Document QkzRB3VdVKNBBLV56g4pxV6ev

FILE NAME: Asbestos Mining (ASMI) DATE: 0000 DOC#: ASMI032 DOCUMENT DESCRIPTION: Unpublished Draft Report - Critique of BraunTruan Report [Limitations Which Lead to False Negative Findings] cue Bra01'"'11"31' Re? Flndi^5 crit a ----- r * short, -- . 1 9 5 0 ^ aS -- ,, of the C^r . , for lung aiV- x *he pe,ioa of i i r o the latent P*~ ufly A * requi^emen . x-^e time the yearS'I"t wTas t-->as - diCal toWleage_n^V tstudiie-s that although cer. i^rmed from many * neriod, ,, ,, - * ' "* occupational cancer can oc " *" *" ' ,, lung cancers ocCU usually the =-- * ^ w .it. an interval of 20 yea or 30 years or -re. ( - - - more. usually an Interval - ^ ^ study o chromate ^ ater an mtervat workers, the f e r i t y of the ^ The long latent _ o 26.36 years and ay ^ others in asbestos period has been observe y occurred after an workers, where ost of the lun, cancer cases interval^of study, the cohort was of those who were i h working in 1050 who had worked five years or mean a period of H years r.he cohort as a whole. < iod o the design of the study limiting th P In essence, the design ^ x t - i n limiting factor subsequent observation to ,evetely limit observations in the study which ^ ^ " t r ' i r . S T l l o w - u p p e r g ^ r e s e n t e d a fundamental ^ i n the study, which ~ 1C^^ 4-Vifa ji *a 1 r\mov\I*. wiaomri +v of the lung cancers. 2 This was evident to the investigators themselves-- as clearly shown in the designation of years of employment for each of the lung cancer cases, shown in Tables 6 and 7 on page 642 of the published paper. (Example: Excluding one case of 16 years, of the total of 12 lung cancer cases, 11 had an employment of 20 or more years, which me'ans that the i" rval of time for these lung cancers since start of employment was 20 or more years.) There is no recognition**^ a d m i s s ion made in the Summary of the paper that the design of the study had a self-limiting factor to make such observations less meaningful. In Essence, the design of the study increased markedly the probability of false negative find ings and i ^ e a s e d the probability that the investigators would come up with artificial, incorrect, low^association of lung cancer and asbestos exposure. Further, the investigators state that 133 members of the cohort could not be found. If the 133 were found and only one or two lung cancers had been located, the conclusion of the s t u d y even in its limited st a t e - w o u l d be in the direction of positive rather than negative findings. A . II. The second limiting factor relates to obvjfbus questionable and inappropriate classification of the cohort' by categories of degree of exposure into. I , I*1 ---- -- of this^classification of exposure is reflected i n ^ both the pre-publication report and in the published report.( ' On page 3 7 of the^pre-publication report, tables show the years of employment for the nine proven lung cancers and three ^0 t ^ w ^ ^ T " r ^V (b^ ^ >U^-7 ' r- ~ ' I' f n . ^ ' ? j, ^ ^ A r - A h ~<- yy yy^ r^ -r^ *f ^ "^ Z f W - n ^- ^ ^ ^^uA^-e./ 90 pt^7 Z - , + -J ^ 4+^*- - ^ ^ ( v ^ t s-? ZZZZ r-~~ r 4~y^ ^ _ _j ftu- i/ ~ ^ * /T^ [ VO^ w wA^7 |7 ( A ^ C4~] 3 "suspected" primary cancers of the lung, classified by exposure. Proven Suspected (A. Asbestos cohort Thetford Exposure Category I ^ Exposure Category II Exposure Category III 34 years 5J years 2^ years 32 years 15 years \ 22 years I 30 years "V / 33 years A 28 years * 42 years ^ \ L 32 years ^ ( 33 years ,, The table demonstrates that for all the workers (who had lung cancer) those with th/ ^ n g 7 s < ? a n d therefore ifite&tperiod of employment exposure to asbestos, p y e a r s a n ^ y e a r s in the a s b e s t o s ^ o r t , were assigned to t h e < W s J ^ a t e g o r y of Ex p o s u r e ^ Three of the total of nine proven lung cancer cases, taeeh had 32_years or more exposure employment, were classified as Category I. For Category II classification of exposure, the highest A. exposure occurring in terms of years of employment of a l W i K ___ ______________ . _ -- _____ **" * , . - and all lung cancers was y e a r s in the Thetford mines. All four 'f' m n g cancers had 22 or more years of employment; three of the four ** had 26 or more years of employment exposure. in the comparison of the f o lung cancers in Category 111 / T \ with C a t e g o r y ^ both asbestos^ohort and Thetford mines, the ( 5 / category ^ h a d equal or higher number of employment exposure years in ^hree' of the comparisons with Category^III. J T t is r e a d i l T ^ p a r ^ T t h a t there are b a s i c 6 w | > i n whatever process was used in the assignment of exposure categories, - evi- I 4 Anyone at all acquainted with work processes would recognize that the constant dust exposure each day, week, month, for 25 or more years for the miners represents a ^ remendous accumulation^ of dust exposure over those years. Even employment exposure of 16 years, the lowest number of years for any of the lung cancer cases, would represent a tremendous cumulative dust exposure. In industrial studies, the usual classification of employment exposure is 0-1 years, 1-2 years, etc. But never in the history of industrial medicine has dust exposure of 34, 37, 32 years been \ classified Category I, as lowest exposure. The contradiction in the classification exposure comparisons of the lung cancer cases in I, II, H I relative to number of years employment exposure, make the analysis of such data unreliable and^ 'J ? therefore the conclusions unreliable. In order for any one to really be able to designate work exposure!^i^iould^haveTbeen necessary to know by air sampling and measurement the dust concentration for each week, month, year for the 30 plus years, for each work area. Sinca^n^ such m e ^ u r ements^ were made, and in face of the great many years of e m p l o y m e n t ^ Category I, it must be concluded that the dust exposure categgoorriieess l made cannot be justified and are misleading, and that there iiss aa \ basic7weakness>tnd /tallac^ relative to the classification oflbhe rely upon this classification of exposure as the b a s i s J o r their conclusion of negative findings, because there was no%orrelation, no increase in death rate for lung cancer by increase in exposure 5 categories. Obviously, if the classifications are in error to G begin with, the correlations themselves are meaningless. ' i ^ ,> ... ' III. Another built-in limitation in the analysis of the data and therefore limitations of the findings, relates to comparison of data in the pre-publication with that in the published article pertaining to theC^ i l u t i ^ Df the cohort , as shown in the percen tage distribution by age and percentage distribution by length o f * employment. r\ \ In the 12 lung cancer cases, had employment of 20 or more years. Only one of the 12 had employment of less than 20 years. ^rove^ lung cancer cases h tion of employment, and seven out of the nine had 25^or mo re. years of employment. This means, according to their study, that the increased risk for lung cancer in their particular miner cohort, * occurs after 20 years of employment. The limitation factor is the dilution factor as shown on page ( Z p b f the preprint, Table 2-- 77% of the asbestos ohort and /67% for Thetford m i n e s `had less than 20 years of employment. It is obvious that a;dilution factor of 70, in the cohort of employ A ___ - ______ sa.--' ;V H~l-i 1-1.1 -u- ees, tremendously lessens the possibility'^ any observations pertaining to lung cancer, which as previously shown required 20 years duration of employment. *` \ k I i-S r,'.- i.) ) Another misleading classification in the study relates to the classification of heavy) smokers as those who smoked five ^gg - H -ift * * e U Ju s ' cigarettes or more a day, which is not the classification defini tion of the major published studies pertaining to cigarette smoking and lung cancer: Richard |ojJL--5(^o more a day; Wynder and Graham-- 35 or more a day; Sodowsky and Gilliam 41^o r more a day; Breslow-- 40 or more a day. It is obvious that the criteria for heavy smoking, stated as 5 cigarettes or more a day by Braun and Truan is incorrect and therefore the analysis that was conducted and reported in the publication relative to smokers (5 cigarettes or more versus none) are invalid. When the designation of heavy smokers is set at 5 cigarettes or more as was done in tsfeis study, it mixes up the truly heavy smokers, those of 40 cigarettes or more a day, with the moderate smokers and very light smokers. Further, there is no way to determine how the smoking his tories were obtained-- no information relative to age^^t. start of smokino, nor duration of period of smoking. --- -- -------- Again in Table 4 of the pre-publication report, the percen tage of smokers is shown to be: 6. Asbestos Cohort Thetford Mines Smokers 85% 7$% However, in Table (20) the lung cancer rate of the Cohort Asbestos "/r\ v-'v Miners is 22.1/100,000'while the Thetford Miners proved 27.3/100,000 The total suspected is 41.0. f\ According to the table, the asbestos miners cohort, with 85% smokers, had the lower lung cancer death rate, 22^1 compared to Thetford miners which had the higher rate for lung cancer of 2 7 ^ 7 with the lower percentage of smokers, 74%. According to these tables, the more you smoke the less is the lung cancer rate. They ,, # do not bring this out in the analysis or in the findings <hy-fr\-x^ft f'-'h j "J1-l Vv^-v^f-sj There is anlootthneer^lximiting factor, and this relates to the use of the "combinedL" the averaging effect when the two indi vidual mining facilities are combined in the presentation of the death rates. The comparison of Table 20 in the pre-publication - " report with Table 16 of the published report illustrates the problem of the averaging effect, where a higher rate is submerged in the average combined rate. In the prepublished report, Table 20-- asbestos cohort had a /' rate 22.1 proven lung cancer and Thetford miners had a rate of 27.3 and for the total lung cancers 41.0. When they combined the proven cases of 22.1 and 27.3, the combined rate became 25.3 which is less -- -v A " than the higher rate observed of 27.3. It is that average combined ' """ ' rramfeer which is compared to the Province rate of 22.5. I n _ e s s e n c e ,.-r* the higher rate of 27.3 is submerged in the average of 25.3. Similarly, the higher rate of 41 is submerged in the average * .. ~ /i of 33.8, which would be statistically significant compared to 22.5, v' This is mentioned because in the published document, it is stated that 33.8 is close to being statistically significant. Obviously, rates above 33 are going to reach that point. By using the combined figures as was done in that report, it lessens the difference between the observed and the control group, which is the Province rate. Whenever an investigator lowers j:he I numbers in the observed group, this of course lessens the chances of noting any positive findings. VI. Further, in the pre-publication, page 29, Tab^e Q ,) the " T , .. . . * 1 ^ ^ distribution^by exposure category of the asbestos raining cohort and the Thetford area differ substantially and therefore, any combination of the two would be inappropriate. This was actually done in the study. U v Exposure Category In Asbestos City Thetford Combined I II III Unknown 969 / 43% 564 / 25% 735 / 32% 5 x 100% 1062 / 29% 1586 / 43% 1037 / 28% 0 -- 100% 34% .y-7 36% 30% 100% ci/*r" It is noted that in Exposure Category I, asbestos miners cohort had 43% and Thetford miners had 29%. The reverse occurred in Exposure Category II: asbestos^cohort had 25% and Thetford 43%. When you combine the two mining facilities who have such marked ^^ferences percentages, by exposure classification, the inves tigators are actually mixing weighted populations in different ~ categories and then combining them as o n e . It is obvious that the use of the combined exposure percen- . -- **" .i.i..11^ tage is not accurate because it submerges important differences. Therefore, the exposure group combinations and designations are misleading. --1^ In the lead-off to the Braun-Truan Report, the authors try to convey the impression that the prior epidemiological work or 9 prior reports relative to asbestos and lung cancer had certain limi tations, and as a means of demonstrating that, they were going to overcome the limitations by having a total population that was followed over a period of time. Their critique in effect said that the prior publications, because they related to autopsies of indi viduals and groups of individuals, were dealing with selective ' groups and therefore, in their terms, their study would be much better because of the fact that they were going to take a total population and study that total population. What they did not say and what they did not include in their analysis was an analysis of their own data in Tables 6 and 7 of the published report. Table 6 refers to proven cases of the primary cancer of the lung; it specifies the exposu r e , meaning the years that the indi vidual was employed; it specifies whether the individual had an autopsy; and the column to the right specifies whether there was an asbestosis or not. Of the ITiri^ cases of proven lung cancers, there wereysevei1* autopsies and there were four cases of asbestosis, which gives you approximately a percentage of 57 percent of lung cancers having asbestosis. -------- :-- --------------- Moving to Table 7, and adding to Table ~ " ' (0 the suspected lung cancer cases and the additional autopsies, you find eight autopsies and five observed cases of asbestosis, for^a percentage of 62 percent What has happened here, and what Braun-Truan unfortunately prior studies of the relationship of asbestosis and lung cancer by overcoming the difficulties they had previously cited other authors had had, by using the method which they called a representative ^ 10 which they took, and they found all the lung cancers in their particular study and determined those who had autopsies of the lung cancers and found ( Q percent in one category a n d ^ i p e r c e n t in the other; 14"percent an\ 16 percent and 17 percent and 13 percent previously reported by ollher authors, which were significant in comparison to the percentage of lung cancers observed among individuals ex posed to silicosis. i U| ,r w./rr ` V v m . U - * * * * * * * - f ~ * * 'v 'r'*"~ 1 t ' n J V --------- - i * - ^ ' V - m j -j-- ^ . -- Itm . . n ~7i. - ^ ***** i r> fv^/br^ C ^ / U ^ f L~\ ^ W ^ - ~ ^ 0 ty^/~^Cl~~ '-'t-- iAs-i~ e*~f ^ 7 ^ "< C^~~i-~ -Ar^C^C^ ^ j ^ o u C U - c L ^ M ^ C ^ - > ^ A-'C. - "c^Zl^ _J'V^ jt^s A-'U.- -'-TM-- 1 26 Dr. Mancuso Direct 81 2 3 MR. HENDERSON: No. I think that we 4 w ant to finish it. You asked him about 5 Braun-Truan, and he has a response for you 6 that's as complete as he has it; and he 9 7 wants to do it and I want him to do it. 8Q All right. Dr. Mancuso. Did you observe anything 9 else in comparing the draft Braun-Truan Report 10 with the final published report? 11 A Yes, I did ( P 'm refe r r i n g now to the publ i s h e d 12 document of the Braun-Truan Report, and I'm 13 r e f e r r i n g s p e c i f i c a l l y to Tables 6 and 7. I'm O' -- a 14 v mix ing it within the following context and for 15 the following reason: In the Braun-Truan Report 16 in the lead-off of his document in the preparation 17 to try to convey the impression that the prior 18 epidemiological work or prior reports relative 19 to asbestos and lung cancer had certain limitation 20 and as a means of demonstrating that he was going 21 to overcome these limitations by having a total 22 population that was followed over a period of 23 time, his critique in effect said that the prior 24 publications, because they related to autopsies 25 of individuals and groups of individuals, he 1 27 2 3 4 5 6 7 8 9 10 Q 11 A 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Dr. Mancuso - Direct 82 implied and stated that this was a selective group and therefore, in his terms, his study would be much better because of the fact that he was going to take a total population and^study that total p o p u lationT*"but he did ^ t ) say and what h ^ d i ( n ^ include in hls_analysis was an_analysi5_ o l his own data on Tables 6 and 7. ^ Is that in the final report, Doctor? His final report published in the c h i v e s p i Tmlnfifrial Health, and in Table 6 it refers to proven cases of the primary cancer of the lung, it specifies the exposure, meaning the years that the individual was employed, it specifies whether the individual had an autopsy, and the column to the right specifies whether there was an asbes.os or not. in looking at that table, Table 6, showing nine cases of proven lung cancers, of those nine lung cancers there were i^vej,'autopsie^ and there V e r e ( t ^ j cases of asbestosis, which gives you approximately a percentage of S T j e r c e n of lung cancers had asbestosis. If you move down then to Table 7 and you add then to Table 7 the suspected lung cancer <=* Dr. Mancuso - Direct ~ ___________ ( n and you add then the additional autopsies, you find eight autopsies and observed cases of five cases of asbestosis, for a percentage of 62 percent. ' so what has happened here and what Braun-Truajn had ^fortunately neglected to emphasize was that he, 'eflfectT;, had $ r o v e n > n d confirmed all the that he had previously cited the other_aut h o r s _ jn"~which"ite said was a ^representative sample, representative popula_tion fallJthe asbesto s miners, which he took, and ^l^the lung cancers in his particular study and determined those who had autopsies of the~lung cancer s~~and~~four^*57^ percent in one category and ^ p e rcent in j ^ T o t h e r . Consequent ly this was much higher than even thej_4 percent anc 16 percent and 17_percent and p e r c e n t previ ous l y r e p o r ted by other authors, which were significant in comparison to the percentage of lung""cancers observed among individuals exposed to ssilicosis 4