ARTICLE IN PRESS

Experimental and Toxicologic Pathology 60 (2008) 141–156

www.elsevier.de/etp

Tobacco smoke: Unraveling a controversial subject$, $$

Anja Thielena,1, Hubert Klusb, Lutz Müllerc,
a
Tabak und Rauchen GmbH, Neustädtische Kirchstraße 8, 10117 Berlin, Germany
b
Oriongasse 9, 3100 St. Pölten-Wagram, Austria
c
Stralsunder Straße 1, 01109 Dresden, Germany

recibido en 4 Diciembre 2007; aceptado en 30 Enero 2008

Dedicated to Dr. Richard R. Baker

Abstract
Cigarettes are a modern and industrial form of tobacco use and obviously involve more than just tobacco. A multitude
of physical processes and chemical reactions occur inside the burning zone of a cigarette. Cigarette smoke is an aerosol of
liquid droplets (the particulate phase) suspended within a mixture of gases and semi-volatile compounds. Two kinds of
smoke with different composition and properties are produced during smoking: mainstream smoke inhaled by the smoker
and sidestream smoke, which is released into the environment between puffs from the lit end of the cigarette.
Several techniques and modifications have altered the design of the cigarette during the last 50 years and changed
smoke composition, with the effect of lower tar and nicotine smoke yields. An enormous amount of research has been
done since the 1950s on smoke composition. With regard to the numerous toxic or carcinogenic constituents identified
in tobacco smoke, there is a strong focus in the industry and with the authorities on the over 40 compounds, called
‘‘Hoffmann analytes’’.
The yields of tar and nicotine in mainstream smoke of a cigarette brand as printed on the pack are measured with
smoking machines under highly standardized conditions. Yields must comply with regulatory limits set in a number of
countries. Smoking by machine is different from the smoking behavior of humans. There is a growing movement to
develop more ‘‘realistic’’ methods to estimate smoke yields. But it is unclear whether alternative smoking regimens are
more representative of human smoking behavior and provide better predictions of human exposure.
Tobacco smoke has strong biological and toxicological effects in vitro and in vivo. There is an obvious need for
developing a unified and validated testing approach particularly for the assessment of additives and the evaluation of
new potentially reduced exposure products (PREPs).
This paper gives a comprehensive overview of cigarette design, the composition and toxicity testing of smoke, and
the way machines and people smoke – with links to the more detailed literature.
r 2008 Elsevier GmbH. All rights reserved.

Keywords: Cigarette design; Cigarette smoke constituents; Hoffmann analytes; Machine smoking; Smoking behavior; Toxicity testing

$
Presented at the congress on alternative test methods in inhalation toxicology, Berlin, Germany, 7–9 May 2007.
$$
Dr. Richard R. Baker was originally invited to be the speaker on tobacco smoke. He worked on his talk until his unexpected death at Easter
2007. Using the material he had prepared, friends of his finished the presentation and one of them (L.M.) gave it at the conference. This paper is
dedicated to an excellent scientist and a very good friend.
Corresponding author. Tel.: +49 351 890 2922; mobile: +49 151 152 15203.
E-mail address: Lutz.Mueller.DD@t-online.de (L. Müller).
1
Present address: Karl-Stieler-Straße 12a, 12167 Berlin, Germany.

0940-2993/$ - see front matter r 2008 Elsevier GmbH. All rights reserved.
doi:10.1016/j.etp.2008.01.014
 ARTICLE IN PRESS
142 A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
Introduction
In 1492, Christopher Columbus returned to Europe
from what was to be called America with the first
tobacco leaves and seeds ever seen in the Old World.
Initially, European doctors recommended tobacco as a
cure for toothache, worms and other ailments. Queen
Catherine de Medici used tobacco to treat her migraines
after the French ambassador in Portugal, Jean Nicot,
had acquired tobacco seeds in Lisbon and introduced
the tobacco plant to France in 1561. In his honor, the
Swedish botanist Carolus Linnaeus named the tobacco
plant genus Nicotiana in 1753.
In the 18th century, snuff and pipe smoking were the
most popular forms of tobacco use. The 19th century
became the age of the cigar. Cigarettes were largely
unknown in the English-speaking world before the
Crimean War when British soldiers began emulating
their Ottoman Turkish comrades, who had the habit of
rolling tobacco in newsprint paper. The invention of
machines for the manufacturing of cigarettes in the
1850s opened the way for mass production and the
development of the modern forms of cigarettes in
the 20th century.
Cigarette consumption has since been rising steadily
to reach a worldwide consumption of nearly 5.6  1012
cigarettes in 2000 (American Cancer Society, 2006).
Tobacco use has always been a controversial issue with
its increasing popularity on the one side and the growing
evidence of health risks on the other. An enormous
amount of research has since been done on this subject
and there are still open questions, which call for
additional research. Today, the regulation of tobacco
products and their use, and the need for tobacco
products with reduced risk are in the focus of the
activities of authorities and the industry.
What follows is a comprehensive overview of the
design of cigarettes, the composition of smoke, approa-
ches to toxicological testing and the way standardized
analytical machines and real people smoke. A number of
references are included for further reading.
The modern cigarette
Commercially manufactured cigarettes consist of a
tobacco blend, paper and usually, but not always, a
filter.
Mainly three types of tobacco are used for the
manufacturing of cigarettes worldwide: Virginia, Burley
and Oriental (Table 1). The distinct tobacco types, and
the methods how the leaves are cured after harvest, lead
to characteristic flavor profiles and composition of
smoke of the finished product. Curing of leaves is the
carefully controlled process used to develop the texture,
Table 1. General characteristics of the three important
tobacco types after curing
Virginia/Bright Burley Oriental
Leaves Large, orange Large, brown Small, orange
Nicotine Middle to high High Low
Sugar High None High
Curing Flue-curing Air-curing Sun-curing
color and overall quality of a specific tobacco type.
During curing, many chemical and physical changes
take place in the tobacco leaf. For instance, starch is
converted into sugar, the green color vanishes and the
tobacco changes its color from light yellow to orange
and brown as tree leaves do in fall.
The Virginia tobacco type is named after the US state
where it was first grown. The large leaves generally
contain a relatively high amount of nicotine. The flue-
curing process is fast (5–7 days) and conserves the
natural high amount of sugar. Flue-curing is the process
when heat is introduced into a barn with the leaves –
indirectly by means of a pipe system or, in former times,
directly as exhaust gases of an external furnace. Because
of its yellow/orange color after flue-curing, Virginia is
also called Bright tobacco.
Growing Burley tobacco requires heavier soils and
more fertilizer than Virginia. After being air-cured in
unheated, ventilated barns to dry naturally for a period
of about 7 weeks, the tobacco turns brown with virtually
no sugar left, giving its smoke a harsh taste.
Oriental is a tobacco type with rather small leaves.
Oriental is densely planted and grown in the hot summer
of the Balkans, Greece, Turkey and the Middle East.
These growing conditions help create its intensely
aromatic flavor and a high sugar content, which are
subsequently enhanced by sun-curing. Sun-cured tobac-
co leaf is strung out on racks and exposed to the sun.
Another type of air-cured tobacco is Maryland
tobacco, which is typically used in small amounts in
cigarettes manufactured in the United States.
The primary component of a cigarette, of course, is
tobacco but today’s cigarettes involve much more than
just tobacco (Fig. 1). Cigarettes are a modern and
industrial form of tobacco use with a rather complex
design providing the consumer with a product of high
and consistent quality.
For the better appreciation of what follows, it is
useful to keep two definitions in mind. ‘‘Ingredient
means any substance or any component except for
tobacco y used in the manufacture or preparation of a
tobacco product y, including paper, filter, inks and
adhesives’’ (The European Parliament and the Council
of the European Union. Directive 2001/37/EC). Addi-
tives are generally understood to be substances added to
 ARTICLE IN PRESS
A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
Permeable Paper Tobacco
Air
Smoke
Fig. 1. The modern cigarette.
the basic components, such as tobacco, paper, filter
materials, inks and adhesives, to impart specific
desirable properties on them and control the perfor-
mance of the cigarette while being smoked.
Finely cut tobacco alone has the tendency to dry up,
break and drop out of the open end of the cigarette. If
cigarettes are packed with higher water content to
prevent such drying losses, they may not burn well and
are at risk of mold infection. To overcome these
problems humectants, such as glycerol, sorbitol or
propylene glycol, are added to the tobacco for the
control of moisture levels. Primarily used in reconstitu-
ted tobacco, preservatives like potassium sorbate,
benzoic acid or sodium benzoate prevent product
degradation. Other additives in reconstituted tobacco
act as binders or fillers (like calcium carbonate).
Combustion modifiers like potassium acetate make the
cigarette burn evenly and ensure that it stays lit between
puffs.
In many countries around the world, American blend
cigarettes have grown popular. They are made from a
mixture of Virginia, Burley – in the United States, with
small additions of Maryland – and Oriental leaf
tobacco. Typical for the taste of this kind of cigarettes
are the presence of Burley tobacco in the blend and the
flavoring of the tobacco. Some Burley tobaccos in
particular need the addition of sugars and a special
processing step, which is called ‘‘toasting’’ (well-contro-
lled heating with the addition of reducing sugars and
other substances like citric acid) to reduce the levels of
ammonia and amines for ameliorating the harsh taste of
this tobacco type.
After these processing steps, the blend is furnished
with flavorings such as cocoa, licorice, fruit extracts and
various sugars together with the already mentioned
humectants, which are known collectively as ‘‘casings’’.
The casing is usually applied at an early stage of
cigarette making to the pre-cut tobacco and improves
the smoke characteristics and the processability of the
tobacco. Casing materials are mostly water soluble with
less volatility. They are applied on the tobacco at a
temperature of 70 1C and in a humid atmosphere. The
level of addition typically is about 1–5% of the total
weight of the cigarette tobacco. At a later stage of
production, perfume-like flavors, known as ‘‘toppings’’
143
Air Vents Filter
Air
Smoke
filtered
and
diluted
with air
or ‘‘top flavors’’ are applied to the cut tobacco mixture
at ambient temperature to improve the consistency in
flavor and pack aroma. Top flavors are usually highly
volatile and, therefore, these aroma substances and
ethereal oils are added in an alcohol base, which
evaporates after the flavor is applied. In the end, the
amount of flavors is about 0.1% of the weight of the
cigarette tobacco.
Although casings and flavorings in some types of
cigarettes include sugars, cocoa and fruit extracts, the
taste of the tobacco smoke is not comparable to the
flavor these additives give to foods.
Cigarettes made from highly aromatic Virginia
tobaccos only are preferred in some parts of the world,
like in the UK, Canada, India or other countries of the
former British Commonwealth. In parts of southern
Europe, there is also a market for Oriental cigarettes.
Virginia or Oriental cigarettes are generally manufactu-
red without the addition of any flavorings.
Special cigarette paper is used to enclose the tobacco.
Besides cellulose fibers of different origins it contains
inorganic substances like sodium citrate and calcium
carbonate that modify the porosity of the paper and the
smoldering rate, but also make the cigarette burn evenly
and control the stability of the ash cone. The porosity of
the paper (its ability to admit air) influences the taste of
the cigarette and the delivery of smoke. The more air
passes through the paper, the more the smoke moving
through the cigarette is diluted, resulting in a lower yield
product.
Seam adhesives – made from starch or polyvinyl
acetate – are used to glue the cigarette paper in place
around the tobacco, and a small amount of dye print ink
may serve to add a distinctive brand mark on the
cigarette paper, near the filter tip.
The filter is a white plug made mainly from cellulose
acetate in the form of fine spun fibers, which retain
particles in the smoke to reduce the delivery of tar and
nicotine. The performance of the filter is fine tuned by
selecting a suitable thickness and stretching the tow to
the right crimp. These filter variations can affect the
efficiency of filtration, and thus the taste and smoke
yields. Charcoal is sometimes included in the standard
cellulose acetate filters as its adsorption properties can
reduce some of the gas components in smoke. Because
 ARTICLE IN PRESS
144 A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
charcoal is mainly elemental carbon, these are some-
times called carbon filters. Making small perforations in
the filter is the basis of the so-called filter ventilation. It
dilutes the smoke with air and has some influence on the
burning temperature during puffing. This leads to less
and slightly changed smoke being produced in each puff
and, consequently, to a reduction of a cigarette’s yield
and sensory strength. Filter ventilation is particularly
important for reducing smoke gases that are not trapped
in the filter. A hardening agent (triacetin, the acetic acid
ester of glycerol) helps the cellulose acetate filter to keep
its shape in the pack and while the cigarette is being
smoked. In addition, triacetin has a specific selectivity
for some smoke compounds and can reduce, for
example, phenols and volatile N-nitrosamines. Plug
wrap is a special paper wrapper put around the filter
to hold it together during manufacture. The filter and
tobacco rod are held together by the tipping paper, often
with an imitation cork pattern printed on it. This paper
may also have holes in it to admit air in low tar
cigarettes. Filter adhesive is used to glue the tipping
paper to the filter and the tobacco rod. The papers used
for tipping the cigarette form the mouthpiece and
surround the filter (for further reading: Davis and
Nielsen, 1999; Hoffmann et al., 2001).
The burning cigarette and the composition of
mainstream smoke
When a cigarette is smoked, combustion takes place
in two ways: during a puff, air is drawn into the cigarette
and mainstream smoke is formed and inhaled by the
smoker while between the puffs the cigarette smolders
and sidestream smoke is released into the environment
from the lit end of the cigarette. In general, the
substances present in mainstream smoke are also present
in sidestream smoke. The relative yield per cigarette in
both kinds of smoke and the mainstream/sidestream
ratios for the different smoke constituents are dependent
on cigarette construction and on the smoke component
in question. Whereas mainstream smoke of American
blend, Virginia and Oriental cigarettes is slightly acidic,
sidestream smoke is slightly alkaline due to the higher
content of ammonia. Fresh sidestream smoke particles
are smaller than mainstream smoke particles, and
sidestream smoke is generated at lower temperatures
(Table 2).
The mixture between sidestream and exhaled mains-
tream smoke in the environment after dilution and aging
is called environmental tobacco smoke or ETS. Aging
may go on for minutes or hours. While aging, the
composition of the gaseous and particulate phases is
changing. For example, the distribution of nicotine
between the gaseous and the particulate phases is
Table 2. Differences between cigarette mainstream and
sidestream smoke
Characteristic Mainstream smoke Sidestream
smoke
Temperature at origin 850–950 1C 500–650 1C
Mean particle size 0.35–0.4 mm 0.15–0.25 mm
Particle concentration 109–1010 cm3 105 cm3
Proportion of burned 30–40% 50–60%
tobacco
Nicotine 499% in particulate 495% in gas
phase phase
Modified from Richter and Scherer (2004).
changed and it is adsorbed onto surfaces. Radicals can
react and smoke cytotoxicity decreases.
Many types of chemical reactions and physical
processes occur in a burning cigarette. There are two
main regions inside a burning cigarette: a combustion
zone and a pyrolysis/distillation/pyrosynthesis zone
(Fig. 2). Inside the combustion zone, oxygen reacts with
carbonized tobacco producing carbon dioxide, carbon
monoxide and hydrogen. Temperatures up to 950 1C are
generated when the cigarette is puffed. Immediately
downstream of the combustion zone is the cooler
pyrolysis/distillation/pyrosythesis zone, where the bulk
of the more than 4700 chemicals in smoke are generated.
The super-saturated vapor rapidly cools within a few
milliseconds in the tobacco rod and condenses into the
aerosol particles that make up the smoke (for further
details, see Baker, 1999).
Cigarette smoke is an aerosol of liquid droplets (called
the particulate phase) suspended in a mixture of gases
and semi-volatile compounds (Ingebrethsen, 1986). The
generally accepted definition is that the particulate phase
(colloquially called condensate) is retained by a specially
designed glass fiber filter (Cambridge filter) at room
temperature, whereas the gaseous phase passes through
the filter. ‘‘Tar’’ is analytically defined as the raw water-
and nicotine-free smoke condensate. A number of
compounds, such as the polycyclic aromatic hydro-
carbons, tobacco-specific nitrosamines (TSNAs), phyto-
sterols and the metals, are found practically in the
particulate phase only. Some compounds, e.g. phenol
and the cresols (phenolics), are partitioned between
the particulate and gaseous phases and are termed
semi-volatiles (Table 3). They can be retained by the
Cambridge filter, but are easily vaporized (Guerin et al.,
1987).
The gaseous phase consists mainly of the air
constituents, nitrogen and oxygen. Several combustion
products, such as carbon monoxide, carbon dioxide and
nitric oxide, are in the gaseous phase exclusively. Also
found there are compounds like 1,3-butadiene, formal-
dehyde, acetaldehyde, acrolein, benzene and hydrogen
 ARTICLE IN PRESS
A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156 145

Sidestream Smoke

Light Gases Air by

by Diffusion Ventilation

Condensation
A B and Filtration Filtration

A: Combustion Zone
B: Zone of Distillation, Mainstream Smoke
Pyrolysis and Pyrosynthesis

Fig. 2. The burning cigarette.

Table 3. Presence of mainstream smoke components in vapor
phase
Compound In vapor phase (%)
Acetaldehyde 98
Acrolein 93
Benzene 78
Formaldehyde 68
HCN 50
Acetone 34
Nicotine o2
Modified from Baker (1999).
cyanide; these are important because of their known toxic
or carcinogenic properties. This is especially relevant
because the traditional substrate for assessing the toxic
effects of cigarette smoke has been smoke condensate.
Only later were methods developed for collecting the
gaseous phase by bubbling it through water or buffer –
approaches with a risk of artifact formation.
The size of particles in fresh mainstream smoke ranges
from 0.1 to o1 mm in diameter and may reach the deep
lung. By weight, only 4.5% of whole smoke is particu-
late phase material (Fig. 3). When water is subtracted
from particulate matter weight, the dry particulate
matter accounts for only 3.8% of the weight of whole
smoke. On the other hand, the gaseous phase including
air represents 95.5% of whole smoke. Only 7% of the
vapor phase (total gaseous phase without air consti-
tuents) is tobacco-derived organic smoke compounds,
equivalent to 1.5% of the whole smoke (Dube and
Green, 1982; Green and Rodgman, 1996).
Because of its effects on health, tobacco smoke has
been studied analytically with very high intensity since
the 1950s with the objective of identifying the compo-
nents responsible for these adverse effects. This is
reflected by the enormous number of publications in
this field. By the early 1980s, more than 4700 tobacco
smoke components had been identified (Table 4). The
high number explains the complexity of tobacco smoke,
but most of the thousands of substances found in smoke
are present only in trace amounts (Rodgman and Green,
2003).
Various carcinogens have been found in tobacco
smoke, such as polynuclear aromatic hydrocarbons
(PAHs) (Wynder and Hoffmann, 1959) and the TSNAs
(Klus and Kuhn, 1974; Hoffmann et al., 1979).
As many as 549 individual PAHs were identified in
tobacco smoke, and 12 of these compounds are
classified by the International Agency for Research on
Cancer (IARC) as carcinogens. PAHs are overwhel-
mingly found in the particulate phase, and the levels of
the different PAHs in tobacco smoke decrease as the
number of C atoms increases (Rodgman and Perfetti,
2006).
Nitrosamines can be divided into those nitrosamines
which are not unique to tobacco and the so-called
TSNAs. Nitrosamines, such as dimethyl nitrosamines
and nitrosoamino acids, are found in many foods and
beer, technical materials and the environment. Some are
also present in the vapor and semi-volatile phases of
tobacco smoke. Some of them are classified by IARC as
possible human carcinogens. TSNAs are found only
in tobacco and in the particulate phase of tobacco
smoke. They are formed by N-nitrosation of tobacco
alkaloids during tobacco curing. Two of the four
TSNAs (NNN and NNK) are classified by IARC as
possible human carcinogens (IARC, 2004). The carci-
nogenic PAHs as well as certain TSNAs are examples of
 ARTICLE IN PRESS
146 A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156

Particulate Whole Vapor Organics

Phase Smoke Phase
Unidentified 4 Particulate Organics
4,5
Pigments 4 7
15 Miscellaneous
Miscellaneous 10 Vapor H2O
22 18 8 Nitriles
Alkanes 5
Terpenoid HC 5
Phenols 3,5
Esters 3,5 CO Aldehydes
Other Alkaloids 3,5 18 27 and Ketones
Nicotine 6

Alkohols and 15,5

Humectants N2
59
Aldehydes and 11
Ketones
CO2 50 Hydrocarbons
57
Carboxyl Acids 13

Water 16 Other
O2
13 gases
1,5

Fig. 3. Whole cigarette smoke composition. 85 mm filter cigarette: 68 mm tobacco rod, 17 mm cellulose acetate filter; cased
commercial blend (Green and Rodgman, 1996).

Table 4. Number of compounds identified in major chemical (Gahring and Rogers, 2005). Nicotine is a naturally
classes occurring alkaloid formed in the roots of the solana-
ceous plant family. In very low amounts, it is also found
Class No.
in potatoes and tomatoes. Probably because of its
Amides, imides, lactams 237 insecticidal effect the plant stores nicotine in the leaves.
Carboxylic acids 227 The solubility of nicotine in water and organic solvents
Lactones 150 and its rate of absorption through the skin or biological
Esters 474 membranes is a pH-dependent process. During smo-
Aldehydes 108 king, nicotine is distilled from the burning tobacco and
Ketones 521 is present mainly in the particulate phase of mainstream
Alcohols 379 smoke. The principal route of nicotine absorption after
Phenols 282
inhalation is through the alveoli of the lung where
Amines 196
absorption occurs regardless of the so-called ‘‘original
N-Heterocycles 921
Hydrocarbons 705 pH’’ of the smoke due to the dominating effect of the
Nitriles 106 physiological pH of 7.5 in the lung (reviewed by Slade,
Anhydrides 11 1993; Rodgman, 2000; Dixon et al., 2000). The smoker
Carbohydrates 42 metabolizes the main part of nicotine (about 80%) to
Ethers 311 cotinine rather quickly (half-life 2–3 h). Cotinine is
Inorganics 50 stable for some time (half-life 18 h) before it is
Total 4720 eliminated in the form of metabolites in the urine
(reviewed in Tricker, 2006).
Dube and Green (1982).

smoke constituents of concern and are potent lung

carcinogens for humans (Hoffmann et al., 1996).
Nicotine is the most important pharmacologically
active compound in tobacco smoke due to its interaction
with nicotinic acetylcholine receptors. These receptors
are ligand-gated ion channels and widely distributed
in the brain but also in other non-neuronal tissues
The ‘‘Hoffmann analytes’’
A great deal of scientific work was and is being done
to identify the constituents of the complex mixture,
tobacco smoke, and to determine their biological
activity. Numerous carcinogens in tobacco smoke have
 ARTICLE IN PRESS
A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156 147

been identified and quantified, and their biological (reductions) of Hoffmann analyte smoke levels and the
activity and relevance for cancer studied. IARC has data from toxicological in vitro testing are often not
classified the complex mixture of cigarette smoke as a convincing (Carmines, 2002; Roeper et al., 2004).
known human (group 1) carcinogen (IARC, 1986). The Hoffmann list of substances clearly represents only
In addition, IARC classified 63 of the chemical a small subset of the over 4700 constituents in cigarette
constituents reported in cigarette mainstream smoke as smoke. There are other toxic or carcinogenic compounds
carcinogens (IARC, 2004). Among them, 11 are known in tobacco smoke which are not on the Hoffmann list,
human carcinogens (group 1). The IARC list of such as furfural, ethylene oxide, propylene oxide,
carcinogens in cigarette smoke is based on one by radioactive elements and radicals. Nevertheless, the list
Hoffmann et al. (2001) where 69 tobacco smoke provides a starting point for identifying and regulating
carcinogens were listed. constituents that should be reduced in, or eliminated
Dietrich Hoffmann, a biochemist, published many from, smoke. This is the reason why Hoffmann analytes
analytical papers on cigarette smoke. He is a leading were selected or proposed for regulatory reporting
authority on tobacco carcinogenesis and one of the most (Canada, 2000; Borgerding et al., 2000; Government of
important contributors to scientific progress in this field. British Columbia, 2007; Burns et al., 2008).
Since the mid-1980s, he and his co-workers at the
American Health Foundation in New York State have
The changing cigarette
published several lists with varying numbers of biologi-
cally or toxicologically active substances in mainstream
Since the early 1950s, epidemiological studies have
smoke (e.g. Hoffmann, 1993; Hoffmann and Hoffmann,
reported an association of cigarette smoking with cancer
1998; Hoffmann et al., 1997, 2001). In recognition of
of the lung (Doll and Hill, 1950; Wynder and Graham,
Dietrich Hoffmann’s work these compounds generally
1950). Both studies revealed a dose–response relationship
are referred to as ‘‘Hoffmann analytes’’. The list of
between the type and number of cigarettes smoked and
analytes compiled of over 40 compounds and 2 other
the risk for developing lung cancer. Dermal tumor assays
parameters is shown in Table 5. Only three of these
and animal inhalation studies supported the dose–respon-
constituents (tar, nicotine, CO) are produced in mi-
se relationship between the amount of inhaled cigarette
lligram-per-cigarette quantities. Twenty-five of the
smoke or tar and the induction of tumors (Wynder et al.,
Hoffmann analytes (including 1,3-butadiene, formal-
1953; Wynder and Mann, 1957; Dontenwill et al., 1973).
dehyde, acetaldehyde and benzene) are produced in
Based on these findings, it was expected that the reduction
microgram-per-cigarette quantities and the remainder in
of cigarette smoke yields of tar and nicotine would be of
nanogram-per-cigarette quantities.
significant benefit for the smoker.
Currently, there is a strong focus on Hoffmann analytes
The introduction of several technological changes and
both in the industry and by the authorities (Borgerding
design modifications during the last 30–40 years resulted
et al., 2000; Gregg et al., 2004). Several scientists and
in considerably lower smoke yields of cigarettes (both
public health authorities believe that the reduction of
tar and nicotine) with concomitant changes in taste
Hoffmann analytes in cigarette smoke would be an
(Fig. 4). The introduction of filter cigarettes in the 1950s
effective step in the development of a less hazardous
was an important step in this direction. The selective
cigarette. At the moment, a major discussion is going on
reduction of phenols in cigarette smoke with cellulose
regarding the measurement of and imposition of ceilings
acetate filters was considered as important for the
on these over 40 compounds (Purkis and Bailey, 2003).
consumer by many scientists. The use of reconstituted
Hoffmann analytes can be considered as representative
tobacco (e.g., a paper-like material made from tobacco
of the compounds in smoke that are responsible for smoke
particles and binding additives) and of expanded
toxicity. Though limited in number, they may be seen as
tobacco (tobacco that has been puffed-up and expan-
indicator substances although the validity of this assump-
ded, and takes up much more volume per weight than
tion is not always clear, as in the case of benzo(a)pyrene for
ordinary tobacco) in cigarette blends contributed further
the whole group of polycyclic aromatic hydrocarbons.
to moving towards lower tar and nicotine yields.
Certain Hoffmann analytes may also reflect the levels, and
Furthermore, porous paper and filter ventilation lead
any changes, of those smoke constituents, which are
to higher dilution of smoke with the result of low tar
actually not measured but have similar structures or
cigarettes (reviewed by Hoffmann et al., 2001).
similar precursors or similar mechanisms of formation.
However, there is little information on single
Hoffmann analytes and their individual contribu-
tion to the in vitro or in vivo toxicological effects of Machine-smoking and human smoking behavior
whole tobacco smoke. In addition, the focus on single
compounds may obscure interactions in biological Because tar as a whole is believed to be the major
systems. It was found that correlations between changes carcinogen and nicotine the addictive agent in tobacco
 ARTICLE IN PRESS
148 A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156

Table 5. The ‘‘Hoffmann analytes’’ currently required or proposed in regulatory reporting schemes

No. Chemical group Smoke constituent Canada Massachusetts Level

1 Ammonia X X mg/cig
2 Hydrogen cyanide X X mg/cig
3 Nitric oxide X X mg/cig
4 Aromatic amines 2-aminonaphthalene X X ng/cig
5 1-aminonaphthalene X X ng/cig
6 4-aminobiphenyl X X ng/cig
7 3-aminobiphenyl X X ng/cig
8 PAH Benzo[a]pyrene X X ng/cig
9 Volatile carbonyls Formaldehyde X X mg/cig
10 Acetaldehyde X X mg/cig
11 Acetone X X mg/cig
12 Acrolein X X mg/cig
13 Propionaldehyde X X mg/cig
14 Crotonaldehyde X X mg/cig
15 Methyl ethyl ketone X X mg/cig
16 Butyraldehyde X X mg/cig
17 Trace metals Mercury X X ng/cig
18 Nickel X ng/cig
19 Lead X X ng/cig
20 Cadmium X X ng/cig
21 Chromium X ng/cig
22 Arsenic X ng/cig
23 Selenium X ng/cig
24 Tobacco-specific nitrosamines N-nitrosonornicotine (NNN) X X ng/cig
25 4-(N-nitrosomethylamino)-1-(3-pyridyl)-1-butanone (NNK) X X ng/cig
26 N-nitrosoanatabine (NAT) X X ng/cig
27 N-nitrosoanabasine (NAB) X X ng/cig
28 Phenols Hydroquinone X X mg/cig
29 Resorcinol X X mg/cig
30 Catechol X X mg/cig
31 Phenol X X mg/cig
32 m- + p-Cresol X X mg/cig
33 o-Cresol X X mg/cig
34 Semi-volatiles Pyridine X X mg/cig
35 Quinoline X X mg/cig
36 1,3-Butadiene X X mg/cig
37 Isoprene X X mg/cig
38 Acrylonitrile X X mg/cig
39 Benzene X X mg/cig
40 Toluene X X mg/cig
41 Styrene X X mg/cig
42 Nicotine X X mg/cig
43 Tar X X mg/cig
44 Carbon monoxide X X mg/cig
45 pH X n/a
46 Filter efficiency X n/a

smoke, measurements of tar and nicotine were chosen as The yields of tar and nicotine in mainstream smoke of
characteristic analytical parameters for each marketed a given cigarette brand, when printed on the pack, are
brand. Carbon monoxide (CO) has recently been added measured by smoking the cigarette in a smoking
as a major smoke constituent of concern. machine under standardized conditions. The first

A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
40
Filters
35
Reconstituted Tobacco
30
Porous Paper
average "tar" mg/cig
25
20
15
10
5
0 0
1954 1960
Fig. 4. The changing cigarette. US sales weighted average Federal Trade Commission tar and nicotine yields and selected product
innovations (Davis and Nielsen, 1999).
smoking machine standard was specified by the Federal
Trade Commission (FTC), a federal government agency
in the USA, in 1966. Soon, other standard methods were
developed, for example the CORESTA (Cooperation
Centre for Scientific Research Relative to Tobacco;
www.coresta.org) Recommended Method in 1969 and
the DIN method in Germany in 1978. They were similar
in many aspects to the FTC method but differed with
respect to butt length and smoke collection (reviewed by
Baker, 2002).
The differences between the existing standards and
methods were harmonized by agreeing on a new set of
ISO (International Organization for Standardization)
Standard Methods in 1991, after a considerable amount
of interlaboratory comparisons of the methods under
development had been undertaken under the leadership
of CORESTA. These ISO methods are not really
different from the FTC methods in use today and define
the procedures for sampling, conditioning, smoking and
the tolerances of the measured numbers. The results
obtained are the basis for comparing smoke yields of
different cigarettes. The parameters of the ISO/FTC
regimen for the machine-smoking of cigarettes are
taking one puff per minute with a 35 ml volume over
2 s and smoking to a defined butt length. The ISO/FTC
testing regimens are mandatory in many countries. ISO
smoke yields also serve as a regulatory limit in a number
of countries and regions, including the European Union,
where brands that generate yields above 10 mg tar, 1 mg
nicotine or 10 mg carbon monoxide (CO) per cigarette
are now prohibited.
ARTICLE IN PRESS
149
5
4.5
4
3.5
average nicotine mg/cig
Expanded Tobacco
3
2.5
2
Filter Ventilation
1.5
Charcoal Filter 1
0.5
1966 1972 1978 1984 1990 1996
"tar" nicotine
Smoking by a machine is different from the smoking
behavior of humans. Smokers are able to vary their
smoke intake over a broad range by adjusting the
following parameters to their personal needs: selection
of brand, number of cigarettes smoked, number of puffs
per cigarette, puff frequency, puff volume, duration of
puff and depth and duration of inhalation.
Fig. 5 shows the considerable inter-individual varia-
tion in nicotine uptake among smokers, as published by
Jarvis et al. (2001). Some people are simply lighter
smokers and desire lower levels of nicotine, whereas
others who smoke the same brand have higher nicotine
needs and either smoke a larger number of cigarettes or
smoke cigarettes more intensely. The nicotine taken up
by smokers does not correlate with the machine-
smoking yields (Borgerding and Klus, 2005). Not only
the uptake of nicotine, but also the uptake of other
smoke constituents is not well correlated with machine-
smoking data (Scherer et al., 2007). This is caused by the
adjustment of individual smoking behavior to different
types of cigarettes, called compensation (Scherer, 1999).
The purpose of the smoking machine standards is to
determine the tar, nicotine and carbon monoxide
content of cigarette smoke when the cigarettes are
smoked under precisely defined conditions, and hence to
allow a comparison of the yields from different
cigarettes. These are technical data important for
quality control by the manufacturer. The yields are
not predictive of which yields humans obtain when
smoking, nor were they ever expected to be that, since
no two smokers smoke exactly the same way nor does
 ARTICLE IN PRESS
150 A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156

900

800

700

600
Saliva cotinine (ng/ml)

500

400

300

200

100

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
Nicotine yield (mg)

Fig. 5. Population-based estimates of nicotine uptake versus ISO nicotine yields. Scatter plot relating cigarette nicotine yields and
saliva cotinine concentrations in 2031 smokers participating in the 1998 Health Survey of England. Cotinine ¼ 173.5+138.7
(nicotine yield); r ¼ 0.19; r2 ¼ 0.034 (Jarvis et al., 2001).

Table 6. Different machine-based smoking regimens

Condition Puff volume (ml) Puff duration (s) Puff interval (s) Ventilation blocking (%)

ISO/FTC 35 2 60 0
Massachusetts 45 2 30 50
Canadian ‘‘Intense’’ 55 2 30 100

a smoker smoke cigarettes the same way on every criticized (Hammond et al., 2006, 2007) and it is unclear
occasion. whether alternative smoking regimens, including the
Nevertheless, there is a growing movement to develop Canadian Intense method, are in fact more representa-
new machine-smoking regimens that are more intensive tive of human smoking behavior and provide better
for a more ‘‘realistic’’ determination of smoke yields. In predictors of human exposure.
the late 1990s, the Commonwealth of Massachusetts Daily average tar intake by smokers is generally
and the Canadian Federal Government implemented overestimated by the more intense smoking regimens,
machine-smoking conditions that are more intensive like ‘‘Massachusetts’’ and ‘‘Canadian Intense’’. The
than the ISO/FTC method (Table 6). The Framework overestimation increases with the nominal tar yield of a
Convention on Tobacco Control (FCTC) developed by cigarette. Yields obtained with whichever standardized
the World Health Organization (WHO) includes provi- machine-smoking regimen do not reflect actual exposure
sions for testing and regulating cigarette emissions under of a smoker. Much rather they allow ranking and
Article 9.13. comparison of different brands (Borgerding and Klus,
Consequently, ISO convened a working group (ISO 2005).
TC 126 Working Group 9) to develop recommendations Measurements on the retention of cigarette smoke
for ‘‘ya robust and practical smoking regime that as far constituents in the human respiratory tract have been
as possible is representative of smokers’ behavior’’. undertaken for more than 100 years. The techniques
Meanwhile, the World Health Organization’s Study used have evolved over the years, and there is a certain
Group on Tobacco Product Regulations (WHO Tob- amount of data, which was reviewed by Baker and
Reg) has developed its own set of recommendations, Dixon (2006). The bulk of studies indicate that, on
which are under consideration by the ISO Working average, 60–80% of the particulate matter is retained in
Group 9 as well. However, all these methods are the lungs after inhalation. Whereas nicotine and nitric
 ARTICLE IN PRESS
A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
oxides are retained in the order of 90–100%, the
retention of carbon monoxide is at the level of
55–65%. The retention of aldehydes is approximately
90%.
These retention levels are dependent on the physico-
chemical properties of the smoke constituents like
volatility or water solubility, and on the respiratory
patterns of smoking, such as hold time in the lungs,
inhalation and exhalation volumes and the depth of
inhalation. Drawing parallels with the size and behavior
of other aerosols, only about 20% of fresh mainstream
smoke entering the respiratory tract would be expected
to be retained. The observed retentions of 60–80% are
due to the rapid growth of the smoke aerosol particles
through water absorption in the humid environment of
the lung.
The need for toxicological testing using
cigarette smoke as the test agent
Cigarette smoke is a thoroughly analyzed and
documented complex mixture with a considerable
amount of research already done in the field of toxicity
testing. As the tobacco industry is increasingly subjected
to national and international regulations with consumer
protection in mind, there is a need to develop a unified
approach for toxicity testing of tobacco products.
Sensible and useful regulations should be based on
scientific data. One rationale for regulation is based on
the results of toxicological testing.
There are different approaches that can be used to
assess the toxic effects of cigarette smoke. These include:
knowledge of the chemical composition, in vitro studies,
animal studies, human clinical studies and epidemiolo-
gical studies. The different approaches have their
strengths and limitations and can be depicted as a
pyramid (Fig. 6).
Chemical studies
Knowledge of the chemical composition of cigarette
smoke is an indispensable requirement for assessing
smoke toxicity and the basis for characterizing some of
the most toxic of over 4700 constituents of the mixture.
Today, the determination of the Hoffmann analytes is
practically indispensable and standard in most industry
laboratories.
Special care must be taken in the preparation of
samples for testing. When fresh, un-aged cigarette
smoke is formed it is an extremely complex, dynamic
and reactive system due to its chemical and physical
properties. While collecting and analyzing the smoke,
the material changes, and the complex nature of
cigarette smoke gives rise to sampling and analysis
151
Epidemiology
Human (Clinical) Studies
Animal Toxicity Studies
In Vitro Toxicity Studies
Chemical Composition
Fig. 6. Different approaches to assess the toxic effects of
chemicals or pollutants.
problems (Borgerding and Klus, 2005). This must
definitely be considered when evaluating the toxicity of
smoke. Mainstream smoke is inhaled fresh and un-aged
by the smoker, whereas inhaled sidestream smoke is
aged and diluted. The sampling conditions and facilities
for collecting particulate matter are different for the
testing of mainstream smoke, sidestream smoke and
ETS (environmental tobacco smoke). Ideally, cigarette
mainstream smoke should be tested as it is generated
and inhaled by the smoker. Historically, there is a wide
range of conditions for the use of mainstream smoke as
a test agent. Convincing work, however, requires
mainstream smoke generated by a smoking machine
according to standards like ISO.
Based on data obtained by chemical analysis, there
have been several attempts to apply toxicological risk
assessment methods to cigarette smoke and to establish
hazard ranking for its chemical constituents (Fowles and
Dybing, 2003; Rodgman and Green, 2003; Levasseur
et al., 2004). For this purpose, available potency factors
from acute or chronic exposure studies – however, with
single substances in ambient air – were used for
components in cigarette smoke. There is some, but
limited, value in using these data for the assessment of a
complex mixture, such as tobacco smoke, under the
special exposure conditions of human smoking. These
approaches caution against calculating the toxicity of
cigarette smoke and the risk for the smoker by simple
addition or subtraction of toxic potencies, cancer
potency factors or risk factors of the individual
components. As a rule, of course, a reduced number
and lower concentrations of carcinogens in tobacco
smoke are considered to be beneficial.
 ARTICLE IN PRESS
152 A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
Pyrolysis of single substances is an approach that
provides some insight into what may happen inside a
burning cigarette. Presently, there are no standards for
the pyrolysis of additives. As studies are generally done
without adding any other materials, possible interac-
tions cannot be observed. As expected, the data
available in the literature demonstrate convincingly that
the outcome of pyrolysis studies with single substances is
strongly determined by the chosen experimental condi-
tions like temperature gradients, highest temperature
used, atmosphere and reaction time. For instance, in the
pyrolysis of pure menthol it is observed that various
pyrolysis products like benzo(a)pyrene, phenols and
others are formed depending on the temperature of
pyrolysis (Schmeltz and Schlotzhauer, 1968). But in the
realistic situation of smoking, volatile tobacco additives,
in particular top flavors, are transferred to a high degree
unchanged into the smoke (Baker and Bishop, 2004).
Data show that practically all of the menthol added to
cigarette tobacco (98.9%) is transferred unchanged into
mainstream smoke during smoking (Jenkins et al.,
1970).
The most realistic approach is the burning of
additives in the presence of the tobacco matrix. Such
data were published for single additives and for diff-
erent mixtures of additives (Rodgman, 2002; Rustemeier
et al., 2002; Baker et al., 2004b, c).
In vitro toxicity testing
The use of in vitro toxicity tests with cigarette smoke is
gaining increased attention. There are several approa-
ches described in the literature. To date, none of these
testing methods are validated for cigarette smoke or
other complex inhalable mixtures.
In vitro toxicity testing methods that are sensitive,
reproducible and show dose–effect relationships are
ideally suited for the screening of many pollutants, not
only cigarette smoke, because they are relatively
inexpensive and rapid to perform compared to in vivo
studies. However, it is difficult to expose cells in a way
that mimics in vivo exposure. Toxicological investigation
of tobacco smoke should be carried out under condi-
tions resembling the in vivo situation of smoking as far
as possible. Clear ideas are required to address the
problems of exposure, dosimetry, the determination of
appropriate and relevant effects, and the extrapolation
of data, especially for in vitro studies.
For reasons of economizing the approach, a number
of screening tests for biochemical reactivity (SH index,
radical index, oxidative potential) may be performed
before starting an in vitro testing program.
Toxicologically relevant changes in the composition
of smoke that are not detected by chemical analysis may
be seen in biological assays. These do not necessarily
provide information on the causative agent. Compared
to chemical analysis they are rather unspecific. One
major limitation of in vitro studies is that cells are
removed from their normal environment (neighboring
cells and their signals, tissues and blood vessels). The
origin of the cells (permanent or primary cells, species
and tissue) and their metabolic competence must also be
considered.
For many compounds, good chemical and toxicolo-
gical data are already available. For example, most
tobacco additives that are in use are ‘‘generally
recognized as safe’’ (GRAS) for use in foods and/or
listed on the GRAS list of the Flavor and Extract
Manufacturers Association (FEMA). However, this
does not necessarily mean that the use of these
substances is safe in tobacco products because the
testing of single substances in unburned form is not a
particularly realistic condition for tobacco ingredients,
which may be burned during smoking. In addition,
with this approach no interactions can be observed
with other compounds of tobacco or with tobacco
smoke.
One driving motive for the development and use of
more specific in vitro toxicity tests is the regulation of
ingredients. The use of additives in tobacco products has
been the subject of intensive debate (Bates et al., 1999;
Müller and Röper, 2000; Deutsches Krebsforschungs-
zentrum, 2005). For many years, governments in
countries such as the UK, Germany, Belgium, Switzer-
land and France had legislation or voluntary agreements
listing the ingredients that can be used in tobacco
products as well as those that are prohibited. In 2001,
the European Parliament adopted Directive 2001/37/EC
concerning the harmonization of the regulations of
tobacco products in the Member States of the European
Union. Inter alia, the Directive requires the disclosure of
ingredients in tobacco products and their toxicological
data.
The national implementation of the Directive in
Germany was the reason for the development of a
guidance document by a working group of DIN, the
German Institute for Standardization (Deutsches Insti-
tut für Normung), which published a report (DIN
Technical Report 133 – Toxicological Evaluation of
Additives for Tobacco Products – A Guide, 2004). The
following set of in vitro assays for tobacco smoke is
recommended by DIN: bacterial mutagenicity tests,
cytotoxicity tests (for example, neutral red uptake) and
genotoxicity tests (for example, micronucleus test), both
the latter with mammalian cells. At the moment, it is
unclear how meaningful the three basic tests really are.
However, they are well-established methods and availa-
ble for the timely regulation of tobacco products. Health
Canada has established these three basic tests as a
requirement for the tobacco industry in Canada in
December 2004. A working group of CORESTA has
 ARTICLE IN PRESS
A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
recommended the same basic test battery for toxicity
testing of tobacco products.
The DIN working group recommended that additive
testing should be as practice-oriented as possible.
Accordingly, the testing of additives should be carried
out in a tobacco matrix and their toxicological effects
assessed in burned form. This gives the possibility to
evaluate interactions with the matrix and other com-
pounds.
Many publications about the effects of additives show
minor changes in the composition of mainstream smoke
due to the addition of additives. However, these changes
are not reflected in the biological response in different in
vitro and in vivo test systems (single additives: Stavanja
et al., 2003, 2006; Carmines and Gaworski, 2005;
Carmines et al., 2005; mixture of additives: Carmines,
2002; Baker et al., 2004a). As tobacco itself exhibits high
biological activity the generally small amount of
additives in tobacco products cannot be expected to
change the overall toxicity of tobacco smoke to a
significant degree.
A comprehensive review of the literature on additives
up to 2002 was published by Paschke et al. (2002).
Animal studies
Animal studies are an important tool for assessing
health effects of chemicals in humans. They have all the
advantages of human clinical studies in terms of
assessing causality, biological plausibility and underl-
ying mechanisms but are subjected to fewer formalities
and are less expensive to perform, even if highly
standardized and controlled studies are required. Two
different approaches are important for tobacco smoke:
dermal skin-painting studies have demonstrated the
carcinogenicity of tobacco smoke condensate (Wynder
et al., 1953; Smith et al., 2006). The various inhalation
exposure techniques have their strengths and weaknes-
ses, particularly problems of high-to-low dose extrapo-
lation following inhalation exposures. All rodents are
nose-breathers. Therefore, inhalation of cigarette smoke
differs between rodents and human smokers. Most
carcinogenicity studies with cigarette smoke in mice,
rats, hamsters, dogs and monkeys yield no apparent
positive results (Coggins, 2001, 2007). Tobacco smoke
inhalation studies with A/J mice suggest that tobacco
smoke is a comparatively weak mouse lung carcinogen
(Witschi, 2005).
An example, published by Brown et al., for the
difficulties to correlate the known levels of a constituent
in smoke with the data obtained in in vitro and in vivo
studies and to evaluate the toxicity and risk of a
substance, which under the real conditions of tobacco
use is part of a complex mixture, is the tobacco-specific
nitrosamine, 4-(methylnitrosamino)-1-(3-pyridyl)-1-bu-
153
tanone (NNK). NNK is classified by IARC as a possible
human carcinogen. NNK in pure form is mutagenic in
vitro (Ames test), is carcinogenic in rodents and
produces lung tumors in A/J mice after a single i.p.
injection. But the minimum amount of NNK required
for A/J mouse tumors (extrapolated to man) is in the
smoke of 11 million cigarettes ( ¼ 37 packs/day for 40
years ¼ 2 packs/day for 739 years). Experimental data
show that the mutagenicity and carcinogenicity of NNK
are inhibited by nicotine (abundant in smoke) and other
unidentified smoke components as well as cotinine.
Therefore, it is not surprising that, after reducing the
NNK level in tobacco, the smoke condensate is not less
mutagenic (Ames test) and carcinogenic (skin painting)
than smoke (condensate) with usual NNK levels (Brown
et al., 2003).
Human studies
Human (clinical) studies represent a logical comple-
ment to epidemiological studies. The controlled expo-
sure to the pollutant of interest establishes causality and
minimizes confounders.
There are specific biomarkers for the exposure to
tobacco smoke like cotinine (the main metabolite of
nicotine) in body fluids and NNAL/NNAL-glucuronide
(a metabolite of the tobacco-specific nitrosamine, NNK)
in urine. For the internal dose of the gaseous phase of
tobacco smoke, CO in the exhalate can be used
(reviewed in Benowitz, 1999; Scherer, 2005). Different
susceptibility factors, which include age, gender, ethni-
city, health status and genetic factors, may modify
biomarker responses. Biomonitoring data allow a
precise understanding of an individual’s uptake of
specific smoke components.
Epidemiological studies
The first epidemiological studies which showed the
relationship of tobacco smoke and lung cancer were
conducted more than 60 years ago (Schairer and
Schöniger, 2001). They were followed by the important
studies published in 1950 (Doll and Hill, 1950; Wynder
and Graham, 1950). In 1964, the Surgeon General’s
Report on Smoking and Health (the so-called Terry
Report) reviewed and summarized, for the first time, the
relationship between smoking and health with several
updates on specific topics (US Department of Health,
1964).
Epidemiological studies investigate people in natural
settings and measure real world exposure scenarios. End
points measured in epidemiological studies are highly
relevant and include mortality and various types of
morbidity. In addition, potentially susceptible popula-
tions or acute and chronic effects can be studied.
 ARTICLE IN PRESS
154 A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
However, epidemiological studies have several limita-
tions. The accurate assessment of individual exposure is,
besides time and cost, one of the most important
challenges in epidemiological studies. Nevertheless,
epidemiology is considered the gold standard also for
tobacco smoke health research but cannot realistically
be expected to be a requirement in a toxicity testing
battery.
Potentially reduced exposure products (PREPs)
While prevention and cessation is the most effective
way to eliminate the health risks of cigarette smoking,
the use of cigarettes and other tobacco products will
continue. In spite of stringent smoking restrictions in the
United States, it is expected that, in 2010, approximately
10–15% of the adult population in this country will not
be willing or able to give up tobacco consumption
(Institute of Medicine, 2001). For these people it is most
important to develop products of harm reduction. The
so-called potentially reduced exposure products are
tobacco products that are designed or modified in some
way to reduce the user’s exposure to tobacco toxins.
However, PREPs are fairly new on the market and have
not yet been evaluated in a way to provide scientific data
to conclude that they are associated with a reduced risk
of disease compared to conventional tobacco products.
Surrogate biological markers, which are associated with
tobacco-related diseases, could be used to offer guidan-
ce. Also, the reduction of toxins in smoke should be
reflected in smoke composition and in changes in the
biological activity in vitro and in vivo.
References
American Cancer Society. Tobacco atlas. 2nd ed. Available at:
/http://www.cancer.org/docroot/AA/content/AA_2_5_9x_
Tobacco_Atlas.aspS; 2006.
Baker RR. Smoke chemistry. In: Davis DL, Nielsen MT,
editors. Tobacco: production, chemistry and technology.
Oxford: Blackwell Science; 1999. p. 398–439.
Baker RR. The development and significance of standards for
smoking-machine methodology. Beitr Tabakforsch Int
2002;20:23–41.
Baker RR, Bishop LJ. The pyrolysis of tobacco ingredients. J
Anal Appl Pyrolysis 2004:223–311.
Baker RR, Dixon M. The retention of tobacco smoke
constituents in the human respiratory tract. Inhal Toxicol
2006;17:255–99.
Baker RR, Massey ED, Smith G. An overview of the effects of
tobacco ingredients on smoke chemistry and toxicity. Food
Chem Toxicol 2004a;42:S53–83.
Baker RR, Pereira Jr. DS, Smith G. The effect of tobacco
ingredients on smoke chemistry. Part I: Flavourings and
additives. Food Chem Toxicol 2004b;42:S3–S37.
Baker RR, Pereira Jr. DS, Smith G. The effect of tobacco
ingredients on smoke chemistry. Part II: Casing ingredients.
Food Chem Toxicol 2004c;42:S39–52.
Bates C, Jarvis M, Connolly G. Action on smoking and health,
London. Tobacco additives: cigarette engineering and
nicotine addiction; 1999. p. 1–22.
Benowitz NL. Biomarkers of environmental tobacco smoke
exposure. Environ Health Perspect 1999;107:349–55.
Borgerding MF, Bodnar JA, Wingate DE, Taylor CR, Reid
JR, Sudholz MA, Podraza KF, Hsu FS, Whidby JF, The
1999 Massachusetts Benchmark Study, Final Report.
Available from the Massachusetts Department of Public
Health under the Massachusetts (US) public records law;
2000.
Borgerding M, Klus H. Analysis of complex mixtures –
cigarette smoke. Exp Toxicol Pathol 2005;57:43–73.
Brown BG, Borschke AJ, Doolittle DJ. An analysis of the role
of tobacco-specific nitrosamines in the carcinogenicity of
tobacco smoke. Nonlinearity Biol Toxicol Med 2003;1:
179–98.
Burns DM, Dybing E, Gray N, Hecht S, Anderson C, Sanner
T, O’Connor R, Djordjevic M, Dresler C, Hainaut P, Jarvis
M, Opperhuizen A, Straif K. Mandated lowering of
toxicants in cigarette smoke: a description of the World
Health Organization TobReg proposal. Tob Control
2008;17,132–41.
Canada Tobacco Reporting Regulations, SOR/2000-273. Canada
Gazette Part II. Available at: /http://canadagazette.gc.ca/
partII/2000/20000719/html/sor273-e.htmlS; 2000.
Carmines EL. Evaluation of the potential effects of ingredients
added to cigarettes. Part 1: Cigarette design, testing
approach, and review of results. Food Chem Toxicol 2002;
40:77–91.
Carmines EL, Gaworski CL. Toxicological evaluation of
glycerin as a cigarette ingredient. Food Chem Toxicol
2005;43:1521–39.
Carmines EL, Lemus R, Gaworski CL. Toxicologic evaluation
of licorice extract as a cigarette ingredient. Food Chem
Toxicol 2005;43:1303–22.
Coggins CR. A review of chronic inhalation studies with
mainstream cigarette smoke, in hamsters, dogs, and
nonhuman primates. Toxicol Pathol 2001;29:550–7.
Coggins CR. An updated review of inhalation studies with
cigarette smoke in laboratory animals. Int J Toxicol
2007;26:331–8.
Davis DL, Nielsen MT, editors. Tobacco: production,
chemistry and technology. Oxford: Blackwell Science; 1999.
Deutsches Krebsforschungszentrum (Hrsg.): Die Tabakindus-
triedokumente I: Chemische Veränderungen an Zigaretten
und Tabakabhängigkeit, Heidelberg, 2005.
Dixon M, Lambing K, Seeman JI. Mini-Review: On the
transfer of nicotine from tobacco to the smoker. A brief
review of ammonia and ‘‘pH’’ factors. Beitr Tabakforsch
Int 2000;19:103–13.
Doll R, Hill AB. Smoking and carcinoma of the lung;
preliminary report. Br Med J 1950;2:739–48.
Dontenwill W, Chevalier HJ, Harke HP, Lafrenz U, Reckzeh
G, Schneider B. Investigations on the effects of chronic
cigarette-smoke inhalation in Syrian golden hamsters.
J Natl Cancer Inst 1973;51:1781–832.
 ARTICLE IN PRESS
A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
Dube MF, Green CR. Methods of collection of smoke for
analytical purposes. In: 36th Tobacco Chemists’ Research
conference. Recent Adv Tob Sci 1982:8.
Fowles J, Dybing E. Application of toxicological risk
assessment principles to the chemical constituents of
cigarette smoke. Tob Control 2003;12:424–30.
Gahring LC, Rogers SW. Neuronal nicotinic acetylcholine
receptor expression and function on nonneuronal cells.
AAPS J 2005;7:E885–94.
German Institute for Standardization. DIN technical report
133 – toxicological evaluation of additives for tobacco
products – a guide. 2004. Beuth Verlag.
Government of British Columbia, Ministry of Health. What
is in cigarettes? Canada. Available at: /http://www.
healthservices.gov.bc.ca/ttdr/index.htmlS; 2007.
Green CR, Rodgman A. The Tobacco Chemists’ Research
Conference: a half century forum for advances in analytical
methodology of tobacco and its products. Fifty years of
tobacco science and technology. Recent Adv Tob Sci
1996;22:131–336.
Gregg E, Hill C, Hollywood M, Kearney M, McLaughlin D,
McAdam K, Williams M, Purkis S. The UK Smoke
Constituents Testing Study. Summary of results and
comparison with other studies. Beitr Tabakforsch Int
2004;21:117–38.
Guerin MR, Higgins CE, Griest WH. The analysis of the
particulate and vapour phases of tobacco smoke. IARC
Sci. Publ.; 1987. p. 115–39.
Hammond D, Fong GT, Cummings KM, O’Connor RJ,
Giovino GA, McNeill A. Cigarette yields and human
exposure: a comparison of alternative testing regimens.
Cancer Epidemiol Biomarkers Prev 2006;15:1495–501.
Hammond D, Wiebel F, Kozlowski LT, et al. Revising the
machine smoking regime for cigarette emissions: implica-
tions for tobacco control policy. Tob Control 2007;16:8–14.
Hoffmann D. Analysis of toxic smoke compounds. In:
Toxicity testing plan, vol. 5, Toxicity testing plan for low
ignition-potential cigarette. Available at: /http://www.
cpsc.gov/LIBRARY/FOIA/Foia00/os/testing2.pdfS; 1993.
Hoffmann D, Hoffmann I. Letters to the Editor. Tobacco
smoke components. Beitr Tabakforsch Int 1998;18:49–52.
Hoffmann D, Adams JD, Brunnemann KD, Hecht SS.
Assessment of tobacco-specific N-nitrosamines in tobacco
products. Cancer Res 1979;39:2505–9.
Hoffmann D, Djordjevic MV, Hoffmann I. The changing
cigarette. Prev Med 1997;26:427–34.
Hoffmann D, Hoffmann I, El-Bayoumy K. The less harmful
cigarette: a controversial issue. A tribute to Ernst L.
Wynder. Chem Res Toxicol 2001;14:767–90.
Hoffmann D, Rivenson A, Hecht SS. The biological signifi-
cance of tobacco-specific N-nitrosamines: smoking and
adenocarcinoma of the lung. Crit Rev Toxicol
1996;26:199–211.
International Agency for Research on Cancer (IARC).
Tobacco smoking. Lyon: World Health Organization;
1986 [Monogr Eval Carcinog Risks Hum 38:1–421].
International Agency for Research on Cancer (IARC).
Tobacco smoke and involuntary smoke. Monographs on
the evaluation of carcinogenic risks to humans, vol. 83.
Lyon: World Health Organization; 2004. p. 1–1452.
155
Ingebrethsen BJ. Aerosol studies of cigarette smoke. Recent
Adv Tob Sci 1986;12:54–142.
Institute of Medicine. Clearing the smoke: assessing the science
base for tobacco harm reduction. Washington: National
Academy Press; 2001.
Jarvis MJ, Boreham R, Primatesta P, Feyerabend C, Bryant
A. Nicotine yield from machine-smoked cigarettes and
nicotine intakes in smokers: evidence from a representative
population survey. J Natl Cancer Inst 2001;93:134–8.
Jenkins RWJ, Newman RH, Chavis MK. Cigarette smoke
formation studies: II. Smoke distribution and mainstream
pyrolytic composition of added 14C-menthol. Beitr Tabak-
forsch Int 1970;5:299–301.
Klus H, Kuhn H. Estimation of the nitrosonornicotine in the
smoke from nornicotine-rich cigarettes. Fachliche Mitt
Österr Tabakregie 1974:275–88.
Levasseur G, Fillion J, Kaiserman MJ. Determination of the
exposure levels of smokers to chemicals present in cigarette
smoke. In: 58th tobacco science research conference, 2004.
Müller L, Röper W. It ain’t necessarily so. Beitr Tabakforsch
Int 2000;19:51–4.
Paschke T, Scherer G, Heller WD. Effects of ingredients on
cigarette smoke composition and biological activity: a
literature overview. Beitr Tabakforsch Int 2002;20:107–247.
Purkis SW, Bailey IA. Current measurement reliability of
selected smoke analytes. Beitr Tabakforsch Int 2003;20:
314–24.
Rodgman A. ‘‘Smoke pH’’: a review. Beitr Tabakforsch Int
2000;19:117–39.
Rodgman A. Some studies of the effects of additives on
cigarette mainstream smoke properties. I. Flavorants. Beitr
Tabakforsch Int 2002;20:83–103.
Rodgman A, Green CR. Toxic chemicals in cigarette
mainstream smoke – hazard and hoopla. Beitr Tabakforsch
Int 2003:481–545.
Rodgman A, Perfetti TA. The composition of cigarette smoke:
a catalogue of the polycyclic aromatic hydrocarbons. Beitr
Tabakforsch Int 2006;22:13–69.
Roeper W, Wieczorek R, Intorp M. ‘‘Hoffmann’’ analytes and
cigarette smoke in vitro toxicity revisited – how do the data
compare? In: CORESTA, Kyoto Congress 2004, APSSPT3.
Rustemeier K, Stabbert R, Haussmann HJ, Roemer E,
Carmines EL. Evaluation of the potential effects of
ingredients added to cigarettes. Part 2: Chemical composi-
tion of mainstream smoke. Food Chem Toxicol 2002;40:
93–104.
Schairer E, Schöniger E. Lung cancer and tobacco consump-
tion. Int J Epidemiol 2001:24–7.
Scherer G. Smoking behaviour and compensation: a review of
the literature. Psychopharmacology 1999;145:1–20.
Scherer G. Biomonitoring of inhaled complex mixtures –
ambient air, diesel exhaust and cigarette smoke. Exp
Toxicol Pathol 2005;57:75–110.
Scherer G, Engl J, Urban M, Gilch G, Janket D, Riedel K.
Relationship between machine-derived smoke yields and
biomarkers in cigarette smokers in Germany. Regul
Toxicol Pharmacol 2007;47:171–83.
Schmeltz I, Schlotzhauer WS. Benzo(a)pyrene, phenols and
other products from pyrolysis of the cigarette additive,
(D,L)-menthol. Nature 1968;219:370–1.
 ARTICLE IN PRESS
156 A. Thielen et al. / Experimental and Toxicologic Pathology 60 (2008) 141–156
Slade J. Nicotine delivery devices. In: Orleans T, Slade J,
editors. Nicotine addiction: principles and management.
New York: Oxford University Press; 1993. p. 3–23.
Smith CJ, Perfetti TA, Garg R, Hansch C. Utility of the mouse
dermal promotion assay in comparing the tumorigenic
potential of cigarette mainstream smoke. Food Chem
Toxicol 2006;44:1699–706.
Stavanja MS, Ayres PH, Meckley DR, et al. Toxicological
evaluation of honey as an ingredient added to cigarette
tobacco. J Toxicol Environ Health A 2003;66:1453–73.
Stavanja MS, Ayres PH, Meckley DR, et al. Safety assessment
of high fructose corn syrup (HFCS) as an ingredient added
to cigarette tobacco. Exp Toxicol Pathol 2006;57:267–81.
The European Parliament and the Council of the European
Union. Directive 2001/37/EC of the European Parliament
and of the Council of 5 June 2001 on the approximation of
the laws, regulations and administrative provisions of the
member states concerning the manufacture, presentation
and sale of tobacco products. Off J Europ Commun 2001;
L194:26–34.
Tricker AR. Biomarkers derived from nicotine and its metabo-
lites: a review. Beitr Tabakforsch Int 2006;22:147–75.
US Department of Health. Education and welfare public
health service. Report of the Advisory Committee of the
Surgeon General of the Public Health Service. Smoking and
health, 1964. p. 1–387.
Witschi H. A/J mouse as a model for lung tumorigenesis
caused by tobacco smoke: strengths and weaknesses. Exp
Lung Res 2005;31:3–18.
Wynder EL, Graham EA. Tobacco smoking as a possible
etiologic factor in bronchiogenic carcinoma: a study of 684
proved cases. J Am Med Assoc 1950;143:329–36.
Wynder EL, Hoffmann D. A study of tobacco carcinogenesis.
VII. The role of higher polycyclic hydrocarbons. Cancer
1959;12:1079–86.
Wynder EL, Mann J. A study of tobacco carcinogenesis. III.
Filtered cigarettes. Cancer 1957;10:1201–5.
Wynder EL, Graham EA, Croninger AB. Experimental
production of carcinoma with cigarette tar. Cancer Res
1953;13:855–64.