A C 0 M P E N D I U M 0 F
E P I 0 E M I 0 L 0 G I C A L S T U D I E S
PART 1: SCIENTIFIC PAPERS
NJ
BAT Industries document for Province of British Columbia 22 April 1999

P R E F A C E
There are many papers which interpret the statistical association between
smoking and various diseases an a cause and effect basis, and there are many
comprehensive reviews of such papers.
This compilation covers papers which relate to the same diseases but
which consider factors other than smoking, and papers which reinterpret some
of the claimed associations between smoking and disease.
In all cited papers the views expressed are not necessarily those of
British-American Tobacco Company Limited or the compiler. The compiler would
like to thank all those who provided papers and/or were willing to describe
and discuss their research with him. The compiler would also like to thank
all those who helped in the preparation of the compendium and particularly
Mrs. Les Piper for her word processing skills.
C)
U,
BAT Industries document for Province of British Columbia 22 April 1999

CONTENTS
VOLUME I SCIENTIFIC PAPERS
1.Introduction
2.Epidemiological methods
3.Biological processes related to diseases
4.Smoking-associated diseases
4.1Lung cancer
4.2Other cancers
4.3Diseases of the heart and blood vessels
4.4Chronic obstructive lung disease
4.5Passive Smoking
S.Other factors which may be related to disease
6.1Diet and Cancer
5.2.Diet and Heart Disease
5.3Geographical and Ethnic variations in disease
5.4Disease in animals
VOLUME 11 MAPS AND GRAPHS, APPENDICES
1. maps
2.Graphs (Cohort Analysis
3. Cohort analysis (methods) CD
-Nj
00
BAT Industries document for Province of British Columbia 22 April 1999

PROLOGUE
Sherlock Rolmes closed his eyes and placed his elbows upon the AP= of
his chair, with his finger-tips together. 'The ideal reasoner" he remarked
'WouId, witen he had once been shown a single fact in all its bearings, deduce
from it not only the chain of events which Zed up to it but also aZZ the
result& which would follow from it. As Cuvisr could correctly describe a
whole animaZ by the contemplation of a single bone, so the observer who has
thoroughly understood one Zink in a "rise of incident* should be able to
accurately state aZZ the other ones, both before and after.'
Sir Arthur Conw& Doyle,
Me rive Orange Pipe.
CD
Qn
BAT Industries document for Province of British Columbia 22 April 1999

1. INTRODUCTION
Tobacco has a long history of Use and in some communities it has been
used in one form or another for several thousand years. Indeed, the history
of man's first use of tobacco are lost in the mists of time. The spread of
its use from countries such as Mexico and from what is now the U.S.A. dates
from the exploration of the .New World" by European Adventurers during the
fifteenth and sixteenth centuries. Once introduced into a country tobacco
use becomes established and it has been said that no civilisation that has
taken up the habit has ever voluntarily stopped using tobacco in one form or
another. The Introduction of tobacco has been reviewed in detail by Mangan
and Golding, (1984).
Although in earlier centuries there were one or two celebrated outcries
against the use of tobacco, and Its use was sometimes proscribed by severe
penalties, these views were largely based on social judgements and it was
not until early this century that the possibility that smoking was related
in some way to the incidence of certain diseases was first raised.
Subsequently many studies have been described in the medical and scien-
tif ic I iterature which claim a statistically significant relationship (wassoc-
iation') between the incidence of a particular disease and the habit of
smoking. This, In turn has led to judgements that smoking actually causes
certain diseases such as lung cancer.
However. there is a well-known scientific principle that without further
evidence one should not proceed to the conclusion that there is a causal
relationship between any two factors that are statistically associated.
Thus, the validity of judgements on smoking have been repeatedly criti-
cised. Independently-minded scientists such as the late Sir Ronald Fisher M.,
CD
(1959) have warned against the uncritical acceptance of the view that Smoking
CD
BAT Industries document for Province of British Columbia 22 April 1999

.2-
smoking causes the diseases with which it is associated and have drawn
attention to inconsistencies in the statistical data and on the lack of
fundamental knowledge of disease mechanisms. Nevertheless, over the last
20-30 years the judgement that the relationship between smoking and disease
is a causal one has become almost a tenet of belief in medical opinion and
is now widely accepted by governments and the general public, such that
legislation and ordinances affecting smoking are widespread.
However, in the period that has passed since the original judgements were
made there have been numerous developments in relevant fields of science and
mathematics, 9.9. epidemiology, data processing and biology. With the use
of powerful computers the statistical analysis of possible associations
between diseases and various factors has advanced considerably. Substantial
progress has been made in unravelling the complexities of the process by
which cells become cancerous and some knowledge has accumulated on the
processes leading to other diseases, e.g. genetic abnormalities leading to
emphysema have been identified. A book describing a complete model for the
development of many diseases has been published by Burch (1976) based on the
concept of spontaneous changes in cells due to their inherent instability.
Recently more conventional models for certain forms of cancer have been
proposed, based on virally or chemically derived alterations to cells,
reviewed in Nature (1984).
It is the purpose of this compilation to list the recent research which
might be considered as 'anomalous' in the sense that the reported results do
not sit comfortably with the causal model of the relationship between smoking
and certain diseases. it is acknowledged that there are many papers which
are interpreted by the authors on a causal basis, but these have been exten-
sively reviewed in documents such as those produced by the Royal College of
Physicians in the UK and the Surgeon General in the USA. Certain summaries
%O
Ul
1-4
co
BAT Industries document for Province of British Columbia 22 April 1999

.3-
from these reports are cited but the basic papers are not listed, nor is it
intended to weigh up the two sides of the arguments about smoking and health
or to make any Judgements.
'Anomalies' should, of course, be regarded as indicative of human ignor-
ance. It must be theoretically possible to develop a comprehensive model
accommodating all the correct facts, but, in the case of Smoking and its
claimed relationship to disease, the complexity of the problem and its intel -
lectual challenge are formidable obstacles to rapid progress.
Epidemiological methods are discussed briefly in Chapter 2, as a necessary
background to the individual studies listed later, while Chapter 3 describes
advances in the knowledge of the biological processes related to diseases.
Chapters 4 and S describe various aspects of disease. A standard format is
used for cited papers, the conclusions or abstracts given being those of the
original authors. The cited papers are readily identified by being printed
on blue paper. In addition, summaries from review papers are cited in
the text.
C:)
Un
__j
00
BAT Industries documentfor Province of British Columbia 22 April 1999