-DRAIT STRICTLY CONFIDENTIAL
Position Paper - For Discussion within Industry
Only
Reduction in Tar Yields
The weight of medical opinion is that
epidemiological evidence suggests
that a reduction in lung cancer rates is
associated with past reductions in
tar yields and this has provided the basis for
regulatory/advisory pressure
to reduce tar deliveries, subject to the premise
that a practicable rate for
this reduction must be related directly to the
rate at which consumers can
adapt to decreasing yields, otherwise, consumer
resistance will negate
the benefits that could be obtained.
Individual Companies are conducting their own
investigations into the
ways in which consumers smoke products of
differing tar yield and their
responses to changes in yie 1d. The ISC may be
kept informed by the Companies
-of the findings from such work and how they
relate to possible rates for
subsequent reductions in tar yield.
We assume that further progress in this area will
continue within the
framework of voluntary agreements.
Despite the significant reductions in
sales-weighted 'tar' achieved to
date the ISC is showing interest in changes in tar
quality as a further
route to product modification. While this may have
an apparent scientific
logic, the great body of knowledge of which the
Industry is aware, must cast
doubt upon the immediate feasibility of such an
approach. The problem is,
of course, the development of a rational
foundation upon which this approach
could be based. The ISC still appears to have
faith in the analytical
approach whereby culpable materials in smoke or
condensate are identified
and then removed. As will be discussed below,
neither arm of this approach
has proved successful in the past nor is there
promise of hope in the near
future. It is sufficient here to recall that a
substantial proportion of
the work of the TRC laboratories was devoted to
this end, via condensate
fractionation studies, as was extensive work in
centres in France, Germany
and the United States. This effort was able to
demonstrate that a
BATCo document for Province of BritiSh Columbia 29
October 1999


proportion of the tumorigenicity to mouse skin of
cigarette smoke condensate
could be concentrated in certain fractions of the
condensate. Not all
activity could be accounted for in this way, as
some remained spread over
a number of fractions, nor were any leads obtained
to indicate how the
limited knowledge obtained could be exploited in
product improvement.
The most encouraging indications for the
qualitative modification
of tar lie in incorporation of materials such as
reconstituted tobacco sheet
and mid-rib stem and related and desired
techniques. Progress in this
direction must be based on a sound knowledge of
tobacco production and
processing and is thus clearly within the compass
of each Company's
commercial activity.
Dr Froggatt has expressed an interest in the
semi-volatile
fraction of smoke. This fraction contains a number
of compounds thought to
be responsible in part for the desirable
organoleptic properties of tobacco
smoke. This fraction is thus of importance in
controlling product
acceptability and smoker behaviour. Dr Froagatt's
interest could lie in the
potential for manipulating flavour characteristics
to enhance the
acceptability of low yield products. The
semi-volatile fraction as such is
not known to have particular toxicological
properties. Although its
relationship to other smoke fractions will vary
according to definition and
methodology, it will tend to be associated in
experimental work to the
particulate phase and to certain compounds which
have toxic properties
when tested in isolation, but this knowledge is,
as discussed below, of
very limited practical relevance to matters of
Smoking and Health.
C=)
The ISC has also (in TA 614) stated that it is
imoortant to reduce
the Imutagenicity content' of tar. This ill-worded
request could mean
BATCO document fo r Province of BritiSh Columbia
29 October 1999


3. ity measured using
two things. Firstly, the ISC may be referring to
act:v
bioassay techniques sensitive to point-mutation in
bacteria, such as the
Ames Test. Such procedures are most frequently
used to give predictions
of carcinogenicity, and as such represent a small
fraction of the array
of bioassay procedures which may be brought to
bear. Their utility is a
matter for scientific debate at the technical
level. Secondly, the ISC may
be concerned that tobacco smoke has mutagenic
properties per se, and this
i=lies a wide new area of interest. The study of
mutation at the population
level is complex and fraught with semantic
uncertainty. A number of aspects
of mutation, acting in the relatively short term,
have been included. in
studies of the teratological aspects of smoking
and it is known that the
immediate effects of maternal smoking on the
progeny are restricted to
effects on birthweight. (As an aside, however, the
Industry should be awarp---.
of current developments in behavioural or
developmental teratology which
are being taken to indicate possible effects of
smoking on stature and
intellectual ability in childhood rather than
infancy). A wider interpretation
of mutation, involving the establishment of gene
mutations (other than
dominant lethals, obviously) or chromosomal
mutations into the human gene
pool, is quite another matter. The only evidence
which has been found in
this area is a number of equivocal reports that
smoking may be capable of
inducing the type of change in question. This is
far removed from a
demonstration of altered gene frequencies in any
human population.
-Detection of 'this would require a highly
sensitive epidemiological approach.
It may be added that there is virtually n o
knowledge of any environmental
agent having effects of this kind.
Uli
C=)
cc
BATCo document for Province of BritiSh Columbia 29
October 1999


Components other than 'Tar', Carbon Monoxide and
Nicotine
The ISC (TA 614), under the above heading,
advocate the gathering of
information on single components of smoke,
specified and unspecified, and
certain forms of biological activity. Their
specific suggestions will be
considered below, but first we must consider the
general value of this
approach.
The committee wished to "seek the Industry's help
to examine the
role of and "... recommend that steps should be
taken by the Industry
to obtain more evidence on these matters". They
are not actuiny clear
about what information they want or, more
importantly, what it would be
used for. As our response to such suggestions must
depend on what they
entail, we must speculate on the possible
applications of the knowledge
which is sought. Possible uses are:-
1. a contribution to a more comolete knowledge of
tobacco smoke
and its properties
2. a toxicological assessment of tobacco smoke
from its
constituents
3. provision of further bases for brand ranking
4. provision of a basis for product modification
and these can be considered in turn.
1. Provision of knowledge, separate from that
related to points 2-4,
is of doubtful concern to the ISC and clearly an
irrelevance to the industry.
while it is true that future advance is sometimes
based on the consolidation
of accumulated knowledge, thisis too uncertain a
postulate to permit the
BATCo document for Province of BritiSh Columbia 29
October 1999


acquisition of knowledge for its own sake to be
anything other than a
frivolous enterprise under our circumstances.
2. A toxicological assessment of tobacco smoke
would attempt to explain
the incidence of diseases associated with smoking
and the range of
phenomena revealed in laboratory experimentation
on the basis of the
composition of smoke. it is possible, but not
likely, that it might also
detect properties not hitherto suggested. To be of
practical value, the
results of such an exercise should have an
application other than the
purely academic, and this could be directed to
either of objectives 3 or 4.
Pursuit of such a toxicological assessment could
take place at a
number of levels. At the simplest, aLnd perhaps
most, -appropriate, level, -
the properties of compounds identified in smoke
could be determined from
the literature and the significance to any
associated disease state
assessed from the quantities in smoke. This can,
of course, be done now
and requires no additional contribution from the
industry. The problem is
that the great body of work conducted in this area
to date has been
fruitless, and there is no indication that further
endeavour would be more
profitable. Since the identification of polycylic
hydrocarbons in
cigarette smoke in 1953? it has been known that
smoke contains compounds
which in appropriate biological test systemshave
effects related to disease
in man. Yet in the subsequent period of nearly
thirty years, it has
not proved possible'to give a quantative
explanation of the incidence
of any disease in man on the basis of chemicals in
smoke. Although
chemicals with the appropriate properties have
been found in smoke,
CD
they are present in insufficient quantity to
account for the QJ1
_;-j
observed activity. The reason for this failure is
not inadequate CD
BATCo document for Province of British Columbia 29
October 1999


toxicology, but simply that the effects of smoke
are not due to single
compounds but to complex interactions and
synergisms between a wide
range of components. Smoke contains for example
chemically diverse
carcinogens (polycyclic hydrocarbons,
nitrosamines, nitriles, metals,
mycotoxins etc) in minute quantities, together
with initiators,
promotors, co-carcinogens, anti-carcinogens and
tumour accelerators. The
toxicology of some is so poorly known that they.
are placed in different
categories according to the tests used (for
example, phenol has been
classed as both co-carcinogen and anti
-carcinogen). The net result of all
these compounds, whose quantities will combine by
addition. and.. by
multiplication, and positively and negatively, is
impossible to predict
by any technique available now or in the
irmnediate future. The difficulty
of this approach is enhanced when it is realised
that epidemiological
evidence now suggests that all diseases.
associated with smoking may also
be related to other environmental factors. Thus
the synergisms of smoke
components extend beyond smoke and into a diverse
range of totally
unknown chemicals.
This last point provides the final blow needed to
destroy the flimsy
Justification existing for another proposed route
of exploration of tobacco
smoke toxicology, namely, the 'seeding' of smoke
with pure chemicals.
It has'been suggested that the role of chemicals
in smoke, including
possibly a limited range of synergisms, could be
explored by adding the
chemical to smoke i.n the hope that, 'in a
suitable test system, the added
chemical would give an added response thus
allowing the identification of
C"D
its specific response. A number of problems attend
this approach. We CD
I-
have already noted that the problem of relating
disease to composition is ~~-j
CD
C-TI .
quantitative not qualitative, so that any changes
in quantity are Qn
immediately unrealistic. This is reinforced by our
knowledge that in
BATCo document for Province of BritiSh Columbia 29
October 1999


past work of this type, small additions are almost
invariably ineffective
and large quantities must be added to induce a
response. The relevance
of such results is obscure, as they tell us
nothing about the quantities
actually present. It should also be noted that
such experiments are
seriously hampered by the lack, in many cases, of
a suitable biological
end-point with which to assess any changes. While
models of varying
quality exist for carcinogenesis or teratogenesis
no reliable bioassays have
been develooed related to functional lung disease
or the range of
cardiovascular problems attributed to smoking.
The foregoing does not deny the real need which
exists to establish
a detailed connection between tobacco smoke and
disease states, for not
only will this provide an approach to alleviatin
an problems which
.9 'Y
exist but it will also give an indication of the
strength of the
connection. It is as important to the ISC as i P
is to tl~e Industry to
have the degree of.relationship between smoking
and disease accurately
established. For the reasons outlined, however,
the toxicological
approach is unlikely to succeed. The problem must
be studied at the level
of the disease, its aetiology, its epidemiolocTy,
its mechanisms and its
experimental models. it is here that the academic
approach has a role and'
where the Industry can collaborate with its
extensive knowledge of the
product as a whole. As it is becoming increasingly
clear for virtually
all diseases associated with smoking that the
habit is not a sale cause
but at best a contributory one, it is evident that
any serious attack on CD-1
the problem should give particular attention to an
honest and complete
clarification of the actiology of the diseases
concerned.
3. The use of toxicological findings to enhance
the list of chemicals
for which public ranking of brands or other
comparisons would be
BATCo document for Province of BritiSh Columbia 29
October 1999


unwelcome and is scientifically unjustifiable. Roe
(R43) has elegantly
Provided reasons why this is so. It is sufficient
to recall here that
the addition of more chemicals to public
information is likely to confuse
rather than inform, partly by volume of
information, and partly because
quantities of some materials will increase while
others decrease. In
addition, of course, the relevance of any
chosencompounds to disease
states may not be*clear. The alternative, of
compounding a single index
from a range of measures is simply unjustifiable,
as no objective basis
for doing so exists.
4. The value of a knowledge of tobacco smoke
toxicology in"product modifi'catic
is more apparent than real. Brieflv, the
reasons-are as follows. Firstly, as
already noted, the connection between any given
compound and a disease state is
at best speculative, and not necessarily an
adequate basis for imnortant
work. Secondly, a relationship between a chemical
in smoke and any
toxicological data can be indirect. The chemical
measured may simply
reflect the quantity of a whole group of materials
within which the
culpable entity lies. Thirdly, many potential
product modifications are
applied at the level of tobacco leaf, and as the
relationship between
leaf and smoke chemistry is not always clear, a
knowledge of undesirable
chemicals in smoke may not be of immediate value.
Fourthly, product
modification is not a. precise process whereby
single chemicals can be
removed at will. The simplest changes to the
product may cause gross
CD
chemical changes, so. that a putative desirable
effect may be accompanied
by a host of undesirable chan'ges. Fifthly, in
extension of the last
point, changes whichare desirable from one
toxicological viewpoint may (__n
be deleterious from another. There is strong
evidence from bicassay work
that adding nitrate to tobacco reduces
tumorigenicity to mouse skin.
it also, however, increases the yield of
nitrosamines which is, in
BATCo document for Province of BritiSh Columbia 29
October 1999


principle, bad. This example also illustrates
another undesirable by-
product of modification. Nitrate. enhanced
cigarettes were quite
unsmokeable'and hence of no practical value..
Finally, the toxicological
route is not the best one for product
modification. Pragmatically based
modifications,-such as the use of tobacco sheet
and stem, have known
worth, lead to a whole series of derived
modifications, and, most
importantly, fit into the known range of
organoleptic and commercial
practicability. As such, the Companies can
undertake modification on a
commercial basis.
We may now deal with the specific proposals which
the ISC made
regarding compounds and Phenomena requiring study.
i. nitrogen oxides in relation to emphysema and
bronchitis: The current theor,
for the aetiology of emphysema, which has enjoyed
wide support for
some time, suggests that the disease is the result
of the destruction
or deformation, of elastin fibres in the lung, as
the result of an
imbalance between proteolytic enzymes and their
inhibitors. The
imbalance may occur in a number of ways. Some
experimental evidence
which has not been repeated recently, has
suggested that nitrogen
oxides could have a contributory role. other
components of smoke,
such as particulate matter as such, kaolinite,
cadmium, iron and
nickel carbonyls and methylindole, have also been
implicated with
varying degrees of credibility. There seems no
reason to single out
nitrogen oxides. Similarly in the case of
bronchitis, for which
the models of aetiology are less clear, other
smoke components may
be involved. It seems less important to study
isolated smoke
components than the diseases and other
environmental correlates
which exist.
U.-
BATCo document for Province of British Columbia 29
October 1999


ii. acrolein, formaldehyde and hydrogen cyanide in
relation to pulmonary
clearance mechanisms: The compounds noted may have
an effect on
certain lung clearance mechanisms, as do many
other substances found in
smoke and the environment. The ISC presumably
considers the effect
of smoke upon lung clearance mechanisms to be
important. If so, it
should be studied as such, and not in the
fragmented manner suggested.
iii. cresols and phenols in promoting
carcinogenesis: The same arguments
apply. if the field of co-carcinogenesis is
thought important, its
significance should be assessed properly, rather
than via the role of
three specific chemicals in one aspect of the
problems.
iv. nitrosamines: A range of nitrosamines are
present in tobacco smoke.
Improvements in* analytical methods in
recent-yea-rs have, by
reducing artefactual synthesis during ~analysis,
given progressively
more accurate measuxes of the concentrations
present. All nitrosamines
detected are present in exceedingly minute
quantities, of the order of
nanogram per cigarette for volatile nitrosamines
and
micrograms per cigarette for the tobacco-specific
compounds.
There is in fact no evidence that any nitrosamines
in any
circumstance are carcinogenic to man, although it
is likely that CZ)
evidence will eventually be adduced. Even
attempting to assess the
CD
significance of these findings from animal data is
difficult however.
It is widely accepted that the quantities of
volatile nitrosamines
are inadequate to give any effect. Assessment of
the tobacco - specific
nitrosamines is more difficult, as the
toxicological data is barely
adequate, but available evidence again suggests
that amounts are too
small to be dangerous-.
BATCO document for Province of BritiSh Columbia 29
October 1999


Clarification of this problem depends not so much
on the quantities
and species in tobacco, on which the Industry has
sound information,
but on the toxicological da%a, which must come
from another source.
Finally, the ISC is interested in the health risks
attending
initiators and promotors, organ specific
carcinogens and ciliatoxic agents.
Starting, from the end of the list, with
ciliatoxic agents, this is a
re-statement of the previous requests for data on
lung clearance mechanisms.
Organ specific carcinogens, that is carcinogens
which affect a specific
organ after transport through the body and not
necessarily the'site of ent~ryl
present a difficult case. Nitrosamines are
carcinogens of this type in
smoke, whose significance is obscure. Others could
be sought, but it seems
at present that there is not sufficient of a
problem.to justify a search
for a cause.
Initiators and promotors, however, may represent a
more significant
issue. Certainly, if the definition is widened to
co-carcinogens in a
general sense, to avoid semantic confusion, such
materials could be
crucial in understanding tobacco carcinogenesis.
The role of synergisms
between different smoke components and between
smoke components and other
environmental chemicals has already been mentioned
as a possible
explanation for the failure of the chemical
components of smoke to account,
quantitatively, for its activity. In addition, a
co-carcinogenic role
for tobacco smoke is. not incompatible with the
epidemiological evidence on
diesease incidence. The clarification of this
point would be useful, but
would also be extremely difficult. Bioassay models
for co -carcinogen es is L__J
exist, but are not satisfactory, especially as a
convincing demonstration U1
(_51
would require an inhalation model. In addition, as
it is at least
possible, if not likely, that other agents than
tobacco are involved in
BATCo document for Province of BritiSh Columbia 29
October 1999


the process suitable candidates would have to be
identified and
incorporated in the model.
'Carreras Rothmans Ltd
May 1982
BATCO document for Province of BritiSh Columbia 29
October 1999