Document 70eE0qk8VjoM9KvwvpRqj67RB
(fjilo : T*?- ^ ^ ^ V-CJ0O-0/6
3*1, :?%h# H-003/sy-csy
'~.-i'L - t^' / ' :***y *
REil'TiC\Tui;y C.`..`,T";: .;;;) .'d..';;;.';; c.\i.S a relations!!:? i:,\;;!-.i) on j*;<oj'o::rio:;.*.tk Mii:;'i7,:.r;v /,::.\i.vs;;;
M. Gerald Ott, M.S., .'ifa j.'.rrin H. Holde-, M.n., Harold L. Cordon, ".D.
The Dov Chenirai Company Midland, Michigan 4SGA0
Send I'roofs To: M. Gerald Ott The Do\i Chemical Cor.p: ny Corporate Medical Pc part rent. 2030 Dow Center Midland, Michigan 4S640
00 -i ca^rnFNT,A,
AEST?-\C"
RESi'i itATf'KY ca::ci:k a.\'u i TwiTi. ::niiiAT.rr
Cordon, Harold I..
s::ni ca:.s : a rfi-attca'shiv base:! om
A'AI.YSLS. Ott ,
Corn It!, Ho! do: , Benjar. id 3.,
The ;>i nport i onatc mortality experience of i73 decedents exposed primarily to lend arsenate and calcium arnenot. was compared with that of 3.809 decedents not exposed to those compounds. A significant increase i_n respiratory cancer was found among the exposed employees. The relation:.hip between accumulated arsenic exposure and thr ratio of observed to expected respiratory malignancy deaths w*as estimated using a least squares approach. `Ihc_pred luted ratio exceeded six-tu-onc for individuals expose d to compounds containing an equivalent level of 1 mg/cm'' arsenic for a period of more than eight years. Tor the more heavily exposed individuals, a liability to respiratory cancer remains 25+ years after the initial exposure, observations based on the proportionate study wore supported by an analysis of the same target population, employing prosper t i ve rather than rctro~ spectivc methods.
Ki'.sri i:.Vio;:v c\:.u:: a:.: a:. :
a Rri.A'i ionsii 11'
o:; r-aii*finv.Ti. ::
.;:.m.v:;is
Arsenica. compounds have Inn" `ht!1 'utspi'.1 L ."I of ruusini, ( anccr la
hutnns,
"ost recently, l.ee and Frau:-, ni
: cp- rt ed an overal three-
fold increase in respiratory cancer a.-ioar, VKP) s-..]ter wer'.ers exposer. m
arsenic trio:-:ide. The increase ran a-, hi-.-h a; sixfold ar.eac e:..n 1 oyees who
were heavilv exposed to arsenic for 12 or r.ore ::iont:.s. ! ae authors ei e carefal to note that the influence of sulfur uinxiue or unideiit i I fed chemi cals vnreiny. concomitantly with the arsenii inposure eouls not he d i scoun ct-1.
In a number of iirirish and Cortina arti 'les, er/'es; cancer nns been re--
. i"', , 5/> 1 ,
ported to lie associated with e:-., osure to arse:. . .a I f.
l.uor, cancer
and clinical arac-'j. i - were found together ,.r..u:r f..--;-.an vineyard wo! .ters
who used Ins. . t ic in.a: .mt.i I n i up .tfenic in .; i ..; ' 111", and .Im.l.i... "I" 1 ,l! `
no tween JVli and ]y42. !n nddiiien, a si ;;ni:icant inerease in 1 one canrer
was found aIrene, i nbnh i ` ant - of tie. ..aw prnd.ie lay areas cf the Vosellc K:ver,
but not .ur.oiv.i residents of the wtne producinc areas of '`he Alir Uiver here
pcf:^ i c
co:;L;iinir.t: `"ir -onic vl'Ti.* hoc i:soii.
This investi;;atiou examines the proportionate mortality experience cl
a primp of industrial workers eiv.-,:'.;,ed ir. the production of i nsec t ir id'-s con-
t ni nine, arsenic. All available exposure informalion was incorporated into
the analysis in an effort to detenr.ine if the proportion of respiratory can
cer deaths was related to unnosure dosap.e. MFTi'OilS
The Fy.nosure larvirmewsit.. burins the perns! fror.'. l^U' through 195f>, a pro
duction unit of the eotpaay was eucaced in for ..mint i us; and packapinc a line
of i c.sc > i. i c i
cnntciiiiny. arsenic . The product line included lead arsenate,
lie:,p i : al arv Cfuneor an.! Aruen i ca 1 s/f't t el al 1
C0NFt0^TTAt
Calcium arucnuU', copper acct n- arse n i 1c .mi! nagm-s i t::. iir'.cinlc. 'll:'' lead
ars male was produced f res; 1919 through ],,y,. <_. i * I urn . i r :-< 11-> t ! i-ni.i i'l'.'U in
3953, copper accio-arseni Le fn-m
through l'.'ts, and nngnos iun ;ir:ii-;vnc
f row 1 9 1 9 I hrougli 19*Vl, The lift . 11 I. .\ei i of arwen i ra i prodm : ien over l hi *::[ i re
37 years was estimated to be 597; lead arsenate, 347 calcium arsenal.'.', 57!
copper ace to-arseni L c, and ?Z mag.'ios i urn arsenate. T!ie noil, .'n addition,
processed and packaged several other products, the most important of which were powdered sulfur and dry lime sulfur. The size of the work force varied from 30 in 192fi to approximately 3 on in ]9iS. Turnover rates were rather high, with less than 237 ol the men remaining w(th the unit for wore than one
year. Tliis was partially due to the seasonality of the week, and due to the
nature of the jobs. The pai haying jobs
red few skills and were often
assigned to new employee::. Transfer to hotter .(Ohs within the company oc curred raniuiv fur most of I h.-se e">p] overs.
The lead arsenate proce- s typified the processes used to manufacture
each of the other arsenical'. This process began by slurrying lead oxide in water. The slurry was purred into a mixing tank., where arsenic acid was added. Thin solution wn: then transferred to an adjacent '"lildine for drying, milling, and happing. In lf1-13, the packaging operations were surveyed, with 15 hreal hi ng zone sai.'s i c`. taken. 1 he readings varied from . JB to 19.0 milli
grams arsenic per runic meter of air, depending on the location of the samp ler. In 1932, the It i giies t exposures occurred in the area of the drum dryers where hreal. a i u:; non.- sample:', r.inred fris.i 1.7 Lo AO.S mill igrar.'S arsenic per cubic meter of air, and in the area of the packaging, machines whore samples
ranged from .26 to 7.5 milligram', arsenic per cubic meter of air. Excessive exposure was also noted during, cleanup operations, such as sweeping of floors.
Heap i rat ory dancer and Ar seni cal s/lii t et al 2
'n'TAI.
Jn addition, bi,;h lead concent rat ; on.. were founl in tin- production a:v.i. The 1952 measurements resulted in ph.nr. im*. out tin? lead .irucnat e ;>rin iv.s over rhe next several years. For analysis purposes, it was nccesearv to reduce the exposure information to a simplified structure, yet one . !;ich did not misrepresent the under lyinv. exposure patterns. The first ton's was to construct a matrix of joh c lass i f tent ions within the production unit by calendar years. This was done from annual personnel lists which enumerated the employees by job classification within department, and from tile work historv card.-, that arc kept for each employee. Over the course o: 3/ years, numerous modifications in job clans i f isi t ion t ermi nol opy ver-' encoun t efed. Fifty di f; erent job cities were identified, including those wiiieit were later found to refer to the same job. Next, job descriptions and exposure measure ments were used to combine the jobs into four exposure classifications. In ana; r.r.iar. t:..e Weighted avcwu;;e concent r,.Lie:..-, to the various proupinvs, the emphasis was placed on arriving at a proper ratio between the exposure groups The estimates shown in Table 1 were obtained through the consensus of two in dustrial iiv^ienists familiar with the processes, the available mdusliial h>pio.xe data, and ion descriptions.
The exposure concentrations experienced by employees while in (.roup I jobs were quite variable, and included tlu likelihood of brie: excursions to very hijjh levels. Respirators were available, but were not worn consistently Individuals on Croup 11 jobs were exposed re dust levels generally at lower, less variable concentrations, but over longer periods of time. Some omp ayit. for example, spent up to six hour!, a day porformir.v, tasks near the packauinp. mncliiner,. "t-v measurements were availanlo for (.roup 111 or Croup 13 joos nocause die expos-urc was apparent 1 '> felt to be of less concern rel.ntj.vo to the other jobs. Tiuis, the greatest subjectivity was used in cs t i ma t i tip. time
Kesp i rat orv Cancer at'*! \rre:> ica 1 s/''11 et a]
--ONh rDFNTT
weighted ti ver,'!)'/' conit*:;! r.Tl j tin;, fur liit/st* groups. Siuco no .id j us l
s witi*
m.ndc for tho wii.irini; of rusp ir-nL nrs , or lIk* fuel Lh.it production nccurrcd on
a soinovhuL interim" LLent basis, it is likely that actual exposure tiosap.es wore overstated rather than understated by tin.* above estimates* However, the i;;i-
precision in cue time wei edited average estimates would not be expected to lead
tc appreciable distortions in the relntiv** dosage exposure rankings since du
ration of exposure, which was determined quite precisely, varied considerably
from individm J-Lo-individ uni.
The dosage exposure to arsenic was calculated from, the following formula:
Dosage in milligrams = (11 T.) X B,
where i represents the exposure group;
13^ represents the number of work days on employee spent
, . . tit
.....
in Lne l exposure cncecory (s: wort me ttnvs nor montn
were assumed in adjusting for nonwork days) ;
T. represents the tine weighted concentration for on S-nour day in t!ie d*' exposure group;
and, B equals 4 cubic neters/8-hour day, the average breathing rate assumed for an 8-hour day.^ (The assumption of
4 cubic metcrs/o-iiour day may be conservative; 10 cubic metera/S-hour day is commonly used.) Study Design. for this study, the samplin'; frame consisted of nearly 2,000 employee'1 who were known, through company records at that m'nufacturing lo cation, to have died between 1940 and 1972. The employee work histories far the entire decedent population were examined to identify those deo dents who
Respiratory Cancer nod Arsenica 1s/Ott it al 4
DO 1 4i?66 1
173 won identified constituted tiio exposed popnl.it ion. The remaining dece dents, 1 css 13 who hod worked with ashestor., served os controls.
The expected proportion of deaths due to respiratory mni innanev in rela tion to year of death and age at death was estimated based on a weighted least squares analysis of the control population. The ]Rf)0 deaths awon;; the controis wore categorized by ten year ape groups and intervals of five calendar years within each ago group. An eight term equation, utilizing age and year of death, explained 5S percent of the variability between the categories anti provided a reasonable fit to the data in the region of interest. Table 2 shows resulting estimates of the expected proportion of deaths due to respira tory malignancy for selected age - at death and calendar vears.
For the decedents in the exposed population, the following items of information were accumulated: a nredicted probability tiiar death would be uue to respiratory malignancy; the observed cause of death; the career dosage ex posure to arsenic in mgs.; the interval between first exposure and death; the age at death; and, the year of death. The strategy of the analysis was to determine if a dose response relationship existed between the proportion of deaths due to respiratory malignancies and exposure dosage, adjusting the ohservations for the expected proportion of respiratory malignancy deaths.
I i 1
j 1 j
J
? **',*!'*"**'. ft'.
A! * _ , y. -
^
am'.
two groups after subdividing tho data into a 3 .'; A table of ages at dealo, by decade of death. TUti purpose of this procedure was to adjust for effects which night be associated with these variables. The normal appro:-: inne i on statistic for this test was 5.3 (P ^.001). Cancer of tiie 1vmphatii and henatopoietic tissues, except leukemia, also occurred more frequently than ex pected. Six deaths were observed ar-onp, tho exposed group versus 2.S deaths expected following the Mantel-linonsxcl procedure (P <.05),
In relating the percent of respiratory cancer deaths to dosage levels, the expected probability that death would be due to respiratory cancer was determined from the regression equation described in the methodology section. The observed versus expected deaths by exposure cat-, gory are presented in Table 4.
A weigh red least squares procedure' was employed ro describe the rela
tionship between the ratio of observed to expected respiratorv cancer deaths
and Lho natural logarithm of the dosage exposure in milligrams. This is shown
in Figure
An approximate test for the significance of the fit yi.lded
a P-valuc of <.01. Weights were assigned on a basis of the estincited vari
ances of the ratio for each exposure group. It is important to note tii.it, if the dosage scale had been expressed in months of exposure at 1 mg/cm 3 , the
ratio of observed to expected deaths would have exceeded unity at one month of exposure.
Tabic 5 shows, for tw-o dosage cat eyerie.'-.. tho observed and expected number of respiratorv cancer deaths by interval between exposure and death. Among tho decedents in the low exposure group, the excess of observed to ex pected deaths occurs 15 to 34 years aftei initial exposure; whereas, in tho high exposure group the cxcens persists among ind i Idua1s dying more than 35 years after their initial exposure to ars-'uic.
Kespi ratery Canter .<nj Arsen i ea 1:,/(U t ut al b
' :JNF'FNTfAi
Si '.Tj*9u ..*w
<.& *
-* 'tf?.'-
Kor the 2S decedents who tli_*ii from, respitory cancer, .1 regr.ion of
interval between in itj.il exposure and death, mi age of enplovee at first ex
posure , vns performed. Tin- age .it first exposure to arsenic w.is no,pat i vel y
ccrrol.11 ed with clio lag tine between cxposuri' ami death di o to respiratory
cancer. The: observed intervals between first exposure and death wvre tiien
adjusted for the age ;il first exposure effect and were plotted ai'.iiie.t tiie
carter dosage levels for c-awh of c.'ie 2S decedentr,. As seen fror;. figure II,
there is no clear association between exposure dosage and tiie interval be tween first exposure and death.
1)1 SCl-'SS .'OS
The positive dose response reiat ionsuip and the overall increase in
respiratory cancer serve to further implicate- arsenical compounds as a cause
of respiratory cancer. N'o other c nr.ro n donor; inat or was found among rhes-c
employers tc explain the ohuervat lens . The
,! papulation , xr.:.:i n.cd pra-
port i oca tel y ('id include four decedents vr.o bud worked wit!; asbosros subse
quent to their arsenic exposure. TVo of these Decedents died from. respira
tory cancer. The effects of includin.; thc-se four decedents in the exposed
group were felt to he iccr.atori.il to the cone -usions. Three of the eight
deaths that have occurred during K'73 were due'to respiratory cancer. These
result.*, were not included in the proportionate analysis. The medical records
for the respiratory cancer decode;!' were -eviewed with respect to smoking
and famlial cancer histories- No analysis of the incor-nlste information
found was attempted. However, in view of tiie partial informal ion avail able,
it was deemed unlikely that those variables would oe highly correlated with tiie arsenic exposure.
Mai 1 p;wi:it ueopl as:v: of the 1 v-p!,.i t : r and hv.-i.l tope i et i e tissue: , except
leukemia, was tiie only other category for :ch t:. expend t-roup di-'fered
Kespiratory f icor ar.d Arsen i, a ! .,/0| t s; .,1 7 1 ^ Tr.'FNT Tfi,i
from the controls. 1 In' ratio of observed to oxpetted deaths for tin's cate-
gory was 2.1. Specifically, tin- sir: cases of lyitpimM vert- classified on the death cert if i cat os an one iynphob las toon , one retin.-ln.u cell sorcomo, and four cases of Hodgkin's Disease. N'o clear dosage relationship could ho draw-n from those observations, nod no association could ho found in the lit
erature relating lymphomas to citron if chemical exposure. The data are so.v.imarixod only ns a clinical observation hy tin- authors.
It was possible to follow 5S0 of the insecticides employees utilising
prospective study in retrospect techniques. Total observed deaths were
9 percent higher than for a rorr c.sm.nding popnlat ion of company employees, hut 11 percent '-elew the number of death.1 expected based on a corresponding U.S. white male population.
On the other hand., respiratory :.nl Ignnncy death rates were 3.3 tine:; high,.!" tV> for other compn--y `-pioy-.-V", and ten time-' i.iy.h.er at.or.p, more heavily exposed e.-uloyecs 35 or more years after initial exposure (7 ob served verse; .7 expected). These results did not conflict with the pro
portionate results to any significant degree, hut did indicate -that the
number of death:, due ro all causes exceeded the expected number by a rela tively small number.
Until the early 1950's, arsenical pesticides had boon sprayed on tobacco
crops in the United States. The use of these insecticides has led to con-
(0) rumination oi cigarettes with arsenic, Satterleo ' suntmar i
the probi.m
in 195b. Tits arsenic content of cigarettes had risen from 13.0 micrograms
per cigarette in 1932 to 12 mi crograrts in 1950-51. These level., revile lead
to rather high concent rations of rscu-.ic in the lang, during, smoking.. in re
cent years, arsenic level-, in ,\-erir:ut cigarettes have ofr ,7.7, ir.i. crop, ra-s per gram of, toe.icco. (10)
, lined Le an average
kospi ra : ory Cane'.:- ,md Arson it .il s/nt t ,-t il
npN-rr 4i
1,1 Peruvian tolioroos, arsenic love]:. were found to Ik- 22 nicru.-.r-.n:: per
ir-'!CUf- as ruconcJy TM
Satterieo a 1 so reported that background
rnf.nc levels of 10 to 40 microprnns per cubic cutter of air were found Jn ,\Vv
ULy- Tfu::: arsenic is a widespread env i r-.ncu-nt ;i I contaminant which, in
:ew of the results of this study, deserves considerate investigation.
I l
Kespir.uory Cancer and Ar-ea :cal s/ott et al
K ,i- */.v`
Tabic 2. Expi'Cici! Proportion i>f ac:;p i r.iLory Malignancy Dcat.li:; !!y Ape nii'i ls.it f "f Death (P..isc`i on Company Control Papulation).
Year of Death
1950 1960 1970
30
.006 .006 .004
Ap.f at Oca tn
40 50
60
70 80
.024 .034 . 06K
.033 .050 .105
.034
.026
.054
.046
* ] 17
.104
--
.010 .026 -065
DO 1-16 rS CONF T OFMI
WIW
Kxposcd Group mil C'"-.! rn 1 ;, 1 *.*- 0--1^72.
C/UIMC Of IK'.'ltil
Al] Cnuseu Mnlip.u.uit Nuop];i;:ins, Tolul
Digestive 0rj;;ins .4 I'critnnoii!:>. ties pi r.n cory Sys 10:1 Gcni^.'ll Or;.; .-ms Urin.iry Or;;.in:', Lyr.-diatic 4 l!rc.'UO?oK'r. if TImum.--; ,
< Nm-Im-!- ^
173
t 100.0
<!nn; 1-0 1>. iVrcrnl
i IS 09 j ino.o
37 32 9 374 20.7
10 5.8 12S 2**: 16.2 102
7.1 5.6
2.3 36 2.0
1 .6 13 .7
3.5 25 1.4
F.xct-pL Lcukc.T.in l.oukei'ti n Al] Others Discosc^. of Cnviliov.'ocuJ.ir Da senses of 'sospir.ntory Sente;-; Disenses of Digestive System lixtcrn.nl In.jury All Others
4>. 1 .2 18 1.0
6 3.5 .82 2.9
SJ
46.8
201-
55.9
11 6.-'. 73 4.0
7 4.0 59 3.3
12 6.9 200 11.1
b 2.9 91 5.0
00 1 a.'1/-,
r.ONFT DFM
Tab 1 e 4. Observed and K::pec L ed Heath;: Dm1 m Kcspi r..t cry Mai i jpianc ios by r.Kpo:.ure date;-/:
1VA Concentration
X Months of Exposure
Avornpe hi (dosnr.e)
mi
Isur'iior of Total Deaths
Pesoi r;i L irv M.il iru.inrv 'i-atiis
j j )bserved
|
Observed i.>:pccU*u j patio l.xpec ,ted.
I
<1 1 - 1-9 2 - 3.9 4 - 5.9 6 - 11.9 12 - 23.9 24 - 50.11 60 - 95-9
96+
3.74 4.84 5.5^ 6.04 6.6S 7.35 8.17 S. 78 10. 30
?6 1 1.78 .6
]7
n 1.02
2.0
z4
4 1.43
2.8
22
3 3.42
2.1
27
3 1.79
1.7
IS
2 .95
2.1
n .75 4.0
13 5
.79
6.3
l:.13 5
6.4 i
U 0 t Q. ]*> -- CONFTOFfJr
no <-ONFr ofntta,
Ill !1!.IU(,;;.M'|!V
1. Kuchnnnn 1962.
: Toxicity of Arsenic Compounds. A1-.- Jersey, Van Nostrum!,
2. I.ec AM, Fraumeni JF Jr: Arsenic and Respiratory Cancer in Man: An
Occupational Study. J Sat Cancer Inst 42:10-5-1092, 1969.
3. Hill AH.Fnniug HI.: Studies in the incidence of cancer in a factory
handling inorganic compounds of Arsenic: I. Mortality experience: in
the factory. Hr .1 Inc! Mod, 5:1-6, 1948.
4. Robson AO, Jell life AM: Medicinal arsenic poisoning and lung cancer.
Brit Med J, 5351:207-9, 27 July 1963. 5. Roth F: The sequelae of chronic arsenic poisoning in Moselle vintagers.
German Med Monthly, 2:172-175, 1957.
6. Roth F: Bronchial cancer of arsenic-poisoned '.image s. Virchow .Arch
Fatii Anat, 331 : 119-37 , JV5R. 7. Fatty FA: Industrial Hygiene and Toxicology, Volume 1, General Princinlvs,
2nd Rev. lid., New York, 1 nt or sc it-nee Publishers, Inc., 195S.
8. Mantel N, llaneuscel * ; Statistical nspe- ts of the analysis of data from
retrospective studies of disease. J Nat Cancer Inst 22:719-748, 1959.
9. Sattcrlee 118: The problem, of arsenic in Amo r-'car. cigarette tobacco.
Now ling J Med, 254 :1149-5-'., 1956. 10. l.cc BK, Murphy G: Determination of arsenic content of American cigarettes
by neutron activation analysis. Cancer, 23:1315-1317, 1969.
11. Arata A1 , Quispe PL: Concentration dc arsenico on t'obneos v cigarillos
Perunnes. Rev Facult (Juirica, 1 7:15-22, 1965.
00 1.4;^ 7"' OONFTDFNTTAI
FIGURE 2. INTERVAL IN YEARS BETWEEN INHIAL EXPOSURE AND DEATH
DUE TO RESPIRATORY CANCER BY DOSAGEADJUSTED FOR AGE AT FIRST EXPOSURE
r> 5 5* ^V ;>