Comparative assessment of HPHC yields in the Tobacco Heating
System THS2.2 and commercial cigarettes
,D.Taﬁn Djoko, O. Moennikes, C. Jeannet, A. Kondylis, M. Belushkin
Philip Morris International R&D, Philip Morris Products SA, Rue des Usines 56, CH-2000 Neuch^
Received 2 June 2017
Received in revised form
7 August 2017
Accepted 12 August 2017
Available online 14 August 2017
Harmful and potentially harmful
Modiﬁed risk tobacco products
Tobacco Heating System
There has been a sustained effort in recent years to develop products with the potential to present less
risk compared with continued smoking as an alternative for adult smokers who would otherwise
continue to smoke cigarettes. During the non-clinical assessment phase of such products, the chemical
composition and toxicity of their aerosols are frequently compared to the chemical composition and
toxicity of the smoke from a standard research cigarette ethe 3R4F reference cigarette. In the present
study, it is demonstrated that results of these analytical comparisons are similar when considering
commercially available cigarette products worldwide. A market mean reduction of about 90% is observed
on average across a broad range of harmful and potentially harmful constituents (HPHC) measured in the
aerosol of a candidate modiﬁed risk tobacco product, the Tobacco Heating System 2.2 (THS2.2), compared
against the levels of HPHC of cigarettes representative of selected markets; this mean reduction is well in
line with the reduction observed against 3R4F smoke constituents in previous studies.
©2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND
In recent years, much research and development has focused on
products that provide an acceptable alternative to continued
smoking of cigarettes while having the potential to present a
reduced risk compared with continued cigarette smoking. Exam-
ples of such products are electronic cigarettes and a novel class of
heated tobacco products eproducts which heat tobacco to tem-
peratures well below that required for combustion, thereby sub-
stantially reducing the formation of harmful and potentially
harmful constituents (HPHC) compared to the mainstream smoke
One such product ethe Tobacco Heating System 2.2 (THS2.2)
was described recently (Smith et al., 2016). It has been extensively
characterized in non-clinical and clinical studies, and has been
demonstrated to provide lower HPHC yields and a lower in vitro
toxicity of the aerosol in comparison to the smoke of a 3R4F
reference cigarette (Schaller et al., 2016). In addition a substantial
reduction in exposure to HPHC excluding nicotine for adult
smokers switching to THS2.2 as compared to continued smoking of
commercially available cigarettes was reported recently (Haziza
et al., 2016a; Ludicke et al., 2017a).
It is important to consider that such products offer an alterna-
tive to continued cigarette smoking, and as such they should be
evaluated in a comparative manner against cigarettes, i.e. there
must be a starting level against which a reduction is to be achieved.
For most assays in the non-clinical assessment steps, there is a need
to select one speciﬁc cigarette comparator and we selected the
3R4F reference cigarette (Smith et al., 2016).
This cigarette is frequently used in non-clinical studies as a
comparator, it is a standard cigarette designed and manufactured
for research purposes. It is distributed by the Center for Tobacco
Reference Products of the University of Kentucky (Anonymous,
2013). Due to the single point in time manufacturing of the 3R4F
cigarettes from a single set of tobacco lots, as well as controlled
storage conditions, it has been shown to elicit long-term variations
in HPHC yields signiﬁcantly lower than those observed in com-
mercial cigarette products (Eldridge et al., 2015; Belushkin et al.,
2015). Due to the standardized design and consistency of main-
stream smoke deliveries, the choice of the 3R4F reference cigarette
as a ubiquitous comparator is reasonable.
A range of different HPHC yields is typically observed in com-
mercial products (Bodnar et al., 2012; Piad
e et al., 2013;Eldridge
et al., 2017). It is due to differences in terms of cigarette designs,
which can impact the mainstream smoke yields (Siu et al., 2013;
e et al., 2013; Hearn et al., 2010).
E-mail address: Guy.Jaccard@pmi.com (G. Jaccard).
Contents lists available at ScienceDirect
Regulatory Toxicology and Pharmacology
journal homepage: www.elsevier.com/locate/yrtph
0273-2300/©2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Regulatory Toxicology and Pharmacology 90 (2017) 1e8
This is also reported to be the case in biological assays, although
the discriminatory power of chemical analysis of HPHC in cigarette
smoke is higher (Oldham et al., 2012), it has been shown that the
in vitro toxicological activity of cigarette smoke is also inﬂuenced by
cigarette design parameters. The impact of the tobacco type and
blend is well established (Bombick et al., 1998; Roemer et al., 2012;
Schramke et al., 2006; Yauk et al., 2012) and for example cigarette
diameter, ﬁlter ventilation or presence of activated charcoal in the
ﬁlter have been shown to also have an impact (McAdam et al., 2016;
Rickert et al., 2007).
This raises the question, however, of how representative the
results of aerosol composition and toxicity comparisons against the
3R4F reference cigarette are when considering commercially
available cigarette products.
To address the question above, we assessed the aerosol
composition of THS2.2 compared against the mainstream smoke
composition of 3R4F and commercial cigarettes from selected
countries, on the basis of common lists of HPHC, using the Health
Canada intense smoking regime (Health Canada, 1999) to generate
The comparative assessment of composition that was per-
formed is based on the Health Canada list of HPHC (Health Canada,
2000) and subsets of this list: The WHO Study Group on Tobacco
Product Regulation list (TobReg) (World Health Organization,
2015), the U.S. Food and Drug Administration abbreviated list
(FDA, 2012), the Health Canada list of HPHC, HPHC which are
classiﬁed as International Agency for Research on Cancer (IARC)
group 1 carcinogenic compounds and the list of ﬁrst priority toxi-
cants proposed by TobReg (Burns et al., 2008). Those lists include
44, 39, 19, 12 and 9 HPHC respectively. The Health Canada intense
machine smoking regime (Health Canada, 2000) was selected
based on the recommendation of TobReg to assess cigarette smoke
yields under such conditions (Burns et al., 2008), and because it
provides a more meaningful basis for the comparison of emissions
between the two different product categories. Countries were
selected to be representative of major tobacco blends and cigarette
designs, with Australia (an essentially Virginia/ﬂue-cured blended
products market), Germany, selected European Union countries
grouped together (essentially American blended products, con-
taining a mix of ﬂue-cured and air-cured tobaccos markets), South
Korea (a market with a high proportion of low ‘tar’cigarettes),
Japan, and Russia (two diversiﬁed markets in terms of cigarette
designs, ie use of ﬁlters containing activated charcoal, reduced
Although the comparative assessment was calculated on a per
article basis and on nicotine-adjusted basis to cope with the
different reporting requirements (e.g. Canada, USA, Brazil versus
TobReg), it is more appropriate to use the data on a per article basis
for the comparison of THS2.2 and commercial cigarettes smoke
constituents’reduction: According to the results obtained in a 3-
month switching clinical study (Ludicke et al., 2017b), the sub-
jects switching to THS2.2 arm and in the continued smoking of
commercial cigarettes arm had no signiﬁcant difference in their
mean consumption of articles (of THS2.2 and commercial cigarettes
2.1. Commercial cigarettes samples
Samples of commercial cigarette products were purchased be-
tween 2008 and 2016 in so-called Market Map studies at the point
of sale. Products were selected to be representative of the market in
terms of different manufacturers, blend types, ISO ‘tar’level and
cigarette designs (cigarette diameter, ﬁlter type). The number of
samples is provided in Table 1.
2.2. HPHC analysis
The analyses of the constituents of mainstream smoke in com-
mercial cigarettes, 3R4F reference cigarettes and in the aerosol of
THS2.2 were conducted by Labstat International (Kitchener, Ont.,
Canada), an independent ISO 17025 accredited tobacco testing
laboratory, under contract to Philip Morris International. For com-
mercial cigarettes, the list of constituents mandated for regulatory
reporting by Health Canada (“Health Canada list”) was assessed. For
THS2.2, the PMI-58 list of HPHC and analytes (Schaller et al., 2016)
was assessed; this list includes the Health Canada list of constitu-
ents. The comparison of emissions of THS2.2 and commercial cig-
arettes is therefore based on the Health Canada list (Health Canada,
2000)ecurrently the most extensive active regulatory reporting
list for cigarette smoke constituents worldwide.
The generation and collection of THS2.2 aerosol necessitates
adequate adaptations to smoking machines: For aerosol collection,
only linear smoking machines could be used in order to accom-
modate the THS2.2 stick holders. The puff number on smoking
machines was set to 12 puffs which corresponds to the 6 min
duration of the heat stick holder battery while applying a Heath
Canada Intense smoking regime, instead of using butt length
detection. The linear smoking machine was equipped with an
activation bar which activated the heat sticks holder by pressing all
activation buttons simultaneously at the start of the process. An
interval of 30 s was taken between the device activation and the
The quantiﬁcation of constituents in the aerosol also requires
speciﬁc considerations, since the composition of the THS2.2 aerosol
is distinctly different to the composition of cigarette smoke. With
respect to ISO parameters, due to the high water content of the
THS2.2 aerosol, accurate water measurements cannot be obtained
with the ISO standard methods due to its evaporation and
condensation (Ghosh and Jeannet, 2014). As such, water is not
considered further. Nicotine-free dry particulate matter (NFDPM),
is not considered in the comparisons, because this quantity
(International Organization for Standardization, 2000), was devel-
oped speciﬁcally in the context of cigarette smoke analyses and is
not meaningful for product categories that do not involve com-
bustion of tobacco and smoke generation. Indeed, the THS2.2
aerosol has a very high water content which requires special
methodologies deviating from the ISO standard methods for its
quantiﬁcation, and the overall composition of the aerosol is
distinctively different from cigarette smoke (Schaller et al., 2016). It
is mainly composed of water and glycerin, the latter acting as an
aerosol former. Thus even if the appropriate analytical methodol-
ogy for the quantiﬁcation of water in the THS2.2 aerosol were
Sampling year Market Number of products
2008 South Korea 13
2010 South Korea 23
2012 South Korea 17
2015 South Korea 35
2008e2016 Germany 59
2008e2015 Russia 204
2008e2016 Japan 169
2010e2016 Australia 44
2015e2016 EU 111
Note: EU (European Union) in the table is limited to the following EU countries:
Germany, Denmark, France, Italy, Netherlands, Poland, Portugal, Romania, Slovenia
G. Jaccard et al. / Regulatory Toxicology and Pharmacology 90 (2017) 1e82
applied, the resulting value would largely reﬂect the glycerin con-
tent of the aerosol, and could not be interpreted in the same
manner as NFDPM for cigarettes. Therefore, NFDPM is not consid-
For commercial cigarettes, all analyses were performed ac-
cording to the ofﬁcial Health Canada methods (Health Canada,
2000), with the exception of the analysis of tobacco-speciﬁc ni-
trosamines, which was performed as of 2010 by a liquid
chromatography-tandem mass spectrometry method, according to
the Labstat International internal method TMS-135. Analyses were
performed in triplicate, except for mainstream cigarette smoke
yields of ‘tar’, nicotine and carbon monoxide (CO) for which 8
replicates were performed.
For THS2.2 and 3R4F cigarettes, all analyses performed were
based on ofﬁcial Health Canada methods (Health Canada, 2000),
with the exception of two methods: analysis of tobacco-speciﬁc
nitrosamines, which was performed by a liquid chromatography-
tandem mass spectrometry method according to the Labstat In-
ternational internal method TMS-135, and the analysis of polycyclic
aromatic hydrocarbons (e.g. benzo[a]pyrene), which was per-
formed by a gas chromatography-mass spectrometry method ac-
cording to the Labstat International internal method TMS-120. The
ofﬁcial Health Canada methods were slightly modiﬁed to either
enhance detection limits or extend the number of compounds
determined by the method. For THS2.2, analyses were performed in
triplicate, on each of three samples, and the aggregated results are
reported. For 3R4F, average results across more than a year of
analysis are reported. All modiﬁed methods have been validated
and are part of Labstat ISO 17025 scope of accreditation.
2.3. Data treatment
The comparison of chemical composition focuses on the HPHC
which are part of the Health Canada list mandated for regulatory
reporting in Canada or subsets of this list. Aerosol constituents of
THS2.2 have been assessed against the yield of each smoke con-
stituent of commercial cigarette products sampled in each speciﬁc
market for a given year. Within each market, the ﬁrst step consisted
of calculating an average yield reduction between THS2.2 and each
cigarette product, as the average in the reductions among the in-
dividual HPHC levels in the aerosol of THS2.2 compared to their
levels in cigarette smoke. In a second step, we calculated the
aggregate mean, median value, and selected additional percentiles
(2.5th, 25th, 75th, and 97.5th) of the distribution of the products’
average yield reduction values. The summary statistics of average
yield reduction of THS2.2 aerosol constituents are reported for
completeness on a per article and on a nicotine-adjusted basis,
however as previously stated the most relevant comparison is the
one on per article basis.
Additionally, the assessment of the average yield percentage
difference among the Health Canada list of HPHC in THS2.2 was
calculated against the weighted (according to the number of
products) yearly median yields of smoke constituents in commer-
cial cigarettes by country, using all data obtained from 2008 to
2.3.1. Treatment of limit of detection/quantiﬁcation values
For several HPHC, the reported yields for some or all replicates
were below the limit of detection (<LOD) or limit of quantiﬁcation
(<LOQ) of the laboratory analytical methods. Some HPHC are below
LOQ only for THS2.2, such as cadmium, lead, hydrogen cyanide,
resorcinol (see also Table 2 for a complete list), some HPHC are
below LOQ for both commercial cigarettes and THS2.2, such as
nickel, chromium, selenium, while some HPHC are very close to
LOQ for THS2.2 only (for example the aromatic amines) or in both
commercial cigarettes and THS2.2 and may be quantiﬁable or not
(for example mercury). This may be due to the analytical variability,
which is higher when the levels are close to LOQ, or to tobacco lot
variability, and may result in large differences in terms of per-
centage reduction for such constituents. In addition, especially
considering the large time period from 2008 to 2016 throughout
which data were obtained, the LOD and LOQ values differ between
different years for some constituents, both for commercial ciga-
rettes as well as for the 3R4F reference cigarette. In all cases in
which a given HPHC yield was reported <LOD or <LOQ for either or
both cigarette product, reference cigarette, or THS2.2, an estimate
of the median value across the replicates is reported for the indi-
vidual HPHC, however the HPHC was omitted from percentage
difference or further quantitative computations. This approach re-
sults in a conservative estimation of the average percentage
reduction of HPHC in THS2.2 towards commercial cigarettes.
3.1. Comparisons to the 3R4F reference cigarette
The results for the Kentucky reference cigarette 3R4F and
THS2.2 for the individual HPHC are provided in Table 2. Except for
nicotine, there is a reduction of more than 90% for most HPHC of the
Health Canada list, covering a broad range of chemical classes, with
an average reduction of about 92% on a per article basis.
3.2. Comparisons to commercial products
The comparison of HPHC in the THS2.2 aerosol and the main-
stream smoke of commercial cigarettes was performed on one
hand for South Korea with datasets obtained for cigarettes bought
in 2008 and in 2015 to assess the potential stability of average
reduction results over time, and on the other hand for a number of
countries (list of countries and number of products provided in
Table 1) for cigarettes bought from 2008 to 2016.
3.2.1. Korean market analysis
The comparison of the THS2.2 aerosol HPHC content with
commercial Korean cigarette products sampled and analyzed in
2008 and 2015 is summarized and reported in Table 3. HPHC are
grouped in ﬁve major HPHC lists, and the resulting reductions are
calculated and reported for each HPHC list in a separate column in
Table 3. HPHC reductions are based either on per article basis
comparisons or on nicotine-adjusted yield comparisons. The esti-
mated reductions are statistically analyzed using the average as
well as ﬁve percentile values covering the whole range of the
The reductions observed in the levels of THS2.2 aerosol HPHC
compared with the HPHC in the mainstream smoke of commercial
cigarettes bought in 2008 and 2015 are very similar, considering
both results obtained on a per article basis or with nicotine-
adjusted results, whatever list of HPHC is taken into account. Us-
ing the Health Canada list of HPHC as an example, the mean
reduction observed in Korea, on per article basis is 90 and 89%
against the cigarettes bought in 2015 and in 2008, while the mean
reduction observed, on nicotine-adjusted yields is 88% and 86% for
the cigarettes bought in 2008 and 2015, respectively. Those results
are well in line with the average percentage reductions calculated
against the weighted median yields obtained by combining the
results for commercial cigarettes in South Korea and bought in
2008, 2010, 2012 and 2014 (average reduction of 90 and 86% on a
per article basis or on nicotine-adjusted basis respectively, using
the Health Canada list, see Table 4).
The range of reduction expressed by the 2.5th and 97.5th
G. Jaccard et al. / Regulatory Toxicology and Pharmacology 90 (2017) 1e83
percentiles is between 88% and 92% for the per article basis results
and between 81% and 92% for the nicotine-adjusted results (using
the Health Canada list). The range of reduction between the HPHC
in the THS2.2 aerosol and the HPHC in the mainstream smoke of
cigarettes is also illustrated in Figs. 1 and 2.
3.2.2. Worldwide markets analysis
The results on the reduction of the levels of THS2.2 aerosol
HPHC as compared to HPHC in the mainstream smoke of com-
mercial cigarettes worldwide are reported in Table 5. Results are
based on both aerosol/smoke yields measured on a per article basis,
as well as yields based on nicotine-adjusted yields. The results
HPHC yields of THS2.2 expressed as a percentage of the HPHC yields of the 3R4F. Comparison of percentage reduction perarticle according to the Health Canada list, under HCI
THS2.2% Reduction vs. 3R4F, Stick
Nicotine (mg/article) 1.86 0.175 1.14 0.0332 -
Mercury (ng/article) 4.77 0.669 2.04 0.104 57.1%
g/article) 29.3 2.88 10.5 1.63 64.3%
g/article) 83.5 5.55 20.3 0.586 75.6%
g/article) 29.7 6.01 6.14 0.423 79.3%
g/article) 89.8 7.14 14.4 0.68 84.0%
g/article) 1641 258 217 7.85 86.8%
g/article) 123 7.75 13.6 0.662 89.0%
g/article) 85.2 16.7 7.98 0.504 90.6%
g/article) 89.1 6.65 7.2 0.391 91.9%
g/article) 14 1.86 1.12 0.0849 92.0%
g/article) 15.4 3.23 1.05 0.145 93.1%
N-nitrosoanabasine (NAB) (ng/article) 30.2 2.61 1.92 0.182 93.6%
Benzo[a]pyrene (ng/article) 15 1.3 0.939 0.0796 93.7%
g/article) 156 25.4 9.63 0.703 93.8%
N-nitrosoanatabine (NAT) (ng/article) 270 22.9 14 1.13 94.8%
g/article) 690 37.5 35.5 1.84 94.9%
Methyl-ethyl-ketone (MEK) (
g/article) 185 12.3 7.59 0.456 95.9%
N-Nitrosonornicotine (NNN) (ng/article) 283 27.8 10.2 0.486 96.4%
Nitric oxide (NO) (
g/article) 504 29 13.8 0.967 97.3%
4-(Methyl-nitrosamino)-1-(3-pyridyl)-1-butanone (NNK) (ng/
264 26.4 6.75 0.493 97.4%
Nitrogen oxides (NOx) (
g/article) 560 30.6 14.1 0.943 97.5%
Carbon monoxide (CO) (mg/article) 30.2 2.76 0.436 0.0811 98.6%
g/article) 137 16.9 1.82 0.163 98.7%
g/article) 4.15 0.494 0.0393 0.00649 99.1%
g/article) 81.1 8.78 0.544 0.0312 99.3%
g/article) 24.5 3.52 0.158 0.0122 99.4%
g/article) 98.5 9.8 0.342 0.0347 99.7%
4-Aminobiphenyl (ng/article) 2.83 0.434 0.00958 0.0014 99.7%
3-Aminobiphenyl (ng/article) 4.18 0.773 0.0112 0.0031 99.7%
g/article) 894 76.7 2.15 0.202 99.8%
1-Aminonaphthalene (ng/article) 21.6 2.28 0.0407 0.0103 99.8%
2-Aminonaphthalene (ng/article) 16.2 2.54 0.0277 0.00909 99.8%
g/article) 50.5 9.42 <3.29 *
Lead (ng/article) 32.1 4 <1.62 * >95
g/article) 0.431 0.0416 <0.011 * >97.4
Hydrogen cyanide (
g/article) 365 31.2 <4.37 * >98.8
g/article) 12.1 0.897 <0.0646 * >99.5
Cadmium (ng/article) 92.9 10.4 <0.28 * >99.7
Arsenic (ng/article) 7.9
*<1.2 * NA
*<0.055 * NA
Chromium (ng/article) <LOQ
Nickel (ng/article) <LOQ
Selenium (ng/article) 1.95
* 1.57 0.123 NA
Average reduction /Health Canada List 92.27%
Notes: LOQ eLimit of Quantiﬁcation. The calculations can be reported on yields per unit mass nicotine, by dividing each HPHC yield by the average nicotine concentration. The
data show that THS2.2 reduced HPHC levels on average, by more than 90% considering their concentrations on a per article basis, and by about 87% considering their con-
centrations on nicotine basis relative to the reference cigarette.
LOQ eLimit of Quantiﬁcation.
The reported Health Canada list of constituents does not include pH, tar, and Total Particulate Matter (TPM).
THS2.2 is designed to deliver similar levels of nicotine as cigarettes.
A‘*’in place of the standard deviation indicates that constituent levels for some or all replicates were below the limit of quantiﬁcation of the analytical method. In these
cases, the median is shown instead of the mean.
10% of values below the limit of quantiﬁcation. Median value reported in place of the mean.
Not applicable ecould not be quantiﬁed in this study.
Few values below the limit of quantiﬁcation. Median value reported in place of the mean.
80% values below the limit of quantiﬁcation.
Not detected (below the limit of detection of the method).
90% values below the limit of quantiﬁcation.
20% of values below the limit of quantiﬁcation. Median value reported in place of the mean.
G. Jaccard et al. / Regulatory Toxicology and Pharmacology 90 (2017) 1e84
provided in Table 4 are calculated from the weighted median yields
obtained per country using all data from 2008 to 2016.
The results provided in Table 5 are statistically assessed using
the average yield reduction of THS2.2 aerosol HPHC across all
Observed percentage reduction of HPHC in THS2.2 compared with commercial cigarettes in South Korean market sampled and analyzed in 2008 and in 2015. Reported
percentage reductions are based on aerosol/smoke yields measured on both per article (top) and a nicotine-adjusted (bottom) basis.
Per Article basis Per Article basis
Health Canada WHO 39 WHO-9 FDA 18 IARC Health Canada WHO 39 WHO-9 FDA 18 IARC
97.5th Percentile 92% 92% 94% 94% 97% 92% 92% 94% 95% 97%
75th Percentile 91% 91% 93% 94% 96% 91% 90% 93% 94% 95%
Median 89% 89% 93% 93% 96% 90% 90% 92% 93% 95%
Mean 90% 89% 92% 93% 96% 90% 89% 92% 93% 95%
25th Percentile 89% 88% 92% 92% 95% 88% 88% 91% 92% 94%
2.5th Percentile 88% 87% 90% 91% 94% 88% 87% 90% 91% 93%
Nicotine basis Nicotine basis
Health Canada WHO 39 WHO-9 FDA 18 IARC Health Canada WHO 39 WHO-9 FDA 18 IARC
97.5th Percentile 92% 92% 95% 95% 97% 89% 89% 93% 93% 96%
75th Percentile 91% 91% 93% 93% 96% 87% 86% 90% 91% 93%
Median 88% 88% 92% 92% 95% 86% 85% 89% 90% 93%
Mean 88% 88% 91% 92% 95% 86% 85% 89% 90% 93%
25th Percentile 88% 87% 90% 92% 95% 85% 84% 88% 90% 92%
2.5th Percentile 83% 82% 86% 88% 92% 81% 80% 85% 87% 88%
Lists of HPHC according to Health Canada (Health Canada, 2000), WHO TobReg nine ﬁrst priority toxicants (Burns et al., 2008), WHO TobReg non-exhaustive priority list
(World Health Organization, 2015), FDA abbreviated list (FDA, 2012) and IARC group 1 carcinogens.
Observed percentage reduction of HPHC in THS2.2 compared with commercial cigarettes sampled and analyzed between 2008 and 2015, using Health
Canada list of HPHC. Reported percentage reductions are based on aerosol/smokeyields measured on both per article (left) and a nicotine-adjusted (right)
basis and on weighted median smoke constituents’yields in commercial cigarettes.
Country Average % reduction on per article basis Average % reduction on nicotine basis
South Korea 90 86
Japan 91 87
Russia 91 88
Germany 92 86
Australia 90 83
Fig. 1. Box plots summarizing the distribution property of THS2.2 average HPHC yield reductions on a per article basis, against the average yields of individual commercial cigarette
products sampled in 2008 from South Korean Market.
G. Jaccard et al. / Regulatory Toxicology and Pharmacology 90 (2017) 1e85
cigarette products. Five critical percentiles covering the whole
range of the observed reductions are also provided.
The mean reduction observed for the THS2.2 aerosol HPHC to-
wards commercial cigarettes bought in 2015 and 2016, using the
Health Canada list is very similar in all countries with values be-
tween 90% and 92% on a per article basis or between 82% and 87%
with nicotine adjusted values.
Fig. 2. Box plots summarizing the distribution property of THS2.2 average HPHC yield reductions on a per article basis, against the average yields of individual commercial cigarette
products sampled in 2015 from South Korean Market.
Observed percentage reduction of HPHC in THS2.2 compared with commercial cigarettes in major markets worldwide in 2015e2016. Reported percentage distribution
reduction statistics are based on aerosol/smoke yields measured on a per article basis (left) and on nicotine-adjusted smoke yields (right).
Per article basis Nicotine basis
Health Canada WHO 39 WHO-9 FDA 18 IARC Health Canada WHO 39 WHO-9 FDA 18 IARC
Australia 97.5th Percentile 93% 93% 94% 95% 96% 86% 86% 90% 91% 94%
75th Percentile 91% 91% 93% 93% 95% 85% 84% 88% 88% 92%
Median 91% 90% 92% 93% 94% 82% 81% 84% 86% 87%
Mean 91% 90% 92% 93% 93% 82% 81% 84% 86% 87%
25th Percentile 90% 89% 90% 91% 92% 80% 79% 81% 83% 84%
2.5th Percentile 89% 88% 89% 91% 90% 78% 76% 78% 81% 80%
Japan 97.5th Percentile 93% 93% 95% 95% 97% 89% 89% 93% 93% 96%
75th Percentile 92% 91% 94% 95% 96% 87% 86% 91% 91% 94%
Median 91% 91% 93% 93% 95% 85% 85% 90% 90% 93%
Mean 90% 90% 93% 93% 95% 85% 84% 88% 89% 91%
25th Percentile 89% 89% 92% 92% 94% 84% 83% 88% 89% 91%
2.5th Percentile 86% 86% 89% 90% 90% 76% 75% 74% 80% 78%
Russia 97.5th Percentile 92% 91% 94% 95% 97% 91% 91% 95% 94% 97%
75th Percentile 91% 91% 94% 94% 96% 87% 87% 91% 92% 94%
Median 91% 90% 94% 94% 96% 87% 86% 90% 91% 94%
Mean 91% 90% 93% 94% 96% 87% 87% 91% 91% 94%
25th Percentile 91% 90% 93% 93% 95% 86% 86% 90% 91% 94%
2.5th Percentile 90% 90% 92% 93% 95% 85% 85% 89% 90% 93%
EU 97.5th Percentile 93% 93% 95% 96% 97% 90% 90% 93% 93% 95%
75th Percentile 92% 92% 94% 95% 96% 87% 87% 91% 91% 94%
Median 92% 91% 94% 94% 96% 86% 85% 90% 91% 93%
Mean 92% 91% 94% 94% 96% 86% 86% 89% 90% 93%
25th Percentile 91% 90% 93% 94% 95% 85% 84% 88% 90% 92%
2.5th Percentile 89% 89% 91% 92% 94% 82% 82% 85% 88% 89%
G. Jaccard et al. / Regulatory Toxicology and Pharmacology 90 (2017) 1e86
The assessment program for THS2.2 was recently described
(Smith et al., 2016) and followed by a series of publications
providing the results of the non-clinical and parts of the clinical
assessment of THS2.2. Investigations on differences in aerosol
composition compared to the smoke composition of cigarettes,
and differences in biological effects in in vitro assays conducted in
the non-clinical part of the THS2.2 assessment used the University
of Kentucky 3R4F cigarette as a reference cigarette (Schaller et al.,
2016). There are several arguments in favor for that selection: The
availability of the reference cigarette from an independent source;
its widespread use for tobacco research purposes, internally and
externally, which provides a solid data base of results for HPHC
analysis and effects in biological assays supported by a homoge-
neity of the reference cigarettes usually exceeding that of com-
mercial products. There are however also limitations, inter alia the
fact that the 3R4F has been designed as reference for an American
blend type cigarette. Consequently, it remained to be demon-
strated that the commercially available THS2.2 achieves a com-
parable overall reduction in yields of selected HPHC whether
compared with the 3R4F or commercial cigarettes from various
In this work, in comparison with 3R4F cigarettes mainstream
smoke HPHC, the average reduction over all analyzed HPHC in
THS2.2 covering a wide range of chemical classes is found to be 92%
on a per article basis and 87% on a nicotine-adjusted basis. The
results are consistent with what was reported previously for
THS2.2 and 3R4F (Schaller et al., 2016) and with recently published
data related to 3R4F mainstream smoke yields (Margham et al.,
2016; Pazo et al., 2016; Eldridge et al., 2015; Roemer et al., 2012).
In comparison with the HPHC in the mainstream smoke of
commercial cigarettes in speciﬁc markets, the mean reduction
observed for the THS2.2 aerosol HPHC is very similar in all markets
and very close to the reduction for 3R4F (e.g., 90%e92% reduction
for the per article basis results using the Health Canada list in in-
dividual markets and 83% and 88% reduction for the nicotine-
adjusted results, using the weighted median values for commer-
cial cigarettes as a comparison). These results conﬁrm that the use
of 3R4F reference cigarette as a comparator provides a value for the
average reduction in aerosol/smoke yields which is representative
of commercial cigarettes available in various markets in studies
with potentially modiﬁed risk tobacco products, such as THS2.2.
The generally lower reduction observed for the nicotine-
adjusted results against the results obtained on a per article basis
is mainly due to the average higher machine smoking yields of
nicotine from cigarettes. The range of reduction observed varies
between around 75% and 97% for individual products for nicotine-
adjusted results and between 86% and 97% on a per article basis.
The observed range among different commercial cigarettes in a
given country is due to differences in terms of cigarette designs,
which can impact the mainstream smoke yields of commercial
cigarettes. Typically, the inﬂuence of the cigarette diameter (Siu
et al., 2013; McAdam et al., 2016), the blend types (Piad
e et al.,
2013) or the ﬁlter types (Hearn et al., 2010; Piad
e et al., 2015;
Shin et al., 2009) have been pointed out.
If we take the same comparative approach for the HPHC yields
between different commercial cigarette products, we obtain for
example a mean decrease in terms of HPHC of 12% on a per article
basis (or 10% on nicotine-adjusted results) between the cigarettes
from Australia (a typical Virginia blended market) and the ciga-
rettes from selected EU countries (American blended cigarette
markets) in this study, using the 2015-2016 median values of each
HPHC in both markets. This difference is driven mainly by the TSNA,
ammonia and aromatic amines levels which are lower in Virginia
blends than in American blends (Piad
e et al., 2013). Considering
further differences in biological activity of smoke from cigarettes
with different tobacco blends (Belushkin et al., 2014), it is well
known, for example, that Total Particulate Matter (TPM) from Vir-
ginia tobacco is consistently less mutagenic than Burley tobacco in
the Ames bacterial mutagenicity assay (Roemer et al., 2004). On the
contrary, in the mammalian cell based Mouse Lymphoma geno-
toxicity assay, TPM from Virginia tobacco cigarettes is more
mutagenic than that from experimental all Burley tobacco ciga-
rettes (Schramke et al., 2006). Those differences do not, however,
appear to reﬂect any difference in terms of cancer and COPD
occurrence in markets consisting essentially of Virginia blended
cigarettes (such as Australia or UK) or essentially American blended
cigarettes (such as Germany) as shown in a review of epidemio-
logical data (Lee et al., 2009). This is also true for the charcoal
containing ﬁlter cigarettes which have lower amounts of mostly
volatile and semi-volatile compounds but not particulate phase
compounds in cigarette mainstream smoke, when compared with
cellulose acetate containing ﬁlter cigarettes (Shin et al., 2009;
Hearn et al., 2010). In a clinical study with carbon-ﬁltered ciga-
rettes with a high loading of charcoal and test cigarettes with
regular acetate tow ﬁlters, there were no signiﬁcant difference in
the measures of biological effect which were performed (Sarkar
et al., 2008), even though there was a signiﬁcant decrease of the
HPHC present in the gas phase. A review of published data,
including non-clinical, clinical and epidemiological studies,
concluded that current charcoal ﬁlter techniques alone may not be
sufﬁcient to reduce smoking-related disease (Coggins and
In addition, it was recently reported that therewas no consistent
change of biomarkers of effect in smokers of cigarettes which were
speciﬁcally designed for reduced toxicant emissions and delivered
around 50% less on average for a wide range of HPHC than reference
cigarettes (Dittrich et al., 2014; Proctor et al., 2014).
In the case of THS2.2, due to absence of combustion, a signiﬁcant
decrease in formation of about 90% is observed on average across a
broad range of chemical compounds when compared against 3R4F
reference cigarettes and commercial cigarettes available in a
number of countries. The reduced formation for this technology is
also consistent with results on toxicity reductions. In comparison
with 3R4F a signiﬁcant reduction of the cytotoxicity determined by
the neutral red uptake assay and the mutagenic potency in the
mouse lymphoma assay has been observed, while THS2.2 aerosol
was not mutagenic in the Ames assay (Schaller et al., 2016).
Furthermore, a signiﬁcant reduced exposure to HPHC (close to the
levels observed for smoking abstinent subjects) has been observed
for subjects switching to this product in comparison with
continuing to smoke commercially available cigarettes in clinical
trials (Haziza et al., 2016a, 2016b; Ludicke et al., 2017a).
In summary, our results conﬁrm that the average reduction in
aerosol yields shown for the THS2.2 in comparison to the 3R4F
reference cigarette are equally valid when considering commer-
cially available cigarette products from diverse markets worldwide.
This leads to the twofold conclusion, ﬁrstly that about 90% average
HPHC emission reductions of the THS2.2 in comparison to the 3R4F
are sufﬁciently representative of commercial market means and
secondly, that the 3R4F is a reasonable comparator for the assess-
ment of aerosols of potentially modiﬁed risk tobacco products such
as the THS2.2 in the non-clinical phase of evaluation.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
G. Jaccard et al. / Regulatory Toxicology and Pharmacology 90 (2017) 1e87
Transparency document related to this article can be found
online at http://dx.doi.org/10.1016/j.yrtph.2017.08.006.
Anonymous, 2013. Reference products used in tobacco and smoke analyses. Tob. J.
Int. 2/2013, 150e154.
Belushkin, M., Jaccard, G., Kondylis, A., 2015. Considerations for comparative to-
bacco product assessments based on smoke constituent yields. Regul. Toxicol.
Pharmacol. 73 (1), 105e113.
Belushkin, M., Piad
e, J.-J., Chapman, S., Fazekas, G., 2014. Investigating predictability
of toxicological assessments of cigarettes: Analysis of 7 years of regulatory
submissions to Canadian regulatory authorities. Regul. Toxicol. Pharmacol.
ISSN: 0273-2300 68 (2), 222e230. http://dx.doi.org/10.1016/
Bodnar, J.A., Morgan, W.T., Murphy, P.A., Ogden, M.W., 2012. Mainstream smoke
chemistry analyses of samples from the 20 09 US cigarette market. Regul.
Toxicol. Pharmacol. 64, 35e42.
Bombick, D.W., Putnam, K., Doolittle, D.J., 1998. Comparative cytotoxicity studies of
smoke condensates from different types of cigarettes and tobaccos. Toxicol.
Vitro 12 (3), 241e249.
Burns, D.M., 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., 2008.
Mandated lowering of toxicants in cigarette smoke: a description of the World
Health Organization TobReg proposal. Tob. Control 17 (2), 132e141.
Coggins, C.R.E., Gaworski, C.L., 2008. Could charcoal ﬁltration of cigarette smoke
reduce smoking-induced disease? A review of the literature. Regul. Toxicol.
Pharmacol. 50 (3), 359e365.
Dittrich, D.J., Fieblekorn, R.T., Bevan, M.J., Rushforth, D., Murphy, J.J., Ashley, M.,
McAdam, K.G., Liu, C., Proctor, C.J., 2014. Approaches for the design of reduced
toxicant emission cigarettes. SpringerPlus 3 (1), 1e23.
Eldridge, A., Betson, T.R., Gama, M.V., McAdam, K., 2015. Variation in tobacco and
mainstream smoke toxicant yields from selected commercial cigarette prod-
ucts. Regul. Toxicol. Pharmacol. 71 (3), 409e427.
Eldridge, A.C., McAdam, K.G., Betson, T.R., Gama, M.V., Proctor, C.J., 2017. Impact
assessment of WHO TobReg proposals for mandated lowering of selected
mainstream cigarette smoke toxicants. Regul. Toxicol. Pharmacol. 86, 332e348.
FDA, 2012. Harmful and Potentially Harmful Constituents in Tobacco Products and
Tobacco Smoke: Established List. US Food and Drug Administration. http://
Ghosh, D., Jeannet, C., 2014. An improved cambridge ﬁlter pad extraction meth-
odology to obtain more accurate water and “tar”values: in situ Cambridge ﬁlter
pad extraction methodology. Beitrage zur Tabakforschung International/Con-
tributions Tob. Res. 26 (2), 38e49.
Haziza, C., de La Bourdonnaye, G., Merlet, S., Benzimra, M., Ancerewicz, J.,
Donelli, A., Baker, G., Picavet, P., Ludicke, F., 2016a. Assessment of the reduction
in levels of exposure to harmful and potentially harmful constituents in Japa-
nese subjects using a novel Tobacco Heating System compared with conven-
tional cigarettes and smoking abstinence: a randomized controlled study in
conﬁnement. Regul. Toxicol. Pharmacol. 81, 489e499.
Haziza, C., de La Bourdonnaye, G., Skiada, D., Ancerewicz, J., Baker, G., Picavet, P.,
Ludicke, F., 2016b. Evaluation of the Tobacco Heating System 2.2. Part 8: 5-Day
randomized reduced exposure clinical study in Poland. Regul. Toxicol. Phar-
macol. 81 (Suppl. 2), s139es150.
Health Canada, 1999. In: Health Canada (Ed.), Ofﬁcial Method-T115. Determination
of Tar, Nicotine and Carbon Monoxide in Mainstream tobacco Smoke. Tobacco
Control Programme, Modiﬁed: 2007-11-28. Ottawa, Canada.
Health Canada 2000. Canadian Ministry of Justice: Tobacco Reporting Regulations.
SOR/200e273. Registration 2000-06-26. Part 3: Emissions from designated
Hearn, B.A., Ding, Y.S., Vaughan, C., Zhang, L., Polzin, G., Caudill, S.P., Watson, C.H.,
Ashley, D.L., 2010. Semi-volatiles in mainstream smoke delivery from select
charcoal-ﬁltered cigarette brand variants. Tob. Control 19 (3), 223e230.
International Organization for Standardization, 2000. Cigarettes eDetermination of
Total and Nicotine-free Dry Particulate Matter Using a Routine Analytical
Smoking Machine. ISO 4387, Geneva, Switzerland.
Lee, P.N., Forey, B.A., Fry, J.S., Hamling, J.S., Hamling, J.F., Sanders, E.B.,
Carchman, R.A., 2009. Does use of ﬂue-cured rather than blended cigarettes
affect international variation in mortality from lung cancer and COPD? Inhal.
Toxicol. 21 (5), 404e430.
Ludicke, F., Baker, G., Magnette, J., Picavet, P., Weitkunat, R., 2017a. Reduced expo-
sure to harmful and potentially harmful smoke constituents with the Tobacco
Heating System 2.1. Nicotine Tob. Res. 19 (2), 168e175 .
Ludicke, F., Picavet, P., Baker, B., Haziza, C., Poux, V., Lama, N., Weitkunat, R., 2017b.
Effects of switching to the Tobacco Heating System 2.2 menthol, smoking
abstinence, or continued cigarette smoking on biomarkers of exposure: a ran-
domized, controlled, open-label, multicenter study in sequential conﬁnement
and ambulatory settings (Part 1). Nicotine Tob. Res. http://dx.doi.org/10.1093/
Margham, J., McAdam, K., Forster, M., Liu, C., Wright, C., Mariner, D., Proctor, C.,
2016. Chemical composition of aerosol from an e-cigarette: a quantitative
comparison with cigarette smoke. Chem. Res. Toxicol. 29:1662e1678.
McAdam, K., Eldridge, K.A., Fearon, I.M., Liu, C., Manson, A., Murphy, J., Porter, A.,
2016. Inﬂuence of cigarette circumference on smoke chemistry, biological ac-
tivity, and smoking behaviour. Regul. Toxicol. Pharmacol. 82, 111e126.
Oldham, M.J., Haussmann, H.J., Gomm, W., Rimmer, L.T., Morton, M.J.,
McKinney, W.J., 2012. Discriminatory power of standard toxicity assays used to
evaluate ingredients added to cigarettes. Regul. Toxicol. Pharmacol. 62 (1),
Pazo, D.Y., Moliere, F., Sampson, M.M., Reese, C.M., Agnew-Heard, K.A., Walters, M.J.,
Holman, M.R., Blount, B.C., Watson, C.H., Chambers, D.M., 2016. Mainstream
smoke levels of volatile organic compounds in 50 U.S. Domestic cigarette
brands smoked with the ISO and canadian intense protocols. Nicotine Tob. Res.
e, J.J., Jaccard, G., Dolka, C., Belushkin, M., Wajrock, S., 2015. Differences in
cadmium transfer from tobacco to cigarette smoke, compared to arsenic or lead.
Toxicol. Rep. 2, 12e26.
e, J.J., Wajrock, S., Jaccard, G., Janeke, G., 2013. Formation of mainstream ciga-
rette smoke constituents prioritized by the World Health Organization - yield
patterns observed in market surveys, clustering and inverse correlations. Food
Chem. Toxicol. 55, 329e347.
Proctor, C., Shepperd, C., McAdam, K., Elridge, A., Meredith, C., 2014. Integrating
chemical, toxicological and clinical research to assess the potential of reducing
health risks associated with cigarette smoking through toxicant regulation. In:
Tobacco Science Research Conference, vol. 2014.
Rickert, W.S., Trivedi, A.H., Momin, R.A., Wright, W.G., Lauterbach, J.H., 2007. Effect
of smoking conditions and methods of collection on the mutagenicity and
cytotoxicity of cigarette mainstream smoke. Toxicol. Sci. 96 (2), 285e293.
Roemer, E., Schramke, H., Weiler, H., Buettner, A., Kausche, S., Weber, S., Berges, A.,
Stueber, M., Muench, M., Trelles-Sticken, E., Pype, J., Kohlgrueber, K., Voelkel, H.,
Wittke, S., 2012. Mainstream smoke chemistry and in vitro and in vivo toxicity
of the reference cigarettes 3R4F and 2R4F. Beitrage zur Tabakforschung Inter-
national/Contributions Tob. Res. 25 (1), 316e335.
Roemer, E., Stabbert, R., Rustemeier, K., Veltel, D.J., Meisgen, T.J., Reininghaus, W.,
Carchman, R.A., Gaworski, C.L., Podraza, K.F., 2004. Chemical composition,
cytotoxicity and mutagenicity of smoke from US commercial and reference
cigarettes smoked under two sets of machine smoking conditions. Toxicology
195 (1), 31e52.
Sarkar, M., Kapur, S., Frost-Pineda, K., Feng, S., Wang, J., Liang, Q., Roethig, H., 2008.
Evaluation of biomarkers of exposure to selected cigarette smoke constituents
in adult smokers switched to carbon-ﬁltered cigarettes in short-term and long-
term clinical studies. Nicotine Tob. Res. 10 (12), 1761e1772.
Schaller, J.P., Keller, D., Poget, L., Pratte, P., Kaelin, E., McHugh, D., Cudazzo, G.,
Smart, D., Tricker, A.R., Gautier, L., Yerly, M., Reis Pires, R., Le Bouhellec, S.,
Ghosh, D., Hofer, I., Garcia, E., Vanscheeuwijck, P., Maeder, S., 2016. Evaluation of
the Tobacco Heating System 2.2. Part 2: chemical composition, genotoxicity,
cytotoxicity, and physical properties of the aerosol. Regul. Toxicol. Pharmacol.
81 (Suppl. 2), s27es47.
Schramke, H., Meisgen, T.J., Tewes, F.J., Gomm, W., Roemer, E., 2006. The mouse
lymphoma thymidine kinase assay for the assessment and comparison of the
mutagenic activity of cigarette mainstream smoke particulate phase. Toxicology
227 (3), 193e210.
Shin, H.J., Sohn, H.O., Han, J.H., Park, C.H., Lee, H.S., Lee, D.W., Hwang, K.J.,
Hyun, H.C., 2009. Effect of cigarette ﬁlters on the chemical composition and
in vitro biological activity of cigarette mainstream smoke. Food Chem. Toxicol.
47 (1), 192e197.
Siu, M., Mladjenovic, N., Soo, E., 2013. The analysis of mainstream smoke emissions
of Canadian 'super slim' cigarettes. Tob. control 22:e10.
Smith, M.R., Clark, B., Luedicke, F., Schaller, J.P., Vanscheeuwijck, P., Hoeng, J.,
Peitsch, M.C., 2016. Evaluation of the Tobacco Heating System 2.2. Part 1:
description of the system and the scientiﬁc assessment program. Regul. Toxicol.
Pharmacol. 81 (Suppl. 2), s17es26.
World Health Organization, 2015. WHO Study Group on Tobacco Product Regula-
tion: Report on the Scientiﬁc Basis of Tobacco Product Regulation. WHO
Technical Report Series, n. 989,.
Yauk, C.L., Williams, A., Buick, J.K., Chen, G., Maertens, R.M., Halappanavar, S.,
White, P.A., 2012. Genetic toxicology and toxicogenomic analysis of three
cigarette smoke condensates in vitro reveals few differences among full-ﬂavor,
blonde, and light products. Environ. Mol. Mutagen. 53 (4), 281e296.
G. Jaccard et al. / Regulatory Toxicology and Pharmacology 90 (2017) 1e88