publication No.94-3644 “A Joint Analysis of 11 Underground Miner Studies,1994”^［1］．The Chinese value in NIH Publication was based on the “cooperative study” of Yunnan Tin Corporation (YTC) miner study between scientists from China amd the United States^［2］,and the recalculated results using the similar data base obtained from YTC.

　　According to NIH Publication the average relative risk (RR) (adjusted for arsenic exposure) of lung cancer was not increased but decreased after exposure to 100-199 WLM,and the lower limit of 95% CI began to reach 1.0 only after as high as 800 WLM;that is quite different to the experience from Chinese uranium miners^［3］．Among the 2 701 lung cancer cases in the 11 cohorts in NIH Publication,980(36.3%) were from China,which made an important input in NIH Publication.

　　Considering that NIH Publication involves the risk assessment and control of Rn-induced lung cancer in China,it is worthwhile to trace back the calculation of the risk coefficient of YTC miners and the adjustment of combined effect of arsenic exposure based on the collected data by the author,hoping that it will be helpful for understanding the background in NIH ERR coefficient.

　　2 YTC miner data used for NIH ERR coefficient

　　ERR coefficient is the ratio of relative increment of lung cancer to the cumulative exposure to radon decay products:ERR/WLM.The NIH prospective YTC cohort was obtained from the occupational health registration of active and retired employees registered in 1976 at five YTC mines (mines L,M,S,K,H),followed up to 1987.The author believe that there was no apparent uncertainty in the numerator of ERR coefficient.The uncertainty,it any,may exist in the denominator,i.e.the estimated WLM,which is the product of exposed concentration of radon progeny and the time of exposure.

　　2.1 Concentration of radon progeny

　　Mines L,M and S were the oldest YTC mines in Gejiu,Yunnan Province.Before 1950,Miners,mostly child labourers,carried ore bag on back from manually developed small crude tunnels.This working style continued up to about 1953.After the foundation of new China,YTC was established in 1950 and began the modern exploitation.Radon monitoring started from 1972 in the above mentioned 5 YTC mines.About 90% of the miners suffering from lung cancer had the history of working underground before 1949,and some miners started working underground from the 1920's.In order to estimate the previous cumulative exposure NIH divided the time period into the following.Before 1953,the estimation of exposure level was based on the direct measurement between 1977 and 1986 in 13 small crude tunnels,which were abandoned or still operated by the local villagers,with a total of 117 measurements,2.3 WL on the average.After 1972,the direct measured results of radon were used.1953-1972 was the period with more uncertainty and estimation was made from the concentrations after 1972 and the previous condition of tunnels.

　　NIH did not describe the estimation of concentrations in detail.Its end results are as follows:2.3 WL before 1953,2.2 WL in the 1960's,1.7 WL in 1971-1975,1.2 WL in 1980 and 0.9 WL in 1985,which were approximately similar to the values reported in China^［4,5］.It is worthwhile to note that,for the period before 1953,all of the above authors used the single average from principally the same set of data measured in small tunnels mainly in 1977-1986 to represent the exposures in all the different small tunnels before 1953,because almost all the original small tunnels described and registered as related to the exposured of each miner had been collapsed and destroyed.For the period of 1953-1972,separated single averages were used for each of the 5 mines,i.e.the value measured in 1972 in each mine was used as the basis to estimate the previous exposure unchangeably for 20 years.

　　2.2 Time of underground working

　　Because of no individual adjustment for actual time of exposure,7 hrs per day and 285 days per year were used by NIH for the calculation of WLM

　　3 Problems in WLM estimation

　　3.1 Radon concentration before 1953

　　Using the available data of actual measurements in the presently existing small tunnels in the only possible choice for the estmation of exposure in previous small tunnels before 1953.The present small tunnels were not quite similar to the previous tunnels,a more important problem is that the radon measurements conducted around 1980 were arranged mostly in the deep part of the tunnels and at the points suspected to be having higher concentration,and were not arranged along the way of the tunnels.The concentration was generally higher in the deep part of tunnel than along the way.Similar condition was also encountered in the recently developed modern tunnels.

　　The previous small tunnels were generally very deep,the deepest end was about 1 000 m downwards.If one step equals 0.5 m,it takes about 2 000 steps to reach the deepest and.Su Rjiang^［6］．investigated 42 small tunnels in Gejiu area in 1942,of which 31 were more than 2 000 steps downwards.A labourer should go and come back 4-5 times per day in the tunnel.It means that the time of working underground for carrying ore was spent mainly along the wasy of walking and climbing but not at the deep face only.So,NIH using the concentrations obtained mostly from the deep part of the tunnel and assuming the whole day time underground there,may have overestimated the exposures.

　　3.2 Radon concentration during 1953-1972

　　In the tunnels of YTC,radon comes mainly from the cracks and exhausted cavities after mining^[7].Because the exits are much different in altitude and the cavities are located mostly at the upper parts,the air pass downwards through the radon contaminated upper exhausted cavity during summer,thereby leading to the increase of radon in tunnel.On the opposite,fresh air will pass upwards through the tunnel during winter,leading to the decrease of radon^[8].This natural ventilation is helpful for the decrease of radon even after the improvement through mechanical ventilation.So,the historical changes of radon concentration depended mainly on the changes of natural ventilation,which was related to the historical changes of formation of altitude differences of the exits and the formation of exhausted cavities.That could be explained from the history of exploitation of the three largest YTC mines.

　　The three largest mines (L,M,S) had not principal difference in the ore formation,mining method,ventilation equipment and production management.But the radon concentration (Bq/L) in 1972 in mine L was apparently higher than those in mines M and S;the averages (and ratio) of L∶M∶S were 28.6∶3.9∶6.4 (1∶0.14∶0.22).Mechanical ventilation and radon exposure were improved during the period 1973-1980,but the radon concentration in mine L was still higher than in mines M and S.That difference may be related to their different natural ventilation which could be expressed by the seasonal difference in radon concentration.The seasonal difference in mines M and S was larger than mine L,indicating that their lower radon levels may be related to their better natural ventilation than in mine L.

　　Then,when did such difference in natural ventilation start in these mines?The answer should be traced back to the history of tunnel exploitation and cavity formation which decided the level of natural ventilation of tunnels.The result of studies shows that,mine L was long exploited with vertical shaft,the terrain was even,and the difference of atomospheric pressure between the exits of tunnels was small.For mine M,the condition was generally similar to L-mine in early years.Then,the exits of adit tunnels increased after 1962,with increment of difference in atmospheric pressure.Meanwhile,there was surface subsidence and collapse,forming many cracks connected with surface,leading to the inprovement of natural ventilation.For mine S,many adits were opened along the precipitous hill,the pressure difference and the effect of natural ventilation were higher,but the surface-connecting cracks were less than in mine M.So,the natural ventilation would not be better than in mine M afterwards^[9,10].

　　The above comparison indicates that,the differences of natural ventilation and radon concentration in the three mines were formed probably since the mid-1960's,but nor from the early 1950's.So,NIH considered that “radon progeny levels were thought to be stable over time at least until 1972” and used the actually measured value of 1972 as the bases to trace back the levels in whole 20 years (1953-1973) may underestimate the exposure in mines M and S,and lead to the distrortion of proper proportionality between estimated WLM,actual exposure and risk of lung cancer.

　　3.3 Time of underground working

　　In previous small tunnels,as NIH noted,no hourly duration of exposure data were availabe.The length of working was registered as number of years,and every year was added up to calculate the successive working years.The results of author's interview with some miners generally conformed with the following early report:“Running of tin mining started after the Spring Festival” (at February) each year,recruiting miners,repairing houses,affanging the site for production.Before July it was the period of mining and carrying ore inside and outside the tunnel.After July it was the period of sluicing ore on the surface.The length of employment seldom exceeded two years,some workers left at any time within one year”^[6].“In winter there was shortage of water,miners decreased in number or discharged”^[11]．So,calculation of the length of underground working time as whole year (285 d) for the successive registered years of each miner,as NIH assumed,may lead to over-estimation.

　　Exposure for 7 hours everday also remains to be discussed.It was true that:“For tunnel deep to 1 km,4-5 times of ore carrying per day was needed may last for 8 hours,sometimes more than 10 hours”.But since “After coming out from the tunnel,miners were greatly exhausted,could not talk and walk for more than 10 minutes”,and “they should walk through rugged mountain path to arrive the site of ore piling”^[6] for acceptance,the 7 or 8 working hours every day were not spent all the time underground.

　　Working time after 1953:The author believe there may be no great error in working time after 1953 assumed by NIH.

　　4 Combined effect of arsenic

　　4.1 Adjustment for arsenic exposure made by NIH

　　Fig.1 is the dose-effect curve of China tin miners given by NIH Publication.The original RR (without adjustment for arsenic exposure) in each WLM groups was 1.79-7.63,reduced to about 1/3 i.e. 0.77-1.76 after adjustment for arsenic.NIH pointed out that the intercept of the fitted original curve was higher than 1.0,and shifted to less than 1.0 after adjustment for arsenic.NIH considered that this shift could have been the result of inherent differences between non-exposed and exposed workers,possibly with regard to smoking habits.

　　Using constant linear ERR model,ERR/WLM coefficient was reduced from 0.61% to 0.16% after arsenic adjustment by NIH.That was similar to the result of cooperative study^[2],but was apparently lower than that of 1.7% reported by Lubin et al^[12]:They pointed out that the reasons for this difference were uncertain,might have differed by chance alone and design.Their previous report presented a case-control study which changed to a cohort study afterwards．

　　The “cooperative study” compared the RR between each radon and arsenic exposure groups.The cumulative arsenic exposure was expressed as arsenic exposure measure AEM mg*y^-1*m^-3:the product of arsenic concentration and years of exposure.The result (Table.1) shows that with the increment of Rn exposure,the total RR increased to 1.9;with the increment of arsenic exposure,the total RR increased to 5.7,the slope of RR for arsenic exposure was substantially larger than that of radon.They noted that the interpretation was problematic since the exposure to arsenic and radon were highly correlated,further evaluation was needed.But,obviously,Figure 1 and Table 1 indicate clearly that the contribution of arsenic is about 3 times higher than that of radon.In fact the Chinese miner values given by NIH(1994) and UNSCEAR (1994) were just the reduced values after adjustment for arsenic exposures.

　　4.2 Effect of environmental arsenic pollution

　　The reference value governing the RR of each Rn-As exposure groups came from lung cancer cases registered as “no Rn- no As” group of surface control workers.The author looked through the lung cancer cases who never worked underground at the 5 specified mines and died during 1976-1987,which was the follow up period of NIH cohort study.The total number was 25 (not 41 as in Table 1,the reason is tobeascertained),among them 13 had the history of working on the surface of mine area during period 1929-1949;7 were employed for ore carrying,5 for ore sluicing,and 1 for other work.In early years “water shortage occurred from November,2 buckets of clean drinking water costed more than 2 yuan there,and miners had to drink the dirty water taken from the pond nearby”^[6].Many old miners had the experience of drinking pond water contaminated by ore sluicing which contained arsenic or the “mud precipitated”sluicing water;8% of miners aged over 40 had arsenic skin keratosis on palms and soles^[10];some lung cancer cases including those working on the surface who had lived previously in the mine area developed arsenic skin keratosis and skin cancer.So,environmental arsenic pollution was an important combined factor in the high incidence of YTC miner cancer.

　　Table 1 RR and cases of lung cancer death amont non-exposed (level 0) and quartiles (level Ⅰ-Ⅳ) of radon and arsenic exposure (from Xuan et al.1993)

　　A:age adjusted original value without adjustment for arsenic;B:adjusted for age and cumulative exposure to arsenic.from NIH Pub.1994,Tab B1(a) and Fig.1(a)

　　Fig 1 Relative risk (RR) of lung cancer by cumulative WLM

　　The situation mentioned above indicates that the “no Rn-no As” group in Tab.1 is unsuitable for providing reasonably the reference value for the miner group.Obviously,the lower reference value of “Rn-no As” (RR 1.3,0.4,1.1,and 0.8) and higher reference value of “As-no Rn” (RR 4.7) are the important reason for the higher increment of RR from arsenic exposure than from radon.It should be noted that these reference values are based on very small number of lung cancer cases,and NIH did not describe their exposure conditions in detail,e.g.who had been exposed to radon without exposure to arsenic ore dust (group Rn-no As) during underground working?Among “no Rn” group,all arsenic exposure groups are blank except As I,whose RR is 4.7,which was used as reference value and played an important role in the adjustment for arsenic,but was based on only 2 cases.They may be workers exposed to ore dust on surface without history of working underground,but whether or not they were exposed to surface water arsenic contamination is unclear.

　　If such unknown and uncertainties are laid aside,and “no Rn-no As” groups are omitted,the differences of RR increments with increasing of Rn and As exposure will be diminished greatly.In Tab.1,981 lung cancer cases are grouped into quartiles,the largest number of cases in group Rn Ⅰ-Ⅳ appeared also in groups of As Ⅰ-Ⅳ respectively,because WLM and AEM are both increased with the year of underground working.Correlation of Rn and As exposure made it difficult to differentiate exactly their etiologic contributions based only on the comparison of RR from cumulative Rn and As exposures.

　　4.3 Histogenetic approach on the etiology of lung cancer

　　As noted above,it was difficult to differentiate the etiologic contribution from exposure to radon and arsenic dust.Another way to deal with this problem is to assume that only two occupational carcinogens existed in tunnels,i.e.radon and arsenic dust.Then each of their risk could be estimated from the total risk and their relative etiologic contribution estimated.The author reported three methods for estimating the relative contribution of Rn/As^[13],based on (a) comparison of ERR coefficient or Rn and Rn+As-induced lung cancer in miners,(b) comparison of the incidence of Rn,As and Rn+As-induced lung cancer in rats^[14],and (c) comparison of the histogenetic probability of radon and arsenic ore dust in the induction of lung cancer^[15]．The results showed that the contribution of radon may be 3 times higher than that of the arsenic ore dust.

　　The principle of histogenetic analysis is that the inhaled radon and its progeny will decay and disappear from the deposited area soon,so it is histologically invisible,but deposited ore dust is visible;local deposition of less soluble arsenic ore dust and its continuous dissociation are the basis of carcinogenesis for arsenic ore dust^[16].So,if the etiologic fraction of arsenic dust could be known by histogenetic study in the total cases,the fraction from the other carcinogen i.e.radon and the ratio of relative probability Rn/As will be obtained on the assumption that the total probability (Rn+As) is equal to 1.0.For that purpose,100 cases of miner lung cancer biopsy specimens were examined^[15]．Through regular sputum cytology screening in YTC miners,27% were confirmed as early lung cancer,mostly with very small incipient foucus,which provided an valuable chance for the histogenetic study.

　　The results of histologic study and X ray microanalysis^[1～19] supported the chemical studies^[16]:the deposited less soluble As-Fe compounds could be dissociated and the local target cells will be exposed to the dissociated arsenic.So,in those cases the cancer focus associated with local ore dust deposition and fibrosis were supposed to be the typical arsenic lung cancer,totaling 11 cases.The probability was 1.0 for arsenic,and 0 for radon.All of them were peripheral lung cancer and mostly typical scar cancer.The total number of peripheral lung cancer was 44,and the other 33 cases were ascribed to the combined effect of arsenic and radon with different fraction,based on their histologic findings.

　　Among 100 lung cancer cases,56 were central lung cancer arising from large bronchus,lacking any dust deposition next to the incipient focus,especially those carcinoma in situ and early cancer with very small dimension and at very early stage.Since it was impossible to be explained with the direct effect of the deposited arsenic dust,the other agent,i.e.radon progeny,should be considered.Then the ratio of relative probability Rn/As for inducing all lung cancer cases could be estimated,based on the histogenetic characteristics of each case with a specific model^[15]:the result was Rn/As=0.76/0.24=3.2.

　　Histogenetic study and previous studies on epidemiology and animal experiments all have some uncertainties,but as a qualitative analysis all those results indicated that the contribution of radon was higher than that of inhaled and deposited arsenic ore dust.At least it is impossible for the ratio of Rn/As to be as low as 1/3.

　　5 Discussion and summary

　　ERR coefficient of radon induced lung cancer in Chinese miners given by US NIH report (No.94-3644) was based on the study of YTC miner cohort.The resultant value was apparently lower than that in ICRP Publications and the data given by other reports including Chinese uranium miners.This paper traces back some materials used by NIH in the calculation of ERR coefficient,indicating that NIH risk coefficient may be underestimated,and the following problems should be reconsidered.

　　Most miner lung cancer cases had worked as “carrying ore bag on the back” in small crude tunnels before 1950.According to the actual estimation of radon in the present small tunnels and supposed working time (7 hours/d,285d/a),NIH calculated the cumulated WLM during the time of small tunnel working.In fact,most of the time of ore carrying was spent on the way of underground walking and climbing;part of time was spent on surface;the site of radon measurment was distributed mostly at the deep of tunnel but not arbitrarily along the way;and the annual underground working time may be less than 285d,therefore,the NIH assumed WLM during this time may be overestimated,and the resultant ERR coefficent may be underestimated.

　　Radon monitoring started from 1972.Radon concentration differed greatly between the YTC mines.That difference began to exist possibly not from 1950 or 1953,but from the mid-1960's following the tunnel development,exhausted cavity enlargement and natural ventilation improvement.So,NIH considered the radon levels to be stable over 1953-1972 in all mines and used 1972 value in each mine as the basis to trace back their levels in whole 20 years.Such a way may lead to the distortion of proper proportionality between estimated WLM and risk of lung cancer.

　　Since exposure to arsenic dust is an important combined factor for radon-induced lung cancer in YTC miners,adjustment for arsenic exposure is needed to obtain the ERR coefficient of radon.The cumulative radon exposure was correlated with cumulative arsenic ore dust exposure,and both increased with the length of underground working.So,it was difficult to differentiate them separately.NIH used the RR of different Rn-As exposure groups to adjust the combined effect of arsenic,so that the original ERR coefficient was reduced to 1/3.But the number of cases used for some reference value was too small.The situations mentioned above will affect the validity of the resultant arsenic-adjusted value.

　　Many lung cancer cases were exposed to previous environmental arsenic pollution at mine area.But few people in the control group had such experience.So,using the data from that control group as the basic value to calculate the dose effect relationship of different levels of Rn-As exposure may affect the compatibility between the miner and the control groups.

　　Based on the studies of epidemiology,animal experiments and histogenesis,the author proposed that the relative contribution of radon/arsenic ore dust in the etiology of YTC miner lung cancer was about 3/1.As a qualitative analysis all of these estimations tended to show that the effect of radon was possibly higher,at least not less,than the effect of arsenic-containing ore dust.