Regulatory Toxicology and Pharmacology 64 (2012) S85–S97

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Contents lists available at SciVerse ScienceDirect

Regulatory Toxicology and Pharmacology

journal homepage: www.elsevier .com/locate /yr tph

Reduced exposure evaluation of an Electrically Heated Cigarette SmokingSystem. Part 8: Nicotine bridging – Estimating smoke constituent exposureby their relationships to both nicotine levels in mainstream cigarette smokeand in smokers

H.-Jörg Urban a, Anthony R. Tricker a, Donald E. Leyden b, Natasa Forte a, Volker Zenzen c,Astrid Feuersenger c, Mareike Assink c, Gerd Kallischnigg a, Matthias K. Schorp a,⇑a Philip Morris International R&D, Philip Morris Products S.A., Quai Jeanrenaud 5, 2000 Neuchâtel, Switzerlandb Hardyville, VA, 23070, USAc Philip Morris International R&D, Philip Morris Research Laboratories GmbH, Fuggerstrasse 3, 51149 Cologne, Germany

a r t i c l e i n f o

Article history:Available online 23 August 2012

Keywords:Biomarkers of exposureDistribution analysisHarmful and potentially harmfulconstituentsHPHCNicotine

0273-2300 � 2012 Elsevier Inc.http://dx.doi.org/10.1016/j.yrtph.2012.08.005

⇑ Corresponding author. Fax: +41 58 242 2811.E-mail address: Matthias.Schorp@pmi.com (M.K. S

Open access under CC BY

a b s t r a c t

A modeling approach termed ‘nicotine bridging’ is presented to estimate exposure to mainstream smokeconstituents. The method is based on: (1) determination of harmful and potentially harmful constituents(HPHC) and in vitro toxicity parameter-to-nicotine regressions obtained using multiple machine-smokingprotocols, (2) nicotine uptake distributions determined from 24-h excretion of nicotine metabolites in aclinical study, and (3) modeled HPHC uptake distributions using steps 1 and 2. An example of ‘nicotinebridging’ is provided, using a subset of the data reported in Part 2 of this supplement (Zenzen et al.,2012) for two conventional lit-end cigarettes (CC) and the Electrically Heated Cigarette Smoking System(EHCSS) series-K6 cigarette. The bridging method provides justified extrapolations of HPHC exposure dis-tributions that cannot be obtained for smoke constituents due to the lack of specific biomarkers of expo-sure to cigarette smoke constituents in clinical evaluations. Using this modeling approach, exposurereduction is evident when the HPHC exposure distribution curves between the MRTP and the CC usersare substantially separated with little or no overlap between the distribution curves.

� 2012 Elsevier Inc. Open access under CC BY-NC-ND license.

1. Introduction changes in physical and chemical composition of the smoke over

Mainstream (MS) cigarette smoke constituent yields are nor-mally quantified and reported using a smoking regimen developedby the Federal Trade Commission (FTC) (Federal Register, 1967),and adopted by the International Organization for Standardization(ISO) (International Organization for Standardization, 2000a,b,c),with minor changes. These two protocols are intended solely toprovide comparative information about the level of smoke constit-uents in different brands of cigarettes by puffing cigarettes accord-ing to a convention of analytical standards, but cannot be used topredict smoke constituent uptake by the ‘‘average’’ smoker (Goriand Lynch, 1985; Baker, 2002; Borgerding and Klus, 2005; FederalTrade Commission, 2008a).

MS cigarette smoke consists of an aerosol containing liquiddroplets (particulate phase) suspended in the gas–vapor phase,and is generated by overlapping burning, pyrolysis, pyrosynthesis,distillation, sublimation, and condensation processes, with

chorp).

-NC-ND license.

time (Baker, 1999; Borgerding and Klus, 2005). Nicotine is mainlypresent (>99%) in the particulate phase of the MS smoke aerosol(Seeman et al., 2004), and the nicotine dose obtained from acigarette is subject to substantial inter- and intra-individual differ-ences in smoking behavior (Scherer et al., 2007; Lindner et al.,2011). Smoking a cigarette can be described by distinct physicalprocesses: Puffing, mouth hold, inhalation, and exhalation(Bernstein, 2004). Little buccal absorption of nicotine occurs fromacidic smoke of flue-cured tobacco (pH 5.5–6.0), even when heldin the mouth (Gori et al., 1986), whereas nicotine from more alka-line air-cured tobacco smoke (pH > 6.5) is efficiently absorbedacross the buccal mucosa (Armitage et al., 1978). Regardless ofthe pH of cigarette smoke, extensive retention (90–100%) ofnicotine occurs when the smoke taken into the mouth is inhaled(Robinson and Yu, 2001; Armitage et al., 2004a,b; Feng et al.,2007). Numerous experimental studies have quantified the pul-monary retention of a range of additional HPHC present in boththe gas–vapor and particulate phase of MS cigarette smoke whichis similar or lower than that of nicotine (Dalhamn et al., 1968; Goriet al., 1986; Armitage et al., 2004a,b; Baker and Dixon, 2006; Feng

S86 H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97

et al., 2007; Moldoveanu and St. Charles, 2006; Moldoveanu et al.,2007, 2008a,b,c).

Retained nicotine is extensively metabolized to a number of dif-ferent metabolites that can be quantified in biological fluids ofsmokers (Hukkanen et al., 2005; Tricker, 2006). However, becauseno single smoking machine regimen will adequately reflect humansmoking behavior (Institute of Medicine 2001; Baker, 2002; Bor-gerding and Klus, 2005; World Health Organization Study Groupon Tobacco Product Regulation, 2007, 2008), only poor correlationsexist between smoke nicotine yields determined using machine-smoking protocols and nicotine-derived biomarker of exposureestimates in smokers (Russell et al., 1980; Rickert and Robinson,1981; Benowitz and Jacob, 1984; Gori and Lynch, 1985; Diding,1987; Andersson et al., 1997; Byrd et al., 1998; Jarvis et al., 2001;Ueda et al., 2002; Scherer et al., 2007; Mendes et al., 2009; Lindneret al., 2011). Consequently, the U.S. Federal Trade Commission(FTC) has officially rescinded its guidance for reported smokingmachine tar and nicotine yields (Federal Trade Commission,2008b), while increased interest has occurred within the tobaccocontrol community to develop alternative protocols which betterreflect smoker exposure to harmful and potentially harmful con-stituents (HPHC) (World Health Organization Study Group on To-bacco Product Regulation, 2004, 2007; Hammond et al., 2006,2007; Marian et al., 2009).

In this study, a ‘nicotine bridging’ method is described whichuses experimentally measured human smoking topographyparameters (i.e., puff volume, puff frequency, and puff duration)and 24-h urinary excretion of nicotine metabolites as a measureof nicotine uptake to develop multiple machine smoking protocolsto provide smoke nicotine yields that more closely correspond toexperimentally determined nicotine uptake estimates in smokers(Urban et al., 2008). This concept is used to evaluate HPHC-to-nic-otine and in vitro toxicity-to-nicotine relationships for 2 conven-tional cigarettes (Marlboro, Philip Morris One) (M6UK, PM1) andthe EHCSS-K6 smoked according to ISO and 15 different ma-chine-smoking regimens to reflect ‘human puffing behavior’(Schorp et al., 2012; Zenzen et al., 2012). Human smoking behaviorwas determined using nicotine uptake distributions derived fromnicotine metabolite excretion data obtained in two clinical studies(Tricker et al., 2012; Lindner et al., 2011). The two approaches arethen combined (‘nicotine bridging’) to model HPHC uptake propor-tional to nicotine uptake distributions as a means of assessment ofexposure to HPHC for which biomarkers of exposure are not avail-able (Urban and Schorp, 2006).

2. Materials and methods

2.1. Smoking protocols and test and comparator cigarettes

Study cigarettes were analyzed for tar and nicotine according toISO methods. All study cigarettes were conditioned according toISO standard 3402 (International Organization for Standardization,1991). Conventional cigarettes were smoked on smoking machinesaccording to ISO standard 3308 (International Organization forStandardization, 2000a). Tar, nicotine and carbon monoxide (CO)were determined according to IS0 standards 4387, 10315, and8454, respectively (International Organization for Standardization,2000b,c, 1995). Mainstream smoke from EHCSS cigarettes was gen-erated on a modified smoking machine with a carousel adapted touse the EHCSS series K lighter. The EHCSS smoke generation con-formed to ISO standard 3308, and some slight technical deviationswere required. The ISO yields as declared on the cigarette packag-ing were as follows: Marlboro (M6UK; 6 mg tar, 0.5 mg nicotine,and 7.0 mg CO), Philip Morris One (PM1; 1 mg tar, 0.1 mg nicotine,and 2.0 mg CO), and EHCSS-K6 (5 mg tar, 0.3 mg nicotine, and

0.6 mg CO). Mainstream smoke was also analyzed for 44 additionalHPHC according to ISO conditions plus 15 experimental smokingregimens reflecting ‘human puffing behavior’ (Zenzen et al.,2012). The data used represent a subset of the data set reportedby Zenzen et al. (2012) to illustrate the principle of the ‘nicotinebridging’ method.

2.2. HPHC yields and in vitro toxicological parameters

Five ‘‘toxicological parameters’’ were determined for main-stream smoke total particulate matter (TPM) using the Salmonellatyphimurium reverse mutation assay with tester strains TA98,TA100, and TA1537 with S9 metabolic activation (Organizationfor Economic Co-operation and Development, 1997), and the neu-tral red uptake (NRU) assay for both mainstream smoke gas-vaporphase (GVP) and TPM (Borenfreund and Puerner, 1985).

2.3. Biomarker measurements in urine

Urinary excretion of biomarkers of exposure to nicotine and se-lected cigarette smoke HPHC (1,3-butadiene, 4-(methylnitrosami-no)-1-(3-pyridyl)-1-butanone [NNK], acrolein, benzene, CO,pyrene, and o-toluidine) were determined in a clinical study com-paring exposure of smokers (N = 32 smokers per group) of theM6UK, PM1, and EHCSS-K6 cigarettes (Tricker et al., 2012). Bio-markers of exposure were determined in 24 h urine samples col-lected on study days 3–8 after randomization to the testcigarettes. Nicotine uptake was expressed as excretion of nicotineequivalents in 24 h urine (Neq; the calculated molar sum of the con-centrations of nicotine, cotinine, trans-30-hydroxycotinine, and theirrespective glucuronide conjugates). Nicotine uptake, expressed as24 h urine excretion of Neq, was also determined in a Europeanpopulation of smokers from the UK, Germany, and Switzerland(Lindner et al., 2011). Smokers were randomized according to differ-ent cigarette ISO tar categories (TC): TC1 (1–4 mg ISO tar; N = 409),TC2 (5–7 mg ISO tar, N = 399), and TC3 (8–12 mg ISO tar, N = 387).

2.4. ‘Nicotine bridging’ method

The nicotine bridging method uses a three-step approach asshown in Fig. 1. This approach is based on: (1) MS smoke HPHCor in vitro toxicity parameter-to-nicotine regressions obtainedusing multiple machine smoking protocols, (2) nicotine uptake dis-tributions from clinical studies, and (3) modeled HPHC uptake dis-tributions (Urban and Schorp, 2006). For a 5 days average estimateof nicotine uptake, determined as Neq excretion per day for eachsmoker, the corresponding smoke constituent uptake was esti-mated based on the corresponding regression equation (i.e., usingthe slope and the intercept of the HPHC-to-nicotine yield). In thisway, a modeled HPHC uptake value for each smoker was obtainedwhich can be displayed as a frequency distribution for all smokers.This method is based on the underlying assumptions that (i) pro-portionality exists between the yield of different HPHC to nicotine(Zenzen et al., 2012), and (ii) that uptake of each HPHC is propor-tional to the nicotine uptake distribution. The latter is a rather con-servative assumption as almost complete uptake of nicotine occurswhen cigarette smoke reaches the small airways of the lung, whilethe retention of other HPHC may be lower (Robinson and Yu, 2001;Armitage et al., 2004a,b; Feng et al., 2007).

For each HPHC/in vitro toxicity parameter the intercept (a) andthe slope (b) of the corresponding regression equation was deter-mined using standard linear regression analysis. The coefficientof determination r2 was used to assess the quality of the linearrelationship. Examples of graphical representations of HPHC-to-nicotine relationships and further interpretation are reported byZenzen et al. (2012).

Fig. 1. The concept of ‘Nicotine Bridging’.

H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97 S87

Statistical analyses were performed using SAS statistical soft-ware V. 9.1.3 and later V.9.2 (SAS Institute, Cary, NC). Linear regres-sions were used for untransformed data and transformed data(natural logarithms). Excretion of Neq was modeled using a normaldistribution for CC (PM1 and M6UK) and a lognormal distributionfor the EHCSS-K6 cigarette. The choice of using either a normal or alognormal distribution was based on the measure of ‘skewness’and the Kolmogorov–Smirnov test. Curve fitting was performedusing SAS routine PROC UNIVARIATE.

The concept of ‘Percentage Change’ was used as an index forcomparison of two overlapping distributions in which the CC wasset as the reference (Fig. 2). The mean, which is identical to themode for both normal and lognormal distributions was used tocharacterize the difference of two distributions for a HPHC fromtwo different cigarettes. Thus, the percentage change of the modeof the EHCSS-K6 cigarette to the mean of a CC was used to providea graphical display in which changes between the EHCSS-K6 and

Fig. 2. Theoretical overlay plot of exposure frequency distributions of a smoke constitOverlap.

CC were displayed as a negative value if the exposure was reducedwhile smoking EHCSS-K6, and as a positive value when exposurewas increased while smoking EHCSS-K6.

A second index, the ‘Extent of Overlap’, was used to comparetwo overlapping distributions, as depicted in Fig. 2. The overlaparea under a normal distribution was calculated using the Z-func-tion in which the mean and standard deviation in the Z-functionare properties of the normal distribution, and ‘‘L’’ was defined asthe value of the x-axis of the normal distributions where the fre-quencies of both normal distributions are equal. The percentagedistribution area of overlap was computed using the CDF-FUNC-TION of the SAS program.

3. Results

Of the HPHC determined in MS cigarette smoke, 15 of 49HPHC were constituents were below the limit of quantification

uent from two different cigarettes. Panel (a) Percent Change; Panel (b) Extent of

S88 H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97

in the EHCSS-K6 cigarette, and eight HPHC were below the limitof quantification in the CC (Zenzen et al., 2012). Therefore,regression analysis was restricted to the remaining 34 smoke con-stituents. Linear relationships were observed for HPHC-to-nico-tine and toxicity parameter-to-nicotine yields of the M6UK andPM1 cigarettes. Nineteen of 33 smoke HPHC-to-nicotine yieldswere linear, while the remaining smoke constituents and allin vitro toxicity parameter-to-nicotine yields were best describedby non-linear (exponential) relationships in the EHCSS-K6

Table 1Smoke constituent yield/toxicity parameter (n = 38) to mean nicotine yield regressions.

Cigarette smoke constituent/toxicity parameter M6UK

r2 Intercept Slope

TPM 0.98 �4.04 20.63Tar 0.99 �1.66 14.57Water 0.90 �2.38 5.06Carbon monoxide 0.99 �0.89 16.68

Aliphatic dienes1,3-Butadiene 0.86 �1.22 41.99Isoprene 0.93 �72.24 553.2

AldehydesFormaldehyde 0.94 �14.68 51.40Acetaldehyde 0.94 �108.7 821.2Acrolein 0.92 �11.86 92.02Propionaldehyde 0.93 �6.05 66.25

Acid derivatesAcetamide 0.98 �3.29 12.32Acrylonitrile 0.99 �12.16 37.61

Nitro compounds2-Nitropropane 0.94 �0.79 11.86

Aromatic amineso-Toluidine 0.98 4.36 63.71o-Anisidine 0.95 0.35 2.362-Naphthylamine 0.84 2.56 5.124-Aminobiphenyl 0.96 0.03 1.61

Inorganic compoundsNitrogen oxides 0.97 20.73 146.6Ammonia 0.94 �1.87 14.75Hydrogen cyanide 0.73 �60.99 166.8

Monocyclic aromatic hydrocarbonsBenzene 0.97 1.16 47.93Toluene 0.94 �11.48 92.66Styrene 0.98 �7.08 15.87

N-nitrosaminesNNN 0.75 25.19 48.95NNK 0.76 17.35 34.25

PhenolsPhenol 0.73 �0.19 15.35Catechol 0.96 11.72 47.96

Polycyclic aromatic hydrocarbonsBenz[a]anthracene 0.95 0.71 14.90Benzo[b]fluoranthene 0.96 0.51 7.44Benzo[j]fluoranthene 0.94 0.05 4.76Benzo[a]pyrene 0.95 1.02 8.76Indeno(1,2,3-cd)pyrene 0.94 0.22 4.28Pyrene 0.94 2.93 51.73

Cytotoxicity1/EC50: GVP fraction 0.95 �45.38 158.71/EC50: TPM fraction 0.97 �6.04 167.2

Bacterial mutagenicity (TPM fraction)Strain TA98 with S9 0.98 4887 23520Strain TA100 with S9 0.93 4264 11786Strain TA1537 with S9 0.90 48.20 5349

Abbreviations: GVP, gas–vapor phase; NNK, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butalinear (NL).

a Linear regression model either on untransformed or log-transformed data. A linear mcigarettes. Linear and non-linear models (Ln(y) = a + b�nicotine) were used for the EHCS

cigarette. This suggests that these HPHC reflect a change in aero-sol composition as a function of puffing intensity. The intercept,slope, and coefficient of determination (r2) of all regressions arepresented in Table 1.

Table 2 presents two-sided 95% confidence intervals for theestimated HPHC concentrations and toxicological effects at a nico-tine concentration of 1.08 mg/cigarette for all regressions. The nic-otine concentration of 1.08 mg/cigarette was chosen because itrepresents the 90th percentile of the nicotine uptake distribution

PM1 EHCSS-K6

r2 Intercept Slope r2 Intercept Slope Regressiona

0.99 �0.61 13.96 0.83 3.64 18.83 L0.99 �0.51 11.60 0.74 1.79 0.08 L0.87 �0.11 1.36 0.79 1.84 7.75 L0.87 1.21 11.83 0.83 0.23 2.34 NL

0.94 2.11 37.39 0.65 1.12 1.95 NL0.97 �6.21 591.7 0.52 10.84 88.18 L

0.93 �0.80 16.52 0.77 9.62 40.65 L0.95 �63.53 816.4 0.83 90.28 268.6 L0.94 �5.64 74.31 0.85 11.70 50.85 L0.94 �4.79 67.37 0.78 3.10 13.89 L

0.97 �0.91 7.08 0.97 �0.39 5.71 L0.99 �2.62 33.69 0.79 0.36 2.25 NL

0.90 �1.55 15.64 0.94 1.65 12.54 L

0.97 �2.08 72.05 0.96 0.26 3.75 NL0.96 0.04 3.03 0.92 �0.08 0.69 L0.92 0.73 7.54 0.91 0.03 3.38 NL0.94 0.13 1.50 0.88 �0.08 0.36 L

0.83 58.04 99.34 0.80 10.86 77.89 L0.20 5.78 3.83 0.85 6.41 26.87 L0.92 �21.15 129.6 0.71 �4.16 33.31 L

0.94 �0.05 53.11 0.77 0.20 2.58 NL0.97 �5.16 103.0 0.75 0.84 1.96 NL0.96 �2.08 13.32 0.79 0.12 2.17 NL

0.69 33.33 32.38 0.98 0.02 73.78 L0.92 10.20 41.98 0.99 �0.89 22.63 L

0.99 �2.28 20.84 0.95 0.05 4.71 NL0.99 1.88 58.06 0.93 �3.24 28.48 L

0.98 �0.21 15.03 0.91 0.010 4.67 NL0.98 �0.14 8.19 0.92 0.010 3.82 NL0.99 0.50 4.17 0.95 0.003 4.83 NL0.98 �0.13 9.38 0.88 0.010 3.97 NL0.98 �0.04 4.29 0.89 0.009 3.44 NL0.99 �0.78 52.90 0.84 0.088 3.73 NL

0.92 �17.66 123.0 0.84 15.80 1.51 NL0.99 �9.22 161.4 0.97 7.69 2.29 NL

0.99 �509.3 29591 0.91 146.9 3.54 NL0.96 �637.3 16138 0.72 103.5 3.12 NL0.97 48.84 4895 0.84 32.79 3.32 NL

none; NNN, N-nitrosonornicotine; TPM, total particulate matter; Linear (L); Non-

odel (y = a + b�nicotine) was used with un-transformed data for the M6UK and PM1S-K6 cigarette.

Table 2Comparison of 95% confidence intervals of smoke constituent concentrations/toxicological parameter (n = 39) at a nicotine concentration of 1.08 mg/cig. (equivalent to the90th%ile of nicotine uptake distribution for EHCSS-K6).

Cigarettesmoke constituent/toxicity parameter

Unit M6UK PM1 EHCSS-K6 EHCSS-K6 vs.

Lower95% CL

Upper95% CL

Lower95% CL

Upper95% CL

Lower95% CL

Upper95% CL

M6UK PM1

TPM (mg/cig.) 15.80 20.68 13.61 15.32 19.29 28.65 = ⁄ HTar (mg/cig.) 13.02 15.14 11.35 12.70 9.39 15.97 = =Nicotine (mg/cig.) 1.02 1.14 1.04 1.12 0.96 1.20 = =Water (mg/cig.) 1.64 4.52 0.96 1.76 8.05 12.37 ⁄ H ⁄ HCarbon monoxide (mg/cig.) 15.83 18.43 10.52 17.45 1.52 5.26 ⁄ L ⁄ L

Aliphatic dienes1,3-Butadiene (lg/cig.) 29.69 58.57 35.56 49.42 3.94 21.12 ⁄ L ⁄ LIsoprene (lg/cig.) 392.7 657.7 549.8 716.0 56.78 155.40 ⁄ L ⁄ L

AldehydesFormaldehyde (lg/cig.) 30.84 50.83 13.71 20.36 41.00 66.04 = ⁄ HAcetaldehyde (lg/cig.) 614.6 941.6 689.5 47.0 312.1 448.7 ⁄ L ⁄ LAcrolein (lg/cig.) 66.65 108.4 61.29 87.93 54.60 78.62 = =Propionaldehyde (lg/cig.) 51.02 79.98 55.93 80.01 14.01 22.19 ⁄ L ⁄ L

Acid derivatesAcetamide (lg/cig.) 8.67 11.37 5.76 7.73 5.16 6.40 ⁄ L =Acrylonitrile (lg/cig.) 25.47 31.45 30.66 36.87 2.12 8.08 ⁄ L ⁄ L

Nitro compounds2-Nitropropane (ng/cig.) 9.59 14.44 11.89 18.79 13.40 16.98 = =

Aromatic amineso-Toluidine (ng/cig.) 64.95 81.39 67.23 84.22 9.58 23.57 ⁄ L ⁄ Lo-Anisidine (ng/cig.) 2.44 3.35 2.85 3.77 0.56 0.76 ⁄ L ⁄ L2-Naphthylamine (ng/cig.) 6.30 9.87 7.26 10.48 0.79 2.27 ⁄ L ⁄ L4-Aminobiphenyl (ng/cig.) 1.50 2.05 1.47 2.01 0.25 0.38 ⁄ L ⁄ L

Inorganic compoundsNitrogen oxides (lg/cig.) 155.8 202.3 130.9 199.8 72.57 117.40 ⁄ L ⁄ LAmmonia (lg/cig.) 11.44 16.68 5.40 14.42 29.36 41.51 ⁄ H ⁄ HHydrogen cyanide (lg/cig.) 34.28 204.0 89.07 148.6 19.51 44.11 = ⁄ L

Monocyclic aromatic hydrocarbonsBenzene (lg/cig.) 45.40 60.45 47.14 67.48 1.40 7.24 ⁄ L ⁄ LToluene (lg/cig.) 69.01 108.2 92.43 119.6 3.63 13.46 ⁄ L ⁄ LStyrene (lg/cig.) 8.01 12.10 10.09 14.51 0.64 2.34 ⁄ L ⁄ L

N-nitrosaminesNNN (lg/cig.) 54.09 102.0 51.54 85.06 73.27 86.12 = =NNK (lg/cig.) 37.95 70.74 45.80 65.27 22.11 24.99 ⁄ L ⁄ L

PhenolsPhenol (lg/cig.) 8.40 24.37 18.30 22.17 4.53 15.49 ⁄ L ⁄ LCatechol (lg/cig.) 55.15 71.89 60.69 68.48 23.14 31.89 ⁄ L ⁄ L

Polycyclic aromatic hydrocarbonsBenz[a]anthracene (ng/cig.) 13.84 19.76 14.55 17.50 0.76 3.39 ⁄ L ⁄ LBenzo[b]fluoranthene (ng/cig.) 7.18 9.91 7.88 9.53 0.38 1.25 ⁄ L ⁄ LBenzo[j]fluoranthene (ng/cig.) 4.18 6.19 4.74 5.28 0.30 1.04 ⁄ L ⁄ LBenzo[a]pyrene (ng/cig.) 8.73 12.25 9.10 10.91 0.34 1.70 ⁄ L ⁄ LIndeno(1,2,3-cd)pyrene (ng/cig.) 3.94 5.74 4.10 5.08 0.21 0.73 ⁄ L ⁄ LPyrene (ng/cig.) 47.35 70.25 52.07 60.65 2.16 11.13 ⁄ L ⁄ L

Cytotoxicity1/EC50: GVP fraction (ml/cig.) 87.37 164.7 87.00 143.2 54.1 120.3 ⁄ L ⁄ L1/EC50: TPM fraction (ml/cig.) 147.1 202.1 153.04 177.2 72.2 114.4 ⁄ L ⁄ L

Bacterial mutagenicity (TPM fraction)Strain TA98 with S9 (rev./cig.) 26390 34189 27945 34953 3134 14472 ⁄ L ⁄ LStrain TA100 with S9 (rev./cig.) 12878 21107 13272 20311 837 10829 ⁄ L ⁄ LStrain TA1537 with S9 (rev./cig.) 3687 7963 4392 6279 450 3072 ⁄ L ⁄ L

Abbreviations: CL, confidence limit; GVP, gas–vapor phase; NNK, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone; NNN, N-nitrosonornicotine; and TPM, total particulatematter.=: Overlap of 95% confidence intervals; ⁄: 95% confidence interval of smoke constituent concentration/toxicological effect at a nicotine concentration of 1.08 mg/cig. eithersignificantly higher (H) or significantly lower (L) for EHCSS-K6 than for M6UK or PM1 (level of significance a = 0.05).

H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97 S89

for the EHCSS-K6 cigarette. The confidence interval for EHCSS-K6was compared to the confidence intervals for M6UK and PM1 foreach HPHC and in vitro toxicity parameter. In cases in which therewas no overlap between two confidence intervals, the differencebetween the corresponding smoke constituent and in vitro toxicityparameters of the two cigarettes was regarded as statistically sig-nificant (a = 0.05). At a nicotine concentration of 1.08 mg/cigarette,

74% of the HPHC concentrations and toxicological effects were sta-tistically significantly lower for the EHCSS-K6 cigarette comparedto both, the M6UK and the PM1, suggesting a significant potentialfor reduced exposure at the ‘product level’.

Examining the ‘smoker level’, nicotine uptake distributions forM6UK, PM1, and EHCSS-K6 cigarettes (Fig. 3) were calculated fromclinical study data (Tricker et al., 2012) as an average daily excre-

Fig. 3. Nicotine uptake distributions of M6UK, PM1, and EHCSS-K6. (a)

Fig. 4. Overlay plots of frequency distributions for selected smoke constituents (a)formaldehyde, (b) carbon monoxide, (c) 1,3-butadiene, (d) acrolein, and (e) NNK.

S90 H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97

tion on days 3–8 (N = 192, 181, and 180 values for M6UK, PM1, andEHCSS-K6 smokers, respectively) using a correction factor of 1.25to convert measured Neq to total nicotine excretion (Schepersand Demetriou, 2002). Daily Neq excretion was then divided bythe reported number of cigarettes consumed to obtain the nicotineuptake distributions.

Fig. 4 present examples of histogram plots of distributions of (a)formaldehyde, (b) carbon monoxide, (c) 1,3-butadiene, (d) acrolein,and (e) NNK uptake determined using the ‘nicotine bridging’ meth-od with normal distribution fits for the M6UK and PM1 cigarettes,and log-normal distribution for the EHCSS-K6 cigarette. Theparameters needed to define the fitted curves (mean/mode andstandard deviation) as well as the 10th and 90th percentile forall HPHC/in vitro toxicity parameters are summarized in Table 3.

Percentage changes and the overlap corresponding to the com-parison EHCSS-K6 vs. M6UK and EHCSS-K6 vs. PM1 are estimatedfor HPHC/in vitro toxicity parameters and summarized in Table 4.

For the 39 comparisons presented in Table 4, the means/modesof the HPHC and in vitro toxicity parameters showed a percentage

change of P90% for 20 (51%) and 18 (46%) of EHCSS-K6 vs. M6UK,and EHCSS-K6 vs. PM1 comparisons, respectively. According to theinverse relationship, a similar pattern was observed for the over-lap: smoke constituent and in vitro toxicity parameters displayedan overlap of 610% for 28 (71%) and 24 (61%) of EHCSS-K6 vs.M6UK, and EHCSS-K6 vs. PM1 comparisons, respectively. Waterand ammonia were higher in EHCSS-K6, compared to M6UK andPM1. TPM and formaldehyde were higher in EHCSS-K6, comparedto PM1. For all other HPHC and in vitro toxicity parameters, EHCSS-K6 had a lower mode than the mean of the two CC.

Examining the proportions of the distributions between the‘lower tail’ (610%) of CC vs. the ‘upper tail’ (P90%) of EHCSS-K6,the 10th percentile of M6UK uptake was above the 90th percentileof EHCSS-K6 uptake for 28 of 39 parameters (72%) (Table 3). In 26out of 39 parameters (67%), the 10th percentile of the PM1 uptakewas above the 90th percentile of the EHCSS-K6 (Table 3). In con-trast, for only one of the 39 parameters determined in CC, the10th percentile of the M6UK uptake was above the 90th percentileof PM1.

Biomarkers of exposure to eight selected HPHC including nico-tine (Tricker et al., 2012) were determined in urine samples andthe percentage change estimates based on mean changes fromdays 3–8 compared to those obtained using the ‘nicotine bridging’method (Table 5). As an index of similarity, the ratio between bio-marker of exposure measurements and ‘nicotine bridging’ esti-mates are also presented; a ratio equal to 1.0 indicates similarity.

The differences between the two methods are 68% for EHCSS-K6 vs. M6UK (except for 1,3-butadiene and NNK) and 613% forEHCSS-K6 vs. PM1 (except for 1,3-butadiene, acrolein, and NNK)supporting the validity of the ‘nicotine bridging’ method. For 1,3-butadiene and NNK larger differences between the two methodswere observed (14% for 1,3-butadiene and 16% for NNK in theEHCSS-K6 vs. M6UK comparisons; 32% for 1,3-butadiene and 20%for NNK for the EHCSS-K6 vs. PM1 comparisons). One outlier isthe difference of 69% for acrolein in the comparison EHCSS-K6 vs.PM1 which cannot be explained based on the available data.

Reduced exposure assessment can be extended to the ‘popula-tion level’ by comparing nicotine uptake distributions of the ciga-rettes smoked in the clinical study ‘test population’ (PM1 andM6UK) (Tricker et al., 2012) to nicotine uptake distributions of cig-

Fig. 4. (continued)

H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97 S91

arettes of similar ISO tar categories smoked in a ‘reference’ popula-tion’ (Lindner et al., 2011). As presented in Fig. 5, median nicotineuptake was 0.86, 1.01, and 1.16 mg nicotine/cigarette for the TC1,TC2, and TC3 ISO tar yield cigarette categories, respectively.

A criterion for similarity (test population/reference population)used was the 90% confidence interval of the median nicotineuptake (ratio of medians of test/reference), which should lie withinthe interval of 0.8–1.25. As shown in Fig. 3, median nicotine uptakein the ‘test population’ was 0.62 and 0.99 mg nicotine/cigarette forPM1 and M6UK cigarettes, respectively. In the ‘reference popula-

tion’ the ratio of median nicotine uptake (90% confidence intervals)were 0.722 (90% CI: 0.695–0.780) and 0.977 (90% CI: 0.929–1.088)mg nicotine/cigarette for tar categories TC1 and TC2, respectively,suggesting that the criteria were met for the latter comparison.The apparent low median nicotine uptake of the ‘test population’smoking PM1 (1 mg ISO tar) cigarettes weighs disproportionallyagainst the higher variability and median nicotine uptake in the1–4 mg ISO tar yield range ‘reference population’ (ratio 0.722(90% CI: 0.695–0.780), thus, falling just short of meeting the crite-rion for similarity in this example.

Table 3Nicotine bridging for 34 smoke constituents and 5 toxicity parameters.

Cigarette smoke constituent/toxicityparameter

Unit M6UK PM1 EHCSS-K6

(mean ± SD) 10thPercentiles

90thPercentiles

(mean ± SD) 10thPercentiles

90thPercentiles

(mean ± SD) 10thpercentiles

90thpercentiles

Mode

TPM (mg/cig.) 16.6 ± 5.41 10.0 23.5 9.2 ± 3.12 6.2 13.8 15.5 ± 4.72 10.5 23.9 13.8Tar (mg/cig.) 12.9 ± 3.82 8.3 17.8 7.6 ± 2.59 5.2 11.5 8.2 ± 2.53 5.4 12.7 7.2Nicotine (mg/cig.) 1.0 ± 0.26 0.7 1.3 0.7 ± 0.22 0.5 1.0 0.6 ± 0.25 0.4 1.1 0.5Water (mg/cig.) 2.7 ± 1.33 1.1 4.4 0.8 ± 0.30 0.6 1.3 6.7 ± 1.94 4.7 10.2 6.1Carbon monoxide (mg/cig.) 15.8 ± 4.38 10.5 21.4 9.5 ± 2.64 7.0 13.5 1.2 ± 1.00 0.5 2.8 0.7

Aliphatic dienes1,3-Butadiene (lg/cig.) 40.8 ± 11.02 27.4 54.9 28.4 ± 8.35 20.4 40.8 4.4 ± 2.83 2.3 9.1 3.0Isoprene (lg/cig.) 481.1 ± 145.1 304.3 667.2 409.1 ± 132.2 283.8 606.2 66.5 ± 22.12 42.8 105.9 57.9

AldehydesFormaldehyde (lg/cig.) 36.7 ± 13.49 20.3 54.0 10.8 ± 3.69 7.3 16.3 35.3 ± 10.19 24.4 53.4 31.8Acetaldehyde (lg/cig.) 712.7 ±215.5 450.2 988.9 509.5 ± 182.4 336.6 781.4 259.8 ± 67.37 187.7 379.9 238.7Acrolein (lg/cig.) 80.2 ± 24.15 50.8 111.1 46.5 ± 16.60 30.8 71.3 43.8 ± 12.75 30.1 66.5 39.4Propionaldehyde (lg/cig.) 60.2 ± 17.38 39.0 82.5 42.5 ± 15.05 28.2 64.9 11.9 ± 3.48 8.1 18.1 10.7

Acid derivatesAcetamide (lg/cig.) 9.0 ±3.23 5.1 13.2 4.1 ± 1.58 2.6 6.4 3.2 ± 1.43 1.7 5.8 2.5Acrylonitrile (lg/cig.) 25.5 ± 9.87 13.4 38.1 21.0 ± 7.52 13.9 32.2 1.8 ± 1.42 0.8 4.1 1.1

Nitro compounds2-Nitropropane (ng/cig.) 11.1 ± 3.11 7.3 15.1 9.4 ± 3.49 6.1 14.7 9.6±3.15 6.2 15.2 8.4

Aromatic amineso-Toluidine (ng/cig.) 68.1 ± 16.72 47.7 89.5 48.5 ± 16.09 33.2 72.5 5.0 ± 7.41 1.0 15.0 1.2o-Anisidine (ng/cig.) 2.7 ± 0.62 2.0 3.5 2.2 ± 0.68 1.5 3.2 0.4 ± 0.17 0.2 0.7 0.32-Naphthylamine (ng/cig.) 7.7 ± 1.34 6.0 9.4 6.0 ± 1.68 4.4 8.5 0.5 ± 0.61 0.1 1.3 0.14-aminobiphenyl (ng/cig.) 1.6 ± 0.42 1.1 2.2 1.2 ± 0.33 0.9 1.7 0.2 ± 0.09 0.1 0.3 0.1

Inorganic compoundsNitrogen oxides (lg/cig.) 167.4 ± 38.47 120.5 216.7 127.8 ± 22.19 106.7 160.8 60.0 ± 19.54 39.1 94.8 52.6Ammonia (lg/cig.) 12.9 ± 3.87 8.2 17.8 8.5 ± 0.86 7.7 9.7 23.4 ± 6.74 16.2 35.4 21.1Hydrogen cyanide (lg/cig.) 105.8 ± 43.76 52.5 161.9 69.8 ± 28.95 42.4 113.0 16.9 ± 8.35 7.9 31.7 12.2

Monocyclic aromatic hydrocarbonsBenzene (lg/cig.) 49.1 ± 12.58 33.8 65.2 37.2 ± 11.86 26.0 54.9 1.3 ± 1.20 0.5 3.2 0.7Toluene (lg/cig.) 81.2 ± 24.31 51.6 112.4 67.1 ± 23.00 45.3 101.4 3.3 ± 2.20 1.7 7.0 2.3Styrene (lg/cig.) 8.8 ± 4.16 3.7 14.1 7.3 ± 2.97 4.4 11.7 0.6 ± 0.41 0.3 1.2 0.3

N-nitrosaminesNNN (lg/cig.) 74.2 ± 12.85 58.5 90.6 56.1 ± 7.23 49.2 66.8 46.6 ± 18.50 26.8 79.6 38.2NNK (lg/cig.) 51.6 ± 8.99 40.7 63.1 39.7 ± 9.38 30.8 53.6 13.4 ± 5.67 7.3 23.5 10.7

PhenolsPhenol (lg/cig.) 15.2 ± 4.03 10.3 20.3 12.4 ± 4.66 7.9 19.3 2.6 ± 4.95 0.3 8.3 0.3Catechol (lg/cig.) 59.7 ± 12.59 44.4 75.8 42.6 ± 12.97 30.3 62.0 14.7 ± 7.14 7.1 27.5 10.9

Polycyclic aromatic hydrocarbonsBenz[a]anthracene (ng/cig.) 15.6 ± 3.91 10.9 20.6 10.3 ± 3.36 7.2 15.3 0.5 ± 0.94 0.1 1.6 0.1Benzo[b]fluoranthene (ng/cig.) 8.0 ± 1.95 5.6 10.5 5.6 ± 1.83 3.9 8.3 0.2 ± 0.34 0.0 0.7 0.0Benzo[j]fluoranthene (ng/cig.) 4.8 ± 1.25 3.3 6.4 3.4 ± 0.93 2.5 4.8 0.2 ± 0.34 0.0 0.6 0.0Benzo[a]pyrene (ng/cig.) 9.8 ± 2.30 7.0 12.7 6.5 ± 2.09 4.5 9.6 0.3 ± 0.39 0.0 0.8 0.0Indeno(1,2,3-cd)pyrene (ng/cig.) 4.5 ± 1.12 3.1 5.9 3.0 ± 0.96 2.1 4.4 0.1 ± 0.18 0.0 0.4 0.0Pyrene (ng/cig.) 54.7 ± 13.57 38.1 72.1 36.4 ± 11.82 25.1 54.0 1.6 ± 2.39 0.3 4.9 0.4

S92H

.-JörgU

rbanet

al./Regulatory

Toxicologyand

Pharmacology

64(2012)

S85–S97

Cyto

toxi

city

1/EC

50:

GV

Pfr

acti

on(m

l/ci

g.)

113.

4±

41.6

462

.616

6.7

68.6

±27

.46

42.6

109.

644

.3±

21.0

627

.380

.335

.51/

EC50

:TP

Mfr

acti

on(m

l/ci

g.)

161.

2±

43.8

810

7.8

217.

510

4.1

±36

.05

69.9

157.

839

.7±

31.7

317

.790

.923

.5

Bact

eria

lm

utag

enic

ity

(TPM

frac

tion

)St

rain

TA98

wit

hS9

(rev

./cig

.)28

414

±61

7120

898

3632

520

260

±66

0913

991

3011

422

76±

3165

531.

766

9162

5.1

Stra

inTA

100

wit

hS9

(rev

./cig

.)16

053

±30

9212

286

2001

710

690

±36

0572

7116

064

1089

±12

9532

0.6

2996

401.

2St

rain

TA15

37w

ith

S9(r

ev./c

ig.)

5399

±14

0436

8971

9834

85±

1093

2448

5115

410.

9±

526.

710

8.9

1168

132.

8

Abb

revi

atio

ns:G

VP,

gas–

vapo

rph

ase;

NN

K,4

-(m

eth

yln

itro

sam

ino)

-1-(

3-py

ridy

l)-1

-bu

tan

one;

NN

N,N

-nit

roso

nor

nic

otin

e;an

dTP

M,t

otal

part

icu

late

mat

ter.

Mea

nan

dst

anda

rdde

viat

ion

(SD

)ca

lcu

late

dfr

omth

en

icot

ine

upt

ake

freq

uen

cydi

stri

buti

ons

ofth

eM

6UK

,th

ePM

1,an

dth

eEH

CSS

-K6

ciga

rett

efo

rea

chsm

oke

con

stit

uen

tan

dto

xici

typa

ram

eter

.10t

han

d90

thpe

rcen

tile

sas

wel

las

the

mod

e(f

orEH

CSS

-K6

only

)of

the

fitt

edcu

rves

are

give

n.

H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97 S93

4. Discussion

Reducing toxicants levels in cigarette smoke, normalized permilligram of nicotine, has been advocated by the World HealthOrganization (WHO) as a public health goal (Burns, 2006; Burnset al., 2008; World Health Organization Study Group on TobaccoProduct Regulation, 2008). According to Hammond et al. (2007),‘‘[t]he use of toxin: nicotine ratios appears to be gaining momen-tum as a potential regulatory strategy.’’ Therefore, we used thisconcept to model HPHC exposure reduction for the EHCSS series-K cigarette using a testing strategy incorporating chemical analysisof cigarette smoke (Zenzen et al., 2012) and clinical study data inwhich excretion of nicotine uptake was determined (Trickeret al., 2012) to predict exposure reduction to selected HPHC forwhich biomarkers of exposure are not available (Schorp et al.,2012).

The measurement of biomarkers of exposure in biological flu-ids has become an accepted method to conduct assessments oftobacco smoke exposure (Institute of Medicine 2001, 2012;Hatsukami et al., 2007; World Health Organization Study Groupon Tobacco Product Regulation, 2007; Food and Drug Administra-tion, 2012). Nicotine and its metabolites meet most of the desiredcriteria for a biomarker of exposure to tobacco and tobaccosmoke (Tricker, 2006). However, no suitable biomarkers of expo-sure exist for the majority of the approximately 5300 compoundsidentified in cigarette smoke (Rodgman and Perfetti, 2009). Inmost cases for which a biomarker of exposure exists, the pharma-cokinetics are not understood adequately to justify that a mea-sured biomarker will reliably measure the uptake of the parentsmoke component. In the few cases in which a specific biomarkerexists, such as nicotine-derived biomarkers of exposure, two inde-pendent measures (i.e., HPHC quantification via smoking ma-chines and biomarkers of exposure measurements from bodyfluids) can be linked using a method, which we term ‘nicotinebridging’ (Fig. 1).

The ‘nicotine bridging’ method depends on the assumption thatall cigarette smoke constituents are retained to the same extent asnicotine. However, several experimental studies investigating theretention of individual smoke constituents (Dalhamn et al., 1968;Gori et al., 1986; Armitage et al., 2004a,b; Baker and Dixon,2006; Moldoveanu and St. Charles, 2006; Feng et al., 2007; Mold-oveanu et al., 2007, 2008a,b,c) suggest that this default assumptionrepresents a worst-case situation, since the extent of nicotine up-take (90–100%) exceeds that of many other HPHC. This is assumedto be valid for all types of smoking behavior (e.g., different puffprofiles and inhalation patterns).

The ‘nicotine bridging’ method requires smoke chemistry datafrom cigarettes smoked according to multiple regimens (Zenzenet al., 2012) to construct HPHC-to-nicotine uptake distributions.As such, the ‘nicotine bridging’ method addresses the limitationof the ‘‘one machine-smoking protocol – one yield’’ (i.e., point esti-mate) approach and incorporates population-based variabilityfrom nicotine exposure frequency distributions (Figs. 1 and 5),and requires that the same cigarette is machine-smoked undermultiple combinations and settings for puff volumes, puff dura-tions, inter-puff intervals; an approach that we term machinesmoking under ‘human puffing behavior’ conditions.

Exposure to smoke constituents other than nicotine for whichbiomarkers of exposure cannot be directly assessed are estimatedby establishing a statistical relationship between both HPHC andnicotine. Since biomarkers of exposure to nicotine can be directlymeasured in clinical studies and nicotine uptake distributions cal-culated (Fig. 1), exposure distributions for other HPHC can be esti-mated. Consequently, differences in exposure to HPHC fromdifferent cigarette designs, e.g., in smokers of CC and smokers

Table 4Percentage change (%) and overlap (%) characteristics for 34 smoke constituents and 5 toxicity parameters.

Cigarette smoke constituent/toxicity parameter EHCSS-K6 vs. M6UK comparison EHCSS-K6 vs. PM1 comparison

% Change % Overlap L % Change % Overlap L

TPM (mg/cig) �16.7 84.9 17.2 50.5 39.9 11.9Tar (mg/cig) �44.0 41.3 10.4 �5.6 98.9 9.4Nicotine (mg/cig) �48.3 42.2 0.8 �26.3 80.3 0.6Water (mg/cig) 126.7 16.6 4.5 617.0 0.0 2.2Carbon monoxide (mg/cig) �95.6 1.0 4.8 �92.6 2.6 3.7

Aliphatic dienes1,3-butadiene (lg/cig) �92.7 1.1 13.8 �89.4 3.3 11.3Isoprene (lg/cig) �88.0 1.3 146.7 �85.8 2.5 137.7

AldehydesFormaldehyde (lg/cig) �13.5 86.4 41.3 194.3 2.6 19.3Acetaldehyde (lg/cig) �66.5 10.0 399.9 �53.2 27.4 358.9Acrolein (lg/cig) �50.9 29.4 59.2 �15.4 84.6 50.8Propionaldehyde (lg/cig) �82.3 2.1 22.9 �74.9 8.6 20.0

Acid derivatesAcetamide (lg/cig) �72.3 20.1 5.3 �38.5 70.0 3.6Acrylonitrile (lg/cig) �95.7 3.1 6.0 �94.8 3.0 5.8

Nitro compounds2-Nitropropane (ng/cig) �24.6 74.6 10.2 �11.4 94.1 10.8

Aromatic amineso-Toluidine (ng/cig) �98.3 1.5 26.0 �97.6 5.3 17.3o-Anisidine (ng/cig) �90.5 0.8 1.1 �88.2 4.2 0.92-Naphthylamine (ng/cig) �98.1 0.3 3.4 �97.6 2.0 2.14-Aminobiphenyl (ng/cig) �94.5 1.1 0.6 �92.3 3.2 0.5

Inorganic compoundsNitrogen oxides (lg/cig) �68.6 7.0 100.4 �58.8 10.9 90.8Ammonia (lg/cig) 63.4 28.3 17.3 148.7 0.5 11.1Hydrogen cyanide (lg/cig) �88.4 8.4 38.9 �82.5 14.8 33.2

Monocyclic aromatic hydrocarbonsBenzene (lg/cig) �98.7 0.1 9.1 �98.2 0.7 6.7Toluene (lg/cig) �97.2 0.4 14.0 �96.6 1.0 12.1Styrene (lg/cig) �96.1 5.3 1.7 �95.3 3.9 1.7

N-nitrosaminesNNN (lg/cig) �48.5 32.0 57.9 �31.9 52.2 46.5NNK (lg/cig) �79.3 2.0 30.1 �73.1 9.2 23.9

PhenolsPhenol (lg/cig) �98.4 7.2 7.0 �98.0 14.5 5.1Catechol (lg/cig) �81.8 3.9 33.2 �74.5 16.8 25.3

Polycyclic aromatic hydrocarbonsBenz[a]anthracene (ng/cig) �99.7 0.6 4.6 �99.5 2.4 2.8Benzo[b]fluoranthene (ng/cig) �99.4 0.3 2.1 �99.1 1.6 1.3Benzo[j]fluoranthene (ng/cig) �99.7 0.9 1.4 �99.6 1.6 1.1Benzo[a]pyrene (ng/cig) �99.5 0.2 2.7 �99.3 1.6 1.5Indeno(1,2,3-cd)pyrene (ng/cig) �99.1 0.3 1.1 �98.7 1.6 0.7Pyrene (ng/cig) �99.3 0.3 14.6 �98.9 1.8 9.1

Cytotoxicity1/EC50: GVP fraction (per cig) �68.7 22.7 69.9 �48.3 52.6 56.31/EC50: TPM fraction (per cig) �85.4 8.9 85.2 �77.4 25.4 62.9

Bacterial mutagenicity (TPM fraction)Strain TA98 with S9 (rev./cig) �97.8 1.8 14895 �96.9 5.7 8457Strain TA100 with S9 (rev./cig) �97.5 0.8 8397 �96.2 4.7 3850Strain TA1537 with S9 (rev./cig) �97.5 1.5 1870 �96.2 5.1 1372

Abbreviations: GVP, gas–vapor phase; NNK, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone; NNN, N-nitrosonornicotine; and TPM, total particulate matter.Change (%) andoverlap (%) for the comparison of means for each smoke constituent and toxicity parameter (EHCSS-K6 vs. M6UK and EHCSS-K6 vs. PM1) as well as the interval length (L)corresponding to equal frequency which was used to estimate the overlap.

S94 H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97

switching to use EHCSS-K6, can be estimated based on distributionanalysis of smoke chemistry data and measured nicotine uptake,contributing to the discussion related to what constitutes ‘substan-tial exposure reduction’ (FDA, 2011), i.e., demonstrating the extentof separation of the exposure distribution curves.

Additionally, we investigated the use of nicotine uptake distri-butions for CC as a measure of similarity between two populations;nicotine uptake determined in a confined clinical study ‘test’ pop-ulation (Tricker et al., 2012) and nicotine uptake determined in a‘reference’ population (Lindner et al., 2011). The clinical data used

in the current study were obtained from subjects smoking onlyM6UK, PM1 and EHCSS-K6 cigarettes. Several population-basedexposure studies demonstrate that variations in nicotine uptakebetween individuals as well as within groups of smokers existdue to complex patterns of individual smoking behavior. Therefore,it is not surprising that nicotine biomonitoring results in a widedistribution of nicotine uptake per cigarette ISO nicotine yield(e.g., Ueda et al., 2002; Scherer et al., 2007; Mendes et al., 2009).Thus, we compared the nicotine uptake distributions for M6UKand PM1 cigarettes in a clinical study (Tricker et al., 2012) to data

Table 5Method comparisons: Percentage change in biomarkers of exposure and estimates based on nicotine bridging.

Cigarette smokeconstituent

Change EHCSS-K6 vs. M6UK (%) Change EHCSS-K6 vs. PM1 (%)

Biomarker ofexposure

Nicotinebridging

Ratio (biomarker/bridging)

Biomarker ofexposure

Nicotinebridging

Ratio (biomarker/bridging)

Nicotine 36.3 36.9 0.98 10.7 10.1 1.06Carbon monoxide 85.8 92.3 0.93 81.1 87.2 0.931,3-Butadiene 76.8 89.3 0.86 57.2 84.7 0.68Acrolein 41.7 45.4 0.92 9.9 5.9 1.69o-Toluidine 99.7 92.7 1.07 99.4 89.7 1.11Benzene 90.6 974 0.93 84.0 96.5 0.87NNK 86.1 74.0 1.16 79.5 66.2 1.20Pyrene 100.0 97.0 1.03 100.0 95.5 1.05Mean 0.99 1.07Standard deviation 0.10 0.30

Abbreviation: NNK, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone.

Fig. 5. Nicotine uptake frequency distributions for smokers of different ISO tar yield cigarettes.

H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97 S95

obtained in a larger population of European smokers (Lindner et al.,2011), concluding that similarity essentially exists.

4.1. Reduced exposure substantiation heuristic

The presented ‘nicotine bridging’ approach closes the loop onassessing exposure reduction (Schorp et al., 2012). Overall, a

demonstration of reduced exposure across the product level(‘yield’ as determined under standard smoking regimen and ‘hu-man puffing behavior’), the smoker level (clinical studies todetermine nicotine uptake and other biomarkers of exposure),and the population level (test of equivalence between test andreference populations) provides factual evidence of exposurereduction.

S96 H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97

In the assessment of the EHCSS-K6 vs. selected CC, the testproduct performed as follows:

(1) 74% of HPHC and toxicity parameters were statistically sig-nificantly lower (p = 0.05, 95% CI) in the EHCSS-K6 vs. CCat the 90th percentile of the nicotine uptake distribution ofthe EHCSS-K6 (1.08 mg/cigarette). In other words, at approx-imately the 1 mg/cigarette median nicotine uptake found inmany population-based biomonitoring studies, the testproduct exhibits a significant potential for exposurereduction.

(2) For HPHC uptake proportional to nicotine uptake distribu-tion, 50% of the ‘percent reductions’ are P90% and ‘overlap’is 610% of EHCSS-K6 vs. CC.

(3) For HPHC uptake proportional to nicotine uptake distribu-tion: 50% of the ‘lower tail’ (610%) of the CC distribution isbelow the ‘upper tail’ (P90%) of the EHCSS-K6 distribution,and

(4) Similarity, as assessed by median nicotine uptake ratios forthe two separate ‘‘tar category’’ (TC) smoker populations(90% confidence interval of ratio of medians of test/referencepopulation is within the range of 0.8–1.25) was 0.722 (90%CI: 0.695–0.780) and 0.977 (90% CI: 0.929–1.088) for TC 1and TC 2, respectively.

These examples highlight both the significant product-specificreduced exposure potential of the EHCSS-K6 vs. CC. alreadyachieved, and the quest for continuous improvement of this plat-form benchmarked against the data presented here.

5. Conclusion

This study introduces the concept of ‘nicotine bridging’ to esti-mate the distribution of HPHC uptake in a population of smokersbased on relationships between analytical smoke chemistry dataand the distribution of nicotine uptake in smokers. Because themethod is nicotine-based, and nicotine is the most studied andwell understood smoke constituent with established pharmacoki-netics and biomarker methodologies, the use of nicotine uptakedistributions from smoker populations provides a scientificallysound basis for further analyses. Under the conservative assump-tion that other smoke constituents are retained to the same extent(or lower) than nicotine, modeling of HPHC uptake proportional tonicotine provides an experimental concept that we call ‘nicotinebridging’ for estimating exposure and uptake of HPHC for whichno biomarkers of exposure are available. Although ‘nicotine bridg-ing’ is still an experimental concept that requires further valida-tion, we intend to use this heuristic to evaluate exposurereduction as a necessary, but not sufficient component to justifythe potential for risk reduction.

Conflict of interest statement

All authors are or were Philip Morris International (PMI) R&Demployees or worked for PMI R&D under contractual agreements.The work reported in all eight parts of this supplement was fundedby PMI R&D.

Acknowledgements

We would like to thank Drs. Wolf Reininghaus, Hans-JürgenHaussmann, and Lynda Conroy-Schneider for their contributionand review of the manuscript.

References

Andersson, G., Vala, E.K., Curvall, M., 1997. The influence of cigarette consumptionand smoking machine yields of tar and nicotine on the nicotine uptake and oralmucosal lesions in smokers. J. Oral Pathol. Med. 26, 117–123.

Armitage, A., Dollery, C., Houseman, T., Kohner, E., Lewis, P.J., Turner, D., 1978.Absorption of nicotine from small cigars. Clin. Pharmacol. Ther. 23, 143–151.

Armitage, A.K., Dixon, M., Frost, B.E., Mariner, D.C., Sinclair, N.M., 2004a. Theeffect of inhalation volume and breath-hold duration on the retention ofnicotine and solanesol in the human respiratory tract and on subsequentplasma nicotine concentrations during cigarette smoking. Beitr. Tabakforsch.Int. 21, 240–249.

Armitage, A.K., Dixon, M., Frost, B.E., Mariner, D.C., Sinclair, N.M., 2004b. The effectof tobacco blend additives on the retention of nicotine and solanesol in thehuman respiratory tract and on subsequent plasma nicotine concentrationsduring cigarette smoking. Chem. Res. Toxicol. 17, 537–544.

Baker, R.R., 1999. Smoke chemistry. In: Davis, E.L., Nielsen, M.Y. (Eds.), Tobacco,Production, Chemistry and Technology. Blackwell Science, Oxford, pp. 398–439.

Baker, R.R., 2002. The development and significance of standards for smoking-machine methodology. Beitr. Tabakforsch. Int. 20, 23–41.

Baker, R.R., Dixon, M., 2006. The retention of tobacco smoke constituents in thehuman respiratory tract. Inhal. Toxicol. 17, 255–299.

Bernstein, D., 2004. A review of the influence of particle size, puff volume, andinhalation pattern on the deposition of cigarette smoke particles in therespiratory tract. Inhal. Toxicol. 16, 675–689.

Benowitz, N.L., Jacob, P., 1984. Daily intake of nicotine during cigarette smoking.Clin. Pharmacol. Ther. 35, 499–504.

Borenfreund, E., Puerner, J.A., 1985. Toxicity determined in vitro by morphologicalalterations and neutral red absorption. Toxicol. Lett. 24, 119–124.

Borgerding, M., Klus, H., 2005. Analysis of complex mixtures – cigarette smoke. Exp.Toxicol. Pathol. 57 (Suppl. 1), 43–73.

Burns, D.M., 2006. Experience ISO Machine Measurements of Cigarette Yields. In:Technical Briefing; Conference of the Parties (COP1): Geneva, 7th February2006. Available at: <http://www.who.int/tobacco/fctc/TobReg_COP_Burns_2.pdf> (accessed 27.09.11).

Burns, D.M., Dybing, E., Gray, N., Hecht, S., Anderson, C., Sanner, T., O’Connor, R.,Djordjevic, M.V., Dresler, C., Hainaut, P., Jarvis, M., Opperhuizen, A., Straif, K.,2008. Mandated lowering of toxicants in cigarette smoke: a description of theWorld Health Organization TobReg proposal. Tob. Control 17, 132–141.

Byrd, G.D., Davis, R.A., Caldwell, W.S., Robinson, J.H., deBethizy, J.D., 1998. A furtherstudy of FTC yield and nicotine absorption in smokers. Psychopharmacology(Berl.) 139, 291–299.

Dalhamn, T., Edfors, M.L., Rylander, R., 1968. Retention of cigarette smokecomponents in human lungs. Arch. Environ. Health 17, 746–748.

Diding, N., 1987. Machine smoking results compared to human uptake of cigarettesmoke. Int. J. Clin. Pharmacol. Ther. Toxicol. 25, 143–147.

Federal Trade Commission, 2008. Rescission of FTC guidance concerning theCambridge Filter Method. Federal Register 73, 74500–74505.

Food and Drug Administration, 2011. Public Workshop: Scientific Evaluation ofModified Risk Tobacco Product (MRTP) Applications. August 25–26, 2011.Center for Tobacco Products; Silver Spring, MD 20993. Available at: <http://www.fda.gov/TobaccoProducts/NewsEvents/ucm259201.htm> (accessed27.09.2011).

Food and Drug Administration, 2012. FDA’s Guidance for Industry ‘‘ModifiedRisk Tobacco Products Applications’’ Request for Comments Docket No. FDA-2012-D-0071. Available at: <http://www.fda.gov/TobaccoProducts/GuidanceComplianceRegulatoryInformation/ucm297750.htm> (accessed 23.05.12).

Federal Register, 1967. Cigarettes: Testing for Tar and Nicotine Content. vol. 32, p.11178.

Federal Trade Commission, 2008. FTC Rescinds Guidance from 1966 on StatementsConcerning tar and Nicotine Yields. Available at: <http://www.ftc.gov/opa/2008/11/cigarettetesting.shtm> (accessed 10.10. 11).

Feng, S., Plunkett, S.E., Lam, K., Kapur, S., Muhammad, R., Jin, Y., Zimmermann, M.,Mendes, P., Kinser, R., Roethig, H.J., 2007. A new method for estimating theretention of selected smoke constituents in the respiratory tract of smokersduring cigarette smoking. Inhal. Toxicol. 19, 169–179.

Gori, G.B., Lynch, C.J., 1985. Analytical cigarette yields as predictors of smokebioavailability. Regul. Toxicol. Pharmacol. 5, 314–326.

Gori, G.B., Benowitz, N.L., Lynch, C.J., 1986. Mouth verses deep airwayabsorption of nicotine in cigarette smokers. Pharmacol. Biochem. Behav.25, 1181–1184.

Hammond, D., Fong, G.T., Cummings, K.M., O’Connor, R.J., Giovino, G.A., McNeill, A.,2006. Cigarette yields and human exposure: A comparison of alternative testingregimens. Cancer Epidemiol. Biomarkers Prev. 15, 1495-1451.

Hammond, D., Wiebel, F., Kozlowski, L.T., Borland, R., Cummings, K.M., O’Connor,R.J., et al., 2007. Revising the machine smoking regime for cigarette emissions:implications for tobacco control policy. Tob. Control 16, 8–14.

Hatsukami, D.K., Joseph, A.M., LeSage, M., Jensen, J., Murphy, S.E., Pentel, P.R.,Kotlyar, M., Borgida, E., Le, C., Hecht, S.S., 2007. Developing the science base forreducing tobacco harm. Nicotine Tob. Res. 9, 537–553.

Hukkanen, J., Jacob, P., Benowitz, N.L., 2005. Metabolism and disposition kinetics ofnicotine. Pharmacol. Rev. 57, 79–115.

Institute of Medicine, 2001. Clearing the Smoke: Assessing the Science Base forTobacco Harm Reduction. The National Academy Press, Washington, DC.

H.-Jörg Urban et al. / Regulatory Toxicology and Pharmacology 64 (2012) S85–S97 S97

Institute of Medicine, 2012. Scientific Standards for Studies on Modified RiskTobacco Products. The National Academy Press, Washington, DC.

International Organization for Standardization, 1991. Tobacco and TobaccoProducts – Atmosphere for Conditioning and Testing, ISO 3402, InternationalOrganization for Standardization, Geneva.

International Organization for Standardization, 1995. Cigarettes – Determination ofCarbon Monoxide in the Vapour phase of Cigarette Smoke – NDIR Method. ISO8454, 2nd ed. Geneva, International Organization for Standardization.

International Organization for Standardization, 2000a. Routine analytical cigarette-smoking machine - definitions and standard conditions. Standard 3308, fourthed. Geneva, International Organization for Standardization.

International Organization for Standardization, 2000b. Cigarettes – determinationof total and nicotine-free dry particulate matter using a routine analyticalsmoking machine. ISO Standard 4387, third ed. Geneva, InternationalOrganization of Standardization.

International Organization for Standardization, 2000c. Cigarettes –determination of nicotine in smoke condensates – gas chromatographicmethod. ISO Standard 10315, second ed. Geneva, International Organizationfor Standardization.

Jarvis, M.J., Boreham, R., Primatesta, P., Feyerabend, C., Bryant, A., 2001. Nicotineyield from machine-smoked cigarettes and nicotine intakes in smokers:Evidence from a representative population survey. J. Natl. Cancer Inst. 93,134–138.

Lindner, D., Smith, S., Martin Leroy, C., Tricker, A.R., 2011. Comparison of exposureto selected cigarette smoke constituents in adult smokers and non-smokers in aEuropean, multicenter, observational study. Cancer Epidemiol. Biomarkers Prev.20, 1524–1536.

Marian, C., O’Connor, R.J., Djordjevic, M.V., Rees, V.W., Hatsukami, D.K., Shields, P.G.,2009. Reconciling human smoking behavior and machine smoking patterns:Implications for understanding smoking behavior and the impact on laboratorystudies. Cancer Epidemiol. Biomarkers Prev. 18, 3305–3320.

Mendes, P., Liang, Q., Frost-Pineda, K., Munjal, S., Walk, R.A., Roethig, H.J., 2009. Therelationship between smoking machine derived yields and biomarkers ofexposure in adult cigarette smokers in the US. Regul. Toxicol. Pharmacol. 55,17–27.

Moldoveanu, S.C., St. Charles, F.K., 2006. Differences in the chemical composition ofthe particulate phase of inhaled and exhaled cigarette mainstream smoke. Beitr.Tabakforsch. Int. 22, 290–302.

Moldoveanu, S., Colmann, W., Wilkens, J., 2007. Determination of carbonylcompounds in exhaled cigarette smoke. Beitr. Tabakforsch. Int. 22, 346–357.

Moldoveanu, S., Colmann, W., Wilkens, J., 2008a. Determination of polycyclicaromatic hydrocarbons in exhaled cigarette smoke. Beitr. Tabakforsch. Int. 23,85–97.

Moldoveanu, S., Colmann, W., Wilkens, J., 2008b. Determination ofhydroxybenzenes in exhaled cigarette smoke. Beitr. Tabakforsch. Int. 23, 98–106.

Moldoveanu, S., Colmann, W., Wilkens, J., 2008c. Determination of benzene andtoluene in exhaled cigarette smoke. Beitr. Tabakforsch. Int. 23, 107–114.

Organization for Economic Co-operation and Development, 1997. Bacterial reversemutation test, guideline 471. In: OECD Guidelines for Testing of Chemicals,OECD, Paris.

Rickert, W.S., Robinson, J.C., 1981. Estimating the hazards of less hazardouscigarettes II: Study of cigarette yields of nicotine, carbon monoxide, and

hydrogen cyanide in relation to levels of cotinine, carboxyhemoglobin andthiocyanate in smokers. J. Toxicol. Environ. Health 7, 391–403.

Robinson, R.J., Yu, C.P., 2001. Deposition of cigarette smoke particles in the humanrespiratory tract. Aerosol Sci. Technol. 34, 202–215.

Rodgman, A., Perfetti, T.A., 2009. The chemical components of tobacco and tobaccosmoke. CRC Press, Taylor & Francis Group, Boca Raton, LA.

Russell, M.A.H., Jarvis, M.J., Iyer, R., Feyerabend, C., 1980. Relation of nicotine yield ofcigarettes to blood concentrations in smokers. BMJ 280, 972–976.

Schepers, G., Demetriou, D., 2002. Determination of nicotine and its metabolites inurine by HPLC after DETBA derivatization. In: Annual Society for Research onNicotine and Tobacco Conference, 20–23rd February 2002, Savannah, GA, US.

Scherer, G., Engl, J., Urban, M., Gilch, G., Janket, D., Riedel, K., 2007. Relationshipbetween machine-derived smoke yields and biomarkers in cigarette smokers inGermany. Regul. Toxicol. Pharmacol. 47, 171–183.

Schorp, M.K., Tricker, A.R., Dempsey, R., 2012. Reduced exposure evaluation of anElectrically Heated Cigarette Smoking System. Part 1: Non-clinical and clinicaltesting strategy. Regul. Toxicol. Pharmacol. 64 (Suppl. 1), S1–S10.

Seeman, J.I., Lipowicz, P.J., Piade, J.-J., Poget, L., Sanders, E.B., Snyder, J.P., Trowbridge,C.G., 2004. On the deposition of volatiles and semivolatiles from cigarettesmoke aerosols: relative rates of transfer of nicotine and ammonia fromparticles to the gas phase. Chem. Res. Toxicol. 17, 1020–1037.

Tricker, A.R., 2006. Biomarkers derived from nicotine and its metabolites: a review.Beitr. Tabakforsch. Int. 22, 147–175.

Tricker, A.R., Stewart, A.J., Martin Leroy, C., Lindner, D., Schorp, M.K., Dempsey, R.,2012. Reduced exposure evaluation of an Electrically Heated Cigarette SmokingSystem. Part 3: 8-day randomized clinical trial in the U.K. Regul. Toxicol.Pharmacol. 64 (Suppl. 1), S35–S44.

Ueda, K., Kawachi, I., Nakamura, M., Nogami, H., Shirokawa, N., Masui, S., Okayama,A., Oshima, A., 2002. Cigarette nicotine yields and nicotine intake amongJapanese male workers. Tob. Control 11, 55–60.

Urban, H.-J., Schorp, M.K., 2006. Nicotine bridging: a new method to extend smokeconstituent biomarker measurements from clinical studies to other cigarettemainstream smoke constituents. In: Society for Risk Analysis Conference, 3–6thDecember, 2006, Baltimore, MD, US.

Urban, H.-J., Gomm, W., Schorp, M.K., 2008. A modelling approach to developmachine smoking protocols reflecting human puffing behaviour forconventional cigarettes. Beitr. Tabakforsch. Int. 23, 8–18.

World Health Organization Study Group on Tobacco Product Regulation, 2004.Guiding Principles for the Development of Tobacco Product Research andTesting Capacity and Proposed Protocols for the Initiation of Tobacco ProductTesting. World Health Organization, Geneva.

World Health Organization Study Group on Tobacco Product Regulation, 2007. Thescientific basis of tobacco product regulation. WHO Technical Report Series 945,Geneva.

World Health Organization Study Group on Tobacco Product Regulation, 2008. Thescientific basis of tobacco product regulation. Second report of the WHO StudyGroup. WHO Technical Report Series 951, Geneva.

Zenzen, V., Diekmann, J., Gerstenberg, B., Weber, S., Wittke, S., Schorp, M.K., 2012.Reduced exposure evaluation of an Electrically Heated Cigarette SmokingSystem. Part 2: smoke chemistry and in vitro toxicological evaluation usingsmoking regimens reflecting human puffing behavior. Regul. Toxicol.Pharmacol. 64 (Suppl. 1), S11–S34.