Z;~~
SUMMARY
The prcnent study was undertaken to exploze the
relationship bet-
ween smoking and mortality in Sweden. The
investigation wac a 10
year follow-up based on a narir,"wiA,~ prohnbility
sample comprisir,g 27 342 males and 27 732 females
between 18 and
69 years of age in 1963. The specific aim was to
make an epide-
miolocical evalulation of the relation3 bctwcen
different z~-cklfiy
habit; and subsequent cause specific mortality and
cancer morbi-
dity. included in the study were also three social
covariables
that were assumed to be related to smoking habits
as well as to
disease expericnce, naraelv alcohol probli~ms,
iir~~n-rltrO rnsf-
dence and income. The study was carried out upon
the-request of
the Central Bureau of Statistics and the Swedish
Tobacco Company.
Data on smoking habits, urban-rural residence and
income were
furnished by the Central Bureau of Statistics
which, upon the re-
quest of the Swedish Tobacco Company, carried out
a mailed ques-
tionnaire survey in 1963 on the sample mentioned
above. Data on
cause of death were obtained from files of death
certificates
and ddtA Mftft"inq cancer registrations were
provided by the
Swedish Cancer Registry. Data on alcohol problems
were obtained
from a nationwide Alcohol Registry which keeps
records concer-
ning misconduct in connection with alcohol use.
The general statistical approach was twofold. on
the basis of
previous reports in the literature a series of
diseases could be
postulated to be smoking related. For these, the
statistical ana-
lysis aimed at verifyingf such a postulation. Flor
other conditiond,
a screening procedure was developed that aimed at
detecting
whether these were significantly related-to
smoking. Fpr both
procedures the results are.presented in sex and
age spdctfic
groups as well as in summarized tabulations
showing age adjusted
risks relative to a standard group constituted by
nonsmokers.
Whenever possible, smokerz; are grouped in ~regard
to type of to-
bacco, amount consumed, duration anq, ~Lcje. at_
st4rt. pf
AnaiY804-4irb also made La order. to study the in-
fluence, of a).cohol registration, urban-rural
residence and income.-
Vhe,b&dkJrouhd of 'Ehe tu:ty and a brief'rundown
of' previouq lite-
kature are preseated is Chapter 1. The saidple is
describe4 in
Chapter 2 and the methodology in Chapter 3.
The results are prq~,entedin Chapterg 4-6. Chapter
4 in composed
of data on smoking in relation to total
mortalilty. lung cancer
ah& ivcbemic heart dJse'.A#e,. C - -e 5: deals
w1th,_dataon
r._ amoklog
in reiatxon to diseases for which a prior
hypotheses could be
postulated. In Chapter'6 data are analyzed in
relation to di-
'00 seasas for which no prior hypotheses could be
postulated but foe
which statistical associations between smoking and
mortality-
were, found in the screenixig procedure.
Throughout Chapters 4-6 *.detailed ulescAotion of
-the. resultw
22 M
302058614
BATCo document for Province of BritiSh Columbia 19
April 1999


di-.cu2sinn. A compariqon iS ~1-o made Wltlh
renlilft~- frnm the -,t.11(iy
by Hammond in 25 U S statos, the only larqe
stticly nt thin kinEl
which includes a suh!;tantial numbor ot r.,,t only
m~l,~s but al~;o
females.
The data showed beyond doubt th3t a clearciat
relationship exists
betwcen smoking and toLal mortality, lung canver
and ischemic
heart disease. The relative risks t:,howed a clear
association with
amount smoked drid, a,- leai;iL total -3 liinq
car-cer.
a consistent dependence on duration of smoking,
and in some gub-
groupings on age at start of smoking. Pipe smokers
had somewhat
lower risks than cignretts --mokers exce.pt for
lung cancer for
which the risk was just as high, while pipe plus
ciqarette smok-
ers had even higher risks than cigarette smokers.
Smokers of
little cigars had increased risks of total
mortality and lung
cancer. 'rhe riskt. L, j:egard to total
tft--~,tality wa-.re lower than
those for cigarette smokers) but were in regard to
lung cancer
just as high. No association between smoking of
little cigars
and ischemic heart disease was found.-
Former smokers were still at risk in most
subgroupings and these
risks were not substantially lower than those for
current smok-
ers. This was seen exclusively in former smokers
who had smoked
more than 20 years and had stopped smoking during
the ten years
iftediAtely priorr to the start of the study.
The risks among females were on an average lower
than those among
males. When amount is taken into consideration,
however, the
risks aModq- fftmles were just as high as among
sales.
The results of the Swedish study agr-ee well with
the, H&ftond
study as concerns male cigarette -- kers, when
amount of smoki-lig
is taken into consideration. The appreciable
amount-spedific
risks for females seen in the Swedish study have
not been obser-
ved in the U S study. Further, pipe smokers in the
U S Ztudy
were generally at a iow riskwhich also is ift
sharp contrast to
the results of the Swedish study.
Besides total mortality, lung caneet eM-ischemic
heart disease,~
-srr.~*retl dinoasWa utere cQrfi=ed to_ 1?e
ass9ciated with smok-
ing. These diseases includO-cancer of pancreas
a~:'c'aftcer of-th"
bladder p,, bronchitis, peptic ulder,' cirrhosis
of the, -liver and
suicides.. For-other diseases a sta~tistit!al
Oonfirmation was not
achieved, these being cancer of 6iccal'cavity -and
'phakynx, cer,6-
brovasculat disease, nonsyphil~tic'-aortic
aa6urysm, art,erial em-
bolism and thrombosis. For still other causes of
death, signifi-'
cant relative risks showed up in =me subgroupings
but not In
others, rendering the overall piota" unclear.
Causes of death,
belonging to this third caLteqo~yjncluda cancer of
lary". eso-
zLnd - rt=mch;- -hy1po*-t"-,Avs- diaease,
te&'iQsqj ~~o is Vid
ax q -a
pulmonary embolism.
Among dise4ses with no prior 'hypothesis about the
relAtion to
smoking, only four turned out to be significant in
the screening
procedureg namely cancer of the.breazt, canqer of
the cervix,
alcohol psychosis.~and
pneumoii~a.-!tlho~',assoolatioii between widkiRg-
-whon d
--dO4.n=04 at4 frcq
302058615
BATCO document for Province of British Columbia 19
April 1999


ir. fitiy-ly c~7r~mTixin, thr~ rrudf! rekt-r
h(ting 9 per 1000 in the
age groups 111-39.
As for the relev,ince of ~_he social --irialAer,
stuilicto IL should
first uf all be pointe,l out that by and large the
same associa-
tions between smoking and excess relative ri.2ks
are seen in the
covariable subgroupn as in the total sample. It is
alno apparent
that fur several diseases, including total
mortality ar;,) ische-
mic heart diRoa~o, th~ i3i-r se effeut c~f thr-,
~r-ioql
is substantial. For example, a nonsmokers
registered in the al-
cohol registry had the same risk of total
mortality aq had a
current cigarette smokers who was nonregistered. A
nonsmoker in
the low income group had about the same risk as a
current ciga-
rette smoker in the high income group. The
association with
urban-rural residenve status is generally not
strong. Exceptions
are cancer of the bladder, cancer of the cervix,
ai)d possibly
also cancer of the pancreas, those living in
cities having con-
siderably higher risks than those living in rural
areas.
4114 .44' 0
BATCo document for Province of British Columbia 19
April 1999