Product Design
Cigarette Yields of Tar and Nicotine and Markers of Exposure to Tobacco Smoke
Abstract
Critically reviews the interpretation of data in a published paper that compares cigarette yields and biological markers of smoke exposure. States numerous smoking factors are ignored including: compensation, number of cigarettes per day, brand/type of cigarette and the time since last cigarette.
User-Contributed Notes
Fields
- Author
- Lee, Peter N. (TAC Biostatistician)Frequently funded by the tobacco industry to criticize and discount published and epidemiological studies that linked between tobacco smoking and health damage.
- Hypothesis
- CompensationIncorporating knowledge of compensation and effects of human smoking behavior into cigarette design.
- FTC machine testing and ratingsDesign changes to achieve altered FTC smoke machine tar and nicotine ratings, with or without measured changes in human intake.
- Inhalation ProfileAre cigarettes designed to cater to individual inhalation profiles?
- Mainstream constituent yieldsModification of selected mainstream smoke constituents in response to health concerns.
- Measuring human intakeDevelopment of scientifically valid procedures for measuring tar and nicotine levels that more accurately reflect human intake.
- Measuring human smoking behaviorMeasuring the effects of changes in human smoking behavior on intake of nicotine and smoke constituents.
- Nicotine transport, transfer, and uptakeDesign changes which alter nicotine delivery or effect how the product causes and maintains dependence, including transfer of nicotine from tobacco to smoke, and uptake into the body.
- Smoke constituent testingDevelopment of methods for measurement of gas and particulate yields in mainstream and sidestream smoke.
- Smoking psychology and behavior
- Keyword
- Aerosol
- Blood nicotine
- Brand differences
- Compensation (Titration)
- Cotinine
- Daily intake
- Human testing
- Metabolite
- Nicotine delivery (Smoke nicotine or nicotine yield)
- Passive Smoking
- Puffing behavior (Human puff parameters)
- Reaction products
- Self-administration
- Tar/Nicotine ratio (Nicotine/Tar Ratio or T/N ratio)
- Total particulate matter (TPM or Tar)
- Smoke Constituent
- Carbon monoxide
- Nicotine
- Total particulate matter
- Named Organization
- American Review of Respiratory Diseases
- Federal Trade Commission (Enforcement agency for laws against deceptive advertising)Enforces laws against false and deceptive advertising, including ads for tobacco products. Ensures proper display of health warnings in ads and on tobacco products;collects and reports to Congress information concerning cigarette and smokeless tobacco advertising, sales expenditures, and the tar, nicotine, and carbon monoxide content of cigarettes.
- Oxford Medical Publications
- Subject
- Compensation (Measures)
- Metabolites (Measures)
- nicotine analogues (Technology)
- Secondhand Smoke/Constituents
- Smoke Delivery/Transport (Measures)
- Smoke Nicotine (Measures)
- T/N Ratios (Measures)
- Tar (Measures)
- Test/Smoke Condensate (Testing)
- Test/Smoke Constituents (Testing)
- Test/Smoke Machine (Testing)
- Test/Smoking Behavior (Testing)
- Transfer to Smoke (Measures)
Document Images
h
'1'
REVIEW 684 CONFIDENTIAL
SubLct ref 6d
"Cigarette yields of tar and nicotine
and markers of exposure to tobacco smoke"
D B Coultas, C A Stidley and J M Samet
American Review of Respiratory Diseases (1993), 148, 435-440
This paper provides a very misleading interpretation of data from a
study of 298 smokers in New Mexico, in which the two smoke uptake
markers, salivary cotinine and
end-expired carbon monoxide (CO), were
related to number of cigarettes smoked per day an& to brand tar,
nicotine and CO yield. The authors note a markedly stronger correlation
of the uptake markers with number of cigarettes smoked than with any of
the indices of brand yield. They conclude that "Because FfC yields of
tar and nicotine are poor predictors of exposure to tobacco consumption
products, subjects' reports of cigarette brand should not be used as
primary marker of exposure in epidemiological investigations".
But are tar and nicotine yields actually poor predictors
a
exposure?' If one looks at their actual main findings presented in Tab1e
3, they show that salivary cotinine is significantly linearly related to
number of cigarettes per day, with a regression coefficient estimated as
about 7.7 ng/ml per cigarette smoked (the
exact value depending on the
model chosen), to tar yisl6with a coefficient of 8.61 ng/ml per mg,,
and
to nicotine yield with a coefficient of 139.6 ng/ml per mg. Since
-2-
averages for the population (not given exactly in the paper) are of order
15 cigs/day, 15 mg tar and 1 mg nicotine, i.e. approximately inversely
related to the regression coefficient, it is clear that for an average
smoker a given percentage reduction, whether it be in number of
cigarettes per day, tar yield or nicotine yield, will be associated with~
a similar reduction in salivary cotinine. It seems highly likely that the
reason the observed correlations for salivary cotinine were much higher
for number smoked than for tar or nicotine yield was that the population
varied much more in number smoked than in yield. (For two variables that
are linearly related, but measured with error, it can easily be
demonstrated that the observed correlation will be higher the greater the
observed range of the variables.): The findings as presented'seem to be
exactly what one might expect if exposure to nicotine was
equal to the
product of number smoked an& bran& nicotine to yield, with measured
uptake (salivary cotinine) for a given~ exposure variable because of
between individual variation in pattern of smoking and metabolism. If it
is in fact true that halving the tar (nicotine yiel& of the brand smoked
(while keeping the number of cigarettes constant) has the same effect on
uptake as halving the number of cigarettes smoked (while keeping
tar/nicotine yield constant), there is no reason for the general tenor of
the whole paper, which is to give the impression that tar/nicotine
reduction is of doubtful value and that information on bran& yields is
misleading to the smoker.
The paper is actually quite remarkable in~that it makes no formali
-3-
attempt at aLl to distinguish between the three major hypotheses of
current interest for studies of this type, namely that for a given
number of cigarettes smoked, salivary cotinine is:
(i) directly related to nicotine yield'- "no compensation";
(ii) not related to nicotine yield at all - "complete compensation"; or
(iii) an intermediate position - "partial compensation".
Although the authors' results appear to exclude the second
possibility, they do not really demonstrate clearly whether there has
been compensation or not. In other studies (e.g. Adlkofer et al (1989),
116-130, in "Nicotine, Smoking and the Low Tar Programme", edited by Wald
and Froggatt and published by Oxford Medical Publications), which
generally show a degree of compensation, a formal mathematical model has
been used in which
UI ,~b
where U is uptake (here salivary cotinine), N'is the number of cigarettes
smoked, Y is brand yield (of nicotine), and a and b are coefficients to
be estimated. A value of b - 1 is equivalent to there being no
compensation, b - 0 indicates complete compensation, and b in between 0
and 1 indicates partial' compensation. By taking logarithms the model not
only becomes linear,, allowing standard regression techniques to be used,
but also takes into account the fact that cotinine values tend to be
log-normally distributed.
The actual model usedi by the authors is totally inappropriate for a
number of reasons:
(!i) It does not allow any estimate of compensation to be made;
(!ii) It does not include any term in number of cigarettes
per day
multiplied by brand yield, which is a priori the most obvious
correlate of uptake;
(iii) It assumes variability in cotinine is independent of cotinine
value, which is clearly untrue (typically values are ± percentage
error, not ± absolute error);
(iv) It very dubiously includes time since last cigarette in the model -
time since last cigarette being obviously highly correlated with
number of cigarettes per day; and finally
(v) It produces totally implausible estimates of cotinine level in
nonsmokers (or smokers of very small numbers of cigarettes).
There is also a fairly important weakness of the study itself in
that the data on brand yield are likely to be inaccurate since they were
based on self-report uncorroborated by evidence from a cigarette product.
For many brand names there are a number of variants, and there is an
obvious possibility of error unless the subject is asked to provide an
empty packet for checking. However this weakness, which the authors
recognize, is minor comparedi with their general misinterpretation of
their data.
One other aspect of the paper deserving mention is that the authors
found no significant relationship between number of cigarettes smoked and
tar/nicotine yiel&. This is consistent with most published evidence
indicating that people switching to lower yiel& cigarettes do not
materially alter the number of cigarettes they smoke. Overall the paper
-5-
does not affect the general conclusion that, although people may alter
their smoking patterns somewhat to compensate partly for reduced
tar/nicotine yield, switching to reduced tar/nicotine cigarettes does
mean a reduction in uptake of tar and'nicotine.
P N Lee
4.11.93
