Document kDRN7Vj6RzpvpDa4qrQ9No9JB
c 1
1 PLAINTIFF'S
i EXHIBIT f UO^SiD
J
An Ipidertoiogical Study of Lung Cancer in Asbestos Miners Zsiniel C. Eraua, M.D., and 7. David Truaa, AKA .Archive*, Juno 1956, Yol. 17, 3o. 6
Reriev by Tirana F. Honcuso, M.D. Chief, Division of Indus trial Hygiene
Jfovccber lb, I960
,`f flrmtare A iiOci.ar_m:
Revieva:
.^e catlye: ^ 'C
TV
rrrscairrG*.Y material
1. Early literature refers to cancer cases vhiea oeenrred
simultaneously vith eabestcala. In other uorda, reprieved thca
cases of esbestoais to find vhat percent of cancer (lisg) occurred
in this Group.
2. Essolc: Herevetber (195^*)
l6.0
Gloyne (1951)
1^1$
1, Eueoer (1952) states l-xT% of cases of aebestcaia, after 1} to
20 j asm latest period, develop lung cancer.
2. Itordarnna and Sorre - nninal experiments- produced lung cancer
in nice vhich wore exposed to eabestca dust.
1. 3auoe (1939), found no cases of Inn cancer among c2D cases cf
asbestos in be examined.
2. Hollsb (1S-U2), found the nusber of cases of asbestosis vith
-"`I US
sS''
)
Mcneuso:
tr
3. Venclius (19^7), in 1^6 x-rayed cases of nsbestssis among ^76, found no lung cancer associated in the group.
Lung cancer occurs in asbestos uoriiers vith and vithout asbeatoais. All studies vers related to cases of lung cancer associated vith asbeatoais and when this simultaneous requirement occurs, it is usually a untiller percentage. The Wegelius study nsJtes the ossurtption
Doll: .-'ancuso: 2raun: ftaneuso:
that for lung cancer to occur, tills would be only when
.
es'oestoais 13 present by x-ray evidence. This say or nay cot
be true. t_v-; cancer can occur in these wariers as a cause of
death, \rlthout asbeetoaia as a secondary cause. 3y considering
only the sinultaneous occurrence of lung cancer and asbestosis,
the'investigator ia automatically excluding a large potential
of c*eatra vttn .to diagnosis cf uabestasis, even though exposure
end duration of exposure have been, established.
Doll (i?55) studied records of men vbo warned for at least
20 years in exposed situations; found that in workers after 20
years of exposure, the risk i3 10 tinea as great aa general population
(Found ol cases of lung cancer-studied 'H ? men. vho bad warred for
20 years cr tore.)
In Doll's study, he excluded these vho died frog, other causes,
after shorter employment. It is evident that lung cancer eases
could have occurred arsons employees vith less tbaa 20 years esplnyoe^
exposure, depending, of course, on vbat is found to be the
number of years exposure required and the latent period.
Cites selective factor of Dr. Gloyne's series of l4.1J of
autopsy coses saying these cans oa part of unusual group of lung
eases to autopsy. Also cites percent of lung cancer at London
Cheat Hospital was 21.37?*
This could very veil relate to all the lung cancer-eases
vhich coce to the hospital, like ccuing to Dr. 0chaner`s Clinic,
so this port of the criticise is valid. However, association
of asbestos ia with lung cancer i3 not a bias that exLsta by
virtue of the Clinic-so observation of Gloyne ia still an
unusual su'd valid suggestive lead.
3 Lung Career la Asbestos Miners
l'
Dmun:
Refers to Klota vbo found only b casea of lung cancer la 173 coses of asbestos!*.
Initial observation la recard to the limitations of this latarprretatlan still bolds. Other cases could fcava occurred but not simultaneously. The alrailtsaeous requlreseat places a restrictive bias la actual situation.
f-PPTlCWT! by Eraua
:Jouree c: Data
L. 2roua used all persons vbo had been process ad through tbs
company elir-ic at Thetford Mines since 15^7 Is Quebec.
2. Established cohort, with total exposure of 5 O" rare years,
vbo vers oa enploytsect rolls In 155^. Tbs office force end
non-exposed vere not included.
Ffcncuso:
I question vbether Ersun is allowing suffioieat suaoer of years
for latent period required for lung cancer. If tbs eutaff date
is 1955, then the cohort latest period is only being tested
against a period of 11 years, 29^*5-1955; question whether there vl;
is enough cf this population in the cohort with exposure of 19 years
or core. For exaeple, if they (the mafrer of wortera with ip or
more years exnosure} represented only
of tbs cohort, the
expectancy of lung cancer would definitely be diluted b7 the other
The cohort followed through 1955vfcetber alive or dead, " >
be statee-ell but a smll number cl thtsc vero accounted for as
either living or dead." This vaa 133 out of 5,091- If tha 133 Vera
found and only 1 lung cancer recorded bad occurred, the edtira
conclusion of the Btudy would havs been different. Certainly a
very slight thread to draw such definite conclusions.
U. r Cancer La Asbestos Kiners
Emun:
1. Through Office of Vital Statistics tabulated lung cancer
rates by o for years 1950-1955, Inclusive.
2. Group policy coverage of one company studied for ar cancers
of lung ood Brnua states: "Very fev cot coveied by this policy-
cod those who leave industry cad are cot covered, ere a erall number
ftaacaao:
Eov "-any? This nay affect results. Oram states "fev cases
la vtiich lung cancer is strongly suspected but test proved as the
cause cf death wtb considered separately, "
Knocuso t
Deaths from ling cancer vere determined from clinical, rscords in
the radical service of industry and checked by rrarn of death
certificates and the Insurance cocaany records* However, he refers
to only the croup Insurance plan cOTerag* of indnstry.
Cohort-5,95<5, placed in three (3) categories according to
increasing decrees of exposure based on veighted average of yeara, ate.
yjzi cso:
Cable l--66i of cohort in 20-kk a^9 rrtnrs*
s'
{ Population veighted to tha young age group'vill not shew natal
incidence of lung cancer found In the older age groups, therefore, '*'
rates of this group as a total for c&aeer vill be lov asd biased,
a dilution factor cf considerable proportion* Severer, the age
specific rates for 20-^fc, if higher than the control, would bo
significant. Again, however, one rust dstez=ine hov may in this
group vorhed for 20 or sore '/ears since this is probably the lataut
period, otherwise the rate becomes xaarhedly diluted* The number cf
this group by the very nature of the ages, 2D-J&, would indicate that i\
the amber of wortera of 20 or core years would be very stall,
; too snail to use" statistically since the cohort vas only followed`for I.
11 years. Sow aaay vere there vith 11 or core years exposure in the
/ age group of 20-U4 comprising 6o> af the total cohort?
5. Cancer la Aaboatoa Miners
!*anruao:
rfcncuso: Msncuso: Itacusoi
Table 2. The .-.iar.ber and percent distribution by length of
-
ec^loywnt 3boved only 30 vlth 20 years or acre exposure; 70^ ld<-
leas than 23 years. If It tikes 33 or sore years to develop 1-^3 nance:
then the dilution factor is 7&p! the bag cancer rates should be based
on the ;p>* Sren should vo oooaider that 10 or rsors years am required,
then them is still TPp of the cohort vhich would not be
in
the population vbea the rates are *-ai~;0-i*wd. licterially, vould
Increase the rate*. Itotlee, however, la addition that the length cf
ersloynent
is
cot
rive^n,, rb 7
are
srcus.
"
Could be as Mgh as 5^ In
the age group vith lass than 10 years of crployrent. Doa't know
actual percent but represents a high H-firrfc.-trv factor.
Table 3. rx:rjnurc categories.
Again,in categories cot related to the age group* lihe 2D-H.
T"-'" 1 r- 1,
Habi*'-'*.
**
Sot misted to age groups*
^
pack a day, etc.
Table 5. rfunber cf Cases.
Definition cf sookess-5 cigarettes
TM ^ "*-*-* * It-
1V
-v
^
Defined proved cases of lung cancer as "diagaoais snggssted by aiztspaj, stsrgical section. Microscopic exaslnatisa." "Suspected csaee-3, diagnosis in doubt but evidence points to cancer."
Basic error here. In **- vi^g this type of "proved" case criteria of autopey or surgical section, vhich does cot apply to
the control of the general population, the ratss for lnag cancer la tba control are derived front lung cancer aa listed ca the death certificates. There is no indication that the sacs criteria vas applied to the general control population. Actually, tM vould be lcgjoeflibls without reviewing all ths lung eaoerr case* of the county cr
ra-establlshlng theIr entire clinical record. Seccaily, the g^ertH.
population nay not have Moan opportunity of r^r*-1^! care. **<. could
ward bcrch ways, excludes other eaae* without such confircatloa.
Controls include both confirmed ard ixcuanfircsed, usually ccncocfir==d
jrcup is larger because of cedi cal or available
care.
la discussion states "If ha uses the 3 suspected cases. Table 12
would be statistically aimpair!cant- "
Be rpes on to acolyte the 3 questionable cases because this
rakes the study so close to the significant level.
:aoi
Table d. lung Cancer Deaths by Ana Crouos.
There was ooJjr 1 case of the 9 proved cases ia-the 23-41 age ^cup.
Beca.il that this age group bed oca of the pcoulntlor. cf the cohere.
The rata for 23-44, per 100,000 core years cf saqpomrre, vu only
4; acd rate for 45-54 vas 15. (Comprised 17S of cohort).
-'
Kates ere based on very mail sunbers, oca each ia 23-44 and
45-54. Cannot xae total average rate cf 25 because of carted dilution
factor of aga 20-44 seta of 4 which ccegirised 6o of popalatlcu;
t
cost take age specific ccamorlscsa with the control-
Table 9. r'~~ eraser Peath3 by Length cf IhoLovsest.
Shows that caly 1 proved case was faad in the leas than 20
year length cf erploysent and this group evarrisart 70 of the cohort.
rates under 2D yearn length of ctployaeat were 7 per 100,000.
This again dilutes total rate and Lowers average rate to 25*
Bote that the rates were high ia 20-29 yr. length of erployrrat
Group which was 54, and the 33-29 yr. length of eeployrmt was 15/100,000 can emloyrant years.
Brava;
Discussion cf 2Uo aca over 4o yearn of age, followed for 6 years,
did cot develop lung cancer* /
Tnbls 13. Lmx Cancer Deaths by Sposure Catscory
states rates do not lacraa.se by catagory of exposure. alia is proof that asbestos Is not a carcinogenic agent."
"Thus
V.
Categsries, however, are not related to age distribution of the
population^ for example, 65^ of population in tha 20-11 ago group.
Do not xnov hov such a contribution Is rade to these various categories, I, II end IU. Hiis =sy ba the reason vhy tha rates did not rise with category Increase fron I to III.
Secondly, workers are so sra.ll, 2 cases in category IU; the audition, of 1 or 2 cases raxes sireabla difference. Table 11.
Categories I, II and. HI were comprised of 1,795 persons with employment only of 5-9 years and comprised 3&P of cohort population ELdd:.-.g tha ID-19 yrt. length of employment group gives 73^ of tha population. However, of these groups at least 66 must have been in age group 20-11; 1300 out of 1772 vere in category m-froa
employment group leas than 20 years; lU00 out of 2,150 vere In category II; 1300 out of 2,COO In Category I. It is obvious that
these categories are biased by contributions of different age groups end higher numbers under 20 years in these categories.
I
Therefore, one cannot use this approach because the categories
are weighted by a larger number in the younger age group; even in category HI which would be less lihely to have cases of lung esnee
occurring. Braun states "Average number of years of employment in each exposure category is almost identical." Ee does cot take into consideration distribution by age group.
ncer in .-.sbeatG* riinere
?nhl3 12.
itatea not a single cslm cf laag caocer ooeurred in 1,265
cna-arokera of both "proved" and "suspected" cases.
Tables 13,1^,15.
_
, 7ears
.
Stauaa -'Itb looker exposure average (2.3 f these archers)
and greater age (data does act show this) coe would expect the
rso-sockiag group to sfccv a higher rate If l;sig cancer vas due
to asbestos."
Actually ha docs cot show ia vfcat ages these cases occur la
the "saokers."
Table l6, Csgoarlaor. cf Cohort Ebpcriencs vlth Province of Quebec.
Gives -
Annual rate-per 100,000
Province
22.5
Cohort
(ZOriudiag Aabeetos workers)
Tbtai
33*3
Proved
25*3
~
Be gives so reason vhy tha suspected cases should cot be
included. This would be statistically slgaifioact since sobs
exclusion factors are cot applied to tbs cootrols. Secondly, be feels
that the general population la Dot studied as diligently as the dinars
and, tie refora, tha rats Day be low In tha control, laylying that
the difference, therefore, is ant read.
fact that the cohort did not necessarily show a progressive increase ia these years is cot a valid ergneat became variations occur depending upon the latent period: and, recalling the 6o ia socager ago groups, coe nay find that ia the next ID years there would be a rise ia lung cancer efter tha latent period 1s fulfilled. '
9 Cancer la Asbestos Xlaera
. Hanguibt m~:
Tabla 17 Brain use* a claaslficatlca of "prinary'' and "proven" a*
datemialas whether death certificats carried the Inclusion of autopsy or surgery, etc. Tills la not a valid cacpariaom There La act Lha a era degree of accuracy ia rocorllng data oa death certificate*. Table 13.
Gives ccrparisocv cf actual ood expected raraber of lung earner death* 07 age aaong asbestos oioers.
'Snrjwt
first, Braun uaea tie
sod 1955 province rates rather thsa the
year* of actual study frsa which the population was stalled-.
bed a higher rate, and with th* control higher, it beecaes lose
lively to plch 19 the difference is the cohort because cf the saall
ember* involved. 3h the 6p year* aad ever grot^i, there vers 5 obierred
deaths compared to the expected cf 2. 3e, however, points out sad
erfhas Its* that this includes 1 cf the "suspected" eases to indicate
what appears to be a suggestion cf statistical evidence. Be adds
that "it 1s rather lihsly that the rate for the general population is
understated In this age group for the obvious reason that the exact
cause of death In the very old ia not o natter cf the asm intensity
of interest as it Is In the younger person." Tils is sot true,
ns expects to irply that the rata for this group would be higher in
the control and, therefore, the differences observed by the cohort
in this age group would not rwin anything.
Brmiat
Si Tibia 18, aaym neabora of cohort have sot died frea btng cancer
at an age earlier than general population
What be has left cut 1s that tbs cohort consists of 6&i of age
Croupe of 20--U*, and k,COO out cf 5T30 have loss t-Sw 30 years
Lung Cancer i Asbestos Mlnsrs
l-'n.^cuao:
Braun: Maneuso:
experience; so, if the 2D-Wj group did rot have a oufTiciant latent-'
a
period, they v*xdd :ot bare lung cancers. Recall too ti*.t tha
v*
cohort was followed only Tor 6 ysar*. A to tha coceant about tha ZhO with oora than lO years of enpinyaent, this aay be c selective factor
witi the tardy currivors. Thera in sow question whether be really got
ail the ainers who retired, ttooa vho left the area, etc.
Discusslor>-cage &<7, Ers.cn naJcea tin case again oa the total rntns-oll ages, 20.3 per
100,000 for Canada. Coeporsa to 25*3 par 100,000 proven cases. He 33-3 for total cases iseluded the suspected.
feel that Table li, aga ruspectad, tend a to grove rather than
distrove, particularly vbcn rccogn 1 ting that 6e of the cohort in 20-^v
Coe group were in tha 72? under 20 years exposure. large populatiga cf
cohort without sufficient follow-up has eutacatlra.1 '!y diluted tha'
results, /tctually by oca specific ratee, the controls are lower then ~
the cohort study ea it no? stands. If cohort was fhUowd longer thin t *'
would shew np core strihinsly that lisitstiea cf tios period of follow-up
tofluesees abscace of cases.
to sot agree with stateseot he cakes that "tbs rs are ao isportant.
differences between the rotes for asbestos reisers oed general
population of Qss'osc and Oscoda. "
Esan proceeds to cccyare vith Halted States rotes which sxe
higher. This Is not valid because of mrkad Industrlallsstioa is
our country cceyarod to Canada. Pae o>3 of dlscriaslrm.
Uovl Speculates that about 5QtQGQ persons eaqxeed througtmt the \Krrld have 1,000,000 nan years of exposure. Since only lyO esses reported in the literature, gives rate 15 per 100,000.
1 1 liiT-a !*- tr--
/.jbcatOe .wln*rS
Iti vraa:
(coot 1 r.'Wri ) Ttt!% is ridiculous because rainy case* of V.r.g cancer never are
reported in the literature alrply because studies vers cot mde in the
aabcetca plant*. Trarpla; Halted States reports in literature are
not an accurate gauge as to ub&t he* or bos not occurred a* fax a*
fhnr.aa:
lung cancer is concerted. Table 22.
This table shows that lung cancer rate of the county iz vfeieb tbs
asbestos plants vers located bus rate 2 t tree tbs adjacent oouErdae, f
13.9 to 9*^*
is a valid eoeparisoa, certainly cakes a better
control. He baa be using the Previses of '^aebtt a* a cactral vthch
I believe includes icr.traa.1 vtlch raises the owra.ll rate cf the
Zrtua;
control. Refers to additional cases net tnclrrirrt in tho cohort, for^pxasale,
the 9
prior to ths beginning of study duo to lung cancer.
Vtmsa: -------------
2rnun refers ngain to "at least 7 cf 23 proved lur>g cencers wore I
sot accoepaalad by asbestaais." This is cot tbs point about dealing
i i rmly eith tbs histologically proven esses. I5a should have used os a i
: control* ths age specific rataa of tbs surrauodlog counties iustend
l*
^ cf tbs total Province ublcb oss weighted with erthsr Industrial areas*
|vhicb ruined the total rats of tha central.
If ha hod casparcd vith tha Adjoining counties on. on age specific tools* It uould buTS been statistically aigvlftcont because the ooe county vas already 2 tiara the rats of ths adjoining oocnty.
Els case is too vaah. Hrtira conclusion hinges on the additioa of
1 case. Ills oocdusiaos should have been either that be bad not
I disproved adequately that lung cancer la cot associated vith asbestosls
\ or that a valid ensver could only be daterainad vbtL ths oohort is
fnllovod for a longer period cf time, equal to tbs lstt .period.
300