NOTES ON STRATEGIES FOR LESS HEALTH-CONTENTIOUS PRODUCTS
By: W. D. E, Irwin
The Case for Reducing Smoke Dose
From epidemiology, lower cigarette consumption correlates with reduced risk of
certain diseases. These notes assume the contention of a causal basis for the
correlation. though this has not been validated.
Lower cigarette consumption is equivaler-.' lo lower total smoke dose, provided it is
not offset by:
1. ' Higher cigarette delivery (machine);
2. Greater human smoking intensity (compensation type A).
Early epidemiological data were obtained at a time when the current range of
deliveries was not available. Therefore the first condition probably held. The second
condition probably held as well, but no justification Is given here.
Total smoke dose = Unit dose x units consumed
Therefore there shouldbea lowerepidemiological risk from lowertotalsmoke delivery
products provided this is not offset by:
1 . Greater smoking intensity (compensation type A);
2. Greater unit consumption (compensation type B).
Hence the need for research to understand what people smoke for, e.g.:
Nicotine dose (pharmacological);
Nicotine impact (sensory);
Flavour;
'Sincke';
Sonielhing to do;
Other, or combination.
Studies of [tie effects of such factors on compensation types A and B behaviour are
-else indicated.
I- lowevef, smolie is a compiex mixture, so when we talk about lower dose or delivery.
do we need to reduce everything, or can we be more selective?
C:@
BATCo document for Province of British Columbia 16 April 1999

Three Strategies
Siralegy I
Identify [lie risk-ciusing substances in tobacco smoke and remove them selectively.
'I leis has ilia advatliage of leaving the positive features, what people smoke lo(,
assuming there is no serious overlap between 'desirable' and 'undesirable'
substances or characteristics.
Strategy 2
The opposite of [lie above. Reduce all deliveries by 'broad spectrum* methods, i.e.
mainly filtration, venlilationandreduction in theweightol tobacco or similar materials.
Build back selectively desirable sensory and other elements. This has the advantage
of not identifying rlsk causes, or finding selective reduction methods. However. it
does imply identification of positive elements for add-back.
Strategy 3 ('Alternative product)
This recognises some of the limitations of 1 and 2, both of which burn tobacco or
smidar to give a highly complex pyrolysis-based aerosol.
Do notbum tobacco, but find away to deliversome or all of the sensory characteristics
of tobacco smoke in a simple chemically defined form allowing similar usage patterns
to a cigarette.
Tile Three Strategies - Further Contrasts/Problems etc.
In reverse order:
Strategy 3
1. It could be labelled a drug delivery system.
2. Added nicotine - new hazards (health or political) - limits to nicotine quantity
and concentration. A non-nicotine option?
3. Other 'now' hazards?
4. Despite above, greatest potential an blo activity.
5. Greatest new technology requirement.
6. Full tobacco sensory property range unlikely to be realised from a relatively
s implo 'smoke' composition?
.M_
CD
NZ
BATCo document for Province of British Columbia 16 April 1999

SD sqLeo y 2
1 - Thq evolutionary approach implemented over 20-30 years.
2. Sorno encouragement from the epidemiology.
3. But compensation is a continuing Issue. A particular problem if consumer
. smokes for 'smoke', see earlier - Smoke -a PMWNFf
4. Points to more radical ellorts In sensory adjustment, e.g. GREENDOT.
5. Benefits from some residual tobacco smoke, hence some residual sensory
cl la unlike Strategy 3.
6. However, residual tobacco smoke means residual biological activity.
7. Less chance of new hazards than Strategy 3, unless sensory adjustment at
very low PMVVNF introduces Strategy 3 issues.
8. Lower new technology requirement than Strategy 3. though 'extreme' options
may move technology towards Strategy 3, e.g. sensory adjustment.
Stratggy 1
This would be the ideal, indeed the only, option it smokers smoke for 'smoke' and
a substitute simple chemically defined aerosol such as glycerol could not be found
which would replace tobacco smoke sensory attributes. However, there are two
major and related problem areas - Identifying the risk causes and removing them
selectively.
A. Identification of the risk causes
I . Smoke contains thousands of chemicals - weighting their contributions to tile
risk.
2. Probability that health effects are not additive. Le. Interaction, the whole not
equal to tile sum of the parts.
3. Paft of the risk may not be chemically based, e.g. physical irritation by smoke
in a more holistic sense.
4. At least 3 disease types.
S. Problem of proof that selective reduction of certain substances or substance
classes lessens risk. Epidemiology may be the most definitive, but would agree
that bio tests oiler some measure of progress. some method of 'weighting'.
B. Selective removal if target substances identified
I li i many cases there are multiple tobacco sources of a given noxa. e.g. catechol
probably from polyphenols, sugar, cellulose and others.
C--
BATCo document for Province of British Columbia 16 April 1999

1
2. In many cases there are multiple smoke components from a given tobacco
source, e.g. ca(echol. CO and aldehydes from sugar. Some of these smoke
components will be desirable as well as undesirable.
3. Ftom I and2, istlienuinberollargetnoxaeincreases, the nu inber of precursor
substances for removal or modification will Increase rapidly with probable rapid
effect on desirable smoke components as well. Indeed, the ability to reduce
noxae relative to tar (-= smok will decrease because some of the precursors
[ray be major tar producers, e.g. cellulose.
4. Limits to specificity in tobacco modification.
5. The above points explain why Strategy 2 has been adopted so far.
Some Conclusions on Programmes Related to Strategy 1
The selective reduction of a finite group of substances will not attract a strong
consistent scientific or political consensus that it Is 'safer' than Strategy 2 products
which give non-selective reduction of all components and therefore must reduce tile
dose of other unknown risk-causing variables, provided compensation is controlled.
However. there will continue to be scientific/political hypotheses creating pressures
for the reduction of certain smoke components, even though it cannot be proved that
such reductions give a safer smoke. An example might be the numerous papers
from the Hoffmann school pressing the case for nitrosamine reduction. oven though
certain epidemiology based considerations may suggest no measurable contribution
of these to the epidemiological risk.
The hypothesised smoke components will change from time to time, In part due to
[tie changing research interests of Influential workers, making prediction difficult and
literature awareness important. However, to allow any focused research activity,
some guestimate must be made of the likely substances in terms of cuff ent and
potential interest, e.g. the SRG list.
Potential "Other Noxae" Goals
I Maintain a Company knowledge base, supported by our own research, on the
health-contentious substances In smoke and the factors which Influence their
levels. This is part of a wider activity to maintain a knowledge base on the
smoking and health partof the environment in which the Company must operate.
and to interact with it as necessary. Another aspect of this goal may be 'due
diligence' in understanding the properties of the product we sell.
2. As an extension of 1. to provide background research information on tile
opportunities for, and limitations to, the selective reduction of smoke levels of
health-conlentious substances.
(_4
BATCo document for Province of British Columbia 16 April 1999

5
3. As an extension of 2, to develop products giving selective reductions In target
substances.
Further Notes
A particular Regulatory Authority whim to 'league table' one specific substance, e.g.
nitricoxide,oronespecificbiologicalparameter,e.g.Amesacflvity, may offer product
opportunities, but is quite different to a "GRAT (generally recognised as safer)
product. The feasibility of meeting a particular whim will depend on whether tile
noxa has finite, preferably one, main source (e.g. nitric oxide) or Is mulUsourc8d (e.g.
aldehydes). For many vapour phase noxas, the simple vapour phase selective
property of carbon filters should not be forgotten. This represents 'cure'. removal
after formation. whereas the other noxae programme concentrates on 'prevention',
by tobacco modification. The sensory and process cost/noxa benefit ratio Is likely
to favour carbon filtration (existing technology) for substances such as acrolein. If
our research goal Is to effect additional reductions beyond current technology, which
it is, then the carbon argument can be Ignored, though not It there were a product
development need now. However, this Is getting a bit detailed and I rest my case.
I hope it provides food for thought.
W. D. E. It win
NIC
BATCo document for Province of British Columbia 16 April 1999